Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
PREFACE
Assalamu’alaikum Wr. Wb.
The A ssociation o f G eotechnical S ocieties i n S outheast A sia ( AGSSEA) co nsists o f t he
Southeast A sian G eotechnical S ociety ( SEAGS), V ietnamese S ociety f or S oil M echanics an d
Geotechnical En gineering ( VSSMGE), H ong Kong Geotechnical S ociety ( HKGES),
Geotechnical Society of Singapore (GeoSS), Thai Geotechnical Society (TGS), C hinese Taipei
Geotechnical S ociety ( CTGS), Indonesian S ociety f or G eotechnical En gineering ( HATTI) an d
Malaysian G eotechnical S ociety ( MGS). It w as d ecided a t t he G eneral C ommittee m eeting o f
SEAGS and A GSSEA on 31 M ay 20 16 a t K uala L umpur t hat t he 20 th Southeast A sian
Geotechnical Conference and the 3rd AGSSEA Conference (20SEAGC-3AGSSEA Conference)
is t o b e h eld a t P ullman Hotel i n Jakarta, from 6 -7 November 2018 . This i s t he f irst t ime in
history of SEAGS that the conference is held at Jakarta, Indonesia, therefore the conference is
also co njunction w ith 22 nd Annual Indonesian N ational Confe rence on G eotechnical
Engineering.
Deep c ondolences for all v ictims d uring t he Lombok - Palu - Donggala ear thquakes.
Liquefaction, flows and deep slide shown as phenomenom during the disaster.
Special gratitude to Menard as main sponsored, Geotekindo and Bauer as co-sponsored and all
sponsored who participate in this conference.
i
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
“Geotechnical Challenge for Mega Infrastructure” become our theme to answer the challenge we
face together from current rapid infrastructure construction. Many geotechnical construction was
built in accordance with good soil mechanics and geotechnical understanding, but we need more
than that. We need all stages in construction industry can work together, from theory, practical
construction, de tailing w orks, c omprehensive s upervision, a nd of c ourse t he qua lity a ssurance
process. We need all parties to collaborate together, to move fast forward without no chance for
mistake nor fa ilure. No time for failure, ISGE/HATTI collaborate with all parties (national and
international) try to ensure there is only one way, and its forward moving.
We understand that there is still many holes in our c onstruction process, not only cause by man
made construction but also by the mother of nature. We saw how horrible the Donggala – Palu
earthquake l ast m onth an d w e m ourn f or t he cau sed; b ut as en gineer, we s tand; w e s eek t he
cause, we learn, we give suggestion and the more important, we stand still, not only as engineers
but also as a nation. We stand for the humanity-our pray for them.
Allow us to convey our gratitude to Menard, our m ain sponsored, Geotekindo and Bauer as our
co-sponsored a nd a ll s ponsored w ho p articipate i n t his event. A lso t o a ll s peakers, w riters,
participants, especially to all committee who works to their best for the success of this event.
Last b ut n ot l east, h opefully t his ev ent will bri ngs g ood a nd ne w poi nt of v iew s eeing t he
geotechnical engineering. Please enjoy the event.
ii
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ORGANIZING COMMITTEE
iii
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
iv
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
TABLE OF CONTENTS
Keynote Speakers :
4. Serious Lessons Learnt from Big Project (Ikuo Towhata, Professor Emeritus,
University of Tokyo, Visiting Professor, Kanto Gakuin University) ………………………… 19-24
5. Stability Analysis of Buried Waste Water Pipeline in Soft Ground (Prof. Eun
Chul Shin-Korea University - South Korea)………………...................................................................................... 25-28
6. New Solutions to Geotechnical Challenges for Coastal Cities (J. Chu1, S.F. Wu1,
H. Chen1, X.H. Pan1, and K. S-L. Chaim2 1School of Civil and Environmental Engineering, Nanyang
Technological University, Singapore, 2Geological and Underground Projects Department, Build
Control Authority, Singapore) …………………………………………………......................................….. 29-34
7. Use of Strut Free Systems in Deep Excavations (Chang-Yu Ou1 and Aswin Lim2)
1NationalTaiwan University of Science and Technology, Taipei, Taiwan, ROC 2Department of Civil
Engineering, Universitas Katolik Parahyangan, Bandung 40141, Indonesia ……..................………… 35-38
v
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
8. Behavior of Single Pile and Pile Group Foundation for High Rise Buildings
on Expansive Soils (Prof. Paulus P. Rahardjo-Universitas Katolik Parahyangan,
Indonesia)………….............…………………………............................................................................. 39-45
5. Study of Bored Pile Capacity in Klang Valley Residual Soil Based on Field-
Performance Data (Allan Chwee Yew Lun, Balakrishnan Etty Gaunder and Nazri
Ali).................................................................................................................. 96-101
8. Pile Settlement Uncertainty in Jakarta, Indonesia (Bondan Satria and Widjojo Adi
Prakoso)…….................................................................................................... 111-114
vi
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
11. ERT-Based Leakage Tracing for Dam Safety and its Potential Sliding
Surface (Helsin Wang, Chih-Hsin Hu, Sheng-Hsiung Hsieh and Yao-Chu Tsai)..............…... 125-135
12. Dynamic p-y Curves for a Single Pile by 1g Shaking Table Tests (Sangseom Jeong).. 136-140
13. Design Method for Bottom Single Blade Steel Rotation Pile Foundation :
Case study in Vietnam (Duy Lam Dao and Thi Tuyet Trinh Nguyen)……….............…. 141-146
14. Active Shaking Tests of Pile Foundation Models in Dry Sand Ground
(Kohei Kenda, Anh Tuan Vu and Tatsunori Matsumoto).................................................... 147-157
16. Volume Measurement for Heavy Punding Induced Craters and Ground
Heave (Amalia Ula Hazhiyah, Hung-Jiun Liao and Chih-Jung Chien) ............................... 168-178
19. Prediction of the Geological Condition for Pipe jacking Base on the Data
Collected in the Shafts (Jian Shou Keh and Kan Tang Fu) ……............................... 188-192
vii
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
23. Plane Strain Ratio and Waling Size Evaluation of Deep Excavation in
Kuala Lumpur Using 3D Finite Element Analysis (Jen Shen Ang, Bin Chen Hsiung
and Ching Hung).................................................................................................. 213-221
26. Dynamic Analysis of Underwater Tunnels (Akhila Babu, Anitha S D and Nimmy
Mariam Abraham)……………………...……………………………………..... 234-237
Session IV.B2 - Soil Mechanics & Lab. (TC 101 & TC 102)
31. Volcanic Cohesive Soil Behavior under Static and Cyclic Loading (Wa Ode
Sumartini, Hemanta Hazarika, Takaji Kokusho and Shinichiro Ishibashi)............................... 258-264
32. Screw Driving Sounding Test for Soil Identification and Classification
(Aminaton Marto, Go Sakai, Naoaki Suemasa, Nor Zurairahetty Mohd Yunus, Siti Norafida Jusoh,
Nadiah Jamaludin, Muhammad Mustakim Ponimin and Muhammad Fakrulnizam Mohd
Tahir)............................................................................................................... 265-270
33. Argillaceous Rock Properties Changes Due to the Weathering Process (Idrus
M. Alatas, Masyhur Irsyam, Ramli Nazir and Pintor T. Simatupang) ..................................... 271-277
viii
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
35. Lessons Learned from Pressuremeter Tests on Stone Columns (Richard Ong). 284-287
44. Field Identification of Active Fault Nearby the Footprint of the Dam (Didiek
Djarwadi, D.H. Natawidjaja and M.R. Daryono)............................................................. 329-333
ix
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
47. Using Time Domain Reflectometry for Monitoring Slope Movement in the
Jiufenershan Landslide (Sheichen Ho, Ihui Chen, Yushu Lin, Junyang Chen and Miaubin Su)
....................................................................................................................... 341-345
50. Numerical Simulation of Some Debris Flow Events in Central Java for
Predicting Run-out Distributions (Imam Achmad Sadisun, Rendy Dwi Kartiko and Indra
Andra Dinata) ..................................................................................................... 357-360
54. Effect of the Initial Suction Boundary on the Slope Failure of Volcanic
Residual Soil (Agus Setyo Muntohar) ................................................................... 375-378
59. Performance of Helix Piled Raft in Tropical Fibrous Peat Soil under Traffic
Loads (Ardy Arsyad, Ahmad Bakri Muhiddin and Lawalenna Samang) ............................... 398-401
x
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
60. Effect of Compaction on Liquefaction of River Sand and Sea Sand in Hai
Phong City, Vietnam (Chau Lan Nguyen, Quang Phuc Nguyen, Hai Ha Nguyen and Duc
Manh Nguyen) .................................................................................................... 402-408
64. The use of the Observational Method in the Deep Excavations for the
Realization of a Residential Compound (Marco Carassini, Filippo Bucci and Andrea
Antiga) ............................................................................................................. 425-432
70. Basement Excavation in Soft Marine Clay in Bukit Tinggi, Klang (Sharmeelee
S.) …............................................................................................................... 461-464
xi
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
73. Study for Structural Performance of Steel Sheet Piles Used for Retaining
Wall (Eka Susanto, Matsui Nobuyuki and Otsushi Kazutaka) …....................................... 475-481
75. Bored Pile Retaining Wall Solutions for Earthquake Slip 6 AT Ohau Point,
Kaikoura, New Zealand (Senthuran Arulanantham, Anthony Fairclough, Sam Glue and Jody
Edwards)…………..........……………………………..………..................…. 487-493
76. Effects of Cross Wall to the System Stiffness of Deep Excavations in Clay
(Zih-Yun Wang, Bin-Chen Benson Hsiung, Hsii-Sheng Hsieh and Louis Ge)…...................…. 494-499
79. Aplikasi Metode Elemen Hingga dan Kriteria Runtuh Papaliangas Batuan
Berkekar : Problem-Problem Analisa Terowongan, Kestabilan Lereng dan
Wellbore Stability dalam Mekanika Batuan (Wilham George Louhenapessy)
....................................................................................................................... 509-514
80. The Application of Active Lime & Cement as Soil Stabilization Material
(Renggo Ginanjar, Wilham G. Louhenapessy, Asrul Ahdar) ….......................................... 515-518
82. On the Weak Limestone Rock Slope Stability Analysis (Wilham George
Louhenapessy) ….............................................................................................. 523-530
84. Perancangan Proteksi Galian Dalam Dengan Soldier Pile. Studi Kasus :
Underpass Pada Pusat Perbelanjaan di Kawasan Jakarta Selatan (Siska Rustiani
and Ryan Alexander Lyman) …................................................................................ 535-538
xii
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
85. Study of Slope Failure and Reinforcement Analysis to Restore and Increase
the Slope Stability in Flores, East Nusa Tenggara Using Plaxis 2D and 3D
(Rendra Priatno1, Ali Iskandar2, and Jo Lian Huat3) ……………………...............................……….. 539-543
87. Pengaruh Tahanan Ujung Tiang dan Faktor Aman Terhadap Lendutan Pelat
Terpaku Tiang Tunggal Pada Lempung Lunak (P. Anas1, O. Ferly2, M. Roza3)
………………………………………………………………………………………………..............………….. 549-543
89. Design of Simple Drapery Systems for Rock Cuts and Natural Slopes (Matteo
Lelli1, Alberto Grimod2, Riccardo Laneri1, Deny1, Vincent Setiawan2, Dinda Mutiara Savitri1)
………………………………………………………………………………………………….....................….. 561-568
93. Estimasi Modulus Resilien dari uji CBR pada Tanah Lempung distabilisasi
Abu Vulkanis dan Kapur (Devi Oktaviana Latif - Dr.,ST.,M.Eng, Ahmad Rifa’i- Dr., Ir.,MT.,
Latif Budi Suparma –Ir.,M.Sc., Ph.D) ………………...................................…………………………….. 592-596
95. Efek EPS Geofoam sebagai Material Pengisi terhadap Nilai CBR
Laboratorium pada Tanah Kohesif Darah Bandung (Lestari, A. S 1 and Julian
Clementio 2 - 1 Dosen Universitas Katolik Parahyangan, Bandung, Indonesia, 2 Mahasiswa
Universitas Katolik Parahyangan, Bandung, Indonesia) ……………………...................………………. 603-606
96. Prediction of Excess Pore Pressure Due to Pile Driving Based on CPTu (A.
Arafianto1 and P.P. Rahardjo2, 1Engineering Faculty, Universitas Katolik Parahyangan Bandung,
Indonesia 2Engineering Faculty, Universitas Katolik Parahyangan, Bandung, Indonesia)................ 607-610
xiii
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
xiv
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: For environmental protection and sustainability, the use of plants and recycled wastes have been investigated in a landfill site
located at Xiaping, Shenzhen under humid climates. The main objective was to validate a novel vegetated three-layer landfill cover system
using recycled construction waste without the need of geomembrane in the field. One section was transplanted with Bermuda grass while the
other section was left bare. To assess the landfill cover performance, the test site was heavily instrumented and monitored for a period of 13
months under natural climatic conditions. The cumulative rainfall depth was about 2950 mm over the whole monitoring period. At the end of
monitoring period, the measured cumulative percolation was only 27 mm and 20 mm at the bottom of the bare and grassed cover,
respectively. It is evident that the vegetated three-layer landfill cover system using recycled concrete without geomembrane can be effective
in minimizing percolation at humid climates.
Keywords: three-layer landfill cover, recycled concrete, vegetation, field monitoring, water infiltration.
1
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
2
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3. CONCLUSIONS
A newly proposed three-layer landfill cover system was constructed
and tested in Shenzhen, China. Based on field monitoring of the
full-scale test over a period of 13 months (i.e., June 2016 to July
2017), the following conclusions may be drawn:
a) Under natural weather variations, pore water pressure at
Figure 4 Variations of pore water pressure at different depths in the shallow depth (i.e., 0.2 m) was mostly affected by rainfall
grass covered three-layer landfill cover system events. However, negative pore water pressure was
retained at all depths (i.e., 0.2m, 0.4m, 0.8 m, 1.2 m and
1.6 m) in the cover after heavy rainfall events.
5.2 Cumulative percolation measured by lysimeters b) At the end of 13-month monitoring period, the measured
cumulative rainfall was 2,950 mm, whereas the
Figure 5 shows the measured cumulative percolation in the bare and corresponding amount of percolation was 27 mm and 20
grass covered three-layer landfill cover from June 2016 to July mm for the bare and grass covered landfill cover,
2017. Cumulative percolation at three different locations (i.e., crest, respectively. The measured percolation in the new cover
middle, toe) in both slopes are included. For ease of comparison, the system clearly meets the recommended criterion of 30
measured cumulative rainfall depth with a total amount of 2,950 mm mm/year for conventional compacted clay covers in the
is also provided in the figure. During the first 6 months of US.
monitoring (i.e., June 2016 to November 2016), measured c) The results of the field monitoring validated the potential
percolation increased at a relatively steady rate and showed little use of a grassed three-layer landfill cover system using
variation in response to daily rainfall events, even those in excess of recycled concrete without geomembrane as a promising
200 mm (i.e., 19 October 2016). This indicates the effectiveness of alternative landfill cover system for humid climates.
the three-layer landfill cover system in preventing excessive
percolation through the cover. No sign of preferential flow was 4. ACKNOWLEDGEMENT
observed through both the bare and grass covered landfill covers. The authors would like to acknowledge the grant 51778166
However, following a long drying period (i.e. December 2016 to provided by the National Natural Science Foundation of China.
May 2017), percolation increased steadily by about 10 mm for both
landfill covers. This may be due to some desiccation cracks which
were observed on the soil surface at the end of this long drying 5. REFERENCES
period. Similar findings were also described by Albright et al.
Albright, W. H., Benson, C. H. and Gee, G. W. (2004). “Field water
(2006). At the end of 13-month monitoring period, the maximum balance of landfill covers.” Journal of Environmental Quality,
percolation measured for the bare and grass covered landfill cover, 33, Issue 6, pp2317-2332.
was 27 mm and 20 mm, respectively. Both landfill covers meet the Albright, W. H., Benson, C. H., Gee, G. W., Abichou, T.,
recommended design criterion of 30 mm/year for compacted clays McDonald, E. V., Tyler, S. W., and Rock, S. A. (2006).
(Benson et al. 2001). It is clear that the newly proposed three-layer “Field performance of a compacted clay landfill cover at a
landfill cover system using recycled concrete without geomembrane humid site.” Journal of Geotechnical and Geoenvironmental
can be effective in minimizing percolation at humid climates. Engineering, 132, Issue 11, pp1393-1403.
3
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Amaya, P., Queen, B., Stark, T. D., and Choi, H. (2006). “Case McCartney, J. S., and Zornberg, J. G. (2010). “Effect of infiltration
history of liner veneer instability.” Geosynthetics and evaporation on geosynthetic capillary barrier
International, 13, Issue 1, pp36-46. performance.” Canadian Geotechnical Journal, 47, Issue 11,
Benson, C., Abichou, T., Albright, W., Gee, G., and Roesler, A. pp1201-1213.
(2001). “Field evaluation of alternative earthen final covers.” Morris, C. E., Stormont, J. C. (1999). “Parametric study of
International Journal of Phytoremediation, 3, Issue 1, pp105- unsaturated drainage layers in a capillary barrier.” Journal of
127. Geotechnical and Geoenvironmental Engineering, 125, Issue
Benson, C., and Khire, M. (1995). “Earthen final covers for landfills 12, pp1057 – 1065.
in semi-arid and arid climates.” Landfill closures, GSP No. Ng, C. W. W., and Menzies, B. (2007). Advanced unsaturated soil
53, R. Dunn and U. Singh, eds., ASCE, Reston, Va., 201-218. mechanics and engineering. Taylor & Francis, London and
Bossé, B., Bussiere, B., Hakkou, R., Maqsoud, A., and Benzaazoua, NY.
M. (2015). “Field experimental cells to assess Ng, C. W. W., Chen, R., Coo, J. L., Liu, J., Ni, J. J., Chen, Y. M. ,
hydrogeological behavior of store-and release covers made Zhan, T .L. T., Guo, H. W., and Lu, B. W. (2018a). “A novel
with phosphate mine waste.” Canadian Geotechnical Journal, vegetated three-layer landfill cover system using recycled
52, Issue 9, pp1255-1269. construction wastes without geomembrane.” Canadian
Bouazza, A., Zornberg, J. G., McCartney, J. S., and Nahlawi, H. Geotechnical Journal. Under review.
(2006). “Significance of unsaturated behaviour of geotextiles Ng, C. W. W., Coo, J. L., Chen, Z. K., and Chen, R. (2016). “Water
in earthen structures.” Australian Geomechanics Journal, 41, infiltration into a new three-layer landfill cover system.”
Issue 3, pp133-142. Journal of Environmental Engineering, ASCE, 142, Issue 5.
Daniel, D. E. (1994). “Surface barriers: problems, solutions and doi:10.1061/(ASCE)EE.1943-7870.0001074.
future need.” Proceedings of the 33rd Hartford Symposium on Ng, C. W. W., Leung, A. K., and Ni. J. J. (2018b). Plant-Soil slope
Health and Environment, Pasco, WA, pp441-497. interaction. Taylor & Francis. In Press
European Environment Agency (EEA). (2013). Managing municipal Ng, C. W. W., Liu, J., and Chen, R. (2015b). “Numerical
solid waste – a review of achievements in 32 European investigation on gas emission from three landfill soil covers
countries, Environmental Assessment Report No. 2, under dry weather conditions.” Vadose Zone Journal, 14,
Copenhagen. Issue 8. doi:10.2136/vzj2014.12.0180.
Garg, A., Coo, J. L., and Ng, C. W. W. (2015). “Field study on Ng, C. W. W., Liu, J., Chen, R., and Xu, J. (2015a). “Physical and
influence of root characteristics on soil suction distribution in numerical modeling of an inclined three-layer (silt/gravelly
slopes vegetated with Cynodon dactylon and Schefflera sand/clay) capillary barrier cover system under extreme
heptaphylla.” Earth Surface Processes and Landforms, 40, rainfall.” Waste Management 38, pp210-221.
Issue 12, pp1631-1643. Rahardjo, H., Santoso, V. A., Leong, E. C., Ng, Y. S., and Hua, C. J.
Iryo, T., and Rowe, R. K. (2005). “Hydraulic behaviour of soil- (2012). “Performance of an instrumented slope covered by a
geocomposite layers in slopes.” Geosynthetics International, capillary barrier system.” Journal of Geotechnical and
12, Issue 3, pp145-155. Geoenvironmental Engineering, 138, Issue 4, pp481-490.
Khire, M. V., Benson, C. H., and Bosscher, P. J. (1999). “Field data Rahardjo, H., Satyanaga, A., Leong, E. C., Santoso, V. A., and Ng,
from a capillary barrier and model predictions with UNSAT- Y. S. (2014). “Performance of an instrumented slope covered
H.” Journal of Geotechnical and Geoenvironmental with shrubs and deep-rooted grass.” Soils and Foundations,
Engineering, 125, Issue 6, pp518-527. 54, Issue 3, pp417-425.
Khire, M. V., Benson, C. H., and Bosscher, P. J. (2000). “Capillary Rahardjo, H., Tami, D. and Leong, E. C. (2006). “Effectiveness of
barriers: design variables and water balance.” Journal of sloping capillary barriers under high precipitation rates.”
Geotechnical and Geoenvironmental Engineering, 126, Issue Proceedings of the 2nd International Conference on
8, pp695-708. problematic soils, Petaling Jaya, Selangor, Malaysia, pp39-
Knidiri, J., Bussière, B., Hakkou, R., Bossé, B., Maqsoud, A., and 54.
Benzaazoua, M. (2017). “Hydrogeological behaviour of an Ross, B. (1990). “The diversion capacity of capillary barriers.”
inclined store-and-release cover experimental cell made with Water Resources Research, 26, Issue 10, pp2625-2629.
phosphate mine wastes.” Canadian Geotechnical Journal, 54, Siemens, G., and Bathurst, R. J. (2010). “Numerical parametric
Issue 1, pp102-116. investigation of infiltration in one-dimensional sand-
Koerner, R. M. and Daniel, D. E. (1997). Final Covers for Solid geotextile columns.” Geotextiles and Geomembranes, 28
Waste Landfills and Abandoned Dumps, American Society of Issue 5, pp420-474.
Civil Engineers, Reston, VA. U.S. Environmental Protection Agency (EPA). (2015). Advancing
Leung, A. K. (2016). “Grass evapotranspiration-induced suction in sustainable materials management: facts and figures 2013,
slope: case study.” Environmental Geotechnics, 3, Issue 3, Report EPA530-R-15-002, Washington DC.
pp155-165. Zornberg, J. G., and McCartney, J. S. (2003). Analysis of
Lim, T. T., Rahardjo, H., Chang, M. F., and Fredlund, D. G. (1996). Monitoring Data from the Evapotranspirative Test Covers at
“Effect of rainfall on matric suctions in a residual soil slope.” the Rocky Mountain Arsenal, Geotechnical Report, US
Canadian Geotechnical Journal, 33, Issue 4, pp618-628. Environmental Protection Agency, Region 8, December.
Zornberg, J. G., Bouazza, A., and McCartney, J. S. (2010).
“Geosynthetic capillary barriers: current state of knowledge.”
Geosynthetics International, 17, Issue 5, pp273-300.
4
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
1. INTRODUCTION The National Center for Earthquake Studies (PuSGeN) and the
Indonesia Association of Geotechnical Engineering (HATTI)
The Palu earthquake occurred on Friday afternoon, 28 September deployed a team to conduct field survey assessing impact of Palu
2018, at 18:02:44 local time (Central Indonesia Time, WITA) with Earthquake, from 8 to 15 October 2018. This paper presents findings
moment magnitude (Mw) of 7.4, centered 26 km north of Donggala, of field surveys with emphasis on geotechnical impact and fault
Central Sulawesi (Figure 1). The earthquake has caused strong rupture offset observed in the field.
shaking, resulted in a tsunami that hit Palu city and disastrous
liquefaction occurrences, particularly in Petobo and Balaroa areas.
2. SEISMOTECTONIC OF SULAWESI
Figure 2 shows the intensity map by the Meteorological,
Climatological and Geophysical Agency (BMKG). BMKG Sulawesi is located in the Sunda block adjacent with three plates;
automated modeling indicates intensity of VI-VIII in Palu city and the Australia plates, Philippines plate and Pacific plate, known as
Donggala regency, Central Sulawesi. Two-day following the triple junction. This triple junction results in a complex tectonic in
earthquake, BMKG updated the intensity to IX-X (i.e. extensive Sulawesi, accommodated by strike slip faulting and thrust faulting.
damage) based on modeling, data instrument, and macro-seismic In 2017, PuSGeN considered 50 active-fault segments in Sulawesi
survey of 30 September 2018. The Palu – Koro fault zone is and the North Sulawesi Megathrust in the north of Sulawesi (Figure
believed to be the source of the Palu earthquake 2018 (Mw 7.4), as 3). Similar tectonic condition in Sulawesi was also previously
observed from its epicenter location and hypocenter depth, and reported by Irsyam, et. al., 2010.
aftershock distribution (Figure 1). The seismicity in Sulawesi is influenced by the subduction along
the North Sulawesi Megathrust (Figure 3). The WBZ from the slab
in North Sulawesi extends to a depth of about 180 km. The major
structures in Central Sulawesi is the Palu – Koro fault system, which
extends NNW-SSE direction and cuts Sulawesi apart more than 300
km, from the North Sulawesi trench pass through Palu bay,
southward turn to the SE connect to the Matano and Lawanopo
faults and further eastward, both faults join to Tolo trench.
Sulawesi moves north-west, suggesting a strong impact from the
Australia plate (Figure 4). Velocities in the northern and southern
part of Palu – Koro show different magnitude, indicating the activity
of Palu – Koro fault zone. To the west of northeast Sulawesi, the
island moves westward, indicating an impact from the Pacific plate.
The Palu – Koro fault is a sinistral strike-slip fault (horizontal fault
striking) in a north-south direction, crossing Palu city. The
maximum slip-rate of Palu – Koro fault is 58 mm/year based on
geological study by interpretation of geological offset, such as river
offset (Daryono, 2016). Recent GPS study also suggests that the
slip-rate of Palu – Koro fault is about 40 mm/year.
Central Sulawesi region is tectonically active as shown in Figure
Figure 1 Epicenters of the Palu earthquake 2018 (Mw7.4) 5. Historical destructive earthquakes along Palu – Koro fault zone
and the aftershock distributions (red dots) along Palu – Koro fault occurred in 1907, 1909, 1937 dan 2012. Paleoseismologi study has
in Central Sulawesi (PuSGeN, 2018) been conducted by Daryono (2016) and obtained that previous
earthquakes occurred in 1909, 1468 and 1338. Abendanon (1917)
This natural disaster caused many losses of lives and other concluded that the 1907 earthquake was followed by a more
material damages. The National Agency for Disaster Management destructive earthquake 2 years later in 1909. Houses that survived
(BNPB) on 9 October 2018 reported that this earthquake has caused during the 1907 earthquake were mostly destroyed by the 1909
2.037 casualties, 671 missing people, and 152 buried bodies. Total earthquake. The damages run along Saluki up to Donggala region.
of 4.084 people are heavily injured, 74.044 evacuees, and 67.310 He reported a large cracking of 7 km with uplift of 1.0 m. Trenching
houses are estimated to be damaged. Although the most damaged in Onu village on Saluki segment showed evidence of sinistral slip
area is in Palu city, these reported numbers are also compiled from of 1.5 m and vertical slip of 1.5 m. Daryono (2016) suggested a
other areas, such as Donggala, Sigi, Parigi Mountong, and plausible recurrence interval of 130 years in Palu – Koro. During
Pasangkayu, West Sulawesi. 2017, two main earthquakes were recorded along the fault.
5
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 2 Location of the main shock of the earthquake on 28 September 2018 and its shakemap (BMKG, 2018). Left is the automatic
shakemap first generated by BMKG on 28 September 2018, Right is the corrected shakemap based on modeling, data instrument and
macroseismic survey as of 30 September 2018
Figure 3 Active fault in Sulawesi as in the National Earthquake Source and Hazard Map launched in 2017
(Irsyam, et al., 2017)
6
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 4 Left: Seismicity of the Sulawesi for Mw ≥ 4,5, right: tectonic and deformation pattern derived from GPS data in Sulawesi
(Irsyam, et.al., 2017)
Figure 5 Historical Earthquake in Central Sulawesi (Daryono, 2016). Focal mechanism from CMT catalogue
3. GEOLOGY SETTING AND SUBSURFACE The morphology of Palu area consists of plains, denudation, and
CHARACTERISTIC hills. Plains area is irregular and weak topography. This area is
formed by fluvial sedimentation. Based on regional geology, Palu
The geology of Palu area is shown in Figure 6. Evolution of area consists of alluvium deposits and beach deposit (Qap) ) that are
Neogene kinematics along the Palu – Koro fault was confirmed Holocene. The unit is composed of gravel, sand, mud, and coral
based on microtectonics approach, i.e., sinistral strike-slip due to limestone.
east-west compression, radial extensions caused by telescoping This unit is formed in shallow river, delta, and sea environments.
vertical movement of Neogene granitoid, and then left lateral with As this unit is a young sediment, the arranged layer has not
normal component displacement due to north-south extension/east- undergone the whole process of lithification or cementation process
west compression which actually is still active. Palu depression area (Soekamto, 1995).
is filled by mostly clay, silt, and sand deposits as alluvial deposit.
The composition of gravel is granitic fragment to the northwest.
West escarpment to the north consists of granite and granodiorite
units, and to the south consists of schist-phyllitic units. East
escarpment consists of molasses (Pramuwijoyo et al., 1997).
7
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
8
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 9 Ground motion and response spectra recorded in Poso (PMCI) and in Sabang (MPSI) by BMKG
9
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
7.2.1. Balaroa
Soil failure in Balaroa area is located about 1 km to the west of Palu
– Koro fault. The crown of soil movement is located in Jalan
Gunung Gawalise, whereas the toe is located in Balaroa village. The
affected area is mainly residential housing comprising of 34.5 Ha
with perimeter of 2.5 km. Aerial view of damages in Balaroa is
presented in Figure 13. Field observation at Jalan Gunung Gawalise
where the crown is located showed that slope failure may also be
Figure 12 Longitudinal Cracks on the Palu airstrip attributed to rotational land slide.
10
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
North
Toe
Damaged road at
Jl. Gunung Gawalise
Crown
7.2.2. Petobo fault. Although this area has slight inclined ground surface, the
Petobo, a populated residential area was severely damaged due to occurrence of mass movement to downslope was triggered by
liquefaction. This area is about 7.2 km to the east of Palu – Koro liquefaction (Figure 14).
11
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
7.2.3. Jono Oge residential housing, and roadway. Practically, damaged roadway
area turned into a large corn farm which initially is located nearby
Massive soil failure in Jono Oge village is located about 6.7 km to the crown. It should be noted that the roadway leading to this
the east of Palu – Koro fault. The crown of soil movement is located damaged roadway section and the surrounding area are also heavily
in the east nearby a water channel, whereas the toe is located in the affected by the earthquake. Damages in a gas station located about 1
west (Figure 16). In general, north boundary of the affected area is km north to the damaged roadway section is shown in Figure 17.
bounded by a river. The affected area is estimated about 185 Ha
with a perimeter of 8 km and is mainly comprised of corn farm,
North
Damaged road at
Jl. Poros Palu - Palolo
3115 m
Crown
Toe
800 m
Soil movement
direction
12
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
7.2.4. Sibaya Selatan located in the west (Figure 18). The affected area is estimated about
29 Ha with a perimeter of 2.2 km and is mainly comprised of paddy
Soil failure in Sibalaya Selatan village is located about 3.6 km to the farm, residential housing, and roadway. Figure 19 shows the impact
east of Palu – Koro fault. The crown of soil movement in this of Palu Earthquake 2018 of translated housing and football field in
location is located in the east nearby a sluice gate, whereas the toe is Sibalaya Selatan village.
North
Damaged road at
Toe Jl. Poros Palu - Kulawi
450 m
840 m
Crown
Figure 19 Translated housing and football field in Sibalaya Selatan after Palu Earthquake 2018
8. CONCLUSION 9. REFERENCES
PUSGEN and HATTI have conducted a field survey to preliminary Daryono, M.R., 2016. Paleoseismology Tropis of Indonesia (Stdu
investigate the earthquake rupture offset and geotechnical impact of case: Sumatera Fault, Palukoro-Matano Fault, and Lembang
the Palu Earthquake 2018. The rupture offset was found as large as Fault). PhD Dissertation Institut Teknologi Bandung.
4-6 meters sinistral offset and about half meter vertical offset. The Irsyam, M., Sengara, I.W., Asrurifak, M., Ridwan, M., Aldiamar,
geotechnical aspects observed after the earthquake are mainly F., Widiyantoro, S., Triyoso, W., Natawijaya, D.H., Kertapati,
related to liquefaction and lateral spreading. In Palu earthquake, E., Meilano, I., and Suhardjono (2010), Summary:
liquefaction resulted in lateral spreading where large horizontal Development of Seismic Hazard Maps of Indonesia for
surface deformation is experienced by large ground blocks above Revision of Seismic Hazard Map in SNI 03-1726-2002,
the liquefied layer. This lateral spreading was observed in relatively reseach report submited to the Ministry of Public Works by
gentle slope surface. Team for Revision of Seismic Hazard Maps of Indonesia, July.
13
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Irsyam, M., Widiyantoro, S., Natawidjaja, D.H., Meilano, I., Pramumijoyo, S., Indarto, S., Widiwijayanti, C. and
Rudyanto, A., Hidayati, S., Triyoso, W., Hanifa, N.R., Sopaheluwakan, J., 1997, Seismic Parameters of the Palu-Koro
Djarwadi, D., Faizal, L., Sunarjito (editor), National Team for Fault in Palu Depression Area, Central Sulawesi. Indonesia,
updating of Indonesia Earthquake Hazard Map 2017 (2017), Journal of South East Asian Earth Science,
“Earthquake source and hazard Map of Indonesia 2017”, Soekamto, R.A.B., 1995, Regional Geological Map of Palu Sheet,
Indonesia National center for earthquake Studies, Research and Indonesia, Scale 1:250,000, Geological Research Center,
Development Agency of Ministry of Public Work and Housing. Bandung.
ISBN 978-602-5489-01-3. Thein, P.Y., Pramumijoyo, S., Brotopuspito, K.S., Kiyono, J.,
PuSGeN (Indonesia National center for earthquake Studies) (2018), Wilopo, Furukawa, A., and Setianto, A., 2014, Estimation of
“Quick Report on Palu Earthquake 2018”, Indonesia National Seismic Ground Motion Induced by the 23 January, 2005
center for earthquake Studies, Research and Development Earthquake in Palu Region, Central Sulawesi, Indonesia.
Agency of Ministry of Public Work and Housing. Journal of Geological Resource and Engineering 4, 200-207.
14
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: During the last 20 years, many high-rise buildings supported by pile foundation are rising up in Astana, the new capital of
Kazakhstan. Following megaprojects are already completed: Ministry of Transportation and Communication, Expo-2017, Housing estate –
Izumrudny Kvartal (Emerald Square), Cultural and Entertainment Center – Khan Shatyr and so on. Many megaprojects are under
construction or in planning. One of the unique projects is the housing estate “Abu-Dhabi Plaza” which started on 2011 in Astana.
The article presents loading tests of large diameter and deep boring piles on the construction site Expo-2017 in new capital city of the
Republic of Kazakhstan. Results of static pile tests using the static compression loading test (by ASTM – SCLT), static loading test (by
GOST – SLT) and bi-direction static loading test (by ASTM - BDSLT or O-cell) methods are presented in this paper. Hereafter the results of
underground testing by the piles with the methods of vertical static testing are presented, which had been made on Expo 2017 projects,
buildings of Pavilion in Astana, Kazakhstan.
Keywords: pile, SCLT, SLT, O-cell or BDSLT, load, settlement.
15
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
16
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
this object. Four bored piles were subjected to static tests (O-cell holding time of intermediate loading stages was 30 minutes,
testing- 2 piles and SCLT- 1 pile and SLT by GOST-1 pile). unloading - 20 minutes. The time for maintaining peak loads was
The target of this tests was obtaining of bearing capacity of piles on 120 and 240 minutes, respectively (Zhussupbekov A.Zh., Lukpanov
problematical soils ground of Expo 2017 (Zhussupbekov and R.E., Omarov A.R. (2016)).
Omarov A.R. (2016))
17
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3. CONCLUSION
The overlay of the curves showed that the convergence of the graphs
is observed only at the initial stage of loading, then a change in the
trajectory of the SLT curve, characteristic of the creeping stage of
soil resistance, is observed, whereas the O-cell curve (at this stage of
loading) is more characteristic of the elastic resistance of the soil.
According to the results of the SCLT unloading curve, elastic work
of the soil is still evident. The reason for the abrupt change in the
trajectory of the SCLT curve, which is not characteristic of the
elastic work of the ground, is the holding time of the loading stages
(lower compared to the O-cell test method), which can also explain
the almost completely elastic work of the soil during O-cell tests.
When testing piles using the SLT method "from top to bottom", a
design load of 6000 kN corresponds to a draft of 2.09 mm, a
maximum test load of 12000 kN is a draft of 10.51 mm. It should be
noted that even with the maximum test load, only the elastic
operation of the pile in the ground is manifested, as evidenced by a
slight residual soil sediment after unloading, which is 1.4 mm.
When testing piles using the O-cell test, a maximum test load of
29000 kN corresponds to a draft of 18.35 mm (for the PTP-1 pile)
and - 14.40 mm (for the PTP-2 pile). During the testing of the piles,
both elastic and plastic deformation of the soil was observed, due to
a greater test load on the pile than in the SLT method.
4. REFERENCES
Zhussupbekov A.Zh., Omarov A.R., Zhukenova G.A. (2017). The
experience in applying of static load and O-cell pile testing
geotechnologies in problematical soil conditions of Astana.
International Mini symposium Chubu (IMS-CHUBU)-
Nagoya, Japan. Japanese Geotechnical Society Special
Publication. Vol. 5 No. 2 pp 167-171.
Zhussupbekov A.Zh., Omarov A.R. (2016). Geotechnical and
construction considerations of pile foundations in
problematical soils. Proceedings of the 8th Asian Young
Geotechnical Engineers Conference (8 AYGEC). Astana,
Kazakhstan. pp 27-32.
18
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Big project is an attractive idea for ambitious engineers. The present level of construction technology allows such big projects
to be carried out and our world may be drastically changed. The problem is that we still do not know how nature would respond to such a
significant human action. We need to know bad consequence was sometimes the case in the past. Although it is difficult, we need to have
some provision by which the worst and catastrophic consequence is avoided by stopping or cancelling the project.
Keywords: Project, nature, disaster, earthquake, slope failure
1. INTRODUCTION The problem was the high rate of water evaporation from the
th lake surface. Together with the significant reduction of water supply
The 20 Century enjoyed a significant development of construction from its two major rivers, Aral Sea started to shrink in 60s.
technology. The source of power changed from human and animal Nowadays, Aral Sea, which used to be the 4th biggest lake in the
muscle in the former centuries to mechanical engines that utilize world, is disappearing. Because the lake water was salty, the deposit
coal and oil. This change gave human an overwhelming power of salt on the dried lake bed is blown by wind and is causing health
above nature. It thus became possible to construct huge canals, man- problem among local residents (Whish-Wilson, 2002). It is not
made islands, bridges etc. possible anymore to turn reverse the clock because cotton
The situation until 1970s was full of optimism. Behind such a agriculture has been well established in the region and the
situation, however, many negative aspects of the overwhelming immigrant farmers cannot go back to Russia because it is a foreign
human power were emerging gradually. Their consequence became country after disappearance of USSR.
evident after 1970s. Consequently, many people such as naturalists There have been several efforts to save the lake from
started to insist on the need to stop big construction projects and disappearance. One of them was the idea to transport water from
protect nature. It is true that human has to be aware of the huge rivers in Siberia to the Aral Sea (Borovskiy, 1980; Aladin and
importance of conservation of nature and control human power not Potts, 1992; Elhance, 1997). Although the idea looked attractive to
to seriously damage the global natural system. On the other hand, some extent, the obvious problems were the difficulty to pump up
the world population is increasing rapidly towards 10 billion. water across high mountains on the way. Later, the environmental
Increased population needs food and drinking water. People desire issues became more important such as the effects of less water flow
healthful life, comfortable living conditions, and safety of life from into the Arctic Sea, and introducing different animal species from
natural disasters. To cope with this new situation, human has to keep those rivers into the Aral Sea. Finally, Siberia and Aral Sea area
seeking for better and wiser use of the natural system. Thus, big became different countries after disappearance of USSR and such a
construction projects will be necessary as well from now on. project became impossible.
However, it is important for human to keep in mind what mistakes The Aral Sea project had both positive and negative aspects. The
happened during the past big projects and try not to repeat them. positive one is the development of agriculture and the negative one
The present paper attempts to provide some information in this is the environmental damages which are disappearance of the Sea
regard. and salt-induced health problems. Kazakhstan constructed a huge
dam in the northern part of the Sea and produced fresh-water Little
2. ARAL SEA Aral Sea (Figure. 2). This may be a solution of the problem to some
The Aral Sea region of the Central Asia is wide in space and enjoys extent but not complete or final. Note that a similar problem is going
ample sunshine. It was supposed to be a wealthy agricultural zone if on in other places of the world. The Urmia Lake in Iran is shrinking
ample water is provided. It was decided in early 1960s to construct quickly as well because of the water of tributary rivers is sent to
big canals and long irrigation channels that transported water of irrigation.
Amu Darya and Syr Darya Rivers, which were two major tributaries
of Aral Sea. It was intended by this water supply to convert the
region from an arid area into a newly developed cotton agricultural
area (Figure. 1). Many farmers immigrated from present Russia and
started successful cotton agriculture.
19
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
20
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5. VAJONT DAM Concrete arch dam was completed in 1960 and impounding
started. Then, the feared left bank started to move gradually.
Construction of Vajont Dam in Northern Italy (Fig. 6) was planned Fortunately, however, this slope movement stopped when the
and designed in 1950s as an extremely important project as the reservoir water level was lowered. This experience gave an
energy resource that would contribute to the development of Italian impression that it is possible to safely control the slope movement
industries and economy. The dam site was situated in a narrow and by raising and lowering the reservoir water level. During the first
deep gorge and looked like an ideally efficient dam site: rigid rock and second raising, this idea worked well (Fig. 7). However, the
mass, narrow and high dam with good amount of impoundment, and third raising triggered very rapid slope movement. The gauge
significant height difference for hydropower generation. became narrower quickly, the reservoir water level came up very
There were several problems at the site (Nonveiller, 1987; fast, and it was impossible anymore to lower the water level. Finally,
Semenza and Ghirotti, 2000). In the valley, there was a trace of an the entire mountain slope jumped into the reservoir and induced
ancient slope failure that suggested possibility of slope failure. high tsunami. This wave overtopped the dam and destroyed villages
During dam construction, the left bank of the reservoir in the downstream area. The number of victims was more than 2000.
(Toc Mountain) started to move and cracks developed in the The causative mechanism was the existence of weak rock layer
mountain slope. This further suggested the risk of natural disaster at depth beyond the capability of subsurface investigation. Reservoir
but was considered minor. In those days, there was not a reliable water pressure propagated into this layer, reduced the effective
technology to collect rock specimens from the depth of hundreds of stress and shear strength, and made the entire slope unstable. Then
meters and measure their mechanical strength. Thus, safety factor of the question is whether or not the senior engineers who feared the
the slope was not precisely assessed. However, the experienced possibility of disaster were able to stop the project. Issues to be
senior engineers felt that something wrong may happen in the dam. considered are;
- This dam project was a national project that was expected to help
develop the economy of Italy.
- Huge money had been spent for the success of this project.
- In spite of fear, it was not possible to verify the risk of slope
failure because deep subsurface investigation was not possible.
- Accordingly, the project was continued as scheduled until the last
moment of tragedy.
Good engineers should not behave against their official missions
because of personal fear. Professional engineers have to respect
confidentiality and should not leak important information to
outsiders. Then how could experts then avoid the tragedy? The
Vajont tragedy still remains an important material for study on
responsibility of engineers.
6. RESERVOIR-INDUCED SEISMICITY
Figure 6 Vajont Dam in Italy The reservoir of Koyna Dam in India (Figure 10) is one of the
earliest examples of reservoir-induced seismicity where significant
earthquake was triggered by impoundment. The 1967 Koyna
earthquake that registered M=6.3 claimed more than 100 victims.
The focal depth of the quake was 30 km. Fig. 8 shows a correlation
between the elevation of reservoir water and the number of
associating earthquake swarms. There is a reasonable correlation
between them, while exhibiting some time delay. The causative
mechanism of reservoir-induced seismicity has been considered
two-fold: stress in earth crust induced by weight of reservoir water,
and/or the increased pore water pressure in unstable faults, leading
to reduced effective stress and shear strength. The author supposes
that the former is not likely the cause of fault rupture because
reservoir is always located at low places (bottom of a valley) and the
increased gravity at the valley bottom tends to equalize the surface
stress distribution. Hence, the subsurface shear stress decreases. In
contrast, the pore water pressure increases the chance of fault
rupture. This situation is same as the cause of slope failure at the
Vajont dam site.
The second well-known example of reservoir-induced seismicity
is that of the Nurek Dam of Tadjikistan for which the induced
earthquake was of M=4.6 at maximum, while the focal depth was 10
km or less (Simpson and Negmatullaev, 1981). The correlation and
Figure 7 Time history of water level in the Vajont Reservoir the reservoir water level (depth) is significant as illustrated in
and rate of slope movement (Semenza and Ghirotti, 2000) Figure. 9.
21
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 8 Correlation between number of earthquakes and level of Koyna reservoir water (Gupta, et al., 1969)
Number of
earthquakes in 10
days
Figure 9 Correlation between reservoir water level and frequency of earthquakes in Nurek Dam (Simpson and Negmatullaev, 1981)
22
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
In contrast, Chen (2009) discussed the occurrence of fore-, main 7. EARTHQUAKES INDUCED BY EXTRACTION OF
and aftershocks and argued that the observed pattern of those shocks OIL AND GAS
is different from those of known induced seismicity. Gahalaut and
The most recent technology for extraction of oil and gas from the
Gahalaut (2010) calculated the stress that was induced by the
earth crust injects pressurized water and other liquids in order to
reservoir water to discuss that the induced stress is oriented in
facilitate the outward flow of the desired resources. The collection
opposite direction to trigger the earthquakes.
of shale oil relies on hydraulic fracturing and hence injects
To date, it is difficult to draw a definite conclusion on the cause
pressurized fluid into earth crust as well. It is obvious that such a
of the Wenchuan earthquake. However, this controversy in
procedure may induce earthquakes as was discussed in the previous
conjunction with other induced earthquakes recommends the
chapter (Hubert, M.K. and Rubey, 1959).
community to be more careful of the possible risk of damaging
One of the earliest earthquakes induced by liquid injection
induced earthquakes.
occurred near Denver, USA, in late 1960s when waste water was
being injected into the earth crust (Healy et al., 1969). Afterwards,
similar phenomena have been reported from such oil/gas projects as
in Colorado (Raleigh at al., 1970). Fig. 15 illustrates a correlation
between gas production and number of earthquakes in Groningen oil
23
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
field in the Netherlands. There is an evident correlation between the Foulger, G.R., Wilson, M.P., Gluyas, J.G., Julian, B.R., Davies, R.J.
amount of gas production and the rate of earthquakes. Because of (2018) Global review of human-induced earthquakes, Earth-
this trend and the seismic risk in the country, the gas production in Science Reviews 178: 438-514.
this region will be terminated in 2030 or so. In spite of the recent Furudoi, T. and M. Kobayashi (2009) Geotechnical issues and
decreasing trend, however, a relatively big earthquake of M=3.4 approach on Kansai International Airport project - Prediction
occurred in January, 2018. and performance of settlement -, Proc. JSCE C 65(4): 998-
1017.
Gahalaut, K. and Gahalaut, V.K. (2010) Effect of the Zipingpu
Gas extraction in Groningen and earthquake swarm reservoir impoundment on the occurrence of the 2008
Netherlands earthquake rate.qpc Wenchuan earthquake and local seismicity, Geophysical
60 Journal International 183: 277–285.
Gupta, H.K. (1992) Reservoir-induced earthquakes, Elsevier.
50 Gas production Gupta, H.K., Narain, H., Rastogi, B.K. and Moha, I. (1969) A study
(billion cubic m/year)
40 of the Koyna earthquake of December 10, 1967, Bulletin of
the Seismological Society of America 59(3): 1149-1162.
30 Healy, J.H., Rubey, W.W., Griggs, D.T. and Raleigh, C. B. (1968)
The Denver earthquakes, Science 161(3848): 1301-1310.
20 Hubert, M.K. and Rubey, W.W. (1959) Mechanics of fluid-filled
Number of earthquakes porous solids and its application to overthrust faulting, Bull.
10 (magnitude>1.5) in 9 months Geol. Soc. Am. 70): 115-166.
0 Lei, X., Ma, S., W en, X., Su, J. and Du, F. (2008) Integrated
2002 2004 2006 2008 2010 2012 2014 2016 2018 analysis of stress and regional seismicity by surface loading -
Year A case study of Zipingpu Reservoir, Seismology and Geology
30(4): 1046-1064 (in Chinese).
Figure 15 Correlation between gas production and number of Liao, Y. (2009) Verification of the effects of Zipingpu Dam
earthquake in Groningen Gas field (after NWO, 2017) reservoir on inducing Wenchuan earthquake, Glacier and
Earth (n Chinese).
Ma, X., Li, Z., Hua, P., Jiang, J., Zhao, F., Han, C., Yuan, P., Lu, S.
8. CONCLUSION and Peng, L. (2015) Fluid-injection-induced seismicity
The present paper addressed several recent big projects by which experiment of the WFSD-3P borehole, Acta Geologica Sinica
some negative aspects became evident. It is thought herein that, English Edition, 89(3): 1057-1058.
although civil engineering has been doing many good projects for Nonveiller, E. (1987) The Vajont reservoir slope failure,
the welfare of people, sometimes mistakes are made to cause Engineering Geology 24: 493-512.
unfavorable consequences. The reason for this is that nature is still Nakase, A. (1992) Settlement of Kansai International Airport Land,
more complicated than what human knows. To cope with this Proc. JSCE(454/III-20): 1-9.
difficult situation, it is important to continuously monitor what is NWO (2017) Aardbevingsrisico's in Groningen (Earthquake risks in
going on in nature and detect small phenomena that suggest more Groningen) Stand van zaken opvolging aanbevelingen (State
serious situations that are going to occur in near future. If the future of affairs follow-up recommendations).
situation seems serious, it is important to stop or cancel, in the Ohtake, M. (1974) Seismic activity induced by water injection at
extreme case, the project. The current society does not have a Matsushiro, Japan, Journal of Physics of the Earth 22: 163-
system that allows such a very serious decision to be made. 176.
Raleigh, C.B., Bredehoeft. IleaJy, J.H. and Bohn, J. (1970)
Earthquakes and waterflooding in the Rangely oil field, Geol.
9. REFERENCES Soc. Am. Abstract 2: 660-661.
Semenza, E. and Ghirotti, M. (2000) History of the 1963 Vaiont
Aladin, N.V. and Potts, W.T.W. (1992) Changes in the Aral Sea slide: the importance of geological factors, Bull. Eng. Geol.
ecosystems during the period 1960–1990, Hydrobiologia Env. 59: 87-97.
237(2): 67-79. Simpson, D.W. and Negmatullaev, S.K. (1981) Induced seismicity
Benedick, R.E. (1979) The High Dam and the transformation of the at Nurek Reservoir, Tadjikistan, USSR, Bulletin of the
Nile, The Middle East Journal 33(2): 119-144. Seismological Society of America 71(5): 1561-1586.
Borovskiy, V.M. (1980) The drying out of the Aral Sea and its Wang, M.R., Yang, M.Y., Hu, Y.L., Li, Z.J., Chen, Y.T., Jin, Y. and
consequences, Soviet Geography 21(2): 63-77. Feng, D. (1976) A preliminary study on the mechanism of the
Carder, D.S. (1945) Seismic investigations in the Boulder Dam area, reservoir impounding earthquakes at Hsinfengkiang, Scientia
1940--1944, and the influence of reservoir loading on local Sinica, 19(1): 149.
earthquake activity, Bulletin of the Seismological Society of Whish-Wilson, P. (2002) The Aral Sea environmental health crisis,
America 35(4): 175-192. Journal of Rural and Remote Environmental Health 1(2): 29-
Chen, H. (2009) Wenchuan earthquake was not induced by 34.
reservoir, World of Science 5: 51-52. Wilson, M.P., Foulger, G.R., Gluyas, J.G., Davies, R.J. and Julian,
Elhance, A.P. (1997) Conflict and cooperation over water in the B.R. (2017) HiQuake: The Human-Induced Earthquake
Aral Sea basin, Studies in Conflict & Terrorism 20(2): 207- Database, Seismological Research Letters 88(6): 1560-1565.
218.
24
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Recently, ground collapse are increasing due to the old pipelines buried under the ground like waste water pipeline. The
purpose of this research was to analyze the causes of pipe deformation during the test construction of the channel pipeline improvement
project with the nature river work. The particle distribution, physical and mechanical characteristics were investigated through laboratory,
and field test using the additional boring exploration. Stability analysis was performed by applying soil cement wall with high strength deep
cement mixing method(SCW+HDCM) and Sheet Pile with HDCM method to two points, No. 84+4.0 and No.100+0.0, which are the most
vulnerable to subsidence of pipeline. The soft ground layer was reinforced by HDCM method for 2m to improve the strength of the clay soil
and to prevent settlement, satisfying the allowable residual settlement of 100mm.
Keywords: Waste water pipeline, stability, temporary earth wall, ground improvement, differential settlement.
25
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The transverse earth pressure starts from the stationary earth passing of sieve number 200. It was carried out water washing
pressure. When the wall is pushed to the excavation side, the earth method and hydrometer distribution test. The particle size
pressure decreases to the main earth pressure. If it is pushed to the distribution of the soil is as shown in Figure 4. As a result of particle
back side, the earth pressure continues to increase but the manual size analysis by unified classification method, the soil type of KG-1
earth pressure can’t be increased. In other words, the minimum and to KG-3 was CL, and KG-4 was identified as ML and KG-5 as GC.
maximum earth pressure limits are set. The ground modeling is
simulated by a spring, and the basic equation of carbon spring is
given by the following Eq. (3).
d4y Ap E p
Ew I w + y = pi ± k h y (3)
dz 4 Lp
Figure 2 Schematics of elastic beam model KG-4 101.40 0.30 0.05 18.05 0.95 1.16
3. SUBSURFACE EXPLORATION ON SOFT GROUND In-situ test was performed. Cohesion of soil was measured with
AND SOIL CHARACTERISTICS a field vane tester in order to confirm the undrained shear strength of
The total length of this construction is 7.9km and 3.55km is undisturbed state and disturbed state. The test results showed that
overlapped with natural river construction. A total of three the boring depth was about 3.5~5.0, cohesion of undisturbed sample
investigations were conducted on the design subsurface exploration with depth was 21.4 ~ 23.3 kPa , cohesion of disturbed sample was
of the waste water pipeline. In this study, the existing ground 2.6 ~ 3.5, and the sensitivity ratio of each boring was 7.15 ~ 8.23.
surveys were combined and re-confirmed the soft ground layer On the design, cohesion was similar with additional survey as 22.0 ~
through additional drilling of 5 holes. Figure 3 shows sewer pipeline 35.3, but sensitivity was not considered. Soil samples of KG-1, KG-
construction map and situation of subsurface exploration. 2, KG-3, KG-4, and KG-5 were very sensitive. Therefore, it is
expected that the ground has large deformation or the settlement
possibility is high due to the ground disturbance during excavation.
The results are shown in Table 2.
Table 2 Result of vane shear test in the construction site
c ur
cu cu
Boring Depth (distur
(undistur (Design, Sensitivity
No. (m) bed,
bed, kPa) kPa)
kPa)
KG-1 3.5 22.9 2.8 22.3 8.08
26
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Cohesion and internal friction angle were compared and examined Structural analysis was carried out a beam on elasto-plastic
by Dunham, Terzaghi-Peck, Meyerhof, Osaki, Schmertmann, and foundation model. It is similar to the beam on winkler foundation
Hisatake using empirical formulas based on N values. The design used to design piles with a foundation or horizontal load. The wall
constants were calculated as shown in Table 3 based on the stability at the final stage was evaluated from the stepwise
laboratory test results of the drilled specimens. excavation analysis. It is assumed that the stress is redistributed due
to the empirical earth pressure over time after the excavation is
Table 3 Soil property of shear strength for each layer in soft ground completed. The stability of the earth retaining walls and support
materials to the earth pressure is evaluated.
Unit weight, Cohesion, Internal friction
Soil type
r t (kN/m3) c(kPa) angle(°)
Soft soil property was revised using the additional ground
survey of 5 boring data, the subsurface exploration data of basic and
Reclaimed layer 19.0 10.0 20 detailed design of the sewer pipeline and the subsurface exploration
data of the basic and detailed design of the natural river
Accumulation
17.0 17.0 5 improvement project. As a result, the soft ground layer was deeper
(clay, N≤4) than the original design ground survey with a maximum of 12m, and
Accumulation
17.6 40.0 5
the ground layer also changed. Based on the revised plan, three
(clay, 4<N≤10) weakest section sites were selected. EXCAV and SUNEX were used
Accumulation to evaluate the stability of the pipeline by prefabricated wall.
18.0 5.0 25
(sand)
Weathered soil 19.0 20.0 30 5.2 Stability analysis for deep wall after ground improvement
Stability of construction depth of temporary wall was analyzed.
Weathered rock 20.0 30.0 33
When the construction depth of temporary wall is 4.0m at the
maximum excavation depth of 5.5m, the safety factor of the
Soft rock 23.0 100.0 33
construction depth is calculated to be 2.46 which satisfy 1.2 as a
standard of safety factor. The results are shown in Table 6.
Table 4 shows that the pre-consolidation load was 70.6kPa to
Table 6 Safety factor of section no.18+11.00
85.1kPa, compression index was 0.325 to 0.522, swelling index was
0.06 to 0.124, consolidation coefficient was 1.96e-3cm2/sec to Constructi Safety factor
Excavation
8.09e-3cm2/sec, and initial void ratio values was 1.032 to 1.311. It Section
depth(m)
on
was applied in the design of sewer pipeline construction. depth(m) Result Minimum Standard
No.18+11.00 5.5 4.0 2.46 1.2
Table 4 Soil property on design around waste water pipeline
Initial
Preconsoli- Compress- Swelling Consolidation Safety for ground settlement was calculated. The section no.
void
Boring No. dation load, ion index, coefficient
ratio, 24+0.0 was 7.2m in the thickness of the soft ground, and no other
P c (kPa) index, c c cs c v (cm2/sec)
eo channel was buried in the vicinity. The section no.84+4.0 is 8.6m in
thickness of soft ground. An existing line appeared at a distance of
BH-1 70.6 0.522 0.124 1.96E-03 1.311 2.98m from the position where the sewer pipeline is constructed.
The section no.100+0.0 point was 12m in thickness of the soft
BH-2 77.1 0.347 0.060 4.44E-03 1.032 ground, and the soft ground thickness of the three sites was the
largest. Figure 5 shows the result of settlement with elapsed time.
BH-3 85.1 0.325 0.063 8.09E-03 1.195
5.1 Elasto-plastic modelling for temporary earth wall As a result of the analysis of the soft ground, it is confirmed that it is
highly sensitive soft clay. Therefore, it is suggested to improve the
27
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ground using the high strength DCM method(HDCM). Figure 6 stability of the pipeline. Therefore, it is suggested that the
shows cross section of No.100+0.0 as original design section and improvement of the soft ground should be done as a way to
reinforced section for ground improvement. minimize disturbance to the lower part of the basement.
28
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Coastal cities like Singapore and Jakarta have undergone rapid economic development in the past decades. There is an
increasing demand for infrastructure upgrading. With climate changes and population increase, geotechnical design and construction are
getting more and more challenging for coastal cities. We need to develop new technologies and new solutions to tackle the challenges and
enhance our ability to fight disasters such as earthquake and flood. In this paper, four different solutions pertinent to geotechnical
engineering for coastal cities are presented. The first is the establishment of a web-based 3D geological map to make better use of geological
and geotechnical data and information for more reliable geotechnical design. The second is a new method for using dredged slurry or soft
materials for land reclamation, as there is a shortage of granular fill materials in many cities. The third is the technology for construction of
seawalls using suction caissons. The last solution is the biogas desaturation method for mitigation of soil liquefaction which is becoming the
most cost-effective solution for prevention of liquefaction for a large area.
Keywords: Coastal cities, geological model, geotechnical engineering, land reclamation, liquefaction, underground construction.
29
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
As an example, the fence diagram constructed for one area in surcharge, or improved dynamic consolidation methods may be used
Singapore is shown in Figure 1. Using this fence diagram, a 3D instead (Chu et al., 2009b; Varaksin and Yee, 2009). Another
geological model can be constructed as shown in Figure 2. method is to premix cement with soft soil or mixing cement and soil
on-site or to use a combined cement mixing and consolidation
method (Chu et al., 2005). However, the use of cement is expensive.
When there are million cubic meters of fill to be treated, a slight
reduction in the unit cost can lead to a huge saving. Clay slurry may
also be dewatered using the geotextile tubes method in which clay
slurry is pumped into tubes made of woven geotextile. The solids of
the soil are retained by the geotextile and the water is consolidated
under the pumping pressure or self-weight. However, this method
may only be useful when the amount of soil to be improved is small.
Rigid inclusions or composite foundation methods by using columns
and reinforcements such as stone columns or deep cement mixed
columns are another type of methods to treat soft ground (Chu et al.,
2009b). However, these methods may not be suitable when the soil
is too soft to provide enough lateral support.
An alternative to vertical drains is to combine vacuum
Figure 2 3D geological model for one area in Singapore preloading with horizontal drains. One application as proposed by
Chu et al. (2005) and Chu and Lim (2008) is to use binders treated
The 3D geological model can not only assist engineers to sewage sludge as fill materials for land reclamation as illustrated
understand or visualise the underground geological conditions better, schematically in Figure 4.
but also help to make better interpretation of geotechnical properties
of the soil or rock. For example, using the 3D geological model
established for Singapore, we could identify the ancient river
channels. One example is given in Figure 3 (Pan et al. 2019).
Knowing the positions and extent of the ancient river channels, we
will be able to construction more accurate soil profiles or ground
models to enable more reliable geotechnical design to be achieved
for underground construction.
30
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
predictable and the reliability of the design and quality control of the of breakwater on soft seabed or in deep water. Suction caissons are
construction can be improved. If there is a need to accelerate the sunk into seabed using suction until sufficient bearing capacity is
consolidation process even further, electrolytes as either anode or obtained. In this way, the treatment of soft seabed soil is not
cathode can be embedded into the HDeGs to create electro-osmosis required. This will save both time and construction cost. This
effect that can also be incorporated into the horizontal drains as method can also be used for disaster rehabilitation or restoration of
shown in Figure 6. failed seawalls. The concrete caissons can be fabricated using
standard modulus in a casting yard connected to the sea. The
fabricated caissons or structures can be towed into the seawall
location and installed into the seabed using suction. With proper
design, the side friction of the caissons will provide sufficient
bearing capacity to safeguard stability and serviceability for the
seawalls. After the installation of suction caissons, the seawalls can
be installed on top of the suction caissons. One example of using
suction caissons as the foundation for a breakwater is shown in
Figure 11. The detail of this project can be found in Yan et al.
(2009).
Figure 6 Use of electro-osmosis together with HDeGs Figure 7 Use of sand key as foundation for a rock bund for
reclamation in Singapore (Chew and Wei, 1980)
The performance of the HDeG product has been evaluated using
model tests and the testing results indicate that the use of HDeG is
effective (Chu, 2016; Chen, 2018). It should also be noted that no
membrane is required in this method as was case for the model tests.
Thus all the problems related to the use of PVDs and vacuum
preloading, such as sand blanket, the formation of a working
platform and the possibility of punching of membrane, were
eradicated. This is the other advantage of using horizontal drains. As
a first practical application, the proposed method has been used in
the design for a project under tendering.
Figure 8 Use of PVDs (not shown) and the weight of the rubble
4. CONSTRUCTION OF SEAWALLS mount to consolidate the soft seabed soil (not shown) before
Seawall construction has become more important for coastal semicircular concrete caissons were placed to form a breakwater
protection against the effects of sea level rising and natural disasters. in China (after Yan et al. 2009).
Seawalls can also be a way to fight flooding. The Giant Sea Wall
Jakarta is one of the examples.
The traditional method for seawall construction is to use sand
key to support embankment as shown in Figure 7. However, the
sand key construction involves dredging of a large amount of soft
clay from the seabed and placement of an equally large amount of
sand fill. Both the disposal of the dredged clay and the supply for
sand fill become problems. An alternative method is to use the so-
called non-dredging seawalls, or in other words, to build seawalls
without dredging of soft soil. This requires the soft seabed soil to be
improved. The methods adopted to treat soft seabed soil in the past
include the use of PVDs plus preloading (see Figure 8), deep cement
mixing or stone columns as shown in Figures 9 for a sand bund
construction in Singapore (Leong and Raju, 2007).
When water depth is relatively deep, the above methods may no Figure 9 Use of stone columns installed into soft clay from a barge
longer be applicable. In this case, suction caissons may become a to support sand bund (after Leong and Raju 2007)
better foundation type for seawalls. The suction caissons are made
of one or a group of steel or concrete cylinders as shown in Figure However, when the water is deep, or in other words, the upper
10 as an example. Four suctions are used as a group and installed structures or caissons to form the seawall is too big or too heavy, it
together to allow the penetrations of the 4 cylinders to be adjusted so will require a special offshore crane to install the upper caissons.
that the four suction caissons can sink into the clay evenly without This can be difficult or expensive. An alternative is to assemble the
any titling. This method is particularly suitable for the construction upper caissons with the suction caissons in the casting yard
31
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
and install the whole segment as one piece as shown in Figure 12. 5. MITIGATION OF SOIL LIQUEFACTION
Upon reaching the site, the whole segment can be ballasted using
water to the seabed level before suction is applied to sink in the Liquefaction of granular materials is one of the major geohazards.
suction caissons. The sea wall formed with back fill behind is shown Conventional methods for mitigating liquefaction include soil
in Figure 13. compaction or deep cement mixing. However, both methods are
expensive for large-scale deployment. It is necessary to develop
methods for liquefaction mitigation to be effective as well as
economical. An experimental study for the development of a more
cost-effective approach for mitigation of liquefaction hazard is
presented in this paper.
It has been found by some researchers that inclusion of some gas
bubbles in a fully saturated sand can lead to an increase of shear
strength in cyclic triaxial tests (Yoshimi et al. 1989; Xia and Hu
1991). Use of the desaturation method to enhance the resistance to
liquefaction has been attempted before (Yang et al. 2004; Okamura
and Soga 2006; Yegian et al. 2007; He and Chu 2014).
A biogas-desaturation method has been developed in this study
as a liquefaction mitigation approach. The denitrification process is
employed to generate nitrogen gas bubbles in sand to reduce the
degree of saturation of a fully saturated sand and thus increase its
resistance against liquefaction. Denitrification is a microbially
Figure 10 Suction caissons used as foundation for seawalls facilitated process of nitrate reduction that produces molecular
nitrogen (N2) through a series of intermediate gaseous nitrogen
oxide products. The complete denitrification process can be
expressed as a redox reaction as Eq. (1).
(1)
A series of shaking table tests were also carried out and the
results also show that the pore pressure generation and settlement in
the bio-desaturated sand was largely contained by using the
proposed method. Figure 14 shows the development of pore
pressure ratio in sand with different degrees of saturation under the
same input acceleration of a = 1.5 m/s2. Ru is defined as the ratio of
maximum excess pore pressure generated by the cyclic load to the
Figure 11 Use of suction caissons for a breakwater construction initial effective overburden stress. The pore pressure increased
in China (after Chu et al. 2012) during the cyclic loading and dissipated afterwards. For fully
saturated sand, there was a considerable amount of increase in
excess pore pressure as the shaking took place. The pore pressure
ratio exceeded 0.9 which indicates that the liquefaction occurred (He
et al. 2013). The pore pressure generated in biogas desaturated sand
were substantially lower than that in fully saturated sand. When the
degree of saturation Sr dropped slightly to 90%, the increase in pore
pressure becomes insignificant. The maximum Ru ratio in the biogas
desaturated sample (Sr = 90%) was less than 0.2 which is far less
than a trigging value of Ru = 0.5 when liquefaction could occur (He
et al. 2013).
32
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
confined within only 1%. This is evident that the biogas desaturated drains accelerate the sedimentation process of the clay mud
sand had strong resistance to liquefaction. This finding corroborates layer and thus shorten the construction time. Thirdly, the
the description of Tokimatsu and Seed (1987), who reported that strength of the clay can be increased before the next layer of
volumetric strains observed in non-liquefiable soils were usually clay is placed. Fourthly, with the use of horizontal drains, all
less than 1%. the fills placed on top becomes the fill surcharge as well. Hence
the fill surcharge load increases with the height of the fills.
3) For seawalls to be constructed in relative deep water, the
suction caisson method is more cost-effective as treatment of
he soft seabed soil is not required.
4) The biogas desaturation method is effective for mitigation of
soil liquefaction as proven by the shaking take tests. When
there is ground water seepage, the stability of the gas bubble
may become a concern. In this case, the biogas desaturation
method can be combined with biosealing method to prevent the
migration of gas bubbles.
7. ACKNOWLEDGEMENTS
The support from a number of finding agencies in particular, the
National Research Foundation and the Ministry of National
Figure 15 Vertical strain development in sand under shaking for Development, Singapore, is gratefully acknowledged. The
fully and desaturated sand with degree of saturation of 90% contributions from several other researchers and students to the
studies presented in this paper are also gratefully acknowledged, in
When there is no or very small seepage flow in the soil, the particularly, Prof Jia He, Prof Wei Guo, Prof Hailei Kou, Dr Bing Li,
biogas desaturation method is sufficient for mitigation of Dr Chuangzhou Wu, and Dr Kok Pang Lam.
liquefaction hazards. However, when there is a relatively big
seepage in soil, the stability of the gas bubbles may become a 8. REFERENCES
concern. In this case, a combined biogas desaturation and Chen, H. (2018). “Prefabricated horizontal drains with vacuum
bioclogging or the so-called “combined biodesaturation and preloading for improvement of soft marine clay.” PhD first
bioclogging” method could be used instead (Wu, 2015). In this year report, Nanyang Technological University, Singapore.
combined approach, the purpose of bioclogging is to “block” the Chew, S.Y. and Wei, J., (1980) “Major Reclamation Scheme for
passage for small gas bubbles to aggregate into bigger bubbles so as Marina City, Singapore”. Coastal Engineering 1980, pp.
to pre-empt the conditions for gas bubbles to escape from the 2245-2260.
ground. The amount of biogrout required for bioclogging in this Chu, J. (2016). “Innovation in Soil Improvement Methods”, S L Lee
case is much less than that for biocementation for the purpose to Lecture, Proceedings 19th Southeast Asian Geotechnical
increase the shear strength of sand. In terms of cost-effectiveness, Conference & 2nd AGSSEA Conf., Kuala Lumpur 31 May –
the construction cost involved in the combined biodesaturation and 3 June.
bioclogging method for liquefaction mitigation will still be Chu, J., Bo, M.W. and Arulrajah, A. (2009a). “Reclamation of a
significantly lower than the cost in the solo biocementation method. slurry pond in Singapore”, Geotechnical Engineering - Proc
The detail of the combined biodesaturation and bioclogging method of ICE, 162(1), pp 13-20.
will be presented in a separate paper. Chu, J. and Guo, W. (2015). “A geotechnical apparatus and method
of usage”, SG Patent Application, Serial No. 10201509039Y
6. CONCLUSIONS on 3 November 2015.
Four different solutions pertinent to geotechnical engineering for Chu, J., Goi, M. H. and Lim, T. T. (2005). “Consolidation of cement
coastal cities are presented: 1) web-based 3D geological map to treated sewage sludge using vertical drains”, Canadian
make better use of geological and geotechnical data and information Geotechnical Journal, 42(2), pp 528-540.
for more reliable geotechnical design; 2) horizontal drain enhanced Chu, J. and Lim, T. T. (2008). “Use of sewage sludge and copper
geotextile method to use dredged slurry or soft materials for land slag for land reclamation.” Geotechnics of Waste
reclamation; 3) technologies for construction of seawalls using Management and Remediation, GeoCongress 2008, ASCE
suction caissons; and 4) biogas desaturation method for mitigation Geotechnical Special Publication No. 177, pp 352-359.
of soil liquefaction. The following conclusions can be made: Chu, J. and Raju, V. (2012). “Prefabricated vertical drains”. Chapter
1) The web-based 3D geological map allows all the existing 4 in Ground Improvement, pp.87-167.
geological data and information as well as geotechnical data Chu, J., Yan, S. W., Li, W. (2012). “Innovative methods for dike
and information to be systematically ultilised for better construction – An overview.” Geotextiles and
geological design. This system makes future site investigation Geomembranes, 30(2), pp 35-42.
much more effective and reliable by providing more data and Chu, J., Varaksin, S., Klotz, U. and Mengé, P. (2009b).
additional geological processes (such as ancient river channels “Construction pro-cesses”. State-of-the-Art-Report, Proc.
or igneous intrusion etc) for geological profiling. The data 17th Int. Conf on Soil Mech. and Geot. Eng., Alexandria,
interpretation also becomes more meaningful when integrating Egypt, 5-9 Oct., 4, pp 3006-3135.
the new site investigation data with the existing geological data He, J. (2013). “Mitigation of liquefaction of sand using microbial
and geological map. methods.” PhD thesis, Nanyang Technological University,
2) When there is a shortage of granular fill for land reclamation, Singapore.
soft slurry or other soft fill materials may have to be used. In He, J. Chu, J. and Ivanov, V. (2013) “Mitigation of liquefaction of
this case, the conventional PVDs plus surcharge or vacuum saturated sand using biogas”. Géotechnique, 63(4), 267-275.
method becomes inefficient or time consuming. The horizontal He J. and Chu, J. (2014) “Undrained responses of microbially
drainages enhanced geotextile sheet (HDeG) method offers desaturated sand under monotonic loading”. Journal of
several advantages over the conventional method. Firstly the Geotechnical and Geoenvironmental Engineering, 140(5),
HDeG method is much faster. The consolidation can take place pp503-515.
as soon as the clay layer is placed. Secondly, the horizontal
33
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Leong, K.W. and Raju, V.R. (2007). “Use of vibro replacement Wu, S.F. (2015). “Mitigation of liquefaction hazards using the
method for the stabilization of an offshore slope on soft clay”. combined biodesaturation and bioclogging method.” PhD
Proc 16th Southeast Asian Geotechnical Conference, 8-11 thesis, Iowa State University, Ames, IA, USA.
May, Malaysia, pp547-550. Xia, H., and Hu, T. (1991) “Effects of saturation and back pressure
Okamura, M., and Soga, Y. (2006) “Effects of pore fluid on sand liquefaction”. Journal of Geotechnical Engineering,
compressibility on liquefaction resistance of partially 117(9), pp1347-1362.
saturated sand”. Soils and Foundations, 46(5), pp695-700. Yan, S.W., Chu, J., Fan, Q.J., and Yan, Y. (2009). “Building a
Pan X., Aung Z., Nyo A.K.K., Chiam K., Wu D.F., and Chu J. breakwater with prefabricated caissons on soft clay”.
(2019). "Interpretation of distribution of ancient rivers in Proceedings of ICE, Geotechnical Engineering, 162(1): pp3-
Singapore using 3D geological model." In the Eighth 12.
International Conference on Case Histories in Geotechnical Yang, J., Savidis, S., and Roemer, M. (2004) “Evaluating
Engineering, Philadelphia, PA, USA. liquefaction strength of partially saturated sand”. Journal of
Pan, X., Guo, W., Aung, Z., Nyo, A. K., Chiam, K., Wu, D., and Geotechnical and Geoenvironmental Engineering, 130(9),
Chu, J. (2018). "Procedure for Establishing a 3D Geological pp975-979.
Model for Singapore." In GeoShanghai International Yegian, ., Eseller-Bayat, E., Alshawabkeh, A., and Ali, S. (2007)
Conference, Springer, Singapore, pp81-89 “Inducd-partial saturation for liquefaction mitigation:
Tokimatsu, K. and Seed, H. B. (1987) “Evaluation of settlements in experimental investigation”. Journal of Geotechnical and
sands due to earthquake shaking”. Journal of Geotechnical Geoenvironmental Engineering, 133(4), pp372-380.
Engineering, 113(8), pp861-878. Yoshimi, Y., T., K., and Kohji, T. (1989) “Liquefaction resistance
Varaksin, S. and Yee, K. (2007). “Challenges in ground of a partially saturated sand”. Soils and Foundations, 29(3),
improvement techniques for extreme conditions: Concept and pp157-162.
Performance”, Proc. 16th Southeast Asian Geot Conf, 8-11
May: Kuala Lumpur, pp101-115.
34
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This paper presents the performance of cross walls, buttress walls and their combination in reducing wall deflection and
ground settlement in excavations. Three dimensional finite element method was used to evaluate the performance. Cross walls have a very
good effect in reducing the movement but expensive. Buttress walls have a moderate effect in restraining the wall deflection. With proper
combination of cross wall and buttress walls, both the wall deflection and cost can be reduced significantly, even to achieve a strut free
excavation condition.
Keywords: Excavation, cross wall, buttress wall, strut free system.
1. INTRODUCTION wall (Ht) was 57.5 m. To reduce the lateral wall deflection and
ground settlement induced by excavation, 3 cross walls of 1.0 m
The huge population in urban areas due to economic development thickness and 26 m intervals were constructed and their depths were
has caused situations such as much denser buildings and deeper between GL-1.5 m and GL-45 m in the north–south direction. The
foundation excavations than ever before. Construction disasters and cross walls between GL0 m and GL-1.5 m were backfilled with the
adjacent building damage often occur due to ground settlement as a in situ soil, those between GL-1.5 m and GL-22 m were cast with
consequence of deep excavation, which not only affects construction 13.7 MPa concrete, and those below GL-22 m were cast with
progress, but also increases public nuisance. The protection of 24.0 MPa concrete. The cross wall were demolished with excavation.
adjacent buildings has become a major concern for designers and Figures. 3(a) and 3(b) show that the analyzed wall deflections at
contractors of deep excavations. It is thus an important hazard SO-1 and SI-8, respectively, for stages 7 and 9. For comparison,
prevention task in geotechnical engineering to conduct a study on analysis with assumption of no cross walls was also performed. The
protection of adjacent buildings when deep excavations are carried maximum wall deflection at the midpoint of two cross walls was
out. The integrity of adjacent buildings can be protected by predicted to have a 67% reduction by the installation of such walls
underpinning. Nevertheless, it is not practical to underpin a large (SO-1) whereas 77% at the cross wall section (SI-8). Installation of
number of target buildings, either from a financial or political point cross walls can significantly reduce lateral wall deflections [3].
of view. Therefore, reducing excavation induced movements to a
tolerable amount by auxiliary measures is a viable alternative. This
paper presents the performance of cross walls, buttress walls and Cross wall
(dismantle along
GL+0m 1FL GL+0.2 with excavation)
combination of cross walls and buttress walls in reducing the wall
Fill t =18.25 kN/m
GL-3.0m
3
B1FL GL-4.4
deflection or ground settlement. Moreover, a strut free excavation SI-8 SI-7 Cross wall
CL
B2FL GL-9.0
t =18.05 kN/m3
system is presented using a combination of cross walls and buttress SI-6
=30o
B3FL GL-13.4
B4FL GL-16.8
walls. Diaphragm
Cc =0.4
C s =0.04
B5FL GL-20.2
B6FL GL-24.8
SI-9 wall SI-5
Inclined
steel strut
2. CROSS WALL Cross wall
GL-32.6m
CL
SI-10 t =18.74 kN/m3 Cross wall
Buttress =31o ~33o
The basic configuration of a cross wall is shown in Figures. 1(a) and wall
SI-4 Cc=0.3; Cs =0.03 GL-45.0
SI-2
1(b), which refers to the construction of a wall, connecting two SI-3
GL-51.0m
SM/GW
retaining walls opposite each other, prior to excavation. SO-1 SI-12 t =19.62 kN/m3 GL-57.5
=35 ~38
o o
Scale Unit:meter
Inclinometer 0 5 10 20 30 (m) N>50
GL-66.7m
Bedrock
(a) (b)
Diaphragm
wall Figure 2 The UPIB excavation project (a) plan (b) profile
Buttress wall
Excavation Excavation
Cross wall bottom bottom
12 10 8 6 4 2 0 10 8 6 4 2 0 12 10 8 6 4 2 12 10 8 6 4 2 0
0 0
Buttress wall Cross wall Buttress
wall 10 10
20 20
Depth (m)
35
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
He
Buttress walls Hbw
Buttress
Buttress
wall
wall
Diaphragm wall Cross
Hcw
wall
Hp
(a) (b)
Depth (m)
The author and his group also demonstrated that when buttress
walls maintained during excavation, i.e., no demolish with Figure 7 Spacing of BW/CW (a) 28 m (b) 8 m
excavation, the combined stiffness of buttress walls and diaphragm
walls plays a major role in reducing the wall deflection and ground
5. STRUT FREE EXCAVATION SYSTEM
settlement and frictional resistance between buttress walls and soil
play little role. The retaining wall behaves as a rigid stiff retaining Different types of Cross walls and cross walls can be combined to
wall and all of the wall, from top to bottom would move laterally. have more effect in reducing the wall deflection and ground
Fig. 6 shows the analyzed wall deflections at the final stage for the settlement. Figure. 8 shows an excavation with different
cases of buttress wall demolished with excavation, inner buttress arrangements of buttress wall where the excavation depth was 9.2
wall but maintains during excavation and outer buttress walls. The m. The maximum wall deflection at the final stage was about 50 mm
wall deflection can be reduced to a certain amount but its effect is [6].
still less than those of cross walls.
36
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
195.7m L = 100 m
Outer buttress wall Type E AH-2
SID 08 SID 10
Length = 5m (33m)
SID 09
North
AH-1
(30m)
Spacing = 8.5m BH-16 SID 11
BH-17 BH-18
Length = 6m (25m)
(25m) (15m) Cross Diaphragm
Type B Length = 8.5m BH-15
wall wall
B = 50 m
Type D (15m)
Ribs wall
88.6m
Middle section
BH-9 BH-10 BH-11 BH-12 BH-13 BH-14
8m
(25m) (25m) (25m) (25m) (15m) (15m)
Type C Cap-
BH-7 Length = 9.7m Buttress
8m
BH-6 (15m) Type A
(15m) Ribs wall wall slab
Inner buttress wall Length = 4.8m
BH-1 SID 07 SID 01
A(25m) A’
9m
SID 02
4m
SID 03
BH-19
SID 06
BH-3
SID 05 SID 04
BH-4 BH-20 AH-4
BH-5 Corner section Middle section
AH-3 (15m)
(40m) BH-2 (25m) (25m) (31.15m) (31.5m)
(30m)
SM 04
SM 03
SM 05
(25m) 13 m 13 m 8m 8m 8m
Figure 10 Plan view of the RFD strut free excavation system
Figure 8 A strut free excavation system by combination of different
shapes of buttress walls 6. CONCLUSION
Figure. 9 shows the RFD system, which comprises four main This paper presents the performance of cross wall, combination of
structures; namely, diaphragm walls, rib-walls, cross walls, and buttress wall and cross wall in deep excavations. Cross walls
buttress walls; and a complementary structure; namely, the cap-slab. function as lateral struts but exist before excavation. Installation of
The characteristics of the RFD system were (1) forming a cross walls in deep excavations can reduce the wall deflection to a
continuous earth retaining wall by constructing diaphragm walls very small amount. However, use of cross walls in a very wide
along a circumference of the excavated zone; (2) forming a rigid and excavation would be costly. Buttress walls can provide moderate
fixed retaining wall system by a series of rib-walls and cross walls improvement in reducing the wall deflection or ground settlement.
as shown in Fig. 9(a); (3) forming a rigid retaining wall by buttress The mechanism of buttress walls in reducing wall deflections
walls and the cap-slab. Fig. 10 shows the maximum wall deflection mainly come from the frictional resistance between the side surface
at the center of the long side is 165.4 mm, which is slightly larger of buttress wall and adjacent soil when buttress walls demolished
than that of the top-down construction. However, the maximum wall with excavation. When buttress walls are allocated outside the
deflection at the center of the short side is 50. 3 mm, which is much excavation, i.e., outer excavation, or buttress walls inside excavation
small than that of the conventional top-down construction. The but not demolished with excavation, the combined bending stiffness
amount of the wall deflection is related to several factors such as the from both diaphragm and buttress walls is very large, so that the
depth of the cross wall, length of the buttress, and length of the wall can deform linearly as a rigid retaining wall. Combination of
excavation side. Interested readers can refer to the references [7, 8]. cross wall and buttress wall can have good effect in reducing the
wall deflection and ground settlement. With proper arrangement of
C buttress wall or cross wall in excavations can even reach a strut free
L
excavation.
Buttress wall
7. ACKNOWLEDGEMENTS
Cap-slab The authors would like thank Ministry of Science and Technology
A A
(MOST) of Taiwan for granting us research funding for many years.
Soil Excavated The work presented in this paper mostly come from my previous
zone
Master and PhD students, especially from the outstanding work of
Dr. Aswin Lim.
Ribs wall
B B
Diaphragm
wall Cross wall 8. REFERENCES
Ou, C.Y. and Wu, C.H. "Effects of soil grouting on deep excavation
(a) behavior in sandy soil", Journal of the Chinese Institute of
Civil and Hydraulic Engineering, 2(2), pp169-182. [1]
Lbw Hsiao, K.J. (2016) "Use of Equivalent Parameters for Ground
Ground surface C
L Ground surface C
L
Cap-slab Cap-slab Improvement Piles in Deep Excavation Analysis", Master
Diaphragm Buttress wall Buttress wall
Thesis, Department of Civil and Construction Engineering,
wall Diaphragm National Taiwan University of Science and Technology,
wall
Taipei. [2]
He
Hsieh, P.G., Ou, C.Y. & Lin, Y.L. (2013) "Three-dimensional
Final excavation level Final excavation level
numerical analysis of deep excavations with cross walls",
Dcw Cross wall Acta Geotechnica, 8(1), 33-48. [3]
Hsieh, P. G. and Ou, C. Y., and Hsieh, W. H. (2016) "Efficiency of
Hp Cross wall Ribs wall
excavations with buttress walls in reducing the deflection of
the diaphragm wall", Acta Geotechnica, 11, 1087-1102. [4]
Lim, A. (2016) "Investigation of Integrated Buttress and Cross
soil soil Walls to Control Movements Induced by Excavation", PhD
Dissertation, Department of Civil and Construction
(b) Engineering, National Taiwan University of Science and
(c)
Technology, Taipei. [5]
Figure 9 The RFD strut free excavation system
37
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
38
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Behavior of Single Pile and Pile Group Foundation for High Rise Buildings on
Expansive Soils
Paulus P. Rahardjo1 and Prieschila Tamsir2
1
Parahyangan Catholic University, Bandung, Indonesia
2
PT Geotechnical Engineering Consultant, Bandung, Indonesia
E-mail: rahardjo.paulus@gmail.com teknik@gec.co.id
ABSTRACT: High rise buildings supported by piles are now increasingly constructed in Indonesia especially in West Java (at Cikarang and
Cibitung area) and West Surabaya. The cases in West Surabaya is interesting because the depth of the expansive clays is more than 80 m.
The characteristics of expansive clays are very specific due to its capacity of swelling and becoming soft upon absorbing water from their
current unsaturated conditions, and yet as clay material, most designers worry about the long term compression. This paper discusses the
results of observation of the single pile and group piles from design, construction and performance in short term and long term settlement of
a complex of high rise buildings with three towers of 51 stories. The results of soil laboratory tests were disturbed except for water content
and index properties, hence the design has rely on the results of insitu testing including SPT, CPT and Pressuremeter Test (PMT). The
estimated settelement of pile groups of 14-15 cm were in fact much less (only 30%) and the short term settlement dominate almost 90% of
the total settlement. This fact may be related to the swelling characteristics and unsaturated soil condition
Keywords : pile foundation, expansive soils, pile group, settlement
1. INTRODUCTION that rich in volcanic material. Boring results in this area obtained
brownish grey clay, which is welknown as expansive material.
Behavior of single pile and pile groups to support high rise buildings
is always of interest to the engineers, and more specific is because
the buildings are founded on expansive soils. The main objective of
this paper is to review the behavior of a single pile and pile group
under three towers. A case study is presented to illustrate the actual
behavior of single pile and measured settlement of group piles. This LOKASI PROYEK
project is a mixed used building that include Apartment, Hotel,
Office and Mall in west Surabaya with 51 stories of Tower and 7
layers of Podium. Three towers and the podiums were connected to
each other as shown on figure 1, i.e. Tower Anderson, Tower
Benson and tower Lavis (tower 7) which was constructed the last.
The authors have been involved in the design and following
construction and performance of the foundation construction as well
as the subsequent settlement
Tower 7
Figure 2 Geological Condition (Badan Geologi Indonesia)
Tower Anderson
39
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Bor Log
Nsp Nsp Nsp
B1 t B2 t
B3
t
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
0 0 0
2
4
2
4
2
4
NSPT (blows)
0 10 20 30 40 50 60
6 6 6
8 8 8
10 10 10
0
12 12 12
14 14 14
16 16 16
18
20
18
20
18
20
2
22
24
22
24
22
24 4 B1 B2
6
26 26 26
Depth
(m)
Depth
28 28 28
Depth
Depth
(m)
B3 B4
(m)
(m)
30 30
8
30
32 32 32
34 34 34
36 36 36
10
38 38 38
B5 B6
12
40 40 40
42 42 42
14
44 44 44
46 46 46
48 48 48
50
52
50
52
50
52 16
18
54 54 54
56 56 56
20
58 58 58
60 60 60
62 62 62
64 64 64
22
Nsp B4
Nsp
B5
Nsp
t B6 Nsp 24
26
t t t
25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
0 0 0
2
4
2 2
4 28
Depth (m)
4
30
6 6 6
8 8 8
10 10 10
12
14
12
14
12
14
32
16
18
16
18
16
18 34
36
20 20 20
22 22 22
24 24 24
38
Depth
26 26 26
Depth
Depth
(m)
(m)
(m)
28 28 28
30
32
30
32
30
32 40
42
34 34 34
36 36 36
38 38 38
40
42
40
42
40
42 44
44
46
44
46
44
46 46
48
48 48 48
50 50 50
52 52 52
54
56
54
56
54
56 50
52
58 58 58
60 60 60
54
62 62 62
64 64 64
Liquid Limit
Depth (m)
• The fourth layer may be described as hard clay found at All data show that the soil is highly plastic clays or clayey silts
depth 50-60 m and increasing values of N SPT > 30m. having liquid limits of 80-130% and plasticity index of 45-85%.
Those are very high values compared to general soil (figure 7)
The distribution and N SPT profile may be shown on figure 5 and we
can see that the strength is consistently increasing with depth. Water IP vs LL
table is not detected through the whole layer, and if water is found in 100
the borehole, they are trap water or perch water table. This condition 90
60
Plasticity Index
30
Laboratory tests conducted for the project include Index Properties, 20
OH and MH
closer to plastic limits and even in some depth less that the plastic 0 10 20 30 40 50 60 70 80
Liquid Limit(%)
90 100 110 120 130 140
limit, this and that tells why the soil consistency has fall in the stiff Figure 7 : Plasticity of the soils
to hard clays.
40
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Liquidity Index Liquidity Index been based on SPT values and for the stress history and soil
-0.8 -0.6 -0.4 -0.2 0 0.2 stiffness, the data from pressurementers have been used.
0
5 5 5
15 15 15
Problems with Expansive soils are due to the swelling and shrinking 20 20 20
Depth (m)
Depth (m)
25 25 25
if triaxial CU are conducted, the saturation takes very long time and 35 35 35
inclusion of water into the soil may change its behavior, the strength 40 40 40
and deriving parameters for analysis. The foundation design has Figure 11 : Variation of PMT parameters with depth
41
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
25 25 25
Depth (m)
Depth (m)
Depth (m)
40 40 40
the structure loads. However since the buildings are 51 storeys, the 45 45 45
65 65 65
However, the development is very close to the surrounding 4. CONSTRUCTION OF PILE FOUNDATION
neighbour buildings, hence on the pheriphery or at the boundary, the
Figure 15 shows the installation of pile foundation using push
foundation selected is bored piles of φ800 mm diameter. In order to
in/jack in method where the pressure reach 250% of the design load
balance between the driven pile and the boredpiles, both foundation
and held 3 times 10 second at end of installation. This method has
system has been design to carry the same axial stiffness at their
proven to be reliable.
allowable load. The allowable load for driven piles is 200 tons and
based on the bearing capacity analysis the required length of the The problems of jack in method (as well as driven hammer) in
driven spun pile is 34-37 m. While for boredpile φ800 mm the same expansive soils are heaving of the neighbour and the problem of
axial stiffness is 35 m with allowable capacity of 300 ton. In some heave of the piles. The second problem was overcome by redriving,
area higher capacity of 400 tons for bored pile is also required and but we have to make sure that all length of piles can be exactly
calculated to be 43 m length pushed into the soils because the movement of the machine requires
that no portion of the pile shall be on ground level.
For cohesive soils, the general formula use for tip resistance is:
Qp = q ⋅ Ap
where : q = D B 9 cu for D/B < 4
4
q = 9 cu jika D/B > 4
Ap = cross section area of pile
The friction resistance is calculated using adhesion factor as
suggested by Kulhawi and Jackson (1989)
Qs = α ⋅ Su ⋅ p ⋅ L
Where α = adhesion factor.
Su = undrained shear sttrength
p = pheriphery of piles
L = length of embedded pile Figure 15 : Pile installation method
The adhesion factor α have used the recommendation of Kulhawy
dan Jackson (1989).
42
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 16 : Preparation for Pile Injection Figure 20 : Concreting of boredpile with tremie
10
20
Pile Settlement in mm
30
40
60
10
20
Pile Settlement in mm
30
50
60
43
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Settlement Analysis 4
Settlement (cm)
Settlement analysis for this project have used two methods, the first 6
one is based on Poulos and Davis (1980) and the second analysis is 10
1
by Finite Element Method (computer software GTS Midas). The 12 400
piles and computed as follows : Figure 26 Settlement profiles cross section 3-3
xi ⋅ ∑ (Pj ⋅ α ij ) + x1 ⋅ Pi Potongan 4-4'
Si = 0 20 40
Coordinates (m)
60 80 100 120
where : 0
2
Si= settlement of pile i due to own load and other piles 4
Settlement (cm)
xi= settlement of single pile due to unit load (mm/ton) 6
1
Pi= load at pile no -i (ton)
8
400
10
about 15 cm.
70.00
60.00 4
50.00
3
40.00
30.00
20.00
10.00 2 2'
0.00
0.00 10.00 20.00 30.00 40.00
4'50.00 60.00 70.00 80.00 90.00
3' 1'
100.00 110.00 120.00 130.00 140.00 150.00 160.00 170.00 180.00 190.00 200.00 210.00 220.00 230.00
N600 spun 200t D600 BP 200t D800 BP 300t D800 BP 400t Series5 Series6 Tambahan D800 35m
4
Settlement (cm)
6
1
8
375
12
1 1
surrounding
400 800
14
Figure 24 Settlement profiles cross section 1-1 The settlement profile can then be evaluated by looking at cross of
settlement profiles as shown on figure 30. It is shown that the results
Potongan 2-2'
0 50 100
Coordinates (m)
150 200 250
of finite element analysis are in line with the results of approach
0
Tower 7
Tower Tower using Poulos and Davis method.
2 Benson Anderson
1
the plot, and differential settlement are more pronounced at the
location of the tower and the podium
8
625
10
12 1
600
14
44
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
14
1000
Design”, John Wiley, New York
16 Rahardjo, P.P., Junita, Brotodihardjo, W., 1993, “Analisis Pondasi
Jarak vs Settlement Proyek SPI Phase 4
Tiang Tunggal dengan Beban Aksial Berdasarkan Metode
Potongan 2-2 (Potongan Melintang)
Jarak (m)
Pengalihan Beban”, Prosiding Pertemuan Ilmiah Tahunan I,
0
0 20 40 60 80 100 120
HATTI, Jakarta
La Viz Tower
2
4
Settlement (cm)
8
1
10
1 650
400
12
8. CONCLUSION SUMMARY
The study of the behavior of piles in expansive soils show that the
expansive characteristics influence the behavior of single pile and
group piles, in terms of there may be inconsistency on the single pile
behavior due to the construction histories. The predicted settlement
is much less than calculated which may be due to the heave and the
unsaturated soil condition. In general the results of insitu testings are
more reliable for design and analysis.
9. REFERENCES
Deep Foundation Research Institute (DRFI). (2017) Manual Pondasi
Tiang edisi 5, Universitas Katolik Parahyangan
45
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Due to scarcity of land, especially in urban areas, the need for basements to optimise the use of land has resulted in
construction of deep excavation works. Deep excavation works pose great challenges to geotechnical engineers, particularly in complex
ground conditions such as in limestone formation, where it exhibits notorious karstic features with irregular bedrock profiles, variable
weathering condition, cavities and slime zones. With proper geotechnical input, costly failures and delays associated with underground
works such as excessive groundwater lowering, occurrences of sinkholes, excessive ground settlements, etc. can be prevented. Suitable
temporary earth retaining system and rock strengthening works have to be properly designed to prevent such failures. This paper presents
design principles of temporary earth retaining stabilising system (ERSS) together with vertical rock excavation as well as three case studies
including from Mass Rapid Transport (MRT) projects, with various challenging geological formations found within the Klang Valley of
Malaysia; Limestone and Alluvium with high ground water table.
Keywords: Deep excavation; temporary earth retaining stabilising system, ERSS, limestone; Kenny hill, alluvium
46
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
8. Structural design
Determine design value (ULS x 1.0 or SLS x 1.4 or WCS x 1.05 whichever is critical) for wall and supports.
Structural serviceability check.
Structural durability check.
Figure 1 Flowchart for analysis and design of deep excavation works modified after Tan et al. (2016)
c) If
47
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
48
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
parameters would be able to define the entire stiffness each may have their own advantages and disadvantages. For
degradation curve. The values of G 0 can be obtained from example, Hardening Soil model is better suited for non-linear
measurement of shear wave velocity from the following formulation of soil behaviour in general (both soft soils and
relationship: harder soil types) as compared to the Mohr-Coulomb model.
However, the Hardening soil model does not capture creep
G 0 = ρv s 2 (1) (secondary compression behaviour). On the other hand, the Soft
Soil Creep model is able to capture the creep or time dependent
where, ρ is mass density of soil and v s is shear wave velocity of deformations; but it is only suitable for soft soils and not for
soil. other soil types.
Proper understanding of the limitations for each soil models are
In addition to using shear wave measurement, the maximum
essential. The quality of a result from numerical analyses
small strain-stiffness can also be estimated using empirical
depends on the user’s understanding of both the problem and the
correlations. Table 1 presents the typical range for G 0 for several
tool, such as the finite element software packages as well as the
generic soil types. The maximum small-strain shear modulus can
soil models within these software packages. One should be aware
be correlated to the SPT N 60 value and to the CPT q c value as
of the following factors which may affect the results of numerical
follows (Sabatini et al., 2002):
analyses :
G 0 = 15,560 (N 60 )0.68 (2) a) Locations of the boundaries of the problem. The
G 0 = 1,634(q c )0.25(σ’ vo )0.375 (3) problem boundary should be located far enough away
such that there is no stress rotation near the boundary.
where, σ’ vo is effective vertical stress of soil. Corresponding For undrained analyses, the extent of the model
units are all in kPa. required will be greater.
b) Details of mesh. Finer mesh is preferred to coarser
It should be noted that empirical correlations G 0 should be
meshing, especially if high strain gradients are
treated with caution. While it may serve as an initial estimate, it
anticipated, or for failure analyses. Nonetheless, the
is highly recommended to evaluate G 0 from shear wave velocity
refinement of mesh should be carried out selectively at
obtained from relevant field or laboratory tests which represents
critical areas within the deformation mechanism to
the actual site condition.
avoid unnecessarily high computation time.
The shear strain at which the shear modulus G is decayed to c) Long, thin partitions will lead to calculation instability.
0.7G 0 for normally consolidated soils can be calculated form the As such, the layout of the model and mesh should
following equation (Benz, 2007): avoid these closely spaced partitions.
d) Stages of construction. As soils are non-linear, stress
0.385 dependent materials, thus, proper modelling of the soil
�0.7 = [2�′ (1 + �
��2∅′) + ��′(1 + �0 )�
��2∅′] (4) at various stages from the past to its construction stages
4�0
need to be carried out.
The values obtained above should also be checked against values
given by Stokoe et al. (2004), where the authors proposed a e) Modelling of interfaces. Improper modelling or use of
interface reduction factors may lead to dangerously
linear increase of γ 0.7 from γ 0.7 ≈ 1 x 10-4 for PI = 0 up to γ 0.7 ≈ 6
unsafe design.
x 10-4 for PI = 100. Where PI is the plasticity index of soil.
f) Use of suitable constitutive soil models to model
Range of typical G 0 values for various soil types are provided in different geotechnical problems.
Table 1. It can be observed that a large range exists between the
lower and upper bound values for each soil type. Vardanega & g) Sensitivity of various soil parameters. For different
Bolton (2011) have shown that the value of G 0 is dependent on constitutive soil models adopted in different finite
multiple factors including PI, overconsolidation ratio (OCR), element method software packages, different soil
mean effective stress at point of consideration. This reiterates the parameters may have different effects on analyses
point that empirical correlations solely to SPT’N is subjected to results.
limitations.
3. CASE STUDIES
Table 1 Typical values of maximum small-strain shear modulus
(Sabatini et al., 2002) Three case studies including from Mass Rapid Transport (MRT)
projects, with various challenging geological formations found
Soil Type Maximum small-strain within the Klang Valley of Malaysia; Limestone, Kenny Hill
shear modulus, G 0 (kPa) formation and Alluvium with high ground water table are
Soft clays 2,750 to 13,750 discussed in the following sections.
Firm clays 6,900 to 34,500
Silty clays 27,600 to 138,000 3.1 Case Study of Deep Excavation for KVMRT in Kuala
Dense sands 69,000 to 345,000 Lumpur Limestone
and gravels
The Klang Valley Mass Rapid Transit (KVMRT) from Sg. Buloh
to Kajang (SBK Line) is the first MRT project in Malaysia. The
project comprises of a total of 9.5km long twin tunnels from
2.2 Important Considerations in Numerical Analyses Semantan to Maluri with seven (7) underground stations and
associated structures such as portals, ventilation shafts, escape
The use of numerical analyses such as finite element method for shafts and crossovers to be constructed over the Klang Valley
simulating deep excavation is common and routinely used by and Kuala Lumpur city areas. Tun Razak Exchange (TRX)
engineers of different levels of experience and expertise. Various Station (known as Pasar Rakyat Station during design
constitutive soil models such as Mohr-Coulomb, Hardening Soil, development), Cochrane Station and Maluri Station are
Soft Soil Creep, etc. are commonly used in practice. The underground stations located in the city area with excavation
characteristics of each soil model varies from one another where
49
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
50
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Table 2 Overlapping length of secant pile wall scenario (e.g. an accident during lifting operations, etc.). It can be
appreciated that this depends greatly on the anticipated so-called
Pile Pile Length Pile Length Pile Length accidental load conditions. The above-mentioned accidental
Diameter <8m <15m <25m impact load was recommended by CIRIA C517 (1999) but site-
(mm) (mm) (mm) (mm) specific risks and hazards may dictate different loads.
880 130 170 -
1000 150 200 340 Table 3 Partial load factors
1180 170 230 360
Load Case EL DL LL TL IL
1500 225 260 380
Working Condition 1.5 1.4 1.6 1.2 NA
Accidental Impact 1.05 1.05 0.5 NA 1.05
All secant piles were founded on competent bedrock with
One-strut Failure 1.05 1.05 0.5 NA NA
minimum rock socket length of 1.5m to 4.0m. The termination
Note:
criteria for rock socket are based on coring in competent bedrock
EL – Earth pressure and groundwater
with verification of point load index strength, I s(50) > 4 MPa
DL – Dead load
(equivalent to average UCS of 44 MPa). Should I s(50) falls below
LL – Live load
4 MPa, an equivalent rock socket length can be calculated to
TL – Temperature effect
compensate for the lower rock strength.
IL – Accidental impact load
Given that the rock head level in limestone formation can be NA – Not applicable
erratic with very steep inclined rock surface, one of the
construction issues that typical arises are on the length of rock
If the strutting design was found to be inadequate for the
socket. To prevent any disputes, the authors recommend the
considered impact load, then the strut is expected to fail resulting
following criteria where it must all be met before it is considered
in a one strut failure condition. The spirit of the one strut failure
as rock for claiming purposes:
as stated in CIRIA C760 (2017) refers to the potential loss of
(a) Change of tools from that of soil to rock support from a single prop rather than the loss of an entire level
coring/excavation tools. of props, unless it is considered as a physical possibility. The
latter is typically simulated in a plane-strain 2D analyses where a
(b) Rock head level is taken as a flat rock surface (i.e. not single prop level is deactivated instead of reducing the stiffness
inclined) that is verified at multiple points by means of of the props; accounting for the larger span between struts.
Kelly bar, etc.
It needs to be recognised that a one-strut failure is in fact a 3D
(c) I s(50) should be > 1.0 MPa problem. For an intermediate strut level, a single strut failure
It is important to ensure that the retaining wall is socketed into would cause the stresses to be distributed to the closest supports;
competent bedrock as the vertical rock excavation is just 1.25m adjacent struts to its sides and to struts above and below it. The
away from the retaining wall alignment. Support system will be redistribution of load is governed by the vertical and horizontal
installed in stages until reaching the bedrock level. A row of tie- stiffness of the wall, and to a lesser extent, the walers. This is
back rock bolts was installed above the bedrock level to enhance because the pressure acting on the wall must first be taken by the
wall toe stability. Toe stability check was carried out with some walls before it is transferred to the waler. A weaker wall could
modification which replaces passive resistance by tie-back force fail even if the walers do not.
to achieve minimum safety factor of 1.2. In addition, vertical Stiffness of the wall in each direction depends on the
stability was checked with resultant vertical load from ground reinforcement design and the spacing of supports. Since most
anchor pre-stress against the rock socket length. walls are constructed vertically with the main reinforcement in
Excavation was carried out in stages facilitated by installing the vertical direction coupled with generally smaller vertical
temporary ground anchors. U-turn ground anchors were used as spacings between levels in comparison to the horizontal spacings,
local authority requires removal of ground anchors after vertical stiffness of the wall would typically be dominant. Hence
construction. The anchor consists of a few pairs of strands with most of the load would be transferred vertically; to the struts
different unit lengths. Adopted strand diameter is 15.24mm with above and below.
U-turn radius of 47.5mm. Proof tests were carried out prior to the If a 3D analysis is not carried out, a conservative approach to
working anchor installation for design verification. Based on the cater for this is to ensure full distribution of loads vertically. This
proof test results, the recommended reduction factor due to is done by removing the strut completely in a 2D analyses. As
bending of strand at U-turn point is 0.65. Working loads of long as the wall stresses are still within limits for the
anchor ranged from 212kN to 1060kN with 2 to 10 strands. reinforcements (i.e. FOS 1.05), the resulting strut forces would
Typical designed pre-stress load is 60-80% of the working load then be conservative. In contrast, the force distribution would be
capacity. Generally, the anchor will be locked off at 110% of less representative if the stiffness of the struts (intermediate strut
designed pre-stress load. All anchors are subjected to acceptance layer) is reduced to account for a larger span for one strut failure.
test up to 125% of working load before lock-off. It is important
to clearly define in construction drawing the anchor working The only exception for this is with regards to the top-level strut
load, pre-stress load and lock-off load to prevent where it’s failure during the initial stages (i.e. when only the
misunderstanding and confusion during construction works. single strut is installed) results in a cantilever condition.
Designing for the wall to survive a cantilever condition with
The design criteria considered in strutting design are earth relatively lower risks at these shallow excavation depths would
pressures and groundwater, material dead load, 1.5 kN/m live be less feasible.
load, eccentric load, temperature effect (changes of 10°C),
accidental impact load (50kN in vertical direction; 10kN in To comply for a single strut failure condition, propping stiffness
horizontal direction), and one-strut failure. Recommended partial is reduced and waler effective length is increased accordingly.
load factors for strutting design are shown in Table 3. This is in-line with the concept proposed by CIRIA C760 (2017).
Ultimately it would result heavier design for the walers at this top
The requirement to consider accidental and one strut failure load level as it must be sufficiently stiff to redistribute the loadings
case is often a subject to debate. Nonetheless, the general idea is horizontally as it remains as the only propped level during this
to provide for a robust design which can prevent catastrophic or initial stage.
progressive failure in the event of a severe but transient load case
51
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The loss of a strut single strut would effectively increase the grouting may be required after reviewing the grout intake from
loadings by 50% for the immediate adjacent struts. This would primary grouting. Rock fissure grouting is also adopted for base
be mostly covered by the partial factors of about 1.5 for the grouting at larger grout hole spacing. If any cavities are detected
working condition. Even if the strut was designed to have 100% during drilling or grouting works, compaction grouting with
efficiency in the basic load case, it would only need marginal cement mortar will be used as cavity treatment. It should be
increase its design to comply with the lower 1.05 partial factor noted that grouting techniques rely much on local experiences
for one-strut failure. and contractor workmanship. The recommended holding
pressures for fissure grouting in limestone in this case study are
However, the same cannot be said for walers. This is because the shown in Table 4.
resulting bending moment of a waler is related to the square of its
effective length. This essentially means that the effective span of Table 4 Holding pressure for fissure grouting
the waler doubles during a single strut failure, causing the
bending moment to increase by 400% as opposed to the 50% Depth (m) Holding pressure
increase in load for struts. The resulting waler design to comply (Bar)
with single strut failure would thus be significantly more 0 to 10 2 to 4
conservative. 10 to 20 6 to 8
20 to 30 10 to 12
In general situations, it can be inferred that a one-strut failure 30 to 40 14 to 16
load case would usually be more critical than the accidental load 40 to 50 18 to 20
case. Thus, if it is considered, accidental impact condition is no > 50 >22
longer relevant. The authors opine that if the adopted accidental Note: Termination criteria shall be satisfied with flow rate less
impact loadings are representative of a credible worst-case than 2 litres per minute or grout volume reaches 10m3 for every
scenario, then the risk of a single strut failure is as low as grouting zone in 5m depth.
reasonably practical (ALARP), where designing for one-strut
failure is not required. The Maluri station is located underneath one of the major public
roads in town. Besides the challenging limestone bedrock, major
The above discussion excludes the scenario where specific risks utilities diversion (e.g. high voltage electrical cables) and traffic
dictates very high accidental impact loads where it becomes less diversion were carried out in stages and carefully planned for
feasible to design the strutting to withstand it. In these cases, a installation of the secant pile wall. Temporary road decking with
one-strut failure check and design would be more appropriate. deck posts were installed concurrently with secant pile wall
Nonetheless for the KVMRT project, it was a client’s installation. Approximately 300m long and 21m wide road
requirement to check for both load conditions. decking covered up the top of the station and crossover area
It is essential to note that groundwater control is essential and it where excavation works beneath are executed concurrently.
is one of the important criteria to be considered in excavation One of the construction difficulties was pile installation under
works. Groundwater drawdown may lead to excessive ground existing electrical transmission lines with safe allowable working
settlement and occurrences of sinkholes surrounding the head room of only 13m. A modified low headroom machine was
excavation. Potential risk of excessive groundwater ingress into used for secant pile installation. In this condition, limit of drilling
excavation pit shall be evaluated especially in limestone size to small diameter is required to fulfil the capacity of the
formation. Natural features of solution channel with cavities and modified machine. Deck posts installation required high capacity
highly fractured limestone connected to excavation pit may cause rig with deep rock drilling which is beyond the typical machine
disastrous flooding inside the excavation pit. Therefore, grouting capacity. Deck posts are supported by four micropiles in a group
in limestone was carried out as a risk mitigation measure for to support the underneath existing electrical transmission line.
groundwater control. Schematic of the excavation works is Figure 11 shows the base slab casting at Maluri station with live
shown in Figure 10. Note that rock slope strengthening indicated traffic on top of the temporary road decking while Figure 12
in Figure 10 is provisional only. The actual locations and extent shows the excavation works with strutting support at the Maluri
of rock slope strengthening are determined after geological Portal. Vibrations from close proximity rock blasting (< 3m clear
mapping works and kinematic analyses. distance) from the deck posts were successfully controlled below
the conservative 25mm/s limit adopted for reinforced concrete
frame structures.
52
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Original Design
53
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Remedial works were carried by installing an additional row of Tan, Y.C. & Chow. C.M. (2008) “Design of Retaining Wall and
sheet piles of about 30m deep, penetrating into the relatively Support Systems for Deep Basement Construction – A
impermeable clayey soil underlying the sandy soils to serve as a Malaysian Experience” Seminar on Excavation and
hydraulic cut off in conjunction with recharging wells. These Retaining Walls by IEM-HKIE. Petaling Jaya, Malaysia.
have effectively restored the water table to its original level. Tan, Y. C., Chow, C. M., Koo, K.S., and Nazir, R. (2016)
Further basement excavation was allowed to be resumed after the "Challenges in Design and Construction of Deep
monitoring confirmed the effectiveness of the remedial measures. Excavation for KVMRT in Kuala Lumpur Limestone
Formation", Jurnal Teknologi (Sciences & Engineering)
4. LESSONS LEARNT October 78:8-5 (2016), pp97-107.
The lessons learnt based on the two case studies above are Twine, D & Roscoe, H. (1999) Temporary propping of deep
summarised below: excavation- guidance on design (CIRIA C517).
Construction Industry Research and Information
(a) Sufficient and detailed SI is required to identify Association (CIRIA), London.
representative soil parameters, particularly in limestone Sabatini, P.J., Bachus, R.C., Mayne, P.W., Schneider, J.A. &
formation where rock level is very erratic in nature. Zettler, T.E. (2002) Geotechnical Engineering Circular No.
(b) Rock probing is crucial for the identification of any 5 – Evaluation of Soil and Rock Properties. Federal
inclined rock levels, potential rock fissures and cavities Highway Administration (FHWA), U.S. Department of
which can be dealt with via grouting at an early stage. Transportation.
(c) Rock head level and socketing criteria should be stated Vardanega, P.J. & Bolton, M.D. (2011) Strength mobilisation in
clearly in construction drawings with prevent disputes. clays and silts. Canadian Geotechnical Journal Vol. 48, pp.
(d) Design of strutting should at least account for 1485-1503.
accidental load impacts if not one-strut failure, where
deemed necessary.
(e) A single strut failure would have significant cost
implications particularly on the waler designs due to its
increased effective span. Therefore, it is possible for a
single layer strut failure approach to yield cheaper
overall design without changing the waler span. This is
not applicable for first layer struts where its failure
would imply a cantilever wall condition.
(f) Deep excavation can be carried out simultaneously
with live traffic on top of the temporary road decking
above the excavation.
(g) Groundwater drawdown induced building settlement
and damages could be as far away as 300m from the
excavation, particularly with soft soils such as alluvium
with interbedded sand layers. Proper design of
hydraulic cutoff is needed to prevent excessive drop in
groundwater table.
6. ACKNOWLEDGEMENT
Many parts of the work in this paper, including the case studies
were carried out together with the following colleagues, Ir. Dr.
TAN Yean Chin, Ir. CHOW Chee Meng, Ir. KOO Kuan Seng
and some others from G&P Geotechnics Sdn Bhd.
7. REFERENCES
Atkinson, J.H. (2000) Non-linear soil stiffness in routine design
(40th Rankine Lecture). Geotechnique Vol. 50(5):487-508.
Benz, T. (2007) Small-Strain Stiffness of Soils and its Numerical
Consequences. Universität Stuttgart, Germany.
British Standards Institution. (1986) BS8004:1986 Code of
Practice for Foundations.
Clayton, C.R.I., Matthews, M.C. & Simons, N.E. (1995) Soil
Investigation (2nd Edition). Blackwell Science, Oxford.
Gaba, A.R., Simpson, B., Powrie, W. & Beadman, D.R. (2003)
Embedded retaining walls – guidance for economic design
(CIRIA C580). Construction Industry Research and
Information Association (CIRIA), London.
Gaba, A.R., Hardy, S., Doughty, L., Powrie, W. & Selemetas, D.
(2017) Guidance on embedded retaining wall design
(CIRIA C760). Construction Industry Research and
Information Association (CIRIA), London.
Geotechnical Engineering Office (2017) GEOGUIDE 2: Guide to
Site Investigation.
Neoh, C.A. (1998) “Design & construction of pile foundations in
limestone formation”. Journal – Institution of Engineers,
Malaysia. Vol. 59, No. 1, pp. 23 – 38.
54
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: The 3rd phase MRT underground project of Bangkok that is being implemented has faced increased difficulties in design and
construction owing to the congested corridor along the alignment and the phenomenon of groundwater rebound in Bangkok subsoils in the
aftermath of the end of land subsidence era from deep well pumping. The rebound is more than 10 meters from the situation during the
initial phase construction and thus it necessitates careful soil investigation and design approach to prevent instability of groundwater ingress
and uplift in bore tunnelling and excavation of station boxes. The tight space along the MRT underground alignment resulted in the twin
bored tunnels being placed at very small clearance in a few locations. It creates high risk in the shield driving, particularly in the saturated
sand layer. The works also involve tunnel driving at close proximity or under a number of existing buildings or road viaducts. In such cases,
various schemes of underpinning are employed, and one of them is strengthening of founding soil layers by cement jet grouting. An
intensive instrumentation program according to the observational design method is employed to safe guard against occurrence of any
undesirable instability in the excavations and impacts to the third party properties.
Keywords: MRT Underground, shield tunneling, deep excavation, underpinning, groundwater rebound, the observational method
55
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ideal condition for tunnelling. However, the soft soil condition Variation of Piezometric Pressure Profile with
necessitates the use of shield method for tunnel construction. The Elevation
(Standpipe Piezometer)
shallow zone consists of a 12-15 m thick layer of soft to medium
stiff clay (Bangkok clay) that is followed by a layer of stiff to hard Short Term Design Piezometric Profile
clay and a layer of sand (the First Stiff Clay and the First Sand). 100 Long Term Design Piezometric Profile
Below them, alternating layers of stiff to hard clay and dense to very Existing Tender Orange Line (2005)
dense sand exist to a great depth. The sand layers are aquifers of Additional Tender Orange Line (2015)
Contract E1 Contract E2
70
Elevation, (m asl)
60
50
40
30
It is well known that Bangkok had suffered land subsidence Figure 1 - Variation of Piezometric Profile with Depth
problem owing to excessive deep well pumping from the sand Figure 3 Piezometric pressure in the subsoils at different times
aquifers for water supply. The decline in groundwater head in
shallow aquifer layers was as much as 25-35 m from the ground
surface during the peak of the crisis in early 1980s, when the
subsidence occurred at a rate as high as 120 mm/year. After the
surface water supply was made adequate in most of the city area
around the turn of the century, the situation improved as deep well
pumping diminished. Since then the rebound of groundwater started
to develop and the subsidence ceased in most city areas. The
rebound continues till present and the accumulative amount is about
10-12 m during the last 15 years. At present the piezometric water
head in the 1st and 2nd sand layers of Bangkok where the MRT
underground works are involved with is about 13-14 m depth below
the ground surface that is 10-12 m higher than the condition when
the 1st Phase MRT Underground line was constructed (Figure 3). It
has been risen at a rate of approximately 0.8-1.0 m per year (Figure
4). Because the groundwater law that has been in effect since 2004
totally bans deep well pumping in the entire area of Bangkok and Figure 4 Changes in piezometric heads in Bangkok sand layers
surrounding suburbs, it is foreseen that the ground water heads in
the sand layers might rebound to the hydrostatic level in the near 4. INCIDENTS IN PAST MRT UNDERGROUND WORKS
future.
In the past the subsidence situation created complication of long Problematic incidents related to the impact of groundwater rebound
term settlement and differential settlement in the design of MRT in the underground works of the past MRT project are presented.
underground structures. On the other hand, the condition of the The recently completed underground structures of the Blue Line
lowered groundwater head in the sand layers resulted in ease in deep Extension were placed in the stiff clay and sand layers as shown in
excavations for station construction and shield tunnelling. Incidents Figure 5. The deepest bored tunnel section is at 37 m depth at tunnel
of hydraulic instability were rare in past deep excavations and invert at a location near the river crossing.
tunnelling. However, at present the situation has changed since the
groundwater head started to show significant rebound about 15
years ago. Difficult and problematic incidents were experienced in
a number of recent deep excavations of building construction as well
as in the works of the Blue Line Extension Project. In addition, the
rise of groundwater head was linked to incidents of difficulties in
EPB shield tunnelling in the city including a severe incident in the
2nd phase MRT Underground project that led to a six-month delay of
the stall and repair of the TBM.
56
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
of the soil with the side of the diaphragm wall. The U-shaped
Facing the situation of higher water pressure in the 1st and 2nd grouted zone as shown in Figure 7 was adopted. The grouting work
sand layers than that indicated by early investigation, the detailed was launched underneath the ground surface from a platform at the
design of two of the three station boxes that were placed at large mid-height of the station height. The scheme was proven to be an
depths needed to be modified from those planned in the pre-tendered effective method to prevent the hydraulic uplift failure of the
design. This is to avoid problems of base instability during excavation base and it allowed a normal construction process of the
excavation from hydraulic uplift of the stiff clay at the excavation concrete base slab casting under a dry condition.
base. The excavation procedure adopted for the third station
remained the normal approach as the designer considered it was still
adequate. Unfortunately, after a portion of the base area was
excavated and concrete casting, the hydraulic uplift occurred in the
next stage of the base excavation that resulted in flooding and
difficulty in control of water ingress in the remaining stages of the
excavation. The work suffered almost 20 months delay in the
construction to fight with the persistent problem of groundwater
ingress despite extensive remedial grouting exercise. A brief outline
on the situations of hydraulic instability involved with the base
excavation design and construction of these stations is given in the
following sections.
The second station was Station BS13 which was 181 m long, 23.4
m wide and 30 m deep. The diaphragm wall was 1.2 m thick and 40
m deep. Cross section and subsoil is shown Figure 9. The
+101.5
piezometric head was observed at approximately 16.0 m below the
ground surface.
Soft Clay
GWL+86.3 +88.0
st
1 Stiff Clay
+77.5
1st Sand
+65.0
2nd Stiff Clay +63.0
2nd Sand
57
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
stages of a small sized excavation zone bounded by high soil BLE project that the EPB shield tunnelling had to be made through
surcharge + cast concrete slab around the perimeter of the the sand layer as deep as 37 m below the ground surface with the
excavation, the stability of the excavated base against the hydraulic piezometric head 10 m higher than that existed during the time of
uplift will be significantly assisted by the side shear resistance of the the first MRT shield tunnelling, an incident of serious water ingress
soil along the perimeter as compared with that of large excavation occurred in August 2013 during the early drive of an EPB shield
area. The Outline Design Specification stipulated at that in case through the interface between the 1st stiff clay and the 1st sand layer.
shear strength of base soil was considered, a minimum factor of The consequence was flooding and severe damage to the TBM that
safety was required at 1.20. This requirement could be satisfied for led to an interruption of the work for over 7 months. The incident
the soil and groundwater condition below the excavation base of the was an uncontrolled water ingress into the tunnel in the tail area of
station. The accurate determination of the density of the soil layers the EPB shield machine. Flooding and excessive movements of
and the spatial variation of the thickness of the stiff clay layer were segmental rings in vicinity of the shield were experienced. It was
a must in the adaptation of this method. The staged excavation reported the ingress initially occurred through a 20 mm wide gap
sequence adopted for BS13 excavation is shown in Figure 10. The formed at the circumferential joint of the key segment placed near
work was successfully completed as planned. the invert of the ring. The groundwater ingress from the 1st Sand
layer shortly led to piping of fine sand into the tunnel, and
subsequently excessive movements and distortion of several
segmental rings behind the shield. During the incident various
remedial grouting measures were urgently carried out to control the
ingress and avoid excessive distortion of the concrete segmental
rings that might ultimately result in a tunnel collapse.
It was obvious that the incident was caused by difficulty of the
EPB shield in controlling stability of the excavated tunnel face while
the tunnel was being excavated in the mixed face stiff clay and sand
layers. At that position the groundwater head in the sand layer was
17 m. The sand was in the lower part of the tunnel and as the
tunnel was progressively advanced, the portion of sand in face area
progressively increased as the TBM drive pitched down. This
condition was one of the most difficult tunnel grounds to deal with
in soft ground tunnelling. It was apparently that with high ground
water pressure in the sand layer in the excavated tunnel face, the
shield machine experienced difficulty in maintaining adequate face
Fig. 10 Sequence of staged base excavation of BS13 pressures. Thus over excavation would significantly occur. After a
careful examination on the TBM operation data it was believed that
Successful base excavation of the BLE Underground station was the triggering factor of the incident was an excessive displacement
not the case at BS11 Station which was the deepest excavation of the key segment of the segmental ring resulted from the over
among all MRT stations in Bangkok. The base of the excavation excavation of the mixed face soil by the TBM under the high
was at 32 m below the ground surface. Prior to the start of the groundwater condition.
excavation it was judged by the designer that the subsoil and
piezometric head in the sand layers data indicated that the base 5. ORANGE LINE EAST
excavation could be with adequate factor of safety against hydraulic
uplift. Unfortunately, after the excavation was made to the lower The 3rd Phase MRT Underground work that is under construction at
level of the bracing of the D-wall and base excavation and concrete this stage is a part of the Orange Line East project (Cultural Center
base slab casting was progressed from one end of the station box to to Minburi) that is implemented as design-build contracts. The
about 15 percent of the length of station, significant water ingress underground portion constitutes 12.1 km length out of the total 21.1
occurred that seriously affected the work. The whole station was km length of the line. There are 10 underground stations along the
flooded by 2-m-depth of water within 12 hours. Extensive underground section that the works are administered in 3 contracts
dewatering pump system was installed and a comprehensive (E1-E3). The underground line is mainly constructed by bored
program of chemical and cement grouting subsequently employed. tunneling using 3 EPB shields except for a short cut-and-cover
The amount of ingress was as high as 1400 cubic meter/hour at one startup section. Owing to the tight space along the corridor of the
stage. Groundwater ingress repeatedly occurred in the subsequent alignment of the line that traverse along two most congested roads
stages of excavation for concrete casting that altogether delayed the of the city that also have foundation of elevated roadway along the
work by 20 months. The investigation pointed that there might be a right or way, the MRT underground twin tunnels are mostly placed
condition of lateral non-continuity of the 2nd stiff clay below the in the vertically stacked configurations and in many places they
excavation base that in the initial design was considered sufficiently need to be placed at very close proximity to each other, particularly
thick and covering the entire area of the excavation base. This when leaving or entering the station boxes (Figures 2 and 11). The
incident of severe and prolonged groundwater ingress also led to tunnels also pass very near to or cross foundations of existing
subsidence problem of the surrounding ground and buildings that buildings or other structures, some of which need underpinning
necessitated extensive stabilizing grouting works throughout the works. The close proximity shield tunneling works for the deeper
period. section of alignment in the sand layer of high groundwater pressure
for the shallow section in soft clay layer necessitate special
4.2 Water ingress in EPB shield tunnelling considerations in the design and construction planning to avert
instability problems during construction and to ensure stability of
Prior to the tunnelling in the Blue Line Extension project, bored the completed structures in long term. Because of the adoption of
tunnelling by means of EPB shields and concrete segmented ring the vertically stacked configuration of the running tunnels most of
support had been commonly used with great success including those the underground stations are excavated at large depth. And with the
in the Blue Line Initial System project, in which the tunnelling was very extraordinary thickness of the 1st sand layer in the subsoils in
made deep in the sand layer at depth as much as 25 m below the the considerable part of the alignment, careful considerations are
ground surface. There had been no incidents of groundwater needed for the design and planning of the station excavation to avert
problem in EPB shield driving in Bangkok sands. However, in the potential basal instability problems as earlier experienced in the
previous project.
58
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
OR14
OR15
Tunnel Profile-Contract E1
Figure 11 Vertically stacked alignment of tunnels in Contract E1
The design of the bored tunnelling and concrete segment lining has
to deal with the situation of very close proximity placement of the
twin tunnels in various places. The clear distance between the
tunnels is as close as 1.6 m which is only 25% of the tunnel
diameter (e.g. Figure 12). For such conditions where the tunnels are
to be excavated in the 1st sand layer, special measures are provided
to avert potential instability problems to the lining of the firstly
excavated tunnel from the TBM weight and any unexpected
excessive over-excavation during the excavation of the second
tunnel. The protective measures generally consist of ground
improvement by jet cement grouting. Where the ground space does
not permit the launching of ground improvement, other measures are
used. They include installation of steel ring bracing inside the
concrete segmental lining of the first driven tunnel, the careful
driving of the shield, and the close monitoring of the movements of
the lining and ground.
1.6 m
Close proximity tunneling of the works also need to deal with the
situation of potential impacts to the piled foundation of the existing
structures and vice versa the impact of the load to the stability of the
tunnel lining. Extensive numerical analyses are employed to check
the situation of all cases in details and necessary preventive and
corrective measures are provided in the design. In some cases, the
excavation may be made without the need for auxiliary measures,
but in some cases, ground improvement by jet cement grouting is
adopted (Figure 13). In the worst case, underpinning with pile
replacement needed adopted.
Besides the cases of close proximity tunneling, there is a case of
the station box excavation placed very close to the piers of the
elevated expressway at a clear distance between the diaphragm wall
and the foundation piles of the pier as small as 2.7 m. In this case
the impact assessment with the aid of rigorous numerical analysis
(PLAXIS) suggests implementation of the mitigation measure to Figure 14 Pre-installed struts for station box excavation next to
control the lateral movement of the viaduct pier and induced flexural expressway viaducts
stress in the piles within the allowable limits by using preinstalled
struts for the station excavation at two levels by means of jet cement
59
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
60
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Singapore is the world’s second busiest port with an average of 60 vessels calling daily. As a testament to Singapore’s
commitment to maintain its position as a global maritime nation, the government is going full steam ahead with the multi-billion dollars
mega-port terminal development at Tuas. The Tuas Terminal will be developed in four phases over 30 years. Phase 1 reclamation of 294
hectares began in 2016 while the works contract for Phase 2 reclamation of 387 hectares was recently awarded in February 2018. When this
entire development is completed by 2040, it is expected to be the largest container terminal in the world. For the land-scarce small city-state,
creating new land by reclaiming it from the sea has played an important role in the country’s maritime industry. Its first reclamation project
for terminal development was in 1967. This was followed by a string of reclamation projects including Tanjong Pagar Terminal (80 hectares),
Keppel Terminal (105 hectares), Brani Terminal (84 hectares) and Pasir Panjang Terminal (456 hectares). While it continues to develop and
expand its maritime industry, it faces serious challenges to land reclamation due to limited availability of sand, among others. As sand is not
readily available in large quantity, it calls for innovative solutions to overcome material constraint. Prior to Tuas, in another mega terminal
development which was completed in 2015, the use of alternative reclamation fill materials was initiated to replace sand. Dredged materials
from deepening of basins and nearby fairway as well as excavated earth from land construction projects were reused as reclamation fill
materials. The use of these alternative materials, which would otherwise be disposed of, managed to save 45% in sand volume and saved
SGD470 million construction cost. Also, it reduced the need for disposal grounds for the dredged and excavated materials. Needless to say,
the innovations first deployed in this development are now being implemented to a greater degree in the upcoming Tuas Terminal. The
concept and application of the innovations used in this earlier terminal development are presented as a case study in this paper. This paper
describes some of the soil improvement techniques and acceptance tests used and it highlights the challenges and lessons learned.
Keywords: Land reclamation, Vibro compaction, Dynamic compaction, Dynamic replacement, Pressuremeter test, Cone penetration test.
61
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
But there are serious challenges to land reclamation for the small
city-state – the availability of sand for reclamation. The problem is
Singapore has used so much sand that it has run out of its own, and
need to import sand from elsewhere to meet its massive land
reclamation needs. Skyrocketing price of imported sand just added
fuel to the fire. The situation got worst when Malaysia banned the
export of sand to Singapore in 1997; followed by Indonesia in 2007
and Vietnam in 2009. In 2016, Cambodia enacted a ban against
exporting sand to Singapore. Hence, aggressive land reclamation Figure 4 Aerial view of the new terminal development
programme using sand is no longer tenable. It calls for innovative
solutions to overcome material constraint. Construction works began in 2007. It showcased several
In 2015, a mega terminal development was completed using innovative engineering feats that were recognised for engineering
several innovative engineering solutions including the use of achievements thought to have the greatest economic, infrastructural
alternative reclamation fill materials to replace sand. Dredged and societal impact. In 2016, it was voted by the public as being one
materials such as marine clay from deepening of basins and nearby of Singapore’s Top 50 Engineering Achievements – making it a
fairway as well as excavated earth from land construction projects construction model for the upcoming Tuas Terminal.
(e.g. roads, MRT, etc.) were reused as reclamation fill materials Some of the innovative engineering solutions used during
(Figure 3). The use of these alternative materials, which would reclamation works include the use of large concrete blocks (caissons)
otherwise be disposed of, managed to save 45% in sand volume and to build the seawall and wharf structure instead of traditional piling
62
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
X
X
Figure 7 Usage of different reclamation fill materials
63
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
After completion of the foundation works, acceptance tests were The results revealed a layer of “softened” sensitive soil of about
carried out. Upon achieving the acceptance criteria, the caissons 1.5m thick at the base of sandkey trench and dredged seabed. To
were installed, backfilled with sand and compacted. supplement the CPT tests, pressuremeter (PMT) tests were carried
out on jack-up pontoon at 15 different locations. The tests were
carried out in the sand fill and clay layer down to 33m or more
below water. A summary of the PMT results is shown in Table 1.
The EP/PL ratios of the disturbed materials (i.e. disturbed sand /
softened clay) were generally less than 7 indicating an altered or
disturbed state while the undisturbed stiff clay shows a ratio greater
than 16 indicating an over-consolidated state (D.60.AN, 1975).
Table 2 shows the estimated strength values – internal friction
angle () and undrained shear strength (cu) of the materials based on
PMT results of Table 1. The internal angle of friction () of 30o to
32o for disturbed sand was less than the design requirement of 35o
while undrained shear strength (cu) of 50 to 80 kN/m2 for softened
clay was also less than the design requirement of 250 kN/m2. Only
DREDGED LINE
the undisturbed stiff clay layer registered undrained shear strength
(cu) greater than the design requirement. These values of ɸ and cu
Figure 9 Cross-section of caisson on sandkey trench were estimated from the equations below (Amar et al, 1991):
PL *
The performance requirement for the foundation of the caisson cu (1)
seawalls was a residual settlement after construction not exceeding
5.5
40mm. To achieve this requirement, the construction of the sandkey φ o 24o
trench called for complete removal of the underlying soft to firm * 4
clay until very stiff clay strata (undrained shear strength (cu) of not PL 2.5* 2 (2)
less than 250 kN/m2) and total replacement with clean sand. The where PL* is the net limit pressure (in bars) which is equal to the
sand fill with fine content (soil particles finer than 75m) less than difference between the measured limit pressure (PL) and the
10% was then compacted to achieve a relative density of minimum horizontal pressure at rest.
70%. The design requirement was an internal angle of friction (ɸ)
greater than 35o and the acceptance criteria based on cone Table 1 Results of PMT tests after vibro compaction of sandkey
penetration tests (CPT) were cone resistance (qc) of 8 MPa at the
surface (i.e. top finished level of sandkey trench) increasing to 22 EP EY PL
Material EP/PL
MPa at depth greater than 20m. (MPa) (MPa) (MPa)
Deep vibro compaction was carried out to compact the sand fill. Disturbed
2.6 – 3.3 7.8 – 9.9 0.70 – 0.72 3.8 – 4.4
Vibroflots with maximum centrifugal force of 35.4 tons and sand
amplitude of 26mm at the tip were used. The compaction was Softened
0.7 – 1.2 1.4 – 2.4 0.49 – 0.51 1.5 – 2.3
carried out with a 3m overburden sand fill placed above the top clay
finished level of the sandkey trench. The compaction spacing was Undisturbed
36 – 39 36 – 39 2.05 – 2.26 17 – 18
3.5m triangular grid with compaction duration of 40 seconds at stiff clay
every 0.5m vertical lift increments.
After completion of vibro compaction and sand trimming, rocks Table 2 Estimated and cu after vibro compaction of sandkey
of maximum 300mm size were placed to form mound ranging from Material EP/PL ɸ (o ) cu (kN/m2)
1 to 4m thick. The rock pieces were compacted to achieve design
requirement of an internal angle of friction (ɸ) greater than 45o. Disturbed sand 3.8 – 4.4 30 – 32
Softened clay 1.5 – 2.3 50 – 80
3.3 Unexpected Soil Softening Undisturbed stiff clay 17 – 18 > 250
Prior to the removal of the overburden sand fill after vibro Note: EP is the pressuremeter modulus; EY is the stiffness modulus;
compaction and before placing of the rock mound, CPT tests were PL is the limit pressure; ɸ is the internal angle of friction and cu is
carried out as routine acceptance tests. On a particular area at water the undrained shear strength.
depth of about 30m, non-compliance results (i.e. measured cone
resistances lower than acceptance values) were recorded. Figure 10 3.4 Remedial Solution
shows a set of non-compliance CPT results after compaction.
From the CPT results, the total thickness of disturbed sand and
softened clay layer ranged from 80 – 150cm. Based on the measured
“c-” values, a re-analysis was carried out for caisson seawall
settlement and stability. Figure 11 (left) shows the original design
assumptions (design requirements) compared to the measured field
values in Figure 11 (center). To meet the performance requirement,
it was necessary to reinstate (“improve”) the measured field strength
values back to the original design assumptions or equivalent. Figure
11 (right) shows the “target” revised design requirements.
A remedial solution consisted of a rock mat over reinforcing
rock columns was adopted. Rock columns were introduced to
Dredged Line improve the composite strength properties (c-of the softened clay
while the compacted rock mat was introduced to increase the
internal angle of friction () of the disturbed sand. It also served as a
load distribution layer to the reinforcing columns below. Dynamic
replacement was carried out to install the rock columns and dynamic
Figure 10 CPT results after vibro compaction of sandkey compaction was carried out to compact the rock mat.
64
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
D n. W.H (3)
Fz
F2 F1
D 2
Z Z
o
2
F1 (4)
3.5 Principle of Dynamic Compaction / Dynamic Replacement The compaction is generally the highest below the pounder.
Maximum increase in density is at about one third of the depth of
Dynamic compaction is a soil improvement technique used for in- compaction from the surface. An increase of the penetration
situ densification (compaction) of granular soil (e.g. sand) by heavy resistance of 300 to 400% can be expected in clean sand and gravel.
impacts. It is done systematically in a pre-determined grid pattern. It Dynamic replacement provides an alternative to dynamic
consists of delivering high energy impacts at the ground surface by compaction when the saturated soil cannot be compacted due to low
repeatedly dropping 10 to 40-ton steel pounder from heights of 10 to permeability or excessive fine contents – generally greater than 10 –
40m using a crane (Figure 12). Deep craters up to 2m are formed as 15%. Dynamic replacement is an extension of dynamic compaction,
a result of the impacts. In loose sand, the heave around the craters is using similar plant and equipment (Figure 14). However, unlike
generally small. The craters are filled with sand after each phase. dynamic compaction which is a soil densification technique for
granular soil, dynamic replacement is a soil reinforcement technique
used mainly for cohesive soil (e.g. clay). Stiffer materials with
higher shear modulus are introduced into the soil mass as
reinforcing columns. The improvement of the cohesive soil mass is
derived from the structural aspect of the stiffer reinforcing columns
via a composite soil-column mass, interacting through friction and
adhesion. It increases the bearing capacity, reduces settlement and
improves stability. Backfilling materials for the columns can be any
free-draining, hard, durable, inert materials such as sand, stones or
even rock pieces (up to 300mm size). The volume of columns
usually represents about 15 – 25% of soil mass volume.
Dynamic replacement process starts out by producing a pilot
crater (“print”) with light pounding. The crater is then backfilled
with suitable materials that will lock together under subsequent
heavy pounding (Figure 14). This pounding process is repeated with
increasing compaction energy until a noticeable decrease in crater
Figure 12 Dynamic compaction rig and impact craters formation occurs. Since the column material is more permeable,
pore water pressure from the underlying and adjacent less permeable
The spacing between compaction points depends on the required cohesive soil will dissipate quickly. Hence, the columns being load
depth of compaction, grain size distribution, permeability and bearing columns (reinforcement) also serve as large vertical drains.
location of the ground water table. The initial spacing of the
compaction points roughly corresponds to the compaction depth. It
is necessary to use highest compaction energy, with the heaviest
pounder falling from maximum drop height for the early phases of
compaction in order to extend the compaction effect as deep as
possible. The compaction energy and the spacing of compaction can
then be reduced for subsequent phases thereby allowing adequate
compaction to be carried out at shallower depths (Figure 13).
65
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
a) DC creates a homogeneous soil compaction mass effect while Since dynamic compaction and dynamic replacement were
DR is a composite soil-column reinforcement effect (Figure 15); carried out below water, the conventional “solid” pounders shown in
b) DC uses the same soil material for compaction while DR uses Figure 16 were not suitable due to water buoyancy. A special
imported material “stiffer than the soil mass” for the columns; offshore pounder was designed and built for this purpose following
c) DC starts with maximum compaction energy to reach design a hydrodynamic impact study (Figure 17). The study was carried out
compaction depth and finished with reduced energy in the final to determine the impact speed below water; water resistance and
ironing phase while DR starts with low compaction energy to inertia; impact stress and durability of the pounder while providing
form the column “plug” and finished with increased energy to sufficient “punching” energy to create the rock columns and
compact the whole column material. “compaction” energy to compact the rock mat.
d) DC uses compaction pounder with larger surface area to avoid
“punching” into the soil while DR uses a punching pounder with
smaller surface area for more penetration effect (Figure 16).
66
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The weight of the offshore pounder was 37.5 tons with its • Based on echo sounding survey before and after DC/DR works,
smaller “punching” surface area of 1.72m x 1.72m (approximately 3 the difference in measured surface elevations was about 38cm.
m2) for dynamic replacement process and a larger “compaction” Considering 1m3 loose rock volume equivalents to 0.87m3
surface area of 2.6m by 2.6m (approximately 6.8 m2) for dynamic compacted volume, the net enforced settlement was about 33cm.
compaction process. The height of the pounder was fixed at 2.3m Hence, the calculated volume of a DR column was 6.7m3 and
determined from the required DR columns penetration. The ratio of the diameter of DR column was 2.5m.
solid surface area over open surface area was determined based on • Based on sea-divers’ “direct” field measurements, the diameter
minimum water resistance against net pounder weight. The pounder of the DR columns varied between 2.4 and 2.6m. These values
was made of special manganese steel plates to withstand the are in agreement with the diagonal length (= 2.43m) of a 1.72m
abrasion caused by the heavy impacts on the granite rocks. x 1.72m pounder – dimensions of the punching surface. Surface
Using this offshore pounder, field calibration tests were carried upheave of 10 – 20cm between the columns was recorded. The
out on an area of 22.5m by 22.5m (506 m2). Twenty-five test columns penetrated about 1.3m into the softened clay layer.
columns were installed at 4.5m square grid with 30 blows per From these observations, the calculated and measured diameters of
column. The drop height was 5m in water. Drop height greater than DR columns exceeded the target value of 2m. Hence, the field
5m was shown not to improve significantly the compaction as operation parameters were satisfactory for full production works.
terminal velocity of 7 – 8 m/s was reached. This was consistent with Figure 20 shows the probable soil profile before and after DC/DR
the results obtained from the hydrodynamic impact analysis. After works but before surface levelling. Due to larger columns, the area
the completion of the columns, the rock mat was compacted with 3 replacement ratio (ARR) had increased to 22% instead of 15.5%.
to 6 blows at an overlapping grid pattern as the final ironing phase.
Based on field measurements, the number of blows was plotted
against column penetration as shown in Figure 19. During “column
penetration” phase, the first 15 blows produced significant column
penetration with an average of 1.3m. After the 15 blows, additional
blows produced marginal penetration (< 10 cm). However, more
blows were needed to increase the density (compactness) of the
columns. During this “column compaction” phase, additional 13 to
15 blows were given until surface heaving around the columns was
recorded.
67
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
computer system. Also, the operation parameters such as drop 0.8 – 1 MPa is adequate for self-bearing condition where settlement
height, drop speed, impact velocity and number of drops at will not occur under its own weight (D.60.AN, 1975).
each column location; the total compaction time and the net
“buoyancy” pounder weight were recorded.
- Production rate was about 350 m2 per day.
Filling of sandkey trench with sand and subsequent compaction.
- The sand filling proceeded immediately after compaction of
the rock mat and acceptance tests.
- The sand layer was compacted by vibro compaction.
- Production rate was about 350 m2 per day.
Acceptance tests
- Acceptance tests using PMT for compacted rock and CPT for
compacted sand were carried out. PMT tests were carried out
before and after DC/DR works. CPT tests were carried out
after completion of vibro compaction works.
- Visual inspections and field observations (including
underwater inspections), surveys, samplings and testing were
carried out at regular intervals.
The offshore DC/DR works was carried out using a 200-ton
Liebherr LB895 hydraulic crane, fixed onto a crane barge measuring
25m by 50m with 4-point anchoring system. The DC/DR rig was
equipped with GPS system and a quality control computer system
for compaction. Figure 21 shows the plant and equipment used for
the offshore DC/DR works. The works was completed in 2009. Figure 22 Operation of PMT test with STAF system
Figure 23 (a) PMT read-out unit; (b) PMT drilling rig; (c) jacked-up
pontoon; and (d) STAF drill bit
Figure 21 Offshore DC/DR rig on barge and offshore pounder
Table 3 Results of PMT tests after DC/DR works
3.9 Post DC/DR Tests Results and Discussions Material EP (MPa) EUR (MPa) PL (MPa)
Settlement and stability for the caisson seawalls were re-analysed Compacted rock 7.3 – 10.4 31 – 34 1.1 – 1.6
using the post-DC/DR strength parameters derived from PMT tests.
PMT tests were carried out using the STAF system. STAF – self- Table 4 Interpretation of PMT test results after DC/DR works
bored tube system, is a slotted tube technique with inside
Material EP/PL EUR/EP EY (Mpa) ɸ (o)
disintegrating tool and mud circulation. This method emplaced the
slotted casing tube while drilling to avoid borehole wall disturbance Compacted
6.7 – 8.4 3.5 – 4.2 29 – 42 47 – 49
(Figure 22). This was deemed to bring improvement to the operation rock
and the test results especially in non-homogeneous and over- Note: EP is the pressuremeter modulus; EUR is the unload-reload
consolidated soil. Figure 23 shows the PMT equipment deployed for pressuremeter modulus; EY is the stiffness modulus; PL is the limit
the offshore testing works. PMT tests were carried out to depths of pressure and is the internal angle of friction.
up to 33m below water on a jacked-up pontoon. After DC/DR
works, PMT tests were carried out at 29 different locations which The internal angle of friction (o) for the compacted rock was
included both static and cyclic tests. A typical cyclic PMT test estimated based on the net limit pressure PL* (bars) using the
results is shown in Figure 24. A summary of the results is given in relationship below (Yee and Varaksin, 2012):
Table 3. Table 4 shows the ratio of the unload-reload pressuremeter
φ 40o
modulus (EUR) over the static pressuremeter modulus (EP). The
EUR/EP value ranged from 3.5 to 4.2 which agreed well with the
7
suggested value of 3.5 – 4 for compacted rock and PL greater than PL * 4 * 2.5 (5)
68
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Equation (5) is an extension of Menard (1970) equation for 4.2 Potential Problems of Loose Sand
granular soil extending to > 40o; presumably the indicative lower
D’Appolonia (1970) reported that loose sand is very unstable when
bound value for the internal angle of friction () for granite rock
subjected to even a modest shock and vibration. Such vibration
pieces of similar size which corresponds to measured PL values of
could emanate from pile installation or from seismic effects. It is
about 0.55 MPa (Varaksin, 2009).
prone to liquefaction. For small strain of the order of 10-5 to 10-3 the
The measured internal angle of friction () of the rock after minimum relative density to prevent liquefaction should be about
compaction was estimated to be between 47o and 49o with an 70% and that fine sand with a relative density less than 50% is
average value of 48.5o based on an average PL* of 1.2 MPa which subject to liquefaction during ground motions with acceleration in
satisfied the target requirement of 45o. excess of 0.1g.
Hydraulic sand filling by pipeline method or rainbow method
has large volume of water. The resulting fill structure is likely to be
loose and it will remain loose because of the capillary retention of
the sand which prevents the sand particles from rolling into a stable
and denser configuration. Typical estimated relative density of sand
fill above water level is 50% to 70% and below water level, it can be
as low as 30% to 50%.
While loose sand is not as compressible as soft clay, it continues
to settle under its own self-weight for many years after their initial
deposition. Creep is often estimated as 1% of the thickness of loose
sand fill per log cycle of time which is sufficiently large that it
cannot be ignored in the design of foundations especially for thick
fill as in this case study. Another inherent problem of loose sand is
sudden instability. Although it may be metastable and change state
readily, its very instability nature makes it possible to alter its loose
Figure 24 A typical cyclic PMT test results structure effectively. Shearing of the sand particles by vibration or
impact to form a denser and stable structure has been the most
3.10 Lessons Learned effective means for densifying loose sand.
Dynamic compaction and vibro compaction have found wide
Soil improvement on land above water is usually an easier task. acceptance, and numerous case histories have illustrated their
When it is carried out at great depth below water, the task becomes practical applications for densification of loose sand for higher
more difficult. It requires more rigorous analyses and more elaborate bearing capacity and lower compression. Dynamic compaction has
construction procedure with specialized plant and equipment. The been described above. Vibro compaction is briefly described below.
works procedure or methodology needs on-site adaptation to suit the
prevailing conditions. Quality control becomes more critical as often, 4.3 Principle of Vibro Compaction
it is beyond our normal “visibility” range.
Some of the challenges and lessons learned in this offshore Vibro compaction uses an electric or hydraulic powered vibroflot
DC/DR works include: suspended from a crane (Figure 25). The vibroflot consists of a
Adaptation of coarser and heavier backfill materials (i.e. rock torpedo shaped horizontally vibrating probe that vibrates at a
instead of sand) for compaction under water due to water frequency of 30 to 50 Hz with amplitude between 8 mm and 48 mm.
buoyancy and “turbulence” effect during impacts and the By shearing of particles caused by the horizontal vibration, the loose
requirement to achieve a high internal angle of friction (); soil particles are re-arranged into a denser configuration.
Adaptation of a “porous” and heavier DC/DR pounder for
underwater compaction due to water buoyancy and terminal
velocity of falling pounder and its limited drop height.
Adaptation of appropriate quality assurance and quality control
tests e.g. using PMT test for rock, cyclic and static PMT test to
verify the compactness of rock pieces;
Visual inspections and field observations above and below water
(using sea divers) during works are necessary especially for
offshore works to provide additional information to assess
“what-is-going-on” below water.
69
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
70
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
For calibration of dynamic compaction (DC) works: layer within the sand fill mass is not included in the net thickness.
- Working platform was at +3.1m CD (i.e. 0.45m above top Table 6 shows the measured enforced settlements.
finished level at +2.65m CD to allow for settlement).
- Spacing of compaction points was 6m square grid with two
phases of compaction.
- Phase 1 consisted of 8 blows with 300 ton.m energy per blow
(total energy per phase of 2,400 ton.m)
- Phase 2 consisted of 6 blows with 300 ton.m energy per blow
(total energy per phase of 1,800 ton.m)
- Total applied compaction energy of 117 ton.m per m2
- Treatment depth was 4.65m; measured from +2.65m CD to
-2.0m CD
For calibration of vibro compaction (VC) works:
- Working platform was at +6.0m CD (i.e. 1.5m above top
finished level at +4.5m CD as overburden fill).
- Spacing of compaction points was at an equivalent of 3.4m
triangular grid using an electric vibroflot of 130 kW; 30 Hz
frequency; 23mm amplitude with 30 ton centrifugal force. A
combination of compressed air (using a 12,000 l/min
compressor) and high pressure water jet (100 m3/h under 1.2
MPa pressure) was used as cutting medium. Figure 30 Pre and post CPT results for vibro compaction
- Compaction was carried out to reach amperage of 200 amps or
40 seconds per 50cm lift whichever comes first.
- Treatment depth was 6.5m; measured from +4.5m CD to -2.0m
CD
The above operation parameters were expected to meet the required
acceptance qc values.
71
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
minimum void ratio varied from 0.45 to 0.51 with an average value For vibro compaction, the compaction grid of equivalent 3.4m
of 0.48. The calculated maximum void ratio varied from 0.99 to triangular spacing was reduced to 3.1m – a 17% reduction in area
1.02 with an average value of 1.01. The following equations were coverage per compaction point. Varying duration of compaction and
used to determine the final relative density: withdrawing criteria together with different combination of water/air
G s, γ w jetting pressures were carried out in the new field calibration.
e 1 (6) To account for “ageing” effect (i.e. increase of qc with time), CPT
γd
tests were carried out 7, 14 and 21 days after compaction.
γ d,max (γ d γ d,min ) (7) All the above produced little improvement (< 10%) and did not
DR
γ d (γ d,max γ d,min ) meet the acceptance qc values. It was true that the operating
parameters were optimised. Inadequate compaction was not the
e max e (8) cause for the non-conformance. The only probable reason left was
DR
e max e min down to the intrinsic property of the sand fill material itself.
VT (9) 4.8.3 Carbonate content
VS
1 e
Sand with significant carbonate content usually has higher initial
where DR = relative density void ratios and lower dry densities as a result of its angularity, poor
e = void ratio grading and intra-particle porosity. The values of emin (0.48) and
Gs = specific gravity (= 2.65) emax (1.01); and d,min (12.9 kN/m3) and d,max (17.6 kN/m3) when
w = unit weight of water (= 9.8 kN/m3) compared with typical values of silica sand fit into this trend (Table
d = dry density 8). These initial void ratios are comparable with carbonate sand
d,max = maximum dry density (= 17.6 kN/m3) found in Hong Kong and Dubai where emin = 0.5 and emax = 0.95 are
d,min = minimum dry density (= 12.9 kN/m3) average values for carbonate content greater than 10% (Figure 32).
emax = maximum void ratio (= 1.01)
emin = minimum void ratio (= 0.48) Table 8 Typical void ratios and dry densities of sand (Das, 2008)
VS = volume of solid compared with hydraulic sand fill (of this case study)
VT = total volume
Void Ratio Dry Density (kN/m3)
Based on the assumption of 45% initial relative density and Soil Type
using the enforced settlements of Table 6, the calculated relative emin emax d,min d,max
densities after VC/DC compaction exceeded the 70% criterion Coarse sand 0.30 0.74 14.9 19.3
(Table 7). This contradicts the non-conformance described earlier Fine sand 0.40 0.85 14.1 18.5
and it could not possibly provide any reasonable explanation. Hyd. sand fill 0.48 1.01 12.9 17.6
Table 7 Calculated relative densities after VC/DC compaction Samples were taken for laboratory tests following BS 1377 Part 3
Enforced Relative Density 1990 (Methods of test for soils for civil engineering purposes:
Methods chemical and electro-chemical tests). Visual inspections revealed the
Settlement (%) (%)
Dynamic compaction 7.8 – 8.3 71.2 – 72.9 presence of calcitic shell fragments that were platy, flaky and
Vibro compaction 7.7 – 12 70.9 – 85.3 porous. The test results confirmed carbonate content ranging from
13.1% to 15.2%. Chang et al (2006) reported a carbonate content of
How reasonable is the assumption of 45% initial relative 8% to 16% in the direct-dumped sand and 4% to 12% in the
density? Radio-isotope (RI) cone penetrometer tests were carried out hydraulic filled sand used in Changi East reclamation. Figure 32
at an adjacent area. Relative densities below water were reported to shows carbonate sand found in Hong Kong, Changi (Singapore) and
be as low as 50%. Choa et al. (1979) reported relative density of Dubai. The test results agreed with the trend shown.
hydraulic sand fill ranging from 40% to 80% at the Changi Airport
reclamation. The lower values were obtained below water. Chu
(2011) reported minimum relative density of 45% to 50% for the
hydraulic sand fill at the Changi East reclamation. Figures 30 and 31
indicated relative density before compaction between 30% and 60%
below water when interpreted using the method of Baldi et al.
(1982). Hence, the assumption of 45% is not at all unreasonable.
emax=0.95
4.8.2 Increased compaction energy
Is the applied compaction energy adequate? Causes of an inadequate
compaction for vibro compaction works include larger spacing of emin=0.5
compaction points, insufficient water/air jetting pressure and quicker
withdrawal of vibroflot during compaction (inadequate compaction
time). For dynamic compaction works, it includes inadequate total
applied compaction energy per m2, compaction energy per blow and
insufficient rest period between phases of compaction.
To address these issues, additional field calibrations were carried
out. For dynamic compaction, a 3rd phase compaction was carried
out with additional 6 blows of 300 ton.m energy. The total applied
Figure 32 Void ratios of carbonate sand
compaction energy increased from 117 ton.m/m2 to 167 ton.m/m2 –
an increase of 43%. The heave and penetration tests showed the 4.8.4 Performance of carbonate sand
penetration per blow was about 10 – 20cm for the 1st two blows and
reduced rapidly to less than 10cm thereafter accompanied by The performance of carbonate sand is strongly influenced by the
substantial surface heaving around the craters. This suggested that crushability and angularity of the fragile carbonate grains, high
the energy saturation point has been reached and further increase in initial void ratio of the fill and cementation between particles.
the number of blows will only cause volumetric displacement Carbonate sand is susceptible to particle crushing and hence, it is
(surface heaving) with no further densification in the soil mass. more compressible than silica sand. For an equivalent relative
72
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
density, the qc values measured in carbonate sand will be less than in ratio (e) relative to its maximum (emax) and minimum (emin) values
silica sand. Vesic (1965) added 10% shells to silica sand and as expressed in Equation (8). The complication with the use of
resulted in a decrease of qc values by a factor of 2.3. The higher relative density is the difficulties in measuring density below water
crushability of the carbonate grains is responsible for the low qc and the uncertainty associated with the determination of emax and
values. Since grain crushing increases with increasing density and emin. Hence, direct testing of relative density measurements is
increasing pressure therefore, a larger ratio of qc(silica) / qc(carbonate) is seldom used these days in favour of correlations developed from in-
obtained for denser states and larger mean pressures. At a denser situ tests such as CPT.
state, silica sand shows significant increase in qc values since the Jamiokowski et al. (2001) reviewed test data obtained from
hard particles are forced aside during cone penetration while for calibration chamber tests (Figure 36). Supplemented with few data
carbonate sand, it implies grain breakage, fracturing and crushing as available on undisturbed frozen sand samples, it shows the relative
the fragile particles are pushed closer together as they approach their density against the log scale of normalised cone resistance is almost
minimum packing arrangement (i.e. emin). Figure 33 presents test a linear relationship but still exhibit quite a wide spread of data.
results that reflect this behaviour. When relative density is less than
30% (loose state), the measured qc is about the same for both
carbonate and silica sand. However, for relative density greater than
30% and increasing, the qc(silica) / qc(carbonate) increases significantly to
reach a maximum value of 3.5 when the relative density is
approaching 100%.
73
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
74
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This paper describes the design and construction of foundation system for Malaysia’s first drawbridge, which is located at the
river mouth of Kuala Terengganu River. The proposed drawbridge aims to as bridge link connecting the northern and southern areas of the
Kuala Terengganu River. With the 76m long main span of bascule bridge decks, it permits no obstruction to the navigation traffic along the
river. The design challenges of the drawbridge foundation with free-standing pile length above the river bed are its stringent foundation
displacement requirements for the operation of the mechanical bascule structures under the critical loading conditions (i.e. wind load, seismic
load, accidental vessel impact and hydraulic load in addition to the structural dead loads, live loads and other imposed operational loads) and
also compliances with maritime structure design code of practice throughout the required designed life period. In general, each of the
drawbridge tower structure is founded on 118 numbers of 1m diameter with vertical and raked pre-casted marine spun piles. A giant pilecap
with dimensions of 76m (long) x 28m (wide) x 4.0m (thick) was designed for load transfer between super-structure and foundation system.
During construction works, the validation of the pile performance was very carefully devised and implemented, which included reference
compressive static maintained pile load test, static lateral pile load test, Statnamic test, high strain dynamic pile tests on working marine piles
over the river. The design and construction processes of ensuring the pile head connection conditions achieving the design rotation stiffness
and the problems encountered during massive staged concrete casting of the giant pilecaps with submerged soffit will also be elaborated and
discussed in this paper.
Keywords: bridge foundation, drawbridge, marine structure, marine piling
3. DESIGN CRITERIA
The design criteria of the foundation system mainly focus on
operation requirements of drawbridge structure, serviceability
Figure 1 Overall layout plan and section view conditions and durability of marine structure throughout the
designed life period. Interaction design coordination was conducted
The foundation system at marine portion consists of 4 piers, among bridge design engineer, operation system suppliers and
namely P3 to P6. Two heavy foundation at P4 and P5 are designed foundation design engineer to establish design criteria for the
to support 55 m high drawbridge towers and 76 m long bascule foundation design. As a result, the foundation system need to design
bridge decks and the end span of the integral bridge. for total settlement less than 10mm and horizontal displacement less
than 20 mm in any directions under service conditions.
2. HYDROLOGY AND GROUND CONDITIONS As the foundations structures are located at river mount near to
The site is located at Kuala Terengganu river mount and subjects to South China Sea and subject to daily seawater tide, therefore the
daily seawater tides from South China Sea. Accordance to the Kuala durability design shall according to the marine structure standard.
Terengganu standard port datum, mean sea level (MSL) of the The foundation structure design including pilecap are in compliance
project site was defined at RL0m (+1.45 m Chart Datum (CD)). with codes of practice (British standard) BS5400 - Steel, Concrete
Meanwhile, the highest and lowest astronomical tide level are at and Composite Bridge, BS6349 - Maritime Structures and BS8500 –
75
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
76
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Pile group analysis was carried out by using soil-structure 6. MARINE PILECAP DESIGN
interaction software, PIGLET developed by M.F. Randolph (2004).
Three dimensional pile group deformation as well as forces along The proposed pilecap for Tower Piers P4 and P5 are approximate in
the pile can be reasonably well estimated by adopting soil-structure capsule-like shape with long dimension of 76 m and width of 28 m
input parameters as summarised in Table 4. (see Figure 4). General pilecap thickness is 4m except the centre
part reduced to 2 m for accommodating the counterweight block of
bascule deck required for operation purpose. Concrete grade
Table 4 PIGLET Input Parameters
adopted is Strength Class C32/40 (cylindrical strength/cube
Input parameter Design value strength) with minimum characteristic cube strength of 40 N/mm2 at
28 days.
Soil Shear Modulus (Axial), G axial 11.6 MPa According to BS8500, exposure class of the concrete for marine
pilecap shall be classified as class XS3 (tidal, splash and spray
Soil Shear Modulus (Lateral), G lateral 2.8 MPa
zone). To obtain concrete grade C32/40 with 75mm for class XS3,
Soil Base Modulus, G Base 2.3 GPa cement type CEM IIB-V (Portland cement with 21-35% fly ash) or
cement type CEM IIIA (Portland cement with 36-65% ground
Pile Diameter, D 1m granulated blast furnace slag) shall be adopted as supplied concrete.
The proposed cement content and water ratio for the concrete mix
Pile Young Modulus, E axial 19.3 GPa design (C32/40 with 75 mm for class XS3) are as follow: -
Pile Young Modulus, E lateral 29.3 GPa
Table 7 Proposed Concrete Mix Design
77
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
78
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Back-analysis was carried out with computer software, PIGLET Pre-joined one-length spun pile was decided with construction
to compare the assumed subsoil lateral shear modulus with the and design consideration. From construction aspect, one-length pile
actual pile lateral deformation. It was observed that actual lateral is commonly use in marine piling with time saving for handling and
shear modulus, Glateral, of subsoil from lateral pile test is avoidance of inconsistency of site welding of pile joint. The
approximately 2.8 MPa (approximate 350 times of average SPT-N delivered pile lengths were in 32 m and 38 m as estimated from
value for PIGLET analysis), which is relatively lower than the adjacent boreholes information plus extra 3m for construction
assumed lateral shear modulus of 4.3 MPa in the design stage. tolerance.
With the two preliminary pile tests under compressive loading Some pile performance verification tests were conducted in
and lateral loading carried out, design assumptions and adopted river. Temporary work preparation for the pile tests were properly
subsoil parameters can be verified and calibrated for foundation planed with a safe temporary platform for the sole purpose of
design. With the obtained design parameters and review of the four testing. In general, working platform level is required to be above
river pier pile group performance, revised detailed design with Highest Astronomical Tide (HAT) level to be free from the tidal and
more piles was needed for Pier P3 to ensure compliance pile group wave impacts.
performance under the assigned loading. The pile foundations are designed with fixed head condition
During the working pile installation, 2 nos. of statnamic tests between pile head and pilecap. Load transfer element with
were carried out for Tower Piers P4 and P5 as part of the quality introducing 3.5 m pile plug inside the spun pile anulus of sufficient
assurance/ quality control (QA/QC) procedure. Outcomes of the anchorage reinforcements into pilecap. Effectiveness of pile plug
statnamic load test results showed mobilised test load with force performance is much dependent on installation workmanship. A
pulse duration satisfied the condition stated in ASTM D7383-10: (1) layer of weak cement laitance with primary cement/lime slurry will
The applied force shall exceed the pre-load for a duration time of at normally found around the pile inner anulus due to segregation of
least twelve times the test pile length (L) divided by the strain wave excessive cementitious slurry and aggregate compaction from the
speed (c), 12L/c, and (2) The applied force shall exceed 50% of the spinning process. Therefore, the contractor shall ensure proper
actual peak force for a minimum duration of four times L/c. removal of this weak cement laitance before casting of pile plug.
HSDPT had also been carried out as QA/QC procedure for During piles installation, unexpected weak soil stratum was
installed working pile. Total of 22 Nos. (9.3% of installed piles at encountered at Pier P4 causing longer pile penetration length than
P4 & P5) HSDPT were conducted with mobilised test load of up to estimated pile length. As such additional extension piles with on-site
2 times working load. pile joint welding was needed. In view of potential insufficient
Meanwhile, verification test for designed concrete is vital before bending capacity and flexural stiffness at pile joint area, longer
the concrete casting work on site in order to prevent concrete quality reinforced pile plug was proposed to enhance the pile joint capacity.
issue of the structure along the structure’s service life. Series of test After completion of piles installation works, an unforeseen
such as concrete slump retention test, concrete casting temperature incident case was encountered at Pier P5. A pile head was observed
monitoring, concrete trial mix and concrete cube strength test were missing during site inspection. As the regular river traffic for public
carried out prior to pilecap execution works on site. has to be uninterrupted between P4 and P5, it was suspected the pile
Large volume concrete casting (casting thickness more than 1m) was damaged by uninformed accidental ship impact occurred during
was anticipated at the initial stage of project and the temperature the period of off-construction works. Underwater inspection had
development during casting was always the major concern for the been carried out and the missing pile head was found broken and
pilecap construction work, thus, temperature monitoring was estimated the damage at 9.5m below pile cut off level (2.5 m below
proposed during initial stage concrete casting at Pier P3. Designated riverbed level). Remedial works had been explored with permanent
initial concrete temperature was checked with estimation of 12˚C steel casing extended up above water level. The material falls into
heat generated from every 100 kg/m3 Portland cement hydration pile annulus required to be cleaned up for placement of
process and thus determining 27˚C as the concrete initial reinforcement cages in the remedial works by tremie concreting in
temperature; to control maximum concrete temperature below 75˚C the cleaned pile annulus from pile toe to the pile cut off level.
during curing. However, the designed temperature is then further The proposed pile cap soffit level of Piers P4 and P5 is RL-0.5
verified with on-site concrete temperature monitoring. m, which is 0.5 m below the MSL. Underwater concrete casting for
Post construction instrumentation and monitoring was proposed the pile caps was anticipated. Several options had been explored and
upon completion of the pilecap casting works. Four (4) settlement proposed by contractor for underwater concrete casting method. The
markers were installed on each pilecap to monitor the foundation final adopted temporary work for pile cap concrete casting is
performance during construction stage of the superstructure. The watertight steel formwork mounted to the group piles with water
instrumentation monitoring results show maximum cumulative seals around the piles. Before placing of steel formwork, vertical
foundation settlement of 5 mm with substantial dead load existed on support at the piles is required to carry the vertical loading from
the bridge in January 2018. The instrumentation monitoring period temporary formwork and self-weight of the lean concrete mentioned
will be extended up to completion of superstructure construction later. The designed supports consist of prefabricated steel gripper on
works. each installed pile head and some secondary girders supporting the
steel form. In order to provide sufficient bonding between the piles
8. CONSTRUCTION and steel grippers, a layer of synthetic rubber was used. Bonding test
was carried at laboratory for design certification. The steel gripper
The proposed Piers P4 and P5 are located in Kuala Terengganu was locked by competent diver under water with bolt and nut
River and the piling works for the foundation system was installed tightening system. After that, steel formwork will progressively
by adopting marine piling method i.e. hammer driven from piling place over the support system with a layer of 150mm thick lean
barge. Several constructions aspects needed to be taken into concrete below pile cap soffit level. Dewatering from inner side of
consideration for the marine piling works such as correct positioning steel formwork was then carried out to create dry working condition
and stability of piling barge during pile handling and installation for pile cap concreting work. For safety consideration, the water
process under the impacts from active river flow, daily tidal changes pressure outside the steel formworks was considered at HAT level.
from South China Sea, unpredictable strong wind gust in the distinct Figure 9 shows the pile cap concrete casting preparation works.
monsoon season at east coast of Peninsular Malaysia. The pile cap thickness of the pile cap Piers P4 and P5 is 4 m
Aforementioned conditions have created work challenges and safety thick with total concrete volume of 7,025 m3 for each pile cap. With
issue of work, thus posing uncertainty in overall construction consideration of local concrete supplier availability and the friction
progress. Addition precaution measures and construction planning limitation of the steel gripper support, concrete casting had been
are considered to prevent any unwanted incident and delay during divided into 4 stages. Stage 1 – cast a layer 700 mm thick concrete
construction works. throughout entire pile cap footprint (concrete volume of 1,353 m3)
to serve as base support for subsequence concreting work. Stage 2 –
79
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
80
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Piled-raft consists of pile group and raft. From the pile group point of view, the presence of raft provides an additional bearing
capacity to the pile group. From the raft point of the presence of piles reduces the settlement of the raft. Bearing capacity and settlement of
piled-raft foundation involve a complex soil-structure interaction. Since the pile configuration, pile stiffness, raft stiffness, and soil properties
may vary, there are several combination of parameters involved in the piled-raft group foundation analysis. Several methods have been
proposed in the analysis of piled-raft foundation. The methods vary in the way of treated the piles, raft, and soil. Generally, the methods
consist of analytical method, numerical method, small-scale foundation test, and large-scale foundation test. This paper presents a review of
the advancement of bearing capacity and settlement analysis of piled-raft foundation. The present as well as the classical methods of analysis
are reviewed and discussed. The review shows that numerical method calibrated with small-scale laboratory lest is a promising method to
analyse bearing capacity and settlement behaviour of piled-raft group.
Keywords: Piled-raft group, bearing capacity, settlement.
1. INTRODUCTION corresponding settlement was then calculated. After the pile group
reach its ultimate capacity, the remaining load was carried by the
Pile is used to sustain load from the upper structure when the ground raft and the corresponding settlement was then computed. The load-
soil strength is low by transferring the load to a soil layer that has settlement curve consisted of two linear sections: (i) The first linear
enough strength (e.g. Poulos and Davis, 1980; Bowles, 1996; section corresponds to the load-settlement of pile group; (ii) The
Tomlinson, 1994). In the practice, the piles are arranged in a group second linear section corresponds to the load-settlement of raft.
with a pile cap. The pile cap can be located above ground surface or Randolph (1983) proposed an analytical method to analyse the
in contact with ground surface. load proportion between the pile group and the raft and the
In the case where the pile cap is in contact with ground surface, settlement of piled-raft group based on average behaviour. The
actually the cap can be considered as a raft. The pile cap may proportion of load carried by pile group and the raft can be
provide an additional bearing capacity to the pile group. Thus the computed incorporating the stiffness of the pile group and the
system can be considered as a piled-raft group. stiffness of the raft.
Piles can also be added to a raft foundation. In this condition, Randolph (1994) introduced a parameter of the stiffness of
piles work to reduce settlement of a raft foundation that has enough piled-raft and an interaction factor. The proportion of load carried
bearing capacity but has an excessive settlement. The system can by the pile group and the raft can be computed using these factors.
also be considered as a piled-raft group.
To obtain a general behaviour the bearing capacity and
3. NUMERICAL METHOD
settlement of piled-raft group is difficult for the following reasons: (i)
A large number of combination of pile group configuration can Poulos (1991) idealized the raft in a piled-raft group as strip and the
occur, (ii) A large number combination of pile group-raft stiffness piles in a piled-raft group as springs. A computer program GASP
can occur, (iii) The high load that is required to perform a full-scale (Geotechnical Analysis of Strip with Piles) was developed to
load test, (iv) Different stress-strain behaviour between a small-scale accommodate the idealization.
and a full-scale tests. Lee (1993) performed a piled-raft numerical analysis. Lade and
This paper presents a summary of the advancement of bearing Duncan-Chang stress-strain model were used to model the soil. Both
capacity and settlement analyses of piled-raft foundation. Several piled-raft and raft foundations were analysed. Settlement of piled-
methods of analysis are reviewed and discussed. Finally the raft group and bending moment of raft were obtained in this analysis.
prospective method of analysis is recommended. Poulos (1994) performed numerical analyses of interaction
among piles and between pile group and raft. The analysis was
2. ANALYTICAL METHOD based on elastic condition. The analysis was incorporated in a
computer program GARP (Geotechnical Analysis of Raft with Piles).
Several analytical methods have been proposed to design piled-raft. Ta and Small (1996) used finite element method to analyse
Hain and Lee (1978) analysed the raft as a flexible elastic plate. The piled-raft group. The analysis was a 2-D analysis incorporating the
piles were modelled as compressible friction piles and the soil were interaction factor. Therefore, the 3-D condition can be
modelled as elastic material. The effect of the piles to reduce accommodated with lesser computation effort. The method was able
settlement was analysed. to model a piled-raft group in layered soil.
Davis and Poulos (1972) introduced an analytical method to Long, D. C. (2016) performed analysis on a building founded on
calculate settlement of piled-raft group. The calculation of piled-raft group. The analysis was performed using a commercial 3-
settlement of piled-raft group was based on the calculation of pile D finite element software. In the analysis, it was assumed that the
group settlement by considering the effect of pile cap. In the load was carried by the raft first. After the raft reached its ultimate
calculation of pile group settlement, the effect of settlement of one capacity, the remaining load was then carried by the piles.
pile to other piles in the group was considered. The method was
developed based on an elastic linear assumption. Thus, the piles, the
pile cap, and the soil were in an elastic condition. Davis and Poulos
(1972) also proposed a method to generate a simplified load-
settlement curve of piled-raft group. In the proposed method, it was
assumed that the load was carried by the pile group first and the
81
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
82
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 1 Preliminary flow diagram for the analysis using numerical 12. REFERENCES
method calibrated with small-scale laboratory test
Altaee, A. and Fellenius, B. H. (1994) “Physical modeling in sand”,
Canadian Geotechnical Journal, Vol. 31, No. 3, pp. 420-431.
In the numerical methods, the pile group configuration can be
Bowles, J. E. (1996) Foundation Analysis and Design, Fifth Edition,
varied and analysed easily. Thus, this method can accommodate the
McGraw Hill, New York, 1175 p.
weakness of the analytical method. However, to obtain
Clancy, P. and Randolph, M. F. (1993) "An approximate analysis
representative input parameters in the numerical analysis is not easy.
procedure for piled raft foundations", Int. Journal for
A calibration process between the numerical analysis and full-scale
Numerical and Analytical Methods in Geomechanics, Vol.
or small-scale model tests is required to obtain the representative
17, pp. 849-869, DOI: https://doi.org/10.1002/nag.
parameters.
1610171203.
Full-scale model test is the technically best among the methods
Davis, E. H. and Poulos, H. G. (1972) "The Analysis of pile-raft
considered in this paper. This method can simulate the stress-strain
systems", Australian Geomechanics Journal, Vol. G2, No. 1,
condition in the actual foundation. However, the method costs the
pp. 21-27.
most among the methods considered in this paper. The amount of
material and the instrumentations require high cost.
83
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Hain, S. J. and Lee, I. K. (1978) "The analysis of flexible raft-pile Proceedings of the Seventh International Conference on Beer
systems", Géotechnique, Vol. 28, No. 1, pp. 65-83, DOI: et al. (Eds.), Balkema, Rotterdam, Vol. 1, pp. 183-191.
https://doi.org/10.1680/geot.1978.28.1.65. Poulos, H. G. (1994) "An approximate numerical analysis of pile-
Hussein H. H., Karim H. H., Shlash K. T. (2018) "Theoretical raft interaction", Int. Journal for Numerical and Analytical
verification for full-scale tests of piled raft foundation", Methods in Geomechanics, Vol. 18, Issue 2, pp. 73-92, DOI:
Advances in Analysis and Design of Deep Foundations, https://doi.org/10.1002/nag.1610180202.
GeoMEast 2017, Sustainable Civil Infrastructures Abu- Poulos, H. G. and Davis, E. H. (1980) Pile Foundation Analysis and
Farsakh M., Alshibli K., Puppala A. (Eds). Springer, Cham, Design, John Wiley & Sons, New York, 397 p.
pp. 200-212, DOI: https://doi.org/10.1007/978-3-319-61642- Randolph, M. F. (1983) Design of Piled Raft Foundations, Dept. of
1_16. Engineering Cambridge Univ, 34 p.
Lee, I. K. (1993) "Analysis and performance of raft and raft-pile Randolph, M. F. (1994) "Design and Performance of a Piled Raft
systems", Proc. 3rd Int. Conference on Case Histories in Foundation", Proceedings of the 13th International
Geotechnical Engineering, St. Louis, Missouri, pp. 1331- Conference on Soil Mechanics and Foundation Engineering,
1345. New Delhi, Vol. 5, pp. 61-82.
Long, D. C. (2016) "Prediction of piled raft foundation settlement – Sengara, I. W., Roesyanto, Krisnanto, S., Jayaputra, A. A., and
a case study", Geotechnical Engineering Journal SEAGS & Irsyam, M. (2018) "Settlement and bearing capacity study on
AGSSEA, Vol. 47, No. 1, pp. 1-6. small-scale piled-raft foundation in sand," accepted for
Long, P. D. (2011) "Piled raft - A new foundation philosophy for publication in 2018 in Geotechnical Engineering, South East
high rise buildings", Proc. 1st International Conference Asian Geotechnical Society (SEAGS) & Association of
Geotechnics for Sustainable Development - Geotec, Phung Geotechnical Societies in South East Asia (AGSSEA),
(Eds.), Construction Publishing House, Hanoi, pp. 267-276. Bangkok, Thailand.
Naylor, D. J., Hooper, J. A. (1974) "An effective stress finite Sengara, IW. (1997) "FE analysis and field load test performance of
element analysis to predict short and long-term behavior of a drilled shaft for 52-story tower foundation design", Proc.
piled-raft foundation on London clay", Proc. Conference on Third Asian Young Geotechnical Conference, Singapore, pp.
Settlements of Structures, Cambridge, pp. 394-402. 295-304.
Nguyen, D. D. C., Jo, S-B., and Kim, D. S. (2013) "Design method Ta, L. D. and Small, J. C. (1996) "Analysis of piled raft systems in
of piled-raft foundations under vertical load considering layered soils", International Journal for Numerical and
interaction effects", Computers and Geotechnics, Vol. 47, pp. Analytical Methods in Geomechanics, Vol. 20, No. 1, pp. 57-
16-27, DOI: http://dx.doi.org/10.1016/j.compgeo.2012.06. 72,DOI:https://doi.org/10.1002/(SICI)1096-9853(199601)20:
007. 1<57::AID-NAG807>3.0.CO;2-0.
Poulos, H. G. (1991) "Analysis of piled strip foundation", Tomlinson, M. J. (1994) Pile Design and Construction Practice,
Computer Methods and Advances in Geomechanics: Fourth Edition, E & F N Spon, London, 411 p.
84
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Piled raft foundations are being used now extensively to support structures on problematic soils successfully in the last two
decades. Large numbers of milestone works have been done but all these works appear to have used only homogeneous layers of sand or
over consolidated clay. Few works have covered the behaviour of the piled raft in compressible clay but perhaps not in an extensive manner.
Naturally available deposits many times have compressible layer in between two relatively stronger layers. The present work studies the
effect of such intermediary compressible layers on the performance of the piles within a piled raft. The study has been carried out using
PLAXIS 3D and the paper presents the results of the study.
Keywords: PLAXIS 3D, compressible layer, pile group.
85
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
below the raft The additional complexity generated by the presence layer is stiff clay followed by weathered rock in gravel form Here
of a clay layer at lower levels is the reason for the present study. the N –value is over 50. Pile load tests data are very scarce in
Surfers Paradise. The Authors have done the analysis and will
3. PREVIOUS WORKS compare with available pile load tests data in the second phase.
Based on all the milestone data number of structures have been
successfully supported on piled raft and also have been monitored.
In most of the cases the strata has been considered as over
consolidated clay or medium dense sand. The study on the
behaviour of piled raft on layered soil appears to be very limited
(Chow and Small,2008 ; Yamashita etal., 2010) had used the piled
raft with ground improvement and also supporting of tall structures
supported on soft ground etc has been discussed and published
results are available. However in reality there are number of
instances wherein piles may have to pass through intermediary
compressible layers as seen in the case of many sites in Brisbane
Gold Coast areas (Oh etal, 2008; Moyes etal.,2006, Min.J.Huang
2006). In such cases many structures have been supported on a piled
raft but appears not to have been monitored Therefore It is felt that a
Figure 1 Typical Soil Profile in Surfers Paradise,
more detailed study needs to be done on the effect of such an
Gold Coast
intermediary compressible layer sandwiched between denser layers,
on the behaviour of piled raft. A detailed study was planned and is
being carried out by the above research group of Griffith University RL (m) BH3 BH4
adopting analytical modelling and this paper presents the results of
part of the study. Further studies are going on. ±0.00
Layer 1: Medium Dense Sand
4. METHOD OF ANALYSIS
Complex three dimensional problems like a piled raft cannot be
Layer 2: Very Dense Sand
analyses have to take into account the soil - structure interaction Scale 1:400
0 3 6 9
within the foundation. Keeping the above in mind, for the present *Layer 6: Clayey Sand and Layer 7: Gravel Sand are missing in this site
study, analytical method the finite element method and the software
Figure 2 Artique Project- Soil Profile along Section
PLAXIS 3D have been used in the rigorous analysis.
(Huang, 2006)
4.1 Basis of Model Definition Depth (m) BH1 BH2
The selection of model was done based on the problem requirement. WT.L
Layer 1: Medium Dense Sand
Since the present study attempts to find the behaviour of the piled
raft with an added complexity, the model has to be computationally
simple and at the same time must be able to produce the desired Layer 2: Very Dense Sand
results with minimum computational time .Hence a simple square
model has been assumed for the study. The model is briefly
discussed in section 4.3
Layer 5: Stiff Clay
4.2 Geotechnical Model Layer 6: Clayey Sand
86
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
It has been observed that the piled raft with an area ratio of around settlement of the piled raft reached 35mm in the case considering
6% produces the best performance (Balakumar, 2008). the intermediary compressible layer.
The raft is 1m thick and the piles are 750 mm dia and 28 m long Figure 6(b) Deformation contour
namely 1.5 times the raft width. The length of the pile has been
taken on 28 m so that there will be sufficient length of pile available 5. LOAD SETTLEMENT RESPONSE
below the compressible layer. The piles are fully fixed to the base of In general in the case of homogeneous layer the applied pressure
the raft. Figure 6 a presents the finite element model and 6b the from the raft enhances the confining pressure below the pile group
deformation contour. The various layers have been identified by the by the inter -granular friction. This enables the pile group to take a
respective Es, density and Poisson’s ratio etc have been derived higher load by friction and once the friction is overcome the system
from the correlations given by Poulos (1988). The parameters are loses its stiffness gradually and then rapidly.(Balakumar , 2008) as
presented in Table 1.No mesh refinement was needed. The piles shown in Figure 7.
have been modelled as beam elements.Four typical piles have been
identified for the study. They are central pile (Pile 1), outer pile in
8MM RAFT 10MM DIA. PILE - 6D SPACING - LOOSE SAND
LOAD, kN
line with the central pile on the X-axis (pile 2),corner pile as pile 3 0 1 2 3 4 5 6 7 8 9 10
and the peripheral pil below the central pile on Y-axis downwards.
0
d = 10mm
4 t = 8mm
S=6d
SETTLEMENT
B = 200 mm
8
SETTLEMENT, MM
NT, mm
12
16
20 Loose
Med dense Bed density
Dense
24
87
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
of 5% which is close to the area ratio of the model studied now thereafter the fall is rapid. But it can be seen that the behaviour is
namely 6% The details of the 1g model tests on square piled raft more uniform and the ratio of axial stress to tip stress works out to
with 4D pile spacing on a poorly graded sand bed is discussed in 20%.
detail elsewhere (Balakumar and Ilamparuthi,2005). It is assumed
that this small variation may not affect the comparison of load
settlement response. Figure 6 presents the load settlement response SQUARE RAFT - LENGTH RATIO Vs SHAFT STRESS (20mm)
Shaft stress N/mm2
of piled raft on a homogeneous layer and Figure 8 presents the load 0.00 0.40 0.80 1.20 1.60 2.00 2.40 2.80
settlement responseof piled raft of the present study. of piled raft of 0.00
the present study.
0.20
Length 3
ratio 0.40
0.60
88
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
tip stress to the head stress to works out to 45% to 53% the pile head , stress at 0.5 L – stress at 0.25L/ stress at 0.25L and so on)
maximum being at the central pile and the minimum being at the is plotted against length ratio to study the effect of layer thickness.
corner pile. The reduction ratio works out to 85 % upto 13 m level; Similarly to study the effect of layer consistency the analyses was
from 13m to 22.5m the reduction in stress with reference to the head repeated for three different N-values namely N=4, N=8 and N=12.
stress is 71% which reduces to around 45%. The trend remains same In the same manner as said earlier the stress reduction ratios are
although the magnitude may vary. This indicates that the frictional tabulated and for one typical pile a plot is made between the stress
resistance of the weaker layer is enhanced by the stronger layers reduction ratio and length ratio for discussion.
above and below the compressible layer.
6.1 Discussion
In the case of load transfer mechanism there are three types of
interaction namely raft - soil soil-pile, pile –pile and raft- pile,and
raft -raft When there is an intermediary compressible layer there
appears to be an interaction between the upper layer (medium dense
to dense sand) and the lower layer namely stiff clay which causes
perhaps an enhancement of the frictional resistance of the weaker
layer in between. There is always a confusion particularly when the
soft layer is found below a denser layer in computing the negative
friction in the sense that when the compressible layer reference to
the pile the upper layer also settles down and hence the upper sand
layer also can contribute for additional drag load. This behaviour is
very evident from the shaft stress distribution from the rate of
reduction in the axial stress with depth. Although the pile group in a
piled raft is intended to be primarily a settlement reducer the pile
design has to recognize this excess load. It is also to be noted here
that the ratio of tip stress to head stress varies from 40% to 45% Figure 12 Bending Moment Distribution with compressible Layer
caused by perhaps the increased frictional resistance of the
compressible layer due to the stronger upper and lower sand layers. Bending Moment (kNm/m) Bending Moment (kNm/m)
-180 -120 -60 0 60 120 180 -100 -50 0 50 100
2 2
In general the piles are always designed for combined axial load and
bending moment due to the lateral loads. The bending moment due 4 4
to axial load is very small, but when the pile passes through 6 6
Depth (m)
is seen from the Figure 12 that the bending moment exhibits both
12 t=0.25m
t=0.4m
t=0.4m
8. PARAMETRIC STUDY
In the present study, the influence of two important factors relating
to the compressible layer, namely the thickness and the consistency To discuss the results, the variation of the stress reduction ratio
on the behavior of the pile group of piled raft need to be considered. to the length ratio is presented for the central pile in Figure 14.
The effect of the layer thickness was studied by repeating the
analyses for three different thicknesses of the layers namely 2.5m,
5m, and 8m. The shaft stress distribution over the length of the piles
is plotted as typically shown in Figure 10. The stress reduction ratio
α which is the ratio of the variation in the stress between two
successive points ( eg : stress at pile head – stress at 0.25L / stress at
89
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
It can be seen that the variation in the shaft stress reduction ratio is
gradual and uniform indicating that N- Value has no significant
influence the shaft stress reduction and the variation is as per the
general behavior. In other words the shaft stress reduces over the
length uniformly unlike in the case of variation in the layer
thickness.
9. CONCLUSION
In order to study the applicability of piled raft under any practical
geotechnical conditions the effect of intermediary peat or
Figure 14 variation of the stress reduction ratio to the length ratio compressible layer was studied. Such profiles are widely seen in the
Gold Coast area as furnished by various authors mentioned in Table.
It can be seen that the rate of increase in the stress reduction Accordingly the study was carried out numerically using PLAXIS 3-
ratio is rapid as the length ratio of the pile increases when the layer D considering a general soil profile. A parametric study was also
thickness is 2.5m. As the thickness increases, the rate of increase in conducted by varying the thickness of the compressible layer and its
the stress reduction ratio reduces between the length ratio 0.5 and consistency. The conclusions are as follows:
0.75 and then it increases. This indicates that the compressible layer 1. The presence of peat layer induces a higher differential
present in this region settles, generating the negative friction which settlement. The settlement varies by 20 to 25% from the centre to
causes an increase in the shaft stress locally. In all the other piles the edge.
also the same trend exists. It can be said that as the thickness of the 2. From the study of axial stress and bending moment variation in
layer increases, the compressible layer contributes for the negative the vicinity of the compressible layer, the fall in the axial stress
skin friction locally, increasing the shaft stress at that section. distribution is influenced by the drag force developed by the peat
layer. The drag force adds an additional load on the pile
8.2 Effect of Variation in the State of Compaction irrespective of its location. In other words the piles are subjected
to additional axial load in the section of the pile passing through
In the same manner as in the previous case, the analyses was the compressible layer.
repeated for three different cases of N-Values namely N=4 ( soft), 3. It is seen from the ratio of tip stress to head stress; the ductile
N=8 ( medium stiff) N=12 (stiff) representing three different behaviour of the pile group is affected by the presence of the peat
consistencies of the compressible layer. The rate of variation of the layer.
shaft stress reduction ratio α is presented in the cases of four 4. From the parametric study it is seen that the thickness of the
selected piles in Table 3. compressible layer has a higher influence on the behaviour than
Table 3 The rate of variation of the shaft stress reduction the state of compaction.
90
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Clancy P. and Randolph M.F. (1993), ‘Simple Design Tests for Poulos H.G., Small J.C. Ta L.D., Sinha J. and Chen L. (1997),
Piled Raft Foundations’, Geotechnique, Vol. 36, No. 2, pp. ‘Comparison of Some Methods for Analysis of Piled Rafts’,
169-203. Proc. 14th Int. Conf. Soil Mech. Foundn. Engg- Hamburg – 2,
E.Y.N OH, M. Huang, C. Surarak, R. Adamec and A.S. pp. 1119-1124.
Balasubramaniam (2008), “Finite Element Modeling for Piled Poulos H.G -( 1998) The Pile - Enhanced Raft – An Economical
Raft Foundation in Sand”, Eleventh East Asia – Pacific Foundation System – Keynote Lecture 11TH Brazilian
Conference on Structural Engineering & Construction conference on soil mechanics and F oundation Engineering.
(EASEC – 11). Russo.G.(1998) Numerical Analyses of Piled Rafts,
Kim H.T., Yoo H.K. and Kang I.K. (2002), ‘Genetic Algorithm Intl.Jnl.Num.and Anl.Methods in GeoMech,Vol 22,pp.477-
Optimum Design of Piled Raft Foundations with Model 493.
Tests’, Journal of South East Asian Geotechnical society, pp. Weisner T.J. and Brown P.T. (1978), ‘Laboratory Tests on Model
1-9. Piled Raft Foundations’, Research Report 318, Sydney
Paran Moyes ; Harry G Poulos ; John C. Small Frances Badelow: University.
Piled raft Design Process for a High-rise Building on the Yamashita, K, Hamada, J., and Yamada, T. (2010),
Gold Coast, Australia. FieldMeasurements On Piled Rafts with Grid-FormDeep
Min.J.Huang -2006. Mixing Walls on Soft Ground, Geotechnical Engineering-
Min Huang. M (2006) Rafts and Piled Raft foundations at Surfers SEAGS, Vol. 42, No. 2, June, 2011.
Paradise Gold Coast, Australia – Analytical study using
PLAXIS Software – A theses submitted in Partial Fulfilment
partial requirement of the degree of Master of Civil
Engineering from Griffith University Gold Coast campus
Australia
91
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: In this study, three cases of in-situ laterally loaded tests of caisson foundations are simulated. The differences of these three
cases are the application position of lateral loads and the scoured depth of the foundations. In the numerical model, a Winkler beam model is
adopted. The Winkler beam model utilizes the beam element to model the caisson and six types of springs to simulate the horizontal
subgrade reaction, the vertical shear stress on the front of the caisson, the horizontal shear stress and the vertical shear stress on the sides of
the caisson, and the normal stress and the horizontal shear stress on the base of the caisson. The simulation results show that the model can
properly capture the responses of the caisson foundations under lateral loading.
Keywords: Caisson foundations, In situ tests, Lateral loads, Winkler model.
2. OVERVIEW OF TESTS
The old Niu-Dou Bridge had two independent bridge structures, as
shown in Figure 1. Both bridge structures had seven spans. In the
test program, columns P2, P3, P4, and P5 of the left-side bridge
were tested, as displayed in Figure 2. The columns were supported
by caisson foundations. The foundations generally had a diameter of
4m and a length of 12m. The three column load tests were
conducted on P2, P3, and P4 and the foundation load test on P5. The
location of load application on P2, P3, and P4 were 9.54, 10.59, and
10.37 m, respectively, measured from the foundation top. The
column sections had a diameter of 1.8 m. These columns had Figure 1 Niu-Dou Bridge (Chiou et al., 2012)
different foundation embedment. The foundation of column P2 was
256.2 m
slightly exposed with a length of about 0.4 m. The P3 foundation
was exposed with a length of 1.2 m. The P4 foundation had a larger 36.6 m 36.6 m 36.6 m 36.6 m 36.6 m 36.6 m 36.6 m
exposed length of 4.0 m. The P5 foundation had an exposed length
of about 1 m.
For the column load tests, cyclic displacement controlled load
tests were performed on columns P3 and P4. A pseudo-dynamic test
and a single-cycle pushover test were conducted orderly on column
P2. A load-controlled monotonic lateral load test was carried out on
the caisson foundation of column P5.
12 m
P1 P2 P3 P4 P5 P6
Figure 2 Test specimens of Niu-Dou Bridge
92
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
4. THE ANALYSIS MODEL AND SIMULATION slightly damaged due to the former pseudo-dynamic test on it.
Figure 7 shows the comparison between the simulation curve and
This section further analyzes the column tests through numerical the pseudo-dynamic test: the initial portion of the simulated curve is
simulation to clarify the behavior of the foundations in the tests in very close to that of the pseudo-dynamic test. In Figure 6, the trend
more detail. of the simulated foundation displacement is consistent to that of the
4.1 Analysis model measured displacement: the simulation curve is a little softer than
that of the experimental one.
Chiou et al. (2012) applied a Winkler beam model to simulate the
lateral response of P5 foundation under loading. In this study, their
model was used to simulate the foundation behavior in the column
load tests. The Winkler-beam model adopted uses beam elements to
simulate the caisson body and spring elements to simulate the soil
reactions. To well simulate different soil reactions around the
caisson, a six-component Winkler spring model, as shown in Figure
5, is adopted. Six types of springs are utilized to simulate the
different components of the soil reactions acting on the caisson with
an equivalent rectangular section of width Be (perpendicular to the
direction of lateral loading) and length Le (parallel to the direction of
lateral loading). Springs kH and kSVB represent the horizontal
subgrade reactions and vertical shear stress levels on the front of the
caisson, respectively; springs kSHL and kSVL represent the horizontal
shear stress and the vertical shear stress on the sides of the caisson,
respectively; springs kV and kS represent the normal shear stress and
the horizontal shear stress on the base of the caisson, respectively. Figure 4 Analysis models for the column load tests
Since the test site was on gravelly soil, Chiou et al. (2012) proposed
a method which modified the load-displacement responses from the
plate loading tests and the direct shear tests to determine the load-
deformation characteristics of the springs. The details for the
method can be found in Chiou et al. (2012).
Beam elements are used to model the columns. In order to
simulate nonlinear flexural behavior of the columns, the distributed
plastic hinge model is applied. According to Chiou et al., (2009), the
properties of plastic hinges are set based on the moment-curvature
curves of the columns. Based on the above settings, the analysis
models for the column load tests using SAP2000 program
(Computer & Structures, 2002) are displayed in Figure 4.
Le
V0
M0
H0
kSVL1
kSVB1 kSHL1 kSVB1
kH1
kSVL2
kSHL2
kSVB2 k SVB2
kH2
kSVL3
kSVB3 kSHL3 kSVB3
kH3
kV
93
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
4.4 Analysis of P4 column load test Figure 12 Simulated load-displacement curves of columns P2, P3,
and P4
In P4 test, the exposed depth of the caisson was 4 m (1/3 length of
caisson). The point of load application on the column was 10.37 m
above the foundation top.
The analysis results of P4 test for the lateral displacement of the
column and the foundation are shown in Figs. 10 and 11,
respectively. In Figure 10, the simulation curve is close to the
experimental one. In Figure 11, the trend of the simulated
foundation displacement is consistent to that of the measured.
From the above simulations, it is seen the model adopted can
reasonably capture the response of the behavior of the columns and
their foundations in the tests.
With the analysis results, Figure 12 compares the lateral load-
displacement curves of columns P2, P3, and P4. As observed in the
test results, the overall curves are close. Their stiffnesses are a little
different, but their strengths are close. However, as shown in Figure Figure 13 Simulated load-foundation displacement curves of P2, P3,
13, their foundations have different responses. It can be seen in the and P4
figure that at the same lateral load P4 foundation has the largest
lateral displacement while P2 and P3 foundations have close lateral 5. CONTRIBUTIONS OF SOIL REACTIONS
displacements. P4 foundation has the lowest lateral stiffness and
In addition to the horizontal soil reaction in front of foundation kH,
strength because of the largest exposed length of foundation. As
the shear reactions around the caisson (kSVB, kSHL and kSVL) and the
those curves are compared with that of P5 foundation, the lateral
bottom resistance also contribute lateral resistance (kS and kV).
stiffness is the largest because of a pure horizontal load on the top of
Figure 14 (a) and (b) compares the horizontal resistance contribution
the caisson and a smaller exposed length.
of P2 and P4. The resistance of the caisson to the horizontal loading
mainly comes from the horizontal resistance of the soil in front of
the caisson. The horizontal side shear resistance of the caisson
94
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(a)
(a)
95
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Study of Bored Pile Capacity in Klang Valley Residual Soil based on Field-
Performance Data
Allan Y.L. Chwee1, E.G. Balakrishnan2, and A. Nazli3
1
Assistant Engineering Manager, GCU Consultants Sdn Bhd, Malaysia
2
Managing Director GCU Consultants Sdn Bhd, Malaysia
3
Faculty of Civil Engineering, Universiti Teknologi Malaysia, Malaysia
E-mail: allanchwee@gcu.com.my
ABSTRACT: Bored pile has been gaining its popularity in this recent years due to its ability to carry higher working load and higher lateral
force resistance. On top of that, bored pile is suitable to penetrate into shallow rock or hard layer. Many designers in the past would ignore
the end bearing contribution in bored pile due to the problem of soft toe at the base. This approach would be too conservative which would
result in longer pile length as the geotechnical capacity would only depends on shaft friction. In Malaysia practise, geotechnical capacity of
bored pile in residual soil is calculated from SPT-N correlation based on modified Meyerhoff approach. This paper presents the study of
relationship between pile capacity obtained from load test results and field performance data in Klang Valley geology. Extrapolation
technique using Chin’s method has been adopted in obtaining the ultimate shaft friction for non-fail load test. A total of 20 bored piles of
diameter varying from 900 mm to 1500 mm were tested on multiple geology formation in Klang Valley. From this research, the relationship
between shaft friction and SPT-N blow in three different types of geology such as Granite formation, Limestone formation and Kenny Hill
formation are explored. The findings showed that fsu/N ratio in Limestone formation and Kenny Hill formation are quite match and
agreeable with studied carried out by other authors. On the other hand, higher fsu/N can be obtained in Granite formation based on the
plotted graph. For the base resistance, interpretation of the collected load test results indicated a matching relationship between fbu and
Pb/Ptop in soil as well as rock where R-squared of more than 90% are obtained. The study of load test results have enable development of
shaft resistance correlation with SPT-N blow as well as correlation between Pb/Ptop and base resistance. This would allow designer to adopt
higher bored pile capacity which would result to shorter pile length and cost saving.
Keywords: SPT-N correlation, shaft friction, end bearing, load test and pile length.
1. INTRODUCTION length optimization if higher shaft resistance factor and end bearing
factor can be used for different layer of soil (Varies in SPT-N).
In Malaysia, the conventional geotechnical practice for pile This paper studied the relationship between shaft resistance/end
foundation design is based on working stress method where the bearing and SPT-N blow in different geology formation and discuss
derived allowable carrying capacity of pile (working load of pile) the possible correlation which can be developed from load test
shall always be greater than the unfactored column load. The results.
working load of a pile is obtained by applying safety factors (FOS)
to the ultimate geotechnical capacity of the pile. Conventionally,
2. LITERATURE REVIEW
different or same FOS will be applied to the shaft and base
resistance of the pile. The FOS and the combination of the different 2.1 General Geology of Klang Valley
FOS are depending on multiple factors such as the experience of the
designer to the similar soil or rock conditions, project design brief, Klang Valley is defined as an area in Malaysia which centered in
client requirement, etc. Technically, the FOS used in foundation is Kuala Lumpur and include the surrounding cities of Selangor state.
not a factor of safety on the strength of the foundation, but is a The largely Lower Palaezoic rocks extend southwards along the
factor to limit the settlement at serviceability based on experience of Western Belt into Selangor and the Federal Territory of Kuala
most soils for which relatively stiff and linear behavior will persist if Lumpur. The Kuala Lumpur Limestone is overlain unconformably
the stress levels are kept below about 30 to 50% of their ultimate by the more gently folded Carboniferous to Permian Kenny Hill
capacities (Atkinson, 2007). Formation, estimated to be about 1200 to 1500 m thick, that is
Commonly, the unit shaft friction and unit end bearing are exposed in several places notably towards the south of Kuala
obtained from correlation with SPT-N value primarily for residual Lumpur, from Petaling Jaya and Puchong to Kajang and Putrajaya.
type of soil. Many studies have been conducted and adopted in The Kenny Hill Formation named after Kenny Hill (Bukit Tunku) in
today practise where the pile geotechnical capacity is calculated Kuala Lumpur is a monotonous clastic sequence of interbedded
based on SPT-N correlation. Pile design is often verified through shale, mudstone and sandstone. The Kenny Hill Formation (Yin,
pile load tests and is essential to ensure that settlement of the pile 1976; Foo, 1983) shows both soft-sediment deformation structures
would comply with the permissible settlements as per the standards and multiple tectonic deformations.
or code requirement. On the other hand, the intrusion of granitic rock, mainly light-
In today construction, instrumented and working pile load test grey, coarse-grained biotite granite occupies the hills and the
such as static load test, high strain dynamic load test and even bi- undulating area around Kuala Lumpur except in the south. The
directional load test have been widely carried out. With the huge granite has intruded into the older sedimentary rocks and has caused
amount of load test data, study can be undertaken to refine the tilting and warping of the Kuala Lumpur Limestones and Kenny Hill
correlation based on field performance data. Commonly in residual Formation. The contacts between the granite and sedimentary rocks
soil, unit end bearing and unit shaft friction are obtained from are generally covered by thick alluvial or superficial deposits during
correlation with field performance data such as SPT-N value. Quaternary period (Komoo, 1989).
Considering the different types of soil in Malaysia, different
stiffness of residual soil could lead to different unit shaft friction 2.2 Existing Correlation with Field Data
value. With the huge pool of instrumented pile load test data in The current design practice in Malaysia is based on the conventional
Klang Valley, Malaysia, there is possibility of developing a realistic approach which is the working stress method. The working load
unit shaft friction and end bearing correlation to suit with Klang acting on the pile shall always be lesser than allowable carrying
Valley ground condition and further enhance the existing capacity of pile. Several correlation for calculating fsu and fbu using
correlation. Study can be carried out to assess the possibility of pile SPT-N values have been proposed by different authors. Meyerhof
96
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(1976) suggested the ultimate unit end bearing, fb in a homogeneous The instrumented pile load tests were collected with the
granular soil for displacement piles as: following information:
a. Load settlement readings at the pile head.
fb = 40Nb (Db/B) < 400Nb (1) b. Instrumentation data comprising strain gauges (resistance wire
and vibrating wire types) and rod extensometers (tell-tale)
where Nb is the average value of SPT-N at the pile base and Db/B is along the length of the piles.
the average depth ratio of the base into bearing stratum. c. Soil Investigation data for each pile test.
Other authors suggested a common equation as below.
3.2 Extrapolation of Data
fbu = Kbu x SPT-N (2)
Geotechnical capacity which consist of shaft friction and end
where Chiu and Perumalswamy (1987) recommended Kbu to be 50 bearing can be obtained from the instrumented pile load test.
and Toh et al. (1989) proposed Kbu values between 27 and 60. Generally, the ultimate capacity is obtained when the pile is loaded
The relationship for shaft friction in correlation to SPT-N values to 3 times of the working load or load to fail which would mobilize
is generally expressed as below. the pile. However, non-fail load test would not have mobilized shaft
friction and end bearing. Thus, the ultimate capacity would not be
fsu = Ksu x SPT-N (3) able to be determined. Hence, it is important to determine the
ultimate or limit load as accurately as possible. Several methods are
Meyerhof (1976) suggested Ksu to be 2.0 while Chiu and available to extrapolate the final value to obtain the peak value or
Perumalswamy (1987) have adopted 2.5. Toh et al. (1989) resistance of the load test results. Four common extrapolation
recommended Ksu ranges between 2.5 and 2.7 for N values up to methods available are Davisson’s, Hansen’s, Chin’s and De Beer’s
120 and a lower values of Ksu should be used for greater N values. methods.
Established local correlations uses 2.0 as Ksu with fsu limited to Davisson’s method allow the engineer to extrapolate the
200 kPa for shaft resistance and Kbu = 40 with fbu limited to 4000 maximum value with consideration to the length and size of the pile.
kPa for base resistance. On the other hand, Brinch Hansen’s method predict the failure based
on the assumption that hyperbolic relationships exist between the
load and the displacement. For Chin’s method, it made assumption
3. METHOD OF STUDY
that the load-settlement (Q vs Δ) relationship is hyperbolic and
This research focused on the study of bored pile geotechnical hence the inverse slope of a plot of Δ/P (vertical axis) vs P results in
capacity in relation to field performance data such as standard the failure value. Lastly in De Beer’s method, load settlement values
penetration test (SPT-N). This research also study the influence of were plotted on double logarithmic chart.
empirical correlation to the different type of rock formation. From the 4 methods discussed, Chin’s method is deemed to be
most suitable for extrapolation of data considering its simplistic
3.1 Data Collection approach and independent from any soil condition. Chin’s method
make it possible to predict the ultimate resistance even if the head
For the study of bored pile performance, instrumented pile load test settlement did not reach 10% of the pile diameter. However, the
results were collected throughout the Klang Valley area. The extrapolation value become more accurate when the head settlement
following data were collected and the summary is indicated in Table approaches 10% of pile diameter.
1. A proposal by Borel et. al (2004) in his research indicated that
Data at 14 sites in Klang Valley were gathered. the Chin method over predicts significantly the shaft resistance
20 numbers of Instrumented Bored Pile were collected with when the load is mainly resisted by the toe. The absolute error made
various geological formation. can exceed 30% of the ultimate resistance which is independently of
Years of testing are from 2012 to 2017. the percentage of load carried by the shaft. Therefore, the value
extrapolated by Chin method has been reduced by 20% in order to
Table 1 Summary of Data Collection represent the actual ultimate capacity.
No Location Pile Year Formation
1 Site 1 3.3 Instrumented Bored Pile Load Test
PTP1 2012 Kenny Hill
2 Site 2 PTP1 2013 Granite Instrumented pile load test has been commonly adopted by designer
3 Site 3 PTP2 2013 Limestone and accepted by most of the client before commencement of actual
4 Site 4 RRSN-TP1 2013 Granite piling work in the recent years. The awareness of instrumented load
test has grown exponentially over the years. Client understand that
5 Site 4 PTP2 2013 Granite
instrumented load test is able to determine the capacity of pile and to
6 Site 5 TP1 2013 Granite verify the design parameters in relation to shaft and base resistance.
7 Site 5 TP3 2013 Granite Thus, optimization can be explored which would result in economic
8 Site 6 PTP1 2013 Granite design.
9 Site 6 PTP2 2013 Granite In the selection of instrumented pile load test, several general
10 Site 6 PTP3 2013 Granite rules were observed as below:
11 Site 7 BN-122 2013 Limestone Pile used in the load test shall be constructed similarly with the
12 Site 7 working pile.
PTP1 2013 Alluvium
The test pile could be smaller than the actual working pile but
13 Site 8 PTP1 2014 Limestone
limited to 50% of the working pile diameter.
14 Site 9 PTP1 2014 Limestone Location of the test pile must be similar to the working pile’s
15 Site 10 P03 2015 Granite ground condition.
16 Site 11 TP1 2015 Kenny Hill Test pile shall be carried out within the same site with the
17 Site 11 TP2 2015 Kenny Hill working pile.
18 Site 12 PTP1 2016 Kenny Hill Instrumented pile shall be tested to 3 times working load or
19 Site 13 PTP1 2016 Granite failure. Otherwise, extrapolation of data shall be carried out.
20 Site 14 PSLT 01 2017 Granite All the test pile selected must be installed with instruments
where the arrangement generally consist of 4 nos of Vibrating Wire
97
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
VWSG LEVEL A
VWSG LEVEL C
VWSG LEVEL D
VWSG LEVEL E
Figure 5 Kentledge (Static) Load Test
VWSG LEVEL F 3.3 Correlation between Pile Load Tests and Field Data
From the large pool of load test results, tabulation can be carried out
PILE TOE LEVEL
VWSG LEVELG to determine the relationship between shaft resistance and base
resistance with field performance data ie. borehole results. In the
process of analyzing the result, several assumption have been made.
Figure 1 Typical Arrangement of Instruments for Pile Load Test i. The piles were vertical and no eccentric loading.
ii. All the strains measured in the piles were due to
compressive force only.
iii. The cross sectional area and circumference of the test
piles were assumed to be constant through the pile length.
iv. Elastic modulus of pile is assumed to be equal to elastic
modulus of concrete material of the pile.
The procedures used in the data analysis of the instrumented
piles and tabulation of data are as follows:
i. Only ultimate/maximum shaft friction and end bearing
values are collected. For load test that have not reach
failure state, extrapolation using Chin’s method are
adopted to obtain the maximum capacity.
ii. Load test selected must be fully instrumented with both
strain gauges and extensometers installed along the length
Figure 2 Vibrating Wire Strain Gauges of the piles.
iii. The depth of the strain gauges were measured from the top
of the pile.
iv. Graphs were plotted for ultimate end bearing against
Pb/Ptop, ultimate shaft friction/N against SPT-N and
ultimate end bearing/N against SPT-N based on different
types of geology.
98
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 6 Relationship between fsu/N and SPT-N for Granite Figure 9 Relationship between fbu and Pb/Ptop in soil
formation
From the plotted graph in Figure 9 and 10, the summary of the
developed equation and R2 percentage are listed in Table 2.
99
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5. CONCLUSION
From the study done on the collected test results, the following
conclusion are drawn:
a. The subsoil profile in Klang Valley consist of varies geology
mainly Kenny Hill, Granite and Limestone formation. With
this varies formation, study has been carried out to determine
the influence toward the fsu/N ratio. From the study, Kenny
Hill and Limestone formation indicated almost similar value
and close to the existing correlation proposed by other
authors but granite formation provided a higher fsu/N ratio.
As Granite is an igneous type of rock while Kenny Hill and
Limestone are sedimentary type of rock, it is believed that
different type of rock formation would influence the fsu/N
ratio.
b. Some of the load tests have not reach the maximum capacity.
Extrapolation using Chin’s method deemed to be suitable to
Figure 11 Bar Chart for variation of Pb/Ptop according to load test obtain the ultimate capacity.
c. Relationship between fsu/N and SPT-N indicated that higher
fsu/N can be obtained for SPT-N < 50 while fsu/N ratio drop
From Figure 11, almost half of the collected load tests indicated to a consistent range of 2 and 3 for SPT-N > 50. Therefore,
Pb/Ptop ranges from 20% to 40% at 12.5 mm settlement while the higher fsu/N can be adopted for soil with SPT-N < 50 but
rest of the load tests showed Pb/Ptop less than 10%. An average verification using instrumented pile load test is highly
value of 12% is tabulated from the 20 numbers of load tests carried recommended.
out. d. For correlation of end bearing, 2 type of conditions are
considered which are pile in soil and pile in rock. Based on
4.2 Discussion the plotted graph between fbu and Pb/Ptop, the correlation of
From the graph plotted between fsu/N and SPT-N for three (3) end bearing are proposed as in equation 4 and 5. Both of the
different types of rock geology as in Figure 6, 7 and 8, a relationship correlation indicated high value of R-squared of more than
can be derived. All the 3 plotted graphs indicated higher fsu/N can 95% which indicate a good fit for all the data.
be obtained for SPT-N < 50. Once SPT-N > 50, the fsu/N ratio e. The obtained end bearing correlation with SPT-N blow for
decreased substantially which shown an almost consistent pattern. soil is within the recommended correlation by other authors.
This study indicates that fsu/N of more than 3.0 can be adopted in In addition, the finding also validate the existing correlation
soil with SPT-N < 50. For soil with SPT-N > 50, fsu/N became used in local practise.
almost consistent and fall within the range of 2 to 3. Therefore,
possibility of using higher fsu/N ratio can be considered for soil 4. REFERENCES
with SPT-N < 50. However, load test shall always be carried out to Abdelrahman, G. E., Shaarawi, E.M. and Abouzaid, K. S. (2003).
verify the assumption made in design. Interpretation of Axial Pile Load Test Results for Continuous
In term of base resistance, study has been carried out for two (2) Flight Auger Piles. Proc. Of the 9th Arab Struc. Eng. Conf.,
condition which are pile socket into rock and pile socket into hard Abu Dhabi, UAE.
layer of soil. Relationship between Pb/Ptop and SPT-N blows for all Atkinson, J.H. 2007. The Mechanics of Soils and Foundations. 2nd
the collected load test results have been plotted. Out of 20 numbers Edition, Taylor and Francis (Publishers)
of load test data, 16 number of load tests terminated at rock while Balakrishnan, E. G. (1994). Performance of bored piles in Kenny
the remaining load test terminated in soil. The findings show a good Hill Formation (weathered meta-sedimentary) in Kuala
match of data where the R-squared for both condition are above Lumpur, Malaysia. M.Eng. thesis, Asian Institute of
90%. The equation are as below: Technology, Bangkok, Thailand.
Balakrishnan, E. G., Balasubramaniam, A. S. and Noppadol Phien-
fbu = 154.69 (Pb/Ptop) for soil (4) wej (1999). Load Deformation Analysis of Bored Piles in
Residual Weathered Formation. J. Geotech. Geoenviron.
fbu = 242.5 (Pb/Ptop) for rock (5) Eng., Feb., 122-131.
Burland, J. F. (1973). Shaft friction of piles in clay—A simple
The bar chart as in Figure 11 indicated a variation of Pb/Ptop ratio fundamental approach. Ground Eng.,6(3), 30–32.
in between 0.5% and 40%. The average value for all the collected Bohn, C., dos Santos, A. L and Frank, R. (2016). Development of
load test data is calculated to be 12%. The average Pb/Ptop value Axial Pile Load Transfer Curves Based on Instrumented Load
obtained is almost in agreement to the industry practice where the Tests. Journal of Geotech. and Geoenviron. Eng.., ASCE,
end bearing capacity of bored pile is normally limited to 20% of 143(1).
working load. At Pb/Ptop = 12%, the fbu in soil is calculated to be Borel S, Bustamante M, Gianeselli L, Ponts et Chaussess LC.
1856 kN while fbu in rock is 2910 kN. Fbu of 1856 kN is correlated (2004). An appraisal of the Chin method based on 50
to be 37N in soil. This correlation is within the range recommended Instrumented Pile Tests. Ground Engineering.
by other authors such as Chiu & Perumalswamy (1987) and Toh et Bjerrum, L. (1953). Les pieux de fondation en Norvege. Ann. Inst.
al. (1989). In addition, 37N correlation is also in agreement with the Tech. Batiment Travaux Public., 6(63/64), 375–376.
local practise of 40N. Thus, this study validate the existing Chan, S. F. (1975). An experimental study of behaviour of end
correlation used for end bearing. bearing cast-in-situ piles. Proc., 4th Southeast Asia Conf. on
Soil Engrg., Kuala Lumpur, Malaysia.
Chang, M. F., and Broms, B. B. (1991). Design of bored piles in
residual soils based on field-performance data. Can. Geotech.
J., Ottawa, Vol 28, 200–209.
100
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Chang, M. F., and Goh, A. T. C. (1989). Design of bored piles Meyerhof, G. G. (1976). Bearing capacity and settlement of pile
considering load transfer. Geotech. Engrg., Bangkok, foundations. J. Geotech. Engrg. Div.,
Thailand, 20, 1–18. 102(3), 195–228.
Chiu, H. K., and Perumalswamy, R. (1987). Foundation for Capital Randolph, M. F., and Murphy, B. S. (1985). Shaft capacity of driven
Square, Phase 1, Kuala Lumpur. piles in clay. Proc., Offshore Technology Conf., Offshore
Proc., 9th Southeast Asian Geotech. Conf., Bangkok, Thailand, 2, Technology Conference, Houston.
177–194. Samuel G. P. and Terry A. T. (1999). Extrapolation of Pile Capacity
Horvath, R. G., and Kenney, T. C. (1979). Shaft resistance of rock from Non-Failed Load Tests.
socketed drilled piers. Proc., Int. FHWA-RD-99-170, U.S. Department of Transportation, VA.
Symp. on Deep Foundations, ASCE National Convention, Atlanta, Tan B. K. and Yeap E. B. (1977). Structure of the Kenny Hill
182–214. Formation, Kuala Lumpur and Selangor.
Karlsrud, K. (2014). Ultimate Shaft Friction and Load-Displacement Geology Society Malaysia, Bull. 8, 127 – 129.
Response of Axially Loaded Piles in Clay Based on Toh, C. T., Ooi, T. A., Chiu, H. K., Chee, S. K., and Ting, W. H.
Instrumented Pile Tests. J. Geotech. Geoenviron. Eng., (1989). Design parameters for bored piles in a weathered
140(12), 04014074. sedimentary formation. Proc., 12th Int. Conf. on Soil Mech.
Komoo, I. (1989). Engineering Geology of Kuala Lumpur, and Found. Engrg., Rio de Janeiro, 2, 1073–1078.
Malaysia. Proceedings of the International Conference on Tomlinson, M. J. (1994). Pile Design and Construction Practise. E
Engineering Geology in Tropical Terrains, Bangi, Malaysia. & FN Spon., London., 99 – 165.
262 – 273.
101
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Jobs involving upgrading of old structures into new modern buildings are common in this country. This paper presents the
value engineering and underpinning design for an abandoned structure located in Southern Malaysia. The alternative design involved
optimization of foundation design by assessing the various design elements including the structural and geotechnical capacity of piles. The
common methods of piling design including the alpha and beta method etc. together with adopted parameters are also presented in this paper.
The acceptance criteria for pile load test commonly adopted in Malaysia will also be shared in for reference. The design of Micropile will be
elaborated for both structural and geotechnical capacity. Recommendations on design parameters and the relation with construction
constraints are also be highlighted. Besides, the common misconception on structural design of micropile will be discussed and pile load test
results are presented to substantiate the derived conclusion. Finally, there are many issues commonly faced in construction projects. Among
them, the limitation of SI in determining the pile length, the importance of proper construction planning, working under limiting headroom
and lesson learned will also be shared in this paper.
Keywords: Underpinning, micropile, structural, geotechnical
102
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
where,
Qall = allowable Geotechnical Capacity
Qsu = ultimate Shaft Capacity, fsu As
Qbu = ultimate Base Capacity, fbu Ab
Fs = factor of Safety for Shaft Resistance
Fb = factor of Safety for Base Resistance
Fg = factor of Safety for Global Resistance
Figure 2 Site layout plan As/Ab = pile shaft area/pile base area
103
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 5 Pile Load vs Pile Top Settlement Curve for Bored Pile
104
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
4. CONSTRUCTION
The foundation construction for the project was carried out between
year 2014 & 2015. Soil investigation works were commenced
slightly ahead the foundation construction in order to enable the pile
design to be finalized. Generally, both bored pile and micropile Figure 9 Micropile Rig Working Under Limiting Head Room
activities started almost in parallel while spun piling works only
started midway through the piling construction. Bored piles were
mainly carried out at the building extension areas and also at
locations where demolishing of buildings was carried out.
Micropiles with smaller machines focused mainly within the
existing building areas due to congested site and limiting head room
condition. Spun piles were later carried out for smaller structures
positioned along the perimeter of the proposed building. Similar to
other constructions, many issues were encountered during the actual
physical works on site. Among those are lack of soil investigation
information for decision on pile termination, tight construction
schedule, congested and low head room and also sloping ground
condition etc.
As there was only limited SI carried out on site, the finalising of
actual pile length can only be confirmed by the site personal with
Figure 10 Micropile Rig Working Under Limiting Head Room
reference made to the nearest soil investigation information. Even
with the presence of geotechnical engineer on site, the identification As the site was surrounded by existing structures and roads, part
of suitable materials for pile termination on site has never been easy. of the site that was founded on higher ground needed to be trimmed
Thus, the termination of pile was purely based on the engineers on to facilitate the piling construction. The maximum difference in
site with feedback from the design office. At location where ground level was recorded to be approximately 7.5m. As such, cut
confirmation of suitable pile length was difficult, the approach by slope and temporary shoring works were required to facilitate the
lengthening the pile has to be taken. piling activities and site works. The slope was check to ensure that
Tight construction schedule has always been an issue for most no excessive movement that may result in damage on the completed
construction projects. As the construction involved demolishing part piles. The design was coupled with instrumentation monitoring
of the existing structure to cater for piling works, it requires proper during excavation to ensure nominal movement on the cut slope.
planning in order to ensure work continuity. Figure 8 showing the Figure 11 shows the cut slope formed during construction to
bored pile rig working on the partially demolished structure. facilitate the piling works.
105
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
6. CONCLUSION
Land scarcity especially in the city centre area has resulted the reuse
and upgrading of abandoned and old building structures. An
abandoned complex in Johor was identified to be upgraded to
service apartment with modern shopping centre. Several
underpinning options have been assessed during the value
106
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: In this paper, conversion of measured strain data into pile loads for tensile load testing of two drilled shafts is studied using the
secant modulus of concrete. A back analysis method, considering the possible effect of concrete cracking or slippage between steel-grout
interface, is used in converting the strain into pile loads. Subsequently, the t-z curves along shaft are obtained based on the pile loads interpreted
from back analysis.
Keywords: drilled shaft, tensile loading test, back calculation, hyperbolic model
107
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
� =−
1 ̅ �(�)
��
(8) Class- Unit
�� 2�� �� SPT C
0 Depth Desc-ription ificati weight
N (kN/m2)
on (kN/m2)
in which Pi(z) is the Spline fitted function. The value of b at 0.0~
various depth can be calculated as �(� ) = ��/� �(�), in which �
Backfill SF 1.5~2 - -
� 2.8
is in general between 0.85 and 0.95. 2.8~ Very soft silty
The displacement at any depth of it’s corresponding loading CL 1.5 17.84 20.6
3.5 clay
step is calculated using the following equation: 3.5~ Very soft silty
CL 1~1.5 17.37 24.5
15.45 clay
� 1
� =� − ∫ � ̅ (�
)�� (9) 15.45~
�� �� ���� 0 Soft silty clay CL 2.5~3 17.47 36.3
18.5
18.5~ Soft silty clay
in which �� and �� are the elastic modulus and the cross- CL 3~4.5 17.69 39.2
22.0 and some sand
sectional area of the pile, respectively. The value of � at various 22.0~
depth can be obtained using the following equation: Silty sand SM 9 - -
23.85
Andesite rock,
(� −�� )(1−�(�)� ) 20
��(�
)= �� �� ��
(10) 23.85~ gravel and
��� - ~ - -
38.6 silty sand
50/8cm
mixture
The initial shear stiffness of pile/soil interface is 38.6~
Sandstone - 50/5cm - -
39.3
���= 1/��(�
) (11) 39.3~ Sandstone
- 50/3cm - -
44.0 /Shale
4. Case Studies
The basic information on six collected tensile load testing is Table.3 Subsurface condition at D3 site
given in Table 1. These two piles were installed by reverse
circulation method. The ground conditions of the D1 and D3 sites Class- Unit
SPT C
are given in Table 2 and 3, respectively. The nonlinear stress- strain Depth Desc-ription ificati weight
N (kN/m2)
relationship of the concrete of the case examples at the uppermost on (kN/m2)
gage level is shown in Fig. 2. 0.0~
Backfill SF 11 - -
In order to determine a and b parameters given in Fig. 1, the 2.5
pile axial load along depth at each corresponding loading steps is 2.5~
Firm silty clay CL 4~5 - -
calculated first as shown in Fig. 3 and 4 for D1 and D3 piles, 4.5
respectively. However, the calculated pile axial force becomes 4.5~ Very soft silty
CL 1~1.5 17.63 45.1
unreasonably high at some gage levels when the applied loading 13.0 clay
becomes higher, as shown in these figures. The reasons for such a 13.0~
Soft silty clay CL 2~4 17.38 33.4
high axial load at these gages are most possibly affected by the bond 25.5
strength between concrete and steel interface, which may have 25.5~
Firm silty clay CL 5~8 17.53 43.2
caused slippage between mortar and steel interface or even caused 37.5
concrete cracking. Hence, the assumption of strain compatibility in 37.5~
Stiff silty sand ML 10~14 19.04- 76.5
Eq. (2) does not apply. The measured strain becomes taken by the 41.7
steel only instead of taking by both steel and concrete. It’s the 41.7~
Silty sand - 14 - -
possible reasonable to have such a high axial force at certain gage 43.0
levels. Andesite rock 50/13cm
An interpolation spline function is used to fit the measured data 43.0~
and some silty - ~ - -
to obtain a load distribution with depth shown in Figs. 3 and 4 for 70.5
sand 50/6cm
piles D1 and D3, respectively. The interpolation is based on the gage
readouts of the uppermost, the second set and the bottom gages for
the fitting. The axial load versus displacement relationships of both
piles is shown in Fig. 5.
Figures 6 and 7 show the results of the calculated initial
stiffness and ultimate strength at the pile/soil interface along shaft.
The t-z curves close to the pile head and close to the pile toe are
shown in Fig. 8 and 9, respectively.
Table.1 Tested pile information
Loading Construction
Pile Diameter Depth
Method Method
Reverse
D1 Tensile 1.5m 45.7m circulation
method
Casing
D3 Tensile 1.5m 49.7m
method
Table.2 Subsurface condition at D1 site Fig.1 Assumed shear stress vs relative displacement at the pile/soil
interface
108
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
D1 (north)
Fig. 3 Axial load along shaft of D1 pile
Fig. 6 Predicted Ksi profile
D3 (south)
Fig. 4 Axial load along dept of D3 pile
109
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5. Conclusions
Back analysis method was used in this paper to interpret the
required parameters based on tensile loading tests of two drilled
shafts. More reasonable axial force along pile shaft was fitted using
hyperbolic model. Subsequently, the t-z curves along shaft are
obtained based on the pile loads interpreted from back analysis.
Based on the back calculation of these two tested piles, the
following conclusions can be drawn:
a) The initial shear stiffness at soil/pile interface Ksi of D1 was
higher than that of D3.
b) The failure strength τf of D1 was also higher than that of D3
especially at depth 22m below ground surface.
c) Based on the t-z curves given in Figs. 8 and 9, the D1 pile was
not yet reached the ultimate value. However, hardening and
softening behavior was observed for the D3 pile close to head
and close to toe, respectively.
6. Acknowledgement
The present study was carried out as part of a research project
funded by Ministry of Science and Technology (NSC 102-2221-E-
019-028-MY3), Taiwan. The author is grateful for the financial
support.
7. References
Lam, C. and Jefferis, S.A., (2011) “Critical Assessment of Pile
Fig. 7 Predicted τf profile
Modulus Determination Methods,” Canadian Geotechnical
Journal, Vol. 48, No. 10, 1433-1448.
Lin, S.S., Wang, K.J., Hsieh, H.S., Chang, Y.H., and Huang, C.S.,
(2007) “Field Testing of Axially Loaded Drilled Shafts in
Clay/Gravel Layer,” Journal of GeoEngineering, Vol. 2, No. 3,
123-128.
Xiao, Z. R. Du, M.F., and Zhang, Z (2003) “A Back Analysis
Method to Determine the Parameters for the Pile Foundation
Analysis in Foundation: Innovations, Observations”, Design and
Practice, Thomas Telford, London, UK, pp.951-960.
110
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Geotechnical engineering is one of the oldest branch of civil engineering, which continues to evolve along with the level of
human civilization. In the practice there are much uncertainty in the geotechnical field, both in terms of design and the inherent variability of
the soil material. This uncertainty, coupled with the limited soil data available, makes the geotechnical engineer assumed over estimates or
under estimate the soil parameter. The safety factor (SF) assumptions is one method to reduce the risk of such design error, but the FS can be
misleading as it is usually recommended or used without specific reference. One of the alternative approaches is the Reliability Based Design
(RBD), RBD is a design methodology that summarizes and incorporates key factors of uncertainty. RBD incorporates the cause and effect of
every variable and hence the possible consequences, making the eventual results more accurate. In this study the authors will discuss about
uncertainty analysis of pile settlement based on soil investigation data. The result is calibrated with the data of static axial loading test based
on existing projects in Jakarta. The use of RBD results in a more optimal and economical design, while maintaining the required safety
design consideration.
1. INTRODUCTION the pile. In this study is assumed that the soil around the pile skin
perimeter is bound or moped perfectly on the surface of the pile
Geotechnical engineering is one of the oldest branch of civil through friction and adhesion, each settlement of the pile is assumed
engineering, which continues to evolve along with the level of by the settlement of the soil around and at the bottom of pile. Based
human civilization. Before the 18th century the period of on classical soil mechanic “Poulus HG, Davis EH. Pile foundation
geotechnical engineering development was divided into 4 main analysis and design. New York: John Wiley & Sons; 1980,” soil can
phase, they are pre-classical, classical soil mechanic phase 1, be assumed to be linear elastic, the decline that occurs will be able
classical soil mechani phase 2 and modern soil mechanic to return to its original position with a force that is proportional to
(Skempton, 1985). In the practice, soil as heterogeneous material the modulus of soil elasticity and decreased magnitude
have much uncertainty in terms of design, the inherent variability of
the soil and also other factor such as design method, variations of 2. METHODOLOGY
loads occurring, variations in the strength of the soil material,
inaccuracies in designing, variations in method analysis, the quality The purpose of this paper to find the coefficient of
of the soil investigation, land degradation, soil expansion and uncertainty factors and determine the empirical relationship between
shrinkage and the use of construction methods . This uncertainty, one variable with other variables, key and factor so it is expected to
coupled with the limited soil data available, makes the geotechnical provide a coefficient factors that can make the calculation method of
engineer assumed over estimates or under estimate the soil pile settlement is more accurate. The field information and data such
parameter. Uncertainty in geotechnics has long been recognized as soil investigation data (SPT) and the pile static axial test of three
(ENR, 1963; Casagrande, 1965). Engineers understand directly or projects in Jakarta will be used as input, parameter and boundary in
indirectly that there is always the possibility of failure to achieve analysing the pile settlement in compression with spring stiffness
design objectives during the design life time. Therefore the safety method and elastic settlement analysis by Poulos & Davis, then to
factor (FS) is used in the design as a solution to reduce the risk of develop distribution data and uncertainty analyzes were analyzed
design failure, Safety factor (SF) generally used 2 ~ 3, this value is using random variable with 100x iteration and lognormal
generally adequately assumed in most foundation designs (Focht distribution for 50%, 100%, 150% and 200% from loading plan.
and O'neil, 1985). Safety factor serves as a solution to reduce design
failure, but (FS) can be misleading as it is usually recommended or 2.1 Pile Settlement Analysis
used without reference and is very subjective
One of the alternative approaches is Reliability Based Spring Method
Design (RBD), RBD incorporates the cause and effect of every The pile settlement formula with spring stiffness method is
variable, key factor of uncertainty and hence the possible calculated based refer to Skempton formula below :
consequences, making the eventual results more accurate. In this
paper, the authors will discuss about uncertainty analysis of pile Pile Stiffness :
settlement based on soil investigation data and the result will be
calibrated with the data of static axial loading test based on some (1)
projects in Jakarta, it is expected by using the key factors of Soil Stiffness :
uncertainty which is the initial step of the RBD will make the result
more optimal and economical design while maintaining the safety of (2)
design.
The paper will discuss the first step of RBD that is about (3)
the analysis of the uncertainty of the elastic settlement of pile (4)
foundation based on the soil data of field investigation and pile
static loading result as calibration on several projects in Jakarta, Where :
random number method and statistic and probabilistic method will
K = Stiffness of material
use to develop the data. When pile is loaded the settlement will
E = Modulus of elastiscity
increased. The settlement increase in the pile is caused by 2 (two)
A = Area of element
factors, the compression due to the structure of the pile itself and the
L = Length of element
settlement of the surrounding soil around pile and also the bottom of
f = Soil resistance
111
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
EA EA Where:
f x1 L u1
EA EAL
s = Standard Deviation
f u 2
xi = X value at-i
x 2
L L n = Sample size
(5) ̅ = Mean
Which then assembled into global coordinates K system which is the
sum of Kpile and Ksoil with the following equation: Number of Iteration Cycle
(8)
Where :
L = Settlement at the top of pile
F = Loading on the pile
Es = Modulus of soil elastiscity
d = Pile diameter
Ip = Settlement influence factor based on soil poison ratio
Random Variable
Figure 1 N-Sample vs COV
Random variable is a variable that has a probable value or chance is
a numerical result of a random phenomenon. The probability
characteristics of the random variable have a probability 3. RESULT AND DISCUSSION
distribution, which determines the possibilities occurring in the
intervals determined by the key factors and the key descriptions Spring Method
specified. Random variable can analysis by formula below, but for Based on the analysis from the spring method, there is has graphic
this paper author use excel formula that refer to algoritma Mersenne correlation between Su and N-SPT value for the three projects
Twister to generate radom variable. reviewed based on 50%, 100%, 150% and 200% from loading plan,
the analysis result have been verified using pile static axial loading
( ) (9) test result.
Where :
α = Multiplier
b = Increment
m = Modulus
112
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Loading 50% 100% 150% 200% Table 3 Mean, Standard Deviation and COV Resume analysis
Mean 11.87 6.33 5.30 3.47
STDV 3.76 0.06 1.44 0.59
COV 31.65% 0.91% 27.217% 16.90%
Based on the analysis result from Poulus & Davis formula (Figure
3), provide the correlation between Es and N-SPT values for the
three projects reviewed for 50%, 100%, 150% and 200% from
loading plan and the analysis result have been verified using pile
static axial loading test result :
Thus, the value of X constant for loading phase 50%, 100%, 150%
and 200% from loading plan below :
̅ =1
Xspring COV = 21.21% ~ 27.69%
̅ =1
Xp&d COV = 13.50% ~ 22.75%
113
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
4. CONCLUSION
5. REFERENCES
114
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Extensive geophysical and geotechnical surveys have been conducted at T field since 2006. The area has been known to show
anomalies that could potentially be shallow gas. The most recent geotechnical investigation took place in June 2016. This included one cone
penetration tests with pore pressure measurement (CPTU), new laboratory tests, four dissipation tests and five headspace tests. Dissipation
tests indicated a low to negligible amount of excess pore pressure. Headspace tests revealed that the gas most likely came from biogenic
source. Based on these new data, it was concluded that the effect of shallow gas on the integrity of the foundations is most likely negligible
and was not taken into consideration in the pile capacity analysis. The pile capacity was then re-evaluated using the NGI-05 method. Six
years of ageing was included in the calculations using the NGI-14 method as well as the effect of cyclic loading. The load was taken as the
largest axial pile head reaction for each pile. Taking the axial pile capacity including ageing and cyclic degradation give resistance factors for
axial loading between 2.07 and 2.97 which meet the requirements in ISO 19902:2007.
Keywords: shallow gas, ageing, cyclic loading.
2. GEOPHYSICAL DATA
From geophysical surveys, it was found that the seabed near T
structures is irregular and mildly undulating due to the presence of
numerous small and large seabed depressions, isolated pockmarks
and pockmark clusters.
The T structures are located above what is believed to be a
shallow channel formation, one of which is directly beneath the
structures.
The geophysical surveys have identified numerous seabed
depressions near the existing T structures. Overall, the depressions
vary between 1.7 m and 4.0 m deep. The geophysical surveys
revealed that these types of seabed depressions are related with the
shallow channels and that shallow gas from deeper source migrates
up towards the seabed from the channel edges.
The geophysical surveys also noted numerous gas seeps within
T area. Within larger seabed depressions, gas seeps were scattered
randomly. Near T structure, gas was identified emanating from two
craters: Crater 1 (GS-1) with size of 12 m x 8m x 2.3m and Crater 2
(GS-2) with size of 7 m x 6 m x 1.7 m. The distance of these craters Figure 1 Borehole, gas seepage, seabed depression and platform
to the structure is 6.6 m and 35 m respectively. locations (not to scale)
115
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
116
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The results show that a majority of the gas detected is nitrogen CPTU data with Nkt = 15 and 20. At greater depths the results from
(N2), followed by methane and a small amount of carbon dioxide UU tests show large scatter, but tend to give lower shear strength
(CO2). The tests detected no hydrogen sulphide (H2S), ethane or values than the CPTU results. The available DSS and CAU tests
propane. The percent of nitrogen is high because the samples were indicate that the Nkt for triaxial compression is likely closer to 15
flushed with nitrogen before being sealed. The samples were flushed than 20. The design profile of su,UU was selected based on the
with nitrogen to remove any gases present in the air that might be available data and engineering judgment.
sealed in with the soil sample. The methane gas most likely comes
from biogenic sources, whereas ethane, propane, and other higher
order hydrocarbons most likely come from a hydrocarbon reservoir.
Because there is no indication of ethane or propane, it was
concluded that the gas is most likely from biogenic sources.
The volume of soil sample and differences in permeability
introduce uncertainty to the amount of gas in the headspace. These
tests should, therefore, only be used as a qualitative indication of
whether gas is present in the sample and whether this gas is from a
biogenic source or from a hydrocarbon reservoir.
Based on the data shown in Sections 2 and 3, it is believed that
the effect of shallow gas to the integrity of T2 platform foundations
at T field is most likely negligible and was not taken into
consideration in the pile capacity analysis.
4. PILE CAPACITY
T2 platform is an eight legged jacket structure with 1829 mm
diameter piles made up of five sections. Wall thicknesses range
between 45 mm and 60 mm. Pile depth is 134 mbsf at piles A1, A2,
B1 and B2 and 142.5 mbsf at piles A3, A4, B3 and B4. An MHU Figure 6 Submerged unit weight and Plasticity Index profile
800S hydraulic hammer was used to drive the piles. From the as
built drawings, it was shown that all piles have a batter of 7.125
(1:8). The configuration of T2 platform piles is shown on Figure 5
below.
117
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
4.3 Effect of Ageing The assessment of the effect of cyclic loading on the ultimate
pile capacity was done by evaluating the cyclic load capacity for
Practical experience has shown that the skin friction on piles various load paths with the computer program PAX2 (NGI, 1994).
increases with time, even after full dissipation of the excess pore The degradation of soil resistance for the piles on the tension
water pressures generated by the pile driving. Comprehensive study side will be greater than for the piles in compression because parts
of pile ageing effects is presented by Karlsrud et al. (2014). of a pile in tension will experience two-way cyclic loading (i.e.
Skov and Denver (1988) proposed the following equation: reversal of direction of shear stresses) during the storm. However,
the available load data indicate that the piles do not experience
(t) = (t0) * [1.0 + 10 * log10 (t/t0)] (1) tension. Therefore, only piles in compression are analysed further.
Pile at leg A4 has the largest maximum storm load and pile at
where, leg B4 has the largest cyclic amplitude. The two piles are analysed
(t) : local skin friction at time t for cyclic effects in PAX2. The calculations in PAX2 evaluate the
(t0) : local skin friction at reference time t0 axial capacity of the pile by following a number of load paths to
t : time since pile driving failure.
t0 : reference time since pile driving The results of cyclic analyses show that cyclic loading may
10 : dimensionless soil parameter determined from pile field reduce the soil resistance for piles supporting T2 jacket by around
tests. 11% for the storm loading with 3-hour peak storm duration and the
most relevant loading direction.
This parameter expresses the increase in skin friction for a 10- Figure 8 compares the same PAX2 results with the data of
fold time increase. observed cyclic capacities for pile load tests on clay (Andersen et
For piles in clay, the reference time t0 is taken as 100 days since al., 2013; Karlsrud et al., 1993). The comparison shows that the
almost all pile tests used to calibrate the different calculation cyclic axial pile capacity estimated with PAX2 is quite consistent
methods were fully consolidated after 100 days. with observations from high quality cyclic pile load tests in clay.
The ageing factor Δ10 is a soil parameter that in principle could
be determined from load tests on piles loaded to failure at different
times after driving. The few tests carried out on identical piles, but
at different times, indicate that previous load tests on piles in clay
lead to Δ10 values that are too high. For piles in sand, the opposite
effect is observed.
Based upon the results presented by Karlsrud et al. (2014), the
Δ10 values for piles in clay are calculated as:
118
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
1)
In compression, submerged weight of pile steel offshore engineering and other applications”, ISSMGE-TC
Table 4 lists the resistance factor for each leg based on the axial 209 Workshop-Design for Cyclic Loading: Piles and Other
capacity and loads of the piles. The capacity includes six years of Foundations. Paris, 4 September 2013.
ageing as well as degradation due to the cyclic nature of loading Huang, N. N., Girsang, C. H. and Mansoor, M. R. (2018)
whilst the load is the maximum storm load for each pile. “Managing shallow gas for offshore platforms via fit-for-
It can be seen from Table 4 that all piles have a resistance factor purpose solutions: a Malaysian case study”, Offshore
greater than 2.0 and as such meet the requirements in ISO Technology Conference Asia, paper OTC-28470-MS, Kuala
19902:2007. Lumpur, March 2018.
International Standard Organization. (2007) Petroleum and Natural
6. CONCLUSION Gas Industries - Fixed Steel Offshore Structures. European
Standard EN ISO 19902, CEN, December 2007.
T structures were installed in T field in 2010. This area has been Karlsrud K., Clausen, C. J. F. and Aas, P. M. (2005) “Bearing
known to show anomalies that could potentially be shallow gas with capacity of driven piles in clay, the NGI Approach”, Proc.
geophysical surveys identify several gas anomalies that extend to International Symposium on Frontiers in Offshore
depths greater than 250 mbsf. The most recent geotechnical Geotechnics, Perth Sept. 2005, A. A. Balkema Publishers,
investigation near T structures was carried out in 2016 comprised ISBN 0 415 39063 X.
one CPTU, new laboratory tests, four dissipation tests and five Karlsrud, K., Kalsnes, B. and Nowacki, F. (1993). “Response of
headspace tests. Headspace tests revealed that the gas most likely piles in soft clay and silt deposits to static and cyclic axial
came from biogenic source whilst from dissipation tests it was loading based on recent instrumented pile load tests”, Volume
concluded that only low to negligible amount of excess pore 28: Offshore Site Investigation and Foundation Behaviour,
pressure exists. As a result, pile capacity analysis does not take Society for Underwater Technology, pp549-584.
excess pore pressure into consideration. Karlsrud K., Jensen, T. G., Wensaas, E. K., Nowacki, F. and
The axial pile capacity of eight piles at T2 platform was re- Simonsen, A. S. (2014) “Significant ageing effects for axially
assessed using the NGI-05 method. This method was selected to loaded piles in sand and clay verified by new field load tests”,
capture changes in capacity correlated with plasticity. The effect of
Offshore Technology Conference, paper OTC-25197-MS,
ageing as well as cyclic loading were included in the calculations. Houston, May 2014.
By applying effect of six years ageing, there is an increase of pile Lunne, T., Robertson, P. K. and Powell, J. J. M. (1997) “Cone
capacity of 16.1 to 16.5% and 17.2 to 17.9% in compression and penetration testing in geotechnical practice”, Blackie
tension respectively. However, by applying the effect of cyclic
Academic & Professional, 312p.
loading, the soil resistance may be reduced by 11%. As a result, the
Norwegian Geotechnical Institute. (1994) “Verification and user’s
resistance factors for axial loading for storm loads was found to be manual for computer program PAX2, version 2.1”, NGI
between 2.07 and 2.97 which meet the requirements in ISO report 525286-1, 10 June 1994.
19902:2007.
Skov, R. and Denver, H. (1988) “Time-dependence of bearing
capacity of piles”, Proc. 3rd Int. Conf. on the Application of
7. REFERENCES Stress-Wave Theory to Piles. Editor B. H. Fellenius, Ottawa,
Andersen, K. H., Puech, A. A. and Jardine, R. J. (2013) “Cyclic Canada, pp879-888.
resistant geotechnical design and parameter selection for
119
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
120
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: In this paper, application of a newly developed real-time website-base GIS (Geographic Information System) automatic
monitoring platform in tunnelling is introduced. Some important functions in this platform have included a user-friendly operation interface,
automatic generation of induced settlement contour and its impacts on adjacent structures together with presentation of ground profile close
to the construction site etc. under a well-recognized GIS, such as Google map. By having these details, potential risk during construction
could be categorized and managed. Furthermore, as all data are provided real-time, site condition could be shown immediately and necessary
engineering judgement are thus eligible to be made without any delay. The application of this system in three tunnelling case histories are
demonstrated; the 1st case is tunnelling in central Kaohsiung which is twin-bored tunnel with depth of tunnel centre is approximately at 15 m
to 17 m beneath a 20 m wide main road in the city. The 2nd one is a case in Taipei metro system which has both cut-and-cover and NATM
tunnels, attentions were put on its application in follow-up of displacement changes as well as pore pressure during the construction. The 3rd
case is TBM tunnelling in a ground consists of very thick clay, occasionally with layers of highly permeable sand and high ground water
level in crowded environment city. It aims to completely monitor and control the surface settlement in order to eliminate impacts on adjacent
traffic, structures and daily life of residence. By applying this newly developed platform, risk of failures could be minimized, as it provides a
clear overall view of geotechnical measurement taken on the site from time to time.
Keywords: real-time website-base GIS automatic monitoring, risk categorized and management, tunneling, settlement
121
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
indicates any measured deformation/stress currently beyond alert In order to generate a full view of ground profile, a full ground
level, “no available data in table” is therefore presented. profile was generated by the platform automatically instead of
Since the platform has to be user–friendly, the output can be manually. By doing so, the opportunity of man-made bias or error
possibly presented in the platform directly (refer to Figure 3) or into should be eligible to be reduced. In addition, potential ground risks
formats of EXCEL, CSV or PDF. In Figure 3, it shows name and could also be predicted and prevented in advance. Figure 5 shows
number of instrument, last measured value and related date as well generated ground profile (light colour means “silty sand” and dark
as alert level etc. colour means “silty clay”), it is aware that more reliable and
As explained previously, the platform has a function of contour intensive site investigation program as well as data should assist to
interpretation which can possibly categorize influence zone induced help to generate a more reliable ground profile. Limited and poor
by tunnel construction. Further details will be presented in later quality site investigation would not be able to achieve the purpose
section. stated above in the aspect of risk assessment.
3. APPLICATIONS
Several examples are presented in this paper in order to demonstrate
the performance of newly developed website-based platform. Figure 5 Generated ground profile
122
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
123
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
124
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ERT-Based Leakage Tracing for Dam Safety and its Potential Sliding Surface
H. Wang1, C. H. Hu2, S. H. Hsieh2, and Y. C. Tsai2
1
Facility Management Research Center, CECI, Taipei, Taiwan
2
HCK Geophysical Company, Taipei, Taiwan
E-mail: herschel39@gmail.com
ABSTRACT: Electrical resistivity tomography (ERT) has been efficiently applied to geological investigation, mining exploration,
groundwater probing, and pollution prevention for years. In this paper, the on-site investigation was required to inspect the possible
groundwater leakage through an earth dam during either its full-level storage period or empty storage period. A potential seepage belt could
gradually develop along the residual permeable sandy/gravelly layer, an old abandoned river channel. A downhole ERT-based brine tracing
technique was conducted on the slopes nearby a spillway and a dam levee widening. The brine seepage flow observation could rapidly
reflect the possible direction and path of leakage flow developed in an earth dam in few hours. Monitoring the ERT distribution variations
can effectively identify the leakage spatial path, accumulation zone, and its potential sliding surface on dam slopes.
Keywords: Electrical resistivity tomography, leakage, brine tracing, dam, potential sliding, seepage.
1. INTRODUCTION vary with their water content. The resistivity response from fresh
water usually overlaps the resistivity ranges of soils. However, sea
Watertightness is one of key factors for planning dam engineering. water, as a better conduction medium, provides much lower
The watertightness investigations cover the seepage conditions on resistivity values below the resistivity ranges for both soils and
not only an earth dam itself but also the dam foundations. Seepage rocks. This implies that infusing brine, like a penetrant, into
is highly associated with permeability in geomaterials, geological subsurface can be a potential method to trace groundwater flows
structures, groundwater conditions, sinkholes, and mining pits when using ERT inspection.
(Hung, 1991). In general, sandstone and limestone are attributed to
relatively high permeable geomaterials. A bedding plane could be
regarded as a relatively high seepage channel when dipping to the
downstream side. Fractured rocks nearby faults or fault belts, and
rocks with tension cracks in anticline crests could lead to the
leakage at dam foundations. Reserving water could impact on the
original hydrogeological condition and promote groundwater to
outflow into adjacent low water table watersheds.
A complete watertightness investigation, including the items of
groundwater table and pressure, consists of geophysical inspection,
geological investigation, borehole exploration, and piezometers
(Hung, 1991). Electrical resistivity tomography (ERT in brief
below), one of near-surface geophysical inspection techniques, has
been used for groundwater flow or seepage investigations for years
(Schuster and Krizek, 1978; TGS, 2011; Wightman, et al., 2003).
In this paper, the ERT method is applied to inspect the possible
leakage through an earth dam in Taiwan. Modified from its Figure 1 A schematic representation for ERT inspection
traditional applications, an ERT-based brine tracing technique is
introduced to efficiently identify the seepage trend of groundwater
Table 1 Resistivity range of common rocks, soils, and waters (after
in potential sliding surface both nearby a spillway slope and on a
Loke, 2000)
dam widening slope.
Material Resistivity (ohm×m)
2. OVERVIEW OF ELECTRICAL RESISTIVITY Rocks
TOMOGRAPHY Granite 5×103–106
Slate 6×102–4×107
The ERT inspection technique can trace geological layers, faults,
Marble 102–2.5×108
groundwater pollutants, carves, mining, burial layers, or landslide
Sandstone 8–4×103
mass. The principle of ERT is to develop an artificial potential field
Shale 20–2×103
by probing one pair of current electrodes around a target zone (as
Limestone 50–4×102
shown in Figure 1). Two extra electrodes are used to measure the
Soils
ground potential difference. The measured apparent resistivity
Clay 1–100
image is usually displayed with a visible-light mode, which
Sand (dry) 5×103–2×104
corresponds to the resistivity values varying from 2,000 to less than
Sand (saturated) 2×102–103
1 ohm×m (Loke, 2000; Society of Exploration Geophysicists of
Gravel (dry) 2×104–8×104
Japan, 2014). The resistivity values for sediments composed of silt,
Gravel (saturated) 103–5×103
sand, and rock are more than 5 ohm×m, corresponding to the pink,
Alluvium 10–800
red, orange, yellow, or green zones. Water, pollutants, or metal
Waters
form low resistivity-content zones, corresponding to the light blue
Groundwater (fresh) 10–100
and grey shades.
Sea Water 0.2
The on-site resistivity values are highly associated with mineral
composition, grain size, mineral formation, water content, and ion
concentration (Loke, 2000; Society of Exploration Geophysicists of There are 4 commonly-used inspection types, including pole-
Japan, 2014). Table 1 shows the typical resistivity value ranges for pole array, pole-dipole array, dipole-dipole array, and Wenner-
geomaterials and water. Rocks usually have relatively high Schlumberger array (Loke, 2000; Society of Exploration
resistivity values. Soils have a wide variety of resistivity ranges and Geophysicists of Japan, 2014). Each inspection type has different
125
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
resolutions along the probing array direction and also expresses findings promote the reservoir agency to engage in the positioning
various inspection depth. investigation of leakage on the dam mass.
In practice, choosing ERT deployment is highly associated with An ERT investigation line is chosen at distances of 0k+603 m~
subsurface characterization and environment (Loke, 2000; Society 0k+783 m at the downstream side along the dam top, the most
of Exploration Geophysicists of Japan, 2014). Inspection depth and possible leakage position (Figures 2 and 3). The total investigation
resolution are in terms of electrode spacing and layout length. length is 180 meters. The inspection is conducted at both the high-
Furthermore, a trial and error process on sectional image is used to level storage and low-level storage periods. The on-site images in
determine the optimal layout deployment. In general, the smaller the reservoir show that the water level and waters coverage have
electrode spacing is arranged, the higher image resolution and the significant changes (Figure 4). If one carries out the ERT survey
shallower inspection depth are. Accordingly, the larger electrode along the dam at the different storage levels, such variations on
spacing is used, the lower image resolution and deeper inspection water head could likely lead to amplify the leakage effect on
depth are. resistivity distribution and probably help investigators identify the
leakage zone at the downstream side of the dam.
3. LEAKAGE TRACING CASES
3.1 Leakage throughout an earth dam Dam Top (Elevation 42 m)
The reservoir located in southern Taiwan was constructed in 1953.
Surrounding with scour-prone mudstone and silty sandstone, deposit
rapidly piled up to 3 quarters of its design volume found in 1996.
From 1997 to 2005, an improvement project was launched for
dredging the reservoir sediments, re-constructing the spill and
irrigation pipes through the earth dam, and dam mass improvement
(Sinotech, 2010). Since 2006, several piezometers, water level Clay
gauges, water level wells, inclinometers, flow weirs, seismographs, Sandy silt~ silty sand
and settlement points have been installed to monitor the dam Silty sand
conditions, including pore water pressure, water level, seepage Muddy sandstone
condition, displacement, and surface conditions (Chiou, et al., 2010; Silty mudstone
Liming, 2014; Sinotech, 2010). Coral stone
The 2,380-m long earth dam consists of one main dam (length
250 m) and two minor dams (right section length 300 m and left
section length 1,830 m) (shown in Figure 2). The main dam is Figure 3 Geological profile along the dam mass (after Sinotech,
composed of concrete core, compact soils, and rockfill. Similar to 2010)
the components in the main dam, the two minor dams are also
constructed with compact soils and rockfill except clay core. The
dam elevation is 42 m. The spill pipe intake and irrigation channel (a)
are also marked on Figure 2.
(b)
Figure 2 Top view photo of reservoir and dam (modified from Spill Pipe Intake
Google Map) Reservoir
The geological conditions along the earth dam is plotted in
Figure 3. The clay core dam overlays a 10~20-m thick silty sand
layer which overlaps a monocline consisting of overlayered pervious
muddy sandstone and silty mudstone. The black dashed line also Irrigation Pipe Intake
indicates the concrete core zone.
The annual monitoring report for dam safety shows: (1) the
measured maximum settlement value less than tolerable settlement
10 mm/month; (2) the maximum lateral displacement less than 10
mm; and (3) eight leakage points found at the downstream side of
the dam at positions of 0K+475 m~0K+825 m (Liming, 2014; Figure 4 Field testing conditions: (a). full-level storage photo; (b).
Sinotech, 2010). The pervious geological condition and seepage
low-level storage photo
126
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figures 5(a) and 5(b) indicate the ERT inspection outcomes and grey shades, represent the silty sand and clay layers. The
during the full-level and low-level storage periods, respectively. significant resistivity variation zone is found at the measurement
The estimated geological layers and groundwater table are labelled positions of 30~70 m as shown in Figure 5(c). This pervious silty
with black and white dashed lines, respectively. The relatively high sand layer could be reasonably inferred as the potential leakage zone.
resistivity contour zones, corresponding to the pink, red, orange, In addition, an old landform sketch before the dam construction
yellow, or green zones, indicate the sand or gravel layers. The reveals that the meandering zone of the old river channel right
relatively low resistivity contour zones, corresponding to the blue overlaps this leakage zone.
Figure 5 Electrical resistivity tomography images: (a). Full-level storage period; (b). low-level storage period; (c). resistivity variations
127
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
128
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
sandstone, and mudstone. The permeable and impermeable layers Figure 11 indicates the resistivity background image measured
alternatively occur throughout the dam mass from the ground along survey Line WH-1S. The dotted and dashed lines represent
surface. the layer boundaries defined by both ERT and borehole sampling
and borehole sampling only, respectively. The ERT background
image consists with the findings at borehole WH-1. The top soil
and weathering sandstone layers correspond to the light blue and
green zones, scattering from the ground surface to depths of 3~5 m.
The yellow, light green, and orange zones could indicate the
sandstone, which overlays the sandstone with mudstone, a relatively
low resistivity value zone.
Figure 12 represents the resistivity variation images (more than -
5%) along survey Line WH-1S up to 840 minutes. The resistivity
variations for brine tracing measure the resistivity differences
between the background and the influence from brine transmission
at specific moment. Infusing brine water could lead to the decrease
of resistivity values in the dam slope when it seeps through more
pervious layers. At the first 60 minutes, the intensive variations
occur at the depths of 2 m on the dam slope around the injection
borehole WH-1. Up to 240 minutes, the brine seepage flows
downward along the permeable sandstone layer. A significant brine
accumulation zone, marked as A, steeply seats upon a relatively
impervious layer. In future, this would possibly develop as a
potential sliding surface, about 3~5 meters deep, around the middle
part of the dam slope when extreme rain occurs. After 14 hours, the
brine water moves downward to the slope toe, labelled as B. This
could be another potential sliding surface.
Figure 13 shows the resistivity background image measured
along survey Line WH-2S. The dotted and dashed lines also
represent the layer boundaries defined by both ERT and borehole
sampling and borehole sampling only, respectively. The ERT
background image also consists with the findings at borehole WH-2.
The extremely low resistivity mudstone layer lays over the relatively
high resistivity sandstone layer, corresponding to the yellow zones.
Another mudstone layer overlays the sandstone interlayered with
mudstone, which covers over a mudstone layer.
Figure 14 presents the resistivity variation images (more than -
5%) along survey Line WH-2S up to 840 minutes. The resistivity
variations for brine tracing indicate the resistivity differences
between the background and the influence from brine transmission
at specific moment. Within the first 240 minutes, the relatively high
variations (-20% maximum) occur at the sandstone and mudstone
Figure 9 Geological column at borehole WH-2 interlayer, around at the elevations of 85~100 m and measurement
positions of 15~30 m. As the infusion time increases, the resistivity
Figure 10 shows the ERT-based brine tracing technique image at variation range expands outward, but limited. However, when
borehole WH-1. The brine water is continuously infused into the stopping injecting brine water into the dam mass (14 hours later),
dam mass through the borehole PVC pipe from water tanks. The the expansion phenomenon on the resistivity variation image stops
ERT instrument records the resistivity images before brine infusion and vanishes synchronously. This possibly reveals that the brine
and traces the resistivity variation images up to 840 minutes. water or groundwater is affected with near-borehole diffusion only.
The coverage on resistivity variation depends upon continuous brine
infusion. There is no significant direction or accumulation zone on
the seepage or groundwater flow along investigation Line WH-2S
installed on the dam slope. In fact, the dam slope is covered with
broomy bushes and grass. There is no significant scour or sliding
scars along survey Line WH-2S.
In brief, observing the brine seepage flow could rapidly reflect
the possible groundwater or leakage flow path in specific geological
layers and accumulation zone based on the ERT inspection
technique. Tracing the brine path can also provide the potential
sliding surface information on an earth dam mass in few hours.
129
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
WH1S-BK
WH-1
110
Top Soil+
Weathering Sandstone
105
Elevation (m)
100
Sandstone A
95
WH01-30-BK WH01-60-BK
105 105
Elevation (m)
Elevation (m)
100 100
95 95
90 90
85 85
0 5 10 15 20 25 30 35 40 45 50 55 0 5 10 15 20 25 30 35 40 45 50 55
Distance (m) Distance (m)
(a). ERT variation image after 30-min brine tracing (b). ERT variation image after 60-min brine tracing
105 105
Elevation (m)
Elevation (m)
100 100
95 95
90 90
85 85
0 5 10 15 20 25 30 35 40 45 50 55 0 5 10 15 20 25 30 35 40 45 50 55
Distance (m) Distance (m)
(c). ERT variation image after 120-min brine tracing (d). ERT variation image after 180-min brine tracing
WH01-240-BK 840min-BK
105 105
Elevation (m)
Elevation(m)
100 100
95 95
90 90
85 85
0 5 10 15 20 25 30 35 40 45 50 55 5 10 15 20 25 30 35 40 45 50 55
Distance(m) Distance (m)
(e). ERT variation image after 240-min brine tracing (f). ERT variation image after 840-min brine tracing
Figure 12 ERT-based seepage brine tracing image along investigation Line WH-1S (variation more than -5% only)
130
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
WH2S-BK
WH2 Mudstone
115
Sandstone
110
105
Mudstone
Elevation (m)
100
95 Sandstone/Mudstone 150
45 Layer boundary defined by ERT and borehole
90 50 60
85 Mudstone Layer boundary extended from borehole
80
5 2 35 50 65 90 12 15 20 26 32 45 80 140 ohm×
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70
0 Mudstone 0 0Sandstone
0 0 0 0 0 0
m
Distance (m) Wet Dry
Figure 13 ERT background image along investigation Line WH-2S
WH2-30-BK WH2-60-BK
105 105
100 100
95 95
90 90
85 85
80 80
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70
Distance (m) Distance (m)
(a). ERT variation image after 30-min brine tracing (b). ERT variation image after 60-min brine tracing
WH2-120-BK WH2-180-BK
110 110
Elevation (m)
Elevation (m)
105 105
100 100
95 95
90 90
85 85
80 80
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70
Distance (m) Distance (m)
(c). ERT variation image after 120-min brine tracing (d). ERT variation image after 180-min brine tracing
WH2-240-BK WH2-840-BK
110 110
Elevation (m)
Elevation (m)
105 105
100 100
95 95
90 90
85 85
80 80
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70
Distance (m) Distance (m)
(e). ERT variation image after 240-min brine tracing (f). ERT variation image after 840-min brine tracing
Figure 14 ERT-based seepage brine tracing image along investigation Line WH-2S (variation more than -5% only)
131
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
132
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
WH3H-BK
110
WH-3
Sandstone
105
Elevation (m)
Sandstone/Mudstone
100
Layer boundary defined by ERT and borehole
Sandstone Sandstone
95 Layer boundary extended from borehole
Mudstone
90 5 2 35 50 65 90 12 15 20 26 32 45 80 140 ohm×
10 15 20 25 30 35 40 45 50 55
0 Mudstone 0 0Sandstone
0 0 0 0 0 0
m
Distance (m) Wet Dry
Figure 18 ERT background image along investigation Line WH-3H
WH3-30-BK WH3-60-BK
105 105
E levation (m )
95 95
90 90
10 15 20 25 30 35 40 45 50 10 15 20 25 30 35 40 45 50
Distance (m) Distance (m)
(a). ERT variation image after 30-min brine tracing (b). ERT variation image after 60-min brine tracing
WH3-120-BK WH3-180-BK
105 105
E le v a tio n (m )
E le v a tio n (m )
100 100
95 95
90 90
10 15 20 25 30 35 40 45 50 10 15 20 25 30 35 40 45 50
Distance (m) Distance (m)
(c). ERT variation image after 120-min brine tracing (d). ERT variation image after 180-min brine tracing
WH3-240-BK WH3-300-BK
105 105
E le v a tio n (m )
E le v a tio n (m )
100 100
95 95
90 90
10 15 20 25 30 35 40 45 50 10 15 20 25 30 35 40 45 50 55
Distance (m) Distance (m)
(e). ERT variation image after 240-min brine tracing (f). ERT variation image after 300-min brine tracing
Figure 19 ERT-based seepage brine tracing image along investigation Line WH-3H (variation more than -2% only)
133
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
WH-3V-BK
110
WH-3
105
Elevation (m)
100
Sandstone
95
Sandstone/Mudstone
90 Sandstone Layer boundary defined by ERT and borehole
Sandstone Layer boundary extended from borehole
85
80
Mudstone 5 2 35 50 65 90 12 15 20 26 32 45 80 140 ohm×
0 5 10 15 20 25 30 35 40 45 50 55
0 Mudstone 0 0Sandstone
0 0 0 0 0 0
m
Distance (m) Wet Dry
Figure 20 ERT background image along investigation Line WH-3V
WH3V-90-BK WH3V-150-BK
105 105
Elevation (m)
Elevation (m)
100 100
95 95
90 90
85 85
80 80
0 5 10 15 20 25 30 35 40 45 50 55 0 5 10 15 20 25 30 35 40 45 50 55
Distance (m) Distance (m)
(a). ERT variation image after 90-min brine tracing (b). ERT variation image after 150-min brine tracing
WH3V-210-BK WH3V-270-BK
105 105
Elevation (m)
100
Elevation (m)
100
95 95
90 90
85
85
80
80
0 5 10 15 20 25 30 35 40 45 50 55
0 5 10 15 20 25 30 35 40 45 50 55
Distance (m) Distance (m)
(c). ERT variation image after 210-min brine tracing (d). ERT variation image after 270-min brine tracing
WH3V-330-BK WH3V-390-BK
105 105
Elevation (m)
Elevation (m)
100 100
95 95
90 90
85 85
80 80
0 5 10 15 20 25 30 35 40 45 50 55 0 5 10 15 20 25 30 35 40 45 50 55
Distance (m) Distance (m)
(e). ERT variation image after 330-min brine tracing (f). ERT variation image after 390-min brine tracing
134
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
135
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: In this study, dynamic p-y curves were suggested for pseudo-static analysis from 1g shaking table tests under various loading
frequencies. Based on the results of the model tests, the dynamic p-y curves are highly dependent on the relationships between the natural
frequency of the soil-pile systems and the loading frequency. The dynamic p-y curves were proposed as hyperbolic functions. The ultimate
soil resistance and the initial soil stiffness were developed as functions of the properties of the pile and soil. Based on comparisons with the
existing p-y curves, the proposed dynamic p-y curve was demonstrated to be capable of predicting the behaviour of a pile under dynamic
loads.
Keywords: Dynamic p-y curve, loading frequency, initial soil stiffness, ultimate soil resistance, Pseudo-static analysis.
136
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
4
d 2M ( z) Dr 80%
p (1) 2
dz 2 Dr 40%
0
M ( z)
y pile (2) 0 0.1 0.2 0.3 0.4 0.5
dz Acceleration amplitude (g)
EI
Figure 3 Natural frequency of soil-pile systems at various
where p is the soil resistance, M(z) is the bending moment at depth acceleration amplitudes
z, ypile is the lateral pile deflection, EI is the flexural rigidity of the
pile and z is the distance along the pile. 60
G.L .-12cm(6D)
The bending moment of the pile foundation were calculated by
Eq. (3) using the stain measured by the strain gauges. 40 G.L.-8cm(4D)
)
m
EI (3)
/c 20 G.L.-4cm(2 D)
M N
(
ec
y na
ts 0
is
where M is the bending moment, y is the distance from the neutral er
li -20
axis, E is the Elastic modulus of the pile and ε is the strain of the o
S G.L.-4 cm (2D)
pile.
G.L.-8 cm (4D)
-40
G.L.-1 2cm (6 D)
2.3 Test results
-60
The natural frequencies of the soil-pile systems were determined -0.1 -0 .0 5 0
Pile deflection (cm)
0 .0 5 0 .1
from sweep tests. Fig. 2 shows the typical fourier transform for a (a)
natural frequency of a soil-pile system using sweep tests results. As
shown in Fig. 2, the natural frequency of this soil-pile system is 80
10.86 Hz under an acceleration amplitude of 0.154 g in dense sand. 0.22g 0.3g
60
Fig. 3 shows change in the natural frequency with the 0.154g
0.4g
acceleration amplitude of the soil-pile systems. As the acceleration ) 40 0.098g
m
amplitude increases the natural frequency of the soil-pile systems /c
N
( 20
decreases due to the reduced elastic modulus of the soil. In addition, ec
the natural frequency of the soil-pile system for the relative density na 0
ts
of 40 %, is smaller than that for the relative density of 80 % due to is
re-20 0.4g
the increased stiffness in the soil with a higher relative density. li
oS 0.3g
Fig. 4 shows dynamic p-y curves under various conditions. As -40
0.22g
shown in Fig. 4(a), the dynamic p-y curves are flat near the surface -60 0.154g
and become stiffer with depth because the soil confining stress 0.098g
-80
increases. Fig. 4(b) shows that the lateral pile deflection increases -0.3 -0.2 -0.1 0 0.1 0.2 0.3
with the acceleration amplitude due to greater inertial forces. The Pile deflection (cm)
soil resistance also increases with the acceleration amplitude. (b)
However, the increments of the soil resistance decreased with the
acceleration amplitude. The secant slope of the dynamic p-y curve 80
0.2 S 1.1fn
1.0fn
-60 0.9fn
0.8fn
0.1 -80
-0.3 -0.1 0.1 0.3
Pile deflection (cm)
0 (c)
0 5 10 15 20 25 30
Frequency (Hz) Figure 4 Experimental dynamic p-y curves: (a) with different depth
Figure 2 Typical Fourier transform for natural frequency of soil-pile (b) with various acceleration amplitudes (c) with various loading
system frequencies
137
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
1.2
50
1
45 Dynamic p-y backbone curve at 2D depth
Measured data at 2D depth 0.8 Loose sand (Dr 40%)
40
Soil resistance (N/cm)
5
Figure 6 Determination of ultimate soil resistance
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 3.2 Initial soil stiffness
Pile deflection (cm)
(a) The variation of the initial tangent modulus with confining pressure
50 is represented by an empirical equation proposed by Janbu(1963):
45
σ'
40 R2 = 0.9277 K = APa ( ) 0.5 (8)
Soil resistance(N/cm)
35 Pa
30
25 where Pa is atmospheric pressure (10.13 N/cm2), is confining stress
20 and A is a curve-fitting parameter.
15 A linear regression analysis was used to obtain the best-fit
10
Dynamic p-y backbone curve at 4D depth values for the parameter A.
5
0
Measured data at 4D depth Fig. 7 shows the initial soil stiffness in the transformed
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 coordinates from the model test results. Through a linear regression
Pile deflection (cm)
analysis, the curve-fitting parameters of A were determined to be
(b)
50
1,414.8 and 995.94 for the relative densities of Dr 80 % and Dr 40
45 %, respectively. Based on these results, the empirical equation for K
40
R2 = 0.9750 can be rewritten as Eqs. (9) and (10).
Soil resistance(N/cm)
10
The initial soil stiffness for loose sand:
Dynamic p-y backbone curve at 6D depth
5
Measured data at 6D depth σ' 0.5
0 K = 995.94 Pa ( ) (10)
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 Pa
Pile deflection (cm)
(c)
Figure 5 Dynamic p-y backbone curve at dense sand: (a) at 2 pile
diameter; (b) at 4 pile diameter; (c) at 6 pile diameter
138
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
1500 Y = 1414.8X
Y = K, X = P a(σ/P a)0.5
20
R2 = 0.9240
1000
Loose sand (Dr 40%)
30
Y = 995.94X
Depth (cm)
500 Y = K, X = P a(σ/P a) 0.5
R2 = 0.9632 40
0 50
0 0.5 1 1.5 2
Pa(σ/Pa)0.5 60
Figure 7 Determination of initial slope
70 Measured data
Diameter (cm) 3.2 American Petroleum Institute (API). (1987) Recommended practice
Soil condition Dr 80%, 40% for planning, designing and constructing fixed offshore
Thickness (cm) 0.5 platforms, API Recommended Practice 2A (RP 2A), 17th
edition, pp 68-71
Frequency Bea, R.G., Littom, R., Nour-Omid, S., and Chang, J.Y. (1984) "A
Pile depth (cm) 110 6
(Hz) specialized design and research tool for the modelling of
Flexural rigidity Acceleration near-field soil interactions", 16th Offshore Technology
2764,424 0.154
(kN·cm2) (g) Conference, Houston, pp249-252.
Boulanger, R. W., Curras, C. J., Kutter, B. L., Wilson, D. W., and
Abghari, A. (1999) "Seismic soil-pile-structure interaction
Bending moment (N·cm)
-1000 1000 3000 5000 7000
experiments and analyses", Journal of Geotechnical and
0 Geoenvironmental Engineering, 125, Issues 9, pp750-759.
Broms, B. B. (1964) " Lateral resistance of piles in cohesive soils ",
10
J. soil mech. Fdns. Div. ASCE, 90, pp750-759.
20 Dou, H. and Byrne, P.M. (1996) "Dynamic response of single piles
and soil-pile interaction", Canadian Geotechnical Journal, 33,
30 Issues 1, pp80-96.
Depth (cm)
139
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Kim, Y. H., Jeong, S. S. and Lee, S. J. (2011) "Wedge failure Reese, L. C., Cox, W. R. and Koop, F. D. (1974) " Analysis of
analysis of soil resistance on laterally loaded piles in clay", laterally loaded piles in sand", Proc. of the 6th Offshore
Journal of Geotechnical and Geoenvironmental Engineering, Technology Conference. Houston, Texas, pp473-484.
ASCE, 137, Issues 7, pp53-77 Rovithis, E., Kirtas, E., and Pitilakis, K. (2009) "Experimental p-y
Kim, Y. H., Jeong, S. S. and Won, J. O. (2009) "Effect of lateral loops for estimating seismic soil-pile interaction", Bulletin of
rigidity of offshore piles using proposed p-y curves in marine Earthquake Engineering, 3, Issues 7, pp719-736.
clay ", Journal of Marine Geosourses and Geotechnology, 27, Thavaraj, T., Finn, W. D. L., and Wu, G. (2010) "Seismic response
Issues 1, pp53-77 analysis of pile foundation", Geotechnical and Geological
Kondner, R.L. (1963) " Hyperbolic stress-strain response: cohesive Engineering, 28, pp275-786.
soils", Journal of Soil Mechanics and Foundation Division, Yang, E. K. (2008) " Evaluation of dynamic p-y curves for a pile in
87, Issues 1, pp115-144. sand from 1g shaking table tests", Seoul, Seoul National
National Cooperative Highway Research Program (NCHRP) (2001) University [Ph. D dissertation].
Static and dynamic lateral loading of pile groups, NCHRP Yang, E. K., Kwon, S. Y., Choi, J. I. and Kim, M. M. (2010) "
Report 461, Transportation Research Board – National Prediction of the natural frequency of pile foundation system
Research Council 2001, pp13-21. in sand during earthquake", Journal of the Korean
Novak, M. (1974) "Dynamic stiffness and damping of piles", Geotechnical Society, 26, Issues 1, pp45-54 (in Korean).
Canadian Geotechnical Journal, 11, pp574-598. Yoo, M. T., Choi, J. I., Han, J. T. and Kim, M. M. (2013) "
Dynamic p-y curves for dry sand from centrifuge tests",
Journal of Earthquake Engineering, 17, pp1082-1102.
140
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Design Method for Bottom Single Blade Steel Rotation Pile Foundation:
Case Study in Vietnam
L.D. Dao1 and T.T.T. Nguyen2
1
Department of Civil Engineering, University of Transport and Communications, Hanoi, Vietnam
E-mail: daoduylam@utc.edu.vn
ABSTRACT: Bottom single blade Steel Rotation Pile (SRP) is a large diameter steel pipe pile with a helical blade welded to the edge with a
lot of advantages such as rapid construction, small construction area, especially less vibration, less noise, environment friendly and high
resistance which has been applied recently in the urban of Vietnam. However, how to design SRP foundation in ensuring the conformity with
the specification for bridge design in Vietnam is an important question and the new Design Specification for SRP has been edited. This
article presents the way to find out the method to design SRP foundation in Vietnam as a case study.
Keywords: Steel Rotation Pile, bridge foundation, design method in Vietnam.
141
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
2.2 Basic equation of LRFD method area, the coastal part of Japan, with the surface layer is soft soil,
very soft clay and very deep bearing layer (Figure 3).
In the bridge design specification [1, 2], the basic equation is: In the US, the typical geological bearing soils are mainly sand
Q R
i i i n Rr (1) and gravel. However, relatively depths are not so big, mainly under
20m. In Europe, the bearing layer is mainly rocky and small depths,
in which: mainly less than 20m.
So in general, the geological conditions of Japan are rather
i = load factor: a statistically based multiplier applied to force similar to the geology of Vietnam when comparing to the US and
effects Europe conditions. Then the loading test data in Japan are quite
useful for calibration of SRP design factors.
=resistance factor: a statistically based multiplier applied to
nominal resistance
i = load modifier: a factor relating to ductility, redundancy and
operational importance [1]
Qi = force effect
Rn = nominal resistance
Rr = factored resistance: Rn
142
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
bridge design specification [1, 2]. The SRP foundation should be 0.16 6
optimal designed to support these loads.
5
In conclusion, the load factors, load modifier as well as load
0.12
combinations for SRP design in Vietnam shall be selected as
4
R e la t iv e f r e q u e n c y
N u m b e r o f ca se s
specified in the bridge design specification 22TCN 272-05.
0.08 3
3.3 New Resistance factors for SRP design
2
Resistance factors for SRP are not mentioned in the 22TCN 272-05
0.04
so need to be assigned by calibration of reliability method based on 1
experimental data [2, 5, 6].
The resistance factor of the foundation can be obtained from 0.00 0
formula (2) [5, 6, 8, 11]: 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Ratio of loading test data over the design value of shaft for clay
1 COV COV
0.16 8
Q
2 2
R D D L
QD QL
L Q 1 COVR2 7
(2)
0.12 6
QD QD
QL exp T ln 1 COVR2 1 COVQ2D COVQ2L
R e la t iv e f r e q u e n c y
N u m b e r o f case s
5
Q L
0.08 4
in which: 3
0.04 2
φ = resistance factor ; R = resistance bias factor
1
COVQ coefficient of variation (the ratio of the standard deviation to
0.00 0
the mean) of load ; COVR coefficient of variation of 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Ratio of loading test data over the design value of shaft for sand
resistance; T target reliability index.
Figure 4 Data analysis for pile resistance
D , L dead and live load factor ; QD / QL dead to live load ratio;
QD , QL dead and live load bias factor Using the formula (2) and the defined method in the items 2.1
and 3.3, the main results are shown in Figure 4 for pile tip resistance
There is relationship between reliability index βT and probability and pile shaft resistance.
of failure Pf [3]. βT of SPT method was calculated and was found to The values of resistance factor corresponding to target reliability
be between about 1.5 and 3.0 [5, 11]. A reasonable value of target index, βT are calculated as in Table 3. Here the value at βT = 2.33
reliability index, βT, for single piles appears to be in the range of 2.0 will be chosen for the next step.
to 2.5, corresponding to Pf between 10-1 and 10-2.
Then a reduction factor RD can be calculated by the index from Table 3 Editorial Instructions
the confidence interval. This index is divided the lower limit of the
Number
confidence interval by the average value, and the reduction factor is Resistance factor
Content of data
the ratio of the index for in case of actual number of data and φ
n
enough data. Confidence interval is calculated following equation:
βT 2.00 2.33 3.0 812
Pile tip resistance 0.70 0.62 0.49 1200
CI t (1.0,n 1) / n (3) Shaft resistance 0.53
0.45 0.32 1350
(sand, gravel)
Equation (3) is for a confidence interval of 95% one-sided, using Shaft resistance 0.58
the ratio of the lower limit of the reliability interval for sufficient 0.48 0.33 1500
(clay, mixed clay)
data as reduction factor.
Each resistance factor is re-calculated by SRP loading data The values of reduction factor with the number of data defined
following LRFD methods. About shaft friction, each data of sand in Table 3, are shown in the Table 4 as follows:
and clay is evaluated individually. Shaft resistance is measured by
stress of pile for each layer, so it is able to get some data by one Table 4. Reduction factor
loading test (Figure 4).
Standard
Mean RD
0.16 6 Resistance deviation
(lower CI)
5 Pile tip resistance 1.10 0.30 0.99
0.12
Shaft resistance
1.22 0.57 1.00
R e la t iv e f r e q u e n c y
4
N u m b e r o f ca se s
(sand, gravel)
0.08 3 Shaft resistance
(clay, mixed 1.48 0.79 0.96
2 clay)
0.04
1 Finally, the recommendation resistance factor is shown in Table
5. This value is obtained by multiplying calculated resistance factor
0.00 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
and reduction factor.
Ratio of loading test data over the design value of pile tip
143
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Table 5. Recommendation resistance factor capacity of the pile shaft resistance. The mobilization of the tip
resistance requires a considerably larger settlement than the
Recommendation
Resistance φ × RD mobilization of the shaft resistance. The effects of the negative shaft
resistance factor)
friction on the geotechnical bearing capacity are checked separately,
Pile tip resistance 0.61 ⇒ 0.60
when negative shaft friction is developed or the pile is plugged. The
Shaft resistance (sand,
0.45 ⇒ 0.45
gravel) possible corrosion of the pile does not lower the geotechnical
Shaft resistance (clay, ⇒ bearing capacity.
0.46 0.45 This instruction is mainly concerned with SRP, the geotechnical
mixed clay)
bearing capacity of the pile can be determined in many different
The uplift resistance factor can be also assigned by the same ways which can be roughly divided into direct and indirect methods.
method based on experimental data. Indirect methods include:
Finally, resistance factors for SRP at the strength limit state [19] - Static bearing capacity formulas
for static analysis shall be taken as specified in Tables 6, unless - Empiric methods based on the penetration resistance
regionally specific values are available. - Stress wave analysis without stress wave measurements.
Direct methods include:
Table 6. Resistance factors for SRP (single pile) - Dynamic test loadings
- Static test loadings.
Resistance In design stage the indirect methods are used in designing of the
Condition
factor pile dimension, penetration depth and the construction equipment.
Pile shaft
These are checked on the site using direct methods, usually with
resistance: all soil 0.45 dynamic test loadings.
Nominal Resistance in and gravel The bearing capacity of the SRP consists of the pile tip
Axial Compression, φst resistance and external shaft resistance.
Pile tip resistance:
0.60
sand and gravel 2.1.4 Geotechnical bearing capacity formula
Pile shaft The studies in Japan [9, 12, 13, 14] and item 3.1 have shown the
resistance: all soil 0.45 sufficient data and scientific basis for determining the bearing
Uplift Resistance, φupst and gravel capacity of Bottom single blade Steel Rotation Pile which is also
specified in the specification JRA 2012 [10].
Pile tip resistance:
0.60 The other studies such as in US, Euro and Australia [3, 4, 8, 15]
sand and gravel
are mainly for small diameter or multi-blades rotation (screw) piles
almost used for buildings, not for bridges. However, these studies
3.4. Bearing capacity of SRP also show the main role of screw blade (rotation wing) in the
3.4.1. Determination of new formula for bearing capacity resistance components of the rotation steel pile.
Consequently, in combination with the geological analysis in
The bearing capacity of a pile is determined as follows: the pile item 3.1, the using a part of the bearing capacity formula of JRA
must sustain with sufficient certainty loadings in different loading 2012 to determine the bearing resistance formula of SRP for LRFD
cases after driving, and settlements and horizontal movements must design is acceptable with the above scientific basic.
be within the permissible structural tolerances. The geotechnical bearing resistance formula for SRP is
The bearing capacity of the pile is determined either based on following:
the structural or geotechnical bearing capacity, and the smaller one
is chosen to the design capacity. RR=φRn= φpstRp+φsstRs (4)
The structural bearing capacity of the pile is determined by the
strength of the pile structure. The structural bearing capacity is in which:
checked for the axial loads, bending moments from the horizontal φpst = Resistance factor of pile tip for SRP, static analysis
loads, eccentricities loads. In addition to the requirements of the φsst = Resistance factor of shaft for SRP, static analysis
supported structure the bearing capacity of the pile should be Rp = Pile tip resistance (kN)
considered for buckling, additional loads, such as negative shaft Rs = Pile shaft resistance (kN)
friction and bending of the inclined piles due to the ground
settlements or bending caused by one-sided soil pressure or lateral Range of considering
resistance. In a completed structure the steel pipe pile is usually shaft resistance
filled with soil. The structural capacity of the pile is formed by the
bearing capacity of the steel pipe considering the corrosion
reduction. Permitted material stresses of the pile are determined on 1.0Dw
the basis of the pile material and soil conditions. In bouldery soil Bearing layer
conditions it may be appropriate to reduce the material stresses over around 1.0Dp
Pile tip depth
permitted in normal situations. The recommended minimum (End of Steel pipe)
thickness of the steel pipe pile driven from the upper head is 9mm
[18, 19]. The corrosion of the pile should be considered when
determining the long-term structural bearing capacity of the pile.
Cutout length
So for SRP foundation, the material resistance shall comply with
steel pile structure calculation by the current bridge specification [1,
18, 19]. In addition, this resistance in most cases is usually greater
than geotechnical resistance so in this paper, we pay more attention
Cutout portion
in the geotechnical bearing capacity.
Dwi
The geotechnical bearing capacity is determined according to the
ground conditions, construction and checking procedures. Dp
The geotechnical bearing capacity of the pile consists of the Dw
bearing capacity of the pile tip resistance, and of the bearing Figure 5 SRP dimension and resistance
144
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The pile resistance can be referred from the bearing capacity Pile construction shall be finished on full examination of the
formula of JRA 2012. finishing conditions so as to ensure the bearing capacity. The
The pile shaft resistance can be determined by: finishing conditions should be determined based on results of test
piling operation, and are represented by such factors as embedded
Rs = qsAs (5) depth of the pile, dynamic bearing capacity. Embedded depth of a
pile specified in design documents or confirmed by results of trial
Where: piling test.
Rs:Pile shaft resistance (kN) For the uplift resistance of SRP, it should be estimated in a
qs:unit shaft resistance of pile (kN/m2) similar manner to determinate the shaft resistance as specified
As : Area of pile shaft surface Dp L (m2) above, and it can be assumed that the steel blade behaves like an
Dp: Pile diameter (m), anchor as shown Figure 6.
L: Length of the section to take into account the shaft resistance of Pullout force
the pile upper 1Dw from pile tip (m), (m2) Ground surface
L
qs = 3N (Maximum 150) (6)
For cohesive soils Pullout resistance
at steel blade
H
Bearing layer
qs = c or 10N (Maximum 100) (7) Blade
Where: Figure 6 Uplift of SRP shaft and steel blade
N: SPT blow count, uncorrected for overburden pressure
2
C : cohesion (kN/m ) 3. CONCLUSION
The pile tip resistance can be determined by: SRP foundation for bridge is suitable for transport works especially
in urban area with the key advantages: easy application in narrow
Rp = qpAp (8) space, fast construction, reduce vibration and noise, high quality
control, small influence to nearby monument, environmentally
Where: friendly, recyclable, etc. The efficacy is shown in many reel projects
Rp:Pile tip resistance (kN) including two overpass projects in the capital of Vietnam.
qp : Unit tip resistance of pile (kN/m2) For application of SRP foundation, the specifications for design,
Ap: Area of pile tip = Dw2/4 construction and acceptance are required. To fulfil with the LRFD
Dw: Blade diameter (m) design of bridge superstructure, abutment and pier, the load factors,
load modifier as well as load combinations for SRP design shall be
The nominal unit tip resistance of SRP in bearing layer, in selected as specified in the specification 22TCN 272-05. The
kN/m2, shall be taken as: resistance factor values for SRP can be assigned based on loading
For bearing layer of sand test data according to LRFD method, in compatibility with the
bridge design specification system in Vietnam. The normal
qp = 120N (Maximum 6,000, Dw is 1.5Dp) (9a) resistance for geotechnical bearing capacity can be referred from the
qp = 100N (Maximum 5,000, Dw is 2.0Dp) (10a) formula in JRA 2012 with the carefully study. These above things
are the most important contents of the new LRFD specification for
For bearing layer of gravel SRP design (TCVN 11520:2016) in the study case of Vietnam.
The Vietnamese case studies show that the new method based on
qp = 130N (Maximum 6,500, Dw is 1.5Dp) (9b) LRFD can be established and used effectively for SRP foundation
design instead of ASD method of Japan if need.
qp = 115N (Maximum 5,750, Dw is 2.0Dp) (10b) The uplift resistance, settlement and other checking for SRP
foundation LRFD method design as well as SRP design examples
N: SPT blow count at bearing layer, uncorrected value for calibration between the different specifications, the construction
requirements and technologies will be discussed in the next papers.
145
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
146
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT:
In this research, behaviours of pile foundation models including piled raft and pile group subjected to dynamic loading were investigated
through small-scale experiments. Foundation models consisting 6 piles, with or without batter piles, were used in the experiments. They were
pile rafts (6PR and 6BPR) if the raft was in contact with ground surface, while they were pile groups (6PG and 6BPG) if the raft was not in
contact with ground surface. To create dynamic load acting on the foundation, a vibro-hammer, placed on the raft, was used. The vibro-
hammer can provide vibration load (active shaking) mainly in the vertical direction (called vertical loading) or in the vertical and horizontal
directions simultaneously (called combination loading) by rotating two discs of eccentric mass synchronously in opposite directions or the
same direction. Active shaking tests were conducted on 4 types of pile foundation models (6PR, 6BPR, 6PG and 6BPG) in a consistent dry
sand ground. It was found from the experimental results that the piled rafts are more effective foundation type to decrease settlement and
inclination under dynamic loading than the pile groups.
Keywords: Piled raft, pile group, dry sand, model experiment
1. INTRODUCTION 40 80 80 40 40 80 80 40
Pile foundations supporting structures such as bridges, buildings,
40
40
wind turbines etc. are often subjected to dynamic loads caused by 4 5 6 4 5 6
traffic load, wind load, or wave load. If excessive displacements
80
80
and/or inclination of a foundation structure are caused by the
1 2 3 1 2 3
dynamic loads, the structure could be damaged and could not be
40
40
used as in normal status. Hence, it is necessary to investigate
behaviours of pile foundations under dynamic loading. In this 240 240
research, behaviours of pile foundation models in dry sand ground
30
30
subjected to active dynamic loading were investigated through
experiments. 80 80 80 80
15
2.1 Pile foundation models
(in mm)
Pile foundation models used in Yesim et al. (2017), Vu (2017) and 20 20
Vu et al. (2018) were used in the experiments of this study. The pile
foundation models consist of 6 piles with or without batter piles and P1 P2 P3 P1 P2 P3
a raft as shown in Figure 1. The model pile is close-ended P4 P5 P6 P4 P5 P6
aluminium pipe having an outer diameter of 20 mm, a thickness of
1.1 mm and a length of 285 mm. In battered pile foundations (6BPR Figure 2 Dimensions of pile foundation models
and 6BPG), inclination angle of batter piles is set at 15 degrees. The
6-piled raft 6-battered piled raft
rectangular duralumin raft has dimensions of 240 mm in length, 160
foundation foundation
mm in width, and 30 mm in thickness. Pile positions are shown in
Figure 2.
Figure 3 shows cases of foundation type in the experiments. The G.L
foundation is piled raft (PR) if the raft base is in contact with the
ground surface. In the case of pile group (PG), a gap of 20 mm is set PR
between the raft base and the ground surface. Each model pile is
mounted with strain gauges along the pile shaft to obtain axial P1 P2 P3 P1 P2 P3
forces, shear forces, and bending moments of the pile. The positions P4 P5 P6 P4 P5 P6
of strain gauges are shown in Figure 4. Physical and mechanical
properties of the model pile are listed in Table 1.
6PR 6BPR
6-pile group 6-battered pile group
foundation foundation
G.L gap
= 20mm
P1 P2 P3
PG
P1 P2 P3
P4 P5 P6 P4 P5 P6
6PG 6BPG
Figure 1 Pile foundation models Figure 3 Cases of foundation type in the experiment
147
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Shear strain
gauge
40
Level 4 Level 2
40
Level 5 80
close-ended 40
Level 6 Level 3
35 35
15
(mm)
D = 20 Figure 5 Dimension of laminar box
P1-P3 P4-P6
In order to grasp the mechanical behaviours of the sand, a
Figure 4 Position of strain gauges series of triaxial CD compression tests were carried out (Vu et al,
2018). Five CD tests were carried out with different confining
Table 1 Physical and mechanical properties of model pile pressures, p0, of 7, 17, 27, 50 and 100 kPa. The results of the CD
tests are shown in Figure 6. The internal friction angle, p', at peak
Property Value strength is 42.8 degrees and the friction angle at residual state, r', is
Outer diameter, D(mm) 20.00 35 degrees. The initial stiffness, q/a, increases almost linearly
with increase in the square root of p0. It is seen from Figure 6b that a
Wall thickness, t(mm) 1.1 small amount of negative dilatancy occurs at a very early stage of
shearing followed by a large amount of positive dilatancy. Positive
Length from raft base, L(mm) 255
dilatancy behaviour weakens after the axial strain, a, exceeds about
Cross section area, A(mm2) 65.31 8%.
Silica sand
Moment of Inertia, I(mm4) 2926.2 500
p0= 100 kPa
Deviatoric stress, q (kPa)
-8 p0= 27 kPa
Volumetric strain,
Minimum dry density, dmin (g/cm3) 1.269 -6 p0= 50 kPa p0= 100 kPa
Maximum void ratio, emax 1.103 cyclic
-4
Minimum void ratio, emin 0.663
-2 p0= 100 kPa
Relative density, Dr (%) 82.0
0
Dry density, d (g/cm3) 1.533
2
0 2 4 6 8 10 12 14 16 18 20
Axial strain, a
(%)
(b) Axial strain a vs volumetric strain vol
Figure 6 Results of triaxial CD tests for the sand
148
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
150mm
Acc7
Acc4 Acc2
Acc3 Acc1
connection plate
530mm
Strain gauge
255mm
Dry slica #6
Dr = 82%
( ρd=1.533g/cm 3)
800mm
Figure 7 Pile foundation with connection plate
149
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 12 shows the natural frequency of each foundation model. Figure 12 Natural frequency, fn, of each foundation model
The four foundation models have almost the same values (fn = 14 to
15 Hz).
3. EXPERIMENTAL RESULTS
In this particular paper, the experimental results of dynamic vertical
Test 1. Static loading by weight loading tests of 4 foundation models (6BPR, 6BPG, 6PR, 6PG) are
of vibro-hammer presented and discussed.
gap
= 20mm
P1 P2 P3 PG P1 P2 P3
P4 P5 P6 P4 P5 P6
150
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
2000
6BPR 30
Vert. force, FV
Hor. force, FH
1000 Dead load of 20
Vibro-hammer
500 15
0 10
5
-500
Vert. freq(Hz) 0
-1000
0 120 240 360 480 600 720 840
Time, t (s)
Figure 13 Time changes of input vibration frequency, fV, vertical force, Fv, and horizontal force, FH, in the case of 6BPR
-2
0 30
30Hz 25
4
6BPR
6 20
8
15
10
12 10
14 Vert. disp.
16 5
Vert. freq(Hz)
18 0
20
0 120 240 360 480 600 720 840
Time, t (s)
Figure 14 Time changes of input vibration frequency, fV, and vertical displacement, w, in the case of 6BPR
2000
30
6BPG
Vert. frequency, fV(Hz)
25
Hor. force, FH
Dead load of
1000 20
Vibro-hammer
500 15
0 10
5
-500 Vert. freq(Hz)
0
-1000
0 120 240 360 480 600 720 840
Time, t (s)
Figure 15 Time changes of input vibration frequency, fV, vertical force, Fv, and horizontal force, FH, in the case of 6BPG
151
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
-2
0 30
25
4
6
6BPG 20
8 Interrupted loading
15
10
12 10
14 Vert. disp.
16 5
Vert. freq(Hz)
18 0
20
0 120 240 360 480 600 720 840
Time, t (s)
Figure 16 Time changes of input vibration frequency, fV, and vertical displacement, w, in the case of 6BPG
2000
30
25
Hor. force, FH
1000 20
500 15
Vert. freq(Hz)
10
0
5
-500
0
-1000
0 120 240 360 480 600 720 840
Time, t (s)
Figure 17 Time changes of input vibration frequency, fV, vertical force, Fv, and horizontal force, FH, in the case of 6PR
-2
30
0
Vert. frequency, fV (Hz)
2 30Hz
Vert. disp. , w (mm)
25
4 6PR
6 20
8 Vert. disp. 15
10
12 10
Vert. freq(Hz)
14
5
16
18 0
20
0 120 240 360 480 600 720 840
Time, t(s)
Figure 18 Time changes of input vibration frequency, fV, and vertical displacement, w, in the case of 6PR
152
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
2000
30
Hor. force, FH 25
1000 20
Vert. freq(Hz)
500 15
0 10
5
-500
0
-1000
0 120 240 360 480 600 720 840
Time, t (s)
Figure 19 Time changes of input vibration frequency, fV, vertical force, Fv, and horizontal force, FH, in the case of 6PG
-2
30
0
30Hz
4
6 20
8 Vert. disp.
15
10 Vert. freq(Hz)
12 10
14
5
16
18 0
20
0 120 240 360 480 600 720 840
Time, t(s)
Figure 20 Time changes of input vibration frequency, fV, and vertical displacement, w, in the case of 6PG
3.2 Comparisons of behaviours of the foundation models Figure 24 shows the horizontal force, FH, in the cases of 6BPR
subjected to 30 Hz loading and 6BPG during 30 Hz loading. Amplitudes of FH of 6BPG are
larger than those of 6BPR. In contrast, amplitudes of vertical force,
In this section, the behaviours of each foundation model in loading FV, of 6BPG are smaller than those of 6BPR (see Figure 21). The
step of fv = 30Hz are compared and discussed. larger amplitudes of FH of 6BPG may be a reason for that the
Figure 21 shows the vertical force, FV, in each foundation model vertical displacement of 6BPG is relatively larger although
with elapsed time from the start of 30 Hz loading. amplitudes of FV of 6BPG are smaller than the other cases.
Figure 22 shows the increment of the vertical displacement, w, Figure 25 and Figure 26 show distributions of bending moments
of each foundation with time from the start of 30 Hz loading. of the piles in 6BPR and in 6BPG, respectively, at elapsed time of
Amplitudes of Fv in 6PR, 6PG and 6BPR were about 1600 N 10s from the start of 30Hz loading. Bending moments of the piles in
(Figure 21). The vertical displacement of 6BPR is the smallest 6BPG are smaller than those in 6BPR, although horizontal forces on
followed by those of 6PR and 6PG, indicating that the batter piles 6BPG are larger than those on 6BPR. Larger vertical load in 6BPR
and the raft base resistance suppress the vertical displacement. It compared with 6BPG (see Figures 13 and 15) is one of the reasons
may be judged from Figure 22 that inclusion of the batter piles is that caused larger bending moments in 6BPR. Another reason could
more effective to suppress the vertical displacement. be the effect of load transferred from the raft base to the ground in
Although amplitudes of Fv in 6BPG were smaller than those in the case of 6BPR. The load transferred to the ground increased
the other cases, w in 6BPG were comparable to those in 6PG. It is stresses and stiffness of the soil beneath the raft, which resulted in
difficult to explain this result definitely at this stage. larger bending moments in the piles of 6BPR.
Figure 23 shows the increment of inclination of the raft, , of It is also seen from Figures 25 and 26 that bending moments of
each foundation model with time during 30 Hz loading. Comparison the centre piles (P2 and P5) are very small in both of PR and PG.
of in 6PG, 6PR and 6BPR indicates that the batter piles and the Figure 27 and Figure 28 show horizontal accelerations of the
raft base resistance suppress the inclination of the foundation, vibro-hammer, the raft and the ground surface in shaking of 6BPR
similarly to the effect for suppressing the vertical displacement. The and in shaking of 6BPG, respectively, with elapsed time from the
inclusion of the batter piles (6BPR) has a great effect on the start of 30Hz loading. Horizontal accelerations on the raft (Acc2 and
inclination reduction of the foundation. Acc4) of PR are smaller than those of PG. This result indicates that
153
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
raft base resistance is effectively mobilised to suppress the that in PG. This indicates that the influence of the shaking of the raft
horizontal acceleration of the raft. on the ground around PR is relatively larger than that around PG.
Focusing on the ratio of horizontal acceleration at the ground Similar result is found from the comparison of the experimental
surface to that on the raft, the ratio in PR is relatively higher than results of 6PR and 6PG as shown in Figures 29 and 30.
2000
1600 30Hz
6PG
1200
800 6BPG
400
Vert. force, FV(N)
0
0 2 4 6 8 10 12 14
Time, t (s)
2000 30Hz 6PR
1600
6BPR
1200
800
400
0
0 2 4 6 8 10 12 14
Figure 21 Time changes of vertical force, FV, during 30 Hz loading
-1
0
6BPR 6BPG
Vert. disp. , w (mm)
1 6PR 6PG
2
3
4
4.45
5
6
6.56
7 6.72
30Hz 7.20
8
0 2 4 6 8 10 12 14
Time, t (s)
Figure 22 Increment of vertical displacement, w, of each foundation model with time during 30Hz loading
(deg.)
0.25
0.46
0.50 0.59
0.75
0.87
30Hz
1.00
0 2 4 6 8 10 12 14
Time, t (s)
Figure 23 Increment of inclination, , of each foundation model with time during 30Hz loading
154
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
1600
30Hz
1200
800
Hor. force, FH(N)
400
-400
Figure 24 Time changes of horizontal force, FH, during 30 Hz loading in cases of 6BPR and 6BPG
Figure 25 Distributions of bending moments of the piles in 6BPR at elapsed time of 10s from the start of 30Hz loading
Figure 26 Distributions of bending moments of the piles in 6BPG at elapsed time of 10s from the start of 30Hz loading
155
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(m/s ) 50
6BPR Vibro-hammer(Acc5)
2
40 Raft(Acc2)
Raft(Acc4)
H
30
Ground surface(Acc7)
Hor. acceleration,
20
10
0
-10
-20
-30 30Hz
-40
0 2 4 6 8 10 12 14
Time, t (s)
Figure 27 Horizontal accelerations of vibro-hammer, raft and ground surface in shaking of 6BPR with elapsed time from the start of 30Hz
loading
(m/s )
50
6BPG Vibro-hammer(Acc5)
2
40 Raft(Acc2)
Raft(Acc4)
H
30
Ground surface(Acc7)
Hor. acceleration,
20
10
0
-10
-20
-30 30Hz
-40
0 2 4 6 8 10 12 14
Time, t (s)
Figure 28 Horizontal accelerations of vibro-hammer, raft and ground surface in shaking of 6BPG with elapsed time from the start of 30Hz
loading
(m/s )
50 6PR Vibro-hammer(Acc5)
2
40 Raft(Acc2)
Raft(Acc4)
H
30
Ground surface(Acc7)
Hor. acceleration,
20
10
0
-10
-20
-30 30Hz
-40
0 2 4 6 8 10 12 14
Time, t (s)
Figure 29 Horizontal accelerations of vibro-hammer, raft and ground surface in shaking of 6PR with elapsed time from the start of 30Hz
loading
156
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(m/s ) 50 Vibro-hammer(Acc5)
2
6PG
40 Raft(Acc2)
Raft(Acc4)
H
30
Ground surface(Acc7)
Hor. acceleration,
20
10
0
-10
-20
-30 30Hz
-40
0 2 4 6 8 10 12 14
Time, t (s)
Figure 30 Horizontal accelerations of vibro-hammer, raft and ground surface in shaking of 6PG with elapsed time from the start of 30Hz
loading
A possible reason for this is illustrated in Figure 31. In the case 4. CONCLUSION
of PR, the ground beneath around the raft is hardened by the raft
base pressure. Hence, vibration of the raft is easily transmitted to the Series of active dynamic load tests on 6-pile foundation models
surrounding ground. In contrast, in the case of PG, the ground (with or without batter piles) in dry sand were carried out at 1-g
around the piles is softened by, for example, generation of gaps field to investigate the behaviours the foundations.
between the pile shaft and the ground. Therefore, the vibration of The experimental results indicate that the piled rafts are more
the surrounding ground is attenuated largely compared to that of the effective than the pile groups in suppressing vertical displacement
foundation. and inclination induced by vertical dynamic load. It is also
confirmed from the results that piled raft with batter piles is the most
effective type to reduce settlement and inclination.
5. REFERENCES
Unsever, Y., Matsumoto, T., Esashi, K., and Kobayashi, S. (2017)
"Behaviour of model pile foundations under dynamic loads in
hardened saturated sand", Bulletin of Earthquake Engineering. Springer,
area Netherlands, 15, pp1355-1373.
Vu Anh-Tuan (2017) "Experimental and Numerical Study on
Behaviours of Pile Group and Piled Raft Foundation Having
Batter Piles Subjected to Combination of Vertical and Cyclic
Horizontal Loading", Doctoral Thesis of Kanazawa University.
Vu Anh-Tuan, Matsumoto, T., Kobayashi, S., and Nguyen Tuong-
Lai (2018) "Model load tests on battered pile foundations and
(a) Piled raft finite-element analysis", Int. Journal of Physical Modelling in
Geotechnics, 18, Issue 1, pp 33-54.
softened
157
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Estimating Pile Axial Bearing Capacity by c- Derived from Pressuremeter Test
Tjie-Liong GOUW1
1
Associate Professor, Universitas Katolik Parahyangan, Bandung, Indonesia
E-mail: gtloffice@gmail.com
ABSTRACT: Due to its rather brittle nature, retrieving undisturbed samples of Jakarta cemented greyish stiff clay, often found at a depth of
30 to 120m, is very difficult. Good and reliable effective shear strength parameters, i.e., c’ and ' values, obtained from triaxial test are hardly
available. By modifying cavity expansion theory, Gouw (2017) was able to derive these effective shear strength parameters through
Pressuremeter in situ test stress strain curve. It was found Jakarta cemented clay exhibiting a drained behaviour when loaded. Its effective
cohesion, c’, values are linearly increasing with depths, averaging from around 95 kPa at 20 m to around 475 kPa at 100m depth, while its
effective friction angle ' values are within 20o – 30o, averaging to around 24o. The values found to be similar to the values derived from CIU
triaxial compression test from relatively good undisturbed samples. This paper presents the methodology in deriving the shear strength
parameters and then applying the derived Pressuremeter c’ and ' values to estimate the pile axial bearing capacity through finite element
simulation and comparing it with the commonly known SPT method applied in Jakarta.
Keywords: Pressuremeter, modified cavity expansion theory, effective shear strength parameters, pile axial capacity
1. INTRODUCTION
By far, Pressuremeter test is the only known in-situ geotechnical
testing device capable to generate a stress-strain curve of in-situ soils,
somewhat similar to the stress-strain curve obtained from triaxial or
direct shear test in soil laboratories. By simulating the Pressuremeter,
hereinafter abbreviated as PMT, test through modification of
cylindrical cavity expansion theory and matching the resulting stress
strain curve with the actual PMT data curve, Gouw (2017) was able
to derive effective shear strength parameters, i.e., c’ and ' values, of
Jakarta cemented stiff clay. His research showed that Jakarta
cemented clay, known of its rather brittle nature, exhibiting a drained
behaviour when loaded under the PMT test. The effective cohesion,
c’, values were found to be linearly increasing with depths, averaging
from 95 kPa at 20 m depth to around 475 kPa at 100 m depth, while
its effective friction angle ' values are within 20o–30o, averaging to
around 24o. The values found to be similar to the values derived from
CIU triaxial compression test from relatively good undisturbed
samples. This paper presents the PMT testing principle, the traditional
PMT parameters, the modified cavity expansion formulas used, a case
study in deriving c and of Jakarta cemented clay, and application of
the values obtained to estimate pile axial bearing capacity through Figure 1 Schematic Diagram of Pressuremeter Test (Briaud, 2013)
finite element simulation, finally comparing the result with the
commonly known SPT method applied in Jakarta local practice.
158
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Traditionally six parameters are obtained from the PMT stress- The expansion of cylindrical cavity can be divided into elastic and
strain curve, i.e.: Po, Py, PL, Km, Em, and G (Baguelin et al, 1972, 1978; plastic zone as illustrated in Figure 4. By using Mohr Coulomb failure
Gambin, 1980, 1995; Gambin and Frank, 2009; Clayton et al, 1982; criterion and radial stress vs modulus of deformation, depicted in
Briaud, 1992; Clarke, 1995). The parameters are described below Figure 5, Mecsi derive equations to calculate the cohesion, c, and
(refer to Figure 2 for some notations): friction angle, , of soils, from PMT test data. His equations are:
Po = ’vo ko + uo (1)
• Yield pressure, Py, is the end of the linear curve and the
beginning of the non-linear or plastic part of the PMT curve,
∆P P -P
Km = = y o (3)
∆R RPy -RPo
RPo +RPy
Em = (1+υ) Km (4)
2
• Shear Modulus, G :
Em
G= (5)
2(1+υ)
3. MODIFIED CAVITY EXPANSION FORMULAS Figure 5 Mohr Failure Criterion and Modulus of Deformation
The cavity expansion theory used in deriving the shear strength Relationship (Modified after Mecsi, 2013)
parameteres from PMT test curve is modified from Mecsi work
(Mecsi, 2013) which is elaborated below.
159
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
2.�
�� = (6) 4000
√� (i)
3500
1−��
��
�= (7)
1+��
��
3000
σ β
Es = Eo ( σ r )
The radius where the soil is still in compression is defined as radius PMT DB-09/66
1000
of compression (plastic) zone, , and formulated as: Mecsi:
Mecsi:b = 0.5, c=0,
= 0.5, f= =
c=0, 32o32o
1+si n ϕ 500
2si n ϕ 39 42 44 47 49
σ r +c.cot ϕ
ρ = rc ( σ ) (9) Cavity Wall Radius, R (mm)
ρ +c.cot ϕ
4000
where rc = cavity radius at cavity pressure r and is horizontal or
radial stress at boundary of compression zone which is defined as:
3500
σ ′ho σu
σρ ≈ [1+ξ-√ (1+ξ)2 -2(1-ξ)2βξ ] +σ′ho (10) 3000
β σ ′ho
Radial Stress, kPA
2500
The radial stress inside the compression zone (at radius r ≤ ) is:
2sinϕ 2000
c ρ 1+sinϕ c
σr = (σρ + ) .( ) - (11)
tanϕ r tanϕ 1500
The radial stress outside the compression zone (at radius r > ) is: PMT DB-09/66
1000
data curve of Jakarta cemented stiff clay, especially in the plastic Eyo = my.Eo = my.Em Mecsi: = 0.8, c=0, =18o
Mecsi_a=0.81_c=0_phi=18deg
phase of the curve, i.e. the part after yield pressure P y. To match the my = yield500factor
test data curve, many trial and error were done. However, every trial cy = cavity pressure
39 at and
42 above yield 44 pressure 47 49
could only partially match the PMT data curve and gave different set aye = rigidity factor after yield pressure
Cavity Wall Radius, R (mm)
160
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
0.6
subsoil. Stiff clay layer is found below 20m depth, it exhibits an
0.5 y = 0.6151x-2.06 increasing SPT blow counts with depth, bulk unit weights vary within
0.4
R² = 0.9249 16.5–18.5 kN/m3 (Figure 9). Plasticity index are mostly within 20 to
60%, water contents fall near the plastic limits, with liquidity indices
0.3
less than 0.30, an indication of stiff clay (Figure 10). Void ratios of
0.2 the stiff clay are found to be within 0.70-1.30, it has specific gravity
0.1
of around 2.63, and water content averaging around 35% (Figure 11).
0.0 SPT NN
SPT Bulk UnitUnit
Bulk Weight, g (kN/m
Weight, 3) 3
g (kN/m )
1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 00 1010 2020 3030 4040 5050 60 60 0 02 24 46 68 810 10
12 12
14 14
16 16
18 18
20 20
22 22 0 20 42
cy /Py 00 0 0 0 0
curve calculated from the modified equation (8), i.e. modified E -30
-30 -30-30 -30 -30
function or modified cavity expansion model. The result shows that -40 -40 -40
-40
when the stiff clay is still in linear “elastic” range, the shear strength
-40 -40
consists both cohesion and angle of internal friction (since the shear -50
-50 -50
-50
-50
-50
strength parameters are derived from Pressuremeter, it is notated as -60
-60
-60 -60
-60 -60
cPMT and PMT). However, once the soil entering non-linear plastic
Depth (m)
Depth (m)
Depth (m)
Depth (m)
Depth (m)
Depth (m)
-70 -70 -70
part, the stiff clay lost its cohesion (cyPMT = 0 kPa), and only the angle -70 -70 -70
of internal friction yPMT is working. It is also found that the angle of -80
-80
-80
-80
-80
-80
internal friction remains constant throughout the elastic and plastic -90 -90 -90
phase, i.e. PMT = yPMT. The same outcomes are found from all the -90 -90 -90
2500
Calculated by Modified Cavity Expansion Wp Wn WL (%) Wp PlasticityWL
Wn Index,(%)
PI (%) Liquidity Index, LIIndex, PI (%)
Plasticity
0 10 20 30 40 50 60 70 80 90 1000 10 20 0 30 40
20 50 4060 70
60 80 8090 100
100 0.000.1 0.2 20
0.3 0.4 0.5400.6 0.7 60
0.8 0.9 1.0
Before yield: 0 0
80 100
0 00
2000
CPMT = 83 kPa - PMT = 21.2o
After Yield: -10 -10 -10 -10
-10
CyPMT = 0 kPa - yPMT = 21.2 o
-20 -20 -20 -20
-20
1500 -30 -30 -30 -30
-30
PMT DB-09/66
-40 -40 -40
-40 -40
"Modified Cavity_Expansion_Theory"
1000 -50
-50 -50 -50
-50
(m)
Depth(m)
Depth (m)
39 41 43 45 47 49
-80 -80 -80
-80 -80
Cavity Wall Radius, R (mm)
-90 -90 -90
Figure 8 Good Match of PMT Test Data vs Modified Cavity -90 -90
161
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Depth (m)
Depth (m)
-60 -60 -60 -60
Depth (m)
-60 -60 -60
Depth (m)
-130 -130 -130 -130 Eff. Overburden Pessure -130 -130 -130
Water Content, Wn (%)
0.0 2.30.5
Specific Gravity, Gs Void Ratio, e
2.4 1.0 1.5 2.72.0
2.8 2.5 3.0
Degree
Water of Saturation,
Content, Wn (%) Sr
Figure 12 Pre-Consolidation
Void Ratio, e
Pressures and Oedometer Modulus
Degree of Saturation, Sr
0 50 60 70 80 90 100 2.2 2.5 2.6 0 103.5
20 30 4040 50 60 70
50 60 100 100 0.0
70 808090 90 0.5 1.0 1.5 2.0 2.5 3.0 3.5 40 50 60 70 80 90 100
00 0 0 0 0
-10
-10 -10 -10 -10 -10
-20
-20 -20 -20 -20 Drained Cohesion _TXCU
-20C'TXCU (kPa) Drained Friction Angle 'TXCU (degree) Undrained Coh
-80
-80 -80 -80 -80 -40 -80 -40 -40
-90
-90 -90 -90 -90 -50 -90 -50 -50
Depth (m)
-100
Depth (m)
Figure 11 Specific Gravity, Water Content, Void Ratio and Degree -100 -100 -100
of Saturation
-110 -110 -110
C'TXCU (kPa) 'TXCU (degree) Undrained Cohesian_TXCU CuTXCU (kPa) Undrained Friction uTXCU (degree)
Figure 12 shows the pre-consolidation pressure and oedometer
Drained Cohesion _TXCU Drained Friction Angle
-120 -120 Angle_TXCU -120
modulus. The0 pre-consolidation
50 100 150 200 250 pressures
300 appear
0 10 increasing
20 30 with
40 50 0 50 100 150 200 250 300 0 10 20 30 40 50
0
depth. Comparing
0
with the corresponding effective
0
stresses, the over 0
-130 -130 -130
consolidation
-10 ratio of the stiff clay layers is
-10found to be in the order -10 -10
of 2.0. The effective and total shear strength obtained from triaxial
CIU tests -20
are shown in Figure 13. -20 -20 -20
162
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
2500 4500
4000
2000
3500
3000
Radial Stress, kPA
2000
1000
1500
0 0
36 38 40 42 44 46 48 50 36 38 40 42 44 46 48 50
Cavity Wall Radius, R (mm) Cavity Wall Radius, R (mm)
1800
4500
1600
4000
1400
3500
1200
3000
Radial Stress, kPA
1000
2500
800
2000
600
1500
"Mod. Cavity_Expansion_Theory"
200 "Mod. Cavity_Expansion_Theory"
500
0
0
34 36 38 40 42 44 46 48 50
34 36 38 40 42 44 46 48 50
Cavity Wall Radius, R (mm)
Cavity Wall Radius, R (mm)
4000
4000
3500
3500
3000
3000
Radial Stress, kPA
2500
Radial Stress, kPA
2500
2000
2000
1500
1500
PMT DB-04/56
1000
1000 PMT DB-03/86
"Mod. Cavity_Expansion_Theory"
500 "Mod. Cavity_Expansion_Theory"
500
0
0
36 38 40 42 44 46 48 50
36 38 40 42 44 46 48 50
Cavity Wall Radius, R (mm)
Cavity Wall Radius, R (mm)
163
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Po, Py, PL (kPa) 'ho(kPa) = P'o = Po - uo depth, with a value of around 95 kPa at a depth of 20 m to 475 kPa at
Em (kPa)
0 1000 2000 3000 4000 5000 0 100 200 300 400 500 600 a depth
0 of 100
20000 40000 m,60000
and 80000
it is 100000
clearly higher than the values obtained
0 0
from
0
CU triaxial test, be the undrained or drained cohesion. The lesser
-10
Po Py PL -10 values of cohesion from triaxial tests are generally attributed to the
-10
-20 -20
brittle nature of the Jakarta cemented stiff clay which tends to suffer
-20
micro cracks resulted from the sampling process by thin wall tube
sampler and during the preparation of the samples in the laboratory.
-30 -30 -30
-40 -40 The higher values of cPMT is attributed to the cemented nature of the
-40
-50 -50
Jakarta stiff clay.
-50
Depth (m)
-60 -60 -60
Depth (m)
Depth (m)
Depth (m)
Depth (m)
-60 -60
Em, Eoed (kPa) -70 -70
Thousands
0 50 100 150 -80 -80
0
-90 -90
y = -0.2106x
-10 -100 R² = 0.9099 -100
Eoed
Em -110 CPMT from PMT data -110
-20
Linear (Em)
PMT from PMT Data
-120 C' from Triaxial CU -120
-30 ' from Triaxial CU
-130 -130
-40
cultimate (kPa) = c after Py ultimate (degree) = after Py
-50 0 100 200 300 400 500 600 0 10 20 30 40 50 60
0 0
-60
Depth (m)
-10 -10
-30 -30
-80
-40 -40
-90
-50 -50
y = -0.0011x
Depth (m)
Depth (m)
-70 -70
-110
-80 -80
-120
-90 -90
rc > Py → cPMT = 0
-130 -100 -100
Py < rc < Py → PMT constant
Figure 15 Pressuremeter Parameters and Oedometer Modulus -110 CPMT from PMT Data -110
PMT from PMT Data
-120 -120
C' from Triaxial CU ' from Triaxial CU
-130 -130
Figure 16 and 17 show the c and values derived from PMT
data, notated as cPMT and PMT, plotted against effective (drained) and
total (undrained) c – from CU triaxial test, respectively. It can be Figure 16 cPMT and PMT vs Triaxial Drained c’ – ϕ’
seen the c- values derived from PMT data by using modified cavity
expansion give a clear existence of soil cohesion when the stress From all the above phenomena, it can be concluded or at least
strain of the stiff clay is still within the linear “elastic’ range, i.e. cPMT postulated that for Jakarta stiff clay, at the initial stage of
and PMT are mobilized at the same time (since the c and are derived Pressuremeter test the soil is in partially or near drained cohesion, as
from PMT, they are given PMT indices). However, once the stress the radial stress and strain reaches its yield pressure, P y, the stiff clay
level reaching and above its yield stress level the stiff clay losses the is already in fully drained cohesion. The explanation is: at the initial
cohesion (cyPMT = cultimate = 0), what remain thereafter is the angle of stage, while the radial stress tends to reduce the soil volume, the
internal friction which remain constant throughout all the stress level concurrent induced tangential strain will expand the soil radially,
(ϕyPMT = ϕ peak = ϕultimate). The same outcomes are found from all the therefore the soil is not in a fully compressive nature, but rather in a
PMT test data. This means Jakarta stiff clay exhibits no dilation radial and tangential ring like shearing nature. Consequently, at this
property (ϕpeak - ϕultimate = 0). stage the soil at least is in a partially drained condition. At and beyond
yield pressure, the induced tangential strain will be large enough to
Comparing Figures 16 and 17, from 27m to 97m depth the PMT cause spacings within the clay particles move to a larger distance one
values are within 21o – 33o and these values fall within the drained another and possibly creates micro cracks within the soil structure,
angle of internal friction (Figure 16) rather than the undrained angle hence the clay start to lose its cohesion and left only with its angle of
of internal friction (Figure 17) obtained from triaxial test. The results internal friction, at this stage the stiff clay is already in a fully drained
also show the cohesion parameter of Jakarta stiff clay increases with condition. This postulated phenomenon is illustrated in Figure 18.
164
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
-60
Depth (m) -60
-70 -70
The shear strength and the deformation modulus of the stiff cemented
-80 -80 clay obtained from PMT data are applied to estimate pile axial bearing
-90 -90
capacity through finite element analysis by using the axisymmetric
y = -0.2106x model in Plaxis 2D software. The input parameters are presented in
R² = 0.9099
-100 -100
Table 1. The finite element model is shown in Figure 19. Figure 20
-110 CPMT from PMT data -110
PMT
shows the resulted pile load settlement curve. By applying the
from PMT Data
-120 Cu from Triaxial CU -120 ultimate load criterion set in the Indonesian Geotechnical standard
u from Triaxial CU
(SNI 8640:2017) which set the ultimate load as the load at pile head
settlement of 4% pile diameter, the ultimate pile capacity can be
-130 -130
-20 -20
-30 -30
-40 -40
-50 -50
Depth (m)
Depth (m)
-60 -60
-70 -70
-80 -80
-90 -90
rc > Py → cPMT = 0
-100 -100
Py < rc < Py → PMT constant
-110 CPMT -110
from PMT Data
PMT from PMT Data
-120 -120
CU from Triaxial CU U from Triaxial CU
-130 -130
165
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
10
20
30
40
50
60
0
0
-5
-10
-15
-20
-25
-30
-35
-40
Deptn, y (m)
-45
-50
-60
-85
Qult (kN) = m Ns As + n Nb Ab (17)
-90
base; As is the pile skin area and Ab is the pile base cross sectional -100
area.
Idealised DB-01 DB-02 DB-03 DB-04
Based on this approximate SPT formulas commonly adopted in DB-06 DB-07 DB-08 DB-12 DB-13
Jakarta practice, the ultimate bearing capacity of the same pile size Figure 21 Idealized SPT Blow Counts
found is:
Qult_SPT = 30,610 kN Further research is necessary to make sure whether the theory
derived in this study can be applied to estimate the strength
It can be seen the PMT and the SPT results give similar values parameters of other soil types. It will be good if PMT test data can be
of estimated pile axial capacity. done in conjunction with instrumented pile load test data tested to
failure, with this the theory can be further verified.
166
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
167
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Volume Measurement for Heavy Pounding Induced Craters and Ground Heave
J. L. Hung1, U. H. Amalia1, H. C. Shih2, and J. C. Chih3
1
Department of Civil and Construction Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan
2
Taiwan Building Technology Center, National Taiwan University of Science and Technology, Taipei, Taiwan
3
Li-Jia Engineering Company, Taipei, Taiwan
E-mail: amaliaula@gmail.com
ABSTRACT: Traditionally, the volume of dynamic compaction induced crater and ground heave is measured by means of level surveying
and ruler measurement. However, since ground heave around the crater and the shape of the crater itself are irregular, it is not only difficult
but also time-consuming to accurately measure the volume of crater and ground heave. This study proposes a method that adopts the up-to-
date image processing (photogrammetry) technology to more accurately measure the crater volume and the ground heave around it. A
commercial software, which is initially used for the drone, is used here to generate point cloud of the crater and its surrounding area using the
images captured with a video camera or smartphone. The accuracy of this method was calibrated with a known volume box in the laboratory
first before it was used in a field trial test. This paper will present and discuss the operation procedure and image processing of this method.
The crater volume measured from the photogrammetry method is compared with that measured from the traditional measuring method. It is
found that the volume of DC crater can be better approximated by cone shape crater than by truncated cone shape crater, which is commonly
used in the DC industry and can be seriously over-estimate the actual volume of DC craters.
Keywords: crater volume, ground heave, dynamic compaction, photogrammetry method.
1. INTRODUCTION
Ground improvement by heavy poundings like dynamic compaction
(DC) and rapid impact compaction (RIC) is a commonly used
method for the densification of in-situ sandy soil to a large depth.
After pounding, a crater on the ground is formed, and surface
heaved. To evaluate the effectiveness of pounding, the volume
change of crater and ground heave before and after pounding needs
to be measured. Traditionally, the volume of crater and ground
heave was measured using level surveying. However, since ground (a) Dynamic compaction (b) Rapid Impact
heaves around the crater and craters itself are irregular in shape, it is (DC) Compaction (RIC)
difficult to be accurately measured with traditional surveying
Figure 1. Dynamic compaction and rapid impact compaction
method. This study will use the up-to-date image processing
methods
technology (photogrammetry) to measure the crater volume and the
ground heave around it with reasonable accuracy. A commercial
software (Pix4Dmapper) which is initially used for the drone is
adopted here to generate point cloud of the crater using the images
captured with a smartphone camera. In this study, the accuracy of
this method was calibrated with a known volume box in the
laboratory first before it was used on a pilot test project of DC and
RIC densification on a land reclamation site, which was reclaimed
by hydraulic fill method.
168
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
10 m 10 m
The traditional survey work carried out on this site includes the
ground surface elevation survey around the crater and the volume of Figure 5. Field surveying work for ground heave and crater volume
the crater. The ground surface survey was carried out by a level
survey on the mark points set on the ground surface (Figure 5). The The evaluation of densification effect is typically done by using
crater depth might be measured when the hammer is still inside the the relationship among accumulated pound-in volume and heave
crater or lifted out of the crater. Its purpose is to calculate the heave volume (Note: plus backfill volume if backfill material is added
and sag volumes around the crater. The area that calculated in the during pounding) and some impacts. Tang (2016) measured the DC
crater volume approach is the area of a grey area in Figure 5. The crater volume in each impact pounding. As shown in Figure 6c, the
heave volume is calculated where the heave of the ground surface is heave volume remained unchanged until it reached the 27th impact;
significant. after 27th impact, the ground heave began to increase, and the
densification effect of DC began to decrease. This information can
be provided to engineer to decide the optimum number of DC
Vh h1 ( D 2 3 a D ) (1)
impact from the pilot test. However, there are two distinct problems
3
here. Firstly, it needs to do the survey work after each DC impact.
Where: As shown in Figure 6b and Eq. 5, to obtain the pound-in volume and
Vh = heave volume (m3) heave volume of the ith impact pounding, an ith run of survey needs
h1 = highest height of measured heave (m) to carry out, and the measurement results of (i-1)th impact pounding
D = distance of affected heave area (m) are needed. It is a very time-consuming surveying process for the
a = radius of the crater (m) pilot test. Secondly, it is not easy to measure the volume of crater
and ground heave with reasonable accuracy using the traditional
So the pound-in volume resulted by the DC pounding (Figure surveying method (Tang, 2016).
6a) can be obtained to evaluate the densification effect of DC to the
ground soil.
V p , i (Vc , i Vc ,i 1 ) Vh , i (3)
V p Vc V h (2)
Where:
Vp,I = pound-in volume at ith number of pounding
Where:
Vc,I = crater volume at ith number of pounding
Vp = pound-in volume
Vc,i-1 = crater volume at (i-1)th number of pounding
Vc = crater volume
Vh,I = heave volume at ith number of pounding
Vh = heave volume
169
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
surveying method for the crater. Preferably, this can be done with
simple devices and at least effort.
2.2 Photogrammetry method
Recently, many image processing programs are available for UAV
or drone to map specific area or landscape from the sky. These
programs use the photogrammetry technology to process the
captured images. The digital images of an area captured by drone
were put into the software to get the 3D digital mapping of the area
(a) including the elevation and plane dimension of the building or
landscape. The process of photogrammetry starts with capturing
images from at least two different locations, then common points are
identified on each image and rays line are developed from each
camera angle to points on the object. These rays are intersected to
produce 3D coordinates of the points of interest. (Li, 2016)
Area captured
Drone by drone
(b)
Landscape Area
3. IMAGE PROCESSING
3.1 Point Cloud Model
The core of photogrammetry method is to generate a 3D object
model by using the captured surface images of the object. The image
processing software converts images into 3D models and point
clouds. Having the point cloud data (e.g., the crater), the volume of
the crater and the elevation change of the surrounding ground
surface can be calculated.
3.2 Image Acquisition Plan
To establish the point cloud of the object, the number of digital
images taken on the site should be from various angles to visualize
its point clouds. Then the point clouds are used to reconstruct the
model from the initial digital image. To obtain good images for
processing, digital images should be taken around the object. It
requires each image to overlap with its neighboring image to get
high accuracy results. The higher the overlapping image rate, the
more clear model can be constructed. The sufficient distance
(ground sampling distance, GSD) of taking digital images should be
carefully considered, and GSD serves as necessary measurement
Figure 7. Problems associated with crater volume measurement in calculation. Therefore, it is crucial to make a good image acquisition
the traditional measuring method plan to capture images at various angles and distances. (Pix4D, 2017)
The GSD is the distance between the center of two consecutive
As shown in Figure 7, the rim of crater appears on the ground pixels on the ground. It influences the accuracy and the quality of
is not the actual rim of the crater. Therefore, the measured crater the results as well as the details that are visible in the final results.
diameter will be larger than the actual crater diameter. As a result, it GSD related to the camera height and camera properties. Therefore,
will overestimate the crater volume and the densification effect of
the image acquisition plan should maintain at least two different
the DC. There is a need to have a more accurate and quicker camera height. While the image acquisition plan for the crater
170
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
depends on the GSD and overlap, the overlap of pictures depends on Figure 10. Generated point cloud model by the image acquisition
the affected area of the crater. In this case, crater volume and ground plan
heave will be measured; overlapped images should contain all the
area affected by the pounding. For crater volume and ground heave Following are the standard operation procedures (SOP) of digital
measurement, the following circular image acquisition plan is image processing adopted in this study:
recommended: (1) Input digital images
(1) Moving around the crater for the first time with a camera dip- Based on the image acquisition plan, digital images are
angle of 60o. obtained from the site. Input digital images to the Pix4Dmapper
(2) Moving around the crater for the second and/or third time while software. Select Processing option of 3D Maps, and start
reducing the camera height and the dip-angle in each round. processing.
(3) It is best to take an image every 5 to 10 degrees when walking (2) Processing Step 1: Initial Processing
around the crater and take pictures to ensure enough overlap. Select the Initial Processing option, and click start. When start
More images should be taken for shorter distances. Step 1, Pix4Dmapper computes the key points on the images. It
Recommended overlap is at least 75% of common points. The uses these key points to find matches between images.
camera should be at a constant height above the crater as much (3) Processing Step 2: Point Cloud and Mesh
as possible to maintain the GSD value. After Step 1 Initial Processing is completed, Step 2 Point Cloud
and Mesh can be processed. Select Point Cloud and Mesh
option. Click start to begin Step 2 processing. In this step, the
3.3 Digital Image Processing
point cloud is generated and can be visualized in rayCloud.
Since the images acquisition needs to change camera heights and (4) Scale the Model
angles, a length adjustable monopod was used to take pictures while After finish Step 2, the initial point cloud data are generated.
walking around the crater. The images taken must cover the whole However, these data need to be adjusted using the scale or
crater and also the possible heaving area around it. To enhance the ground control points (GCPs) on site. Once the point cloud has
stability of smartphone/digital camera and to assure better image been adjusted with the scale, it needs to be reoptimized. After
overlapping, a DJI OSMO Mobile® was used here. OSMO re-optimization, the processing of Step 2 is re-generated.
Mobile® can also maintain the camera angle by its built-in gimbal (5) Processing Step 3, DSM, Orthomosaic and Index
and facilitate the process of image taking on site. DSM (Digital Surface Model) is a 2.5D model of the mapped
area. Each pixel and each point of the vector point cloud
contain (X, Y, Z) information. Running Step 3 means that there
is information about the coordinate of the measuring object.
60°
After processing Step 3, the volume of the crater can be
obtained. Select the area of interest and click Compute.
Start
45°
h2 Image acquisition in
the site
Point of image taking
h1
Input images to
Area affected by pounding Pix4Dmapper
Processing Step 3.
DSM, Orthomosaic
Number of image and Index
input with different
height and camera Assign area to
dip-angle
calculate volume,
compute.
Finish
171
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
4. VOLUME MEASUREMENT Table 3. The standard deviation of the box volume measurement
Standard % of
4.1 Calibration Box Volume Volume Mean Deviation
Deviation Error*
No.
To calibrate the accuracy of volume measurement, a box with (cm3) (cm3) (cm3) (cm3) (%)
known dimension is used here as the sample object to check the 1 2947.32 11.95
2 2939.70 4.33
accuracy of volume measured by this photogrammetry method. The
3 2880.56 2935.37 -54.81 42.13 1.46
dimension of the box is 28.5 x 16.5 x 6.8 cm (Figure 12) with the 4 2914.69 -20.68
inside volume of 2894.74 cm3. Unfortunately, the default unit 5 2994.58 59.21
“meter” used in Pix4Dmapper cannot be changed on the computer 42 . 23
* % of error 100 1 . 46 % 1 . 5 %
screen, and the output value of measured volume can only be 2894 . 74
displayed to two decimal places. As a result, the output volume of
the calibration box from Pix4Dmapper is 0.00 m3. It is unacceptable
4.2 Photogrammetry Measurement of Craters
for the volume measurement here. To deal with this problem, the
default unit of the software is disregarded, and the metric unit in the In the pilot test site, both DC and RIC craters were measured with
software is regarded as “cm”. Fortunately, this change on the default the photogrammetry method following the above mentioned image
unit did not affect the operability of volume calculation of this acquisition plan. Not only the crater volume but also the ground
software. Figure 13 shows the volume determined for the box using heave were measured. The captured images covered an area which
the Pix4Dmapper software. was large enough to cover the ground control points (GCPs) around
the crater and also the surrounding ground heave. In addition, both
the digital images before and after DC pounding were taken and
compared. GCPs adopted here were with known coordinates and to
6.8 cm be used for creating the initial coordinates and measure the elevation
change of ground surface caused by DC pounding.
16.5 cm Following are the steps of this photogrammetry procedure adopted
for the crater measurement:
(1) Taking images of initial site condition. In each impact
location, at least three points are marked on the ground surface
25.8 cm and used as GCPs. One of the GCPs is used as the benchmark
and assigned the initial coordinate of (0,0,0). The coordinates
Figure 12. Point cloud image of the calibration box of the other GCPs should be decided based on their distances
and directions to the benchmark GCP. It is essential to make
sure that these GCPs cannot be moved during DC pounding.
They are supposed to work as permanent GCPs (Figure 14).
172
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(3) Processing images of the initial site condition. Input the measure the top diameter of the crate during traditional crater
images data of initial site condition to the software and run the measurement. The intersection of the cross was used to locate the
process. Get the point cloud data of the initial site condition. measuring location for the crater depth. The yellow ball inside the
Input the coordinate of benchmark GCP (0,0,0) and also the crater was used to keep the camera focus on the crater during images
other GCPs. Each GCP needs to be precisely clicked on at least capturing process.
three images of the point cloud. Input the adjusted GCPs then
re-optimize the point cloud (Figure 16).
Crater R1 Set 1
(4) Set elevation of the initial site condition. After the point
cloud has been re-optimized, the elevation of the ground
surface can be set. Choose the area of interest which will cover
the crater and its surrounding area and export as shapefiles
(.shp). The elevation of the selected area will be set as the
Crater R1 Set 1
initial elevation for the calculation of crater volume and ground
heave afterward.
(5) Processing after-pounding condition. Input the images
captured after the first pounding. Since the GCPs do not move, Figure 18. (a) Photo taken at the site, (b) Point cloud model (bird’s
the GCP coordinates for after-pounding are the same as the eye view), (c) Point cloud model (cross view) of the third stage of
initial site condition. Repeat the process adopted for the initial RIC crater (R1) at first set
site condition to get the point cloud of the first pounding.
(6) Crater volume measurement. Import the shapefiles (.shp) of
the initial elevation and then compute the crater volume. At the
same time, the ground heave can also be calculated (Figure 17).
Crater R1 Set 2
(a) (b)
Figure 17. (a) Selected elevation area on initial condition that
exported as shapefiles (b) Volume measurement with imported
shapefiles after pounding
close to a conical shape which is different from the real shape of the
crater formed by RIC pounding. Because, after each set of RIC
pounding, the hammer assembly of RIC was lifted up from the
crater. In the process, it also brought out some soil with it and
formed a crater with an upside down conical shape.
Figure 20 and 21 show the pictures of third stage DC craters Crater R1 Set 2
and their point cloud models (Crater D3). It can be seen from the
picture that the shape of the crater is irregular, but is close to conical
or truncated conical shape. The conical shape assumption can be Figure 19. (a) Photo taken at the site, (b) Point cloud model (bird’s
confirmed by the cross-section of the point cloud model shown in eye view), (c) Point cloud model (cross view) of the third stage of
the figures. The white cross shown in the picture is the ruler used to RIC crater (R1) at the second set
173
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
r 2h
Vc1 (4)
Crater D3 Pounding 7 3
h 2
Vc 2 ( a ab b 2 ) (5)
3
Crater D3 Pounding 7 Where:
Vc1 = volume of the crater with a conical shape
Vc2 = volume of the crater with a truncated conical shape
r = radius of crater measured on site
a = on-site measured top radius of the crater
b = effective radius of hammer footprint = 1.0 m (converted
Figure 21. (a) Photo taken at the site, (b) Point cloud model (bird’s from square to circular shape by means of equivalent area
eye view), (c) Point cloud model (cross view) of the third stage of method)
DC crater (D3) at seventh pounding h = measured depth of the crater
174
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Crater R1 Set 1
Crater R2 Set 2
Crater R3 Set 1
Photogrammetry
Traditional Field Measurement
Measurement
Crater Heave Top Vc1 Vc, p Figure 23. Comparison graphs between photogrammetry and field
Crater No. Depth
Volume Volume dia. Vc1 measurement of RIC craters
(h) Vc , p
(Vc,p) (Vh,p) (2r)
For the DC cases, each pounding stage had 20-25 poundings, but
m3 m3 m m m3 %* only the first ten poundings are shown in Table 5. The DC hammer
R1 1 1.60 0.08 0.88 3.00 2.073
used in this case had a square shape footprint (Figure 24). However,
29.6%
during the free fall process, the hammer tends to rotate a little during
2 1.25 0.02 0.70 2.40 1.056 -15.6% each pounding and forms a more or less circular crater on the
ground surface after a few impacts (Figure 24). The crater resulted
R2 1 1.57 0.01 0.84 2.55 1.430 -8.9% by this DC hammer can be somewhat in a cone shape or truncated
2 1.67 0.05 0.80 2.93 1.798 7.7%
cone shape except for the 1st impact. As the number of impact
increases, the crater depth increases. However, the crater shape
R3 1 1.40 0.02 0.83 2.38 1.231 -12.1% remains close to conical. Together with the on-site measured depth
and a top diameter of the crater, Eq. 4 and 3 were used to calculate
2 0.95 0.16 0.53 1.85 0.475 -50.0% the crater volume. Craters (D1, D2, & D3) from different pounding
* “-” means crater volume is underestimated stages were chosen for comparison here.
175
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Crater D1 Pounding 2
Crater D2 Pounding 5
Crater D3 Pounding 7
Crater D1 Pounding 1
176
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
* “-” means crater volume is underestimated measuring method can under-estimate the crater volume by 12~50%
(Crater R3 in Figure 23).
As for DC case, the crater volumes calculated from Eq. 4 and 5
are compared with those calculated by photogrammetry method. As
shown in Table 5 and Figure 26, the crater volume calculated from
Pix4Dmapper software is in better agreement with Eq. 4 for conical
shape crater with the maximum difference of 65%. The larger
difference appears in the initial impacts. As the number of impacts
increased, the crater shape became closer to conical; then the
difference can be reduced to about 20%. In comparison, the crater
volume calculated from Eq. 5 (Note: the most commonly used
formula used by the DC contractors) tends to overestimate the
Figure 26. Comparison between photogrammetry and field volume of DC crater by more than 65%. It implies that if the
measurement of DC crater volume (contd.) hammer was not in the crater during crater measurement, the
truncated conical shape assumed for the crater by Eq. 5 could be
quite different from the apparent crater shape and overestimates the
volume of crater. So if the crater volume is used to evaluate the
effectiveness of compaction as shown in Figure 6, such an
overestimation on crater volume should be taken into account.
Also shown in Figures 23 and 26, the RIC and DC induced
ground heave is not apparent in this pilot test. In other words, most
of the compaction energy had been absorbed by the ground. It
indicates that the in-situ ground and groundwater conditions
underlying the site are suitable for either DC or RIC method.
6. CONCLUSIONS
Among the parameters monitored for DC or RIC method, the
volume of the crater is the most significant one to evaluate the
effectiveness of pounding. This paper presents a photogrammetry
method which can carry out the volume measurement of the crater
and the surrounding ground heave with better accuracy. The
following conclusions are made based on the results of a field trial
test on a hydraulically filled reclaimed site.
1. Before the field test, an accuracy of 98.5% on volume
measurement of a calibration box with known volume had been
obtained using the photogrammetry method and Pix4Dmapper
software in the laboratory.
Figure 26. Comparison between photogrammetry and field 2. By using the smartphone or digital camera and the
measurement of DC crater volume photogrammetry technology, the point cloud data of the ground
surface can be established. Then, the point cloud data were
5. DISCUSSION ON MEASUREMENT RESULTS further calibrated with the ground control points placed around
the crater. The change of crater volume before and after each
It can be found from the RIC craters that for the first and second pounding can be determined using the calibrated point cloud
stages poundings, the crater volume determined from the traditional data with reasonable accuracy.
measuring method yields a larger volume for Set 1 and a smaller 3. By assuming a conical shape for the RIC crater, the crater
volume for Set 2 compared to those measured from Pix4Dmapper. volume calculated from Eq. 4 for the traditional measuring
Apart from the overestimation of top diameter by traditional method tends to underestimate the crater volume by about 20%
measuring method, one other possible reason for this phenomenon is compared to that measured with the photogrammetry method.
that the traditional measuring method used the ruler to measure the However, the conical shape may not be very representative to
crater depth (h). The contractor tended to push the ruler with force the real shape of the crater formed by RIC pounding due to the
into the bottom of the crater. So the measured crater depth (h) was disturbance caused by lifting up the hammer from the crater.
larger than that measured by photogrammetry method which only 4. DC crater volume measured by the photogrammetry method is
measured the surficial shape of the crater. Thus, the volume in good agreement with that calculated from Eq. 4 (for conical
calculated with either Eq. 4 or Eq. 5 showed a larger volume using shape crater). Although the overestimation by Eq. 4 is large in
the traditional method. After one set of RIC pounding, the crater the initial impacts, the difference is reduced to about 20% as
was backfilled with soil before performing next set of pounding. the number of impacts increased and the crater shape became
Since the ground became denser as the pounding increased, it closer to conical. In comparison, if Eq. 5 (truncated cone) is
became more difficult to push the ruler into the bottom of the crater. used to calculate the crater volume, more than 65%
As a result, the measured depth of crater was closer to the overestimation than that measured with photogrammetry
photogeommetry method, so was the crater volume approximated by method may result. However, it should be noted that the craters
Eq. 4 and 5. Therefore, the crater volume measured with traditional in this study were measured after the hammer was lifted up
method reduced and became less than that obtained from the from the crater and the side of crater might fall off to the
photogrammetry method (Figure 23). At the third stage, the above bottom and changed the apparent shape of the crater from
mentioned effect of over-estimating crater depth disappeared. The truncated conical to more or less conical (Figure 25). If the
volume (Vc1) of R3 crater became less than Vc,p for both set 1 and crater were measured with the hammer still in the crater, the
set 2. However, the assumption made on the shape of the crater may shape of the crater might more likely be in truncated conical
be the other source of volume underestimation. In other word, shape. Then the overestimation in volume by Eq. 5 may not be
assuming a conical shape for the crater and using the traditional as much as those shown in Table 5 and Figure 26.
177
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
7. ACKNOWLEDGMENTS
Li, L., Zhang, X., Chen, G. and Lytton, R. (2016). “Measuring
The Authors wish to thank the I-San construction company for
Unsaturated Soil Deformations during Triaxial Testing Using
carrying out the DC and RIC operation for the trial test of this crater
a Photogrammetry-Based Method”, Canadian Geotechnical
volume measurement program. The financial support from the
Journal, NRC Research Press, 53, pp. 476.
Ministry of Science and Technology (MOST) of Taiwan
Pix4D SA. (2017). Pix4Dmapper 4.1 User Manual, Switzerland, pp.
Government and Li-Jia Engineering Company are also greatly
11-46.
appreciated.
Simpson, L. A., Jang, S. T., Ronan, C. E. and Splitter, L. M. (2008).
“Liquefaction Potential Mitigation using Rapid Impact
8. REFERENCES Compaction”, Geotechnical Earthquake Engineering and Soil
Dynamics IV, ASCE, pp. 5.
Chow, Y. K., Yong, D. M., Yong, K. Y. and Lee, S. L. (1992). Tang, I. E. (2016). “Effect of Dynamic Replacement on the
“Dynamic Compaction Analysis”, Journal of Geotechnical Engineering Properties of Coal Ash Pond”, Master’s Thesis,
Engineering, ASCE, 118(8), pp. 1141. NTUST, Taipei, Taiwan
178
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: The study focused on the potentiality of Boehmeria Nivea as alternative material in the production of geotextile. The
researchers used Ramie's bark in the production of geotextiles and series of tests were performed at the Philippine Textile Research Institute
(PTRI) of the Department of Science and Technology (DOST) located at Taguig, Metro Manila, to compare ramie geotextiles and
commercially available coconet geotextiles. Each test specimen was tested for their nominal thickness, mass per unit area, and tensile
strength. The results were analyzed and gave presentable results. After analyzing the results, the researchers reached at acceptable findings.
All the test conducted of ramie geotextile did meet the minimum tensile strength of the commercially available coconet geotextile for 400,
700, and 900 and considerably met the polyfelt specification filtration 58 for non - woven geotextile. Therefore, ramie fiber can be used as a
raw material in the production of a high strength and sustainable geotextile.
Keywords: Geotextile, Boehmeria Nivea, Ramie, polyfelt.
1. INTRODUCTION jute, coir, and wood shavings. Ramie (Boehmeria Nivea), a plant
that is vastly found in the Philippines especially in Mindanao and
Geosynthetics have been characterized by the American Society some parts of Batan and Babuyan Islands. It has a great potential
for Testing and Materials (ASTM) committee D35 as “a planar as an alternative material for geotextile production due to the
product manufactured from polymeric material used with soil, high tensile capacity of its fiber.
rock, earth, or other geotechnical engineering related material as Ramie are subtropical bast fibers, which are obtained from
an integral part of a human-made project, structure, or system” their plants five to six times a year. The fibers have silky luster
(https://www.slideshare.net/pparida/geosynthetics). They are and have white appearance even in the unbleached condition
used for purposes of separation, reinforcement, drainage, and (http://textilelearner.blogspot.com/2012/12/ selection-of-fiber-
filtration. According to Dr. Robert M. Koerner, “there are eight for-geotextiles.html). They consist of pure cellulose fiber which
types of geosynthetics: geogrids, geotextiles, geonets, possess highest tenacity among all plant fibers”. Thus, ramie can
geomembranes, geosynthetic clay liners, geopipe, geofoam, and be considered a great source of natural fiber.
geocomposites.”(http://www.acegeosyntheticsecopark.com/geosy Textiles are primarily and traditionally used for the design
nthetics#WHAT%20IS%20GEOSYNTHETIC?). Among these and production of garments but recently natural fibers and
geosynthetic products, the commonly used product is geotextiles. textiles were used as an alternative to synthetic fiber to promote
Geotextiles have been defined by the American Society of environmental awareness and conservation for the benefit of the
Agricultural and Biological Engineers (ASAE) as, “a fabric or generations ahead of us. Knowledge with regards to the
synthetic material placed between the soil and a pipe, gabion, or utilization of natural-fiber geotextiles should be further
retaining wall to enhance water movement and retard soil encouraged and recognized.
movement, and as a blanket to add reinforcement and The study aimed to determine the potentiality of Ramie
separation."(http://www.apparelsearch.com/education/research/n (Boehmeria Nivea) as an alternative raw material in geotextile
onwoven/2001_kermit_duckett/education_research_nonwoven_g production.
eotextiles.htm). Geotextiles are commonly utilized in
geotechnical engineering to reinforce soil and to construct strong
2. METHODS AND MATERIALS
bases for roads. Geotextiles are the most popular type of
geosynthetic material due to their versatility and affordability 2.1 Preparation of Sample
compared to other types of geosynthetics.
In civil engineering applications, the geotextiles used are The natural fiber used in this study came from Boehmeria Nivea
generally polymeric material. They offer longer life span and are (Ramie) a wild plant generally abundant in Barangay Manila de
not subject to biodegradability. However, synthetic geotextiles Bugabos, Agusan del Norte, Mindanao and other parts of the
are not eco-friendly and could possibly create environmental Philippines. Initially, the Ramie’s bark were taken from the stem
problems in the long run. Also, the raw materials used in of the plant. The collected stems were then cleaned and prepared
synthetic geotextiles are not readily accessible. Thus, prices of for the succeeding processes.
raw materials needed for production do not reduce resulting into The bark were removed with by means of peeling using a
expensive synthetic geotextiles. Due to the problems in relation knife, precautions were done to avoid damages in the fibers.
to the use of synthetic geotextiles and with the awareness and Cellulose layer of the bark were then removed and cleaned in
efforts to provide sustainable development, the use of preparation for drying the fiber.
biodegradable natural geotextiles has emerged. Natural fiber- Ramie’s fiber were then sun-dried to remove excess water.
based geotextiles are widely accepted because they are eco- Oven-drying can also be performed if available to thoroughly
friendly, renewable, economically viable, abundantly available, remove excess moisture of the extracted fiber.
very cost-effective and cheaper in comparison to their synthetic Figure 1 shows the process flow of ramie geotextile
counterparts. production.
Natural fiber-based geotextiles are commonly in the form of
179
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
are the weft or filling. The strategy in which these strings are
interwoven influences the quality of the textile.
Collection of
Ramie’s bark 2.2 Equipment
Several materials and equipment were used in the preparation of
sample specimen needed for testing the properties of the natural
fiber geotextile.
The J.A. king pneumatic sample cutter SASD-692 was used
Extraction of Fiber to cut eight (8) circular test specimens that were used for the test
From Ramie’s bark for mass per unit area and nominal thickness. (Laroza, J.P et al.
2014).
Eight (8) test specimens were tested using the Zwick/Roell
tensile strength tester Z050 (CRE) using a 50 kN full safe load
with a wide-width strip method in order to determine the average
tensile strength of sample.
Drying of Ramie’s bark The SDL digital thickness gauge M034A with 25 cm2
Fiber pressure foot area subjected to 2 kPa was used to find the
nominal thickness of the eight (8) circular test specimens by the
J.A. king pneumatic sample cutter.
180
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The tensile strength of geotextiles and related materials is a 3.1.3 Tensile Strength of Ramie Geotextile (Machine
very important property as virtually all applications rely on it Direction and Cross-Machine Direction)
either as the primary or secondary function. This test is useful for
quality control and can also be used for design purposes Figure 3 presents the graphical representation of the Tensile
Table 1. Shows the number of specimens required for every Strength Machine Direction (MD) of Ramie Geotextile based on
test. the test results.
Table 1. Number of Test Specimens
Geotextile Nominal Tensile strength Number of
sample thickness test test
test/Mass per specimens
unit area MD CMD
Ramie 8 4 8
Figure 2 Nominal Thickness of Ramie Geotextile 3.2 Comparison of Properties of Ramie Geotextile and
Coconet Based on Test Results.
The above figure presents the average nominal thickness of
ramie geotextile which is about 13.39 mm (blue bar) and 3.2.1 Nominal Thickness of Ramie Geotextile and Coconet
comparing to no. of trials (8 trials) done in test specimens Geotextile.
(orange bar). Based on results, Ramie Geotextile has larger cross For verification of the strength and potentiality of Ramie
sectional area compared to the commercially available Coconet Geotextile, comparisons were made against commercially
Geotextile. available geotextile Coconet. Figure 5 presents the graphical
comparison between the average nominal thickness of ramie
3.1.2. Mass Per Unit Area geotextile and coconet geotextile.
Table 2 presents the mass per unit area of Ramie Geotextile.
Formula:
4
mx 10 (1)
M
A
Based from the data drawn, Table 3 shows the average mass
per unit area of Ramie Geotextile was 1684.73 g/m2 after 8 trials
were made.
181
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
182
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Laroza, J.P. et. Al., Viability of Banana Sheath as Natural Fiber Ministry for the Environment. (2009, August 24). Soil intactness
Geotextile, Batangas State University, 2014. of erosion-prone 5555land. Retrieved June 28, 2015,
Mahuya Ghosh, (2009). SUITABILITY OF NATURAL FIBRES from new Zealand. Govt.nz: 5555www.mfe.goct.nz/
IN GEOTEXTILE 5555APPLICATIONS. Retrieved environmental-reporting/land/erosion-risk/
June 28, 2015 , from Scientist, Geotech Cell, Parida, P. (2009). A Seminar on Geosynthetics. Viewed January
5555Indian Jute Industries’ Research Association, 2018,
Kolkata–700088, https://www.slideshare.net/pparida/geosynthetics.
India:5555http://gndec.ac.in/~igs/ldh/conf/2009/articles Selection of Fiber for Geotextiles: Fibers Used in Geotextiles,
/T07_40.pdf Viewed January 2018,
http://textilelearner.blogspot.com
/2012/12/selection-of-fiber-for-geotextiles.html
183
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: A major geotechnical challenge for large and mega infrastructure developments is to economically appraise the geotechnical
conditions of the development area early enough for concept design, with sufficient resolution for detailed design and with sufficient spatial
coverage to provide flexibility for layout changes. In active geological settings, characterisation and mitigation of potential geohazards
presents a significant additional challenge. This paper describes the use of a multidisciplinary ground model as a tool to support both
geotechnical site appraisal and geohazard appraisal. The combined use of a geographical information system with a ground model is also
described, to manage the spatially extensive data associated with large infrastructure developments and to perform geotechnical and
geohazard spatial analysis over large areas. Examples of successful application of this approach from major offshore developments are
described and the importance of a competent and experienced geoteam of discipline specialists is highlighted.
Keywords: Ground model, geographic information system, geohazard, spatial analysis
184
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
may provide short-term savings, but the flexibility afforded by a level of data manipulation and combined data and spatial query is
wider site understanding can save significant costs associated with not possible unless data are integrated within a GIS-based ground
late-stage remobilisation of data acquisition equipment if more data model.
are required. Non-spatial data such as reports, photographs, diagrams, etc. can
Newly acquired data are typically used to update the ground be made accessible via hyperlinks from a relevant geographical
model to reflect the improved understanding of the site. Therefore, location within the GIS ground model to provide single-interface
an integrated multidisciplinary approach should be applied to data access to all information.
acquisition planning, monitoring and results interpretation to
optimise the resulting ground model iteration. For example, this 2.3 Visualisation and Communication
includes the specification of combined sample acquisition for
geotechnical testing, geochronological dating and sedimentological A multidisciplinary ground model is an effective way to visualise
logging, either from within a single composite borehole or via the the three-dimensional spatial relationships between datasets and to
specification of clustered boreholes to target multiple requirements understand the link between the terrain, subsurface, geotechnical
from the same location. Where combined geophysical and conditions and geohazard processes. Communication of this key site
geotechnical data acquisition is planned, boreholes and in situ tests understanding is possible via both two-dimensional and three-
should be sited to lie on geophysical acquisition lines for optimal dimensional GIS interfaces. Typically, specific features or processes
correlation and interpolation potential. will also be illustrated using annotated conceptual block models
As well as the potential to overlap geotechnical and geohazard which make use of a graphical approach to clearly communicate
data acquisition targets, geohazard-specific data acquisition may observations and interpretations (Figure 1).
also require a focus on areas distal to the development area. For
example, a pipeline route may be planned along shallow gradient
valley floor, surrounded by steeper slopes of the valley sides.
Without proper characterisation of the valley sides, which may
include data acquisition, the geohazard risk associated with slope
instability cannot be fully assessed.
Whilst careful data acquisition planning helps avoid costly late-
stage remobilisation of data acquisition equipment, it can be
beneficial to adopt a phased approach whereby an initial limited
quantity of ‘reconnaissance’ data are acquired and used to make an
interim ground model update, from which a further gap analysis and
detailed acquisition scope are created.
A further alternative is to ensure the results of the data
acquisition are reviewed by a multidisciplinary team, or ‘geoteam’
(described later in the paper), as the acquisition progresses. This
approach avoids multiple iterations of data acquisition but provides
flexibility for changes to the data acquisition scope in almost real-
time. New data are compared with predictions from the existing Figure 1 Example of a conceptual block model which uses a
ground model; where conditions are consistently as predicted the graphical approach to clearly communicate key observations and
density of data acquisition locations can be reduced and conversely interpretations
where conditions are not as expected additional data can be
specified. Specification of additional data locations during data 3. GEOTECHNICAL SITE APPRAISAL
acquisition can be expedited by pre-defining option locations before 3.1 Soil Model
mobilisation.
A predictive soil model focused on infrastructure engineering
2.2 Geographical Information System Framework requirements is the primary interpretive output from a
multidisciplinary ground model for geotechnical engineering
A successful multidisciplinary ground model requires integration of application. The predictive soil model comprises soil units and soil
data from each discipline in a common framework. All disciplines provinces. Soil units are used to divide the three-dimensional zone
of ‘geo-data’ for infrastructure developments have a position in of interest into volumes of similar soil properties and geotechnical
geographical space and therefore a GIS platform provides an conditions, with characteristic geotechnical parameters defined for
intuitive spatial framework in which to integrate all data of each unit. Soil provinces are surface zonations which group areas
relevance to a ground model. with a similar sequence and thickness of soil units between ground
A GIS has a primarily map-based interface built from an ordered level and the depth of interest below ground level. Soil unit
set of layers which clearly demonstrate the spatial interrelations boundaries are interpolated between boreholes and, where
between datasets and the planned infrastructure. Continuous surface geophysical data are available and if acoustically significant, can be
data are represented as grids and triangular irregular networks mapped as continuous surfaces across a development area (Figure 2).
(TINs), and discrete features as points, lines and polygons (typically This simple but versatile approach allows the geotechnical
known as vector layers). All vector layers have associated attributes conditions to be predicted anywhere in a development area with as
which are used to capture tabular detail about each feature and much accuracy as the available data will allow and with
vector layers are typically symbolised to highlight a specific geologically-informed interpolation where data are absent.
attribute.
Most GIS software has a powerful set of in-built tools and can 3.2 Geotechnical Application
typically be customised to develop additional tools. In addition to
The predictive soil model can be used directly to provide soil
basic navigation tools, some of the most useful tools for GIS-based
parameters for foundation design associated with a fixed
ground model manipulation allow users to make composite spatial
infrastructure layout and, if sufficiently detailed, will include lower
and attribute queries. A well-structured GIS ground model can be
bound parameters for capacity calculations, upper bound parameters
used to make queries such as “select all locations within 2000m of a
for installation calculations, etc. However, where the infrastructure
planned structure where the soil friction angle is greater than 30° at
layout is not defined, spatial analysis informed by the GIS-based
less than 10 m below ground level” or “select all samples on slopes
ground model can help refine the layout for optimal geotechnical
greater than 10° with a unit weight greater than 21kN/m3”. This
placement.
185
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
4. GEOHAZARD APPRAISAL
4.1 Geohazard Model
A geohazard model is the primary interpretive output from a
multidisciplinary ground model for geohazard appraisal and
assessment. The format of the geohazard model will vary according
to the application, but typically includes a geohazard database and
register.
The geohazard database records the number of observations of
each geohazard occurrence in the ground model datasets, and
records key metrics associated with each event. For example, a
database of past offshore landslides would be built from
observations on hillshaded seafloor elevation data and sub-bottom
profiler or seismic data, where available, to characterise the number
of events visible, the depth below ground level (and hence inferred
age) of each event, and key metrics such as the length, width,
thickness, orientation, etc. Where possible, regional
seismostratigraphy and geochronological dating will be used to
Figure 2 Example soil model for a wind farm development refine estimates of the age of events, with the aim of determining a
comprising soil units visible on the cross section and soil provinces rate of reoccurrence.
visible on the plan view map (Clare et al, 2012) The geohazard register documents all geohazards considered to
be credible at the site, summarises the properties of each geohazard
Rushton et al (2017) present examples of geospatial analysis, based on observations and metrics from the geohazard database,
which includes a method developed to derive GIS-based lateral pile captures initial estimates of the likelihood of each geohazard and, if
capacity and map the required monopile length sufficient to provide possible, makes some inference of the consequence of the each
capacity for an offshore wind turbine across an entire licence zone. geohazard impacting the planned infrastructure. This geohazard
The method involved development of a regular mesh of analysis register is an important step in documenting which geohazards
points and use of the GIS-based ground model to assign design soil require further assessment in a full quantitative risk assessment
parameters and lateral load-displacement (p-y) soil springs to each (QRA) and which geohazards are considered of sufficiently low
mesh node, for use in a one-dimensional finite element analysis of concern to the planned infrastructure to not require further
the monopile (Figure 3). An iterative procedure was performed to consideration.
determine the required monopile lengths: if the monopile was able
to withstand the applied lateral load and moment, and remain within 4.2 Quantitative Risk Assessment
the prescribed displacement and rotation tolerances, the pile length A QRA takes the largely qualitative geohazard register and
was reduced. As soon as either of the tolerances were exceeded, the advances the assessment of the main threats by further quantified
previous monopile length was taken as the required length. analysis, which should include consideration of all components
described in Table 1.
Table 1 Quantitative Risk Assessment Considerations
Component Description
P(event) Probability of a geohazard event
P(spatial) Spatial probability: infrastructure exposure to
a geohazard event and likelihood of it being
hit given an event occurs
P(hit) P(event) × P(spatial)
P(damage|hit) Probability of infrastructure damage in the
event of it being hit by a geohazard event.
This is assessed via vulnerability analyses
P(damage) P(hit) × P(damage|hit)
Risk P(damage) × Consequence, where
consequence may be in terms of health and
Figure 3 Graphical depiction of ground model sampling via a safety, financial, reputational or
regular mesh in GIS to provide inputs to an automated sequence of environmental and is typically assessed by the
monopile lateral capacity analyses infrastructure developer
In another offshore example for a major oil and gas company, a Probability of a geohazard event may be inferred from rates of
very shallow GIS ground model was built to integrate cone reoccurrence as described in the Geohazard Model section. In
penetrometer test (CPT) data with the reflected amplitude of the addition to reliance on such historical frequency estimates, it is
acoustic signal from a geophysical source. The aim was to locate typical to perform forward modelling to estimate the likely future
186
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
7. REFERENCES
Clare, M., Rushton, D. and Balthes, R. (2012) “A ground model-
based approach to efficient assessment and management of
risk for pile installation and behavior”. Proceedings of the
Hans Lorenz Symposium, Hamburg, Germany.
Dimmock, P., Mackenzie, B. and Mills, A. (2012) “Probabilistic
Slope Stability Analysis in the West Nile Delta, Offshore
Egypt”. Proceedings of the 7th International Offshore Site
Investigation and Geotechnics Conference: Integrated
Technologies – Present and Future, London, Volume 1, pp
535-542
Evans, T., G., (2010) “A Systematic Approach to Offshore
Engineering for Multiple-Project Developments in
Geohazardous Areas”. Keynote Paper, Proceedings of the 2nd
International Symposium of Frontiers in Offshore
Geotechnics, Perth.
Fawcett, T. (2006) “An introduction to ROC analysis”. Pattern
Figure 4 Graphical depiction of GIS-based slope stability Recognition Letters 27, pp 861-874.
assessment which uses geotechnical properties extracted from the Fookes, P., G. (1997) “Geology for engineers: the geological model,
ground model as pixel surfaces and generates a further pixel surface prediction and performance”. Quarterly Journal of
of factor of safety Engineering Geology and Hydrogeology, 30, pp 293-424.
Spatial probability is typically a function of the size of the
Hill, A., Prakash, A., Hampson, K., Gray, T., Rushton, D.,
geohazard event relative to the size of the infrastructure and
Mackenzie, B. and Puzrin, A. (2015) “Geotechnical
therefore is usually calculated within the same GIS framework as
Challenges in the Caspian Sea”. Proceedings of the Offshore
the ground model. The other QRA components are largely assessed
Technology Conference, OTC-25988-MS.
outside of the ground model in a risk workbook, but the interim
Knill, J., L., (2003) “Core values: the First Hans Cloos Lecture”.
result of P(hit) and the final results of P(damage) and Risk are most
Bulletin of Engineering Geology and the Environment, 62,
effectively communicated as mapped layers hosted in the GIS
pp 1-34.
ground model and visualised relative to the planned infrastructure
Mackenzie, B., Hooper, J. and Rushton, D. (2009) “Spatial analysis
layout (e.g. Hill et al, 2015),
of shallow slope instability incorporating an engineering
5. GEOTEAM geological ground model”. Proceedings of the 4th
International Symposium of Submarine Mass Movements and
The approach described for the development of a multidisciplinary Their Consequences, Austin, Texas.
ground model for application in geotechnical and geohazard site Rushton, D, Gray, T., Puzrin, A. and Hill, A. (2015) “GIS-Based
appraisal relies heavily on large and diverse datasets, powerful GIS Probabilistic Slope Stability Assessment Using Shear Band
software applications and cutting-edge analysis. However, the most Propagation”. Proceedings of the Offshore Technology
important component in the development of a multidisciplinary Conference, OTC-25871-MS.
ground model is a competent and experienced geoteam of discipline Rushton, D., Jones, L., Nott, T. and Rattley, M. (2017) “Geospatial
specialists able to work collaboratively and combine their expertise. Analysis for Geotechnical Design”. Proceedings of the 8th
The specific skillsets within a geoteam will depend on the International Offshore Site Investigation and Geotechnics
requirements of each individual project, but may comprise Conference: Smarter Solutions for Future Offshore
geologists, geomorphologists, geotechnical engineers, geophysicists, Developments, London, Volume 1, pp 231-238
risk analysts, sedimentologists and facility engineers.
187
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Prediction of the Geological Condition for Pipe jacking Based on the Data Collected
in the Shafts
Keh-Jian Shou1, Fu -Kang Tang1
1
Department of Civil Engineering, National Chung-Hsing University, Taiwan
E-mail: kjshou@dragon.nchu.edu.tw
ABSTRACT: No -Dig method often encounters difficult geological problems, which makes the excavation in difficulty and thus affecting
the construction schedule. Central Taiwan area is majorly covered by gravel alluvium, in which the strata contains large gravels. Those large
size gravels often result in excessive wear and increase in construction cost. In this paper, we took samples from the launch shaft and arrival
shaft pipe-jacking and carried out sieve analyses. Then we transformed the size distribution into borehole data by the newly derived formula.
The obtained borehole data was applied to the T-PROGS geostatistic analysis. The purpose of the geostatistic analysis is to building a 3D
geological model from data from pipe-jacking shafts, then we can obtain the geological and size distribution along the pipe-jacking
alignment. In this study, the analysis was performed for different scenerios and the impact of material changes on the pipe jacking rate. The
results show that the change of gravel particles larger than 20 cm is highly correlated with pipe jacking rate. The results suggest that it is
feasible to apply the image sieve analysis in the geostatistcs analysis.
Keywords: pipe jacking, geostatistics, geological condition, T-PROGS, Markov chain model.
188
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
x and y axes into four sectors of equal area at the xy plane of the T-PROGS is composed of many sub-programs, including
projected area (z = 0). The equation for the sector area is: GAMEAS for calculation in bivariate statistics (transition
probability and indicator variogram), MCMOD for development of
R 2 y 2 Markov chain model, TSIM for three-dimensional and cross-
Ac (1) correlation conditional simulation, etc. The theoretical background
4 of T-PROG is briefly described as below.
Transition probability tjk (h) is defined (Carle and Fogg, 1996)
as follows:
The sector area Ac is integrated along the z-axis and divided by 4. RESULTS AND DISCUSSION
1/8 the volume of the sphere to obtain the distance from the center
of sphere to the three axes; which, after some calculation 4.1 Analysis of the Actual Cases
adjustments, we obtain y =3R/8, as shown in Fig. 1(b). Applying The analysis of the pipe jacking project with a configuration plane
the geometry, we obtain R' (distance EF from the starting point to diagram, as shown in Fig. 2. The diameter cross-section corresponds
the end point). Applying the Pythagorean Theorem, we obtain: to an actual drill machine head of 0.6 m as actual case 1
(extrapolated) and actual case 2 (interpolated). The sieve analysis
R 2 EF y
2 2
(2) data of shaft Bi43 and shaft Bi46 were used to establish the 3D
geological model and estimate the probability of jacking through
large-size gravel particles. Then these probabilities of encountering
Substituting the known
y =3R/8 into Eq. (2) and obtaining: the large-size particles (greater than 20 cm) were used to explore
their correlation with the advancing rates of actual pipe jacking
work.
55
EF R ' R 2 (3R ) R (3)
8 8
189
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The pipe jacking project records for the sections Bi42 to Bi43 and
Bi43 to Bi46, including daily work record, the history of pipe
jacking, and the number of jacked pipe count, were used to obtain
the advancing rate, as shown in Fig. 5~6. The data were also
converted into advancing rates in hours per pipe jacked. Such that
we can analyse the relationship between the number of hours spent
and the change in the geological materials can be obtained.
However, the figures show that the advancing rate increases as the
end of the work nears.
(a)
(a)
(b)
(b)
Figure 5 Progress and advancing rate of pipe jacking section Bi43 to
Bi46
(c)
(d)
Figure 4 The gradations of the four data blocks, (a) Bi43 -5.5 m, (b)
Bi46 -6.5 m, (c) Bi43 -9 m, and (d) Bi46 -8.5 m (a)
190
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(b)
Figure 6 Progress and advancing rate of pipe jacking section Bi42 to
Bi43 Figure 8 The results of three-dimensional geological modelling for
case II, section from Bi42 to Bi43.
The three-dimensional geological model for case I, i.e., section
from Bi43 to shaft Bi46, was illustrated in Fig. 7, with a regression 5. CONCLUSIONS
trend line of particle size change. Comparison of the results found
that the tendency of large size particle (greater than 20 cm) to
gradually decline (representing a decreasing number of grids for The alluvium in Taichung area is mainly gravel deposits, containing
particle size greater than 20 cm) is very consistent with the tendency a substantial amount of large sized gravel, which has a significant
of the advancing rate to increase. In contrast, grids for particle size impact on the success of any no-dig project. This study conducted
between 2 cm and 20 cm tend to rise at the same position an on-site sieve analysis of geomaterial launch and/or arrival shafts.
(representing an increase in finer materials). This finding is quite After obtaining sieve analysis data, a derived formula is applied to
consistent with fact that the reduction of large pieces of gravel in the adjust and convert the data into drilling core log. Then a GMS sub-
strata, leading to an increase in advancing speed. program T-PROGS is used to perform three-dimensional geological
Similarly, the material change for the section from Bi42 to Bi43 modelling to analyse the vertical and horizontal material changes.
in the three-dimensional geological model (Fig. 8) added to the The analysis results were compared with pipe jacking data of the
trend line obtained Fig. 5-6. Comparison of results showed that the study cases.
position of grids greater than 20 cm tended to increase gradually Comparison of the simulation results with advancing rates of
(represent an increasing number of grids greater than 20 cm), which actual cases showed that the trends in material change and
is very consistent with the tendency of the position of the drilling advancing rate have a high degree of correlation. When the number
rate to increase (representing a decrease in drill rate speed). of grids of size greater than 20 cm decreased, and those grids of less
However, the trend of grids for size between 2 cm and 20 cm is to than 20 cm but greater than 2 cm increased, the advancing rate also
decline at the same place (representing a decrease in finer increased. Conversely, when the number of grids of size greater than
materials), which is quite consistent with the increase in large size 20 cm increased, and those grids of less than 20 cm but greater than
gravel in the strata, leading to a decrease in advancing speed. 2 cm decreased, the advancing rate also decreased. They highly
correspond to the actual pipe jacking situation: when the cutter head
of a shield machine encounters large sized gravel rock, its speed will
slow down due to cutting wear.
From the diameter point of view, the larger the diameter cross
section, the greater the number of grids encountered of size greater
than 20 cm; the smaller the diameter cross section, the smaller the
number of grids encountered of size greater than 20 cm. Since the
parameters for the shafts of the two study cases are the same, the
trends in material change in the longitudinal section from the two
cases are consistent. The results show that the material change in
grids of size less than 2 cm in each case is scattered. The reason for
the inaccuracy might be due to the size of the analysis grid, i.e., the
unit grid size of 0.2 m by 0.2 m by 0.2 m far exceeds the particle
size of less than 2 cm.
191
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Li, C.Q., Shao, J.L., Jin, P., Cui, Y.L., 2009. Conditional simulation
of hydro-geological structure and its application in plain
6. REFERENCES
area: A case study of North China Plain. Geoscience,
Vol.23, No.1, pp.137-143.
Carle, S.F., Fogg, G.E., 1996. Transition probability-based indicator
geostatistics. Mathematical Geology, Vol. 28, Seeboonruang, U., 2006. An application of transition probability
No.4,pp.453~476. approach to geostatistical simulation: A case study in the
Lower Chao Phraya Basin, Thailand. Proceedings of the
Carle, S.F., Fogg, G.E., 1997. Modeling spatial variability with one 2nd IMT-GT Regional Conference on Mathematics,
and multidimensional continuous-lag Markov chains. Statistics and Applications Universiti Sains Malaysia,
Mathematical Geology,Vol.29,No.7, pp.891~917. Penang.
Carle, S.F., 1999. T-PROGS:Transition Probability Geostatistical Theodossiou , N., Latinopoulos , D., 2009. Economic aspects of the
Software, Version 2.1, delineation of well head protection areas under conditions
of uncertainty. Proceedings of the 2nd International
Hydrologic Sciences Graduate Group, University of California, CEMEPE & SECOTOX Conference, Mykonos.,
Davis. pp.309~314.
Jones, N.L., Walker, J.R., Carle, S.F., 2002. Using transition Weissmann, G.S., 2003. Application of transition probability
probability geostatistics with geostatistics in a detailed stratigraphic framework,
MODFLOW. Proceedings of Model CARE 2002, Prague, Czech University of New Mexico, USA.
Republic, pp.359~364.
192
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This study examined monitoring and numerical analysis results of a dip slope case in New Taipei City, Taiwan. The
displacement data collected over the years by the inclinometers are used to study the creep of the slope, and combined with the influence of
the typhoon rainfall. The time lag and groundwater level changes and the effectiveness of the implementation of two catchpits in the slope
are reviewed. The Shape Acceleration Array (SAA) measurement is taken as the reference basis for hazard prevention and maintenance of
the slope. Geostudio is applied to analyze the relationship between rainfall infiltration, groundwater level, and the displacement of the entire
slope.
Keywords: Groundwater variation, inclinometer, SAA, numerical analysis.
The main exposed stratum under the site is the Mushan
1. INTRODUCTION Formation, with the bedrock being mainly interbedded with
sandstone and shale. It is a dip slope striking toward the east,
This study deal with the data obtained from monitoring system of dipping southward about 10°–20°. Huang and Jeng (2004) pointed
dip slope area at the Huafan University in northeastern Taiwan as out that through the observation and comparison of drilling cores
shown in Figure 1. and ground resistance images, as well as topographical
characteristics, there are two local small faults: a south-trending
Nanshihkeng Fault and a northwest-trending A Fault, as shown in
Figure 3.
193
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
comparison of different rainfalls shows that the peak rainfall of post- Figure 5 shows that the greater the peak rainfall, the shorter the
peak rainfall was greater than the peak of other types of rainfall. reaction time lag is for the groundwater level to rise. With respect to
During the typhoon season, rainfall will cause greater rise of the the rainfall type perspective, the peak rainfall of the pre-peak
groundwater level, and the lag time between the peak rainfall and rainfalls is concentrated less than 20 mm and has a long
peak groundwater level will significantly shorten. When the groundwater-level time lag. The peak rainfall of the middle-peak
groundwater level in W1 rises to about ground level (GL) -30 m, rainfalls ranges from 20 mm to 40 mm, and the peak rainfall of the
W2 and W3 will then increase more noticeably. post-peak rainfalls is concentrated at around 50 mm with a short
time lag. It indicates that if this study area has rainfall greater than
50 mm, it should be closely observed and the observation time
should be stretched to about 40 hours. Further, Figure 6 shows that
for greater peak rainfall, the groundwater level will significantly
increase, and the post-peak rainfall will cause the most significant
groundwater level rise. According to the auto-recording
groundwater-level gauge, the groundwater rise amount was between
1 and 23 m. Lastly, Figure 7 shows that when the total average
cumulative rainfall to peak point exceeds 200 mm, the groundwater
level in auto-recording groundwater-level gauge in W1 will increase
up to 20–23 m.
As mentioned above, when the peak rainfall is greater, the time
lag of the groundwater level rise will be shorter. The peak rainfall of
the post-peak rainfall type is greater than that of other rainfall types,
which makes the change of its groundwater level significant. Thus,
observation for up to 40 hours is needed for the post-peak rainfall.
The data from the auto-recording groundwater-level gauge shows
that the groundwater level W1 is most affected by rainfall, with a
change between 1 m and 23 m. A groundwater-level change of 0.6–
5.8 m was found in the auto-recording groundwater-level gauge in
Figure 3 Site geological map W2, 0.25–3.1 m in W3, and 0.1–2 m in W4.
0
0 100 200 300 400 500 600 700 800
60 160
level respond in W1 (hr)
-5 w1
140
-10 50 w2
120
-15 w3
80 Forward type
Hourly rainfall (mm/hr)
-25
20
groundwater-level gauge region, and (5) seasonal factors. Analyses
of these factors are illustrated in Figs. 5–7, which contain only
analysis data from observation well W1. These figures are, in this 15
post-peak type
order, the groundwater level and rainfall peak time lag diagram, the pre-peak type
peak rainfall and groundwater-level rise diagram, and the 10
medium-peak type
accumulated rainfall and groundwater-level rise diagram for W1 linear regression
groundwater-level gauge from 2013 to 2015. Figure 5 is the 5
groundwater level and rainfall peak time lag diagram of automatic
groundwater-level gauge W1. The x-axis is the peak rainfall, and the 0
y-axis is the time lag for groundwater level to reach its peak. Figure 0 20 40 60 80
6 is the peak rainfall and groundwater level rise diagram of Peak rainfall (mm)
automatic groundwater-level gauge W1. The x-axis is the peak Figure 6 Peak rainfall and groundwater level rise diagram of
rainfall, and the y-axis is the amount of groundwater level rise. automatic groundwater-level gauge W1.
Figure 7 is the accumulated rainfall and groundwater-level rise
diagram of automatic groundwater-level gauge W1. The x-axis is
the accumulated rainfall, and the y-axis is the groundwater-level rise
amount. The round dots are post-peak rainfalls, the square dots are
pre-peak rainfalls, and the triangle dots are middle-peak rainfalls.
There are two types of accumulated rainfall: accumulated up to the
peak (dark-colored symbols) and accumulated for the entire rainfall
(light-colored symbols).
194
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
10
Linear regression of
segment 1
5
Linear regression of
0 segment 2
0 100 200 300 400 500 600 700 800
-5
Figure 7 Accumulated rainfall and groundwater level rise diagram
of automatic groundwater-level gauge W1.
SIS-40
25
20 Segment 1
y = 0.0341x - 8.6631
15 R² = 0.7358 Segment 2 Figure 10 Comparison of groundwater level before and after
catchpits implementation.
10
Linear regression of
segment 1 3.4 Discussion of SAA displacement measurement results
5
Linear regression of This study sought to understand slope displacement caused by
0 segment 2 typhoon rainfall, so the collection of rainfall data and SAA
0 100 200 300 400 500 600 700 800 displacement data continued until February 2017 to observe and
-5 analyze the relationship between typhoon rainfall and displacement
Figure 8 SIS-40 displacement and cumulative rainfall amount from of SAA. In this paper, only part of the data was taken and enlarged,
typhoon rainfall. shown in Figs. 11–12. The x-axis is the date, the left y-axis is the
rainfall amount, and the right y-axis shows the accumulated
3.3 Comparison of Groundwater Level Change to Explore displacement of inclinometer SIS-11 and inclinometer SIS-20.
the Effectiveness of Catchpits As shown Figures 11 to 12, there were two typhoons in August
2015, the medium typhoon Soudelor and the strong typhoon Goni.
Figure 9 is the relation graph of rainfall data and groundwater level Accordingly, SIS-11 had displacement of 6.048 mm and 1.489 mm,
from the auto-recording groundwater-level gauge from April to July respectively, and the displacement of SIS-20 was 4.839 mm and
2016. The x-axis is time, the left y-axis is the water level, and the 1.64 mm, respectively. In October 2016, there was light typhoon
right y-axis is the rainfall amount of the day. The groundwater level Aere along with its peripheral circulation. SIS-11 had a consequent
from the auto-recording groundwater-level gauge in W1 has a red displacement of 11.321 mm, while SIS-20 had a displacement of
curve, the water level in W2 has a green curve, the water level in 3.384 mm. The above data are plotted in Figure 13 as an SAA
W3 has a purple curve, and the groundwater level in W4 has a light displacement increment graph during typhoon rainfall.
blue curve. The dark-blue curve shows the daily rainfall amount Figure 13 shows that when the typhoon cumulative rainfall
reaches 184 mm, significant displacement will occur; when the
from April to July. The figure shows that, similar to aforementioned
cumulative rainfall reaches 300 mm or more, or when the peak
data, W1 is much more affected by the rainfall amount than the
rainfall exceeds 156 mm, the displacement will begin to accelerate,
other three auto-recording groundwater-level gauges. The data from
producing displacement larger than 3 mm. This result agrees with
the auto-recording groundwater-level gauge in W1 from 2013 to that in Section 3.2 of this study. Therefore, the study sets typhoon
2016, as shown in Figure 10, shows that the normal groundwater rainfall of more than 300 mm as the alert threshold. In other words,
level from the auto-recording groundwater-level gauge in W1 has a
lowest value of about −43.67 m from 2013 to 2015. After the
195
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
200 165
120 155
100
(mm)
80 150 Rainfall
60 SIS11A_12.5m-X
40 145 SIS20_15.0-X
20
0 140 Figure 14 Cross-section profile of SEEP/W.
2015/8/15
2015/8/23
2015/8/11
2015/8/13
2015/8/17
2015/8/19
2015/8/21
2015/8/25
2015/8/27
2015/8/29
2015/8/1
2015/8/3
2015/8/5
2015/8/7
2015/8/9
Typhoon Aere
80 178 Unit weight
Rainfall 19.31 25.52 23.86 23.3
60 176 SIS11A_12.5m-X
(kN/ m2)
40 174 SIS20_15.0-X Cohesion, C
20 172 18.5 41.8 38.7 0
0 170
(kN/ m2)
Internal friction
2016/10/1
2016/10/13
2016/10/25
2016/10/3
2016/10/5
2016/10/7
2016/10/9
2016/10/11
2016/10/15
2016/10/17
2016/10/19
2016/10/21
2016/10/23
2016/10/27
2016/10/29
Figure 12 SIS-11 and SIS-20 displacement versus rainfall amount in This study used rainfall data during typhoons in August 2015
October 2016. and October 2016, the amount of groundwater-level variations from
four auto-recording water-level gauges, as well as two auto-
14 recording inclinometer SAAs, to observe and analyze the
relationship between rainfall amount during typhoon, slope
SAA Displacement increatement (mm)
196
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
6. REFERENCES
Huang, C. S. and Jeng, C. J.: A supplementary geological survey
Figure 16 Profile of potential sliding surface analysis and analysis of the Talun area around the Huafan University.
Journal of Art and Design of Huafan University, (1), 59–69,
2004.
Japan Landslide Association for Slope Disaster Management
Factor of Safety (JASDiM): Essential implementation of technical design for
2 landslide countermeasures, (2), 22, 1978 (in Japanese).
1.8
Typhoon Aere Jeng C. J., and Lin, T. A.: A case study on the in-situ matrix suction
1.6 monitoring and undisturbed-sample laboratory test for the
Factor of safety
197
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Numerical simulations of a top-down constructed excavation in Central Jakarta was carried out in this study. The excavation
supported by the diaphragm wall has been constructed in a medium-stiff Central Jakarta clay. Measurements on wall deflections were made
during the construction. The numerical simulations were conducted by using a hypoplasticity model (HC model) for clays which is capable
of modeling small strain non-linearity, soil anisotropy, and recent stress-history effects. Properties of the medium-stiff Jakarta clay were
collected and studied carefully to calibrate the HC model. Numerical simulations with consideration of pre-excavation recent stress history
effect is thus suggested herein and parameters of HC model used for excavations in Central Jakarta are thus recommended also.
1. INTRODUCTION of a very thick (more than 100 m) greenish silt with consistency
from very stiff to hard. Laboratory test results of Jakarta clay such
Jakarta is a rapidly developing megacity that demands to maximize as index properties, oedometer tests, consolidated undrained tests
the use of underground space. More and more large deep and seismic down-hole test had been taken. Field measurement of
excavations are under construction and their deformations will affect wall deformation is also collected. Soil properties for Central Jakarta
adjacent infrastructures such as buildings, road, bridges, and public clay at the site are shown in Figure 1.
utilities. To properly estimate the interactions among the affected The depth of the excavation is 24.85 m, with 1.2 m thick and
soil, excavation support system and adjacent structures in these 33.7 m deep of diaphragm wall. The groundwater level was at a
projects, advanced soil constitutive models must be used to depth of 2.8 m below the ground surface (GL-2.8 m). A road
represent soil behavior during design. Features of soil behavior that embankment with total length 66 m is located at one side of
affect the calculation of ground movements including soil embankment. Figure 2 shows the excavation profile.
nonlinearity at very small strains, stiffness anisotropy, soil
compressibility and recent stress history effects should be well
simulated in the soil model.
A non-symmetric excavation in Central Jakarta was used as a
case history. A 3.1 m height of embankment with width about 66 m
was located at one side of the excavation. The construction of the
embankment altered the stress-history of soils subjected to
excavation activities later. The so-called recent stress history effect
is defined as the pre-excavation events that generated the current
state of effective stresses of the clays. The effect of recent stress
history of soils have been studied mostly at the laboratory scale for
London clay (Atkinson et al. 1990), Bothkennar clay (Smith et al.
1992), Boston blue clay (Santagata et al. 2005), and Chicago clays
(Cho 2007, Finno and Kim 2012). Significant differences on soil
stiffness in the small strain range were found as a result of those
experimental programs, and it was concluded that the observed soil
responses depends on the recent stress history of the material prior
to shearing.
Figure 1. Soil properties
An advanced soil constitutive model, hypoplasticity model for
clays (HC model) developed by Masin (2005, 2013 and 2014), is
adopted in this study. HC model is capable of capturing the features 3. SOIL PARAMETER
that represent the small strain stiffness of soils, stiffness anisotropy The basic hypoplasticity model have 5 parameters: ’, *, *, N and
and recent stress-history effects. (Arboleda et al 2017, Teng et al. pp. These parameters are similar to parameters used in Modified
2018). Numerical simulations of the hypothetical excavation made Cam Clay model. The model is not based on a conventional
through the Central Jakarta clays are presented to illustrate the effect elastoplastic framework decomposing strains in elastic and plastic
of pre-excavation recent stress history effects on the computed components, and consequently does not involve complex yield-
ground movements. The performance of HC model, with well- surface and plastic-potential definitions. It captures the nonlinear
calibrated parameters, on the magnitude and distribution of the behavior of soils at large strains in the framework of critical-state
computed excavation-induced ground movements is presented soil mechanics by means of the Matsuoka-Nakai shape of the
herein. critical-state locus in stress space. HC model used herein is
enhanced by intergranular strain concept which can represent the
2. PROJECT BACKGROUND soil behavior at small strain. Since this model is using critical soil
Soil in Jakarta is generally composed by quaternary and tertiary mechanic concept, the friction angle is determined by stress at
deposit (Firmansyah & Sukamta 2000). Quartenary deposit formed critical state. There are 9 additional soil parameters for intergranular
by volcanic ash which divided into 3 layers: 3-5 m thick of upper strain parameter: G, E, , Ag, ng, mrat, R, r, and . Table 1
lahar; alternate silty clay, silty sand and sandy silt; and about 5 m summarized the parameters of the hypoplasticity model for clays.
thick of lower lahar that consists of cemented silty sand. Tertiary
deposit is located 35 m below the ground surface. This layer consists
198
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Table 1. Parameters for HC model Table 2. parameter of Central Jakarta Clay for HC model
No. Symbol Parameter name Upper Lower
1 N Position of normal compression line No. Symbol clay clay Source
2 * Slope of normal compression line Value
3 * Slope of unloading line 1 N 1.52 1.285 Oedometer tests
4 ’ Critical-state friction angle 2 * 0.126 0.098 Oedometer tests
5 pp Stiffness fitting parameter 3 * 0.018 0.011 Oedometer tests
Anisotropy coefficient of shear, 4 ’ 38 39 CU tests
6 GE Optimization
Young and Poisson's moduli 5 t 15 5
7 Ag Stress dependency of Gvh on mean from CU test
8 ng normal stress 6 pp 0.1 0.1 (Masin 2014)
Very small strain behaviour upon 7 15.46 16.67 Test data
9 mrat
strain path reversals 8 e0 1.81 1.62 Test data
10 R Size of elastic range 9 G 1 1 Assumed value
Material constant controlling the rate 10 Ag 15500 15500 Downhole tests
11 r of evolution of the intergranular strain 11 ng 0.46 0.46 Downhole tests
tensor (Arboleda et al
Material constant controlling the rate 12 mrat 1 1
12 2017)
of degradation of the shear stiffness
13 5E-05 5E-05 (Masin 2014)
13 OCR Over consolidated Ratio
14 r 0.08 0.08 (Masin 2014)
Since available data for Central Jakarta Clay at small strain level is 15 0.9 0.9 (Masin 2014)
limited, some parameters were evaluated from several references (as listed Optimization
16 OCR 6.25 3.37
in Table 3). G is set to be 1 to ignore the anisotropy behavior. The from CU test
parameter mrat is an intergranular strain parameter that controls the very
small strain behavior of clays upon strain path reversals. A value of 1 is
chosen for this parameter as a material independent constant and the very 3.1 Parameter *,*, and N
small shear stiffness was controlled with the parameters Ag , ng and the
remaining three intergranular strain parameters (Arboleda et al 2017). R as Thirteen oedometer test results around excavation site were
size of elastic range was taken equal to 5E-5 (Masin 2014). In fact, collected to determine *,*, and OCR. *,*, Nare obtained
Central Jakarta Clay could be further categorized into 2 layers: upper clay from ln(1+e) vs ln p space from oedometer test results. * and , *
and lower clay. Some parameters calibrated in this simulation mainly are slope of normal compression line and unloading line
focus on lower clay layer because this layer dominates to wall
respectively as shown in Figure 3.
deformation. Table 2 summarized parameters used in this study.
199
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(c) (d)
Figure 5. Oedometer test results
200
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
201
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The stiffness of structural parameters is reduced by 20%, (a)1st excavation stage (b) 2nd excavation stage
considering that the stiffness of the concrete retaining wall
reduces
when subjected to large bending moment of diaphragm wall
causes the occurrence of the crack in the concrete. The axial
stiffness of the concrete floor slabs is also reduced by 20%. Table
4 contains material properties of structures used in numerical
simulation and t, E and means thickness, elastic modulus and
Poisson ratio.
202
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
203
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This paper presents a study on 3D finite element analysis of a large-scale deep excavation in Central Jakarta. The soil, Central
Jakarta clay, is modeled by both Hardening Soil and Mohr-Coulomb model in Plaxis 3D. The total and effective stress analyses under undrained
conditions were used to model the short-term undrained conditions of the clay during excavation. The input soil properties were established
using data based on an extensive site investigation program including in situ and laboratory tests, and empirical relationships with standard
penetration numbers (SPT-N). The top-down construction method and a supporting system of concrete slabs are also considered in the
numerical model. The numerical results are compared with measured wall deformation from inclinometers at a certain distance from the corner
of excavation. The simulations using drained and undrained soil shear strength (Undrained A and B functions in Plaxis 3D) are also compared
and discussed.
Keywords: Jakarta, deep excavation, 3D finite element, total and effective stress analysis, wall deformation.
204
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Table 1 Stage construction of the excavation Wall deformation (mm) Wall Deformation (mm)
-20 0 20 40 -20 0 20 40
Elapsed 0 0
Stages Construction sequences 1st exc
days 1st exc
1 Diaphragm wall installation 101 5 2nd exc 5
2 1st excavations to the depth of GL. −1.5m 18 2nd exc
3 Deck slab installation at ground level (slab 20 10 3rd exc
10
thickness, t=0.4 m) 3rd exc
Depth (m)
Depth (m)
4 2nd excavation to the depth of GL. −5.08m 30
5 Top slab construction at GL. −4.18m (slab 14 15 7.8 mm
15
18 mm
thickness, t=0.8 m)
6 22
th
5 exc 20
3rd excavation to the depth of GL. −11.18m 20
7 Middle slab construction at GL. −10.48m 20 5th exc
(slab thickness, t=0.4 m) 25 25
8 4th excavation to the depth of GL. −17.53m 28
9 28 Toe movement = 10.2 m
5th excavation to the depth of GL. −18.93m 30 30
10 Bottom slab construction at GL. −18.13m 21 (a) (b)
(slab thickness, t=1.0 m)
Figure 3 Monitoring data of wall deformation DW-15 (a) before
2.2 In situ observation and instrumentation correction; (b) after correction
Several monitoring instruments were installed around the excavation
site, including inclinometers, surface settlement points, observation Wall deformation (mm) Wall deformation (mm)
wells, rebar stress transducers, and kingpost strain gauges to monitor
-20 0 20 40 -20 0 20 40
wall displacement during construction. The observation well 0 0
summarized the groundwater level generally located at depth 2 m
beneath the ground inside the excavation zone. Then, all monitoring st
1 exc 1st exc
data were carefully monitored to obtain the reliable and representative 5 5
2nd exc
data were chosen for this study. However, only the representative
2nd exc 3rd exc
measurement from inclinometer would be presented. The
10 3rd exc 10
inclinometers were inserted on the left and right sides of the
Depth (m)
Depth (m)
205
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
close or higher than the plastic limit, with a liquidity index (LI) of SPT-N S- wave velocity (m/s)
0.19–0.63. A gradual decrease in the LI relating to depth was also 0 20 40 60 0 300 600
observed in this excavation, suggesting that shear strength may rise 0 0
with depth. A high void ratio was connecting with the high water BH-11
content as shown in Figure 6 (c), especially for the soil layer Soft BH-12
5 5
occupying top 10 m. BH-13 BH-15
Permeability is one of the most significant index properties of soil Very stiff
in the advanced simulation of deep excavation. In situ permeability 10 10
tests, by the falling head method, have been performed in selected
boreholes (BH15-17). After firstly observing the initial level of the Hard
groundwater table, then the casing is loaded with water until 15 15
achieving the top of the casing pipe. The water drop-down is
Depth (m)
Depth (m)
measured at particular time intervals until attaining the stable or up to
the first initial water level. The permeability test results represent that 20 20
Stiff
the hydraulic conductivity of the in-situ soil in the research Section is
in the range of 3×10−8 to 7×10−8 m/s. Hard
25 25
Average
Fine Content (%) value
Sand Content (%)
30 30
0 50 100 0 50 100 Very stiff
0 0
35 35
5 5
10 10 40 40
(a) (b)
15 15
Depth (m)
Depth (m)
20 20 follows.
25 25 Su = 6 × N (kPa) (2)
206
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Undrained A purposes. The K0 values were defined by using Soil modulus, E50 (kPa)
equations from Jaky. (1944) for normally consolidated soils and from 0 200000 400000
Mayne and Kulhawy. (1982) for overconsolidated soils, as shown in 0
Interpreted from PMT
Eq. (3).
CU test
2800N
K0(NC) = 1−sin ϕ' (3a) 5
4000N
Depth (m)
20
Therefore, the value of K0(NC) in Plaxis 3D follows the default
setting by Jaky. (1994) equation.
25
The soil modulus measured from PMT was expected to be close to
the initial soil modulus (Ei). For comparison, the measurements were
converted to the soil modulus at 50% stress level (E50) by using Eq.
30 E50 = 833 (z+76)
(4), assuming a hyperbolic curve for the stress-strain relationship.
2−𝑅𝑓
𝐸50 =𝐸𝑖 (4) 35
2
Su = 170×e-4.6LI (kPa)
15
length of 40 m was performed by symmetric model to represent
the excavation case (E-E) which is located in the plane-strain
condition of the excavation. This was because the influence of corner
20 effects becomes inconsiderable once the distance is more than 30 m
Kulhawy & Mayne (1990) from the corner. At the distance of 30 m away from the corner, the
Su = 6N (kPa) wall is technically under plane-strain conditions and the wall
25
≈ 0.32'v deflection remains almost the same (Ou. 2006, Ou et al. 1996, and
Hsiung et al. 2016). Thus, the total length of the excavation in the
design was 80 m. Moreover, shortening the excavation length can
reduce the dimension of the 3D model, saving computational time and
30 cost without affecting the numerical results.
Ten construction phases were performed with the total of 5
excavations and 4 slabs as listed in Table 1. The groundwater table
35 was initial set 2.0 m below the ground surface and lowered to 1.0 m
below excavation surface at each excavation phase. The distance from
Figure 8 Undrained shear strength with depth the lateral boundaries of the model to the diaphragm wall in the
(Hsiung et al. 2018) excavation was 80 m, which was about four times the excavation
207
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Note: m = 1.0; Rf = 0.9; v' = 0.3; vur = 0.2; 𝐸′50 = 0.867 𝐸50 and 𝐸′ = 1.5 𝐸′50
depth. Standard fixed conditions were used to the FE model; Two undrained functions, namely Undrained (A) and Undrained
horizontal movement was fixed at the lateral boundaries, and both (B), were specified in Plaxis 3D: Undrained (A) requires effective
horizontal and vertical movements were fixed at the bottom boundary stress parameters for both soil modulus and shear strength, whereas
of the model. The soil model, structural elements (diaphragm walls Undrained (B) is performed using the effective soil modulus and
and floor slabs), and soil-structure interface elements are discussed in undrained soil shear strength.
the following Section. The explanation of the undrained soil shear strength was
discussed in Section 2.3.2. The numerical analyses results were
3.2 Input parameter performed using PMT as shown in Eq. (5) were compared with field
measurement wall deformation
3.2.1 Soil parameter
As noted, in the Plaxis 3D manual, when selecting Undrained B,
Mostly, previous researchers in Indonesia often used plasticity index the soil modulus in HS model becomes stress-independent. It is
(PI) correlations to obtain the effective friction angle for Central entirely different with Undrained A which needs to convert the value
Jakarta clay as shown in Figure 10 that was conducted by Bjearum 𝑟𝑒𝑓
of the drained reference soil modulus (𝐸50 ) by Eq. (7).
and Simons. (1960). This concept is applicable for soft clay, but
misleading with the Central Jakarta clay which possibly gives the 𝑟𝑒𝑓 𝑐′ 𝑐𝑜𝑠 𝜙' + 𝜎 ′ 3 𝑠𝑖𝑛 𝜙'
𝑚
underestimated value of strength parameter. Moreover, establishing 𝐸′50 = 𝐸50 ( ) (7)
𝑐′ 𝑐𝑜𝑠 𝜙' + 𝑃𝑟𝑒𝑓 𝑠𝑖𝑛 ϕ'
the correlation value of soil stiffness corresponding to the depth has
to be appropriate to achieve the better result compared with field 𝑟𝑒𝑓
Therefore, E′50 can be used directly as 𝐸50 in the Undrained B
measurement.
model because the value of effective soil strength becomes
In this research, Hardening Soil (HS) result was compared with
meaningless. However, the stress dependency of the soil modulus was
Mohr-Coulomb to evaluate the effective soil strength used in
still modeled by manually inputting various soil modulus values for
Undrained A with previous research conducted by Hsiung et al.
soil layers to consider the change of the soil modulus with depth.
(2018) using Undrained B performed by Plaxis 3D. Table 3 shows the 𝑟𝑒𝑓
input soil parameters in the HS model for the FE analyses. The soil Table 3 lists the 𝐸50 values used in the simulation. As known in the
strength is divided into two layers. The upper layer performed with 6 Mohr-Coulomb soil model which only has a single soil modulus value
kPa of effective cohesion and 38o of effective friction angle obtained that can be estimated to be 𝐸′50 = 1.5 E′ in this model. The HS model
from high-quality of CU test. The lower layer used 46 kPa of effective allows to input separate soil modulus values to distinguish the soil
cohesion and 41o of effective friction angle. The effective stress behavior under loading and unloading conditions. According to Lim
analysis under undrained conditions was simulated to model the et al. (2010) and Calvello and Finno. (2004), the reference modulus
short-term undrained conditions of the clay during the excavation for unloading/reloading and oedometer loading were estimated to be
𝑟𝑒𝑓 𝑟𝑒𝑓 𝑟𝑒𝑓 𝑟𝑒𝑓
process. The assumption of soil undrained conditions can be selected 𝐸𝑢𝑟 = 3𝐸50 and 𝐸𝑜𝑒𝑑 = 0.7𝐸50 . Similar procedures were suggested
by low soil permeability, as described previously. Moreover, the real by Surarak et al. (2012) to determine input soil parameters in the HS
construction revealed that no pumping required, which is suggested model. Moreover, the sensitivity of each parameter on the wall
that the soil in the research remained soil to undrained conditions. deformation of an excavation was discussed by Gebreselasse and
Kemfert. (2005).
208
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Table 4 shows the input material properties of the diaphragm wall Plaxis 3D has the interface menu to design the plate elements.
used in Plaxis 3D. According to Ou. (2006), the axial stiffness of floor Interface elements were used to model the correlation between the
slabs in the top-down construction method must also be reduced by soil and structural elements such as the diaphragm wall and bottom
20% of the total stiffness considering the construction process. This concrete slab. The value of the interface reduction factor (Rinter)
is because the compressive strength of the constructed concrete might influences both the stiffness and strength of the interface. According
differ from the design strength during the construction; defects and to Ou. (2006), Rinter = 0.67 is a typical value for the interface between
cracks in the concrete can also be considered through this approach. clay and concrete interface.
The slab thickness for each floor level and the parameters for the slabs
are listed in Table 5.
4. RESULTS AND DISCUSSIONS
Table 4 Input parameter of diaphragm wall 4.1 Comparison of prediction and field measurement
Figure 12 describes the comparisons of the measured and predicted
Parameter Symbol Value Unit wall from several soil constitutive models included HS Undrained A,
Compressive strength of concrete f'c 21 Mpa HS Undrained B, MC Undrained A and MC Undrained B. Central
Thickness d 1 m Jakarta excavation was selected to model the simulation and it is
located in the plane-strain condition as shown in Figure 3.
Young's modulus E 21700 Mpa
As displayed in Figure 12, the wall deformation kept raising
Young's modulus 70% 70%E 15200 Mpa corresponding to the excavation phase until reaching the maximum
Unit weight 𝛾 6 (kN/m3) value in the last stage of the excavation. The comparison indicates
that the input parameter used performed well to all soil constitutive
Poisson’s ratio v 0.15 - models. However, the 3rd excavation shown in Figure 12 (c) describes
that the numerical analysis was overestimated compared with field
Table 5 Input parameter of concrete slabs measurement data. Numerical analysis results in 5th excavation as
displayed Figure 12 (d) are located between the field measurement
Slabs d (m) v 80%E (MPa) data both sides (DW-15 and DW-171).
The predicted wall deformation for soil near wall toe of HS results
Deck slab 0.4 0.15 17400 was lesser than MC in the below of excavation zone because the
Top slab 0.8 0.15 17400 stiffness moduli in the HS could take the elastic unloading process
𝑟𝑒𝑓
Middle slab 0.4 0.15 17400 due to the excavation process by 𝐸𝑢𝑟 together with comparatively
Bottom slab 1 0.15 17400
short embedded wall depth. Moreover, the simulation results of
209
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5 5
GL. -5.28 m
(a) (b)
10 10
Measurement
Depth (m)
Depth (m)
DW-15
15 15
Measurement
DW-171
HS
20 Undrained A 20
HS
Undrained B
25 MC 25 (c) (d)
Undrained A
MC
Undrained B Figure 13 The shading of relative stress at different excavation
30 30 stages from MC Undrained A: (a) 1st stage; (b) 2nd stage; (c) 3rd
(a) (b) stage; (d) 5th (final excavation) stage
GL. -11.38 m
15 15 analyses at every stage toward the soil strength. Eq. 8 and 9 can be
Bottom slab rewritten regarding of the effective stress. Moreover, Eq. 10 can
express the relationship between the Kf line and the Mohr failure
20 20 envelope:
GL. -19.03 m
(𝜎 ′ 𝑉 +𝜎 ′ 𝐻 )
p' = (kPa) (8)
25 25 2
(σ′ V −σ′ H )
q= (kPa) (9)
30 30 2
(c) (d)
sin ϕ' = tan ψ (10)
Figure 12 Comparison of predicted and measured wall All those ESP points were taken at every excavation depth and the
displacement from several soil constitutive models at the excavation toe of diaphragm wall as shown in Figure 14. Each point was
at various excavation stages: (a) 1st stage; (b) 2nd stage; (c) 3rd stage; observed for every stage of excavation. Points E and F were observed
(d) 5th (final excavation) stage at the end of the embedded wall to represent the effective stress path
because in the last excavation process, shown in Figure 13 (d),
Undrained A for both HS and MC recorded the consistent result and illustrates the highest stress occurred.
located in the between of measurement data which mean the effective The stress path values are corresponding to the location of
soil strength and total stress parameter are similar and further excavation depth and stress path points. The Kf line was obtained by
discussion will be carried out in the later section. the high-quality data of effective strength. When the stress path points
Figure 13 shows the shading of relative shear stress for each are above the excavation depth, then the value of overburden stress
excavation process. The relative shear stress is the ratio of the would increase and would be in reverse when the stress path points
mobilized shear strength and the maximum shear strength which under the excavation depth as shown in Points A, B, and C. Those are
illustrates the proximity of the stress point to failure envelope. The due to the horizontal stresses 𝜎𝐻 in the retaining zone decreases when
figure used MC Undrained A to see the evolution of relative shear the soils are excavated above the stress points and 𝜎𝐻 should increase
stress in different phases. The shading of relative stress at 1st stage when the soils are excavated below the stress path points. Moreover,
represented if the wall had no experience of high ground pressure in the point E shows that the overburden pressure increases
against the wall as shown in Figure 13 (a). The transformation corresponding to the excavation stages due to unloading process.
occurred in Figure 13 (b) and (c) which the wall has taken on the soil Figure 15 indicates that those values, based on numerical analysis
pressure. Eventually, in Figure 13 (d) indicates that the soil elements results, were located below the Kf line and far enough to reach the Kf
around the excavation site and the embedded length of diaphragm line which indicates that the models were in the safe zone and
wall in plane-strain location became critical in the design of the appropriate for the structure design. This indicates that by having a
numerical analysis. high-quality data of effective strength parameter and selecting
stiffness moduli are necessary to obtain the consistent simulation
210
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
results of Undrained A and Undrained B, respectively. However, clay during the excavation process. The following conclusions were
unless having a good quality test result, relative loading between soil illustrated by the results of this research:
stress status and failure envelope may not be the same which could
lead totally different results from Undrained A and Undrained B 1. The ground condition in Central Jakarta generally dominated by
analyses. Then, HS model with Undrained B analysis still can give clay/silt which consisted of soft clay to very stiff clay. The natural
acceptable results and it means as long as a set of proper soil stiffness water content of the clay in the range of 34.6 to 89.3% with LI of
can be given, HS model with total stress parameter can give a between 0.19 and 0.65. The SPT-N values were lower than 10 to a
reasonable result. depth of 7 m (soft clay) and continuously in the average of 30 at
depth 20 m (stiff clay).
2. The soil model was separated into upper soil layer and lower soil
layer to simplified the input soil parameter. The upper layer
performed with 6 kPa of effective cohesion and 38o of effective
friction angle. Moreover, the lower layer used 46 kPa of effective
cohesion and 41o of effective friction angle.
3. For Central Jakarta clay, a linear relationship between the soil
modulus (from PMT) and depth was suggested to use E 50 = 833
A (z+76), were measured in kilopascals and meters. This empirical
equation results were located in between of soil moduli obtained
based on 2400N and 4000N.
4. The results of numerical analyses were compared between
predicted and measured. The results indicate that the effective soil
strength parameter used in Undrained A was reliable to be used in
B the analysis because the wall deformation results of predicted,
located close to both sides of measurement data and Undrained B,
were consistently good in numerical performance.
5. The stress path was conducted to confirm the result of diaphragm
wall in the excavation site. The effective stress path points based
on numerical results were located below the Kf line indicate that
Undrained A and Undrained B could be performed equally well to
C be used.
211
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Hsiung, B.C.B., Yang, K.H., Aila, W., Hung, C., 2016. Three-
dimensional effects of a deep excavation on wall deflections in
loose to medium dense sands. Comput. Geotechn. 80, 138–151.
Hsiung, B.C.B., Yang, K.H., Aila, W., Ge, L., 2018. Evaluation of
the wall deflection of a deep excavation in Central Jakarta using
three-dimensional modeling. Tunnelling and Underground
Space Technology. 72 (2018) 84-96.
Hwang, R.N., Moh, Z.C., 2007. Deflection paths and reference
envelopes for diaphragm walls in the Taipei Basin. J. GeoEng.
1, 1–12.
Jaky, J., 1944. The coefficient of earth pressure at rest. J. Soc.
Hungarian Archit. Eng. Budapest Hungry 355–358.
Kulhawy, F.H., Mayne, P.W., 1990. Manual on Estimating Soil
Properties for Foundation Design. Electric Power Research
Institute, Palo Alto, California.
Lim, A., Ou, C.Y., Hsieh, P.G., 2010. Evaluation of clay constitutive
models for analysis of deep excavation under undrained
conditions. J. GeoEng. 5, 9–20.
Lim, A., Ou, C.Y., 2017. Stress paths in deep excavations under
undrained conditions and its influence on deformation analysis.
Tunnelling and Underground Space Technology. 63 (2017)
118-132.
Likitlersuang, S., Surarak, C., Wanatowski, D., Oh, E.,
Balasubramaniam, A., 2013. Finite element analysis of a deep
excavation: A case study from the Bangkok MRT. Soils
Foundat. 53 (5), 756–773.
Mayne, P., Kulhawy, F.H., 1982. K0-OCR relationships in soil. J.
Geotechn. Eng. Divis. 108 (GT6), 851–872.
Moh, Z.C., Hwang, R.N., 2005. Geotechnical considerations in the
design and construction of subways in urban areas. Seminar on
recent developments on mitigation of natural disasters, urban
transportation and construction industry, Jakarta, Indonesia.
Muir-Wood, A., 1983. Index properties and critical state soil
mechanics. Paper presented at the Proceeding of the
Symposium on Recent Developments in Laboratory and Field
Tests and Analysis of Geotechnical Problems, Bangkok,
Thailand.
Ou, C.Y., Chiou, D.C., Wu, T.S., 1996. Three-dimensional finite
element analysis of deep excavations. J. Geotechn. Eng. 122
(5), 337–345.
Ou, C.Y., Shiau, B.Y., Wang, I.W., 2000. Three-dimensional
deformation behavior of the Taipei national enterprise center
(TNEC) excavation case history. Canad. Geotechn. J. 37 (2),
438–448.
Ou, C.Y., 2006. Deep Excavation: Theory and Practice. Taylor &
Francis, Netherlands.
Schweiger, H.F., 2009. Influence of constitutive model and EC7
design approach in FEM analysis of deep excavations. In:
Proceedings of ISSMGE International Seminar on Deep
Excavations and Retaining Structures, Budapest, Hungary, pp.
99–114.
Surarak, C., Likitlersuang, S., Wanatowski, D., Balasubramaniam,
A., Oh, E., Guan, H., 2012. Stiffness and strength parameters
for hardening soil model of soil and stiff Bangkok clays. Soils
Foundat. 52 (4), 682–697.
Wang, J.H., Xu, Z.H., Wang, W.D., 2010. Wall and ground
movements due to deep excavations in Shanghai soft soils. J.
Geotechn. Geoenviron. Eng. 136, 985–994.
Yang, K.H., Uzuoka, R., Thuo, J.N., Lin, G.L., Nakai, Y., 2017.
Coupled hydro-mechanical analysis of two unstable
unsaturated slopes subject to rainfall infiltration. Engineering
Geology. 216 (2017) 13-30.
Yong, K.Y., 2015. Learning lessons from the construction of
Singapore Downtown line (DTL). In: Proceedings of
International Conference and Exhibition on Tunneling and
Underground Space, Kuala Lumpur, Malaysia.
212
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Plane Strain Ratio and Waling Size Evaluation of Deep Excavation of Kuala Lumpur
Using 3D Finite Element Analysis
Jen Shen Ang1*, Bin Chen Benson Hsiung1, Ching Hung2, A. J. Li3, J. W. Chen2
1
Department of Civil Engineering, National Kaohsiung University of Science and Technology, Kaohsiung, Taiwan
2
Department of Civil Engineering, National Cheng- Kung University, Tainan, Taiwan
3
Department of Construction Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan
E-mail: 1102102171@gm.kuas.edu.tw
ABSTRACT: This paper presents a study based on 3D Finite Element Analysis of deep excavations in Kuala Lumpur, Malaysia. The
ground in Kuala Lumpur mainly consists of a layer of highly permeable sand, occasionally with some thin layers of clay above rock which
includes interbedded sandstone, siltstone, shale and mudstone for so called “Kenny Hill Formation” or limestone with sinkholes for so called
“KL limestone formation”. A common feature is the depth of rock varies in a widely range which leads to challenges of design and
construction of excavations. By using 3D finite element analysis together with observational data, it aims to explore plane strain ratio (PSR)
of deep excavation of Kuala Lumpur which can indicate impacts from the corner effect associated with distance to the corner. Influences on
PSR from certain factors, such as centre- to- centre distance of struts and hard soil stratum/rock depth are also covered and examined in this
study. Moreover, individual secant bored piles (SBP) wall have to be adopted for retaining structure of deep excavation in Kuala Lumpur as
hydraulic- grab type diaphragm wall can’t be constructed in very hard rock, such as limestone and thus steel waling has to be installed in
order to connect each SBP, not only providing the function of strut- wall connection. It is not possible to apply 2D analysis to examine
waling size in the aspect of function of SBP connection so 3D analysis has to be conducted instead for said purpose. The evaluation of
waling size is thus included in this study.
Keywords: Deep excavation, Kuala Lumpur, 3D finite element analysis, Plain Strain Ratio, SBP wall, walling size
213
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
machine) need to launch from this site, that is the reason there is one
section having a very large spacing of strut, which is 17m in order to
fit the whole TBM into the pit.
214
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The numerical modelling of Case KLCC is built associated with the of this model is used “normal” mesh setting. The model consists of
information which provided by Law et al. (2016), such as mesh 416,761 nodes and 284,581 ten-node tetrahedral elements.
boundary and geometry of deep excavation side. The groundwater
table is observed at approximately 2.5 m below the ground surface
and set the groundwater table into PLAXIS 3D. The mesh boundary
and phreatic level are shown as Figure 5. The total nodes in the 3D
model are 147,229 and 99,446 is the number of total ten-node
tetrahedral elements. The “normal” mesh is applied to the whole
model.
(b)
Figure 8 Detail supporting system in PLAXIS 3D: (a) Plate element
(retaining wall); (b) Node to node element (struts)
Figure 6 Dimension of retaining wall and struts spacing 3.2 INPUT PARAMETERS
For Case Waler, since it aims to evaluate whether the capacity of Hardening soil (HS) model is adopted to simulate the soil behaviors
waler is enough to take the stresses from retaining wall or not so the in Case KLCC. HS model is an advantage model for simulating the
geometry of whole excavation is not simulated by using PLAXIS behavior of different types of soil, both soft soils and stiff soils
3D in order to shorten time of generation of model and running time reported by Schanz (1999). Law et al. (2014) highlighted that the
of computer operation. A rectangular shape of 3D model is general problem in the analysis and design of deep excavations in
conducted to undertake the analysis. The mesh boundary of PLAXIS loose to medium dense sand with shallow meta-sedimentary hard
3D is length of 290 m (x), width of 200 m (y), depth of 104.5 m (z) layer which soil tests information is limited and low-quality
presented as Figure 7. The geometry of excavation is length of 120 sampling. Normally, the shear strength of soil can be obtained from
m and 30 m for width in plan view and the strut spacing of 6 m with laboratory triaxial tests and direct shear tests but not on the stiffness
an area of 17 m strut spacing (Figure 8). The purpose excavation of soil. Due to the reasons, one of the field tests which called
level is 31 mbgl and the toe of wall is located at 28 mbgl. The mesh Standard Penetration Test (SPT) is conducted to obtain the
215
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
information of subsoil for geotechnical design and analysis purpose. calibration by experience from previous study on similar ground
However, the empirical correlation between stiffness of soil and condition. Table 4 summarizes the soil parameters of Case Waler.
number of SPT (SPT-N) are used for prediction the ground HS model is used for upper alluvium soil, Mohr-Coulomb (MC)
movement induced by excavation. model is used for rock simulation. The secant Young’s modulus of
The constitutive soil model of HS model used for case history is soil is also taken with STP-N as below:
studied by Law et al. (2016), the performance between numerical
analysis and field observation reading are similarly, the empirical E50ref = 3000N (4)
correlation is suitable for simulating in the formation. Triaxial
secant modulus E50ref with SPT-N is taken as below: For this study, 6 of STP-N is selected.
E50ref = 2000N (1) Table 2 Diaphragm wall elastic properties (Plate element)
The ratios between E50ref, Eoedref and Eurref are as follows: Parameter Name Value Unit
Compressive strength of concrete fc’ 40 MPa
Thickness d 0.6 M
E50ref = Eoedref (2)
Young’s modulus E 28 x 106 kPa
Young’s modulus x 70% 70% E 19.6 x 106 kPa
E50ref = Eurref (3) Unit weight w 9 kN/m3
Poisson’s ratio v 0.2
The simple assumption as suggested by Schanz et al. (1999) and
Brinkgreve et al. (2012), the effective stress strength (c’ and ф’) and Table 3 Steel struts properties (Anchor element)
stiffness parameters adopted in the 2D numerical study are
summarized in Table 1. Based on Tan (2010), the ground properties Level Strut size Section EA 60%EA
of excavation in residual soils is fully drained condition. Therefore, Area (m2) (kN) (kN)
effective stress analysis (drained material) is performed in the 3D Level 1 H400 x 400 0.0219 4.483 x 106 2.6898 x 106
numerical back analysis.
Table 4 Input soil parameters of case CO
Table 1 Input soil parameters of case KLCC
(a) Sandy soil
Symbol Unit S1 S2 S3 S4 S5 S6
c’ kPa 1 5 5 8 15 50 Layer Depth Drainage Unit weight ϕ' E50ref
ϕ' o 28 30 31 33 35 35 (m) type (kN/m3) (kPa) (o )
(kPa)
E50ref MPa 15 30 45 75 150 225
Eoedref MPa 15 30 45 75 150 225 1 0.0-25 Drained 19 1 29 18 x 103
Eurref MPa 45 90 135 225 450 675
m - 0.5 0.5 0.5 0.5 0.5 0.5 (b) Limestone (Mohr- Coulomb model)
vur - 0.2 0.2 0.2 0.2 0.2 0.2
pref kPa 100 100 100 100 100 100 Depth Soil c' ϕ' E'
Layer
Rinter - 0.67 0.67 0.67 0.67 0.67 0.67 (m) type (kN/m3) (kPa) (o ) (kPa)
2 25-104.5 Limestone 24 400 32 1 x 106
Linear elastic plate element is used to simulate the diaphragm
with 6-node. The Young’s modulus of plate element is assumed 19.6
GPa which 70% of the concrete compression stiffness of 28GPa. Three types of plate element are adopted in this study and the
Assuming the diaphragm wall is “wished-in-place”, which means input parameters of plate are summarized in Table 5. There are
this work does not consider the stress and ground movement during primary piles (a), secondary piles (b) and average piles (c). Equation
the construction diaphragm such as trench excavation and (x) for secondary piles and average piles, equation (x) for primary
concreting. Table 2 is the input parameters of plate element wall caused of no steel rebar in the pile. Corresponding to the
(diaphragm wall). A required input parameter is unit weight, due to British Standard Institute (BS 8110), Young’s modulus of concrete
the overlapping of unit weight and volume between soil element and can be obtained by:
plate element, so need to subtract the surrounding unit weight of soil
for the unit weight of plate element (diaphragm wall). The node to E = 20 + 0.2 x fc’ (MPa) (5)
node anchor is used for simulating the steel struts to support
diaphragm wall. The input parameter of anchor is EA, where E is 70% of Young’s modulus is used for input parameters. The primary
the Young’s modulus of steel struts, and A is the cross-sectional piles and secondary piles are modeled at the 17m spacing of struts.
area of steel struts. 60% of Young’s modulus of steel is adopted for Universal beam (UB) and universal column (UC) are selected
the input parameters. The parameters of node to node anchor are for strutting system in Case CO, the steel grade of S 355 for UB and
summarized in Table 3. As the plate element is assumed to be fully UC. Struts are also simulated by node to node anchor. Table 6 is the
connected to each other in the model, waler is thus not simulated in input parameters required by anchor element (strut).
this research.
According to Ang et al. (2017), only two layers of soils and Table 5 Various retaining wall properties
rocks are categorized of Case Waler, which is karst bedrock
(limestone) underlying by loose sand. For simulating the behavior of (a) Primary piles
soil and bedrock, two constitutive soil models were selected, HS Parameter Name Value Unit
model for the upper loose sand with drained and MC model for the Compressive strength of fc’ 40 MPa
limestone with undrained A. Advanced model is not eligible to be concrete
adopted due to limit of available and reliable site investigation data. Thickness d 0.88 M
The modeled soil profile is based on available borehole logs and site Young’s modulus E 28 x 106 kPa
survey information. Due to limited results of in-situ tests and the Young’s modulus x 70% 70% E 19.6 x 106 kPa
condition of soil during the time of testing and sampling quality may Unit weight w 9 kN/m3
be highly disturbed and the result by using empirical method to Poisson’s ratio v 0.2 -
216
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(b) Secondary piles Similarly, bottom-up construction method is also adopted for the
Case Waler. Number of total excavation stages is 11 and the purpose
Parameter Name Value Unit excavation depth is 31.5 mbgl. There are 9-layers strut for
Compressive strength fc’ 40 MPa supporting the retaining wall. For strutting excavation, every depth
of concrete of excavation stage is 1 m below strut level and the excavation
Thickness d 1.48 M without any supporting system when the excavation starts from top
Young’s modulus E 28 x 106 kPa of bedrock.
Young’s modulus x 70% E 19.6 x 106 kPa The actual prestress of the struts applied is around 20% to 30%
70% of design strut load. Point load is used to simulate prestress of the
Unit weight w 9 kN/m3 struts at both ends of struts. Table 9 is the construction phase
Poisson’s ratio v 0.2 - description for Case Waler.
(c) Average piles Table 4.9 Construction phases of case CO
Parameter Name Value Unit Phase Construction
Compressive strength fc’ 40 MPa 0 Initial phase (stress regeneration)
of concrete 1 Installation of second bored piles wall
Thickness d 1.072 M 2 1st stage excavation to 3.5 mbgl, dewatering to 3.5 mbgl
Young’s modulus E 28 x 106 kPa 3 Installation of 1st layer of struts at 2.5 mgbl
Young’s modulus x 70% E 19.6 x 106 kPa 4 2nd stage excavation to 7 mbgl, dewatering to 7 mbgl
70% 5 Installation of 2nd layer of struts at 6 mgbl
Unit weight w 9 kN/m3 6 3rd stage excavation to 9 mbgl, dewatering to 9 mbgl
Poisson’s ratio v 0.2 - 7 Installation of 3rd layer of struts at 8 mgbl
8 4th stage excavation to 11 mbgl, dewatering to 11 mbgl
Table 6 Steel strut properties 9 Installation of 4th layer of struts at 10 mgbl
Level Strut size Section EA x 60%EA x 10 5th stage excavation to 14 mbgl, dewatering to 14 mbgl
Area 106 106 11 Installation of 5th layer of struts at 13 mgbl
(m2) (kN) (kN) 12 6th stage excavation to 16 mbgl, dewatering to 16 mbgl
13 Installation of 6th layer of struts at 15 mgbl
1 2-UB 610x324 0.106 21.73 13.038
14 7th stage excavation to 18 mbgl, dewatering to 18 mbgl
2 2-UB 610x324 0.106 21.73 13.038
15 Installation of 7th layer of struts at 17 mgbl
3 3-UB 610x324 0.159 32.56 19.557
16 8th stage excavation to 21.5 mbgl, dewatering to 21.5
4 3-UB 610x324 0.159 32.56 19.557
mbgl
5 3-UB 610x324 0.159 32.56 19.557
17 Installation of 8th layer of struts at 20.5 mgbl
6 3-UB 610x324 0.159 32.566 19.557
18 9th stage excavation to 23.5 mbgl, dewatering to 23.5
7 3-UB 610x324 0.159 32.56 19.557
mbgl
8 2-UB 610x324 0.106 21.73 13.038
19 Installation of 9th layer of struts at 22.5 mgbl
9 2-UB 610x324 0.106 21.73 13.038
20 10th stage excavation to 25 mbgl (bedrock), dewatering
to 25 mbgl
3.2 Computational sequences
21 Excavation to purpose depth (31.5 mbgl), dewatering to
The construction method of Case KLCC is bottom-up 31.5 mbgl
construction. Since the site only has one-layer of steel strut to
support the retaining wall, so the total of construction phases in 4 RESULTS AND DISCUSSIONS
PLAXIS 3D are 5 phases, the depth of first excavation stage is 0.5
From the Figure 9 is the comparison between the observation and
meter below the first struts layer, the simulation of phases is
numerical analysis of Case KLCC. The inclinometer name of IN-3
described at Table 7.
and IN-8 are installed at the approximately center of long walls and
the lateral wall movement is no significant difference between
Table 7 Construction phase for case KLCC
inclinometers readings and result of numerical model analysis due to
Phase Construction the cross-section might be in the plain strain condition. As
0 Initial phase (stress regeneration) illustrated in figure 9 (IN-1 and IN-10), modelling the wall
1 Installation of diaphragm wall underestimates the lateral wall movement possibly due to
2 1st stage excavation to 4.5 mbgl, dewatering to 5.5mbgl comparatively worse quality of vertical joint between the
3 Installation of 1st layer of struts rectangular panel and L-shaped panel.
4 Excavation to purpose depth (13.2mbgl), dewatering to The location of IN-5 and IN-6 are stood along the
14.2 mbgl complementary wall, the wall movement decreases while decreasing
the distance from the corner. Although IN-5 is at the centre of
For the preloading of steel struts is already created in the complementary wall, but the wall movement at IN-5 is smaller than
structure section as a point load instead of the prestress function of IN-3 and IN-8. The reason is IN-5 is located on a relatively short
PLAXIS 3D by using point element. In calculation analysis, the wall. Refer to the Figure 9, majority of inclinometers reading are in
calculation type of K0 procedure for initial phase and plastic good agreement with the prediction expect the ones near the corner
analysis for the rest of construction phases, the calculation of pore by reasons stated above.
pressure in all the construction phases are used phreatic. The For Case Waler, this is the Class B analysis which means to do
calculation types are described as below (PLAXIS): the analysis at the same time during occurrence of the event.
1. K0: Direct generation of initial effective stresses, pore Currently, the construction progress of the site is going to second
pressures and state parameters; excavation stage where excavate to 7 mbgl. This study only shows
2. Plastic: Elastoplastic drained or undrained analysis. the comparison of first excavation stage between inclinometer
Consolidation is not considered readings and numerical analysis (Figure 10). There is significant
3. Phreatic: Direct generation of steady-state pore pressures difference between the readings and result. The difference between
from phreatic level and cluster-related conditions. numerical analysis and inclinometer readings around 10 mm, the
217
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(a)
Figure 9 Comparison between observation and numerical analysis of
Case KLCC
218
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(b)
(a)
(c)
219
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
220
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
due to the wider excavation has a low stiffness at the in Soil Mechanics and Geotechnical Engineering,
corner. Proceedings of the 5th International Young Geotechnical
For Case Waler: Engineers’ Conference, pp559-563.
Law K. H., Ismail Z. and Roslan H. (2016) “3D Finite Element
1. From the comparison between results of the numerical
Analysis of a Deep Excavation Considering the Effect of
analysis and observations, it is likely the stiffness of soil is
Anisotropic Wall Stiffness” 19th Southeast Asian
underestimated. As the result, the analysis might
Geotechnical Conference & 2nd AGSSEA Conference
overestimate the lateral wall deflection.
(19SEAGC & 2AGSSEA) Kuala Lumpur, pp659-664.
2. Due to 3- dimensional characteristics of the model and
Law K. H., Hashim, R. and Ismail Z. (2013) “Performance of
software, the result of shear force and bending moment on
Multipropped Deep Excavation in Kenny Hill Formation”,
wall from numerical analysis of PLAXIS3D is possible to
Proc. 18th Southeast Asian Geotechnical Conference,
be adopted for the evaluation of waler sizes.
Singapore, pp705-712.
3. The envelope of suitable waler size is suitable for prelim
Ou, C. Y., (2006) “Deep Excavation: Theory and Practice” Taylor &
selection of the situation with 17 m of strut spacing in
Francis, Netherlands
various depths of bedrock with upper alluvium soil.
Ou, C. Y., Chiou, D.C., and Wu, T. S. (1996) “Three-dimensional
4. The reason why having the 17 m of strut spacing is caused
Finite Element Analysis of Deep Excavations”, Journal of
by the need to put a whole tunnel bored machine (TBM)
Geotechnical and Geoenvironment Engineering, pp337-345
from ground surface to purpose excavation level. The
Schanz, T., Vermeer, P. A., and Bonnier, P. G. (1999) “The
envelope of suitable waler size is a reference for selecting
hardening soil model: formulation and verification”, Beyond
waler sizes in this kind of situation.
2000 in Computational Geotechnics, Balkema, Rotterdam,
pp281-296.
6 REFERENCES
Skempton, A. W. (1986) “Standard Penetration Test Procedures”
Ang J. S., Hsiung B. C. B. and Wu, M. R. (2017) “Performance of Geotechnique, vol. 36, no. 3, pp.425-557.
Deep Excavation in Kuala Lumpur” Proceedings of the 17th Tan, S. A. (2010) “One North Station Excavation in 30m of Jurong
Conference on Current Researches in Geotechnical Residual Soils in Singapore”, Earth Retention Conference,
Engineering in Taiwan. Washington, pp732-739.
Bishop, A. W. (1966) “The Strength of Soil as Engineering Tan Siow Meng and Simon (2011) “Karstic Features of Kuala
Materials” Geotechnique, pp91-128 Lumpur Limestone”
Brinkgreve, R.B.J. et al., Editors (2012) “Plaxis 3D 2012 user’s Waltham, A. C. and Fookes, O.G. (2003) “Engineering
manual”, Netherlands, Plaxis bv. classification of karst ground conditions”. Quarterly Journal
BS 5950: Part 1 (2000), Structure use of Steelwork in Building, Part of Engineering Geology and Hydrogeology, pp101-118
1: Code of Practice for Design and Construction, British Zdravkovic, L., Potts, D. M., and St Jogn, H. D. (2005) “Modelling
Standards Institution of a 3D Excavation in Finite Element Analysis”,
Chiew, S. P. and Leow, B. H. (2006) “Common Mistakes in Steel Geotechnique, pp497-513.
Design” International Symposium on Worldwide Trend and
Development in Codified Design of Steel Structures. Kuala
Lumpur, Malaysia.
Chogueur A., Abdeldjalil Z. and Reiffsteck P. (2018) “Parametric
and Comparative Study of a Flexible Retaining wall”
Periodica Polytechnica Civil Engineering, pp295-307
Eurocode 3 Part 1-1 (2005), Design of Steel Structures: Part 1-1
General Rules for Building, British Standards Institution
Ganno J. (2016) “Primary Firm Secant Pile Concrete Specification”
Proceedings of the institution of Civil Engineers, pp110-120
H. Zabidi, M. Termizi, S. Aliman, K. S. Ariffin and N. L. Khalil
(2015) “Geological Structure and Geomorphological Aspects
in Karstified Susceptibility Mapping of Limestone
Formations” 5th International Conference on Recent
Advances in Materials, Minerals and Environment & 2nd
International Postgraduate Conference on Materials, Mineral
and Polymer, pp659-665
Hsiung, B. C. B. (2009) “A Case Study on the Behavior of a Deep
Excavation in Sand. Comput. Geotechn. Pp.665-675
Hsiung, B. C. B., Hwang, R.N. (2009) “Evaluating Performance of
Diaphragm walls by Wall Deflection Path. SEAGS. Special
Issue on Excavation and Tunneling in Geotechnical
Engineering, pp.81-90
Hsiung, B. C. B., Wang, C. L., Lin, H. T. and Chen C. H. (2013)
“Design and Performance of a large-scale Excavation
Adjacent to Sensitive Structures in Urban area” Proceedings
of the 2nd International Conference on Geotechnics for
Sustainable Development-Geotec. Hanoi, Vietnam
Hsiung, B. C. B., Yang, K. H., Aila, W., Hung, C., 2016. Three-
dimensional Effects of a Deep Excavation on Wall
Deflections in Loose to Medium Dense Sands. Comput.
Geotechn.
Koo Kuan-Seng (2013), “Design and Construction of Excavation
Works for Klang Valley Mass Rapid Transit Underground
Station at Cochorance, Kuala Lumpur, Malaysia” Advances
221
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: The mechanical behavior of soils is highly complex due to the heterogeneous behavior of these deposits. There are various
experimental and empirical methods followed to assess the behavior of these granular masses. However, field tests, laboratory tests and empirical
relationships have inherent uncertainties associated with them. The empirical correlations developed are based on the data collected from various
field tests. The quality of the data collected from various field tests like Standard Penetration Test (SPT) and Cone Penetration Test (CPT)
depends on the field practices which may be inconsistent due to different test methods followed in the field. Also the quantity of data available for
adopting these correlations are not sufficient in some cases. Hence for a sophisticated evaluation of various properties, analytical methods are a
must. In addition to this, the behavior of saturated soil during sudden loading is highly influenced by the development of excess pore water
pressure. The sudden development of excess pore pressure in soils is directly related to the field condition of that soil deposit. Analytical methods
help to understand the gradual accumulation of permanent strain. In this research, various models based on theory of plasticity is adopted for
assessing the behavior under drained static conditions. The data obtained from the reported experimental results are collected for various soils and
are utilized to obtain the elastic and hardening parameters required for the numerical and analytical models. Numerical modeling of the sample
and various loading conditions are simulated using a finite element software. The stress strain relationships which accounts for the accumulation
of plastic strain and the subsequent plastic flow is defined and a comparison of the different models considered is reported.
Keywords: Drained Triaxial test, Cam Clay Model, Drucker Prager Model, Shear stress, Strain
1. INTRODUCTION
2. BACKGROUND STUDY
Granular material is made up of discrete particles. The shape and
Jefferies (1993), developed constitutive model for sand based on
size of particles influence the mechanical behavior of granular critical state theory. The study indicated that soil moves to critical
materials. Also it depends upon particle orientation, friction state regardless of initial conditions. Khalili et. al (2005),
among particles, pore spaces and saturation level of particles. formulated Bounding surface plasticity model for granular soils
When external forces such as static and cyclic loads act on these
subjected to static and cyclic loads. They developed the model
materials, deformation takes place by particle sliding leading to using the available experimental data from literature. It is reported
changes in mechanical behavior. Therefore, understanding the that the developed model predicts stress softening and dilatancy
mechanical behavior is important in designing the structures. during drained loading of dense sand. Arvelo (2005) evaluated the
Because any changes in behavior of soil affects the structures
behavior of dense granular soil using Modified Cam Clay Model
directly and may cause failure of structure (Brenda, 2003). (MCCM). The results of the analytical model are compared to the
When granular material such as sand or silty sand (saturated laboratory drained triaxial test data. The study suggests that
cohesion less soil) is subjected to rapid rate of loading, positive MCCM with some modifications, is applicable to dense sands.
excess pore pressure is developed. This causes decrease in
Thomas Oommen et. al (2010) compared the predictive
effective stress leading to the reduction on shear strength of soil performance of empirical liquefaction models. Deterministic and
and subsequently soil behaves like a liquid. This phenomenon is probabilistic empirical liquefaction models are developed using
known as liquefaction which is usually associated with earthquake SPT and CPT data. But it is found that identifying gaps in data
loads. But liquefaction can occur due to static sudden loads also.
sets is extremely important for improving empirical models. Xilin
This phenomenon is known as static liquefaction. Static & Maosong (2014) proposed a model to analyze the stress strain
liquefaction causes damage to the infrastructure from landslides behavior of soil when it is subjected to static liquefaction. Mohr
and bearing capacity failure. So it is important to study the Coulomb elasto-plastic hardening model is proposed to predict
behavior of soil before infrastructure construction. (Ellison & static liquefaction. It is concluded that static liquefaction is
Andrade, 2009). Triaxial tests are generally conducted to initiated in loose sands when the undrained stress path occurred
understand the the behavior of granular soil. But it is difficult to along with potentially unstable stress path. If sand is dense, it fails
collect the undisturbed sample of loose sand to conduct triaxial when state of sand becomes potentially unstable. Results show
test. To overcome this difficulty, empirical methods are developed that static liquefaction occurred for selected sample at the
based on SPT, CPT and Borehole Penetrometer Test (BPT) hardening stage only before it reaches plastic limit failure. Rani. et
(García et.al, 2012). But empirical method depends upon material al. (2014) studied the prediction accuracy of Mohr-Coulomb and
conditions and is limited to specific topographic conditions. Drucker Prager models in evaluating the behavior of clayey soil.
Accuracy of prediction of behavior of soil adopting empirical It is reported that the Mohr- Coulomb model showed lower
methods depends on knowledge of soil properties such as predictions compared to Drucker-Prager model.
strength, stiffness and sampling method (Beaty and Byrne, 2000).
These all limitations affect the accuracy of evaluation of shear 3. CAM CLAY MODEL
failure of granular soil under various loading conditions. To Cam Clay model is developed to represent the soil behavior based
overcome the limitations of empirical method, analytical methods on critical state. This model considers strength, compression and
are adopted which are simpler, easier and more accurate than critical state aspects while formulating the behavior under various
traditional empirical methods (Lee, 2009). In this study, Cam Clay loading conditions. The basic parameters which describe this
model and Drucker Prager models are used to analyze the model include the effective mean stress, deviatoric stress and
behavior of granular soil subjected to drained static loading specific volume. The initial state of stress is also considered
conditions.
222
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
significant in this model which can be considered in terms of over developed using SOLID 185 element in ANSYS (Figure 2). This
consolidation ratio or pre-consolidation stress. element is suitable for 3-D modeling of solid structure which is
3.1 Collection of experimental data defined by 8 nodes. The material properties given in Table 1 are
The various input parameters required for the model has to be used for the modeling. To replicate the experimental conditions,
obtained from experimental data. For this, drained triaxial test fixed base is given at the bottom of sample by selecting the
data reported by Arvelo (2005) is adopted for this study. From this bottom surface as shown in Figure 3. In this study, the gradual
test, material properties and parameters for Cam Clay model are application of loads is done by providing 600 load steps. These
selected to evaluate the stress and strain behavior of the granular load steps were sufficient to capture the gradual development of
material. This data can be used as the input for both numerical and the non-linear behavior of sand.
analytical modeling. The properties collected from the triaxial test
are given in Table 1. The sand sample is subjected to a confining An isotropic loading condition is imparted to the sample by
pressure of 100kPa. Figure 1 shows the variation of shear stress applying a confining pressure of 100kPa in all directions and is
with shear strain for the drained triaxial laboratory test. shown in Figure 4(a). After isotropic stage, shear stage is
developed by giving vertical displacement at the rate of 2mm per
Table 1 Material Properties of Dense Sand Obtained from minute throughout the last load step. i.e. up to 600 seconds as
Triaxial Drained Test shown in Figure 4(b).
Internal friction angle (Φ in degree) 34.96
Initial voids ratio (eo) 0.28
Swelling index (κ) 0.0021
Slope of critical straight line (M) 1.416
Poisson’s ratio (μ) 0.35
Young’s modulus (E) 48 MPa
Density of sand (γ) 2700 kg/m3
Specific gravity (Gs) 2.7
223
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(a) Deformation
224
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
numerical and analytical solutions is less than that of experimental Table 3 Material properties of Sand
results. This also points to the fact that this model underestimates
the residual value or fails to predict the strain softening behavior. Peak State Internal friction angle (Φ in degree) 39.5
225
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(a) Deformation
226
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Beaty, M. H., and Byrne, P. M., (2000) “A Synthesized Approach Ellison, K. C., and Andrade, J. E., (2009) “Liquefaction Mapping
For Predicting Liquefaction”, WCEE, pp1589-1597. in Finite-Element Simulations”. Journal of Geotechnical and
García, S., Ovando-Shelley, E., Gutierrez, J., and Garcia, J., Geoenvironmental Engineering, November, 135, pp1693-1701.
(2012) “Liquefaction Assessment through Machine Learning”, Lee, Shuh-Gi Chern., and Ching-Yinn., (2009) “CPT-Based
WCEE. Simplified Liquefaction Assessment By Using Fuzzy-Neural
Xilin Lu., and Maosong H., (2014) “Static Liquefaction of Sands Network”. Journal of Marine Science and Technology, 17(4),
under Isotropically and K0-Consolidated Undrained Triaxial pp326-331.
Conditions”. Journal of Geotechnical and Geoenvironmental Sandhya Rani, R., Nagendra Prasad, K., Sai Krishna., (2014).
Engineering, 141(1), pp. 04014087(1-9). “Applicability of Mohr-coulomb & Drucker Prager Models for
Jefferies, M. G., (1993) “Nor-Sand: A simple critical state model Assessment of Undrained Shear Behaviour of Clayey Soils”.
for sand”. Géotechnique, 43(1), pp. 91-103. International Journal of Civil Engineering and Technology,
Khalili, N., Habte, M. A., and Valliappan, S., (2005) “A bounding 05(10), pp. 104-123.
surface plasticity model for cyclic loading of granular soils”. Thomas Oommen, Laurie, G. B., and Richard, V., (2010)
International Journal For Numerical Methods In Engineering, 63, “Validation and Application of Empirical Liquefaction Models”.
pp. 1939-1960. Journal Of Geotechnical And Geoenvironmental Engineering,
136(12), pp1618-1633.
227
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Good and accurate soil investigation is necessary to calculate pile performance with considerable accuracy. Inaccurate or
insufficient soil investigation can lead to underestimation or overestimation of pile capacity. For the former case, the design becomes
uneconomical, while for the latter case, it can lead to unacceptable settlement, and in worst case scenario, even failure. Recently, there was
one such case of an oil tank in East Java. Initial investigation found very dense layer 8 metres below the ground surface until end of
investigation depth (30 m depth). Over 300 driven piles were installed to support the tank. Ten piles were tested with Pile Driving Analyzer,
and 8 out of 10 did not meet the design capacity. Additional soil investigation later revealed that the very dense layer was only 3 m thick.
Below the very dense layer, loose sand was found until 45 m depth. The design was re-evaluated with PLAXIS 3D, and installation of
additional piles were required to meet the settlement criterion of 100 mm. However, since it was possible to jack up the oil tank after its
settlement, it was decided not to do any remediation.
Keywords: pile, serviceability, soil investigation, numerical analysis.
228
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
229
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
liquefied layer. If there is sufficient thickness of non-liquefied soil Table 1 Soil Profile and Parameters Used for Analysis
above the liquefied layer, there will not be liquefaction-induced Young’s
damages. Based on the chart proposed by Ishihara (1985) shown in Depth SPT Unit weight Friction
3 o Modulus
(m) Values (kN/m ) angle ( )
figure 4, for 0.25 g acceleration, as long as the non-liquefied layer is (MPa)
thicker than 4.5 m, there is no liquefaction-induced damages. Hence, 0.0-2.5 15 18 30 18000
the oil tank is deemed safe from liquefaction problems. 2.5-5.5 50 19 27 60000
5.5-47.0 8 16.5 28 9600
3.2 Settlement Analysis 47.0-70.0 60 20 40 72000
To evaluate the settlement of the oil tank, a finite element software,
PLAXIS 3D was used. The soil profile used was idealised using the
3 new SPT data, shown by the dashed line in Figure 2.
Figure 2 Existing and new SPT N values Figure 4 Boundary curves for site identification of liquefaction-
induced damage (Ishihara, 1985)
230
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3.2.2 Numerical Model For the plate element, the required inputs are thickness, unit
weight, Young’s modulus and Poisson’s ratio. The thickness used
The numerical model used is shown in Figure 5. The model had a was 0.6 m, the same as that planned. Unit weight and Poisson’s ratio
dimension of 150 x 150 x 70 m (length x width x height), with the are 28 GPa and 0.2 respectively, typical values for reinforced
oil tank located in the center of the model. The piles were modelled concrete.
as embedded beam elements (Brinkgreve et al., 2017). The broken
piles were not modelled, while the short piles were modelled as they
3.2.4 Modelling Procedure
were on site. The raft was modelled as plate elements, and the
weight of the oil tank was modelled as a distributed load. The modelling of the oil tank followed the following procedure:
1. Generation of initial stress by K0 procedure
3.2.3 Properties of Embedded Beam and Plate 2. Installation/activation of piles
3. Installation/activation of plate
In PLAXIS, the required inputs for embedded beam are Young’s 4. Activate surcharge load
modulus, unit weight, diameter, maximum axial skin resistance and
base resistance. The young’s modulus and unit weight are 28 GPa,
3.2.5 Evaluated Settlement
and 24 kN/m3 respectively, typical values for reinforced concrete.
The diameter of pile was 0.4 m as previously mentioned. In Figure 7 shows the settlement contour of oil tank if the oil tank was
PLAXIS, the axial skin resistance is automatically calculated, but to be constructed as it was. The maximum settlement is located in
not the base resistance. The base resistance has to be calculated the middle of the tank and minimum settlement is located in the
manually and inserted as input. For the axial skin resistance, only edge of the oil tank. The maximum and minimum settlement are 330
the limiting value is required. mm and 230 mm respectively. The settlement was higher than the
Base resistance of pile was calculated using correlations tolerable settlement of 100 mm. Therefore, improvement was
tabulated by Poulos (1989). Decourt (1982) gave a correlation for required.
driven pile in sand as:
where fb is the unit base resistance and N is the average SPT values
in the local failure zone.
4. PROPOSED REMEDIATION
In order to reduce the settlement to the acceptable settlement, it was
Figure 5 Finite element mesh used proposed to install additional piles with longer length in between the
existing piles. The additional piles were to be driven pile with pre-
drilling to 6 metres depth. This was planned to ensure that no piles
were damaged while penetrating the very dense layer from 2.5 to 5.5
m depth. Several configurations were attempted to optimize the
design, while meeting the settlement criterion. Only the final
configuration is presented in this paper.
Figure 8 shows the final proposed layout of additional piles. One
center pile and four ‘rings’ of additional piles were designed. The
additional piles were designed to be 0.6 m diameter. The center pile
and innermost ring of piles (1st ring: shown by blue diamond in
Figure 6) had a depth of 47 m, i.e. founded on the very dense sand
layer. The innermost ring of piles consisted of 12 piles and were
located 5 m from the center of the foundation. The second innermost
ring of piles (2nd ring: shown by green squares) consisted of 24, 40-
m long piles to be located 10 m from the center of the foundation.
The second outermost ring of piles (3rd ring: shown by blue
diamond) had the same depth as innermost ring of piles, i.e. 47 m.
Figure 6 Zoomed in view of foundation piles and pile cap
The second outermost ring of piles consisted of 36 piles, located 15
m from the center of the foundation. The outermost ring of piles (4th
The limiting shaft resistance was also calculated based on
ring: shown by orange triangle) had a depth of 30 m and consisted
Decourt (1982), whereby he gave the correlation
of 48 piles. This gave a total of 121 additional piles.
The embedded beam properties of the additional piles were
fs = 3.3 x N kPa (5)
evaluated using the methods laid out in section 3.2.3. The numerical
procedure was identical to those written in section 3.2.4.
where fs is the unit shaft resistance and N is the SPT values at
certain pile depth.
231
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 8 Plan view of the oil tank with additional piles installed
Figure 9 shows the settlement contour of oil tank with the their design capacity. Additional soil investigations revealed that the
additional piles. The maximum and minimum settlement were 90 previous soil investigation was not accurate. PLAXIS3D was used
mm and 80 mm respectively, with the maximum settlement (shown to re-evaluate the oil tank design, and remediation was suggested to
in red) located at the corners of the tank and minimum settlement meet the initial design criterion.
(shown in blue) located at the second outermost ring of piles. With From the analysis conducted, in order to satisfy the settlement
the added piles, the settlement was reduced from 330 mm to 90 mm. criterion of 100 mm, another 121 piles would be needed, in addition
The differential settlement was also reduced from 100 mm to 10 to the existing 304 piles. Further analysis shows that by using only
mm. 121 piles, the settlement criterion can be met. In fact, without the
existing piles, the 121 piles would perform better and settle less than
5. EVALUATING EFFECTIVENESS OF ADDITIONAL 425 piles. This shows that more piles do not necessarily means
PILES ONLY better performance, optimization is required.
This case also highlights the importance of good and accurate
To evaluate the effectiveness of ‘additional’ piles, additional run soil investigations to an effectiveness of geotechnical design. A
was conducted without the existing piles (i.e. the oil tank is to be good practice is to have the design consultant to supervise the driller
founded only on the 121 piles). to countercheck the borelog. This practice is common in countries
Figure 10 shows the settlement contour of the oil tank, if it were with established geotechnical practice, e.g. Australia. Unfortunately,
to be founded only on the ‘additional’ 121 piles. It can be seen that such practice is rarely practiced in Indonesia. Another important
the settlement magnitude is similar, in fact the oil tank founded on thing to note about soil investigation are the numbers of soil tests.
the 121 piles without the existing 304 piles had slightly lower Many project owners belittle the importance of soil investigation
settlement than the one with existing 304 piles (refer to Figure 9). and try to save cost by limiting the number of soil tests. However,
This is due to pile group efficiency (Poulos, 1989). For the analysis the cost saved by reducing the soil tests is often insignificant
with existing piles included, the center to center spacing between 2 compared to the potential cost saving from design optimization. In
adjacent piles were as close as 1.25 m, which is only twice the worst case scenario, the lack of information can lead to failures, and
diameter of ‘additional’ piles. Whereas, if the existing piles did not lead to heavy losses in terms of money and time. It is an engineer’s
exist, the piles would have 2.5 m spacing, which is more than 4 duty to inform project owners the importance of soil investigation,
times the diameter. Therefore, the ‘additional’ piles would be more as it is the very basis of a geotechnical design.
efficient without the existing piles, and hence less settlement. Due to the urgency and heavy cost to remediate the oil tank, the
project owner decided to build the tank as it is. Another reason is
6. CONCLUSIONS AND DISCUSSION that the oil tank can be built to accommodate the settlement. The
In this paper, an oil tank to be founded on 304 piles encountered accommodation can be done by jacking up the oil tank after its
problem during pile installation. Problems encountered include piles settlement, and the foundation can be releveled (Wit, 1991).
not reaching target depth, damaged piles, and piles not reaching Furthermore, before the oil tank operation, hydrostatic test (Morera
232
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
and Makar, 2009) would be conducted, whereby the oil tank would 4. REFERENCES
be filled to its full capacity with water. Fortunately, the oil tank is
founded on sand, hence the settlement will occur very quickly, and Bowles, J. E. (1977). Foundation Analysis and Design. McGraw-
the releveling can be conducted after the hydrostatic test. Little or no Hill Companies, United States.
settlement is expected after the hydrostatic test, as oil has lower Brinkgreve, R. B. J., Kumarswamy, S., Swolfs, W. M., and Foria, F.
density than water, i.e. the operational weight of oil tank is lower (2017). PLAXIS 2017 Manual. Plaxis bv, Netherlands.
than the hydrostatic test. Decourt, L. (1982). Prediction of the bearing capacity of piles based
In conclusion, a good and accurate soil investigation is required exclusively on N values of the SPT. Proceedings of ESOPT2,
for any construction. Without a good soil investigation, the Amsterdam, 1, 29-34.
foundation design will not be optimum, and cost more than a design Ishihara, K. (1985). Stability of Natural Deposits During
with proper soil investigation. From this case study, one can also Earthquakes. Proceedings of the Eleventh International
learn that remediation may not be necessary, even though the initial Conference on Soil Mechanics and Foundation Engineering,
design criteria is not met. It is dependent on the structure, as well as 1, San Francisco, 321–376.
soil condition. However, careful evaluation need to be conducted Irsryam, M., Widiyantoro, S., Natawidjaja, D. H., Meilano, I.,
before making the decision. Rudyanto, A., Hidayati, S., Triyoso, W., Hanifa, N. R.,
Djarwadi, D., Faizal, L. and Sunarjito (2017). Peta Sumber
Bahaya Gempa Indonesia Tahun 2017. Pusat Penelitian dan
Pengembangan Perumahan dan Permukiman.
Morera, J. and Makar, G. (2009). Specification for Tank Hydrotest.
TransCanada.
Peck, R. B., Hanson, W. E. and Thornburn, T. H. (1974).
Foundation Engineering. 2nd edition, Wiley, New York, 113-
115.
Poulos, H. G. (1989). Pile behaviour – theory and application.
Géotechnique, 39(3), 365–415.
Seed H.B, K. Tokimatsu, L.F. Harder, and Riley M. Chung (1985).
Influence of SPT Procedures in Soil Liquefaction Resistance
Evaluations. Journal of Geotechnical Engineering, ASCE,
111(12), 1425-1445.
Stroud, M. (1989). The SPT – Its applications and interpretation,
Penetration testing in the UK. Thomas Telford, London.
Figure 9 Evaluated settlement distribution with additional piles Terzaghi, K. (1943). Theoretical Soil Mechanics. Wiley, New York.
Wit, J. (1991). Safe jack up method permits repair of tank bottoms
and foundations. Oil & Gas Journal, 89(44).
233
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Tunnels have become an important mode of transportation due to increase in population and density of traffic all over the
world. An underwater tunnel is a tunnel which is partly or fully constructed under a body of water. Underwater structures are not absolutely
immune against dynamic waves and are prone to damages. Thus it is significant to perform the dynamic analysis of underwater tunnels. In
the present study, the dynamic behaviour of underwater single and twin tunnels are studied numerically with different tunnel materials. Study
includes static and dynamic analysis of twin tunnel models with varying overburden, varying position along depth and varying spacing. The
dynamic analysis is done adopting the load data of Kutch earthquake. It is observed that twin tunnels with varying overburden developed
maximum stresses when compared with uniform overburden.
Keywords: Underwater Twin tunnels, Dynamic analysis, Numerical modelling.
234
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
235
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
this, the effect of spacing between the tunnels in the lateral direction
as well as in the vertical direction is also analysed.
When the tunnels are located at two different depth levels from the
uniform soil bed, it was observed that the increase in stress was
Figure 11. Stress in twin tunnels when located at same depth
around 40.89% for the copper twin tunnel model as shown in Figure
from uniform soil bed and at different depths: Dynamic analysis
11. The stress increase was found to be in the range of 40% for the
tunnels with 2d spacing compared to 4d spacing.
5. CONCLUSION
In this study, the behaviour of single and twin tunnels buried below
the water bed is analysed under static and dynamic loading
236
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
conditions. Parametric studies involving the effect of varying Anirban De, Anthony Niemiec, S.M., and Thomas F. Zimmie.
overburden, spacing in lateral direction and depth of tunnels on the (2017), “Physical and Numerical modelling to study
seismic response of twin tunnel models are carried out. Twin tunnels effects of an underwater explosion on a buried tunnel”,
with varying overburden indicated that the maximum stresses Journal of Geotechnical and Geoenvironmental
developed was more for the varying overburden situation when Engineering, 143(5), https://doi.org/10.1061/(ASCE)
compared with uniform overburden and can be quantified to an
GT.1943-5606.0001638.
increase of 88% for the case considered. The results of dynamic
analysis indicate that the twin tunnels located in varying overburden Boldini,D., Amorosi, A., and Misano, F.P. (2010), “Analysis of
terrain showed an increase of 7.3% in the stress values when Tunnel Behaviour Under Seismic Loads by Means of
compared with the tunnels located below uniform overburden. Simple and Advanced Numerical Approaches”,
When the tunnel models are subjected to dynamic loading with 2D International Conferences on Recent Advances in
and 4D spacing, concrete showed significant increase in stresses Geotechnical Earthquake Engineering and Soil Dynamics,
compared to the other two models. (6), pp.1-12.
Bazaz, J. B, and Besharat, V. (2008), “An Investigation on Seismic
6. REFERENCES Analysis of Shallow Tunnels in Soil Medium”, 14th
World Conference on Earthquake Engineering, 50, pp.
Anastasopoulos, I., Gerolymos, N., Drosos, V., Kourkoulis, R.,
368-382.
Georgarakos, T. and Gazetas, G. (2007), “Nonlinear
Response of Deep Immersed Tunnel to Strong Seismic Navarro, C. (1992), “Seismic analysis of underground structures”,
Shaking”, Journal of Geotechnical and Geoenvironmental Earthquake Engineering Tenth World Conference
Engineering, 133(9), pp.1067-1090. Balkerna, Rotterdam, 60, pp. 480-496.
An, J., Tuan, C.Y., Cheeseman, B.A., and Gazonas, G.A. (2011), Shrikhande, M., Basu, S., Kumar, A., Das, J., Thakkar, S.K. and
“Simulation of Soil Behaviour under Blast Loading”, Chandra,B. (2001), “Analysis of Strong Motion Data of
International Journal of Geomechanics, 142(9), pp.1053- Bhuj Earthquake of January 26”, Journal of applied
1064. Geophysics, pp.324-338.
Anitha Kumari, S. D., Vipin, K. S., and Sitharam, T. G. (2014), Tiwari, R., Chakraborty, T., and Matsagar, V. (2016), “Dynamic
“Effect of shape of twin tunnels during seismic loading”, Analysis of Underground Tunnels subjected to Internal
Journal of Rock Mechanics and Tunnelling Technology, Blast Loading”, 11th World Congress on Computational
20(1), pp 49-59. Mechanics, pp.1-12.
237
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: In this research, the authors carry out a numerical study through three-dimensional finite element method to investigate the
behaviours of pile foundations subjected to cyclic horizontal loading. Three-pile pile foundation models (with or without batter piles) and
six-pile pile foundation models (with or without batter piles), which were used in the experiments by the authors, are considered in the
numerical analyses. The foundations work as pile group foundations if the raft base is not in contact with the ground surface, while they are
piled rafts if the raft base is in contact with the ground surface. The foundations are modelled as linear elastic. Interface elements are
employed to simulate the slippage between the foundations and the soil. In this study, the hypoplastic model, an incrementally non-linear
constitutive model, is used to model the ground. The analysed results indicate that the piled rafts have higher horizontal resistance than the
corresponding pile groups and the horizontal resistance of the foundations are improved by inclusion of batter piles. The results also show
that the resistance of six-pile pile foundations is not equal two times the resistance of the corresponding three-pile pile foundations due to the
influence of interaction.
Keywords: Piled raft, pile group, numerical analysis, cyclic loading
1. INTRODUCTION
40 80 80 40 40 80 80 40
Pile foundations including pile group and piled raft foundations are
usually applied to structures subjected to large horizontal load, such
40 40
40 40
as bridges, wind-turbine towers or offshore structures. Horizontal 1 2 3 1 2 3
loads acting on these structures are wind load and/or water wave
load and can be considered as cyclic load.
Piled raft has been increasingly used as an effective foundation 240 240
type to reduce average and/or differential settlement, e.g. Poulos et
30
30
al. (2011), and Yamashita et al. (2011). Experimental studies as well
as numerical analyses on piled raft foundations having vertical piles 80 80 80 80
alone have been conducted, e.g. Randolph (1994), Horikoshi et al.
(2003), Matsumoto et al. (2004), Reul (2004), Unsever et al. (2014),
1.1 1.1
255
40
80
40
30
238
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
30
30
50
50
load was measured by 2 load cells (LC-R and LC-L) arranged in the
level 1 right (positive) direction and in the left (negative) direction. Both
the horizontal and vertical displacements of the foundations were
35 40 40 40 40 40
level 2 measured by horizontal and vertical dial gauges.
120
level 3
285
VDG-L VDG-R
level 4
80
level 5 Winch
Winch
LC-R
level 6 LC-L
35
HDG
20 (in mm) 20
(a) P1, P2 and P3 (b) P4, P5 and P6 Figure 3 Experiment setup with measuring instruments
Figure 2 Model piles with strain gauge instrumentation
3. FEM MODELLING
The rafts were made of duralumin with the dimensions as shown
in Figure 1 and can be regarded as rigid. The sand particles were 3.1 FEM simulations of the triaxial tests
adhered on the raft base surface to increase the friction between the A series of triaxial CD tests of the sand having a relative density Dr
raft and the ground during horizontal loading. = 82% were conducted under different confining pressures, p0 = 7,
Table 1 Geometrical and mechanical properties of the model pile 17, 27 and 50 kPa, in order to obtain the mechanical properties and
and the raft to investigate the behaviour of the sand. To select an appropriate soil
model and to evaluate the soil parameters, simulations of the triaxial
Property Value tests were carried out, prior to the analyses of the load tests. The
Outer diameter, D (mm) 20.00 experimental and simulation results of the triaxial tests are shown in
Wall thickness, t (mm) 1.1 Figure 4 (Vu et al., 2018).
Length from raft base, L (mm) 255 250
Cross section area, A (mm2) 65.31
Deviatoric stress, q (kPa)
Solid: experiment
Moment of Inertia, I (mm4) 2926.2 Open: simulation
200
Young’s modulus of the pile, Ep (N/mm2) 70267 p0= 50 kPa
Young’s modulus of the raft, Er (N/mm2) 68670
Poison’s ratio, v 0.31 150
p0= 27 kPa
2.2 Model ground 100 p0= 17 kPa
The soil used for model ground in this study is a dry silica sand
50 p0= 7 kPa
having the properties shown in Table 2. The model ground with a
relative density, Dr, of about 82% (d = 1.533 t/m3) was prepared in
a soil box having dimensions of 800 mm in length, 500 mm in width 0
0 2 4 6 8 10 12 14 16 18 20
and 530 mm in depth. In order to control the density of the model
Axial strain, a (%)
ground, the model ground was prepared by 11 layers (10 layers of
50 mm and 1 layer of 30 mm). In each layer, the sand was poured (a) Deviatoric stress q versus axial strain a
and compacted by tapping so that the target relative density of 82% -12
Solid: experiment
was achieved.
Volumetric strain, vol (%)
239
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
r 0.5 200
0.5 0
pt (kN/m2) 3 -200
e 0.739 -400
-600
3.2 FEM modelling of loading tests -800
Numerical analyses were carried out using a three-dimensional FEM -1000
-1.2 -0.8 -0.4 0.0 0.4 0.8 1.2
program, PLAXIS 3D. Only a half of the foundation and the ground
was modelled owing to symmetric conditions. Figure 5 shows an Normalised horizontal displacement, u/D
example of the finite element mesh of the modelling. (6a) Experimental results
1000
3PG FEM
800
3BPG
Horizontal load, H (N)
600 3PR
800 250
400 3BPR
200
530
(mm) 0
-200
-400
-600
-800
Figure 5 Finite element mesh -1000
-0.4 -0.2 0.0 0.2 0.4
Boundary conditions are applied as follows: Normalised horizontal displacement, u/D
- Vertical model boundaries with their normal in x-direction (i.e. (6b) FEM results
parallel to yz-plane) are fixed in x-direction (ux = 0) and free in y- Figure 6 Horizontal load vs. normalised horizontal displacement for
and z-directions. 3-pile pile foundations
- Vertical model boundaries with their normal in y-direction (i.e.
parallel to xz-plane) are fixed in y-direction (uy = 0) and free in x- Similar results are also obtained in the cases of the 6-pile pile
and z- directions. foundations, in which the piled rafts have much higher horizontal
- The model bottom boundary is fixed in all directions (ux = uy = resistances than the corresponding pile groups and the resistances of
uz= 0). the foundations are enhanced by inclusion of batter piles, as shown
- The ground surface is free in all directions. in Figure 7. It is seen from the above results that the FEM
calculations simulate the experimental results very well.
The raft and the piles were considered as linear elastic materials.
2000
In order to model the pile, a hybrid model in which beam element 6PG Exp.
surrounded by solid elements was used, according to Kimura and 1600 6BPG
Horizontal load, H (N)
Zhang (2000). The raft was modelled by using solid elements. 1200 6PR
Interface elements of Mohr-Coulomb type were assigned at the 800 6BPR
raft base (in the cases of the piled rafts) and along the pile shafts. 400
Interface cohesion was set as 0, and the interface friction angle was 0
set as 40.2˚ following Unsever et al. (2015). The analysis procedure
is as follows: -400
Step 1: Self-weight analysis of the model ground alone, where -800
K0= 1-sin ( is internal friction angle of the soil) was -1200
assumed. -1600
Step 2: Setting the foundation in the ground, and self-weight -2000
analysis including the foundation. -1.2 -0.8 -0.4 0.0 0.4 0.8 1.2
Step 3: Analysis of loading process using displacement control Normalised horizontal displacement, u/D
manner. (7a) Experimental results
240
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
2000 7
1200 6PR 4
800 6BPR 3
2 6BPG
400 1
0 0
-400 -1
-2
-800 -3
-1200 -4
-5
-1600 -6
-2000 -7
-0.4 -0.2 0.0 0.2 0.4 -1.2 -0.8 -0.4 0.0 0.4 0.8 1.2
Normalised horizontal displacement, u/D Normalised horizontal displacement, u/D
(7b) FEM results (10a) by experiments
Figure 7 Horizontal load-normalised horizontal displacement for 6-
pile pile foundations 1.2
1200
6PG Figure 10 Inclination of the raft of 6PG and 6BPG during cyclic
6PR horizontal load the experiments and FEM
1000
6PR
800 7
2 x 3PG
Inclination of the raft, (deg.)
600 6
Positive (+) 6PR
6PG 5
400 4
200
3
2
0 1 6BPR
0.00 0.05 0.10 0.15 0.20 0.25 0
Normalised horizontal displacement, u/D -1
Figure 8 Horizontal load vs. normalised horizontal disp. curves for -2
-3
6PG, 6PR, 2×3PG and 2×3PR (Experimental results) -4
-5
1600
-6
FEM -7
1400 2 x 3PG 2 x 3PR -1.2 -0.8 -0.4 0.0 0.4 0.8 1.2
Horizontal load, H (N)
1200 2 x 3PR
6PG Normalised horizontal displacement, u/D
1000 6PR 2 x 3PG (11a) by experiments
800 6PR
1.2
Inclination of the raft, (deg.)
600 6PR
6PG 1.0 FEM
6BPR
400 0.8
200
0.6
0.4
0 0.2
0.00 0.05 0.10 0.15 0.20 0.25
Normalised horizontal displacement, u/D 0.0
-0.2
Figure 9 Horizontal load vs. normalised horizontal disp. curves for -0.4
6PG, 6PR, 2×3PG and 2×3PR (FEM results) -0.6
-0.8
Figure 10 shows comparisons of inclination of the raft during -1.0
cyclic horizontal loading between 6PG and 6BPG by the -1.2
experiments (Figure 10a) and FEM (Figure 10b). Figure 11 shows -0.4 -0.2 0.0 0.2 0.4
the corresponding results of 6PR and 6BPR. It is indicated from Normalised horizontal displacement, u/D
both experimental and FEM results that the inclination of raft (11b) by FEM
increases almost linearly with the increase of normalised horizontal
displacement in all cases, and the inclination is suppressed by Figure 11 Inclination of the raft of 6PR and 6BPR during cyclic
inclusion of the batter piles. horizontal load by the experiments and FEM
241
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
400 5
0
0 Level 1
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 -5
Level 2
Inclination of the raft, (deg.) Level 3
-10 Level 4
Figure 12 Inclination of the raft vs. horizontal load during the initial Level 5
loading stage for 6-pile foundations (Experimental results) -15
Level 6
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
1600
Normalised horizontal displacement, u/D
6PG FEM 6BPR
(b)
Horizontal load, H (N)
6BPG
1200 6PR
6BPR 6PR 15
Level 1
Bending moment, M (N.m)
Level 2
6BPG 10 Level 3
800
Level 4
5 Level 5
6PG
Level 6
400 0
-5
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 FEM
-10
Inclination of the raft, (deg.)
P3 in 3PR
Figure 13 Inclination of the raft vs. horizontal load during the initial -15
loading stage for 6-pile foundations (FEM results) -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
Normalised horizontal displacement, u/D
Figures 14 and 15 show the numerical results for 3PR and (c)
3BPR, respectively, in which changes of bending moments with
normalised horizontal displacement, u/D, at different levels (see Figure 14 Bending moments of piles for 3PR (FEM)
Figure 2) of each pile during horizontal loading are given. Note that
P3 is the front pile and P1 is the rear pile for positive loading, and 15
vice versa for negative loading. These numerical results are in good FEM
Bending moment, M (N.m)
242
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
15 35
FEM 30 Level 1
15
35
Bending moment, M (N.m)
25 P2 in 3BPR
35 20
30
Bending moment, M (N.m)
P1 in 3PR 15
25 10
20 5
15 0
10 -5
5 -10 Level 1
0 -15 Level 2
-5 Level 1 -20 Level 4
-10 Level 2 -25 Level 5
-15 Level 3 -30 Level 6
-20 Level 4 -35
-25 Level 5
-1.2 -0.8 -0.4 0.0 0.4 0.8 1.2
-30 Level 6 Normalised horizontal displacement, u/D
-35
-1.2 -0.8 -0.4 0.0 0.4 0.8 1.2 (b)
Normalised horizontal displacement, u/D
(a) 35
30 Level 1
Bending moment, M (N.m)
25 Level 2
35 20 Level 3
30 15 Level 4
Bending moment, M (N.m)
25 P2 in 3PR 10 Level 5
20 5 Level 6
15 0
10 -5
5 -10
0 -15
-5 -20
Level 1 -25
-10 Level 2 P3 in 3BPR
-15 -30
Level 3 -35
-20 Level 4 -1.2 -0.8 -0.4 0.0 0.4 0.8 1.2
-25 Level 5
-30 Level 6 Normalised horizontal displacement, u/D
-35
-1.2 -0.8 -0.4 0.0 0.4 0.8 1.2 (c)
Normalised horizontal displacement, u/D Figure 17 Bending moments of piles for 3BPR (Experiment)
(b)
243
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5. CONCLUSION
A numerical study through three-dimensional finite element method
was carried out to investigate the behaviours of pile foundations
subjected to cyclic horizontal loading and the results were compared
with the corresponding experimental study.
The analysed results indicate that the piled rafts have higher
horizontal resistance than the corresponding pile groups and the
horizontal resistance of the foundations are improved by inclusion
of batter piles. It is worth to notice that the piled raft with batter
piles is the most favourable foundation type to minimize the
inclination induced by horizontal loading.
The results also show that the resistance of six-pile pile
foundations is not equal two times the resistance of the
corresponding three-pile pile foundations due to the influence of
interaction.
6. REFERENCES
Hamada, J., Tsuchiya, T., Tanikawa, T., and Yamashita, K. (2015)
“Lateral loading tests on piled rafts and simplified method to
evaluate sectional forces of piles”. Geotechnical Engineering
Journal SEAGS & AGSSEA, 46(2), pp29-42.
Horikoshi, K., Matsumoto, T., Hashizume, Y., Watanabe, T., and
Fukuyama, H. (2003) “Performance of piled raft foundations
subjected to static horizontal loads”. Int. Jour. of Physical
Modelling in Geotechnics, 3(2), pp37-50.
Kimura, M., and Zhang F. (2000) “Seismic evaluations of pile
foundations with three different methods based on three-
dimensional elasto-plastic finite element analysis”. Soil and
Foundation, 40(5), pp113-132.
Matsumoto, T., Fukumura, K., Pastsakorn, K., Horikoshi, K., and
Oki, A. (2004) “Experimental and analytical study on behaviour
of model piled rafts in sand subjected to horizontal and moment
loading”. International Journal of Physical Modelling in
Geotechnics, 4(3), pp1-19.
Poulos, H.G., and Small, J.C., Chow, H. (2011) “Piled raft
foundations for tall buildings”. Geotechnical Engineering
Journal SEAGS & AGSSEA, 46(2), pp78-84.
Randolph, M. F. (1994) “Design methods for pile groups and piled
rafts”. Proc. 13th ICSMGE, Vol. 5, New Delhi, pp. 61-82.
Reul, O. (2004) “Numerical study of the bearing behaviour of piled
rafts”. International Journal of Geomechanics, 4(2), pp59-68.
Unsever, Y.S., Matsumoto. T., Shimono, S., and Ozkan, M.Y.
(2014) “Static cyclic load tests on model foundations in dry
sand”. Geotechnical Engineering Journal SEAGS & AGSSEA,
45(2), pp40-51.
Unsever, Y.S., Matsumoto, T., and Ozkan, M.Y. (2015) “Numerical
analyses of load tests on model foundations in dry sand”.
Computers and Geotechnics, 63, pp40-51.
Vu Anh-Tuan, Matsumoto, T., Kobayashi, S., and Shimono, S.
(2017) "Experimental study on pile foundations having batter
piles subjected to combination of vertical and horizontal loading
at 1-g field", Geotechnical Engineering Journal of the SEAGS &
AGSSEA, Vol. 48(3), pp12-24.
Vu Anh-Tuan, Matsumoto, T., Kobayashi, S., and Nguyen Tuong-
Lai (2018) "Model load tests on battered pile foundations and
finite-element analysis", Int. Journal of Physical Modelling in
Geotechnics, 18(1), pp33-54.
Wolffersdorff, P.A. (1996) “A hypoplastic relation for granular
materials with a predefined limit state surface”, Mechanics of
cohesive-frictional materials, 1, pp251-271.
Yamashita, K., Yamada, T., and Hamada, J. (2011) “Investigation of
settlement and load sharing on piled rafts by monitoring full-
scale structures”. Soils and Foundations, 51(3), pp513-532.
244
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This paper introduces the Finite Difference formulas derived from the thin-plate theory on raft foundation with the influences
from boundary conditions. The proposed solution was established for a raft located at the ground surface, where the bending moments and
shears along the edges of the foundation were presumed vanished. The solutions were examined with the three-dimensional Finite Element
analysis on numerical foundation subjected to uniformly load. Computer program WERAFT-S was developed for raft foundation on elastic
soils under vertically static load. Effectiveness of a couple of soil spring models was discussed with the validation of WERAFT-S to three
dimensional Finite Element solutions.
Keywords: Finite difference analysis, raft foundation, plate theory, boundary conditions, soil springs
245
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3. VALIDATION
Validation of the proposed analysis is presented next. Assuming that
a square raft has width (W) of 26 m and thickness (D) of 1 m.
Figure 2 Allocations of the corresponding equations of the proposed Young’s modulus (E), Poisson’s ratio (), and unit weight () of the
analysis concrete raft foundation are 3×104 MPa, 0.15, and 24 kN/m3,
This assumption matches closely with the flexible foundations respectively The soils underneath the foundation are assumed to
where the raft size is relatively large. For smaller raft foundation have a shear wave velocity (Vs) of 150 m/s, Poisson’s ratio (s) of
that behaves more rigidly, this assumption should be not applicable. 0.4, and unit weight (s) of 19 kN/m3. Note that Vs2×s/g=Gs, where
One can adopt other types of soil spring model. For example, the g is acceleration of the gravity, Gs is Shear modulus of the soil. Note
Lysmer Analog soil spring can also be used. The total soil spring that Es=2Gs(1+s), therefore Es can be calculated. The uniform load
constant of the whole raft foundation could be calculated as applied on top of the foundation is 100 kPa. Three dimensional FEA
4Gsro/(1s), where Gs is the shear modulus of the soil, s is the solutions of the numerical model were first obtained from Midas-
Poisson’s ratio of the soil, and ro is the equivalent radius of the raft GTS (Midas 2017) varying the dimensions of the FE zone. Essential
foundation. boundaries (i.e., roller and hinge) were adopted in the modeling.
246
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Eight-node brick elements were mainly used to generate the 4. EFFECTS OF SOIL SPRINGS
structural system.
Fig. 5 shows that the foundation settlements obtained from FE The influences of effective thickness of the soil spring were studied.
analysis will gradually converge by increasing the FE zone at It was found that the shear wave velocity of the soil is the most
60~200 m (note that the length, width and thickness of the zone important parameter if the compressional spring was adopted. Fig. 7
were kept the same for the analytical zone). The settlements became depicted the results. More details of the discussions can be found in
stable when the analytical zone was increased to 200 m x 200 m x Chang et al. (2018). Fig. 8 shows the matching of the displacements
200 m. Therefore the FEA of the study was conducted using the at centre of the foundation with the FEM ones on numerical model
optimized core. It can be seen that the foundation settlements are varying the shear wave velocity of soil. Corresponding
different at various locations. In general, the center exerts the largest interpretation of the appropriate thickness of the soil springs with
settlement whereas the corner has the smallest settlement. The soil respect to shear wave velocity of the soil can be obtained. Note that
mass involved in the FE analysis seems to have significant influence when Vs is at 120~180 m/s, the effective thickness of the soil spring
on foundation settlements. Validation of the WERAFT-S analysis will be around 32~17 m. The effective thickness of the soil spring
was then conducted to compare with the FE solutions. The seems to be much smaller than what was expected (for square
influences of the area ratio (Ask/Ark=n) of the edge node, the shear foundation, effective depth of the foundation is around 2B). The
reason is because that the force acting along the spring is the same
wave velocity (Vs) and the Poisson’s ratio (s) of soil, the thickness
while the real stress magnitudes of the foundation will gradually
of raft (D), and the load intensity (q) were studied to see the
decrease with the depth.
appropriateness of the length (l) of soil spring.
247
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Comparisons of the displacement contours of the standard Chang, D.W., Lien, H.W. and Wang Tzuyu (2018) “Finite
model foundation from both WERAFT-S and FEM can be found in difference analysis of vertically loaded raft foundation based
Fig. 9. To show the possible influences of the soil springs using on the plate theory with boundary concern,” J. of
Lysmer Analog model, the results were listed in Table 1. It can be GeoEngineering (under review)
found that the deviations between the WERAFT-S and FEM are Chen, C.N. (1998) Solution of beam on elastic foundation by
smaller when finding the optimized solutions of simplified soils DQEM, Journal of Engineering Mechanics, ASCE, 124(12),
springs. The predictions using Lysmer Analog springs need to pp1381-1384.
further examinations.
Chiou, J.S., Lin, H.S., Yeh, F.Y., and Sung, Y.C. (2016) “Plastic
Table 1 Comparisons on Foundation Displacements Obtained from settlement evaluation of embedded railroads under repeated
FEA and WERAFT-S using Different Soil Spring Models train loading”, Journal of GeoEngineering, Taiwan
Shear w/ Lysmer Geotechnical Society, 11(2) pp97-107.
w/ EA/l Clancy, P. and Randolph, M.F. (1993) ″Simple design tests for piled
wave Analysis Midas Analog
model raft foundations.″ Geotechnique, 36(2), pp169-203.
velocity model
(m/s) location foundation displacement (mm) Dinev, D. (2012) “Analytical solution of beam on elastic Foundation
by singularity functions”, Engineering Mechanics, 19(6)
center 42.3 40.3 28.7 pp381-392.
120 edge 30.0 31.8 21.7 Gupta, S.C. (1997) Raft Foundation Design and Analysis with a
corner 22.5 22.7 13.2 Practical Approach, New Age International Publisher.
center 18.8 18.6 18.6 Hemsley, J.A. (1998) Elastic Analysis of Raft Foundations, ICE
150 edge 13.1 13.7 13.6 pub.
corner 9.7 9.2 7.9 Horikoshi, K. and Randolph, M.F. (1996) ″Estimation of overall
center 10.6 11.1 13.0 settlement of piled rafts.″ Soils and Foundations, 39(2), pp59-
180 edge 7.5 7.9 9.2 68.
corner 5.6 5.1 5.2 Itasca (2017) FLAC Version 8.0, Itasca Consulting Group. Inc.
Jones, M. (1997) Analysis of Beams on Elastic Foundations: Using
Finite Difference Theory, ICE pub.
5. CONCLUDING REMARKS
Kitiyodom, P. and Matsumoto, T. (2002) ″A simplified analysis
This paper presents a newly proposed finite difference analysis method for piled raft and pile group foundations with batter
WERAFT-S for the surface raft foundation under vertical loads piles.″ Int. Journal for Numerical and Analytical Methods in
using the Thin-Plate Theory. The effects of boundary conditions Geomechanics, 26, pp1349-1369.
were considered to derive the formulas of the raft foundation Kitiyodom, P., Matsumoto, T. and Kawaguchi, K. (2005) ″A
settlements. Elastic compressional soil springs were used to simulate simplified analysis method for piled raft foundations subjected
the soil resistances underneath the raft. Solutions of the proposed to ground movements induced by tunneling.″ Int. Journal for
analysis were verified on a numerical raft foundation located at the Numerical and Analytical Methods in Geomechanics, 29,
surface of a ground consisting of elastic soils (where Vs=150 m/s pp1485-1507.
and s=0.4). Vertical uniform load of 100 kPa was mounted on the Kobayashi, H., Nishio, H., Nagao, T. Watanabe, T., Horikoshi, K.,
foundation. Validation of the analysis was made with those from Matsumoto, T. (2009) ″Design and construction practices of
three-dimensional finite element analysis (FEA) using Midas-GTS. piled raft foundations in Japan.″ Proc., Int. Conf. on Deep
The effective thickness of the compressible soil springs underneath Foundations - CPRF and Energy Piles, pp101-135.
the raft was studied. It was found that the effective thickness of the Kukreti, A.R. and Ko, M.G. (1992) “Analysis of rectangular plate
soil at 17~32 m (for Vs at 180~120 m/s) can provide agreeable resting on an elastic half space using an energy approach”,
solutions with the FEA. Although the compressible soil spring Applied Mathematical Modeling, 16(7), pp338-356.
model has some defects in the real applications, it has been shown Lien, H.W. (2018) Finite Difference Analysis of Piled Raft
that the adequate solutions can be still achieved. On the other hand, Foundations under Vertically Loads, Master Thesis, Dept. of
the Lysmer Analog springs seem to provide much smaller Civil Engineering, Tamkang University, Taiwan (in Chinese).
foundation displacements on stiffer site. The results need further Mathews, P.M. (1958) “Vibrations of a beam on elastic foundation”,
studies with the concerns of uneven soil stiffness distributed Journal of Applied Mathematics and Mechanics, 38(3-4),
underneath the foundation. If the raft foundation is rigid enough pp105-115.
where the boundary effects can be neglected, then the solution Midas (2017) Midas GTS NX User Manual, Midas IT Co.
proposed by Bowles (1977) can be used assuming uniform Omer, J.R. and Arbabi, A. (2015) “Evaluation of finite element,
settlement. Vertical loads such as uniform, non-uniform, and point finite difference and elasticity methods for hypothetical raft
loads need to be taken into account with pre-calculations. foundations installed on layered strata”, Geotechnical and
Geological Engineering, 33(4), pp1129-1140.
6. ACKNOWLEDGEMENT Poulos, H.G. (1991) “Analysis of piled strip foundation”, Computer
Methods and Advances in Geomechanics, ed. Beer et al.,
The paper presents the partial results of the research grant MOST Rotterdam, 1, pp183-191.
106-2221-E-032-025-MY2. Finance support is greatly appreciated. Poulos, H.G. (1994) “An approximate numerical analysis of pile-raft
7. REFERENCES Interactions”, Int. J. for Numerical and Analytical Methods in
Geomechanics, 18, pp73-92.
Biot, M.A (1937) “Bending of an infinite beam on an elastic Randolph, M.F. (1983) “Design of Piled Foundations”, Cambridge
foundation”, Journal of Applied Physics, 12, pp155-164. Univ. Eng. Dept., Res. Rep. Soils TR143.
Bowles, J.E. (1977) Foundation Analysis and Design, 2nd ed., Timoshenko, S. and Woinowsky-Krieger, S. (1959) Theory of Plates
McGraw-Hill Companies, Inc. and Shells, 2nd ed., McGraw-Hill. New York.
Chang, D.W., Lee, M.R., Hong, M.Y. and Wang, Y.C. (2016) “A Ting, B.Y. and Mockry, E.F. (1984) “Beam on elastic foundation
simplified modeling for seismic responses of rectangular finite element”, Journal of Structural Engineering, ASCE,
foundation on piles subjected to horizontal earthquakes”, J. of 110(10), pp2324-2339.
GeoEngineering, Taiwan Geotechnical Society, 11(3) pp109- Tomlnison, M.J. and Boorman, R. (2001) Foundation Design and
121. Construction, 7th ed. Addison-Wesley Longman Ltd.
248
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Vlasov, V.Z. and Leontev, U.N. (1966) Beams, Plates and Shells on
Elastic Foundation. (Translated from Russian), Israel Program
for Scientific Translation Jerusllem, Israel.
249
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
250
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: The stability of slope is known to be affected by rainfall infiltration, especially when groundwater table is relatively deep. Thus,
careful analysis of slope stability requires the determination of unsaturated properties of the soil. Laboratory tests for the determination of
hydraulic properties and unsaturated shear strength are tedious and time consuming. Thus, numerous fitting equations were proposed by
researchers for determination of soil water characteristic curve (SWCC), permeability function and shear strength parameter of the unsaturated
soil. This paper presents derivation of unsaturated soil properties from simple laboratory tests using relevant parameter estimation equations.
The laboratory data include the suction measured using filter paper, and the shear strength parameters obtained from Triaxial UU test on
compacted clayey soil under different gravimetric water contents. Results show that hydraulic properties and unsaturated shear strength
parameters could be derived from available data and the slope stability analysis could be carried out using the estimated properties.
Keywords: Compacted soil; Suction; SWCC; permeability function; rate of strength increase with respect to suction.
251
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
saturation, e is the natural logarithmic base number, 2.71828, a, n, m
are fitting parameters related to the air-entry value of the soil in kPa By extending the classical Mohr-Coulomb envelope, the three-
(a), the slope at the inflection point of the SWCC; and residual water dimensional representation of Equation 4 can be visualized in Figure
content (r) respectively; = (ua–uw) is the matric suction (kPa); and 2.
C() is a correction factor given as follows:
�
ln(1+ � )
�(�) = 1 − �
(2)
ln (1+1000000
� )
�
Unsaturated shear strength from direct shear tests (Gan et al., 1988)
indicated that the relationship between the shear strength versus and
suction is not linear. However, Rahardjo et al. (1997) suggested it is
possible that the relationship over a selected range of suction is linear.
This is applicable for results of Undrained Unconsolidated (UU)
Triaxial test performed on different water contents (Fredlund et al.,
1978). As the undrained compressive strength (qu) obtained from this
test can be defined as half of the maximum deviator stress (i.e.
qu=(1–3)/2).
Many shear strength equations were developed to predict the rate
of strength increase with respect to suction (b) (Goh et al’ 2010;
Abdullah et al. 2013), among them is the equations proposed by
Vanapalli et al (1997) is as follows:
f c' ( n uw ) f tan ' (ua uw ) K tan ' (7)
Figure 1 Typical soil water characteristic curve showing zones of
desaturation (Fredlund et al, 2012).
where is normalized volumetric water content (w/s) and is an
unknown fitting parameter. This equation is a fitting type of equation
Permeability function can be plotted based on saturated
developed for sand and fine-grained soil. For sand, is equal to one,
coefficient of permeability (ks) and a few data on water coefficient of
but the value increases with soil’s plasticity. The equation was found
permeability (kw) obtained from triaxial test with fitting parameters of
to fit results of UU test by (Gofar & Lee, 2014) for both coarse and
the SWCC curve (a, n, and m). In the absence of data on the water
fine-grained soil for suction less than 100 kPa.
coefficient of permeability, statistical method proposed by Childs &
Goh et al (2010) equation is basically the improvement of
Collis-George (1950) and presented in Fredlund & Rahardjo (1993)
Equation (7). Based on their study on eleven sets of published data as
could be used. Alternatively, equation (3) which took the same form
well as laboratory tests, they found that the relationship between shear
as Equation (1) could be used to predict the permeability function
strength and suction changed at suction equal to air entry value AEV
(Leong and Rahardjo, 1997b).
= (ua–uw)b. therefore:
ks
k w C ( ) f c' ( n uw ) f tan ' (ua uw ) tan b (8)
n mp (3)
if (ua–uw)< (ua-uw)b or <AEV
ln e
a f c' [( n u w ) ua u w b ] tan '
(9)
with p value ranges from 2.4 to 25 depending on the type of soil. The [(ua u w ) (ua u w )b ]b K tan '
estimation of p value could be made based on the similarity of the
if (ua–uw)≥ (ua-uw)b or ≥ AEV
permeability function with the SWCC.
where
A practical formulation for shear strength of unsaturated soil was
proposed by Fredlund et al. (1978) as follows: [log (un uw ) log( ua uw ) b] y (10)
f c' ( n u w ) f tan ' (u a u w ) f tan b (4) where b and y are fitting parameters.
This equation was validated with published laboratory data which
where c’ is the effective cohesion; ’ is the effective angle of shearing mostly fine grained soil (classified as CH, CL, MH, ML and SM
resistance; (n–uw)f is the effective normal stress at failure; b is the according to USCS).
rate of change in shear strength relative to change in suction. For ua
equal to atmospheric, Equation (4) can be simplified as:
252
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3. METHODOLOGY
The study presented in this paper made use of the available data in
Arif (2005) for the evaluation of unsaturated soil properties required
for slope stability analysis. The soil for his study was retrieved from
Babarsari, Yogyakarta. The data were originally used for evaluation
of bearing capacity of the soil. Index properties of the soil comprising
grain size distribution, Atterberg limits and falling head test for
determining the coefficient of saturated permeability of the soil were
tested in laboratory. Compaction test was carried out to define the
maximum dry density and the optimum water content of the soil.
Advanced laboratory tests were performed on the compacted soil
including suction measurements and shear strength tests.
The SWCC was obtained experimentally by measuring suction of
the soil samples using Filter Paper method following the procedure
outlined in ASTM D 5298. The calibration of the Whatman filter
paper No. 42 follows the following equation: Figure 3 Grain size distribution of the soil evaluated in this study
�= 105.327−0.0779 �� for wf < 45.26 (11a)
3.412−0.0135 ��
Filter paper test was performed on two sets of samples. Each set
�= 10 for wf ≥ 45.26 (11b) comprises four (4) samples were prepared at gravimetric water of
where is the suction (kPa) and wf is the water content of the filter about 22, 19, 16, and 14% (denoted as sample A, B, C and D). In other
paper. However, direct correlation between the gravimetric water words, two samples were prepared below the optimum water content
content of sample with suction should be derived empirically based while the other two were prepared above the optimum water content.
on a number of tests performed in the study. The corresponding suction for the samples are calculated based on
The shear strength parameters were obtained using conventional empirical correlation developed in this study is as follows:
Triaxial testing apparatus under Unconsolidated Undrained (UU)
condition. Frelund & Rahardjo (1993) suggested that the procedures �= 107 × �0.6519 �� (12)
outlined for conventional Triaxial test on saturated soil can be applied The corresponding suctions for the gravimetric water content for soil
for undrained loading condition of unsaturated soil. In this case, the preparation is shown in Table 1.
suction stress can be linearly correlated with the undrained strength
of soil (Chae et al., 2010). The unsaturated soil specimen is tested at Table 1. Suction corresponding to the water content of soil at
its initial water content or suction. In this case, the initial matric preparation
suction is not relaxed or changed prior to commencing the test by
Sample Gravimetric Volumetric Suction
replacing the porous stone with metal or plastic discs on top and
water water (kPa)
bottom of specimen and by closing the specimen with rubber
content content
membrane during the test. It is assumed that the variation is negligible
A 22 33.26 5.9
despite the suction may decrease due to the increase of pore-water
pressure during shearing stage of the soil sample. Soil specimens (38 B 19 28.73 41.8
mm dia. and 76 mm height) were prepared at the same initial C 16 24.18 295.2
gravimetric water content as for the suction measurement tests. The D 14 21.17 1087.3
shearing rate was controlled at 1.50 mm/min or equivalent to a strain
rate of 0.033% per second for the specimen height of 76 mm. The resulting points from filter paper test were plotted in Figure 4 in
Fredlund & Xing (1994) equation was used to develop SWCC terms of volumetric water content. The laboratory data points were
curve based on the suction measurement data using filter paper test. fitted using Fredlund & Xing (1994) equation to form the SWCC as
The permeability function was plotted using the coefficient of shown in Figure 4. The fitting parameters of the equation a, n, m are
saturated permeability data and the SWCC curve. Subsequently, the 23 kPa, 0.8 and 0.5 respectively. The air entry value (ψa) and residual
rate of shear strength in-crease with respect to suction (b) was suction (ψr) are 8 and 3700 kPa respectively. The saturated volumetric
obtained by applying non-linear failure envelope equation (Vanapalli water content (θs) and the residual volumetric water content (θr) were
et al., 1996 and Goh et al, 2010) on the experimental test results. The 0.37 and 0.15 respectively while the slope of SWCC curve m2w was
performance of both predicted types equation was evaluated for this 0.083. The SWCC was the typical of drying SWCC of silty soil
soil by comparing with the results of Triaxial UU test. (Fredlund and Rahardjo, 1993). The permeability function was
calculated based on ks and the SWCC using statistical method (Childs
4. RESULTS AND ANALYSIS & Collis-George, 1950) and Leong and Rahardjo’s fitting equation is
Following ASTM D2487, the soil used in this study was classified as shown in Figure 5.
low plasticity silt (ML). Specific gravity of the soil was 2.67. The
grain size distribution of the soil is presented in Figure 3. The soil
contains more than 60% fines with 53.5% silt and 10.1% clay
fractions. Liquid limit was 47.8% while plastic limit was 29.2%, thus
the plasticity index of the soil was 11.6. The coefficient of saturated
permeability of the soil is quite low (i.e. 5.6×10-6). Compaction test
results in a maximum dry density of 1.68 Mg/m3 and optimum water
content of 18.4%. Soil samples were prepared at initial relative
density of 90% or dry density of 1.51 Mg/m3.
253
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5. CONCLUSIONS
254
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Ching, R.K.H., Sweeney, D.J., & Fredlund, D.G. (1984) “Increase in Vanapalli S.K., Fredlund D.G., Pufahl D.E, & Clifton A.W. (1996).
factor of safety due to soil suction in two Hong Kong slopes”. Model for the Prediction of Shear Strength with Respect to Soil
Proc. 4th Intl. Symp. Landslides. pp.617-623 Suction. Canadian Geotechnical Journal, (33) pp. 379-392.
Childs E.C., Collis-George G.N. (1950) “The permeability of porous Yunusa, GH., Kassim, A., and Gofar, N. (2014) Effect of Surface
materials”. Proc. Royal Society of London 210A pp. 392-405. Flux Boundary Conditions on Transient Suction Distribution
Fredlund, D.G. & Xing, A. (1994) Equations for the Soil-Water in Homogeneous Slope. Indian Journal of Science and
Characteristic Curve. Canadian Geotechnical Journal, 31 Technology, 7(12): 2064–2075
pp.521-532. Zhai, Q. and Rahardjo, H. 2012. Determination of Soil-Water
Fredlund, D.G., Xing, A. & Huang, S. (1994) Predicting the Characteristic Curve Variables. Computer and Geotechnics,
Permeability Function for Unsaturated Soils Using the Soil- 42: 37-43.
Water Character Curve. Canadian Geotechnical Journal. 31(3)
pp. 533-546
Fredlund, D.G. & Rahardjo, H. (1993). Soil Mechanics for
Unsaturated Soils. New York: John Wiley & Sons, Inc.
Fredlund, D.G., Morgenstern, N.R., & Widger, R.A. (1978) “The
Shear Strength of Unsaturated Soil”. Canadian Geotechical
Journal. 15 pp. 313–321.
Gan, J.K.M., Fredlund, D.G., & Rahardjo, H. (1988) “Determination
of the Shear Strength Parameters of Unsaturated Soil Using
Direct Shear Test”. Canadian Geotech. Journal. 25 pp.500-510
Gasmo, J.M., Rahardjo, H., & Leong, E.C. (2000) “Infiltration effects
on stability of a residual soil slope”. Computers and
Geotechnics, 26 pp. 145–165.
Gofar, N., Lee M.L., & Rahardjo, H. (2015) “Field Monitoring of
Suction Distribution in Coarse-Grained and Fine-Grained
Residual Soil”. Proc. 6th AP-UNSAT2015 pp. 679-683.
Gofar, N. & Lee, ML. (2014) “Unsaturated Shear Strength of Coarse
Grained and Fine Grained Tropical Residual Soil”. Proc. ISGE
Annual Science Meeting PITHATTI XX: 257-262
Gofar N. Lee, M.L. & Kassim, A. (2008) “Response of Suction
Distribution to Rainfall Infiltration in Soil Slope”. Electronic
Journal of Geotechnical Engineering. 13(E) pp.1-13
Goh, S. G., Rahardjo, H., & Leong, E.C. (2010) “Shear Strength
Equations for Unsaturated Soil Under Drying and Wetting”.
ASCE Journal of Geotechnical and Geoenvironmental
Engineering. 136(4) pp.594-606.
Ho, D.Y.F., & Fredlund, D.G. (1982) “A Multi-Stage Triaxial Test
for Unsaturated Soils”. ASTM Geotechnical Testing Journal.
5(1-2)pp.18-25.
Kassim, A., Gofar, N., Lee, M.L., & Rahardjo, H. (2012) “Modeling
of Suction Distribution in an Unsaturated Heterogeneous
Residual Soil Slope”. Engineering Geology 131-132 pp.70-82.
Leong, E.C. & Rahardjo, H. (1997a) “A Review on Soil-Water
Characteristic Curve Equations". ASCE Journal Geotechnical
& Geoenvironmental Eng. 123(12) pp.1106-1117.
Leong, E.C. and Rahardjo, H. (1997b) “Permeability Functions for
Unsaturated Soils”. ASCE Journal Geotechnical &
Geoenvironmental Eng. 123(12) pp.1118-1126.
Likos, W.J., & Lu, N. (2002) “Filter Paper Technique for Measuring
Total Soil Suction”. Transportation Research Record 1786.
Paper No. 02-2140 pp. 120-128
Liu, C. & Evett, J.B. 2005. Soils and Foundation SI Edition by Nurly
Gofar, Prentice Hall.
Rahardjo, H., Leong, E. C., & Rezaur, R. B. (2008) “Effect of
antecedent rainfall on pore-water pressure distribution
characteristics in residual soil slopes under tropical rainfall”.
Hydrological Processes, 22(4) pp.506–523.
Rahardjo, H., Lim, T.T., Chang, M.F., and Fredlund, D.G. 1995.
“Shear Strength Characteristics of a Residual Soil in
Singapore”. Canadian Geotechnical Journal, 32 pp.60 - 77.
Sillers, W.S. and Fredlund, D.G. (2001) “Statistical assessment of
soil-water characteristic curve models for Geotechnical
Engineering”. Canadian Geotechnical Journal. 38(6) pp.1297-
1313
Toll, D.G., Rahardjo, H., & Leong, E.C. (1999) “Landslides in
Singapore”. Proc. 2nd Intl. Conf. Landslides, Slope Stability &
the Safety of Infrastructures. pp.269–276.
Tsaparas, I., Rahardjo, H., Toll, D.G., & Leong, E.C. (2002)
“Controlling parameters for rainfall-induced land-slides”.
Computers and Geotechnics, 29 pp. 1–27.
255
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: There are a number of man-made island in Osaka Bay. In order to maintain and manage the man-made island, we need to
predict consolidation behavior. The Pleistocene clay layer of Osaka Bay has few conservation of ground survey and has characteristic
consolidation characteristics of pseudo overconsolidation. Therefore, it is difficult to predict its settlement behavior only from indoor test
results. Incidentally, in recent years, data assimilation methods have been spreading in various fields. Data assimilation is a method of obtaining
a highly accurate predicted value by correcting a simulation by observation value. In this study, data assimilation was carried out using the
measured value of the pore water pressure for the subsidence amount for the upper Pleistocene clay (Ma12) layer. And we derived Ma12’s
future settlement as a probability distribution.
Keywords: Data assimilation, Particle Filter, Consolidation settlement, Pleistocene clay
e :settlement amount
:pore water pressure
Figure 1 the conceptual diagram of the compression curve
:non drainage boundary
Table 1 the range of each parameter set for analysis :free drainrage bounfdary
k (cm /s) Cc OCR
M axim um m agnification 1.50 1.40 1.04
M inim um m agnification 0.67 0.67 0.96 Figure 2 the one-dimensional analytical model
256
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
determined within this range. Also, we assumed that the error of the Kuroku clay layer (Ma12). And we derived the probability
measured values follows the normal distribution. In addition, we distribution of future sinking amount of Ma12. As a result, the
decided that standard deviation is 1 (cm) for settlement amount and settlement amount of Ma12 at 2051 can be represented by the
0.3 (kgf/cm2) for pore water pressure. In addition, we decided that the probability distribution with an average of 115.38 cm and a standard
error covariance of subsidence and pore water pressure is 0. deviation of 16.07 cm.
6. REFERENCES
3. MEASURED VALUE USED FOR DATA ASSIMILATION 1) T. Higuchi,
In this study, we used the data of settlement and pore pressure 2) S. Ito, K. Oda, K. Koizumi, Y. Usuki,
measured at Ma12 (layer thickness of about 27m) of Osaka Bay
Miocene clay from 2002 to 2014 as actual measurement values used 3) T. Shuku, A. Murakami, S. Nishimura, K. Fujisawa, K. Nakamura,
for data assimilation. Figure 2 shows the one-dimensional analytical Data assimilation of the Settlement Behavior of Kobe Airport
model used in this study. The red circle on the analysis model shows Constructed on Reclaimed Land Using the Particle Filter, Journal of
the measurement position of the settlement amount (1 depth). The applied mechanics, Vol.72, No.4, pp.354-367, 2016.
blue circle shows the measurement position of pore water pressure (9 4) Doucet, A., Godsill, S. and Andrieu, C.: On sequential Monte Carlo
depth, 0, 2, 4, 6, 13, 20, 22, 24, 26 m from the top). sampling methods for Bayesian filtering, Statistics and Computing,
Vol. 10, pp.197-208, 2000 .
Settlement amount(cm)
Figure 3 shows the temporal change of measured values of settlement date
amount and settlement amount identified by data assimilation. In
addition, we decided the settlement amount that weighed average was 20
identified from the weight distribution obtained after data
assimilation. Also, Figure 4 shows the probability distribution of the
subsidence amount after data assimilation at 2014. And, Table 2 40
shows weighted average and standard deviation of the probability
distribution. The probability distribution of the settlement amount,
which we have derived is a weighted average 67.15cm, the 60 Actual m easurem ent value
probability distribution of the standard deviation of 4.2cm. W eighted average
W eighted average±σ
Using this result, we derive the probability distribution of the 80
subsidence amount of Ma12 layer about 40 years later. Figure 5
Figure 3 Change with time of settlement amount
shows the temporal change of the predicted value of the settlement
amount up to 2051 and the actual measurement value of the settlement 0.4
amount. Figure 6 shows the probability distribution of the settlement
amount at 2051. Table 3 shows the average of the predicted values of
the settlement amount at 2051 and its standard deviation. The 0.3
Probability
5. CONCLUSION
0
In this study, we performed data assimilation based on measured
100
40
45
50
55
60
65
70
75
80
85
90
95
values of settlement amount and pore water pressure for Osaka Bay
Settlement amount (cm )
Figure 3 Weighted average and Standard deviation at 2051 Figure 4 probability distribution of the subsidence amount
after data assimilation at 2014
2051 W eighted A verage S tandard D eviation
Table 2 weighted average and standard deviation at 2014
S ettlem entA m ount 115.38 16.07
W eughted A verage S tandard D eviation
(cm )
A fter D ata A ssim ilation 67.15 4.20
0.2 1 A ctual M esurem ent V alue 70.90
Cumulative distribution
(cm )
Probability distribution 0.8
0.15 1999/12/6 2013/8/14 2027/4/23 2040/12/30 2054/9/8
Probability
Settlement amount(cm)
0
0.6 date
20 Actual measurement value
0.1 Weighted average
0.4 40 Weighted average±σ
0.05 60
0.2
80
0 0 100
120
124
134
105
115
144
153
163
66
76
85
95
140
Settlement amount(cm)
Figure 6 Probability distribution of settleatment amount 2051 Figure 5 Change with time of settlement amount to 2051
257
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: The objective of this study is to evaluate the behavior of reconstituted samples of a volcanic soil under static and cyclic
loading using series of undrained static and cyclic test. The static test results show that at a low confining pressure, the soil exceeds
contractive behavior followed by dilative behavior. This behavior indicates that the pore water pressure development is not higher than the
soil shear strength. Otherwise, the soil shows contractive behavior. The cyclic test results show cyclic mobility behavior under an
investigated cyclic stress ratio. In one hand, at low cyclic stress ratio, the shear strain increased slowly, and after a certain number of cyclic, it
significantly increases. In another hand, the shear strain increases gradually at high cyclic stress ratio. These results indicate a contradictive
behavior of the soil under different confining stress and cyclic stress ratio.
Keywords: Static triaxial test, cyclic triaxial test, volcanic soil, reconstituted samples, soil behavior.
1. INTRODUCTION in this paper). This Orange soil deposit, which acted as the slip
surface of the landslide, is located on the top of Pre Takanoobane
Historically, in Japan, some damages triggered by an earthquake on Lava pumice deposit (referred as Blackish soil). In one hand, from
volcanic soil areas has been reported (Hazarika et al., 2018; Song et the map of resistivity distribution of the slope (Fig. 6) that has been
al., 2017; Kazama et al., 2012; Miyagi et al., 2011; Sassa, 2005). drawn by Kochi et al. (2018), it can be concluded that the Orange
One of those area is Aso caldera in Kumamoto prefecture. The area soil which is located on the ground waterbed, is in saturated
suffered from widespread landslides which were triggered by the condition. In another hand, the resistivity map also shows that the
2016 Kumamoto earthquake (Fig. 1). The landslides were Blackish soil is an impermeable deposit. Based on these fact, the
concentrated in the Mount Aso area, within a 64 km radius of the authors presume that the Orange soil deposit was liquefied during
epicenter. The 2016 Kumamoto earthquake is series of earthquakes the earthquake and becomes the main reason of the occurrence of
which struck Kumamoto Prefecture of Kyushu Island on 14–16 the landslide.
April 2016. The foreshock earthquake occurred at 21.26 JST on 14
April 2016 at an epicentral depth of ~11 km at a magnitude (Mw) of
6.5 and the main shock struck at 01.25 JST on 16 April at an
epicentral depth of ~10 km at an Mw of 7 (USGS). The source of
the earthquakes was the activity of the Hinagu and Futagawa faults
(GSI). These two faults experienced more than 2 m of strike-slip
displacement at shallow depth. One of the landslides occurred near
Aso Volcanological Laboratory of Kyoto University (Fig. 2). This
landslide brought damage to houses (Fig. 3), public spaces and roads
(Fig. 4). The inclination of the slope is about 10–15° Kochi et al.
(2018) which is consistent with the value of 12° found by Song et al.
(2017). In those landslides, several volcanic soil deposits have been
found scattered on those slope. Sumartini et al. (2017) reported that
the slope is composed of volcaniclastic deposits (Fig. 5) which
visually have different colors and characteristics. The volcanic soil
deposits came from different places as listed in Table 1. Song et al.
(2017) and Kochi et al. (2018) stated that the landslide is composed
of Kusasenrigahama pumice tephra beds (referred to as Orange soil
Figure 2 A massive landslide near Aso Volcanological Laboratory
of Kyoto University
258
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Several studies related to the Orange soil were done and reported
by Sumartini et al., (2017-2018). They studied chemical, mineral
and microstructure characteristics as well as the behaviour of the
Orange soil under static and cyclic loading. According to those
studies, the landslide occured because the earthquake breakout the
soil structure of the Orange soil deposit and led to liquefaction on
that deposit. Several researchers have been studied the behavior of
those volcanic soils due to earthquake by conducting triaxial cyclic
tests (Ishikawa et al. in 2011, Suzuki and Yamamoto in 2004,
Hatanaka et al. in 1985 and Sumartini et al. from 2017 to 2018).
However, the behaviour of the Orange soil under cyclic loading in
disturbed condition has not been done yet. Even though, the
landslide occurred during the main shock, the deposit is not
completely in undisturbed condition due to the foreshock. Thus, it is
necessary to understand the behaviour of the Orange soil in
disturbed condition. For that reason, series of undrained cyclic
triaxial tests were performed to evaluate the behaviour of disturbed
samples under cyclic loading. Finally, the results were compared
with previous study and presented in this paper.
Figure 5 Schematic profile of the slope in Aso caldera (Modified
from Sumartini et al. 2017)
Figure 6 Resistivity distribution of the slope in Aso Volcanological Laboratory (Kochi et al., 2018)
259
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
2. MATERIAL PROPERTIES
The Orange soil is taken from the scarp of the slope near Aso
Volcanological Laboratory. This soil contains about 60 percent of
fine particles (Fig. 7) and based on its properties (Table 2) it can be
classified as volcanic cohesive soil type II. It also contains 97 % by
weight of feldspar mineral (Table 3) and has vesicular fabric. The (a) Petal-type (b) Flower type
fabric is composed of crystal flakes (Fig. 8). Sumartini et al. (2018b) Figure 9 Idealized crystal flake structures found in the deposits of
idealized the flakes of the volcaniclastic deposit of the slope into the Aso Volcanological Laboratory Landslide. (a) Petal-type
flower type (Fig. 9a) and petal-type (Fig. 9b). structure and (b) Flower type structure (Sumartini et al., 2018b)
3. TESTING METHOD
Undrained cyclic triaxial tests were conducted on the Orange soil
with 60 kPa of confining pressure. The disturbed samples were
reconstituted from the undisturbed samples which have been tested.
About 100 mm in height and 50 mm in diameter of the specimen
have been produced using a steel mould. Double negative pressure
and appropriate back pressure were applied to the samples and
isotropically consolidated at the target effective pressure. B-values >
0.95 was ensured for all samples before shearing. The frequency of
the cyclic axial load was 0.1 Hz for the undrained triaxial cyclic
tests. To decide whether liquefaction occurred in this study or not,
the pore water pressure ratio (ru), define as the ratio of the pore
water pressure to the normal stress, was used. When ru ≥ 0.95, the
specimen was considered to have liquefied.
260
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 11 Stress versus axial strain of undisturbed samples Figure 13 Pore water pressure versus axial strain of undisturbed
(Sumartini et al., 2018a) samples (Sumartini et al., 2018a)
Figures 12 and 13 display respectively the pore water pressure- The behavior of the disturbed and undisturbed samples can be
strain relation of disturbed and undisturbed samples in the undrained clearer by looking at the stress path of the soil as respectively shown
static triaxial test. In low confining pressure, the pore water pressure in Figures 14 and 15. In one hand, under a low confining pressure,
is lower than the deviatoric stress and in high confining stress, the the disturbed sample is contracted and then dilated with no sign of
pore water pressure is higher compared to deviatoric stress. temporary liquefaction. In contrary, under low confining pressure,
Consequently, when the confining pressure is low, the soil the undisturbed sample is dilated, and when reaching the peak of the
susceptibility to liquefaction is low. Otherwise, under high confining soil strength, the soil contracted. In another hand, under high
pressure, the soil has a high susceptibility to liquefaction. confining pressure, disturbed and undisturbed samples show a
contraction which means that they have a high susceptibility to
liquefaction.
Figure 12 Pore water pressure versus axial strain of disturbed Figure 14 Stress path of the disturbed Orange soil samples
samples (Sumartini et al., 2018a) (Sumartini et al., 2018a)
261
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
φ (°) 11 7
φ’ (°) 29 30
c d
262
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
a b
c d
263
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
4. ACKNOWLEDGMENT
The financial grant for this research under the J-RAPID program
(Principal Investigator: Hemanta Hazarika) of Japan Science and
Technology Agency (JST) is gratefully acknowledged. The first
author is grateful to the Indonesia Endowment Fund for Education
(LPDP) for financially supporting her study.
5. REFERENCES
264
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Screw Driving Sounding (SDS) test had been developed about a decade ago and has now been widely used in Japan together
with the Swedish Weight Sounding (SWS) test in site investigation work, particularly for housing development project and investigating the
performance levels of housing lots. Unlike SWS, studies in Japan and New Zealand show that the results from SDS tests can be used to
identify and classify soil types and therefore it is able to give the soil profile without the needs of taking any soil samples. Recently this test
has been introduced to researchers in Universiti Teknologi Malaysia to characterize various soil types. This paper discusses the results of the
SDS tests in identifying and classifying various soil types in Malaysia. Test sites around Kuala Lumpur, Selangor and Johor had been chosen
in performing the SDS tests. Results from three of the sites are compared with the borelogs obtained from the study area comprising of soil
profile, Standard Penetration Test and laboratory tests results. Analysis shows that SDS is able to classify soil types in Malaysia besides its
ability to identify soil layers in a more comprehensive manner. SDS data also correlates well with SPT-N values.
Keywords: Screw Driving Sounding, Standard Penetration Test, site investigation, soil profile, soil classification
1. INTRODUCTION rate of 25 rpm. The 7 load steps for SDS are 0.25, 0.38, 0.50, 0.63,
0.75, 0.88 and 1 kN. The load is increased for every revolution of
In Malaysia, deep boring with Standard Penetration Tests, SPT the rod. For each 250 mm penetration, the rod will move up about
(with disturbed sampling) and undisturbed sampling for laboratory 10 to 20 mm and rotate back down to calculate the rod friction.
tests have been the common method for determining the subsurface
soil profile and geotechnical engineering properties for foundation
design purposes. SPT measures the resistance to penetration offered
by soil at various depths (Wazoh and Mallo, 2014). SPT in Malaysia
is carried out in accordance with MS1056:2005, using a hammer
weight of 65 kg and a drop height of 750 mm. Total penetration is
450 mm and the number of blows for the last 300 mm is the SPT-N
value.
Since performing deep boring either through wash boring, rotary
drilling or percussion drilling are expensive, other tests such as (a) SDS equipment (b) Data can be sent (c) Graphs of collected
Mackintosh or JKR probe is also carried out to complement the soil to office through data can be viewed from
exploration work. In Japan, Swedish Weight Sounding, SWS has cloud system mobile phone
been widely used instead of Mackintosh or JKR probe. Recently, the Figure 1 Screw driving sounding test
Screw Driving Sounding, SDS test had been developed in Japan for (https://www.j-shield.co.jp/english/sds.html)
value added to SWS. SDS is able to classify soil types as already
proven for soils in Japan (Tanaka et al. 2012 and 2015, and Suemasa Measurement items during SDS test includes torque of rod,
et al. 2018) as well as in New Zealand (Mirjafari, 2016 and Mirjafari applied load, penetration velocity and rotation speed at every
et al. 2016). An attempt to implement the test in Malaysia has also revolution of the rod. Both load and torque are applied to the rod at
been reported by Marto et al. (2018a) and (2018b). the same time during the SDS test which is done automatically
A comparison made by Orense et al. (2013) between SDS and (Suemasa et al. 2018).
SWS, SPT as well as Cone Penetration Test, CPT indicated that
SDS test has many advantages including simpler system, faster 2.2 Theoretical Assumptions and Analysis Method
procedure, lighter reaction weight and better cost efficiency than
other sounding tests. SDS also needs only small working area. Not Suemasa et al. (2005) proposed a plasticity model for the SDS test
just that, the data from the SDS machine could be sent to the ‘cloud’ from the results of SDS miniature tests. This plasticity model has
and would be later downloaded from the G-web system for analysis. been illustrated in Figure 2 by Tanaka et al. (2007). According to
Once results are sent to the ‘cloud’, anyone with the given password Tanaka et al. (2015), the combination of torque and vertical load
could access to the analysed data. While the test crew is still at the measured in the SDS test forms a yield locus and the corresponding
site, one could instruct them to do more tests if upon checking incremental components of a rotation rate and a settlement rate obey
through G-web system it shows questionable or contradictory results. the plastic potential rule. The interactive relationship between the
This is the technology in line with the Industrial Revolution 4.0 that combined loads and the corresponding displacement of the soil
is by using automation in work and making use of the cloud element had been described as a constitutive equation.
computing technology (Marto et al., 2018a and 2018b).
265
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
SDS takes into consideration on rod friction during the test. The The Soil Classification Chart is a plot of the coefficient of
concept of estimating the rod friction has been explained by Tanaka plastic potential, Cp’ versus the slope of corrected torque divided by
et al. (2012) and shown in Figure 3. When the rod penetrates into screw point diameter and corrected load, dT/dWD graph. Cp’, which
the ground while being rotated during SDS test, two components of represents the difficulty of penetration, is obtained from
rod friction occurred, which are vertical component, Wf and approximate slope of the graph between the number of normalised
horizontal component, Tf. The frictions are measured after each 250 half turns, NSDD and normalised torque ratio, πT/WD (Tanaka et al.,
mm penetration when the rod is lifted about 10 to 20 mm and then 2015). For New Zealand’s Soil Classification Chart, it is the plot of
rotated back to the previous position. Through the Eq. (1) and Eq. Cp’ against average change of torque, Ave (T) with the equation as
(2) below, the corrected torque, T and corrected load, W at the follows (Mirjafari, 2016):
screw point are calculated for each 250 mm penetration. It is
necessary to deduct the friction to obtain the actual force applied to
the rod. 1 n 1
Ave(T ) (Ti Ti )
n 1 i 1 1
(3)
Wa = Wf + W (1)
266
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Table 1. Site Locations with Respective SDS and Borehole Number Setapak and Figure 8(c) for Batu Pahat area. Results from the
borelogs are checked with the laboratory tests from the disturbed
Borehole soil samples obtained through SPT tests. However, there is a
No Site location SDS No.
No. Obtained from disadvantage of soil profile plotted from borelogs such that the SPT-
1 Cheras SDS28 BHJ1 PWD N values and soil laboratory results are obtained only every 1.5 m
2 Setapak SDS37 BH2 PWD depth, unlike the SDS test which is a continuous result throughout
3 Batu Pahat SDS13 BH2 Project owner the soil depth. In this case a more comprehensive results will be
Note: PWD – Public Works Department, Malaysia obtained from SDS tests.
Setapak
Cheras
(b) Setapak
Batu Pahat
267
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The variations of dT/dWD and Cp’ for the soils with depth are
shown in Figs.11(a), (b) and (c) for the sites of Cheras, Setapak and
Batu Pahat, respectively. From these, the results for respective soil
types are plotted in the Soil Classification Charts in which some are
shown in Figure12. From the graphs in Figure12(a) and (b), it can be
seen that generally the Silty SAND from Cheras occupied the right
hand side of the chart with positive values of dT/dWD while for
(a) 1.75 to 2.0 m (stiff sandy CLAY) (b) 7.0 to 7.25 m (SAND) Cp’, majority of the points spans from 0.9 to 7. For stiff Clayey
SILT in Figure12(c), the points scattered on the left side of the chart
with negative values of dT/dWD and Cp’ of between 0.5 and 2.
268
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(b) Setapak
(a) Silty SAND (b) Silty SAND with traces of gravel Cheras, E0.25 = 0.2NSPT (6)
Setapak, E0.25 = 0.16NSPT (7)
Batu Pahat, E0.25 = 0.27NSPT (8)
269
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
6. ACKNOWLEDGEMENT
SPT-N The authors gratefully acknowledge the financial support and
contribution given by the Japan Home Shield Corporation through
E0.25
the Research Contract No. 4C134, in undertaking this research.
Acknowledgements are also conveyed to respective researchers
from the School of Civil Eng., Universiti Teknologi Malaysia
(UTM) for their assistance in this research. Lastly, the first author
would like to give a special thanks to the Tokyo City University for
appointing her as a Visiting Professor when this paper was written.
270
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Idrus M. Alatas
Civil Engineering Department of ISTN Jakarta, Indonesia
Masyhur Irsyam
Civil Engineering Department of ITB, Bandung, Indonesia
Ramli Nazir
Centre of Tropical Geoengineering UTM, Johor Bahru, Malaysia
Pintor T. Simatupang
Civil Engineering Department of Mercubuana University, Jakarta, Indonesia.
ABSTRACT : Argillaceous rock is a clay rock in fresh condition has a very high shear and crushing
strength. This is because the index properties of argillaceous rock are included in the soft rock category.
However, if weathering occurs due to contact with atmosphere and hydrosfere, then argillaceous rock will
quickly reduce its strength. In this paper, wetting-drying cycle process and drying process were done on
argillaceous rock in laboratory were reported. Changes in the indes and physical properties of argillaceous
rock show a complete adverse behavior due to wetting-drying cycle process and drying process. Due to
wetting-drying cycles, from disintegration ratio parameter (DR) of Semarang-Bawen argillaceous rock
appears to have completely weathered on 24th days and completely non-durable. While due to the drying
process until the 80th day, the physical properties of argillaceous rock seems to be reliable and completely
durable.
Keywords : argillaceous rock, index properties, physical properties, wetting drying cycle, drying process
The 20th SEAGC – 3rd AGSSEA and 22th IGC (PIT) HATTI, Jakarta, November 2018
271
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The 20th SEAGC – 3rd AGSSEA and 22th IGC (PIT) HATTI, Jakarta, November 2018
272
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Triaxial
Walled Tube to Moulding,
Testing
Sampling Core/ Laboratory Extruder,
Barrel Sampling/ Triming
Box Sampling
ARGILLACEOUS SAMPLING DEVELOPMENT
METHODS
Undistubed
Mobilization
Sampling using Triaxial
to
Split Single Core Testing
Laboratory
Barrel
The 20th SEAGC – 3rd AGSSEA and 22th IGC (PIT) HATTI, Jakarta, November 2018
273
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
With this new improvised developed method, process. The change in volume and moisture
then undisturbed sample in the laboratory can be content of the argillaceous rock samples results
eliminated and only done in the field. also changes in other properties index (Alatas I.
M. 2017).
IV. CHANGES ARGILLACEOUS ROCK Initial Hambalang argillaceous rock water
PROPERTIES DUE TO content between 4.1% and 5.3%, during the 80
WEATHERING PROCESS days drying process it reduces to 2.1%. While
the initial water content in Semarang-Bawen
Changes in the properties of argillaceous between 15.9% to 17.6% in 80 days drying time
rock due to drying process as shown in Figure it reduces to 8.9%. Both samples have shown
11 for wetting-drying process include changes in similar behavior to water-level reduction, where
index properties such as moisture content, bulk significant water content reduction occurred
density, dry density, saturation density, void early in the drying process (Alatas I. M. 2017).
ratio. Changes in soil properties of argillaceous
rock due to wetting-drying cycles show a very
Figure 11. Changes in water content on argillaceous rock where (a) Hambalang (b) Semarang-Bawen due to drying process
up to 80 days.
The 20th SEAGC – 3rd AGSSEA and 22th IGC (PIT) HATTI, Jakarta, November 2018
274
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 12. Swelling and shrinkage behavior of argillaceous rock due to wetting-drying cycle process
(a) Hambalang, (b) (Semarang-Bawen)
frequently within a certain time, it will cause a The smectite mineral from the montmorillonite
faster weathering process. Wetting process is mineral family has a greater swelling effect
done 1 soaking, 2 soaking and 3 soaking every 8 when compared with other clay minerals
days and only 5 minutes soaking (Pineda J.A (Mitchell J.K 2005).
2014; Alatas I. M. 2017). Measurements of
vertical and horizontal deformations of the Table 2. Mineralogy content for two types argillaceous
Hambalang and Semarang-Bawen argillaceous rock (LEMIGAS 2015)
TOTAL
rocks were performed with free swells, without Location
CLAY MINERAL (%)
(%)
pressure. . Smg-Bwn
Smectite
50
Illite
3
Kaolinite
2
Chlorite
1 56
From the results of the above properties Hambalang - 12 30 20 62
change research, argillaceous rock from Calsite
CARBONATE MINERAL (%)
Dolomite Siderite -
TOTAL
(%)
Semarang-Bawen has a slightly larger vertical Smg-Bwn 30 - 2 - 32
strain and horizontal strain compared to the Hambalang - - 4
OTHERS MINERAL (%)
- 4
TOTAL
argillaceous rock of Hambalang. This is due to a Quarts Feldspar Plagioclase Pyrite (%)
dominant clay mineral element that is different Smg-Bwn 8 - 2 1 12
Hambalang 30 - 4 - 34
from the two argillaceous rocks. Argillaceous
rock Hambalang contained kaolinite and
The 20th SEAGC – 3rd AGSSEA and 22th IGC (PIT) HATTI, Jakarta, November 2018
275
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Disintegration ratio is represented as an When due to drying process until day 80, DR
argillaceous rock weathering parameter though equal to 0.9162 show both sample still
Grained Size Distribution testing with sieve completelly durable.
shaker. By plotting between the cumulative 1,00
percentage against the grained size as in Figure 0,9162
13. The next process is to determines the
disintegration ratio (DR) by utilizing the graph in 0,90 Hambalang- Drying Process
Ratio ( DR )
8 days Smg-Bwn- Drying Process
80
Cumulate Retained
16 days
0,50 Smg-Bwn- Wetting-Drying
60 Process with 1 soaking in 8
24 days
days
Disintegration 0,40 Smg-Bwn- Wetting-Drying
32 days Process with 2 soaking in 8
days
40
40 days Smg- Bwn- Wetting-Drying
0,30 Process with 3 soaking in 8
48 days
days
20
64 days 0,20
0,1467
80 days 0,1368
0
100 10 1 0,10 0,0965
Diameter ( mm )
Figure 13. GSD Curve Hambalang argillaceous rock due to -
wetting-drying proses with 1 soaking every 8 days in 5 0 8 16 24 32 40 48 56 64 72 80
munites (Alatas I. M. 2017). Time (days)
Figure 14. Changes disintegration ratio (DR) Hambalang
From Figure 14 it can be seen that and Semarang-Bawen argillaceous rock. (Alatas I. M.
Hambalang argillaceous rock is more resistant to 2017).
weathering than Semarang-Bawen argillaceous
rock. On the 24nd day it appears that argillaceous
rock Semarang-Bawen has been completelly non V. CONLUSSION
durable with DR approaching the number 0.00
with a frequent of 2 and 3 soaking in every 8 From the discussion about the behavior of
days until the undisturbed sample completely argillaceous rock properties due to the
non durable on day 24th . However, for weathering process, it can be concluded as
Hambalang argillaceous rock samples on day follows:
24th , DR was 0.67 on 1 soaking , 0.55 on 2 1. The process of wetting and drying is more
soaking and and 0.38 on 3 soaking. Soaking influential to accelerate weathering rather
period 1, 2, 3 times every 8 days, and every than the drying process on argillaceous
soaking period only takes 5 minutes. (Shakoor A rock.
2011; Alatas 2015 c). Until the end of day 80, 2. When wetting-drying cycles occur more
Hambalang argillaceous rock has a DR frequently in certain periods, it is more
respectively of 0.15, 0.14 and 0.10 on the same likely to accelerate the occurrence of
soaking day as above. The difference between weathering in argillaceous rock.
the argillaceous rocks from two samples is due 3. Acceleration of properties change during
to the different dominance of clay minerals. the weathering process is also highly
The 20th SEAGC – 3rd AGSSEA and 22th IGC (PIT) HATTI, Jakarta, November 2018
276
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
dependent on the main dominant clay LEMIGAS, Laboratory. 2015. "SEM and XRD Report for
minerals from argillaceous rock. Semarang-Bawen and Hambalang Clay Shale."
In. Jakarta: LEMIGAS.
4. From shrinkage with a frequency of 2 Mitchell J.K, Soga K. 2005. Fundamentals of Soil
soaking in 5 minutes per day for every 8 Behaviour 3rd edition (John Wiley & Sons, Inc.).
days on undisturbed sample, both samples Pineda J.A, Romero E, De Gracia M, Sheng D, . 2014.
from Hambalang and Semarang bawen 'Shear strength degradation in claystones due to
environmental effects', ' Geotechnique, 64 493-
argillaceous rock can still be measured by 501 (2014).
shrinkage until day 40. After 40 days both Shakoor A, Gautam T P. 2015. 'Influence of Geologic and
samples have collapsed, and the test is not Index Properties on Disintegration Behavior of
can proceed. Clay-Bearing Rocks', Environmental &
5. from disintegration ratio testing using Engineering Geoscience, Vol. XXI, No. 3, August
2015, pp. 197–209.
distubed sample, Semarang-Bawen Shakoor A, Tej P Gultom 2011. 'Assessing The Slaking
argillacesous rock completelly non durable Behaviour of Clay- Bearing Rock', 10th Annual
on day 24 with DR approaching 0.00. While Tecnical Forum Geohazards Impacting
on Hambalang on the same day its DR Transportation In The Appalachian Region,
Columbus Ohio USA.
ranges from 0.38 to 0.67. So it can be
concluded that Hambalang argillaceous
rock is more resistant to weathering than
Semarang-Bawen argillaceus rock.
REFERENCES
The 20th SEAGC – 3rd AGSSEA and 22th IGC (PIT) HATTI, Jakarta, November 2018
277
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: As in civil engineering field, the intricate problem emerges when subsurface system undergoes the vibration, isothermal and
non-isothermal effect in double-porosity soil. This will influence the migration of aqueous phase liquid (APLs) in the subsurface system. A
series of laboratory experimental were carried out to monitor and observe the characteristics of the soil structure and APLs migration in
deformable double-porosity soil under vary temperature. In order to monitor the changes, a unique digital image processing technique was
used. The contour pattern of APLs migration was plotted based on the generated Hue Saturation Intensity (HSI), image analysis using Matlab
and Surfer software. The results of the experiments show that both samples do not have a uniform migration pattern. In addition, the
migration of fluid in soil sample 2 (isothermal condition) was faster than sample 1 (non-isothermal condition). This is mainly because the
condition of heat imposed into the soil could influence the duration required for the migration. It is found that the heat will slow down the
speed of migration. It can be concluded that factors that significantly influence the APLs migration (under vibration, isothermal and non-
isothermal effect) were the soil structure, soil fractured pattern, and soil temperature.
Keywords: Non-isothermal porous media, isothermal porous media, APLs migration, digital image analysis technique, geo-environmental.
278
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The dried kaolin powder was constantly poured into the distilled
water to control the moisture content. The reason for chosen 30%
moisture content for this study was because the optimum moisture
content for the used kaolin soil was 28%. Next, as in the sample
curing period, the moisture kaolin was kept in a cool condition for a
minimum of 24 hours. In order to keep the moisture constant during
the cooling the samples were placed in a re-sealable plastic bag.
Thereafter, the cured mixtures were broken simply by hand and
passed through a 2.36 mm sieve (sieve No. 8) to obtain kaolin
granules to create the double-porosity soil structure. The kaolin
granules were placed in an acrylic soil column and compressed to a
height of 100 mm using a compression machine. The 100 mm height
of soil sample was chosen to ensure uniformity through the sample
depth. The prepared double-porosity kaolin soil sample is given in
Figure 1. Figure 3 Soil sample after vibration process with fracture
A Nikon D90 DSLR digital camera was used for APL migration
image acquisition at specific time frequency interval. The setup of
digital camera during the experiments was set at minimum shutter
speed of 1/640 second and ISO speed set at ISO-2500, which has
Figure 2: Heater and thermocouple setup been implemented for all the soil sample experiments. To soften the
problem of insufficient image acquisition, this experiment used a
The vibration frequency for the vibration table was set on the single digital camera, and two mirrors for reflection positioned
control panel with 0.98 Hz frequency with 60 seconds duration for behind the circular soil column were adjusted to enable a clear
vibration process based on the previous literature. The result of image facing the digital camera. Thus, the image of the APL
fractured double-porosity soil sample with 30% moisture content migration throughout the whole area of soil column circumference
after the vibration process is displayed in Figure 3. The concept of can be initially captured by just a single click on the DSLR digital
fracture double-porosity soil was to prove that the vibration effect camera. The light source for soil sample come from linear
had caused the double-porosity soil to fracture. The fractures on the fluorescent lamp-40 watt that was placed slightly above the circular
top of soil sample can be clearly seen. soil column.
279
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Both soil samples were first sheathed in white paper with pre- 3. RESULTS AND DISCUSSION
drawn gridline (20 mm x 20 mm) onto the soil column as a control
point on the reference image. Once the reference image was taken, After the APL migration process, the top view of the fractured soil
the pre-drawn gridline paper can be removed from the circular soil samples 1 and 2 as shown in Figure 6. The actual size measurement
column. The adjustable heater machine starts to heat the soil with 30 for the column circumference zone was divided to aid visualize the
Degrees Celsius and verified by the thermocouple to confirm the crack position for samples 1 and 2.
temperature of 30o C for soil sample 1 with the non-isothermal
condition. Meanwhile, sample 2 for isothermal condition was
implemented in the room temperature of 23o C without used the
adjustable heater machine. Both experiments began by pouring the
dyed water instantaneously onto the top centre of fractured
aggregated soil sample in acrylic soil column. The quantity of 70 ml
dyed water was used in samples 1 and 2. After the dyed water had
covered the whole surface area of the fractured soil sample, the first
digital image of aqueous phase liquid migration was taken. The
dyed water migration pattern at a specific time frequency interval
was captured for the rest of the subsequent digital images. A total of
115 images were subsequently recorded in 60 minutes for soil a) Soil sample 1 (Non-isothermal) b) Soil sample 2 (Isothermal)
sample 1, while 118 images were captured in 90 minutes for soil Figure 6 APL migrated soil surface with measurement of
sample 2. actual column circumference zone for samples 1and 2
2.3 Digital Image Processing Analysis Setup The downward migration pattern of HSI contour plot of APLs in
The captured colour digital images were saved in JPEG format and fractured double-porosity soil sample with 30% moisture content for
transferred from Nikon D90 digital camera to computer for further sample 1 (non-isothermal) and sample 2 (isothermal), respectively,
image processing using Matlab routine and Surfer Software. A are shown in Figures 7 and 8. The hole appear on sample 1 was the
developed Matlab routine for digital image processing was used to effect of the rods removed from the sample. When the HIS plot in
extract the area of interest from captured image and to transform the curve joint the right and left boundary has formed the circular shape
area of interest from distorted image to a scale image via affine was clearly apparent in two-dimensional shape; however, the actual
transformation method, which involves converting the JPEG scale aqueous phase liquid migration in the acrylic soil column was one-
images to Red Green Blue (RGB) and HSI images; extracting HSI dimensional.
digital value from HSI image and saving the HSI value in a text file
using American Standard Code for Information Interchange (ASCII)
format.
First, the surfer software was used to digitize the control point
from reference image to extract actual image coordinates of control
points to serve as control coordinate. While, area of interest refers to
pre-determined migration boundary area (front image and V shape
refection image) for sample 1 and 2 that contained the APL as
shown in Figure 5. Matlab routine was then used to convert area of
interest into RGB and HSI image format, where the values were
extracted and saved in ASCII format. Matlab routine was used to (0.5 Minute)
loop three times for the subsequent digital image to extract and save
the intensity values for all three section area of interest of the acrylic
soil column. Finally, the contour plot pattern of the APL migration
in fracture porous media based on the generated HSI value by digital
image processing analysis was plotted. The HSI contour plot of APL
migration behaviour in the condition of non-isothermal and
isothermal condition can provide detailed information to facilitate
researchers to understand the pattern of APL.
(6 Minutes)
(40 Minutes)
Figure 5 Digitization of the control point for area of interest
280
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Based on the observation soil sample 1, the soil condition was dried,
and process of liquid evaporation occurred due to the heat induced
in the soil sample. Both samples were not fully migrated due to
water has higher viscosity compared to previous research by Sa’ari
et al., (2015) was fully toluene migrated to the bottom of soil sample
with the lower toluene viscosity. Besides, this could be because the
physical bonding between toluene and soil weaker than water and
soil. The physical bonding between toluene and soil was attributed
to Van Der Waals Force, which are weaker than hydrogen bonding,
(90 Minutes) which has stronger physical bonding between water and soil.
Figure 7 HSI plots of downward migration in fractured
double-porosity soil for non-isothermal condition
281
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
continue after 6 minutes to gradually incline horizontally until the the other high speed rates occurred at 270 mm and 0 mm column
end of the experiment. Meanwhile, the rest of the column circumference with the value of 0.322 mm/s and 0.308 mm/s,
circumference positions continue a slow, decreased migration from respectively. Thus, the overall average dyed water migration speed
start until the end of the experiment. rate for soil sample 1 is 0.043 mm/s. The highest average migration
speed rate was at 240 mm column circumference zone at 0.061
mm/s. This could be because the larger fractured soil structure after
the process of vibration was at the position of 240 mm.
The soil sample 2 displayed the higher migration speed rate
from initial to 30 seconds at 90 mm column circumference zone
with the speed rate of 2.234 mm/s. Meanwhile, the column
circumference zone at 150 mm and 120 mm also have a high speed
rate with the values of 2.231 mm/s and 2.224 mm/s, respectively.
Therefore, the overall average dyed water migration speed rate for
sample 2 is 0.097 mm/s. This scenario occurred because the soil
sample 2 with isothermal condition without the temperature effect,
could cause the penetration to be faster than soil sample 1.
282
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
283
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: The Menard pressuremeter was developed by Louis Menard in the 1950s. Since its availability as a tool to the engineers for
the last 60 years, pressuremeter has been used extensively for ground improvement projects. It is proven to be a valuable tool throughout the
life cycle of a project, from the geotechnical investigation for design works to acceptance test conducted to verify the construction completed
for a ground improvement project. This paper presents the application of Menard pressuremeter in the construction of an LNG tank in South-
East Asia using stone columns as the foundation system. Some of the important lessons learned from the pressuremeter tests conducted for
this project are presented and discussed in this paper.
Keywords: Menard pressuremeter, ground improvement, stone column, LNG tank.
1. INTRODUCTION tank on the project site. This former tank applied some 150 kPa on
the ground. Thus, the area in Layer 2 that was under the old tank
Stone columns are often used as reinforcing and stiffening elements footprint is over-consolidated, whereas it is normally consolidated
in ground improvement works to increase the bearing capacity and everywhere else. Groundwater level is at about 1.0 m below the
reduce the settlement of the treated ground. Numerous methods existing ground level.
(e.g., Priebe (1995), Goughnour and Bayuk (1979), Balaam and
Booker (1981)) are available to calculate the settlement of ground Table 1 Soil characteristics
improved by stone columns. These methods rely on the modulus of
the stone column, Ec and that of the soil, Es to determine the SPT CPT
Layer
settlement reduction factor following ground treatment with stone N-value qc (MPa)
columns. In recent years, numerical modelling such as finite element 1b: Medium dense to dense sand 23 14.5
analysis has also been used in the design of stone column projects. 1c: Medium dense sand 16 8.2
These types of modelling also require reliable soil and stone column 2: Very soft to soft clay 2 0.5-0.9
moduli to be utilised in the analysis. One of the in-situ tests that can Note: The SPT N-value and qc shown are average values.
provide a direct measurement of the soil modulus is the Menard
pressuremeter test (PMT). This paper demonstrates the use of PMT 2.2 Description of the Ground Improvement Works
in the construction of stone columns for the foundation of a large
LNG tank in South-East Asia. The difficulties encountered to The overall ground improvement works include the removal and
conduct the PMT inside stone columns for the project and the replacement of Layer 1a (i.e., loose sand layer) with well compacted
solution adopted are highlighted. Finally, some lessons learned from granular material and the wet top feed stone column method. Under
examining the results of the PMT by considering the nature and the LNG tank footprint, Layer 1a will be excavated and replaced
characteristics of the soils, the intensity of the stone column with well compacted granular fill (with minimum oedometric
compaction and the measured Menard pressuremeter modulus, EM at modulus, Eoed of 10 MPa) from +1.2 m EL to +2.2 m EL. From +2.2
different depths are presented. m EL to the bottom of the tank slab (+2.9 m EL), the tank pad will
be constructed with well compacted granular material with Eoed of
2. LNG TANK PROJECT 40 MPa. The scope of the stone column works includes penetration
of the vibroflot down to 10.7 m depth from working platform level
2.1 General Description of the Project (i.e., from +2.7 m EL to -8.0 m EL). This is followed with the
The purpose of the ground improvement works is to reinforce the installation of stone columns with average diameter of 1,200 mm
soft compressible cohesive soil (i.e., Layer 2 in Figure 1) with within the clay layer and 600 mm in the sand layer as shown in
stones and to improve the sand material on top of it (i.e., Layers 1b Figure 1.
and 1c in Figure 1) to construct a Liquefied Natural Gas (LNG) The tank dead load is shown in Figure 2.
storage tank of 77.9 m diameter on top of the treated ground.
Figure 3 shows the live load during hydrotest, which is the most
critical stage of the live load applied to the tank. The ground
improvement was therefore designed for it based on the live load
Figure 1 Cross section of ground improvement with stone column during hydrotest.
284
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The ground improvement scheme for the LNG Tank foundation has ensure the probe is within the stone column at the soft clay layer.
been designed with the aim to minimise two differential settlements: Thus, the idea of constructing “test” stone column within the clay
a. Settlement between centre and edge of slab must be lower than layer was adopted. The “test” stone column was constructed by
1/300 (dishing criterion); installing stone column only within the clay layer with stone
b. Settlement between two diametrically opposed edge points consumption and compaction energy similar to those applied for the
must be lower than 1/500 (tilting criterion). construction of normal stone columns. The difference was that
The final design is for the stone columns to be installed with a 2 m beyond the clay layer, the hole created for the construction of the
square grid spacing pattern, giving the area replacement ratio, ar of “test” stone column was backfilled using sand with little
7% within the sand layer and 28% within the clay layer. compaction. This ensured the pressuremeter probe can penetrate to
the “test” stone column within the clay layer and meaningful test
2.3 Testing Programme can be conducted.
Apart from the digital recordings of the construction parameters
0 0
(primarily depth and compaction effort versus time) for every stone
column using a data logger during construction of stone columns, -1 -1
plate load test (PLT) and pressuremeter test (PMT) were specified as 2.3 15
acceptance tests. The PLT was carried out on top of some stone -2 -2
2.6 18
columns using 600 mm diameter steel plate to check on the -3 -3
deformation modulus based on the settlement versus load graphs 2.6 15
obtained. The PMT was carried out within some stone columns at 1 -4 -4
m interval to obtain the limit pressure, PL and Menard pressuremeter 2.6 16
-5 -5
modulus, EM. The acceptance criteria for PMT were to achieve
Depth (m)
2.4
Depth (m)
13
minimum PL of 1.5 MPa and EM of 15 MPa at each test location. -6 -6
These values were derived based on French recommendations NFP 1.9 11
11-212 (2005), which recommends the Young’s modulus of the -7 -7
stone column (average value over the stone column), Ec to be 60 2.4 12
-8 -8
MPa. 1 6
-9 -9
0.2 2
-10 -10
3. PRESSUREMETER TEST RESULTS 0.2 2
-11 -11
3.1 Menard Pressuremeter Test (PMT)
The Menard pressuremeter was developed by Louis Menard for -12 -12
0 1 2 3 4 0 10 20 30 40
commercial application through his first Menard company known as PL (MPa) EM (MPa)
“Centre d’Etudes Menard” in 1957. The original pressuremeter was
developed in an attempt to solve the problem of sampling Figure 4 Initial pressuremeter test result
disturbance and to ensure that the macro-fabric of the soil can be
adequately represented from an in-situ test. To date, the Menard 3.3 PMT Result for Stone Column within Clay Layer
pressuremeter has evolved as part of an international standard
procedure to give design parameters directly. This procedure covers Figure 5 shows the typical PMT result obtained from a “test” stone
the type of probe, installation, testing and interpretation as provided column formed within the clay layer. It is shown that the PL is about
in the publication “Interpretation and Application of Pressuremeter 1.1 MPa and the EM is around 8.5 MPa. This indicates that the PL
Test Results to Foundation Design” published in Sol-Soils (Menard, and EM are below the specified values despite being much higher
1975). The pressuremeter test provides two main soil parameters, than the natural soil’s PL and EM. The EM of 8.5 MPa translates to
namely Menard pressuremeter modulus, EM and limit pressure, PL, Young’s modulus within stone column, Ec of 34 MPa (i.e., Ec
which can be used directly for settlement and bearing capacity EM/; where 0.25 for gravel). This is about 11 times higher than
calculation, respectively. the Young’s modulus of the natural clay (i.e., Es of about 3 MPa)
and it is well above the typical modular ratio (i.e., Ec/Es) for the
3.2 Initial PMT Result application of stone columns in soft soil. In fact, for the design
approaches using elastic methods that consider vertical deformation
Figure 4 shows the plot of PL and EM versus depth from the initial only, Barksdale and Bachus (1983) have suggested that Ec/Es in
pressuremeter test (PMT) conducted from the working platform practice to be ranging from 3 to 10, depending on the area
level at +2.7 m EL. The result suggests that the upper 5 m of the replacement ratio, ar adopted. The reasons why such a high stiffness
stone column was well constructed. However, below that level it ratio was obtained for this project are likely to be due to the fact that
appears as though the stone column was not properly compacted high area replacement ratio, ar was adopted with column diameter of
within the medium dense sand layer (i.e., Layer 1c) and the test 1.2 m and high compaction energy was applied during construction.
result within the clay layer is even worse where the PL and EM
obtained are not any better than the PL and EM of the natural soil.
-8 -8
This kind of PMT test result was not expected in the project as 1.2 8
D e p th (m )
-9 -9
D e p th (m )
285
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3.3 PMT Result for Stone Column within Sand Layer 20 (i.e., Ec/Es 60/3). However, as noted from the PMT results, the
modular ratio achieved is only about 11. Clearly, a constant Ec
Figure 6 shows the typical PMT result for the test conducted on a
should not be specified and a limit on the modular ratio should be
stone column within the sand layer when the pressuremeter probe
imposed in the design of stone columns. The following limits can be
did not walk out from the centre of the column. The PL is about 2.8
considered:
MPa and the EM is around 22 MPa, which are higher than the
required values in the specification. The Young’s modulus of the In sand, despite the fact that relatively high Ec is achievable, the
stone column, Ec is calculated to be 88 MPa, which is 3 to 4.5 times modular ratio is likely to be around 3 to 5 due to the relatively
higher than the Young’s modulus of the original sand layer (i.e., Es high Es for sand.
of 20 MPa to 30 MPa). In clay, although relatively low Ec is the norm, the modular ratio
is likely to be around 5 to 12 due to the relatively low Es for soft
0 0 clay.
The above limits are similar to the findings by Barksdale and
-1 -1 Bachus (1983), which suggested Ec/Es in practice is ranging from 3
1.8 11
-2 -2
to 10 and Han (2012) which noted that the modular ratio is from 1 to
2.8 16 5 for flexible columns and 5 to 10 for semi-rigid columns.
-3 -3
2.4 13 4.3 Settlement Calculation
Depth (m)
Depth (m)
-4 -4
3.1 26 Based on the experience in this project, it appears that it is better to
-5 -5 consider the minimum modular ratio between stone column and soil
2.9 22
-6 -6 (i.e., Ec/Es) in the design of stone column than specifying a constant
3 25 stone column modulus, Es to be achieved. In fact, designing and
-7 -7 specifying stone column projects using the modular ratio should
3.4 34 come quite naturally as most of the analytical methods to calculate
-8 -8
28
settlement of the soil treated with stone columns are using this ratio
3.3
-9 -9 to determine the settlement reduction factor for the settlement
0 1 2 3 4 0 10 20 30 40 calculation. Settlement reduction factor, n is the ratio of the
PL (MPa) EM (MPa) settlement of the untreated ground to the settlement of the ground
improved with stone column.
Figure 6 PMT result of stone column within sand layer
Zooming into the Priebe’s (1995) method, which is one of the
most popular methods used for calculating the settlement of ground
4. DISCUSSION ON THE PMT RESULTS treated with stone columns, the maximum allowable settlement
4.1 Stone Column Modulus reduction factor is given by the equation below:
In this project, the modulus of the stone columns, Ec had been
specified to be 60 MPa throughout the length of the column. The
PMT results show that the Ec in the clay layer and sand layer is 34
MPa and 88 MPa, respectively. Clearly, the value of Ec is not where:
constant throughout the length of a stone column. From various nmax maximum allowable settlement reduction factor
literature, the value of Ec is shown to range from 25 MPa to 100 ar area replacement ratio Ac/A
MPa and it is a function of the confining pressure (Castro, 2017). In Ac cross section area of the stone column
fact, the value of Ec is likely to be interdependent on factors such as A tributary soil area of the stone column (or unit cell area)
the stiffness of the natural soil, the compaction energy provided, the Ec/Es modular ratio between stone column and soil
area replacement ratio (i.e., spacing and diameter of stone columns),
etc. Therefore, it may not be suitable to specify or design a stone It should be noted that the above equation (1) is not normally the
column project using a constant Ec throughout the treatment depth, settlement reduction factor to be adopted in the calculation of
especially when the subsoil is heterogeneous. settlement after stone columns installation using Priebe’s (1995)
method, but it is the cap of the settlement reduction factor that can
4.2 Modular Ratio be adopted in the analysis. To illustrate quickly the effect of natural
Table 2 shows the modular ratio between stone column and natural soil and stone column moduli in settlement calculation, the above
soil (i.e., Ec/Es) obtained from the pressuremeter test conducted in equation is adopted to calculate the maximum settlement reduction
this project. factor for this project as tabulated in Table 3.
286
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
also within the proposed modular ratio for sand (i.e., 3 Ec/Es 5). 5. CONCLUSION
Indeed, the proposed modular ratio seems to be a good tool to
This paper has presented the application of Menard pressuremeter
estimate the range of settlement reduction factor following the
test as part of the quality control measures for the application of
installation of stone columns to treat the soils in a project.
stone columns as the foundation system for a large diameter LNG
tank. The challenges associated with conducting pressuremeter test
Table 3 Calculation of settlement reduction factor
inside the stone column is presented and the solution to overcome
Natural Ec Es these challenges is provided. The pressuremeter test results from this
ar Ec/Es nmax Note project are presented showing the limit pressure, PL and
soil (MPa) (MPa)
pressuremeter modulus, EM obtained inside stone columns within
34 3 11.3 3.92 (a) the sand and clay layers. It is shown that the PL and EM inside stone
60 3 20 6.37 (b) columns are interdependent of the nature and characteristics of the
Clay 28%
- - 5 2.13 (c) natural soils treated with stone columns, the area replacement ratio
- - 12 4.11 (d) (i.e., spacing and diameter of the stone columns) and the compaction
energy adopted during construction of the stone columns. The
88 30 2.9 1.14 (a) pressuremeter test results also suggest that specifying or designing
Dense 60 30 2 1.07 (b) using a constant Ec throughout a stone column might not be
7%
Sand - - 3 1.14 (e) appropriate, especially for heterogeneous soil. It is best to utilise the
- - 5 1.28 (f) modular ratio between stone column and soil (i.e., Ec/Es) for the
settlement analysis of soil treated with stone columns.
Note:
(a) Actual data from project
6. REFERENCES
(b) The Ec is specified to be 60 MPa
(c) Using lower bound of modular ratio (i.e., Ec/Es 5) Balaam, N. P., and Booker, J. R. (1981). “Analysis of rigid rafts
(d) Using upper bound of modular ratio (i.e., Ec/Es 12) supported by granular piles”. International Journal for
(e) Using lower bound of modular ratio (i.e., Ec/Es 3) Numerical and Analytical Methods in Geomechanics, Vol. 5,
Issue 4, pp379-403.
(f) Using upper bound of modular ratio (i.e., Ec/Es 5)
Barksdale R.D. and Bachus R.C. (1983). “Design and Construction
of Stone Columns Volume I”. Federal Highway
4.4 Recommendations Administration Report FHWA/RD-83/026, National
The above section shows that the important parameters for the Technical Information Service, Springfield, Virginia.
calculation of settlement of a ground treated with stone columns are Castro, J. (2017). “Modeling Stone Columns”. Materials, Vol. 10
the area replacement ratio, ar and the modular ratio, Ec/Es. The area (7), Article 782 (https://doi.org/10.3390/ma10070782).
replacement ratio, ar is related to the diameter and spacing of stone Goughnour, R. R., and Bayuk, A. A. (1979). “Analysis of Stone
columns which can be designed or determined with relatively high Column – Soil Matrix Interaction Under Vertical Load”,
precision. The difficult part in the design or settlement calculation of Proceedings of International Conference on Soil
ground treated with stone columns is determination of the modulus Reinforcement, Vol. 1, Paris, pp279-285.
of stone column, Ec to be adopted. It has been shown that the Han, J. (2012). “Recent Advances in Column Technologies to
approach of using a constant Ec in the settlement calculation of Improve Soft Foundations”, Proceedings of International
ground treated with stone columns might not be a good idea, Conference on Ground Improvement and Ground Control,
especially for heterogeneous soils (with mixture of sand and clay or Wollongong, Australia, pp99-113.
silt). As such, the modulus of stone column, Ec should not be Menard, L. (1975). “The Menard Pressuremeter: Interpretation and
specified with a constant value, especially with a high value that Application of Pressuremeter Test Results to Foundation
might not be achievable for stone columns constructed in soft soils Design”. General Memorandum D.60.AN. Sols Soils, N° 26,
due to the low confining pressure from the soft soils. Thus, it is pp5-43.
recommended for the modular ratio, Ec/Es with the limits proposed Normes Françaises NFP 11–212. (2005). “Recommandations sur la
on the above to be used in the settlement analysis to get the idea of conception, le calcul, l’exécution et le contrôle des colonnes
the likely improvement factor of installing stone columns for a ballastées sous bâtiments et ouvrages sensibles au tassement”.
project. Following this, a verification at site can be done by Revue Française de Géotechnique, N° 111, pp3-16.
conducting in-situ test such as the pressuremeter test. Finally, if Priebe, H. J. (1995). “The design of vibro replacement”. Ground
necessary, improvements or changes in the design can be conducted Engineering, December, pp31-37.
before full production works is carried out.
287
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: A structural reclamation embankment material generally requires minimum sand content of 50%, maximum gravel of 30%
and silt & clay of 20%. Related to quality control material in the field, it is desirable to only do sieve analysis with some gradation coefficient
and water content test to know the correlation some parameters of soil density such as dry density, void ratio, effective cohesion, internal
angle friction & CBR. Thus it can be determined at that time in the field whether the embankment of reclamation has been qualified or not.
To study the correlation between gradation and CBR with physical and mechanical parameters of embankment material then performed
laboratory tests, among others CBR, sieve analysis, volumetric gravimetric, proctor and direct shear tests. Research through laboratory tests
to study the correlation between the CBR with the physical parameters of the soil embankment uses reclaimed embankment material
consisting of sand (pass of sieve No.2) and clay-silt (pass of sieve No. 200). Samples are made of 5 types with composition sand and silt-clay
as follows 100% : 0% , 95% : 5% , 90% : 10% , 85% : 15% , 80% : 20%. The next step is the compaction of modified proctor and CBR test
with saturation-unsaturated condition and then volumetric-gravimetric and specific gravity test. Other tests such as sieve analysis and
hydrometer test are performed to determine Cu (coefficient of uniformity) and Cc (coefficient of curvature). This test is done using
reclamation embankment material from river, beach and hill quarry, while the material from the selected quarry is material containing
maximum 20% clay and sand 80% minimum. The next step is compacted by Modified Proctor and Direct Shear testing with saturated water
content (on Zero Air Void line) and unsaturated (on wopt). The results of the tests indicate that the increasing value of Cc/Cu then the value of
the maximum dry unit weight (dmax), effective cohesion and effective internal friction angle decreases and increases the optimum water
content. In addition the test results show that the increasing value of d then the CBR value increase and the void ratio (e) and porosity (n)
decrease. The correlation between sieve analysis and CBR with soil physical and mechanical parameters is applicable for reclamation
material with minimum sand composition 80% and clay maximum 20% and the limit of Cu is 1<Cu <50.
Keywords: embankment, reclamation, density, shear strength, CBR.
1. INTRODUCTION - What is the effect of water content (w) on dry density (d) and
soil shear strength ( and C).
Implementation of coastal reclamation requires very large
embankment material. The general requirements of embankment
This research is expected to be a solution in the selection of
material must be fulfil maximum 30% gravels, minimum 50% sand
reclamation material. The expected solution is related to the material
and maximum 20% clay and density must be comply with
composition and parameters based on the required CBR, in addition
requirements. Generally the field practitioners want to know quickly
to estimate the required density based on the composition of
and easily when the embankment material arrives in the field by
embankment material grain size.
only observe at the composition of sand and clay-silt. Based on the
composition can be determined whether the material involves the
requirements or not as a material embankment. Furthermore, the 2. LITERATURE STUDY
required density can be determined by performing one test only. The 2.1 Compaction
test can be a CBR test or a sieve analysis test.
Determination of the density level of a soil can be seen from Lee and Suedkamp (1972) concluded that the soil compaction curve
three parameters: relative density (Dr), dry unit weight (d) and void can be divided into four types. Type A compaction curve has bell-
ratio (e). Relative density is only used for granular soil type, while shaped. Type A is generally present in most clay soils with a liquid
dry unit weight (d) and void ratio (e) for all types of fine and limit value (LL) between 30 - 70. Type B curve is one-half peak,
generally present in the sand with LL<30 (type B curve is a more
coarse-grained soils. Therefore the dry unit weight (d) and void
suitable result with condition of our predominant sand test sample).
ratio are more suitable for use in reclamation work because the soil
Type C has double peak curve, found on soil with LL<30 or LL>70.
type of embankment consists of fine and coarse grained soils. The
The D-type curve is odd-shaped, generally present in soils having
determination of grain size distribution is very important in order to
LL>70.
determine including into well graded, poorly graded or gap graded.
These gradations are determined from the value of uniformity
The relationship between moisture content (w) and dry unit
coefficient (Cu) and the curvature coefficient (Cc), which greatly
weight (γd) and weight of zero air void volume (γZAV) can be
affects the density of the embankment. In addition, the density is
formulated as follows:
also influenced by water content. Water content in the field is
influenced by the rainy season and reclamation work is always sat
associated with tidal conditions. This submerged and non-
d (1)
1
submerged condition affects the compaction. Such conditions affect
the compaction process and the stability of embankment. G
S 100% W s
(2)
The existing problems based on above condition are as follows : r ZAV 1 G
s
- How does the composition of sand and silt-clay affect the dry
unit weight (d), and the correlation of soil composition with the
where :
CBR value.
- How is the influence of water content and the relationship sat = saturated unit weight of soil
between void ratio and CBR value. w = unit weight of water
- How is the relation between Cu, Cc, gradation of grain size with Sr = degree of saturated (%)
dry unit weight (d), internal friction angle () and cohesion (C), Gs = spesific gravity
for saturated and unsaturated initial conditions. w = water content
288
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Day (1997) says that the requirement factors decrease of void the contrary it is said to be negative if the increase (decrease) x is
ratio (e) at the compaction are the distribution of grain size, the ratio generally followed by a decrease (increase) y.
between d100/d0 (ratio between the largest and smallest diameter) The strength of the relationship between x and y is measured by
clay (with low activity) to fill the smallest void and compaction the correlation coefficient (denoted r). The correlation coefficient
process, to compress soil particles into more dense arrangement. value is -1 ≤ r ≤ 1, which means:
If r = 1, the relationship of x and y is perfect and positive (close to 1,
2.2 Swelling ie very strong and positive relationship).
R = -1, the relationship of x and y is perfect and negative (close to -
Swelling calculation used the equation below: 1, ie very strong and negative relationship).
H R = 0, the relationship of x and y is very weak and there is no
1 100% (3)
relationship.
H0
It says x affects y if the change in value of x will cause a change
where : in the value of y; that means the rise and fall of x will make the y
ε1 = axial strain (%) value also up and down, so the y value will vary, either to the y
ΔH = change of sample height (mm) average or to the linear line representing the scatter diagram.
Ho = initial of sample height (mm) However, the rise and fall of y is such that the y value varies, not
only due to x, because there are other factors that cause it.
Previous studies have been conducted to predict the magnitude The magnitude of the contribution of x to the rise and fall of the
of swelling. One of which is the research conducted by Seed et al. y value is obtained from the coefficient of determination, denoted by
1962, it discuses the correlation between plasticity index (IP) with R2. If r = 0.92 for example, then the value of R2 = (0.92)2 = 0.85
swelling potential (Table 1). (85%), ie the amount of factor x contribution to the rise and fall of y
is 85%, while 15% by the other factor. In addition to the value of r
Table 1 Correlation between swelling (ε1) and Plasticity Index (IP) and R2 is also required standard deviation value (s) to determine
(Seed et al., 1962) whether valid or not valid data from the distribution of data
obtained. The smaller standard deviation value, the result of the
IP (%) εs (%) Swelling potential
existing analysis is closer to mean of linear regression curve, which
0 – 10 0 – 1.5 Low
means the existing data is more valid. Correlation regression theory
10 – 20 1.5 – 5 Mean
is used to prove that the relationship between parameters sought is
20 – 30 5 – 25 High valid.
>35 >25 Very high
2.5 Correlation of CBR with Physical Parameter of Soil
2.3 Gradation
Several studies related to CBR correlation and other soil physical
The distribution of soil grain size is known as gradation (Bowles, parameters have also been performed by several researchers. Shirur
J.E. Hainim, Johan K, 1984). Gradation is usually known on coarse et al. (2014) with submerged soil having LL between 20 - 70 gives
grained soils. The gradient indication can be calculated numerically CBR relationship and parameters as follows:
from the grain size curve by using the uniformity coefficient (Cu)
defined as : CBR = 5.09477-0.09323 (LL) + 0.10939 (SL) + 0.022566 (SI)
D CBR = 5.813 – 0.007826 (LL) + 0.12097 (PL)
Cu 60 (4) CBR = -4,8353 – 1,56856(OCMC) + 4.6351 (MDD)
D10 CBR = -3.2353 – 0.06939 (PI) +2.8(MDD)
The shape of the curve defined as the coefficient of curvature (Cc), CBR = 6.5452 – 0.07703(OMC) – 0.10395(PI)
Cc
D30 2 (5) where :
D60 D10 CBR = california bearing ratio
LL = liquid limit
where :
PL = plastic limit
D10 = grain diameter at 10% through sieve SL = shrinkage limit
D30 = grain diameter at 30% through sieve PI = plasticity index
D60 = grain diameter at 60% through sieve OMC = optimum moisture content
MDD = maximum dry density
When the Cu value gets larger (≥4), the soil grain size
distribution becomes heterogeneous, indicating that the gradation of Zohib et al. (2016) with soil type SW-SM (well graded sand
the soil is well graded. And the other hand, if the value of Cu is containing silt) submerged condition has the following relation:
getting smaller then the grain size is getting the same (uniform). CBR = 0.142 (%fines) + 0.0262 (LL) +0.0283 (OMC) + 1.043
The curvature coefficient (Cc) is used as a reference of the grain (MDD) -17.029
size distribution curve shape between D60 and D10. The ideal grain The parameter relation has been valid since it has R2 = 0,7 - 0,836
size distribution curve shape of the Cc value lies between 1 and 3. If
the value of Cc <1 then the shape of the grain size distribution curve 3. RESULTS
shows the curved shape above the ideal curve and if Cc >3 then the
shape of the soil grain size distribution curve shows the curved Based on previous research, the authors conducted laboratory
shape below the ideal curve. experiments on two types of soil samples i.e soil samples made so as
to have the composition of sand and clay 100%: 0%, 95%: 5%,
2.4 Regression Correlation Theory 90%: 10%, 85%: 15%, 80%: 20% and soil samples taken from
various locations as in Table 1. In all the compositions of sand and
Correlation is the relationship degree between several variables that clay, the greater the moisture content (wc) the greater the dry
aim to know how well a linear equation or other equations. density (γd) as shown in Figure 1. At certain water content γd will
Relationships x and y are said to be positive if the increase decrease, because the percentage of water that fills the pores
(decrease) x is generally followed by an increase (decrease) y. On between the grains enlarged. Therefore the percentage of incoming
solid granules is not maximal. The greater the percentage of clay-silt
289
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
in the mixture, the optimum water content will be greater, because Figure 2 shows for all the composition of sand and clay tend to
the content of large clay (SiO2) can absorb water (H2O). In terms of greater the dry unit weight (γd) the greater the CBR value.
density (γdmax), more greater the percentage of clay-silt then more This is due to the greater unit weight of soil thus the more dense soil
greater γdmax will be but in the mixture clay-silt greater than 10%, and the soil bearing capacity increase, which is seen with the value
the γdmax decreases. It can cause instability such as low bearing of CBR increase.
capacity and large settlement.
From the sieve and hydrometer test results (Table 1) shows that
the embankment material from rivers and hills content of the fine
Figure 4 Relationship CBR vs porosity (n) for unsoaked condition
fraction (silt-clay) is greater than 2%, except in Lanang River
Ngrambe Ngawi containing only the fine fraction 0.18%. When
Figure 4 shows that the CBR value increases hence the porosity
compared with the location of embankment material from the beach
(n) decreases. This is due to the increasing value of CBR thus the
fine fraction has ranged between 0.13% - 0.83%.
soil occur granules reposition. Therefore, the porosity is smaller.
Based on Figure 2 to 4 can be obtained the relationship between
CBR with γd, e, n for unsoaked condition as shown in the following
equation :
290
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
CBR = 38.524(γd) – 46.992 (R2 = 0.9991) 1. From the results of this study concluded that optimal
CBR = -29.128(e) + 39.851 (R2 = 0.9977) composition of reclamation embankment material to density is
CBR = -0.709(n) + 48.836 (R2 = 0.9992) 80% sand and 20% clay. If the content of clay is multiplied (>
10%) or increases slightly (<10%), resulting in decreased
The correlation between γd, e, n with CBR for soaked condition density.
can be seen in the following equation : 2. The results of this study indicate that the value of CBR, physical
parameters and material density in unsaturated conditions is
a. Sand 100% dan Silt-Clay 0% more optimal than with saturated conditions.
3. The most dominant soil parameter of the effect on the CBR value
CBR = 134.95(γd) – 214.94 (R2 = 0.9956) is the dry unit weight parameter (γd), this is because the CBR
CBR = -75.451(e) + 67.337 (R2 = 0.9788) value increase if the soil density increases.
CBR = -2.2052(n) + 105 (R2 = 0.9699) 4. The larger the Cc/Cu ratio the smaller dry unit weight maximum
(dmax), the shear strength of the soil (effective cohesion, C' and
b. Sand 95% dan Silt-Clay 5% effective shear angle, ') while the larger optimum water content
(wopt). This is due to the greater Cc/Cu ratio, the embankment
CBR = 55.967(γd) – 90.89 (R2 = 0.9453) material tends to be uniform and the smaller the grain diameter
CBR = - 66.287(e) + 51.681 (R2 = 0.9699) of the soil. Therefore, there is void resulting in density, bonding
CBR = -1.6763(n) + 74.39 (R2 = 0.9829) and interlocking between granules getting lower. At the same Cc
/Cu value, the shear strength of the soil (effective cohesion and
c. Sand 90% dan Silt-Clay 10% shear angle in effective) in the unsaturated initial state is greater
CBR = 58.014(γd) – 92.613 (R2 = 0.9969) than in the saturation initial state. The presence of water-
CBR = -70.884(e) + 56.679 (R2 = 0.9998) saturated conditions will make the distance between the bonding
CBR = -1.7141(n) + 78.045 (R2 = 0.9991) granules getting away and consequently reduced cohesion and
can block interlocking or friction between grains.
d. Sand 85% dan Silt-Clay 15% 5. Materials from the ocean tend to have greater optimum water
content to reach their density compared to materials from rivers
CBR = 37.995(γd) – 54.207 (R2 = 0.9685) because the embankment material on the coastal site is
CBR = -42.771(e) + 40.517 (R2 = 0.9884) dominated by fine sand with a grain diameter smaller than 0.425
CBR = -1.1253(n) + 56.692 (R2 = 0.9760) mm, requiring larger water to be solid compared with the
embankment material of the river with an average content of fine
e. Sand 80% dan Silt-Clay 20% sand is less than 40%.
291
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: The Municipal Solid Waste Disposal in Ngipik – Gresik (near Surabaya, in East Java Province, Indonesia) has become an
overflowing landfill due to the indiscriminate dumping of solid waste at the site. Leachate migration away from the landfill boundaries
creates serious environmental problems which include, and are not limited to vegetation damage, groundwater pollution and soil pollution.
This present paper is meant to study the possible impact of the leachate percolation to the adjacent soil and groundwater. The soil samples
and leachate samples were collected from this landfill-site and its surrounding area to study the possible impact of leachate percolation on
soil characteristics. The geophysical (Electrical Resistivity Tomography – ERT) investigations were carried out on this landfill-site and its
adjacent area, to better understand the sub-soil characteristics. The water quality data were collected to study the possible leachate
penetration on the ground water. Concentration of various physicochemical parameters and engineering properties were then be determined.
The study indicated that leachate percolation quantity has strongly affected the groundwater properties. There is a general change in
groundwater properties which are attributed to the chemistry of leachate.
1. INTRODUCTION
Ngipik Landfill is one of the landfills in East Java Province,
Indonesia, which still applies the open dumping method for waste
disposal (Figure 1). Although Ngipik landfill was designed for
sanitary landfill in 2002, this landfill still applies the open dumping
method until 2016. Recently, Ngipik landfill is shifting to be a
controlled landfill. The amount of waste carried and dumped in
Ngipik Landfill was around 624,72 m3 per day or 187,42 tons per
day with waste density 300 kg/m3 (Agustina, 2012).
The common problem of landfills is the leachate deployment.
Leachate means any liquid percolation through the deposited waste
and emitted from or contained within a landfill. The leachate
consists of many different organic and inorganic compounds that
may be either dissolved or suspended. They will bring potential
pollution issues for groundwater and surface waters in nature. The
landfill leachate is a secondary contamination related to landfills
(Monroe, 2001). Landfill deployment problems have been studied
by several researchers such as Suparmanto et al (2006) in Benowo
Landfill Surabaya and Endah and Pudjiastuti (1997) in Keputih Figure 1. Location map of the Ngipik landfill site
Landfill Surabaya. They found out that the disposal method was the
main cause of the groundwater contamination. Based on geological setting, the geological formation at
Sutra et al.(2016) found that water from the monitoring wells Ngipik Landfill is in the Lidah Formation (Figure 2) with Quarter
(Figure 1) has been influenced by the leachate from Ngipik Landfill. age, composed of clay, black and blackish, hard when dry, little
It means that the clay soil has also been contaminated due to the fossil content and sandstone clay. Characteristics of the Lidah
contaminant migration by the water. The increasing use and Formation are to have fairly thick clay content on the layer (Sukardi,
expanding of this site will seriously affect the quality of 1992). This can be proved by drilling results, up to 30 m indicating
groundwater in the region besides the major influence on the air that the research area is dominated by silt and clay. Detailed soil
pollution. characteristics, will be discussed further in this paper.
292
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3. RESULTS OF MEASUREMENTS
3.1. Soil characteristics
293
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
294
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
high value of TDS may be due to the leaching of various pollutants worth less than 2.4 ohm m. This data will be used to interpret
into the groundwater (Suman Mor, et al, 2006). resistivity measurement results in the field
COD is a measure of oxygen equivalent to the organic matter
content of the water susceptible to oxidation by a strong chemical
oxidant and thus is an index of organic pollution. The COD level in
the groundwater samples higher than 24 mg/L, indicating the
presence of organic contaminants in the water and can be used as
organic indicators to assess the groundwater pollution caused by
landfill (Suman Mor, et al, 2006).
The N total (NH3) and Cd concentration was also found to be
remarkably high at all samples. The Cd is one of heavy metal that
undesirable metal or toxic in drinking water. Thus, it can be said that
the groundwater around the landfill could not be used for drinking
water purposes and likely indicate that groundwater quality is being
significantly affected by leachate percolation. Strictly speaking one
should avoid using groundwater drawn from the wells located in
proximity of the landfill sites.
γd Ww Ws wc opt Wt d t
(gr/cm3) (g) (g) (%) (g) (cm) (cm)
295
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
g. Line 7
7. REFERENCES
Agustina, Ummy Arofah. (2013). Effect of Leachate Recirculation
on Solid Waste Degradation Rate In Ngipik Landfill Gresik. ITS
Undergraduate Thesis of Environment Engineering, RSL 628.42
Umm p.
Alloway, B.J; Ayres, D.C. (1997) Chemical principles of
environmental pollution (Second edition). Waste and other
multipollutant situation. P 357. Blackie A &P. ISBN
0751403806.
Figure 7. The 2D resistivity tomography on line 1 – line 7 Ayolabi, E.A., Folorunso, A.F., Idem, S.S. (2013) “Application of
electrical resistivity tomography in mapping subsurface
Furthermore, from the inversion 2D resistivity results, 3D cross-
hydrocarbon contamination”, Earth Sci. Res. 2 (1), 93-104.
section is made by displaying only the target resistivity value which
Awni T. Batayneh. (2005) “2D Electrical Imaging of an LNAPL
is a low anomaly that is the presence of considerable water content
Contamination, Al Amiriyya Fuel Station, Jordan”, Journal of
in the soil (leachate water flow). The 3D result (as illustrated in
Applied Sciences, 5: 52-59.
figure. 8) combined with the direction of the aquifer stream (figure
Braja M, (2013) “Principles of geotechnical engineering”, 7th edn.
3), shows that the leachate plume follows the aquifer stream.
Domenico PA, Schwartz FW, (1990) “Physical and Chemical
This indicates that the leachate distribution pattern that occurs
Hydrogeology”, Wiley, New York.
from the overloaded area of eastern garbage, the leachate spreads to
E. Al-Tarazi, J. Abu Rajab, A. Al-Naqa, M. El-Waheidi. (2008)
the lower groundwater level that is towards the Southeast -
“Detecting leachate plumes and groundwater pollution at
Northwest.
Ruseifa municipal landfill utilizing VLF-EM method”, Journal
of Applied Geophysics, 65. 121-131.
Endah, N. and L. Pudjiastuti. (1995) “The effect of Leachate on
Surabaya MSW in Sukolilo Landfill to Water Quality of
Resident’s Wells Around”, Research Institution, Institut
Teknologi Sepuluh Nopember.
Fetter CW. (1988) “Applied hydrogeology”, Nerrill Pub Co., A well
and Howell information Co., Columbia, p 529
Loke, M.H,. (1999) “Electrical imaging surveys for environmental
and engineering studies”: A practical guide to 2-D and 3-D
surveys.
Monroe, Michele. (2001) “Landfill Leachate Treatment: VSEP
Offers a Revolutionary Solution Introduction”, New logic
Research, Inc. Company News.
Rabinove C. J., Long Ford, R. H. and BrookHart, J. W. (1958)
‘Saline water resource of North Dakota U.S.’, Geol. Sur. Water
Supply Paper 1428, 72.
Suman Mor, Khaiwal Ravindra, R. P. Dahiya, A. Chandra. (2006)
“Leachate Characterization and Assessment of Groundwater
Pollution Near Municipal Solid Waste Landfill Site”,
Environmental Monitoring and Assessment (2006) 118: 435–
Figure 8. The 3D perspective view of leachate plumes 456.
Sukardi. (1992) “Peta Geologi Lembar Surabaya & Sapulu, Jawa”.
Pusat Penelitian dan Pengembangan Geologi, Bandung.
4. DISCUSSION AND CONCLUSION Sutra, N., Soemitro, R.A.A., Warnana, D.D. (2016) “Preliminary
Assessment to Contaminated Soil Characteristics on Ngipik
From the groundwater testing and resistivity measurement, it is Open Dumping Waste In Gresik Residence”, Proceedings of 2nd
clearly seen that the leachate generated from the landfill site is International Conference of Problematic Soil, Bandung,
affecting the groundwater quality in the adjacent areas through Indonesia.
296
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This paper applied digital elevation models of different periods and long-term field monitoring data to compare elevation
changes of a spillway for remedial works of landslide-dammed lakes in the Jiufengershan landslide. Digital aerial photogrammetry and
LiDAR techniques were used to rebuild orthophotos and digital elevation models in the area from 1998 to 2015 in order to review the
changes of deposit and erosion in topographies and profiles of the spillway. Through many torrential rainfalls in the past, it was particularly
effective for related spillway remedial works to suppress the production and flow of collapse debris from the bed and bank of the spillway
channel. It is the most important for spillway remedial works to remain the stabilization of two dammed lakes.
Keywords: Landslide-dammed lake, spillway, remedial works, digital elevation model.
297
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 2 3D Geological Structure in the Jiufengershan landslide Figure 4 Relationship between rainfall and elevation of lake level in
the Sezaikeng dammed lake
Table 1 Relative elevations, depths and volumes of two landslide-
dammed lakes 3. REMEDIAL WORKS OF THE LANDSLIDE AND
Jiutsaihu Sezaikeng DAMMED LAKES
dammed lake dammed lake
3.1 Emergency treatment
Area (ha) 4.43 6.48
Dammed lakes were vulnerable to catastrophic failure by
Before completed spillway overtopping and breach, landslide emergency treatments involved in
the potential failure of these dams need to be carried out for
Elevation (m) 585 577.5 protection of relative people’s living and property. Thus, there were
two items of remedial works for the landslide and two dammed
lakes.
Max lake depth (m) 29 37.5 Firstly, one of emergency mitigative measures was an
innovative method, namely temporary geotextile dams of debris-
Storage volume (m3) 678,000 1,089,700 filled shipping containers, in the end part of accumulation area in
order to increase strength of slope stable and avoid headward
After completed spillway erosion. The measure of debris-filled shipping containers was the
first for the slope toe loading of a landslide in Taiwan whose height,
width and height were 122.7 m, 8.2 m and 7.5m, respectively, and
Elevation (m) 575.89 570.05
the aims of the measure were for four functions as marked in Figure
5(1) (they are: ①slope stable, ②slope toe loading, ③headward
Max lake depth (m) 15 29 erosion control, ④ protection of stream bank).
298
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The other emergency measure is a spillway of dammed lakes 4.1. Landform changes with photogrammetry and LiDAR
that is an emergency drainage channel with 1450 m in length and 8 survey
m in depth in order to discharge the overtopping of dammed lakes
and decease the storage volume of two lakes in October 1999, as Digital aerial photogrammetry and analysis tools of ERDAS
shown in Figure 5(2). IMAGINE were used to rebuild orthophotos and DEMs for different
periods before and after the 1999 Chichi earthquake. The result is
shown in Figure 7. According to the topographic analysis (see Table
3.2 Soil and water conservation treatment
2), the average depth of the landslide is 34 m with a collapse volume
After emergency mitigative measures of dammed lakes, it is of 34.92 million cubic meters and a deposit volume of 36.58 million
important for remedial works of a permanent spillway so as to avoid cubic meters.
severely incised channel, losing a lot of debris and dammed-lake
failure. One of the remedial works was a slit-type sabo dam that was
located at 1K+230 m from the overflow outlet of the spillway in
2000. The aim of the dam was for a direct barrier to debris flows
which could occur. The location is shown in Figure 6.
4. PERFORMANCE EVALUATION OF SPILLWAY Figure 8 (1) Landslide change between 2004 and 2015; (2)
WORKS Landslide change between 2011 and 2015. (Erosion area is in green
color and deposit area is in orange color.)
The landslide blocked two streams which caused two landslide-
dammed lakes so a spillway was built to discharge water volume of 4.2 Profile changes of the spillway with photogrammetry and
the two lakes. Thus, it is important to evaluate remedial works of the LiDAR data
spillway. The paper applied DEMs for different periods and long-
term monitoring data to compare elevation changes of the spillway DEMs of the area were built by the data of photogrammetry in 1998,
and water surfaces of dammed lakes. Digital aerial photogrammetry 2002 and 2005. Then, profiles of the Jiutsaihu stream before and
and Light Detection and Ranging (LiDAR) techniques were used to after the 1999 Chichi earthquake were drawn in Figure 9 by using
rebuild orthophotos and DEMs in the area from 1998 to 2015 in the data of DEMs. The graph of profiles in the figure shows
order to review the change of accumulation and erosion in profiles elevation changes of the stream bed before and after the earthquake
of the spillway. in which the maximum landslide depth is 87 m at 2K+250 of the
profile line. It is clear for the profiles that there was an erosion
299
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
phenomenon at 3K+500 of the upstream and a deposit area occurred Through some heavy rainfall events from 2004 to 2015, the highest
from 0K+775 to 1K+365 at the downstream that was possibly lake surface went up approximately 1~4 meters, which means that it
caused by excavated spillway or collapsed debris of the upstream. is successful for the spillway to control the changes of lake surfaces
in two dammed lakes. According to Schuster (1986), 91% of 63
global barrier lakes broke down within a year after having formed.
As a result, it is valuable for the spillway in the landslide that two
dammed lakes can be maintained and there was no severe disaster in
the past.
Compared to elevation changes of the spillway by DEMs of
different periods with photogrammetry and LiDAR techniques, there
are some results of the analysis. Firstly, the average depth of the
landslide is 34 m with a collapse volume of 34.92 million cubic
meters and a deposit volume of 36.58 million cubic meters while the
Chichi earthquake occurred. Secondly, the data of DEMs in 2004,
2011 and 2015 shows that landform changes in the landslide were
nonstop in which the landslide volume is estimated about 267,000
m3 with a maximum erosion depth of 27 m and maximum deposit
depth of 24 m. Then, there was a significant downcutting in the
elevation change by 20 m between 0K+870m and 1K+80m in the
spillway from 2004 to 2011, which caused some submerged dams
damaged but the changes gradually decreased from 2011 to 2015
after some check dams were completed from 2008. Finally, the most
Figure 9 Location and profile of the Jiutsaihu stream before and
important performance is to remain the stabilization of two dammed
after the Chichi earthquake
lakes after emergency treatment and a series of spillway remedial
works.
Other DEMs of the area were built by the data of LiDAR in
2004, 2011 and 2015 with the grid precise of 5 m. Then, profiles of
6. REFERENCES
the spillway were drawn by using the data of DEMs in the three
periods, as shown in Figure 10. Meanwhile, the locations of Schuster, R. L. (1986). Landslide dams: processes, risk, and
remedial works, including a sabo dam, 20 columnar-type submerged mitigation. ASCE.
dams and check dams, are plotted in the figure. Shou, K. J., & Wang, C. F. (2003). Analysis of the Chiufengershan
landslide triggered by the 1999 Chi-Chi earthquake in
Taiwan. Engineering Geology, 68(3), 237-250.
Chen, X. Q., Cui, P., Li, Y., & Zhao, W. Y. (2011). Emergency
response to the Tangjiashan landslide-dammed lake resulting
from the 2008 Wenchuan Earthquake, China. Landslides,
8(1), 91-98.
Delaney, K. B., & Evans, S. G. (2017). The evolution (2010–2015)
and engineering mitigation of a rockslide-dammed lake
(Hunza River, Pakistan); characterisation by analytical
remote sensing. Engineering Geology, 220, 65-75.
Xu, Q., Fan, X. M., Huang, R. Q., & Van Westen, C. (2009).
Landslide dams triggered by the Wenchuan Earthquake,
Sichuan Province, south west China. Bulletin of engineering
geology and the environment, 68(3), 373-386.
Cedergren, H., 1989 “Seepage, drainage & flow nets”, McGraw
Figure 10 Elevation changes of the spillway for different periods Hill.
Li, T., Schuster, R. L., & Wu, J. (1986). Landslide dams in south-
Figure 10 illustrates that there was a significant downcutting in central China. In Landslide dams: processes, risk, and
the elevation change by 20 meters between 0K+870 m and 1K+80 m mitigation (pp. 146-162). ASCE.
in the spillway from 2004 to 2011, which caused falling and damage Sager, J. W., & Chambers, D. R. (1986). Design and construction of
of some columnar-type submerged dams. Nevertheless, the the Spirit Lake outlet tunnel, Mount St. Helens, Washington.
elevation changes gradually remained stable from 2011 to 2015 after In Landslide Dams: Processes, Risk, and Mitigation (pp. 42-
remedial check dams and groundsills from 2008. The other point is 58). ASCE.
that the sabo dam effectively blocked the massive debris against its Code, J. A., & Sirhindi, S. (1986). Engineering implications of
flowing to downstream. All spillway works could have a capacity of impoundment of the Indus River by an earthquake-induced
debris pool by 6 million cubic meters. Dammed-lake failure and landslide. In Landslide Dams: Processes, Risk, and
debris flow did not occur in the landslide area while all spillway Mitigation (pp. 97-110). ASCE.
works completed. Cameron, C. P., Cato, K. D., McAneny, C. C., & May, J. H. (1986).
Overall, through many torrential rainfalls and typhoons from Geotechnical Aspects of Rock Erosion in Emergency
2004 to 2015, related remedial works which have cost about 12 Spillway Channels. Department of the Army, Waterways
million USD were particularly effective suppress the production and Experiment Station, Corps of Engineers, Geotechnical
flow of collapse debris in the bed and bank of the spillway channel. Laboratory.
The most important performance is to remain the stabilization of Hansen, D. C., & Morgan, R. L. (1986). Control of Thistle Lake,
two dammed lakes. Utah. In Landslide Dams: Processes, Risk, and Mitigation
(pp. 84-96). ASCE.
5. CONCLUSION
From the lake-surface monitoring data of two dammed lakes, an
average drop of the lake surface in the Jiutsaihu and Sezaikeng
dammed lakes was 8.28 m after completed spillway works in 2000.
300
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: In this paper, a comparison is shown between site factors derived from the conventional VS30 parameter based on SNI 1726-
2012 and from the soil profile natural frequency (f0) referred to Cadet et al (2012). Three series of stochastic site response analyses (SRA) as
a benchmark for the former two approaches have been performed considering linear viscoelastic condition from three large data sets. The
data sets were stochastically generated to represent different site classification uniformly: soft, medium, and hard soil (SE, SD, and SC soil
class, respectively). For each stochastic model set, 1000 soil profiles were generated and its linear responses were computed to derive the
spectral ratio at 1 Hz (FV – SA 1 s), 5 Hz (FA – SA 0.2 s), and 25 Hz (PGA). Correlation scatter charts were then developed to observe and
to check the performance of each approach. The charts showed correlation of site factors based on the VS30 and f0 parameters with the SRA
modeling. Based on the correlation, the best site amplification factor estimation approach for each spectral acceleration (FPGA, FA, FV) and
for each soil class (SE, SD, SC) is provided.
Keywords: site factor, site amplification, local site effect, natural frequency, VS30
301
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
value of dimensionless frequency. The equation of their SAPE Model 3 (soil class SC) High range site frequency:
method is given by the following formula: approximately f0 > 5 Hz
The purpose of this division is to have a clear verification on the
reliability of FA (0.2 s – 5 Hz) and FV (1 s – 1 Hz) for sites with
(1)
broadband f0 and VS30.
The artificial soil profiles are modeled by assuming three-
where Gx is the estimated site amplification factor, X is the soil layered profile overlying a halfspace (bedrock, soil class SB). The
natural frequency, Xref and β are the coefficients derived from linear thickness and the velocity of each layer in each model set are
regression at a considered dimensionless frequency, v. The determined stochastically using truncated Gaussian distribution with
coefficient table of Xref and β is given as an appendix in their the parameter set as presented in Table 2.
publication.
Table 2 Statistical parameters for developing artificial soil profiles
2.2 VS30-based site factor (SNI approach)
VS (m/s) thickness (m)
In this paper, the other site proxy for comparison is the 30 meters mean std mean std
averaged shear wave velocity (VS30) which is determined by: layer 1 100 10 10 2
model 1;
layer 2 120 30 10 4
class SE
(2) layer 3 150 30 10 5
layer 1 160 20 10 2
model 2;
layer 2 350 40 10 4
where di is the layer thickness up to 30 meters depth and VSi is class SD
layer 3 500 60 10 5
the shear wave velocity of layer i (in m/s). The calculated VS30 then layer 1 550 20 10 2
corresponds to five soil classifications (Table 1), each has its own model 3;
layer 2 650 40 10 4
site amplification factor which was determined based on Borcherdt class SC
layer 3 700 50 10 5
(1994). The linear site amplification factors as applied in SNI 1726-
2012 are represented in the following table. Aside from layer thickness and shear velocity, unit masses and
material damping are also required for linear viscoelastic analysis.
Table 1 Site classification and its corresponding linear site In this study the unit masses and material damping of each layer is
amplification factor (SNI 1726-2012) determined by using an approach proposed by Cadet et al (2012).
Soil VS30 (m/s) FPGA FA FV The unit masses of each layer are assumed to be increasing with
Class shear wave velocity, as follow:
(PGA<0.1g) (SS<0.25g) (S1<0.1g) ρ = 1.7 g/cm3 for VS < 180 m/s
SA > 1500 0.8 0.8 0.8 ρ = 2.0 g/cm3 for 180 < VS < 360 m/s
ρ = 2.2 g/cm3 for 360 < VS < 1500 m/s
SB 750 – 1500 1.0 1.0 1.0 ρ = 2.5 g/cm3 for VS > 1500 m/s
The damping ratio of each layer is represented as quality factor
SC 350 – 750 1.2 1.2 1.7
(Q) and is derived using the following formula, with VS is the shear
SD 175 – 350 1.6 1.6 2.4 velocity in m/s.
SE < 175 2.5 2.5 3.5 Q = VS/10
2.3 Stochastic analysis Based on the statistical parameter set and assumptions as
described above, 1000 artificial soil profiles are created for each
The simulation framework of this study is to divide the stochastic model set, and the histogram of the site frequency (f0) and VS30 are
model of soil profile into three different VS30 ranges: presented in the figure below. VS30 is determined using Equation (2),
Model 1 (soil class SE) Low range site frequency: while f0 is determined from the transfer function of each profile.
approximately f0 < 1 Hz
Model 2 (soil class SD) Medium range site frequency:
approximately 1 Hz < f0 < 5 Hz
302
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 1 Histogram of natural frequency and VS30 for all model sets
The above histograms show that not all generated soil profiles regression coefficient tables as in Cadet et al (2012) is used for
comply within the desired frequency/VS30 range, for instance not SAPE approach.
all profiles of model 1 have f0 lower than 1 Hz, as there is still about Linear viscoelastic site response analysis is then performed for
50% which has larger than 1 Hz. However, large portion of the each generated soil profile using a recorded outcropping rock
profiles have VS30 value lower than 175 m/s which put them to soil ground motion during Kobe (1995) earthquake, scaled down to PGA
class SE. Most of soil profiles in model 2 lie within the desired of 0.05g to ensure a linear condition. The response spectrum of the
range, for both f0 and VS30. Similar to model 1, part of soil profiles calculated motion at ground surface is calculated, and compared
in model 3 have f0 lower than 5 Hz while the rest has f0 exceeding with the response spectrum of the input motion at three specific
beyond 5-10 Hz. However, most of the profiles have VS30 between frequencies, correspond to the site factor value of FV, FA, and
350-750 m/s (soil class SC). It is not clearly depicted in the figure FPGA. The calculated site factors from site response analysis are
but there is one soil profile in model 3 which lies in soil class SB then compared with the estimated site factors from both SAPE and
classification and two soil profiles lie in soil class SD. SNI approaches. The analysis procedure in this study is better
After VS30 and f0 have been determined for each generated soil illustrated by the following flowchart as below.
profile, the next step is to determine the corresponding site
amplification factor (FPGA, FA, FV) based on the two approaches: 3. RESULTS
SNI (using VS30) and SAPE (using f0). Table 1 is used for SNI
approach, only considering the site factor for the smallest spectral The correlation charts between both estimation approaches and the
acceleration (linear condition). Equation (1) along with several site response result are given in Figure 3 and Figure 4. The
observation result from the charts is summarized in this section.
303
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 3 SAPE approach correlation scatter chart for all soil classes
Figure 4 SNI approach correlation scatter chart for all soil classes
304
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3.1 Soil class SE For SNI approach, FPGA has a good correlation while FA
underestimate the site response factors. FV-SNI does not give a
With the exception of FV-SAPE, it is clear from the figures above good result for this soil class since the estimated site factors over-
that both approaches do not give a satisfactory correlation with site predict the value if compared with site response result.
factors from site response analysis as the scatters are deviating from
the correlation line. FPGA-SAPE tends to have a narrow value while
4. DISCUSSION
the value distribution from the site response is broad. F A-SAPE has
a broader distribution compared to FPGA, but the deviation is still Figure 5 below shows an illustration of how SAPE was derived by
high. One interesting observation is for FV-SAPE, as the trend and Cadet et al (2012). As explained above, hundreds of spectral ratio
the distribution of the scatter is good for profiles with site frequency from KIK-Net database was collected and the frequency of each
(f0) higher than 1.0 Hz (FV is in the range of 2.2 - 3.9). For profiles curve is normalized with the natural frequency of the corresponding
with f0 lower than 1.0 Hz, the scatter becomes worse and deviating soil profile, leading to a set of “bell-shaped” curves with maximum
from the correlation line (FV-SAPE range is narrow, from 3.5-4.0 amplitude located at abscissa of unity. Then, a correlation is made
while FV-site response has very broad distribution). for natural frequency and its amplification at a specific
FPGA, FA, FV determined from SNI all show bad correlation with dimensionless frequency value.
site response site factor. Having that in mind and by observing Figure 5, by logic it is
safe to say that the correlation produced for dimensionless
3.2 Soil class SD frequency less than one (left side of black dashed line) will be better
than for dimensionless frequency beyond unity (right side of dashed
For this class, FPGA-SAPE has a bad correlation similar to soil class line). The left side correlation will be better because each curve is
SE. SAPE gives a significantly narrow distribution of FPGA values always asymptotic at the beginning and eventually reach its maxima
while site response gives a broader distribution. FA-SAPE has a at the dashed line, thus the scatter distribution is more regular. On
good correlation for profiles with f0 larger than 5 Hz, while the the contrary, for dimensionless frequency larger than one, the scatter
distribution becomes worse for profiles with f0 lower than 5 Hz. FV- distribution is more irregular due to the presence of amplification at
SAPE in general has a satisfactory correlation, although the scatter higher modes. This will have implication on the quality of the
slightly deviates from the correlation line, but the trend is good and correlation between amplification value and natural frequency.
correlates well with site response FV. It can be observed that all soil The explanation above explains Figure 3 on how SAPE
profiles in this model set have f0 larger than 1 Hz. performs at a certain frequency (FA / FV / FPGA) in each soil class.
Site factors from SNI approach still have bad correlations with For soil class SE, SAPE gives good result for FV only, for sites with
site response site factors, however the distribution of site response f0 larger than 1 Hz. Recall that FV is a site factor at 1 Hz frequency
site factors are now become narrower if compared with the (period 1 sec) and dimensionless frequency is a ratio between the
distribution from soil class SE. considered frequency of spectral ratio value and the natural
frequency of a soil profile (f0). Therefore, the calculated
3.3 Soil class SC dimensionless frequency for sites with f0 larger than 1 Hz is less
FPGA-SAPE has a good correlation (values varying between 1.0 - than one, thus lies on the left side of the dashed line in Figure 5.
1.5). A bad correlation is observed for FA-SAPE, since the Meanwhile, sites with f0 lower than 1 Hz will have dimensionless
distribution of FA by the approach is very broad ranging from 1.0 - frequency value which is larger than one, thus plotted in the right
3.0 while the site response factor have values ranging from 1.0 - 2.0. side of the dashed line and it explains the bad correlation observed
Similar to FA, FV-SAPE also seems to overestimate the site for these soil profiles.
response result.
For soil class SD, SAPE gives good result for FA (frequency 5 performance might be due to the difficulty in obtaining a good
Hz; period 0.2 sec) for profiles with f0 larger than 5 Hz. SAPE also transfer function with a clear amplification by using field
performs well for FV (frequency 1 Hz; period 1 sec) since generally instrumentation at rocky sites with high VS30 (soil class SC, SB).
all soil profiles in this model set has f0 larger than 1 Hz. For FPGA It is confirmed that for “hard soil” sites, a good quality
(frequency 25 Hz, period 0.04 sec), SAPE leads to bad correlation amplification spectrum is really difficult to get either by using
since at this frequency (25 Hz), all soil profiles have dimensionless geophysical tool (H/V microtremor) or by deriving the spectrum
frequency value greater than one. from surface/borehole accelerometer since the impedance contrast
By using the same logic as described above, for soil class SC, between the soil profile and the reference rock site is relatively
SAPE should give good result for FA and FV and bad result for weak (Pierre-Yves Bard - personal communication). However,
FPGA since almost all of the soil profiles have f0 larger than 1 Hz despite the bad data scatter of FA-SAPE for this soil class, in fact
and 5 Hz, but lower than 25 Hz. However the result is the contrary; the FA correlation is satisfactory enough for sites with f0 larger than
FPGA has a satisfactory correlation while bad correlation is 15 Hz.
achieved for FA. The main reason behind this anomaly in SAPE
305
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Although the trend of FV-SAPE correlation for soil class SC such method since site estimation method always comes with
seems to be not good enough, the fact that: uncertainty (epistemic and aleatory).
1. FV correlation achieve good result for soil class SE and SD for 2. It appears that SAPE approach is not fully reliable, as it has bad
sites with f0 larger than 1 Hz, and correlation for soft soil sites and for specific spectral
2. All soil profiles within soil class SC has f0 larger than 1 Hz acceleration. However, SAPE result is better than SNI approach
lead to an argument that in fact FV -SAPE also performs well for in several conditions.
class SC. 3. In-line with the finding of Zhao and Xu (2013) and Cadet et al
The bad correlation of SNI approach as observed for FPGA, FA, (2012), natural frequency of a soil profile (f0) turns out to be a
FV for SE sites explains very well the complexity of a site response better site proxy compared to VS30. The very root of the reason is
at soft soils, which has broad variability and cannot be simplified because VS30 cannot include the effect of deeper soil stratum
into a single scalar site factor value (this is also the case for SAPE below 30 m. Furthermore, amplification is also a function of
result). However, it seems that for soil class SD, the correlation is impedance contrast of the soil profile with the underlying
not becoming any better. The SNI site factors achieve satisfactory halfspace, which makes an even worse estimation for VS30
result for hard soil sites (class SC) for PGA. For F A, good result is approach for deep-bedrock profiles.
only valid for sites with VS30 larger than approximately 450 m/s. 4. The study presented in this paper is only performed for linear
Compared to SAPE site factors for class SC, SNI factors tend to condition (low PGA, small strain), therefore non-linearity is not
have low value which can be explained by two reasons. First, the taken into account. The inclusion of such variable will make it
reference stations within this soil class, as used by Borcherdt (1994) more challenging since the uncertainty is increased.
in his study, is basically consist of hard soils and soft rocks, which Based on the finding and discussion, a summary table is
eventually will lead to less amplification. Second, as mentioned prepared which consist of suggested approach for different site
above, the difficulty of obtaining f0 in hard soil sites leads to conditions and different spectral accelerations. The authors hope
overestimation of site factors. In this case, class SD and SE are the that this paper can give an insight for practicing engineers and
governing soil profiles and lead to correlation bias. stakeholders, that there is much into it behind the site factors of SNI
1726-2012 as adopted from NEHRP provision 1997. In the end, the
5. CONCLUSION AND RECOMMENDATION selection of any method to estimate site amplification will largely
depend on many aspects, e.g. project stage, project budget, project
Through stochastic site response analyses performed on varying risk, etc.
artificial soil profiles, the application of site factor estimation It is important to note that this table is made solely based on the
procedure by VS30 (as adopted in SNI 1726-2012) and by f0 (Cadet finding of this study. Improved aspects of this study e.g. more
et al, 2012) has been compared and several conclusions can be taken uniform distribution of artificial soil profile, deeper bedrock depth,
as follow: different method to estimate unit weight and material quality factor,
1. There is no perfect site factor estimation approach. Site proxy is etc may lead to different recommendation as presented herein.
just a proxy of a soil profile, a simple representation of a
complex presence. Although it seems that one method
outperforms the other, it does not guarantee the reliability of
6. REFERENCES Castellaro, S., Mulargia, F., Luigi Rossi, P., (2008) “Vs30: Proxy
for Seismic Amplification?”, Seismological Research Letters,
Borcherdt, R.D., (1994) “Estimates of Site-Dependent Response Vol. 79, pp. 540-543.
Spectra for Design (Methodology and Justification)”, Yogatama, B.A. and Lengkeek, H.J., (2017) “Comparison of several
Earthquake Spectra, Vol. 10, No. 4, pp. 617-653. non-linear soil models in a ground response analysis for a
Cadet, H., Bard, P-Y., Rodriguez-Marek, A., (2012) “Site effect typical sand and clay profile”, Proceedings of the 3rd
assessment using KiK-net data: Part 1 - A simple correction Performance Based Design in Earthquake Geotechnical
procedure for surface/downhole spectral ratios”, Bulletin of Engineering Conference, Vancouver, Canada.
Earthquake Engineering, Vol. 10, pp. 421-448. Zhao, J.X. and Xu, H., (2013) “A Comparison of VS30 and Site
Cadet, H., Bard, P-Y., Duval, A-M., Bertrand, E., (2012) “Site Period as Site-Effect Parameters in Response Spectral
effect assessment using KiK-net data: Part 2 - site Ground-Motion Prediction Equations”, Bulletin of the
amplification prediction equation based on f0 and VSZ”, Seismological Society of America, Vol. 103, No. 1, pp. 1-18.
Bulletin of Earthquake Engineering, Vol. 10, pp. 451-489.
306
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: The seismic response of deep soil surface layers is of concern in developing the seismic microzonation in Jakarta. The present
studies show that the thick sedimentary layer in Jakarta may locally up to 750 m, which means the effect of bedrock depth on site
characterization for this typical of site is importance. To develop design response spectra for structures, the present building codes are now
based on short-period (0.2 sec) and long-period (1.0 sec) spectral accelerations instead of basing design spectra on the effective peak
acceleration (EPA) and effective peak velocity (EPV). For differences in this basing design spectra, the corner period (Tc) of the response
spectra provides the most stable information on the frequency contents of the ground motion. From our study, for 2475 years design response
spectra, the general tendency appears to consist in larger values of Tc (Tc > 1.0 sec) in the several parts of Jakarta, in particular in the eastern
region. These results are based on 1-D site response analysis performed at 383 borehole sites. The modified earthquake acceleration time
histories were adopted in this study.
Keywords: Deep soil layer; site response analysis; design response spectra; corner period.
1. INTRODUCTION 2. METODOLOGY
To construct the smooth response spectra for design purposes, the For estimating ground acceleration at ground surface in Jakarta, the
use of a constant Sa for short periods of response, constant Sv for site response analysis is performed. One of the most important steps
the mid range, and constant Sd for long period response had been in the analysis is site characterization. The average of standard
compensated for the peaks, valleys, and shape variations in actual penetrate blow count numbers (N) value up to 30 metres is a
response spectra (Newmark and Hall, 1982). In earlier, ATC-3-06 common parameter for classifying sites. The classifications
proposed the effective peak acceleration (EPA) and effective peak according to the Indonesian seismic building codes (SNI-1726-
velocity (EPV) to account the differences in the high and low 2012) show that most of location in Jakarta (based on 383
geotechnical boring sites) is classified as the soft soil site (SE) and
frequency characteristics of ground motions to be used in
medium soil (SD) with average N value less than 15 and ranging
developing of the design response spectra for structures. In present
from 15 to 30 respectively (Figure 2).
time, the building codes are now based on short-period (0.2 sec) and
long-period (1.0 sec) spectral accelerations. For example, SD1 and
SDS are the spectral response design values adopted in FEMA 302.
The value of corner period (Tc) and the spectral acceleration value
(SA) in mid range of design spectrum shall be as given by the
following equation:
Tc = SD1/SDs (1)
(a)
-6.1
-6.15
-6.2
Tc
-6.25
-6.35
Then, the curve will be constructed from the parameter above.
The equation force the Sd1 value to be lower than Sds value. So, Tc 106.7 106.75 106.8 106.85 106.9 106.95
value will not greater than 1.0. From our study, for 2475 years (b)
design response spectra, the general tendency appears to consist in Figure 2 The distribution of borehole points in Jakarta and the
larger values of Tc (Tc > 1.0 sec) in the several parts of Jakarta. contour of site classification based on N values (SE site = triangle
These results are based on 1-D site response analysis performed at symbol or grey colour; SD site = square symbol or black colour)
the borehole sites.
307
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
0.35
Vs Profile for All Borehole Sites 0.3
Vs (m/s) 0.25
0.2
0 200 400 600 800
0.15
0
cm/sec
40
Dep th (m)
65
55
52.5
60 50
47.5
45
42.5
80 40
2.2
using estimated Vs along boreholes. All of the boreholes failed to
2
reach the engineering bedrock. Therefore a layer representing 9315000
1.8
engineering bedrock deposits was included into each shear wave
1.6
velocity profile, with Vs of 760 m/s. In this case, extrapolation of Vs 9310000 1.4
values from end of boreholes to layer representing engineering 1.2
bedrock has to be done. The base of the soil columns constructed 9305000 1
down between 350 to 750 m depth as information from the depth to 0.8
308
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5. REFERENCES
Applied Technology Council (1978) “Tentative Provisions for the
Development of Seismic Regulations for Buildings,” ATC 3-
06, NSF 78-8, Applied Technology Council, Redwood City.
Bardet, J.P. & Tobita, T., (2001) NERA, A Computer Program for
Nonlinear Earthquake site Response Analyses of Layered
Soil Deposits, Departement of Civil Engineering USC, Los
Angeles.
Federal Energy Management Agency (1997) "FEMA 302- NEHRP
Recommended Provisions for Seismic Regulations for New
Buildings and Other Structures.
Imai, T. & Tonouchi, K., (1982) Correlation of N-value with S-
wave Velocity and Shear Modulus, Proceeding of The 2nd
European Symposium of Penetration Testing, Amsterdam,
57-72.
Newmark, N.M., and Hall, W.J., (1982) “Earthquake Spectra and
Design”, Earthquake Engineering Research Institute,
Oakland, California, U.S.A.
Ohta, Y. & Goto, N., (1978) Empirical Shear Wave Velocity
Equations In Term Of Characteristic Soil Indexs, Earthquake
Engineering Structural Dyanamics, 6(2),167-187.SNI-1726-
2012, (2012) Tata Cara Perencanaan Ketahanan Gempa untuk
Struktur Bangunan Gedung dan Non Gedung, Badan
Standarisasi Nasional, Jakarta.
309
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Analysis of Bedrock Synthetic Ground Motion on Bandung City using PSHA Method
Arifan Jaya Syahbana1, Anggun Mayang Sari2, Eko Soebowo2, Masyhur Irsyam3, Muhammad Asrurifak4, Hendriyawan5
1
Student of Doctoral Program of Civil Engineering, FTSL ITB, Bandung, Indonesia
2
Researchers LIPI, Bandung, Indonesia
3
Professor of Civil Engineering, FTSL ITB, Bandung, Indonesia
4
Researcher of PPMB ITB, Bandung, Indonesia
5
Marine Engineering Lecturer, FTSL ITB, Bandung, Indonesia
E-mail: arifanjaya@s.itb.ac.id
ABSTRACT: It is a necessity to calculate ground motion in large and dense urban areas such as Bandung City. In order to calculate the
threat of earthquake hazard and risk both primary analysis (for soil or structural dynamic stability) and secondary analysis (for liquefaction)
requires bedrock ground motion. The synthetic ground motion could be an alternative method if the recording data is not attained yet. This
study used Probabilistic Seismic Hazard Analysis (PSHA) method to generate the synthetic ground motion on the bedrock. This study used
earthquake sources in the radius area of 500 km from the survey location, using the latest earthquake data in order to obtain results closer to
real condition. These ground motions were useful for identifying site soil response and mitigation. The results of peak ground acceleration at
the bedrock are 0,502g; 0,505g; 0,510g and 0,539g for BH-1, BH-2, BH-6, and BH-7 site, which is similar to the analysis result of
Indonesian Earthquake Source and Hazard Map 2017 at 2475 years return period.
Keywords: bedrock synthetic ground motion, PSHA, Bandung City
1. INTRODUCTION Seismic sources used in this study are included all mechanism
(shallow crustal and megathrust) within 500 km from studied
1.1 Hazard Analysis location (Table 1 and Table 2). These sources are obtained from
Analysis of ground motion for large and dense urban areas such many catalogue (USGS, Engdahl, BMKG, and PusGen).
as Bandung City is very needed. The purpose of calculating the Table 1. Seismic Source for Megathrust
earthquake hazard is to determine the primary hazard (earthquake
shock) and secondary hazards (phenomenon). In seismic analysis for W M
structural building, sometimes ground motion is required in the L b- a-
Structure Segment
dynamic analysis (Argyroudis & Pitilakis, 2012; Azzam, 2014; (km) (km) PSHA val val
Houshmandan, Rasouli, & Hajiamiri, 2014; Sugianti, Sari, & Sumatran
Syahbana, 2015; Syahbana & Sugianti, 2013; Xiaodong, Weiyuan, Selat Sunda 290 100 8.7 1.15 5.99
Megathrust
& Ruoqiong, 2001). Problems arise if there is no data recorded
Sunda West-Central
ground motion at that location. Thus, an approach is needed to 700 150 8.7 1.08 5.55
Megathrust Java'
obtain the ground motion, which is referred to synthetics ground
Sunda
motion. East Java 280 150 8.7 1.08 5.63
Megathrust
In this study, earthquake hazard in bedrock will be analyzed
using PSHA approach. It is compulsory to gain earthquake data with
a radius of 500 km from the study location to get spectral curves Table 2. Seismic Source for Shallow Crustal
from various sources (Uniform Hazard Spectra/UHS). The next
L W M
analysis is deaggregation to determine the dominant earthquake Segment Type Dip
(km) (km) Max
source to obtain the response spectra. The next step is generating
spectral matching from UHS and RS, and the last is running spectral Kumering-
SS ~90 111 20 7.5
matching using ground motion database. The result is synthetics North
ground motion. Kumering-
SS ~90 60 20 7.1
South
1.2 EZ-FRISK Semangko
SS ~90 90 20 7.4
EZ-FRISK ™ is one of earthquake analysis software that contains Barat-A
several key capabilities which are Seismic Hazard Analysis, Semangko
SS ~90 80 20 7.3
Spectral Matching, and Site Response Analysis. This capability Barat-B
allows various seismic hazard issues to be solved with direct input Semangko
SS ~90 12 20 6.5
specifications using a graphical user interface. EZ-FRISK ™ is Timur-A
designed to be a user-friendly tool for beginners who often interact Semangko
SS ~90 35 20 6.9
with seismic issues. This allows seismic risk analysts focused on Timur-B
identifying key inputs and making decisions that affect earthquake Semangko
Normal 60 50 20 6.5
hazard evaluation, rather than in routine preparing input files, Graben
running command line programs, and plotting out the calculated Ujung
SS ~90 80 20 7.3
results. EZ-FRISK ™ helps analysts make a better risk mitigation Kulon A
planning and decisions in dealing with earthquake threats (Risk Ujung
Engineering, 2011). SS ~90 150 20 7.6
Kulon B
Cimandiri R E-W 23 23 6.7
2. PROBABILISTIC SEISMIC HAZARD ANALYSIS
(PSHA) Nyalindung-
R 30 15 6.5
Cibeber
2.1 Seismic Sources Rajamandala SS 45 22.5 6.6
310
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Lembang SS E-W 29.5 29.5 6.8 and unbranched sandstone with Holocene age. This formation has a
fascinating relationship with the upper Cibeureum Formation.
Subang R NW 33 16.5 6.6 Cikapundung Formation is the oldest rock unit exposed in the study
Cirebon-1 R NW 15 7.5 6.5 area (Koesoemadinata & Hartono, 1981), and consists of a
conglomerate and compact breccia, tuff, and andesite lava. The age
Cirebon--2 R NW 18 9 6.5 of this formation is thought to be early Plistocene. The compactness
Karang of the lithology of this formation can be used as one of the
R EW 22 22 6.7
Malang distinctions with Cibeureum Formation, as well as the basis for
Brebes R 22 11 6.5 determining this formation as bedrock in this area. According to
(Silitonga, 1973), this formation is equivalent to Qvu. In addition to
Tegal R ENE 15 15 6.5 this formation, based on its lithological properties, Qvl, Qvb, Qob,
Pekalongan R NE 16 16 6.6 and Qyl can be included as bedrock. Other units that form the
bedrock are Quaternary volcanic rocks (except Cibeureum
Weleri R 17 17 6.6 Formation and Cikapundung Formation), Tertiary volcanic rocks,
Semarang R EW 34 17 6.6 Tertiary sedimentary rocks, and breakthrough rocks encompassed in
geologic maps composed by Alzwar et al. (1992) and Sudjatmiko
Rawapening R NW 18 9 6.5 (1972). Based on the gravity data on the basin of Bandung reflect
Demak R EW 31 15.5 6.6 basement / basement configuration, it is seen that graben structure
pattern of low and high that describe has happened deformation
Purwodadi R EW 38 19 6.7 process which is controlled by the appearance of structure of
Cepu R ESE 100 50 7.1 alignment of northwest trending structure -southeast and northeast-
southwest. Lower Zone such as in Dayeukolot, Cimahi, Rancaekeh
and altitude in Gedebage, Banjaran, Cililin. Water drilling data and
geotechnics drilling by Soebowo et al. (2016) and Tohari et al.
2.2 Attenuation (2015) show that bedrocks are found at depths -50 to -80 m in the
eastern part of Gedebage, apparently in some locations may be more
Attenuation used to calculate fault/shallow crustal mechanism are than 90 meters. The seismicity recorded in the Bandung Basin in
Bore-Atkinson (2008) NGA USGS 2008, Campbell-Bozorgnia 2011 by Sulaeman & Hidayati (2011) is 3.4 SR. In addition, the
(2008) NGA USGS 2008 and Chiou-Youngs (2007) NGA USGS history of seismicity in the Bandung Basin was also observed in
2008. Then, atttenuation for subduction mechanism is Zao et all 1999, 2000, 2003, 2004, 2005 (Marjiyono, Soehaimi, & Kamawan,
(2006) USGS 2008, Atkinson-Bore (2003) Worldwide Subduction 2008).
USGS 2008 and Young (1997) Subduction Rock All.
The study area consists of four locations in Bandung area. This
2.3 Site Location location is a point that belongs to the eastern area of Bandung and
still lies in the basin of Bandung. The point coordinates to be
Bandung basin plains represent depression in West Java that shows reviewed can be observed in Table 3 and can be seen in Figure 1.
the basin between the mountains (van Bemmelen, 1949 in Dam,
1994). The basin is surrounded by Quarter volcano cones and lava Table 3. Coordinates of Site Location
domes that are characterized by active volcanic activity. Basin
Location Latitude Longitude
Bandung has a length of approximately 60 km starting from
Padalarang-Nagrek and width 40 km from Lembang-south Soreang. BH-1 -6.981917 107.700945
This height is between 650 m to 2200 m. The Bandung Basin is
composed of lake sediments, lake fans, swamps, alluvial fans, BH-2 -7.01091 107.725492
floodplains, paleosoils of the Quaternary, in the middle of the basin BH-6 -6.995883 107.693052
there are four runways of lake clay with layers of volcaniclastic and
fluvial facies (Dam, 1994). These sediment conditions are soft and BH-7 -6.981313 107.65068
consolidated. The Bandung basin plain is a historical record of
environmental change as it grows in continuous physical changes
such as changes in river flow, flooding, degradation and
neotectonics on different time scales at the time to illustrate some
types of sedimentary facies, where the condition of geological
structures is a determinant of the sedimentation process (Dam, 1994).
This process is characterized by the looping of the silt-sand unit
(flood plains), the clay of lakes and gravel (alluvial fans) with
varying thickness and geometric shapes in several locations
mutually paying and slightly decreasing and rising and cutting.
Based on the geological history of the Bandung Basin area in the
Quaternary (Dam, 1994), there are several volcanic fans sourced
from Mount Malabar and Wayang Mountain Complex in the south
and Manglayang Mountain in the north, there are also 4 phases of
material that fill the ancient lake basin of Bandung. Formation
Cibeureum lithology compilers of volcanic fans with lithology tufa
sand, lapili, breccia, lava and aglomerate (Qyu) and the results of
old volcano collapse (Qc) mainly found in the north and south
(Silitonga, 1973). Kosambi Formation proposed by Koesoemadinata
& Hartono (1981) to substitute the name of the Lake Deposit used
by Silitonga (1973). Kosambi Formation is also used for the name of
Lake Deposits used by Alzwar, Akbar, & Bachri (1992) and
Sudjatmiko (1972). The distribution of this formation on the surface
is in the middle. Its lithology consists mainly of claystone, limestone,
311
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3. DEAGGREGATION
Deagregation is the process of determining magnitude (M) and
distance (R) which is dominantly based on the result of total
probability which gives the biggest hazard contribution to the
location reviewed in the time of earthquake request and period of
building structure. The distance and magnitude are determined
based on the relationship between magnitude, distance and rate
(incidence/year over the acceleration being reviewed) to then
searched for emphasis on the 3D deaggregation curve (Asrurifak,
2010; Irsyam, 2012). The equations used can be seen in Equations 1
and 2 (Sunardi, 2015).
(1)
(2)
The pair of one magnitude (M) and the distance from the site to
the dominant source (R) and the hazard due to the earthquake can be
expressed in one function. This concept obtains on seismic
deaggregation (McGuire, 2001). Deaggregation can provide an
overview of earthquake magnitude and distance for a particular
earthquake source, which is likely to have a major effect on the site.
One deaggregation purpose is in the dynamic analysis of a building
with a large number of floors. In the dynamic analysis of such high-
rise buildings, it is not possible to use PSHA's ground motion input,
for which the PSHA result mechanism is a combination of various
earthquakes mechanism. For that, it is needed separation of
earthquakes that represent each mechanism. Referring to the
Earthquake Indonesian National Standard 2012, it is required to
simulate the earthquake at least 5 pieces of ground motion input
representing all existing mechanisms (Anonim, 2012).
In deaggregation analysis, the selected point is PGA period
(period close to 0 sec) and return period of 2475 years with damping
5%. Analysis result using Ez Frisk software can be observed in Figure 2. The result of deaggregation at BH-1, BH-2, BH-6
Figure 2. and BH-7 site.
Once the result of deaggregation at certain location obtained,
then it is necessary to define what earthquake mechanisms affected
the hazard predominantly. Here the mechanism could be determined
by observing the graph relationship between hazard and PGA
curved (often referred to as hazard curve). The function of this graph
is to obtain earthquake spectra with modification of Uniform Hazard
Curve. This graph of hazard and PGA relationships can be observed
in Figure 3.
312
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
313
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 6. Peak acceleration (PGA) at 2475 year old basalt rock (2%
exceeded in 50 years) from PusGen (2017)
6. ACKNOWLEDGMENTS
The authors would like to thank the Indonesian Endowment Funds
of Education (LPDP) for the scholarship being granted to carry out
doctoral studies in ITB. Acknowledgments are also addressed to the
LIPI research team of the Center for Geotechnology Research on
data and cooperation in this paper. Last and not least, the authors say
many thanks to the promoter for the science and guidance that has
been given and to others who helped this project whom cannot be
mentioned one by one.
7. REFERENCES
Alzwar, M., Akbar, N., & Bachri, S. (1992). Peta Geologi Lembar
Garut dan Pameungpeuk, skala 1: 100.000. Pusat Penelitian
Dan Pengembangan Geologi, ESDM, Bandung.
Anonim, A. (2012). SNI 1726:2012 Tata Cara Perencanaan
Ketahanan Gempa untuk Struktur Bangunan Gedung dan Non
Gedung, (1), 149.
Argyroudis, S. A., & Pitilakis, K. D. (2012). Seismic fragility curves
of shallow tunnels in alluvial deposits. Soil Dynamics and
Earthquake Engineering, 35, 1–12.
Asrurifak, M. (2010). Peta Respon Spektra Indonesia Untuk
Perencanaan Struktur Bangunan Tahan Gempa Dengan
Model Sumber Gempa Tiga Dimensi Dalam Analisis
Figure 5. Ground motion at locations at BH-1, BH-2, BH-6 Probabilitas. Disertasi Doktor Teknik Sipil ITB, Indonesia.
and BH-7 site. Azzam, W. R. (2014). Seismic Response of bucket Foundation and
The value of PGA produced in this study is valid compare to the structure under earthquake loading. Electronic Journal of
result PusGen Research Team on Indonesia Earthquake Source and Geotechnical Engineering, 19, 1477–1489.
Hazard Book of 2017 for the area around Bandung City which is Dam, M. A. C. (1994). The Late Quaternary Evolution of the
worth between 0,5-0,6g as presented in Figure 6. Thus, PGA value Bandung Basin, West-Java, Indonesia. Thesis Vrije
into the top of the spectra in each synthetic ground motion can be Universiteit, Amsterdam, 252.
used as input on the analysis. Houshmandan, H., Rasouli, R., & Hajiamiri, M. (2014). The Effect
of Near Fault and Far Fault Earthquake on Dynamic Analysis
of Concrete Face Rockfill Dam with Case Study on the Neka,
Gelevard CFRD. Journal of American Science, 10(1s).
Hutasoit, L. M. (2009). Kondisi Permukaan Air Tanah dengan dan
314
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
315
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This study focuses on the effect of “cyclic” behaviour of pile installation methods on the penetration resistance and bearing
capacity. A series of laboratory model test was conducted to investigate the cyclic effect by comparing three kinds of piling methods;
monotonic jack-in, pseudo-dynamic push-in and pull-out, i.e. “surging” and vibratory driving in dry or saturated sand grounds. It was found
that surging or vibratory pile driving decreased the pile penetration resistance due to the negative dilation caused by the cyclic shearing of the
soil surrounding the pile. The static load test showed that surging and vibratory pile driving provided the same or larger pile head load than
jack-in method. Furthermore, the fluctuation of the pore water pressure strongly indicated the change of dilation. It was seen that surging and
vibratory pile driving prevent the positive dilation than jack-in method due to the difference of cyclic shearing and monotonic loading.
Keywords: Pile installation method, jack-in, surging, vibratory pile driving, pore water pressure.
1. INTRODUCTION
70 SG1
Recently, it is rather difficult to employ the impact hammer method
in urban areas and residential area, because it causes noises and 175
ground vibrations. Vibratory pile driving and rotary pile installation
(jack-in) methods have been developed as the alternative methods,
SG2
which can install the pile with less noises and ground vibrations.
Previous studies about these methods mainly focused on the soil 100
resistance during pile penetration. In contrast, only a few studies SG3 595mm
focused on the bearing capacity of the pile installed by these 100
methods. If the bearing capacity is clarified, these methods will SG4
become more available.
100
The purpose of this study is to investigate influences of the
different piling methods on the pile performance, such as SG5
penetration resistance, and bearing capacity and load-settlement 50
relation during the pile load test. Significant factor is “cyclic”
movement during pile penetration. Moriyasu et al. (2016) compared strain gauge
the pile behaviour between the monotonic jack-in and surging in dry (a) location of strain gauges (b) photo of the pile
model sand. Surging refers to the cyclic push-in and pull-out Figure 1 Model pile
movement of a pile during installation. Cyclic shearing by surging
caused negative dilation around the pile shaft and soil compaction 2.2 Model ground
below the pile tip.
Thus, this paper focused on the influence of the dynamic surging The material of the model ground was Silica sand #6. Table 1 lists
on the pile penetration resistance and bearing capacity by comparing the physical properties of the sand. The model ground was prepared
to with the influence the pseudo-dynamic surging. A series of in a rigid cylindrical soil container having a diameter of 566 mm and
laboratory tests were conducted in this paper. The laboratory tests a height of 580 mm. The model ground was prepared with 11 soil
included three kinds of piling method (monotonic jack-in, surging layers of 50 mm thick and one layer of 30 mm thick. The sand for
and vibratory pile driving) in two sand conditions (dry or saturated each soil layer was put into the soil box, and the sand layer was
ground). compacted by hand tamping to have a designated relative density, Dr
= 80%. The procedure was repeated to complete the model ground
2. EXPERIMENT DESCRIPTION of 530 mm high.
For the preparation of saturated ground, the soil box was filled
2.1 Model pile with water. Then the sand was poured into the soil box and
compacted, similarly to the case of the dry ground. Although Dr =
An aluminium pipe pile with an outer diameter of 32 mm, a wall
80% was intended also for the saturated ground, Dr = 70% was
thickness of 1.3 mm and a length of 595 mm was used for model
achieved in the saturated ground.
piles. Regarding the pile tip condition, six cases were open-ended
and one case was close-ended. In order to obtain the distribution of
Table 1 Physical properties of the sand
axial forces, strain gauges were attached to the pile shaft at 6
different levels, two strain gauges in the opposite faces at each level, Soil particle density, s (t/m3) 2.679
as shown in Figure 1(a). Minimum dry density, dmin (t/m3) 1.366
The pile surface was coated with an acrylic adhesive to protect Maximum dry density, dmax (t/m3) 1.629
the strain gauges, and glued with silica sand to increase the pile Maximum void ratio, emax 0.962
shaft resistance (Figure 1(b)). The sand was the same material as the Minimum void ratio, emin 0.645
model ground. Mean particle size, D50 (mm) 0.52
316
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
2.3 Experimental procedure and cases For each ground condition, three kinds of penetration methods
were employed. Jack-in method in Case 1 and Case 5 is the
Figures 2 and 3 show the experimental apparatuses. During pile monotonic pile installation at a penetration rate of 0.2 mm/s.
penetration tests (PPT), the pile was installed by means of a motor Surging method in Case 2 and Case 6 means the cyclic installation
jacks or a vibratory hammer. When the pile was penetrated to a with repetition of 2 mm jack-in and 1 mm pull-out. In these methods,
depth of 400 mm, PPT was finished. Subsequently, during static the pile was installed by using the jack without pile rotation. In the
load test (SLT), the pile head was pushed in two or three loading vibration method in Cases 3, 4 and 7, a vibratory hammer model
cycles. At the end of SLT, the pile extraction force was measured by was employed. As shown in Figure 3, the hammer has a tandem
extracting the pile slightly. Finally, Cone penetration test (CPT) was rotating eccentric weights which generates vertical vibrations and
carried out to investigate the soil strength profile of the model cancels horizontal vibrations. The vibratory hammer has a weight of
ground. 300 N and a maximum frequency of 60 Hz. The close-ended pile
Table 2 shows the test cases. Case 1 to Case 4 were carried out was used in Case 4, while the open-ended pile was employed in the
in the dry ground, while Case 5 to Case 7 were carried out in the other cases.
saturated ground. SLT was carried out subsequent to PPT. In all cases, the pile
penetration rate was 0.1 mm/s.
Screw jack
Vibratory
hammer
Pile
Sand box
Static Static
Static loading
loading loading
by V.H. *4
by V.H. *4 by V.H. *4
weight
PPT*1 PPT*1 weight weight PPT*1 PPT*1
↓
↓ ↓ ↓ ↓ ↓ ↓
Test Vibratory
SLT*2 SLT*2 Vibratory Vibratory SLT*2 SLT*2
sequence penetration
↓ ↓ penetration penetration ↓ ↓
↓
CPT*3 CPT*3 ↓ ↓ CPT*3 CPT*3
SLT*2
SLT*2 SLT*2
↓
↓ ↓
CPT*3
CPT*3 CPT*3
*1 PPT: Pile Penetration Test, *2 SLT: Static Loading Test, *3 CPT: Cone Penetration Test, *4 V.H.: Vibratory Hammer
2.4 Measurements axial force was estimated from the strain measured near the pile
head (i.e., SG level 1 in Figure 1(a)). Furthermore, a pair of
During PPT and SLT, the pile head load, pile head displacement and acceleration transducers was attached near the pile head. In the
pile strains were measured. In the cases of jack-in and surging, the saturated sand cases, pore water pressures were measured by the
pile head load was measured by a load cell set between the pile head transducers located as shown in Figure 4.
and motor jack. On the other hand, in the cases of vibration, the
317
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
PWP3 PWP3
545 mm
540 mm
50 50 50
PWP4 PWP5 PWP4 PWP4
PWP3 PWP3 PWP3
100 100 100
PWP6 PWP5 PWP5
100 100 100
Coupler PWP7 Coupler PWP7,PWP6 Coupler PWP7,PWP6
50 50 50
80 Monotonic
600 Cyclic
60
400 r' = 30 kPa
r' = 30 kPa 40
200 Monotonic
20
Cyclic
0 0
0 2 4 6 8 10 12 14 16 18 20 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Axial strain, a (%) Axial strain, a (%)
(a) General view (b) Enlarged view
Figure 5 Relationship between the axial strain, a,, and the devlatoric stress, q
-150 5
-200 0
0 1 2 3 4 5 6 7 8 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Axial strain, a (%) Axial strain, a (%)
(a) General view (b) Enlarged view
Figure 6 Relationship between the axial strain, a, and the excess pore water pressure, ue
2.5 Monotonic and cyclic triaxial shear tests of the sand stress paths. The confining pressure of these tests was about 30 kPa.
The internal frictional angle was 38 to 40 degrees, and the shift
In order to investigate the mechanical property of the sand, a phase angle was 28 degrees.
monotonic and a cyclic consolidated undrained (CU) shear tests of
During cyclic loading, it is seen from Figure 5 that a increased
the sand were carried out.
with accumulation of the number of cycle loading with each
Figure 5 shows the relationship between the axial strain, a, and constant q. At the same time, cyclic loading generated excess pore
the deviatoric stress, q. Figure 6 shows the relationship between a water pressure higher than that in monotonic loading. At the
and the excess pore water pressure, ue. Figure 7 shows the effective monotonic loading stage after the cyclic loading, the increment of ue
318
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
60
5&7
50 500
Case 6
40
F.L.
600
Figure 8 Vertical effective stress
30
CPT tip resist, tip (kPa)
20 r'=30kPa 0 1000 2000 3000 4000
0
10 (Dry)
100 (Saturated)
5
3.2 Results of the pile penetration tests
Figure 10 shows the time history of the accelerations of the pile 0
head, the ground surface and the vibratory hammer during the
vibratory pile penetration in Case 4. The downward acceleration is
taken as positive. As shown in the figure, the acceleration of pile -5
20H z 25H z 30H z
head was much larger than those of the ground surface and the
vibratory hammer. The frequency of the vibratory hammer was -1 0 35H z
changed from 20 Hz to 35 Hz to enhance the pile penetration ability
of the vibratory hammer during PPT. Figure 10 Acceleration during vibratory penetration
Figure 11 shows the relationship between the pile head load, Ph,
and pile head displacement, wh, in Cases 1 to 4 (dry ground). It can Figure 13 shows the relationship between the pore water
be seen that Ph in Case 3 and Case 4 during PPT was almost similar. pressure, p, and pile head displacement, wh, during PPT. In Case 7, p
In Cases 3 and 4, at first, the pile was installed by the self-weight of increased with increases in wh, and p increased dramatically when f
the vibratory hammer from the ground surface. When the pile could was increased to 18.3 Hz at wh = 400 mm. It is seen that soil
not be penetrated by the self-weight at wh = 75 mm, the operation of liquefaction was caused, because the excess pore water pressure
the vibratory hammer was started. Until the pile reached wh = 400
corresponded to the effective vertical stress, v, shown in Figure 8.
mm, the frequency of the hammer was increased if the pile
Therefore, Ph decreased with the reduction of v related to the soil
penetration was degraded. Ph in Case 3 and Case 4 (vibration) was
liquefaction. As mentioned earlier, the CU test shows that cyclic
smaller than that of Case 1 and Case 2 after wh exceeded 200 mm.
shearing generated higher excess pore water pressure than
Figure 12 shows the relationship between Ph and wh in Cases 5
monotonic loading, as shown in Figure 6. If the degree of the excess
to 7 (saturated ground). During PPT, Ph in Case 6 (surging) was
pore water pressure depends on the number of cyclic shearing, the
slightly smaller than that in Case 5 (jack-in). In Case 7 (vibration),
vibration provided a large number of cyclic shearing (3800 times)
Ph decreased significantly, when f was increased to 18.3Hz at wh =
than surging (350 times).
400 mm.
319
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
200 200
300 300
400 400
500 500
(a) Case 1 (Dry, Push-in) (a) Case 5 (Saturated, Push-in)
SLT
100
200
200
300
300
400
400
500
(b) Case 2 (Dry, Surging) 500
(b) Case 6 (Saturated, Surging)
Pile head load, Ph (N)
-1000 0 1000 2000 3000 4000 5000
0 Pile head load, Ph (N)
Static penetration
-1000 0 1000 2000 3000 4000 5000
Vibration 0
Pile head disp., wh (mm)
29 Hz (n = 435) SLT
100
29.6 Hz (n = 858)
200
33.5 Hz
(n = 720) 200
300
34.5 Hz 15.5 Hz (n = 1603)
(n = 1397) 17.1 Hz (n = 982)
35.3 Hz
300
400 (n = 918)
17.7 Hz (n = 1047)
320
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 14 (a) shows the comparison of load-settlement relations As Bolton et al. (2013) pointed out, the cavity created below the
between Case 1 (jack-in) and Case 2 (surging) in dry sand condition. pile tip collapses during the upward movement of the pile, and the
It should be noted that the origin of vertical axis of Figure 14 (i.e, soil below the pile tip is compacted during the following downward
pile head displacement during SLT, wh_SLT) is the end of wh during movement. This cyclic compaction enhanced the pile base resistance
PPT due to adjust the difference end of wh among cases. #1 in each largely.
figure means the yield point in the 1st loading cycle, and #2 is the Figure 14 (b) shows the comparison of load-settlement relations
end point of increasing load. As shown in Figure 14 (a), the initial among Case 1 (jack-in), Case 3 and Case 4 (vibration). The yield
pile head stiffness in Case 2 (surging) is higher than that in Case 1 loads (#1) in Case 3 and Case 4 (vibration) are larger than that in
(jack-in), and the yield point (#1) in Case 2 is larger than that in Case 1. However, a softening behaviour occurred: i.e., Ph increased
Case 1. Figure 15 shows the change of pile shaft resistance and pile to 3878 N at #1, and dropped immediately to Ph = 3604 N with a
base resistance from #1 to #2. The pile base resistance, Pb, is the little increasing of wh_SLT. It can be seen from Figure 15 (a) that the
axial force obtained from strain SG5, and the pile shaft resistance is pile shaft resistance in Case 4 decreased with increasing wh_SLT
the difference between the pile head load, Ph, and the pile base (from #1 to #2), while the pile base resistance increased. This is not
resistance. The pile base resistance in Case 2 (surging) is higher than related with soil plugging because the behaviour was occurred both
that in Case 1. This is similar to Moriyasu et al. (2016). condition of open-ended pile (Case 3) and close-ended pile (Case 4).
100 100
300
200
200
350
300
300 400
400
400 450
500
PWP1 PWP2 PWP3 PWP1 PWP2 PWP3 PWP4 PWP1 PWP2 PWP3 PWP4
PWP4 PWP5 PWP7 PWP5 PWP6 PWP7 PWP5 PWP6 PWP7
(a) Case 5(Jack-in) (b) Case 6 (Surging) (c) Case 7 (Vibration)
Figure 13 Relationship between the pile head displacement, wh, and the water pressure (P.W.P.), p, during PPT
Pile head load, Ph (N) Pile head load, Ph (N) Pile head load, Ph (N)
Pile head disp. during SLT, wh_SLT (mm)
-1000 0 1000 2000 3000 4000 5000 -1000 0 1000 2000 3000 4000 5000 -500 0 500 1000 1500
0 #1 0 #1 0 #1
#1
#1 #2 #1 #1 #2
#1
10 10 #2 10
#3
20 #2 20 20 #2 #2
30 30 30
#2 #2
40 40 40
50 50 50
60 60 60
case1(Push-in)
case1(Push-in) case3(Vibration) case5(Push-in)
70 70 70
case2(Surging) case4(Vibration_C.P.) case6(Surging)
case7(Vibration)
80 80 80
(a) Cases 1 and 2 (dry ground) (b) Cases 1, 3 and 4 (dry ground) (c) Cases 5, 6 and 7 (saturated ground)
Figure 14 Relationship between the pile head load, Ph, and the pile head displacement, wh_SLT
5000 1500
base shaft base shaft
4000
Load (kN)
Load (kN)
1000
3000
2000
500
1000
0 0
#1 #2 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2
Case 1 Case 2 Case 3 Case 4 Case 5 Case 6 Case 7
(Jack-in) (Surging) (Vibration) (Vib._C.P.) (Jack-in) (Surging) (Vibration)
321
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
420 435
322
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: In high seismic areas, liquefaction is a major issue when a designer dealing with loose sand below groud water level and low
fines content. To solve this problem, densification by means of compaction is the most simple approach to be accomplished. Dynamic
compaction technique by repeated impacts of a heavy weight on to the soil can successfully densify the soil and thus, dismiss the liquefaction
issue. The project case described in this paper is the development of New Yogyakarta International Airport, which is located in Kulon Progo,
Indonesia. It is located in high seismic area with Peak Ground Acceleration (PGA) of 0.4g. Sand with fines content ranging from 0.25% to
17.23% was found all over the project site. Furthermore, the design criteria was stated to mitigate the liquefaction during the designed
earthquake. The evaluation of project achievement to mitigate the liquefaction was performed based on In-Situ Testing Parameter; in this
case is Cone Penetration Test. This paper presents a brief discussion of the entire project case which is using Dynamic Compaction method
to mitigate the liquefaction issue.
Keywords: Ground improvement, dynamic compaction, liquefaction mitigation, high seismic area, cone penetration test, new airport.
323
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
W ∙H ∙N ∙P
AE =
G2
(2)
AE = applied energy
N = number of drops at specific drop point
W = mass of tamper in tons
H = drop height in meters
P = number of passes
G = grid spacing in meters
Figure 2 Project location
More recently, Varaksin and Racinais (2009) have proposed Eq.
(3) which considers the degree of improvement as a function of
depth.
݂2 - ݂1
f(z) = (ݖ− ܰ)ܮܩ2 + ݂1
ܦ2
(3)
where,
f(z) = improvement ratio at elevation (z)
z = elevation in meters
NGL = natural ground level
D = depth of influence of dynamic compaction
f1 = maximum improvement ratio at surface Figure 3 Project ground improvement area
f2 = improvement ratio at maximum depth of influence
Source n-values Soil Type The project was located within high seismic area as shown in
Menard & Broise Indonesia seismic map, Figure 4, with Peak Ground Acceleration
1.0 All soils (PGA) value of 0.4g. The design criteria was to mitigate the
(1975)
Leonards, Cutter & liquefaction, with moment magnitude, Mw, of 7.8 and safety factor
0.5 - against liquefaction of 1.2.
Holtz (1980)
Soil with unstable
0.5
structure
Smoltczyk (1983)
0.67 Silts and sands
1.0 Pure frictional soils
Lukas (1980) 0.65-0.8 -
Mayne, Jones, &
0.3-0.8 -
Dumas (1984)
Gambin (1985) 0.5-1.0 -
0.65 Fine sand
Qian (1987) 0.66 Soft clay
0.55 Loess
0.65 Silty sand
Vsn Impe (1989) 0.35 Municipal waste
0.5 Clayey sand
Yee, Setiawan & Calcareous sand/coral
0.5
Baxter (1998) sand
Faisal, Yee &
0.33-0.59 Municipal waste
Varaksin (1997)
324
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
325
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 10 Dynamic compaction rig and pounder on site Figure 12 Ground settlement and heave with distance
326
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
all over the site. The work acceptance flow chart was also
estabilished as presented in Figure 16. The first-rate CPT rig with
capacity of 20 ton and auto-generated CPT result was used to carry
out the tests (Figure 17).
327
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(Figure 19). The 0.5 m thick layer on top is loosened certainly due 4. CONCLUSION
to the mechanism of heaving. The improvement ratio increase until
reach the peak at 3.2 m, so-called optimal depth, then decrease as Dynamic compaction has demonstrated to be a successful technique
the depth getting deepened and the existing qc value getting higher. to densify and mitigate the liquefaction issue in high seismic area.
The improvement ratio is insignificant after the existing qc value This paper has highlighted the effective implementation of dynamic
higher than 25 MPa. The average improvement ratio is compaction to treat liquefiable ground condition for new
approximately 1.8. international airport in Indonesia with compaction area of
approximately 900,000 m2. The calibration test is required to be
carried out prior to dynamic compaction work. Proper quality
assessment and quality control is essential to maintain the accurate
result and evaluation.
5. ACKNOWLEDGEMENTS
Special acknowledgement is made to PT. Pembangunan Perumahan
(Persero) as our client and to PT. Angkasa Pura I (Persero) as our
project owner for their positive response towards the use of non-
conventional solution using dynamic compaction for the new
international airport to mitigate liquefaction problem – first time in
Indonesia.
For the unforgettable work, acknowledgement is also made to
all Menard project team of New Yogyakarta International Airport
which has delivered outstanding efforts to be able to achieve this
work smoothly executed.
6. REFERENCES
Chow, Y.K., Yong, D.M. and Yong, K.Y. (1992) “Dynamic
Compaction of Loose Sand Deposits”, JSSMFE, Soils and
Foundations, Vol.32, No.4, p.93-106.
Figure 18 Comparison of pre-test and post-test CPT Krinitzsky, E.L., Hynes, M.E., (2002) “The Bhuj, India, earthquake:
lessons learned for earthquake safety of dams on alluvium”,
Engineering Geology 66 (3– 4), 163– 196.
Lukas, R.G. (1980) “Densification of Loose Deposits by Pounding”,
ASCE, Journal of Geotechnical Engineering, GT4, p.435-
446.
Mayne, P.W., Jones, J.S. and Dumas, J.C. (1984) “Ground Response
to Dynamic Compaction”, ASCE, Journal of Geotechnical
Engineering, Vol. 110, No.6, p.757-774.
Menard, L. and Broise, Y. (1975) “Theoretical and Practical Aspects
of Dynamic Compaction”, Geotechnique 25, No.1, p.3-17.
Obermeier, S.F., (1998) “Liquefaction evidence for strong
earthquakes of Holocene and latest Pleistocene ages in the
states of Indiana and Illinois, USA”, Engineering Geology 50
(3 – 4), 227– 254.
Schneider, J.A., Mayne, P.W., Rix, G.J., (2001) “Geotechnical site
characterization in the greater Memphis area using cone
penetration tests”, Engineering Geology 62 (1– 3), 169–184.
Tuttle, M.P., Sims, J.D., Dyer-Williams, K., Lafferty, III, R.H.,
Schweig, III, E.S., (2000) “Dating of liquefaction features in
the New Madrid seismic zone”, Technical Report, US
Nuclear Regulatory Commission NUREG/GR-0018, 78 pp.
Youd, T.L., (1991) “Mapping of earthquake-induced liquefaction
for seismic zonation”, Proc., Fourth International Conference
on Seismic Zonation, vol. 1, pp. 111 –147.
Youd, T.L., Idriss, I.M., Andrus, R.D., Arango, I., Castro, G.,
Christian, J.T., Dobry R., Finn, W.D.L, Harder, L.F., Hynes,
M.E., Ishihara, K., Koester, J.P., Liao S.S.C., Marcuson,
W.F., Martin, G.R., Mitchell, J.K., Moriwaki, Y., Power,
M.S., Robertson, P.K., Seed, R.B. and Stokoe, K.H., (2001)
“Liquefaction Resistance of Soils: Summary Report from the
1996 NCEER and 1998 NCEER/NSF Workshops on
Figure 19 Improvement ratio of dynamic compaction works Evaluation of Liquefaction Resistance of Soils.”
328
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Tigadihaji dam located in the mountain range and have only 18 km from the Komering section of the Sumatra Fault Zone,
which known have a high seismic activity, and has already known as an earthquake sources. The seismic hazard assessments reveal that 2
faults located on the downstream of the dam as secondary faults which parallel to the Sumatra Fault Zone. The Pauh and Sukarena faults
named after the nearby villages have a distance of 1.85 and 8 km from the dam site respectively. Since the faults were closed to the dam, a
further detail investigation was carried to reveal the level of activity of the Pauh fault. An electrical resistivity tomography (geo-electric)
method and paleoseismic trenching were carried out to confirm the level of activity. The carbon dating test was also carried out from the
samples taken from the fault zone at the depth of 2,00 m from the ground surface. The result indicated that the Pauh fault considered as an
active fault with low sliprate, and shall be account in the seismic hazard analysis of the dam.
Keywords: Active fault, electrical resistivity tomography, paleoseismic trenching, carbon dating test.
329
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
parts of SFZ where it runs toward the Andaman spreading center elevation map (DEM) and then processing using GIS (Geographic
and the Sunda Strait grabens respectively. Information System) software to obtain detail topographic map
The SFZ map that has been significantly revised is In the which may shows the geologic structures and suspected fault
northern and the southern parts (Natawidjaja, 2017). In the northern location. Tigadihaji dam was built on the Gumai and Air Benakat
part of Sumatra, Renun – Tripa segments of SFZ runs along the west geologic formation which covered by thick Ranau tuff as product of
side of Lake Toba toward north and make a compressive bend mega eruption of ancient Ranau volcano, which at present left
before it merges with the Batee fault that ascending from the west Ranau lake as their huge crater. Figure 2 shows the distribution of
coast line. Previously, in Sieh and Natawidjaja (2000) as well as in Ranau tuff as product of mega eruption of ancient Ranau volcano
the 2010 Indonesian PSHA map, the Batee fault was not taken into (Natawidjaja, 2017).
account since it was previously considered not active. Recent data
and analysis indicates vice versa (Natawidjaja et al., 2017).
3. SEISMICITY CONDITION OF TIGADIHAJI DAM During field survey, we confirmed that these faults are indeed
presence, marked by topographical breaks, fault-water falls and
Geological regional survey was carried out within radius of 50 km, lineaments of scarps and streams. We also conducted GPR surveys
to include the extent of the Pauh and Sukarena faults, while detail two ascertain the locations of the faults. Similar to the Ranau fault,
geological survey was carried out surroundings the Tigadihaji dam however, the geomorphological evidences of fault movements have
site in order to identify the fault strands and the activity of the been subdued and unclear. Luckily, on the main road between
Sukarena and Pauh faults. The basic data in this survey was digital Ranau lake and Muaradua town, we found a fresh cut of road cliffs
330
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
exposing the southern end of Sukarena fault outcrop as shown in Southeast of Ranau lake, Kumering segment of Sumatra Fault
Figure 4 (Natawidjaja et al, 2017). Zone straightly runs through the thick, flat-top Ranau-tuff filling
Liwa basin, marked by linear alignment of Way Robok deep
canyons and major deflections of major river canyons crossing the
fault line. Five major river canyons are right-laterally offset about
the same amount along the single major traces of Kumering fault
indicating they are formed isochronously or nearly so. Three of the
displaced canyons are across the fault at right angle, thus give
excellent measures for fault offset of about 320 to 380 meters, as
shown in Figure 5. Uncertainty of the amount offset are
appoximated by the width of displaced river canyons. We do not
include the other two offsets since the river cut at low angle to the
fault, so they yield less accurate measure of the offsets. The average
of these three river offsets is 350 ± 50 meter, thus yields a sliprate of
10.4 ± 2.4 mm/year for the Kumering fault segment.
331
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
popular following the successfully study of palaeoseismology in Geological observation, geo-electric survey and data evaluation
Pallet Creek, California (Sieh et al, 1989., Sieh, K , 1977). on the paleoseismic trenching as shown on Figure 8, indicated that
Palaeoseismic trench was carried out on the approximate the Pauh fault was still active, even there was not clearly exposed on the
Pauh fault line. Trenching was made over the line 1 of the ground surface. The Ranau tuff layer as a product of mega eruption
geoelectric survey, cut the Pauh fault perpendicularly. The above the Gumai formation indicated a shifting in vertical direction
dimension of the trench are; 20 meter in length, 1.50 in width and (left lateral normal fault). Carbon dating results on the soil samples
2.00 m in depth. Logging on the both (left and right) walls were taken from the soils on the fault location indicated that the age of the
made using different colours to identify the soil layers and fault soils was 114 to 249 yaears. This indicated as the last earthquake
structures. Logging were made to identify colour, texture, type or resulting from the Pauh fault, even there was no historical record
nomenclatures of soils. Trenching in Pauh faults have five (5) was found. The calculation on the Mmax based on the length of the
different soil layers and described as follows: fault based on several GMPE (ground motion prediction equation)
Layer 100 was the first soil layer, identified by black colour was resulting the possibility Mmax 6,4 earthquake may occured from the
top soil consists of sandy silt soils, dark brown in colour, many roots Pauh fault. Based on this result, the earthquake which may occured
found. from Pauh fault shall be included in the Seismic Hazard Analysis for
Layer 200 was the second layer, gravelly sandy silt with some Tigadihaji dam.
lithic fragment up to 4 cm, brown in colour. This founding will change the Seismic Hazard Analysis which
Layer 300 was the third layer, coarse sand from Ranau tuff, with has been made previously. In the previous Seismic Hazard Analysis,
some lithic fragment up to 2 cm, yellowish brown in colour. the Pauh fault was considered as an inactive fault, and the Seismic
Layer 400 was the fourth layer, coarse sand from Ranau tuff with Hazard Analysis were made based on the Probabilistic Seismic
some sand layers interbedded, greyish brown. Hazard Analysis (PSHA). But considering that Pauh fault was an
Layer 500 was the bottom layer, coarse sand from Ranau tuff active fault and the distance to the proposed Tigadihaji dam was
with some sand layers interbedded, grey colour. very closed, and also the length of Pauh fault has already known, the
On the trench walls there is an indication of fracture and fault new Seismic Hazard Analysis was made based on the Deterministic
plane that shear the soil layers with the parallel direction of the Pauh Seismic Hazard Analysis (DSHA).
fault. Although the fault structure did not appear on the surface,
indication of Pauh fault found at least on the soil layer 400 that 5. CONCLUSION
shear the soil layer, so that the existence of possible ttectonic
fracture structures deforming this soil layer is classified as an active The identification of active fault nearby the foot print of the
fault. Tigadihaji has been presented. The presence of the Sumatra Fault
Figure 7 shows the logging of the east and west walls of the Zone, the sliprates along the fault and paleoseismic on the Pauh fault
palaeoseismic trenching, while Figure 8 indicated the Pauh fault in has been carried out to study the condition of the fault. The
the palaeoseismic trenching wall. conclusions of the investigation are summarized as follows;
i. The Tigadihaji dam location was covered by Ranau tuff as a
product of mega eruption of Ranau volcano which have an
age of around 31.000 years,
ii. Pauh fault which located at 1.85 km in the downstream of
the dam with 14 km in length shall be accounted as an
active fault, based on the finding from paleoseismic
trenching
Logging of the left wall iii. The Seismic Hazard Analysis for Tigadihaji dam shall
included the Pauh fault in the analysis.
iv. There was a change in Seismic Hazard analysis for
proposed Tigadihaji dam, where before Pauh fault was
consideren as inactive fault, the Seismic Hazard Analysis
was made base on the probabilistic (PSHA), while after the
Pauh fault was considered as an active fault, the Seismic
Logging of the right wall Hazard Analysis was re-calculated based on Deterministic
(DSHA).
Figure 7. Logging of the palaeoseismic trenching walls
6. REFERENCES
Bellier, O. and Sebrier, M. (1995) Is the slip rate variation on the
Great Sumatran fault accommodated by fore-arc stretching?
Geophys.Res.Lett. Vol.22, No.15. pp1969-1972.
Bellier, O., Sebrier, M., Pramumijoyo, S and Forni, O. (1997)
Paleoseismicity and seismic hazard along the Great Sumatran
fault (Indonesia). Journal of Geodynamics. Vol.24, Nos 1-4,
pp169-183.
Curray, J. (2005) Tectonics and history of the Andaman Sea region.
Journal of Asian Earth Sciences.
Diament, M., Harjono, H., Karta, K., Deplus, C., Dahrin, D., Zen,
M. T., Gerard, M., Lassal, O., Martini, A., and MAlod, J.
(1992) Mentawai fault zone off sumatra - a new key to the
geodynamics of western Indonesia. Geology. Vol.20, No3.
pp259-262.
Fitch, T. (1972) Plate convergence, transcurrent faults, and internal
deformation adjacent to southeast Asia and the western
Pacific. Journal of Geophysical Research. No.77. pp4432-
Figure 8. Indication of Pauh fault on the palaeoseismic trench wall 4462.
332
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Genrich, J.F., Bock, Y., McCaffrey, R., Prawirodirdjo, L., Steven, Natawidjaja, D.H., Bradley, K., Daryono, M.R., Aribowo, S. and
C.W., Puntodewo, S.S.O., Subarya, C and Wdowinski, S. Herrin, J. (2017) Late Quarternary eruption of the Ranau
(2000) Distribution of slip at the northern Sumatran fault Caldera and new geological slip rates of the Sumatran Fault
system. Journal of Geophsical Research. Vol.105(B12): p. Zone in Southern Sumatra, Indonesia. Geoscience Letter. pp
28,327-28,341. 4-21.
Huchon, P., and Pichon, X.L. (1984) Sunda Strait and Central Prawirodirdjo, L., Bock, Y., Genrich, J.F., Puntodewo, S.S.O., Rais,
Sumatra fault. Geology. Vol.12. pp668-672. J., Subarya, C., and Sutisna, S. (2000) One century of tectonic
International Commission on Large Dams (ICOLD). (2016) deformation along the Sumatran fault from trianggulation and
Selecting Seismic Parameters for Large Dams. Bulletin no. Global Positioning System survey. Journal of Geophysical
148. 2016. p75. Research. Vol.105(B12): p. 28,295-326.
Katili, J.A. and F. Hehuwat. (1967) On the occurence of large Sieh, K., (1977) Late Holocene displacement history along the
transcurrent faults in Sumatra, Indonesia. Journal of south-central reach of the San Andreas fault.. Stanford
Geoscience, Osaka City University. Vol.10, pp 5-17. University, Stanford, California.
McCaffrey, R. (1991) Slip vectors and stretching of the Sumatran Sieh, K., Stuiver, M and Brillinger, D. (1989) A more precise
fore arc. Geology. Vol.19. pp881-884. chronology of earthquakes produced by the San Andreas fault
Natawidjaja, D.H. and Triyoso, W. (2007) The Sumatran Fault in southern California. Journal of Geophsical Research.
Zone: From source to hazards. Journal of Earthquake and Vol.94. pp603-623.
Tsunami, Vol.1. Sieh, K., Bock, Y and Rais, J. (1991) Neotectonic and paleoseismic
Natawidjaja, D.H. (2017) Updating active fault map and slip rates studies in West and North Sumatra. AGU 1991 Fall Meeting
along the Sumatran Fault Zone, Indonesia. in Global Program & Abstacts. 72(44). p460.
Qolloqium on Geoscience and Engineering LIPI 2017. Sieh, K., Zachariasen, J., Bock, Y., Edwards, L., Taylor, F and
Jayakarta Hotel, Bandung: Puslit Geoteknologi LIPI. Gans, P. (1994) Active tectonics of Sumatra. Geo. Soc. of
Natawidjaja, D.H., Sapiie, B., Daryono, M.R., Gayatri, I.M., Astyka, Amer. Bull. Vol.26. p A-382.
P., Mukti, M.M., Supartoyo, Hidayati, S and Solikhin, A. Sieh, K. and D. Natawidjaja. (2000) Neotectonics of the Sumatran
(2017) Geologi Gempa Indonesia, in Peta Sumber dan fault, Indonesia. Journal of Geophysical Research, 105(B12).
Bahaya Gempa Indonesia tahun 2017, M.Irsyam, et al., pp295-326.
(Editors). Kementrian PUPR: Bandung. p400. Yeats, R.S. (1996) Introduction to special section: Paleoseismology.
Journal of Geophsical Research. Vol.101. pp5847-5833.
333
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This paper presents the dynamic analysis of El–Agrem concrete face rockfill dam using the finite element software Plaxis-2D.
A parametric study is performed to identify the degree of influence of some parameters such as peak ground acceleration, rockfill shear
modulus and constitutive model on the dam behavior. The results indicated that these parameters have an important effect on the response of
the dam under seismic loading
Keywords: Behaviour of concrete face rock fill dam, stress - strain analysis, earthquake behaviour, performance of CFR dams.
334
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Rockfill materials were modelled with Mohr Coulomb model. Table 4.1 Effect of constitutive model
1 lists the parameters of the dam materials used in the numerical
The effect of rockfill constitutive model on the dam response to
analysis. The dilation angle was estimated by the relation = - 30
seismic loading was investigated using Mohr Coulomb model and
(Bolton, 1986).
elastic model.
The elastic model indicated a crest settlement (2.6 mm) less than
200
that obtained with Mohr Coulomb model (6.8 mm). Moreover, the
A c c e le r a t io n [ c m / s e c 2 ]
100
acceleration at the crest obtained with elastic model (0.92g) is about
0
1.4 times the acceleration obtained with Mohr Coulomb model
-100
(0.66g). Thus the elastic model resulted in a decrease of the crest
-200
settlement by about 62% and an increase of the acceleration at the
-300
crest by about 39% compared to Mohr Coulomb model.
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Time [sec] 4.2 Effect of shear modulus
Figure 3.21 : Accélérations du séisme de Boumerdes composante de la direction Est-Ouest. The effect of the rockfill shear modulus on the response of the dam
Figure 2 Seismic excitation record in the east-west direction
(Boumerdes 2003 earthquake, Algeria) to seismic loading was investigated using shear modulus values of
400 MPa, 450 MPa and 500 MPa. The results shown in Figure 5
indicate that the increase of the shear modulus leads to the decrease
of the settlement and increase of the acceleration at the dam crest.
5 CONCLUSIONS
The dynamic analysis of El–Agrem concrete face rockfill dam using
Plaxis-2D with Mohr Coulomb model indicated that the
amplification factor of the acceleration at the dam crest is about 1.94
and this amplification increases with the increase of the peak ground
acceleration.
The elastic model resulted in a decrease of the crest settlement
and an increase of the acceleration at the dam crest compared to
Mohr Coulomb model.
The increase of the rockfill shear modulus led to the decrease of
Figure 4 Horizontal accelerations at the base and crest of the dam settlement and increase of acceleration at the dam crest.
335
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
REFERENCES Makdisi F., Seed B. 1978. Simplified procedure for estimating dam
Bayraktar, A., Kartal, M.E. and Adanur, S. 2011. The effect of and embankment earthquakes induced deformation.
concrete slab –rockfill interface behaviour on the earthquake ASCEJGE, 104(7), 849-867.
performance of a CFR dam. International Journal of Non- Materón, B. and Fernandez, G. 2011. Consederations on the seismic
linear Mechanics, 46, 35-46. design of high concrete face rockfill dams. The second
Bolton, M.D. 1986. The strength and dilatancy of sands. International Symposium on Rockfill Dams, Rio-de-Janeiro,
Geotechnique, vol. 36, pp. 65-78. Brazil
Brinkgreve, R.B.J.; Swolfs, W.M. & Engin, E. 2010. Plaxis 2D Newmark N. 1965. Effects of earthquakes on dams and
User’s Manual. Plaxis 2010, Netherlands, 199 p. embankments. Geotechnique 15(2), 139-160.
Bureau, G. et al. 1985. Seismic analysis of concrete facerock fill Núñez E. 2007. Behavior of coarse alluvium slopes subjected to
dams”, Proc. Of Symp. On rockfill concrete face dams – earthquakes. Proc. XIII PCSMGE, Margarita (Venezuela),
design, construction and performance. ASCE, New York., 862-867.
pp. 479 – 508 Sfriso, A. O., 2008. Numerical assessment of the deformation of
Cooke J. 1984. Progress in rockfill dams. ASCE, JGE, 110(10), CFR dams during earthquakes. The 12th International
1381-1414. conference of international association for computer methods
Cooke J. 1997. The concrete face rockfill dam. Proc. 17 USCOLD and advances in geomechanics. Goa, India., pp. 4054 – 4061.
Lect., San Diego (USA), 117-132. Sherard, J. L., and Cooke, J. B., 1987. Concrete face rockfill dam:
Haciefendiglu , K. 2009. Stochastic response of concrete faced strong seismic excitation. Journal of Geotechnical
rockfill dams including partially ice-covered reservoir – Engineering, ASCE, 121(2). 185 – 197
foundation interaction undr spatially varying seismic waves. Uddin, N. and Gazetas, G. 1995. Dynamic response of concrete
Cold regions Science and Technology, 58, 57-67. faced rockfill dams to strong seismic excitation. Journal of
Kartal, M.E., Bassaga, H.B. and Bayraktar, A. 2010. Probabilistic Geotechnical Engineering, ASCE, 121(2). 185 – 197.
nonlinear analysis of CFR dams by MCS using Response
Surface Method. Applied Mathematical Modelling, 35, pp.
2752-2770.
336
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This paper presents experimental data on the change of density in a cohesionless fill due to vibratory compaction at different
vibration frequencies. A variable-frequency plate compactor with a frequency range of 10 ~ 100 Hz was designed and constructed. Air-dry
Ottawa sand was used as fill material. The initial relative density of the uncompacted fill was about 33 %. Vibratory compaction was applied
on the surface of a 1.5 m-thick lift. The cyclic normal stress applied on the surface of the fill was kept a constant (cyc = 30 kN/m2) under the
compaction frequencies of 45, 58, 64, 74, 83, 89 and 100 Hz. Test results indicated that within the depth of 0.3 m, which was the width of the
compaction plate, the density increase was most significant. The density increase effect reduced with increasing depth. The effective
compaction depth varied from 0.43 to 0.54 m, which was deeper than the maximum permissible lift thickness 0.15 ~ 0.30 m suggested by
NAVFAC DM-7.2 (1982). Under the same amplitude of cyclic normal stress, both the peak relative density of the fill and the effective
compaction depth were not significantly influenced by the applied compaction frequency.
Keywords: Compaction, Frequency, Relative density, Plate compactor, Sand.
1. INTRODUCTION
Model Wall
In the construction of highway embankments, earth dams, and many
0
other engineering structures, engineers are often required to compact 150
45
loose soils to increase their densities. The purpose of compaction
operation is to improve the engineering properties of soil such as Sidewall
increasing shear strength, reducing permeability and 15
Fo 00
compressibility. Various soil improvement techniques had been used End
otb
oar
d
in the past. Wall
Vibration is especially effective to compact granular soils.
1600
Typical vibrational methods included hand temper compaction,
20
roller compaction, vibroflotation, dynamic compaction, and
compaction sand pile. However, in the literature, there were few
research regarding the relationship between vibration frequency and
soil density change. Limited laboratory and field studies were made,
and no explicit conclusion was drawn. The objective of this study
was to investigate how does the vibratory frequency affects the Steel Column
Steel Base Plate
compaction results.
Baidya and Krishna (2001) conducted experiments to investigate
the effects of loading frequency and amplitude of a vibrating footing
on a layered soil system. It was concluded that, the frequency Unit : mm
corresponding to maximum surface displacement amplitude from Figure 1 NCTU non-yielding model retaining wall and soil bin
the response curve was considered as resonant frequency. For the (after Chen and Fang, 2008)
layered soil system, the observed natural frequencies varied from
8.3 to 31.3 Hz. Wersäll and Larsson (2013) conducted small-scale
tests to investigate the influence of loading frequency on
compaction effects. For the test with large-amplitude the resonant
frequency was about 48 Hz, and for small-amplitude tests the
resonant frequency was about 68 Hz.
Handle
2. EXPERIMENTAL FACILITY
2.1 Soil Bin
The soil bin was designed to minimize the lateral deformation of
soil specimen during testing. In Figure 1, the soil bin was fabricated
of steel plates with inside dimensions of 1,500 mm × 1,500 mm ×
1,600 mm. In Figure 1 the model wall is 45 mm-thick, and the
sidewalls are 35 mm-thick. To achieve an at-rest condition, the wall
material should be nearly rigid.
In Figure 1, twenty-four 20 mm-thick steel columns were
welded to the four sidewalls to reduce any lateral deformation Eccentric
during loading. In addition, twelve C-shaped steel beams were also Motor
337
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Q(t) was produced by the eccentric motor with rotating mass. The
force amplitude applied was:
Qo = me e 2 (2)
where me was the eccentric mass, e was the eccentricity, and was
the angular frequency. The amplitude of cyclic normal stress applied
from the plate compactor to the fill was cyc = Qo/A, where A was
the area of the loading plate. To keep the cyclic normal stress a
constant with increasing frequency, the rotating mass was decreased
by remove the rotating eccentric plate attached to the central axis of
the motor shown in Figure 4.
338
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Lane 2
4. DENSITY INCREASE DUE TO VIBRATORY
COMPACTION
Lane 3
4.1 Density Distributions at Low Vibration Frequency
Figure 8 shows the density distribution with depth after two passes Lane 4
of the vibratory plate compactor at the loading frequency of 45 Hz.
In the figure, for the uncompacted fill, the relative density was 33.3 Lane 5
1.8 %. After 2 passes of the compactor with the rotating rotor
frequency of 45 Hz, the relative density increased up to the peak
relative density of about 116 %. For the upper 0.53 m of fill, the
relative density of the compacted fill was greater than the Dr = 75 % Steel Base Plate Steel Beam Steel Column
which was required by the US Navy Design Manual 7.2 (1982).
Top-View
Figure 7 Compaction path on fill surface (after Lin, 2015)
339
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5. CONCLUSIONS
For the cohesionless fill improved with the compaction frequencies
of 45, 58, 64, 74, 83, 89 and 100 Hz, the density increase was most
significant within the depth of 0.3 m. The density increase effect
was reduced with increasing depth.
The measured peak relative densities varied from 114 to 119 %.
The peak relative density of the fill was not significantly affected by
the compaction frequency.
The effective compaction depths varied from 0.43 to 0.54 m,
which was deeper than the maximum permissible lift thickness 0.15
~ 0.30 m suggested by NAVFAC DM-7.2 (1982).
Under the same amplitude of cyclic normal stress, both the peak
relative density of the fill, and the effective compaction depth were
not significantly influenced by the applied compaction frequency.
6. REFERENCES
Baidya, D. K. and Krishna, G. M. (2001). “Investigation of resonant
frequency and amplitude of vibrating footing resting on a
layered soil system.” Geotechnical Testing Journal, 24(4),
Figure 9 Distribution of relative density with depth at different 409-417.
compaction frequencies Chen, T. J. and Fang, Y. S. (2008). “Earth pressure due to vibratory
compaction.” Journal of Geotechnical and Geoenvironmental
Engineering, 134 (4), 437-444.
Das, B. M. (2010). “Principles of geotechnical engineering.” 7th Ed.,
Cengage Learning, Stamford, CT, USA.
Lin, W. H., (2015). “Effects of Compactor Vibration Frequency on
Ground Settlement and Density Change of Sandy Soil,”
Peak Relative Density, Dr, peak (%)
340
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This paper applied time domain reflectometry (TDR) technology to monitor slope movement of a landslide. The TDR system
was installed from 2012 to 2017 in order to monitor slope movement in the Jiufenershan landslide. TDR waveform signature showed a cable
ruptured by extension at 7.51 m and 35.72 m in depth of TDR_N2 station between 2012 and 2013. Compared to the inclinometer data of the
area only appears some sliding phenomena in a range of 0 m and 8 m in depth. Overall, the TDR technology shows that the locations of
failure planes and the magnitude of the displacements can be determined by the changes and quantification of TDR reflected waveforms in
the landslide.
Keywords: Time Domain Reflectometry, slope movement, landslide monitoring.
1. INTRODUCTION and the magnitude of a cable deformation (Su and Chen, 2000).
Furthermore, whether shearing deformations are big or small, the
This paper uses Time Domain Reflectometry (TDR) technology to magnitude of TDR reflected waveforms is significant correlation
monitor slope movement of a landslide and compares other on-site with the deformation (Drusa and Bulko, 2016). They point out that
measures such as inclinometers to TDR monitoring data. there is a correlation between TDR reflection coefficient and shear
It is the most important for monitoring slope deformation to deformation in a laboratory test so that a regression equation of
determine the location of the sliding surface and the magnitude of them can be determined.
displacements in a landslide. However, traditional monitoring The paper analyzes the waveforms interrogated by TDR
equipment, such as an inclinometer, is time-consuming and difficult monitoring stations in a landslide area and compares to the data of
to interpret an accurate location of sliding surface (Dowding and on-site inclinometers in order to understand the location, type and
Pierce, 1994; Dowding et al., 2003). Thus, TDR technology is a new deformation magnitude of slope movement in the field.
method of slope movement monitoring which is a continuous sensor
to detect any deformation along its cable length when sliding
2. JIUFENERSHAN LANDSLIDE
deformation occurs (O’Connor and Dowding, 1999). They argue
that TDR has been used for slope deformation by comparing to A case study is located in the Jiufenershan landslide in Taiwan with
mechanical inclinometers and the differences of TDR reflected TDR technique where occurred severe landslides by the 1999
waveforms can monitor the locations of cable deformation. Su et al. earthquake. The area of the landslide is 195 ha and the amount of
(2009) demonstrate that TDR would be applied to long-term the collapse is 35 million cubic meters that blocked two creeks,
monitoring of underground sliding planes in alpine landslide areas namely Jiucaihu and Sezikeng creeks, and caused two dammed
effectively, fast and economically. lakes. The area of the Jiucaihu and Sezikeng dammed lakes is 4.4 ha
However, the magnitude of the cable deformation cannot be and 6.4 ha, respectively. Since 2003, there has been a monitoring
directly determined in the field although the instrumentation of TDR landslide project in the Jiufenershan landslide with some monitoring
would monitor suitably deep underground movement (Su et al., equipment, including extensometers, inclinometers, groundwater
2009; Osasan and Afeni, 2010). The paper explains the data of TDR lever gauges, etc. TDR was installed in the area from 2012 in order
monitoring in a case study compared to the data of inclinometers; to monitor the slope movement.
moreover, the paper measured the magnitude of rock displacement The geology of the landslide is underlain mainly by Miocene
corresponding to TDR reflected waveforms in laboratory sedimentary rocks where the strike is N36°E and the dip is 21°SE
experiments. Generally, drill-log reports can determine in situ with an N–S trending synclinal axis of Daanshan syncline in the
geological conditions. According to Su, et al. (2009), they indicate eastern part; the geologic formation of the syncline axis is the
that the TDR system of coaxial cables grouted inside a drill hole can Kueichulin Formation which unconformably contacted with the
detect sliding surfaces and their movements as the function of a underlain Changhukeng Shale; the lower formations include
traditional in-place inclinometer, which showed sliding planes Miocene strata of Shihmen Formation (Shou & Wang, 2003).
between layers of colluvium and strongly weathered slate compared
with the reports of rock cores. TDR has been used for rock and soil
movements in slope monitoring (Federico et al., 2012; Lin, 2014;
Drusa and Bulko, 2016).
In terms of the qualification of TDR monitoring slope
deformation, the technique has been employed to identify zones of
rock mass deformation and blasting performance (Dowding et al.,
1988; Dowding et al., 1989). Then, much research has been
completed with good results in relation to TDR indoor experiments,
including the effect of TDR cable length on reflection signal and
different types of deformation of TDR waveforms (Su, 1990).
Furthermore, the methodology to quantify TDR waveform change
and correlate to ground deformation has made the TDR monitoring
system used in landslides more useful and advanced (Su and Chen,
1998). Su and Chen (1998) declare that an integrated area method is
proposed to quantify the magnitude of a TDR reflected waveform on
the deformation of rock masses and concrete structures through
TDR theories and bench tests. By this method, the study applies
laboratory experiments of shear tests to determine equations in the Figure 1 Location map of the monitoring stations of the
relation between the integrated area of TDR reflection coefficient Jiufenershan landslide
341
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
For the shear test, a test specimen was installed on the bench test
shown in Figure 3 and was design as two shearing units, S1 and S3.
While the amount of shear displacement at S1 location was fixed to
0, 10, 20, 30 and 40 mm, the amount of shear displacement at S3
location incremented by every 2 mm until the cable ruptured,
respectively. Meanwhile, a TDR tester interrogated the reflected
waveforms of the cable displacements at the two locations. The
results of TDR interrogated waveforms are plotted in Figure 5. The
TDR cable signature (reflection coefficient, unit: millirhos)
represents the amplitude of reflected waveform at a given distance
(unit: meter) of grouted cable. The zoom-in graph of the reflected
waveforms is shown in Figure 8 that illustrates the increase of the
shear displacement on S3 plane until the cable ruptured by 54 mm.
All shear tests are the same tendencies as the graph of Figure 5
at S3 plane when the amount of S1 displacement is 10, 20, 30 and
40 mm, respectively. Above all shear tests, the average maximum
Figure 2 Diagram of on-site TDR instrumentation magnitude of cable displacement by shear is about 60 mm.
According to the results of the shear tests, there is a significant
There are two TDR monitoring stations in the area where coaxial relationship between shear displacements and the magnitude of
cables were grouted. Also, there are two inclinometers, as shown in reflection coefficients. As can be seen in Figure 6, a quadratic
Figure 1. A diagram of on-site TDR instrumentation is shown in regression equation is built up in the relations between the reflection
Figure 2. A TDR cable tester (HL 1101) sends a voltage pulse coefficient and shearing deformation (The R-Square of each
waveform which travels along a metallic coaxial cable through a equation is more than 0.99). As a result, it is applied to determine
connector. The cable was grouted into a borehole in a landslide with the magnitude of shear slips in field corresponding to the magnitude
cement. When the waveform of TDR encounters a change in the of reflection coefficients interrogated by on-site TDR system.
characteristic impedance of the cable, it is partially reflected by the
deformation of the cable in a potential sliding surface. The change
could be determined as the location of the shear or extension zone in
a landslide.
342
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
343
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
7. REFERENCES
Blackburn, J. T. and Dowding, C. H. (2004). Finite-element analysis
of time domain reflectometry cable-grout-soil
interaction. Journal of geotechnical and geoenvironmental
engineering, 130(3), 231-239.
Dowding, C. H. and Huang, F. C. (1994). Early detection of rock
movement with time domain reflectometry. Journal of
Geotechnical Engineering, 120(8), 1413-1427.
Figure 10 Results of BH8 inclinometer measures Dowding, C. H., Dussud, M. L., Kane, W. F. and O’Connor, K. M.
(2003). “Monitor deformation in rock and soil with TDR
5.2 N1 monitoring station sensor cables.” Geotechnical Instrumentation News, June,
51–59.
There was not a significant change of the waveforms at TDR_N1 Dowding, C. H. and Pierce, C. E. (1994). “Use of Time Domain
station from 2012 to 2017, which means no slope movement Reflectometry to Detect Bridge Scour and Monitor Pier
occurred. Meanwhile, the inclinometer data did not detect any Movement,” Proceedings of the Symposium on Time Domain
sliding phenomenon in Figure 11. The station could be located at the Reflectometry in Environmental, Infrastructure, and Mining
outside of the landslide area but the monitoring data still needs to be Applications, Evanston, Illinois, Sept 7-9, U.S. Bureau of
detected for long-term observation. Mines.
Dowding, C. H., Su, M. B. and O'Connor, K. M. (1988). Principles
of time domain reflectometometry applied to measurement of
rock mass deformation. International Journal of Rock
Mechanics and Mining Sciences & Geomechanics
Abstracts, 25 (5), 287-297.
Dowding, C. H., Su, M. B. and O'Connor, K M. (1989).
Measurement of rock mass deformation with grouted coaxial
antenna cables. Rock Mechanics and Rock
Engineering, 22(1), 1-23.
Drusa, M. and Bulko, R. (2016). Rock Slide Monitoring by Using
TDR Inclinometers. Civil and Environmental Engineering,
12(2), 137-144.
Federico, A., Popescu, M., Elia, G., Fidelibus, C., Internò, G. and
Murianni, A. (2012). Prediction of time to slope failure: a
general framework. Environmental Earth Sciences, 66(1),
245-256.
Kane, W. F. (2000). "Monitoring Slope Movement with Time
Figure 11 Results of BH7 inclinometer measures
Domain Reflectometry (TDR)", Geotechnical Field
Instrumentation: Applications for Engineers and Geologists.
To summary, there were some extended displacements at
different depths over time at N2 TDR monitoring station in the Sponsored by: ASCE Seattle Section Geotechnical Group
Spring Seminar and the University of Washington
landslide area, which was the same as the interpretation of rock core
Department of Civil Engineering, Seattle.
data. Meanwhile, the locations of failure planes and the magnitude
Kane, W. F. and Beck, T. J. (1996). Rapid slope monitoring. Civil
of the displacements can be determined by the changes and
quantification of TDR reflected waveforms. However, the data of Engineering, 66(6), 56.
Lin, C. (2014). Slope’s automatic monitoring and alarm system
inclinometers was not significant result of monitoring. Inclinometers
based on TDR technology. Civil Engineering and Urban
may not detect smaller changes in the displacement of slope
Planning III, 371.
movement. It is obvious that TDR is an efficient and economical
O’Connor, K. M. and Dowding, C. H. (1999). GeoMeasurements by
method of locating the depth and measuring the magnitude of slope
Pulsing TDR Cables and Probes. CRC Press, Boca Raton,
movement in the landslide.
Florida.
Osasan, K. S. and Afeni, T. B. (2010). Review of surface mine slope
6. CONCLUSIONS
monitoring techniques. Journal of mining science, 46(2), 177-
The research applied TDR to detect the displacement of slope 186.
movement and to measure the type and magnitude of the Shou, K. J., & Wang, C. F. (2003). Analysis of the Chiufengershan
deformation in the Jiufenershan landslide. Compared to the data of landslide triggered by the 1999 Chi-Chi earthquake in
on-site inclinometers, it proved that there was an extended Taiwan. Engineering Geology, 68(3), 237-250.
displacement at a depth of 35.72 m at N2 TDR monitoring station Su, M. B. (1990). Fracture Monitoring within Concrete Structure by
by 0.31 meters and TDR waveform signature showed the cable Time Domain Reflectometry. Engineering Fracture
ruptured by extension at a depth of 7.51 m from 2012 and 2013. Mechanics, 35 (1/2/3), 313-320.
344
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Su, M. B. and Chen, Y. J. (1998). Multiple Reflection of Metallic Su, M. B. and Chen, Y. J. (2000). TDR Monitoring for Integrity of
Time Domain Reflectometry. Experimental Techniques, Structural Systems. Journal of Infrastructure Systems, 6(2),
22(1), 26-29. 67-72.
Su, M. B. and Chen, Y. J. (1999). MTDR monitoring systems for Su, M. B., Chen, I. H. and Liao, C. H. (2009). Using TDR Cables
the integrity of infrastructures. Journal of intelligent material and GPS for Landslide Monitoring in High Mountain Area.
systems and structures, 10(3), 242-24 Journal of Geotechnical and Geoenvironmental Engineering,
ASCE, 135(8), 1113-1121.
345
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This paper presents the applicability of data mining techniques to improve the accuracy of photogrammetric roughness data.
Close range photogrammetry (CRP) has employed to create competent 3D slope models and can be used to highlight geological features on
rock slopes. However, when a high degree of accuracy is required, this technology has also shown inconsistent values with respect to its
accuracy because of various influencing factors induced by camera structures and survey environment. Thus, the uncertainty of this
technology demands complex post image processing frequently. Data mining (DM) techniques are increasingly being used to identify valid
and potentially useful correlations and patterns in existing data. In this study, a data mining technique is applied to reduce the noise of 3D
images and to analyse multiple images effectively. Conclusions showed that a better performance was achieved using DM technique and this
data mining approach was able to improve the accuracy of photogrammetric roughness data.
1. INTRODUCTION al., 2003; Thompson et al., 2005). For example, 3D images are used
in a medical image analysis area, sets of 3D images of human brains
Data mining (DM), which is also known as knowledge discovery obtained from MRI scanning have been analysed using a fully
from data (KDD) is the process of discovering interesting patterns
automatic method using a data mining program (Uher et al., 2013).
and knowledge from large amount of data. As a fast-growing field,
DM has been applied effectively to solve multiple problems This study presents applicability of data mining techniques to
connected to risk management, including error detection and predict the accuracy of photogrammetric roughness data. Data sets
quantification, detecting fraud etc. in our data-rich world. Higher and attributes of joint roughness coefficient (JRC) and asperity
level of this technique can be to build learning models to heights obtained from a set of photogrammetric laboratory tests of a
automatically extract knowledge from big and complex data. This previous study are employed as input data. Also, the results of
technique has been progressively introduced to various industry image mining approaches using 3D images corresponding to the
sectors like marketing, manufacturing and construction. data sets are discussed.
In geotechnical engineering area, since early 1990s’ Artificial 2. OVERVIEW OF DATA MINING TECHNIQUES
Neural Network (ANN) has been increasingly employed to analyse
various engineering issues such as constitutive modelling, bearing Data mining applies statistical and logical methods to large data sets
capacity of pile and slope stability. This technique was regarded as and allows people to locate and interpret data patterns, helping them
the most appropriate and effective tool among other data mining better informed decisions. The analysis process, called Knowledge
techniques as it is known as the most relevant to engineering Discovery in Database (KDD), defines the procedure for
disciplines. Recently, Tinoco et al., (2010) approached to predict transforming raw data into useful knowledge. The method can be
compressive strength of jet grouting columns using a data mining used to categorize the data, or this can be used to create predictive
technique. A data mining approach for the volcanic rock models. As shown in Fig. 1, a general KDD analysis process
classification has been reported by Miranda et al. (2018). consists of the following multiple steps: [1] Data selection, [2] Data
cleaning, [3] Data transformation, [4] Pattern extraction and
Close distance photogrammetry has been increasingly adopted to interpretation.
create 3D models to investigate detailed aspects of rock slopes, as
well as large-scale surveys in rock engineering and mining. Recent
high-end equipment in the field of digital photographing can create
high resolution images and this technology has encouraged the
spread of close range photogrammetry to obtain surface features
such as undulation and roughness which require higher degree of
accuracy of 3D models. However, there are several known and
unknown factors that can affect the accuracy of the results and the
accuracy of photogrammetric models has insufficiently discussed.
The authors have been studied on the accuracy of photogrammetric
roughness data for many years (Kim et al., 2015; 2016). However, Figure 1 Process of data mining
the relevant data have still been insufficiently explored and analysed In general process of data mining, firstly, we need to identify the
by advanced data analysis techniques. objects of interests in the data and secondly, the quality of the data
As the pixels of digital photographs embody large number of can be improved by techniques such as de-noising and cleaning.
features, images represent a significant assembly of measurements. Third, the multivariate data can be converted into suitable forms for
As a part of data mining, image mining as a specific term has been pattern recognition by pre-processing. Then, the patterns in the data
used for analysing various images of DNA microarrays, can be identified using techniques such as classification, clustering
astronomical observations, satellite maps, medical images etc (Van and association rules. The overall data mining is interactive process.
der Walt et al., 2014). With regard to image data, data mining is Data analysts involved in every step starting with an initial
associated with four broad problems: finding association, formation of data sets in the data collection process to validating the
classification, sequential patterns and time series patterns (Ordonez results obtained at each step. The process of this study is similar to
and Omiecinski, 1998). The applications of image mining the basic procedure of data mining shown in Figure 1.
techniques have been reported in the literature (Soh, 1999; Uher et
346
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3. METHODOLOGY In the program ˋRapidminerˊ, there are many operators for image
transforms and manipulation. As a feature extraction algorithm, this
This study analyses rock surface roughness data obtained from a set study uses “local-level” feature extraction. This is suitable for
of photogrammetry laboratory tests. The image data used in this extracting information such as pixel brightness integer values after
study is parts of image sections extracted from the 3d surface image enhancement from each point in the image. Then, the image
models. JRC values and the corresponding measurements such as data can be transformed from image into a structured table form.
camera-to-object distances, pixel sizes and focal length of lenses are The first operator used as a median filtering with a size of 3 pixels.
also employed in this study. To analyse the data in each process, this In this stage, multiple image operator was also employed. This
study uses a data mining program, RapidMiner studio (ver. 8.1) operator iterates the next analyses over all image files in a folder.
which is an open-source and flexible platform implemented in Java This provided convenience for the reiterative procedure for dozens
(RapidMiner Inc.). Also, in the pre-processing stage, the 3D images of images in this study.
are filtered by smoothing to reduce the influence of the noise of raw
images.
A digital image can be regarded as a group of discrete pixels, each
of which has various ranges of colour and brightness information.
Colour images with various geological features can be more
complex than greyscale images. In this study, pixels converted from
colour to 8-bit grey-scale are used as the input data. Thus, the
processed values are in the range from 0 to 255. Common approach
in data mining is to mine information from structured data with
tabular form. However, image is an example of the unstructured data
form. Thus, a feature extraction process which transforms an image
into the tabular data is needed. In this study, Image mining
extension (IMMI) operator is employed in RapidMiner for the
feature extraction process. Figure 3 Concept of image mining using image processing operators
with multiple images
3.1 Data cleaning
There are several operators which can enhance the quality of images Then, a learning process can be applied to reduce noise, recognize
in the IMMI operator. As a representative tool, median filtering is and identify a pattern of pixel information for multiple images.
one of effective solutions to reduce the influences of the noise from There are several learning algorithms in IMMI’s operators such as
the obtained photogrammetric roughness data. Poropat (2008) support vector machines (SVM), decision tree and random forest. In
showed a clear effect of median filters from some simulations using this study, SVM with a Kernel method was selected for training
roughness profiles (Fig. 2). To enhance the accuracy of the stage based on the results of trials with other algorithms. It has been
photogrammetric roughness data, the raw 3D images are processed reported by Burget et al. (2012) that SVM with Kernel method was
with a median filter in this study. also successfully used for analysing a biomedical image. The
designed process of Rapidminer in this study is shown in Figure 4.
The results obtained from the learning process can be presented in
tabular forms for further analysis steps. These structured data were
used for statistical analyses using the differences of integer numbers
between neighbouring points of interest. This data collection method
was applied to the pixels at the given points of interest at the same
locations on several sets of different resolution images.
347
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
where Into is the brightness integers obtained from the highest values between two consecutive pixels along the profiles of interest
resolution image; and Intp is the brightness integers obtained from obtained from different resolution images obviously indicate
compared images. downward trends as the camera moves farther from the rock sample,
corresponding to the decrease of image resolution. This is directly
dependent on the sizes of pixels of the created images.
4. CASE STUDY
4.1 Photogrammetry laboratory data
The data used in this paper were collected from a previous
laboratory work on accuracy of photogrammetry roughness data.
The data were images of a rock sample with different resolutions as
shown in Table 1. The photographs were taken in the range of 1.0 ~
7.0 m, using three different focal length of lenses: F=24mm, 50mm
and 85mm. The length and width of the targeted area on the rock
surface were 20 cm and 8 cm, respectively. Details of the tests are
presented in the previous literature by the authors (Kim et al., 2015).
The spatial data of the 3D surface models were created using a
photogrammetry code, Sirovision (CAE). JRC values were
Figure 6 Schema of image profile
estimated along 8 measurement profiles on the rock sample.
This study selected high resolution images with a range of pixel
sizes less than 1.0 mm as summarized in Table 1. The selected area
of the sandstone surface was appeared to be reasonable in this study
because in observations it was found that the range of brightness on
the surface was quite simple as an indicator of the change of
roughness based on its simple rock-forming. Figure 5 compares the
greyscale images of the sample area with different resolutions.
Using the targets on the images, 8 profiles were selected at the same
locations. The data of pixels with 1 mm interval along the profiles
were extracted by using an operator of IMMI. A total of 28,000
integer values were collected from 14 images and employed for the
learning process.
Table 1 Summary of image data used in this study
FL c-to-o Pixel size No. No.
(mm) distance (m) (mm) images data
24 1~3 0.5 ~ 1.0 3 6,000
50 1~5 0.1 ~ 1.0 5 10,000
85 1~7 0.1 ~ 1.0 6 12,000
Figure 5 Comparison of images of the rock sample of pixel size = Figure 7 Data distributions of integer values
1.5 mm a), 1.0 mm b) and 0.2 mm c)
4.2 Integer values of pixels The histograms of all training showed an interesting trend. Using all
resolution images, Figure 7 indicates the differences of integer
The designed learning process produced integer values of the values obtained from comparison between highest resolution (pixel
smoothed images by a median filter in a tabular form. This image size = 0.1 mm, FL= 85mm) and target images. The data formed an
mining process enabled us to collect tens of thousands of data in a exponential distribution for each profile. Mean values of the integer
short period of time from hundreds of the selected points on differences was ranged from 13.7 to 15.7 for all image profiles. Also,
multiple images. As presented in Figure 6, for the continuous integer a high range of standard deviation which ranges from 11.6 to 21.2
of a profile on an image, absolute values of the difference between was calculated from the set of data values. The ranges of statistics
two consecutive integer values along the profile is regarded as a were quite similar to each other in the different resolution images.
roughness parameter in this study. Sets of the differences of integer
348
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
In Figure 7, the histograms and the cumulative distribution curves - Using RMSEINT, a range of JRC values can be estimated and
distributed by their corresponding JRC values also indicate this method can complement when photogrammetry is
interesting trends to evaluate the accuracy of photogrammetric JRC employed to quantify JRC values.
values. Overall, bigger JRC values formed wider dispersion of data.
The range of integer values for smaller JRC values is composed of 6. REFERENCES
smaller |Into-Intp|values. The threshold values which composes
American Society for Photogrammetry & Remote Sensing (2014)
50% of data are 10 (JRC= 9.9), and 14 (JRC= 11.5) and 17
"ASPRS positional accuracy standards for digital geospatial
(JRC=13.5). This may be attributed to the fact that the standard
data". PE & RS, 81, 3, ppA1-A26.
deviations increase with the JRC values. The inspection of
Burget, R., Vaclav, U. and Masek, J. (2012) "Trainable
cumulative distribution plots indicates that the integer values
segmentation based on local-level and segment-level feature
obtained from different resolution images can be used to verify
extraction". IEEE International symposium on biomedical
photogrammetric JRC values.
imaging: from nano to macro, pp1-8.
4.3 Integer values and JRCs Han, J., Kamber, M., Pei, J. (2012) "Data mining concepts and
techniques". 3rd edition, Elsevier Inc., p.703.
Using the RMSEINT values obtained as described in Section 3.2 and Hofmann, M. and Klinkenberg, R. (2014) "RapidMiner dataming
using the equation (1), statistics analysed a relationship between use cases and business analytics applications". Taylor &
RMSEINT and JRC values as presented in Figure 8. The ranges of Francis Group, LLC, p.518.
RMSEINT obviously increases as the corresponding JRC values are Kim, D-H., Poropat, G., Gratchev, I., Balasubramaniam, A. S.
raised. This is an interesting finding because the integer values are (2015) "Improvement of photogrammetric JRC data
obtained from 2D images which exclude any information of distributions based on parabolic error models". International
coordinates of the rock surface. Journal of Rock Mechanics & Mining Sciences, 80, pp19-30.
In the authors’ previous study, the data distributions of normalized Kim, D-H., Poropat, G., Gratchev, I., Balasubramaniam, A. S.
JRC values comparing photogrammetric JRC values and manual (2016) "Assessment of the accuracy of close distance
measurement were fairly scattered showing the ranges of the photogrammetric JRC data". Rock Mechanics and Rock
coefficient of determination, R2 from 32% to 48%. The uncertainty Engineering, 49, pp4285-4301.
of this technology to estimate JRC values need assistants to improve Miranda, T., Sousa, L. R., Gomes, A. T., Tinoco, J. and Ferreira, C.
its accuracy to quantify roughness data of rock surface. The integer (2018) "Geomechanical characterization of volcanic rocks
values obtained from different resolution greyscale images can using empirical systems and data mining techniques". Journal
complement the shortcoming of photogrammetry. of rock mechanics and geotechnical engineering, 10, pp138-
150.
Ordonez, C. and Omiecinski, E. (1998) "Image mining: a new
approach for data mining". Georgia Institute of Technology
technical report, GIT-CC-98-12.
Poropat, G. V. (2008) "Remote characterisation of surface
roughness of rock discontinuities". In: Proceedings of the 1st
Sourthern hemisphere international rock mechanics
symposium, Perth, Australia, pp 447-458.
Soh, L-K. (1999) "Segmentation of satellite imagery of natural
scenes using data mining". IEEE Transactions on geoscience
and remote sensing, 37, 2, pp1086-1099.
Thompson, D. R., Smith, T. and Wettergreen, D. (2005) "Data
mining during rover traverse: from images to geologic
signatures". In: Proceedings of 8th International Symposium
on Artificial Intelligence, Robotics and Automation in Space,
Germany.
Uher, V., Burget, R., Masek, J. and Dutta M. K. (2003) "3D brain
tissue selection and segmentation from MRI". In: Proceedings
of 36th International conference on telecommunications and
signal processing (TSP), Rome, Italy.
Figure 8 Correlation between RMSEInt and JRC values
5. CONCLUSION
This paper presented an applicability of image mining technique to
analyse multiple images of rock surface for rock surface
investigation. A learning process has been designed to perform the
image mining process using a data mining program, Rapidminer.
Using an image data set collected from a previous photogrammetry
laboratory tests, following conclusions can be drawn:
- Image mining techniques can be effectively used for
analysing multiple images including a large volume of image
data. In this study, IMMI operators of Rapidminer was
employed.
- Differences of integer values of pixels are used to identify
rock joint roughness coefficient (JRC) in two different types
in this study; [1] |Inti+1-Inti|and [2] |Into-Intp|. The
cumulative distribution of|Into-Intp|indicated a relationship
between JRC values and the integer values.
349
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: It has been a common practice in Malaysia to adopt reinforced soil walls as one of the common wall types particularly at
housing earthworks platforming. This paper will discuss one case study of the failed reinforced soil walls used in the housing project. The
paper will discuss on the background on the construction of the site development works leading to the construction of the RS wall on a deep
fill ground. The paper will highlight on the wall failure and the postulations on the likely reasons for the failure. The subsoil and groundwater
details from extensive ground investigation together with the short term and long term remedial solutions to strengthened the fill slopes and
wall will be discussed. The paper will also present all the instrument monitoring results validating the remedial works performance. Lessons
learned from these experiences will be highlighted and discussed in detail.
Keywords: Failed reinforced soil wall, remedial works with soil nail, case study.
1. INTRODUCTION
The project site where the wall collapsed is located about 15km
from Kuala Lumpur City within a refutable suburb housing project.
Existing bungalow houses are located on the south side of the site
with a road dividing the collapsed wall and the houses. The north
side is predominantly undeveloped and comprise of flat ground after
the slopes reserved for future development of Muslim cemetery and
school. The forest reserve is located on the eastern side. A stream is
located at the toe of the slope flowing from the forest reserve to the
west.
Based on the topographic map obtained from the relevant
authority of Malaysia, the site was located at the eastern boundary of
an estate on a north trending valley along the eastern slope of a
ridge. The map also indicates a stream together several ephemeral
streams originally passed through the site. Three streams or rivulets
coming from the north and east directions converged into a bigger
stream north of the site. As the three streams shown in the map
initiated just outside or within the development site, this may not
contribute to any underground seepage. The stream from the north
side originating within forest serve may contribute to the
underground seepage as this stream was not diverted during the site
development stage in year 2000 to 2001.
During the site development (earthworks), the western and
south-eastern parts of the site were being cut whilst the north part
being filled with the cut materials which comprise of earth and rock
fragments. Between 2001 and 2004, the southeastern hill was further
cut and the north valley was filled to form the building platform
level. The fill slopes were formed with slope drainage and
vegetation based on the satellite imagery. In year 2005, a 120 m
long RS wall with heights varying from 4m to 10m was constructed
on the northern most slope to create space for the construction of the
access road at RL 63.20 m to the residential units.
Part of the wall about 11m stretch collapsed on 20 May 2011.
The collapsed wall is 9m high with exposed wall height of 25 m.
The top 6 m of the RS wall with 4 panels vertical and 7 panels
horizontal collapsed. The location and photograph of the collapsed
wall is shown in Figure 1 and Figure 2. Figure 1 Location of collapse wall
350
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
This paper will address the subsoil conditions together with the
ground water conditions, the topographical changes over the period
at this site, subsoil design parameters interpreted from the laboratory Figure 5 SPT-N versus Depth plot
tests, likely reasons for the wall collapse, the short term & long term
351
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Layer Description SPT – N Thickness Moisture Bulk Atterberg Limits Grain Size (%)
Blows (m) Content Density (%)
(%) (kN/m3) LL PL PI Coarse Fine
1 Fill 2 – 31 1 - 25 10 - 25 19.7 – 20.8 20 - 15 - 5– 45 - 90 10 - 55
(varies) 40 25 15
2 Residual Soil < 50 4 - 12 10 - 35 15.8 – 21.5 25 - 15 - 10 - 20 - 75 25 - 80
50 35 15
3 Weathered Soil > 50 > 10 15 - 25 - 20 - 15 - 10 - 50 - 80 30 - 50
40 20 25
The typical t-s plots for the fill material and the residual Block samples from the trail pits were also obtained to validate
weathered formation to obtain the effective strength parameters are the effective strength parameters consolidated undrained (CIU) and
shown in Figure 6a and 6b. drained (CD) triaxial tests. The summary of these results shown in
Table 2 validated the values used in the analysis.
RESIDUAL SOIL
(SPT-N<50)
γ = 18kN/m³
c' = 5kPa
φ’ = 32˚
RESIDUAL SOIL
(SPT-N<50)
Adopt: γ = 20kN/m³
c' = 5.0 kPa c' = 7kPa
φ’ = 34˚
φ’ = 32˚
Seismic refraction survey was carried out along the wall line as
shown in Figure 8 and the results in shown in Figure 9. The seismic
refraction survey indicates two distinct layers based on the contrast
of the velocities. The upper layer is characterized by low p-wave
velocity representing the fill materials of sandy SILT or silty SAND.
Underlying this fill layer is the highly weathered meta-sedimentary
profile with higher p-wave velocity. These results conform closely
Figure 6b t-s Plot for cut (SPT-N < 50) to the findings of the boreholes.
352
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Three lines of resistivity survey were carried out at the first berm
from the base of slope, base of the RS wall and at the roadside near
to the houses. Typical resistivity profile is shown in Figure 10.
These profiles show the presence of partial saturated soil layers
indicating presence of groundwater seepage.
4. REMEDIAL WORKS
The remedial works were carried out in two stages viz short term
and long term. The details of the remedial works are elaborated in
the following sections.
353
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
FOS: 1.06
354
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
To mitigate erosion problems on slopes, close turfing and Instrumentation were installed during the remedial works to
mortared stone pitching were used on slopes not covered by the monitor the performance of the slopes stabilization measures during
rockfill. and upon completion of construction works. The instrumentation
As failure of the reinforced soil wall involved leakages of the plan is shown in Figure 17.
water carrying services and soaking of the cohesive soil behind
the wall, free draining material was used to backfill behind the
wall. By doing so, water will be easily drained out from behind
the wall and the probability of another failure would be much
reduced. The existing precast drainage system along the wall and
part of the site road, which is currently leaking was replaced with
cast in situ drains to reduce or avoid leakages. The water supply
pipes along the wall was totally exposed and installed inside a
RC channel with measure to drain the water from the pipe if any
leakage. Sewage pipes along the wall was also redesigned and
replaced with a more robust system that could withstand large
differential settlement.
The damaged water and sewer pipes were repaired with special
allowance for damage. The final completed remedial work is shown
in Figure 16.
355
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
6. LESSONS LEARNED
356
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: The occurrence of several landslides followed by the debris flow, such as Sijeruk Landslide on January 4, 2006, Jemblung
Landslide on December 12, 2014 and Pasir Panjang Landslide on Februari 22, 2018 has make a lot of casualties and terrible destruction in
some parts of Central Java. In general, debris flow can be divided into source area, flow track and depositional area. This study will focus on
modelling the run-out of debris materials, starting from the source area to the distribution of debris materials at the depositional area.
In this study, debris flow simulations were used to better understand with regard of predictions of volume, run-out mechanisms, flow rate of
materials and the distribution of debris materials. In general, the results of this study will be very useful to predict the run-out of any potential
similar debris flows in some areas of Indonesia which might experience the similar events as well as the determination of mitigation
measures to minimize the negative impact of these events.
Keywords: landslide, debris flow, run-out, modeling.
1. INTRODUCTION Debris flows always occur in an area which has hill morphology
with high intensity of rainfall. Slope angle and river stream are
The occurrence of several landslides followed by the debris flow, factors that determine the process of debris flow. Debris flow will
such as Sijeruk Landslide on January 4, 2006, Jemblung Landslide happen if a slope has huge amount of debris material with huge
on December 12, 2014 and Pasir Panjang Landslide on Februari 22, amount of water supply as its stream. Debris flows always occur
2018 has make a lot of casualties and terrible destruction in some after or when heavy rain. Debris flows has high specify gravity
parts of Central Java. which make a boulder flow through the stream to the toe of
From Stiny (1910) in Jacob and Hungr (2005), debris flow depositional area.
begins with the description of a flood in a mountain torrent, carrying
suspended load and transporting quantities of bedload. As the
2. METHODS
amount of sediment carried by the flow increases, at certain limit it
has change into a viscous mass consisting of water, soil, sand, This study will focus on run-out modeling of debris materials, since
gravel, rocks and wood mixed together, which flows like a lava into the flow of debris materials started to distribution of debris materials
the valley. Meanwhile, Varnes (1978) in Blijenberg (2007) describe in the depositional area, which is usually found in the plain areas of
that debris flows are mass movements consisting of granular solids, foot slope. Some of debris flow events in Central Java are selected
water and air moving as a viscous flow. to be modeled in this research by mean of numerical simulation
In general, debris flow zonation can be divided into three parts, technique with Graphical User Interface (GUI). The software that
which are source area, flow track and depositional area of debris used in this simulation is kanako ver.2.01. From Takahashi and
materials. The Geometri of debris flow can be seen in Figure 1. The Nakagawa (1991) in Nakatani, et al. (2008), the modeling concept in
first part is source area which an area where the accumulation of 2D based from the equations for momentum, continuation, riverbed
debris material occurs. The accumulation of debris materials might deformation, erosion/deposition and riverbed shearing stress.
be consist of materials of slope itself or material that make natural Input data parameters obtained from field conditions and the
dam. The second part is flow track which an area where debris results of physical properties analysis of materials, geomorphology
materials flow through stream to the valley. The third part is conditions and hydrogeological conditions. These parameters are
depositional area which a plain for sedimentation of debris used to modeling source area of debris flow. Shape of river was
materials. The Distribution of debris flow in depositional area used as analogy of flow track. These parameters which used as input
depends on topography. If in the depositional area still has a data can be seen in Table 1. Based on input data parameters, a series
difference in slope angle with slope below, secondary flow track of numerical simulation techniques performed by several stages.
maybe occurs which will make secondary depositional area First stage is conversion contour elevation into the form of a grid
(multiple depositional area). system with a scale adapted to the area affected. Second stage is
define engineering properties of rocks as the model’s variable input
and flow tracks. Third stage is analysis of supply hydrograph for
indicating the influence of water that would affect flow behavior.
Last stage is simulate to predict the volume, run-out mechanisms,
the flow rate, as well as the distribution of debris materials that
formed.
Table 1 Input Parameters
Parameters Unit
Simulation Duration second
Calculation time interval second
Diameter of materials m
Mass density of bed materials kg/m3
Mass density of fluid phase (water, mud, and silt) kg/m3
Consentration of movable bed unitless
Gravity acceleration m/s2
Figure 1 Block diagram of debris flow geometry Coefficient of erosion rate unitless
(Modified from Nettleton et al., 2005) Coefficient of accumulation rate unitless
357
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
where:
Qsp = Peak of the sediment supply per second (m3/second)
ΣQ = Total amount of moveable material that define in Figure 3 The distribution of debris flow with viscosity 0.45 at 891
equation 1 (m3) seconds that seen from southeast
Vdq = Volume of the sediment (m3)
C* = Consentration of moveable bed
3. CASE STUDY
3.1 Sijeruk Village, Banjarnegara, Central Java, January 4
2006
Debris flow occurred in Pawinihan Hill, Sijeruk Village,
Banjarnegara Regency at coordinates S 07° 19 '16.1"and E 109° 42'
08.3'' with elevation 994 m. Over the incident, as many as 77 people
were killed, 8 people missing. Pawinihan Hill composite rock
consists of flow breccia flow, pyroclastic, lava, and alluvium
(Condon W.H., et al., 1996). The slopes of Mount Pawinihan have a
slope of between 20° to 60°. The direction of landslide at the crown
on the slopes of Pawinihan Hill is towards the southeast (N 145° E).
In the former landslide area, remnants of breccias that experienced
landslide still visible.
Debris flow in the Sijeruk Village has uniqueness in the flow
path, which is encounter a sharp deflection. It is alleged that the
deflection is caused by a ridge that blocks the flow track of the
flowing material that move southeast. The ridge causes the direction
of the flow that originally toward the southeast transformed to the
east. The depositional area tends to the southeast following the slope Figure 4 The distribution of debris flow with viscosity 0.45 at 891
of the valley of Pawinihan Hill. The topographic of the Pawinihan seconds that seen from top
Hill can be seen in Figure 2 with 1 grid representing the 25x25
meter area. The length of debris flow in Sijeruk Village reached 3.2 Jemblung Village, Banjarnegara, Central Java, December
about 1.1 km. 12 2014
The modeling of the debris flow using scenario of the viscosity The location is on Jemblung Village, Banjarnegara Regency which
value 0.45. The process of direction deflect in flow track and the has steep topography that has angle reached 75°. Landslide occurs in
distribution of sediment in the deposition area can be seen in Figure the southern slope of Jemblung Village with coordinate 7° 16' 51.6"
3 and 4. When the material enters the flow track, material moves at S and 109° 43' 12.6" E. The victims died reached 95, 24 people
speed 21.17 km/hour. The flow of materials is slowly moving in the missing, and dozens more injured. From December 9 to 11,
deposition area at speed 5.4 km/hour. The time it takes the debris Banjarnegara has high intensity of rainfall reached 100 mm.
flow material to reach the depositional area is 891 seconds. The Material compilers in the slopes of Jemblung Village are Jembangan
volume of debris flow material which flow reaches 127,735 m3.
358
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Volcanics (Qj) which consists of Andesitic lava and volcaniclastic 3.3 Pasir Panjang Village, Brebes, Central Java, February 22
(Condo et al., 1996) and have largely been strongly weathered. 2018
These materials became source of the debris material. Figure 5
shows the debris flow area and surrounding area with 1 grid Landslide incident occurred in Pasir Panjang Village, Brebes
representing an area of 14.55 (west-east) x 15.31 (north-south) Regency with coordinates 108° 47' 19.68" E and 07° 7' 6.59" S.
meters which seen from northeast. Victims that recorded are 18 peoples who work as farmer in the
southeast landslide location. Landslide occurs on hills morphology
with steep slope. The hill consists of Tapak Formation which
composed of greenish sandstone in the lower part that gradually
grades upward into greenish sandstone with some intercalation of
grey to yellowish sandy marl (Kastowo, 1975). The Landslide that
flows toward southeast in slopes that have angle between 35° - 45°.
Soil cover in the landslide location is light-reddish brown sandy clay
with 5 m depth on top of dark grey marl. Figure 8 shows the
geometry of the debris flow from southwest direction and
topography around Pasir Panjang Village with 1 grid representing
area 29.57 (east-west) x 25.35 (north-south) meters. The length of
debris flow in Pasir Panjang Village reached about 2.3 km.
Figure 5 Topographic area northern Jemblung Village
359
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 10 The distribution of debris flow with viscosity 0.38 at 828 5. REFERENCES
seconds that seen from top which have been overlayed with Centre
of Volcanology and Geological Hazard Mitigation (CVGHM) and Blijenberg, H.M. (2007) "Application of physical modelling of
Geospatial Information Agency (GIA) aerial mapping (2018) debris flow triggering to field conditions: Limitations posed
by boundary conditions". Engineering Geology, 91, pp 25-33.
4. CONCLUSION Centre of Volcanology and Geological Hazard Mitigation
(CVGHM) and Geospatial Information Agency (GIA) (2018)
All of the debris flow simulations were used to better understand "Mosaic Aerial Image, Recording dated February 26, 2018".
with regard of predictions of volume, run-out mechanisms, flow rate Condon, W. H., Pardyanto, L., Kentner, K. B., Amin, T. C., Gafoer,
of materials and the distribution of debris materials. Table 2 shows S., and Samodra, H. (1996) "Geologic Map of The
the viscosity, flow rate, volume of material and the length of Banjarnegara and Pekalongan Quadrangle, Java, Scale 1 :
flowing material closest to debris flow conditions of all cases. 100,000", Geological Research and Development Centre.
Table 2 Summary of simulations were closest Jacob, M., and Hungr, O. (2005) Debris-flow Hazard and Related
to debris flow conditions Phenomena. Praxis Publishing Ltd., Chichester.
Kastowo (1975) "Geologic Map of the Majenang Quadrangle, Java,
Sijeruk Jemblung Pasir Panjang Scale 1 : 100,000", Geological Research and Development
Results Centre.
Landslide Landslide Landslide
Viscosity 0.45 0.5 0.38 Nakatani, K., Wada, T., Mizuyama, T., and Satofuka, Y. (2008)
Flow rate in "Development of “Kanako” a wide use 1-D&2-D debris flow
flow tract 21.17 66.14 25.58 simulator equipped with GUI", 2nd Intl. Conf. on Debris
(km/hour) Flow, The Wessex Institute, UK.
Flow rate in Nakatani, K., Wada, T., Matsumoto, N., Satofuka, Y., and
depositional 5.4 16.53 8.68 Mizuyama, T. (2011) "Development and Application of GUI
area (km/hour) Equipped 1-D and 2-D Debris Flow Simulator, Applied to
Volume of Mixed-Size Grains", 5th International Conference on Debris-
127,735 111,962 275,295 Flow Hazards Mitigation: Mechanics, Prediction and
material (m3)
Length of Assessment, Itali, pp 735-744.
flowing material 1.1 1 2.3 Nettleton, I. M., Martin, S., Hencher, S. and Moore, R. (2005)
(km) Debris Flow Type and Mechanism in Winter, M. G.,
Macgregor, F. and Shackman, L (editor) "Scottish Road
Debris flow in the Sijeruk Village has uniqueness in the flow Network Landslide Study", Scottish Executive, Edinburg.
path, which is encounter a sharp deflection. It is alleged that the Pleiades (2004) "Pleiades Image, Date: December 16, 2014 (After
deflection is caused by a ridge that blocks the flow track of the Landslide)", Centre National d'Etudes Spatiales.
flowing material that move southeast. The ridge causes the direction
of the flow that originally toward the southeast transformed to the
east.
360
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Several flexible ring net barriers against debris flow have been installed worldwide during the last 10 years mainly for
retention and a few for erosion control in the way by reducing the inclination of the river bed. The increasing number of projects showed the
economic benefit of this solution. Special applications to retain big volumes with single barriers have been successfully installed in
Switzerland and Japan. This paper gives examples of different special applications like one big barrier providing big retention capacity,
several barriers in line or filled up barriers for erosion control. Advantages and challenges for the use of flexible ring net barriers are
discussed on a technical and economic level. Furthermore, needs for maintenance and replacement works are addressed.
Keywords: landslides, debris flow, protection, barriers, experience, maintenance
1. INTRODUCTION
Since 2005, over 250 flexible debris flow barriers have been
installed in more than 25 countries. Between 2005 and 2008, full
scale experiments at the test site Illgraben in Switzerland proved the
feasibility of retaining debris flows. The efficiency of some of the
first reference projects, mostly installed in Switzerland, was
analysed and a load design was then established together with the
Forest, Snow and Landscape Federal Institute (WSL). Standard
systems were then developed with the simulation software FARO.
Data from real-scale testing were used to verify and calibrate the
software outputs. Following this development, the flexible ring nets
became increasingly an alternative to classical debris flow
protections in Europe, USA and South America. In large scale
projects, where nets were installed in a row in the same channel, the
efficiency of retaining large volumes and the feasibility of this type
of installation in a row were proven as well. The nets are
appreciated, by designers and engineers, as a practical and
economical addition or alternative to existing classical debris flow
protections. Ten years of experience with flexible ring net barriers
signify that their advantages have been recognised and their
efficiency in the field have been established. The increasing
knowledge of single barriers, barriers in a row and large-scale
barriers have allowed to understand the advantages but also the Figure 1 Testing of debris flow retention system with ring net in the
limits of such a netting system for debris flow retention. This Illgraben channel, 2006. Retention volume approx. 1000 m3
acquired knowledge is presented in the following paper,
accompanied by case studies. On the dimensioning side, the weight acting on a debris flow net
during an event were better understood, thanks to an extensive
2. REAL-SCALE TESTING AT ILLGRABEN, measuring concept on and around the system (Wendeler, 2006),
DEVELOPMENT OF STANDARD BARRIERS AND which lead to the final dimensioning concept (Wendeler, 2008).
CE-MARKINGS
2.1.1 Remarks about 1:1 field tests results
2.1 Real-scale testing at Illgraben
Based on the described test site and recording facilities, real impact
Between 2005 and 2008, real-scale testing was conducted in the
forces on flexible barrier systems could be derived and used for the
Illgraben debris flow channel in Wallis, Switzerland (Wendeler,
development of a load model for the interaction between debris
2008). Prior testing it was observed that rockfall protection nets
flows and flexible, permeable barriers. Herein, the measured rope
were retaining some slides but the dimensioning concept was
forces during impact and overflow of debris flows present the most
missing to prove that flexible ring nets could retain larger debris
important results of the test barriers. An example is given in Figure
flows in a channel without sustaining damage. In Illgraben, a middle
2. Clearly, one can follow the filling process with the increasing
to large debris flow is occurring at least once a year naturally and
rope forces. Sudden load reductions show the activation of the brake
therefore a flexible ring net could be tested yearly (Figure 1).
elements. Even large blocks could be measured by higher weight at
Two key characteristics were defined and analysed with testing.
the balance and later by single peaks in the rope measurements. An
On one hand, a single barrier could, depending on the channel
example of a large block can be seen in Figure 3.
geometry, retain over 1000 m3. On the other hand, it was observed
If a barrier stays in use fully filled e.g. for river bed stabilization
that over 10’000 m3 were flowing over the barrier without damage.
one should consider that the remaining barrier height in the middle
This led to planning and constructing a debris flow retention system
of the barrier is only ¾ of the original barrier height.
with several nets in a row to retain successfully most of the material.
361
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
2.3 CE-marking
The real scale testing was also basis for certifying all standard
barriers. Certification was achieved in 2017 (EAD document No.
340020-00-1062). The CE marking is based on a “European
Assessment Document” which defines precisely the suitability, the
type classification and yearly quality controls necessary to
Figure 2 Rope forces and the filling height over time for the filling correspond to a certain standard. This states that the products with
event in 2006 CE marking fulfil the European guidelines for product quality and
field appropriateness (ETA 17/0268-17/0276 and ETA 17/0439).
Figure 3 demonstrates also that a debris flow can overflow a
filled flexible barrier. The overflow material is the remaining 3. DIMENSIONING
volume of a surge exceeding the retention capacity of a single
barrier which passes the barrier without causing any damages. The 3.1 Results of laboratory tests
barrier shown in Figure 3 was overflown by several 10’000 m2 In general, it is difficult to compare such laboratory tests with field
without any damage to the barrier. situation. Usually the physical modelling only provides an
informative basis on typical tendencies. After a detailed analysis of
the performed test results with a dimensional analysis the most
important result is that the stopping process of the front behaves like
an impulse of a pressure surge. Slow and friction dominated fronts
hitting the barrier showed a reflected surge. These observations are
in support of the physical behaviour of a stopping pressure surge.
The maximum dynamic force is dependent upon the velocity
squared and the flow consistency. For muddy flows, the pressure
surge needs a lower pressure coefficient Cw than for granular and
viscous flows. Additionally, flow density influences the pressure
Figure 3 Large block (left) flowing over the balance and afterwards experienced by the barrier linearly. Faster traveling fronts are
over the barrier (right) deviated vertically when they interact with the barrier and result in
material overtopping the barrier without filling it. The same effect
2.2 Development of standardized barriers also occurs for rigid barriers resulting in less retention capacity. This
observation could be confirmed because the rigid barriers were
The dimensioning concept as well as the distribution of the loading greater affected by smaller impact forces than the more flexible ones.
on the net were integrated in the finite element software FARO It therefore can be concluded that the flexible barriers are able to
(Volkwein, 2004) and first projects, mostly in Switzerland, were stop the debris flow whereas the rigid barriers can only deviate it.
dimensioned with it. Tests comparing the retention capacity of different mesh sizes in
Following the first projects, standard barriers were designed relation to the maximum particle size were conducted (Figure 5).
with a given load capacity in kN/m2. VX barriers are conceived for This enabled the determination of a good retention behaviour with a
channels up to 15m in width und barrier height of up to 6m, taking mesh size as big as the d90-grain size (90% of the grains are smaller
loads up to 160 kN/m2. UX-barriers find their application in larger than d90).
channels, are installed with additional posts, a barrier height up to 6
m and taking up loads of 180 kN/m2 (Geobrugg, 2016, Figure 4).
The dimensioning concept for debris flows is now state of the art
and freely accessible through the software DEBFLOW on Geobrugg
website. After registration on the website, everybody can use this
software and produce a first estimate for the dimensioning of a
barrier.
Figure 5 Influences of different mesh sizes on the retaining volume
362
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 7 Second surge hiding the barrier while the filling process.
The dynamic component acts one surge higher
Figure 9 FARO simulation software output when barrier number 2
3.2.3 Overflowing is impacted by an avalanche in Hasliberg, Switzerland
If a barrier is filled completely, any following surges spill over the
top of the barrier adding an additional load to the system by both In this special case, the upslope guy wires are loaded up to 70%
their weight (σ) and the acting shear forces (т). The shear force is of their capacity. Figure 10 shows the snow load on the barrier in
winter.
usually ten times smaller than the normal force and is neglected in
An easy predetermination of the dimensioning of a standard
this loading approach for flexible barrier design. The retained
material behind the barrier changes now from the hydrostatic state to barrier up to 6 m in height can easily be performed with DEBFLOW
an active earth pressure for wet material. The time it takes to drain software. A more complicated scenario can still be dimensioned by
Geobrugg or WSL with FARO simulation software. A few special
the material depends on the debris flow composition and the water
cases regarding construction are described in section 6.
content. Figure 8 illustrates the loads that must be applied.
363
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5. PLANING ASPECTS
Often debris flow barriers are installed close to the source zone of
the debris flows while greater structural measures such as a retaining
basin or deviation measures are constructed further down.
Net barriers and large steel and concrete construction can
therefore be perfectly combined. The advantages of both methods
can be specifically used together. Examples of this combination are
the streams Trachtbach in Brienz and Milibach in Hasliberg, both in
Switzerland. In both projects, the combination of the nets upstream
and the larger construction measures downstream allowed to
increase the retained mass upstream and diminish the erosion in the
stream bed. Therefore, the capacity of the concrete protection
Figure 11 Washed out anchoring of the debris flow barrier number measures could be lowered and constructed at smaller scale and
25 in the Illgraben channel. Anchoring partially in loose material existing protection structures were easily and cost effectively
and partially in disused concrete debris flow barrier renovated and added to the protection measures series.
364
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5.1 Protection nets as an immediate solution 5.3 Passages for small animals and greening
Protection nets installed in the source zones of debris flows, slow The relatively large openings of the ring nets allow for passage of
these down, which allows for longer warning and evacuation time in small animals, when the barrier is not filled, even fishes when the
the endangered areas. This is especially of importance in small barrier is immersed in water, in contrast to a concrete structure
catchment zones where debris flows are rapid and travel along short (Wendeler, 2008). There are examples where this was an expressed
distances only. The easily installed protection nets are therefore wish of the developer. Ring nets are as well appropriate for greening
practical for an immediate protection solution. They increase the and blend perfectly into the landscape.
safety of the infrastructure downstream and even allow for the
protection of the construction crew building a retaining basin for 6. DIFFERENT TYPES OF DEBRIS FLOW BARRIERS
example. These protection nets can be equipped as well with a
warning system (more details are given in section 7). 6.1 Single barriers
Most barriers installed are single barriers along roads and railway
5.2 Visual and landscape protection aspects tracks or above settlements (Figure 15).
Debris flow protection nets instead of concrete dams are more and
more an alternative regarding landscape protection and visual
aesthetics. The filigree design is almost invisible from far away and
a primary argument for protection measure construction in
landscape protection zones. An example is the UNESCO World
Cultural Heritage along the Rhine close to Koblenz (Figure 13). At
the back of the village debris flow nets are installed and even with
one barrier partially filled in 2017, the nets are still barely visible
but fulfilling their purpose (Figure 14).
Figure 14 partially filled debris flow barrier above the German Figure 16 11 debris flow protection barriers, successfully filled in
Railway close to Koblenz Hasliberg in 2011
365
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 17 Filled debris flow barrier in 2017 in Chosica Peru, 8. MAINTENANCE AND CLEANING OF BARRIERS
protecting successfully a large city downstream As any protection structure, debris flow barriers require
maintenance from time to time. It is recommended to undertake
6.3 Large debris flow retention with single barrier (special regular, for example yearly, checks of the protection system if no
construction) event (debris flow, slides, …) occurred during that time span.
In special cases, an adapted design higher than 10m and larger than Working with a checklist and a maintenance scheme, such as for any
40m can be constructed. A typical example is the debris flow barrier other protection structures, should facilitate regular controls. After
in Hüpach, next to Oberwil in the canton Berne in Switzerland an event, the barrier needs emptying and replacement of certain
(Berger et al., 2016). This barrier has a retaining capacity of more components. A filled barrier can for example be cleaned from
than 12’000 m3. Such a construction necessitates strong abutments behind with an excavator. It is essential, when planning for the
of steel reinforced concrete, long anchors and needs special ropes system, to consider what happens to the material of the debris flow
used for cable cars which need precise adjustment (Figure 18). and to organise a deposit area. Budget wise, it should be considered
Special calculations for the netting and the ropes, adjustment to the that after a fully filled barrier, parts should be replaced, whereas the
anchoring and special foundation engineering in exposed terrain was anchoring can often be reused, as explained earlier. A net can be
necessary to complete the project. emptied from the front when certain conditions are fulfilled. The
material of the debris flow should be dry and stable and the netting
must be stabilized upslope and safety aspects for the working crew
should be respected.
The decision to install a large retaining structure with netting 10. CONCLUSION
was based on the topography, the difficulty of access and lack of
Since the publication of the load design of flexible protection nets
alternatives to protect the village below. The debris flow barrier has
and their appropriateness tests in the Illgraben in Switzerland, many
not been filled yet. Another special construction is situated in
projects have been successfully installed in the last 10 years. Several
Sitäbach along the stream Lenk, in Switzerland. The construction is
construction details have been revised and improved. When
based on concrete slices and netting in between (Figure 19).
considering the hydrological processes affecting the stability of the
stream banks and planning for reinforcement, the flexible ring net
7. SURVEILLANCE
systems can be considered as equivalent to classical large concrete
Protection nets can be monitored with sensors (Sentinel System). In protection structures. The lighter conception of the barriers makes it
larger systems, some components can be monitored such as the ring an unavoidable solution when easy handling, environmental
brakes and when a loading threshold is reached, an alarm is requirements and landscape protection are key issues of a project.
triggered. An example is the debris flow net, installed as an The dimensioning concept developed at the WSL, in use worldwide,
immediate protection solution, in Magnacun in Switzerland. The has been verified by several fillings and successfully retaining
railway tracks of the Rhaetian Railway are perfectly protected since events. A further adaptation and refining of the dimensioning
2009, with the surveillance system working faultlessly, according to concept could be achieved with more testing, but is hampered by
the developer. lack of funding.
366
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
367
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: There is little study on failure of aeolian sand roadbed slope. Field test is more difficult, so the author prepares a
representative sample of aeolian sand, and makes the failure of aeolian sand roadbed slope through static load test simulation in laboratory to
analyze the aeolian sand roadbed slope stability. Test results indicate that the roadbed slope is stable when the roadbed slope ratio is 1:2 and
test load is under critical load, the aeolian sand is compacted by water dumping and mechanical compaction or mechanical compaction. The
allowable bearing capacity of roadbed increases significantly when the slope ratio is 1:2 and especially the roadbed is compacted by the
water dumping and mechanical compaction.
Keywords: aeolian sand, roadbed slope, static load test, slope stability.
50cm
roadbed filling
erosion, rain erosion and man-made damage. The design of desert
highway subgrade slope is necessary to design a reasonable slope condition 1
ratio, also to adopt reasonable and reliable protective measures. It is
difficulty to do destruction test of aeolian sand subgrade slope (Liu compaction
Dapeng, Li Zhiping), and the information of destruction test
50cm
roadbed filling
currently is lack. Therefore it is necessary to carry out simulation
tests indoors. The author choose a representative of aeolian sand condition 2
specimens, studying the sand subgrade slope under different Figure 1 Subgrade construction of two conditions
loadings and different compaction conditions in indoors. The
study’s purpose is to analyse the wind sand roadbed slope stability,
then to guide the design and application of roadbed slope Work. longitudinal center line
That is to ensure the stability of subgrade slope, to extend the useful
life of roads. The research have some economic and social benefits.
slope top
Figure 2) set the provisions when the bearing plate settlement reach
2
500 mm Load test stops. Simulation test on basis of subgrade filling
slope top 3
methods are divided into two different conditions: (1) The subgrade
4
is layered buried, and each layer thickness is 50 cm. The subgrade
have not used mechanical compaction, The simulated subgrade
slope ratios are 1:1, 1:1.5, 1:2.0 and 1:2.5 respectively; (2) The
simulation subgrade is divided into four layers, each layer of 50 cm 5
50cm
368
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(6) Jack load, using step by step loading method, is to measure force
design control load, is less than 0.1 MPa, each level increases 0.02
MPa, after each increase of about 0.04 MPa. In the experiments,
experimenter should pay attention to the top of the subgrade and
slope cracks. Then begin to the next scheduled load, and continue to
read displacement.
369
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3. COMPREHENSIVE ANALYSIS
In condition 1, 1# measuring point displacement under load larger.
When the settlement is 500 mm, the corresponding external load
varies with slope ratio Increase the trend. Load ratio is 1:2.0 and
1:2.5 when the corresponding load containing the basic approach,
mainly due to the subgrade slope ratio increased to a certain extent.
370
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
When the top of the roadbed to improve the carrying capacity less. Based on the analysis of the stratified settlement of subgrade
Slope ratio 1:1 and 1:1.5, the subgrade is first sent under the stress under two conditions, the subgrade settlement of the simulated
of the carrier board deformation of local depression, with the embankment is larger when the conditions are filled. When the
increasing of external load, the slope occurred slip and slippery vibratory compaction the subsidence of the subgrade filler in the
surface through the foot. Slope ratio of 1:2.0 and 1:2.5, the surface simulated subgrade is small. These are caused for static load on the
of the slope is 50cm away from the top of the roadbed and the aeolian sand subgrade compaction less impact, vibration load on the
bottom of the roadbed. The measuring point of the horizontal subgrade packing density had a significant impact.
displacement relative to the first two when the slope ratio is smaller,
subgrade mainly under the action of the bearing plate partial 5. REFERENCES
depression deformation (Liu Cheng, Wang Xiaodong).
In condition 2, the initial load carrying capacity of the settlement Liu Dapeng, Yang Xiaohua, Wang Jing and Li Yuan. (2016)
plate relative to the work. Condition 1 significantly reduced. This “Experimental Study on Dynamic Stress-strain Relation of
shows the compacted subgrade under the action of less than its Aeolian Sand in Xinjiang Desert Region”, Journal Of Wuhan
critical load the overall stability is good, when the upper part University Of Technology, 38, Issues 3, pp71 ~ 374.
pressure is greater than the critical load, the top of the slope of the Li Zhiping. (2003) “Experimental research on the strength of lime-
overall collapse occurred. Slippery does not pass slope foot, load is flyash stabilized aeolian sand” Transportation and
different, the shape of the sliding surface is not the same. Second Technology, Issues 2, pp30 ~ 31.
subgrade slope than 1:1 when the load-bearing plate allowed to Li Zhiyong, Cao Yuanwen, Liang Naixing and Mei Yingjun.(2006)
carry force of 81 ~ 121 kPa, slope bearing capacity of 1:1.5 “Compaction mechanism of aeolian sand” , Journal of China
allowable 64 ~115 kPa. Slope is than 1:2.0 allowable bearing Highway Journal, 19, Issues5, pp 6 ~ 11.
capacity of 75 ~ 112 kPa, slope ratio 1:2.5 allowable bearing Liu Cheng , Wu Guanling , Zhang Yanzhao , Yang Youhai.
capacity of 73 ~ 129 kPa. When the slope ratio is 1:2.5, the top of (2016)“Research on Features and Compaction Coefficient of
subgrade is under additional load sink deformation occurs, the slope Subgrade Aeolian Sand Filling for Heavy Haul Railway in
has obvious uplift (Zhang Yang). Mu Us Desert”, Subgrade Engineering, Issues 6, pp138 ~
141.
Wang Xiaodong.(2016) “Influence of The Perfomance of aeolian
4. CONCLUSIONS AND RECOMMENDATIONS Arenaceous Cemented Filling Materials on Response of
Water-Solid Ratio”, Coal Geology And Exploration, 44,
In summary, the static load on the slope of aeolian sand subgrade,
Issues 3, pp106 ~ 112.
the results of simulating indoor tests show that aeolian sand is in the
machinery Compaction or water fall plus mechanical compaction Zhang Yang , Zhang Yanzhao , Wu Guanling , Yang Youhai.
state, slope ratio of about 1:2.0, the stability of subgrade slope is less (2017)“Compaction Characteristics of Aeolian Sand and
than the critical load. When slope ratio is 1:2.0 the subgrade slope Compaction Technology for Filling Subgrade of Heavy Haul
stability is not only good, but also allow the roadbed allowing Railway”, Railway Engineering, Issues 3, pp77 ~ 80.
bearing capacity increased significantly.
371
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: In this paper, stability of an overall failure excavation case in Hang Zhou, China was analyzed using the finite element method
(FEM). The retaining system of the excavation was fully modelled, including walls, horizontal struts, and vertical center posts. For
comparison, the structural elements were simulated using plates with both elastic and elastoplastic behaviors. The soil response nesar failure
was assumed to follow the Morh-Coulomb model. Results showed that the FEM using the elastoplastic retaining system gave a more
reasonable estimate of stability of the excavation than that using the elastic support system. With the elastoplastic retaining system, yielding
firstly occurred on the wall and then on the struts, which caused large movement of surrounding soil toward the excavation. On the other
hand, with the elastic retaining system, failure of the excavation was only due to the great plastic heave of soil at the excavation bottom. The
predicted movement of soil and wall was nearly one meter as using the ealstoplatic support system but several meters as using the elastic
one.
Keywords: Deep excavations, stability analysis, finite element method, failure mechanism.
1. INTRODUCTION horizontal spacing of 2.9 m. Center posts were embedded into 0.8-
m-diameter bored piles. The site stratigraphy consists of eight
Stability of deep excavations is one of the main concerns of subsoil layers, at which most of them are soft clay, ranging from GL
practical engineers. Failure of excavations is often characterized by -2.0 m to -50.0 m. The undrained shear strength of soil obtained
collapse of the support system and the large inward movement of from field Vane shear test at two boring holes (V01 and V07) (Chen
surrounding soil. Therefore, occurrence of failure would result in et al., 2013) are also included in the figure. The ground water table
not only economic losses (e.g., destruction of neighboring facilities) is located at GL -2.0 m. As shown in Fig. 2, when collapse of
but also casualties. Most recently, a 15.7-m-deep excavation in excavation occurred, the wall had a deep inward deformation and
Hangzhou, China, collapsed because the lowest strut level was not was broken into two parts. Connections between struts and walls
installed timely, causing twenty one casualties. Based on an were failed and center posts were tilted toward the right wall. The
intensive investigation, Chen et al. (2013) predicted the failure soil heave at the excavation bottom was 2.5 m and the subsidence of
surface of soil from the disturbance of surrounding soil. the adjacent road was 7 m (Chen et al., 2013).
Stability problem of excavations has been studied by many
researchers using the FEM with reduced shear strength. A B = 21.2 m
11.9 m 9.3 m Right wall su (kPa)
parametric study by Goh (1990) showed that the factor of safety 0 20 40 60 80 100
Left wall
would be increased with the embedded depth and the stiffness of the 1 f 609, d 12 GL -0.0m 0
FILL
wall but reduced with the thickness of the clay layer below the 2 f 609, d 16 CL
3 f 609, d 16
excavation bottom. Faheem et al. (2003) also performed a OL
4 f 609, d 16
parametric study and gave consistent results with those by Goh OL 10
(1990). Furthermore, closed form equations considering the wall GL -15.7m
Depth, m
2. HANGZHOU CASE
2.1 Ground Conditions and Construction Sequence
The Hangzhou case was a 21.2- × 107.8-m internally braced
excavation, which was a part of an underground subway station. The
construction sequence and subsurface conditions of the case are
plotted in Fig. 1. The excavation had a maximum depth of 15.7 m
and was performed in five stages. The support system was
composed of a 0.8-m-thick and 33-m-deep diaphragm wall and four
steel pipe strut layers at GL -0.5 m, -4.2 m, -8.5 m, and -12.4 m with Figure 2 Failure of excavation
372
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
2.2 FE Analysis The maximum wall deflection developed near the final excavation
grade (GL -15.7 m). The maximum soil heave occurred at 4 m away
Since most of the subsoil is saturated clay, it was assumed that from the left wall.
the behavior of soil is elastoplastic and well-described by the Mohr-
Coulomb model. The model requires five input parameters,
including Young’s modulus (E), Poisson’s ratio (), friction angle
(), cohesion (c), and dilatancy angle (). Clayey soils were
simulated as undrained materials, at which u = 0.495, u = 00, cu =
su, and Eu = 450su. su was taken from the average results of Vane
shear tests.
The structural elements (struts, walls, and center posts) were
modeled using plate elements, which required four input parameters
including axial stiffness (EA), flexural rigidity (EI), maximum
(plastic) bending moment (Mp), and maximum (plastic) axial force
(Np). The input parameters of the elastoplastic structural elements
used for analysis were listed in Table 1. When the elastic structural
elements were employed, the Mp and Np values would be very high,
e.g., Mp = 1015 kNm and Np = 1015 kN (per meter width). For the
reinforced concrete wall at this case, the compressive strength of
concrete (fc’) was assumed to be 21 MPa. The Young’s modulus (E)
was equal to 15000(fc’)0.5. The Mp value was calculated using a
cross section analysis program, namely XTRACT, from wall
thickness, area of steel reinforcement, and properties of concrete and
steel. The Np value was estimated according to ACI 318-11. Since
the wall may have cracks due to bending during excavation and the
quality of concrete of the wall cast in stabilizing fluid properly does
not meet the design requirements, a reduction factor of 0.8 was
applied to the input parameters (EA, EI, Mp, and Np) of the wall. On Figure 3 Wall deflection and soil heave as using the elastoplastic (a)
the other hand, EA, EI, Mp, and Np of struts were calculated based and elastic (b) support systems
on the assumption that the Young’s modulus (E) and the yield
stress (y) of struts were 2.04x108 kN/m2 and 250 MPa, respectively. When the elastic support system was employed, the final
Due to the improper installation of struts in the field (i.e., they are excavation stage was computed completely and the SRmax value
not totally straight as splicing H steel together), their parameters was 3.07975. The wall deflection and the soil heave corresponding
were reduced by 10%. EA and Np of center posts were determined to 98%, 99%, and 100% of SRmax are plotted in Fig. 3b. As shown
similarly to those of struts. EI and Mp of center posts were not in the figure, when SR was increased, the wall deflection remained
considered to avoid any restriction on soil movement below the constant and the maximum wall deflection, about 3×104 mm,
excavation bottom. happened at the wall toe. On the other hand, the soil heave
During the strength reduction procedure, the soil strength at the developed gradually with the maximum soil heave occurring at the
final excavation stage was reduced successively by increasing the center of the excavation, up to 4×104 mm. It is observed that the
SR ratio. The divergence of numerical solutions was defined as the constant wall deflection is due to its elastic behavior whereas the
failure of the excavation. The maximum ratio, namely SRmax, at increasing large soil heave indicates the plastic behavior of soil,
which numerical solutions still converge, was treated as the factor of which then leads to the divergence of numerical solutions.
safety of the excavation. For understanding the percentage of load Fig. 4 shows the interaction diagrams of internal forces (M, N)
being applied at the final stage, Mstage is defined as the ratio of of struts and walls in the numerical analysis as using the
the load applied successfully in calculation to that caused by elastoplastic support system. As shown in the figure, when
excavation at the final stage. excavation was performed from the 1st stage to the 5th stage (final
stage), bending moment had a greater impact than axial load on the
Table 1 Input parameters (per meter width) of structural elements behaviors of struts at the 1st and 2nd layers and the wall because
Strut y f c’ EA EI Np Mp N, kN/m
1 -26000
layer No. (MPa) (MPa) (kN) (kNm2) (kN) (kNm) 2
1st strut layer
-18000 2nd strut layer
1 250 1.4x106 6.4x104 1750 256 3
4 3rd strut layer
5 -2000 wall
2 and 3 250 1.9x106 8.3x104 2310 334
5
Center post 250 2.4x105 0 288 0 4
-1000
3
Wall 40 1.9x107 106 24260 1186 4 5
1 2 2 34 5
0 M, kNm/m
-1200 -800 -400 -200 0 34 5200 400 8001200
When the elastoplastic support system was employed, the final
excavation stage was not calculated successfully and the Mstage 1000
max value was 0.6652. Fig. 3a plots the wall deflection and the soil BL of 1st strut layer
heave as Mstage is increased to 98%, 99%, and 100% of Mstage 2000 BL of 2nd and 3rd
max. It is noted that since center posts stayed closer to the left wall strut layers
than the right wall, as shown in Fig. 1, the deformation of the left 18000 BL of wall
wall was greater than that of the right wall in this analysis. Note:
26000
Therefore, only the left wall deflection was presented in Fig. 3a. As BL represents boundary lines.
shown in Fig. 3a, when Mstage increased from 98% to 99% of At the 5th stage, internal forces corresponding to SMstage
Mstage max, the wall deflection and the soil heave did not develop max are plotted.
significantly. But when Mstage reached the maximum value, the Figure 4 Interaction diagram of internal forces of elastoplastic
deformations of soil and wall increased rapidly, up to 800 mm each. structural elements
373
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
their internal force curves developed along the M axis. On the other Fig. 5b is the plastic point plot of soil at Mstage max as using
hand, since the internal force curve of struts at the 3rd layer grew the elastoplastic support system. It is observed that plastic points of
closer to the N axis than the M axis, their behaviors were strongly soil distribute in front and back of the wall but do not spread below
affected by axial load. The wall started to yield at the 4th excavation the wall toe, which is consistent with the soil movement mentioned
stage, which was earlier than the strut system. At the final stage previously. Also, the distribution of plastic points is in good
(Mstage max), both the wall and struts at the 3rd layer yielded agreement with the failures surface predicted by Chen et al. (2013)
whereas struts at the 1st and 2nd layers remained elastic behaviors. based on the soil investigation after failure of the excavation.
Fig. 5a shows the incremental displacements of soil and
elastoplastic structural elements at the last calculation step of the 3. CONCLUSION
final excavation stage corresponding to the Mstage max value. As
The following conclusions can be drawn on the basis of the work
shown in this figure, because of the soil weight, the soil behind the
presented herein:
wall moved downward and toward the wall but did not pass below
i. With the numerical analysis of the Hangzhou case, it is found
the wall toe. The wall was pushed to deform and yield at the final
that when the elastoplastic support system is employed, the
excavation grade, as marked with a rectangle and a square in the
yielding of struts and walls firstly causes a sudden increase in
figure. Due to the asymmetric characteristics of the excavation, the
the wall deflection and the soil heave, and then the failure of
left wall yielded more seriously than the right one and had an
the excavation. On the other hand, as using the elastic support
additional plastic hinge at the 3rd strut level. The inward deformation
system, the failure of the excavation is caused by the large
of the wall mainly caused the soil heave at the excavation bottom.
soil plastic heave at the excavation bottom. The amount of the
The downward sinking of the wall bent the horizontal struts, as
wall deflection and the soil heave obtained in the elastic case,
presented in the enlarged area, but the bending effect was not
i.e., soil plastic heave, was much higher than that in the
enough to cause yielding of struts at the 1st and 2nd layers. Yielding
elastoplastic case, i.e., failure of structural system, by several
of struts at the 3rd layer, as marked with a circle, on the other hand,
orders from the point of view of numerical solutions.
was a result of the large axial load transferred from the wall. It is
ii. For the Hangzhou case (with an insufficiently installed strut
noted that since center posts were supported by bored piles, which
system), the soil behind the wall moved directly toward the
were embedded deeply into subsoil, the upward movement of center
wall, pushing the wall to have a bulging deflection and yield.
posts under the effect of the soil heave was not significant. These
Then, struts yield due to the large axial load transferred from
observed deformations are in good agreement with those in the site
the wall. Yielding of struts and walls leads to the movement
(Fig. 2), at which the wall was broken, struts were destroyed, and
of surrounding soil toward the excavation zone.
center posts were tilted toward the right wall.
4. REFERENCES
American Concrete Institute (ACI), (2011) “Building code
(a) requirements for structural concrete (ACI 318-11) and
Commentary (ACI 318R-11)”. Detroit.
Left wallRight wall Chen R.P., Li Z.C., Chen Y.M., Ou C.Y., Hu Q., and Rao M.,
Legend: (2013) “Failure investigation at a collapsed deep excavation
Plastic hinge of strut in very sensitive organic soft clay”. Perform. Constr. Facil.
Plastic hinge of wall
04014078.
Spread of plastic hinge of wall
Do T.N., Ou C.Y., and Lim A., (2013) “Evaluation of factors of
Left wall Right wall safety against basal heave for deep excavations in soft clay
using the finite element method”. Geotech Geoenviron Eng
2013, 139, pp2125–2135.
Faheem H., Cai F., Ugai K., and Hagiwara T., (2003) “Two-
(b) dimensional base stability of excavations in soft soils using
Failure surface (Chen et al., 2015) FEM”. Computer and Geotechnics, 30, Issue 2, pp141–163.
Goh A.T.C., (1990) “Assessment of basal stability for braced
Figure 5 Incremental displacement plot (a) and plastic point plot (b) excavation systems using the finite element method”.
at Mstage max as using the elastoplastic support system Computer and Geotechnics, 10, Issue 4, pp325–338.
374
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Effect of the Initial Suction Boundary on the Slope Failure of Volcanic Residual Soil
Agus Setyo Muntohar1
1
Department of Civil Engineering, Universitas Muhammadiyah Yogyakarta, D.I. Yogyakarta, Indonesia
E-mail: muntohar@umy.ac.id
ABSTRACT: Rainfall-induces slope failure has been intensively investigated to obtain the failure mechanism and its affecting and controlling
factors. This paper presents a numerical analysis of the rainfall and limiting pore water pressure on the slope instability of volcanic residual
soil. The slope was natural slope where was located in Kedungrong, Kulonprogo Regency. The slope was modeled using 2D finite element
method to anlayze the rainfall infiltration and slope stability. In this study, the initial suction at the surface was limited from 1 to50 kPa with
initial groundwater depth up to 20 m below the slope surface. In general, a higher initial suction induces a delayed rainwater infiltration and
result in delayed slope failure. When the initial suction was generated from ground water table, limiting the initial suction in the numerical
analysis can provide a reasonable result to evaluate the slope failure
Keywords: initial suction, rainfall, slope stability, infiltration, pore water pressure, volcanic soil.
1. INTRODUCTION
Many landslide cases in Indonesia occurred in rainy session during
November to February. At this period, the rainfall continuously
precipitated and triggered the slope failures. Indonesia receives
significant rainfall year-round but experiences a wet season that peaks
in January and a dry season that peaks in August (Hendon, 2003; Lee,
2015). Evaluation of the rainfall triggering slope instability shows
complexity in the numerical model. Several studies indicated that
slope stability was affected by the boundary on the slope surface and
initial condition such as flux, initial suction, rainfall intensity and
duration (Tsaparas et al., 2002; Rahardjo et al., 2013; Muntohar et al.,
2013). The researcher also investigated the controlling parameter on
the rainfall-induced slope failures. Rainfall characteristics and
hydraulic properties of the soil were the most controlling parameter
that should be paid attention to the numerical model (Tsaparas et al.,
2002; Rahardjo et al., 2007).
In the 2D numerical analysis for infiltration or seepage requires the
initial conditions of pore-water pressure at the soil surface. Through
this initial condition, then pore pressures are calculated for each
change in volumetric water content during infiltration. Many
Figure 2 Location of Kedungrong landslide
researchers (e.g. Ng and Shi, 1998; Tsaparas et al., 2002; Lee et al.,
2009; Rahardjo et al., 2010) suggest that initial negative pore water
2. RESEARCH METHOD
pressure at the soil surface needs to be restricted, especially when
very deep groundwater is encountered. In a very deep groundwater 2.1 Soil and Slope Properties
table, when the initial conditions of pore water pressure are based on
hydrostatic pressure, they tend to produce unrealistic pore pressures. Both field and laboratory tests were carried out to obtain the soil
layers and the geotechnical properties. Figure 3 shows the slope
A landslide case in Kedungrong village occurred in November 2001 profile which was interpreted from hand-auger work. The slope was
after five days of intense rainfall (see Figure 1). This landslide is covered with residual soil of tuffs, weathered-breccia, and breccia-
interested to be studied since the area is populated. Figure 2 shows rock at the lowest layer. The geotechnical properties of each layer are
the location of landslides in District of Kulonprogo, Yogyakarta. This presented in Table 1. The soil-water characteristic curve and
paper is addressed to investigate the effect of limiting of initial pore hydraulic conductivity – suction relationship are shown in Figure 4.
water pressure on the rainfall-induced slope instability.
375
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 4 (a) The soil-water characteristics curve, (b) Hydraulic conductivity – suction curve, (c) numerical modelling condition
Pressure (kPa) Pressure (kPa) Pressure (kPa) Pressure (kPa)
-100 0 100 200 300 -100 0 100 200 300 -100 0 100 200 300 -100 0 100 200 300
280
Elevation, Y (m)
260
Mid Mid Mid Mid
240
220
200
Toe Toe Toe Toe
180
160
Figure 5 Initial condition of suction at the top, mid, and toe slope (a) ua-uw = 50 kPa, (b) ua-uw = 10 kPa, (c) ua-uw = 5 kPa,
and (d) ua-uw = 1 kPa
2.2 Numerical Analysis these analyses, a finite element software SEEP/W was used
(GeoStudio, 2004a). The slope stability analyses were used to study
Figure 4c illustrates the idealization of the slope for numerical the effect of seepage conditions on the factor of safety of the slope.
analysis. The finite element used quadrilateral meshes of 3 m size. The analysis was conducted by SLOPE/W (GeoStudio, 2004b).
This size was suitable to obtain numerical convergence result. For
376
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
In this study, the initial suction at the surface was limited from 1, 5, Bishop method. This analysis allowed the factor of safety varied with
10, and 50 kPa with initial groundwater depth up to 20 m below the the elapsed time as the pore water pressure changed.
slope surface (see Figure 5). The rainfall was applied on the slope
surface as unit flux (q) and the seepage review was allowed on the
surface. The pore water pressure are reviewed at the top (top), middle 3. RESULTS AND DISCUSSION
(mid) and toe slopes (toe). The pore water pressure obtained from The typical pore pressure profile of the numerical simulation is
SEEP/W analysis was then directly linked to SLOPE/W analysis. The presented in Figure 6 and 7 for initial suction 1 kPa and 50 kPa
factor safety (FS) was calculated according to limit equilibrium respectively. The pore water pressure profile clearly shows the
Figure 6 Pore water profiles for limiting suction 1 kPa (a) at top, (b) mid, and (c) toe slope
Figure 7 Pore water profiles for limiting suction 50 kPa (a) at top, (b) mid, and (c) toe slope
60
(a)
Rainfall (mm/day)
50
Infiltration (mm/day)
40
30
Infiltration rate
20
10
0
2.0 Initial Suction:
(b) 1 kPa
5 kPa
Factor of Safety
10 kPa
1.5 50 kPa
1.0
0.5
0 1 2 3 4 5 6 7 8 9
377
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
change in pore pressures against the initial conditions of suction on was generated from ground water table, limiting the initial suction in
the slope surface. After a day of rainfall (t = 1 day), it was observed the numerical analysis can provide a reasonable result to evaluate the
that the surface was saturated for all initial suction condition. The slope failure.
saturation advanced to a deeper depth with the elapsed time of
rainfall. Positive pore water pressure was build up on the interface of 5. REFERENCES
the Layer B and Layer C, because of the Layer C is almost
impervious. It allows the rainwater to flow along the interface and GeoStudio, (2004a). "Seepage modeling with SEEP/W." GEO-
induced seepage at the toe of slope (see Figure 6c and 7c). This SLOPE/W International Ltd., Canada
mechanism was introduced in Muntohar and Soebowo (2013). GeoStudio, (2004b). "Stability modeling with SLOPE/W." GEO-
SLOPE/W International Ltd., Canada
At low initial suction (Figure 5), the wetting front zone is rapidly Hendon, H. H. (2003) “Indonesian Rainfall Variability: Impacts of
achieved at low rainfall intensity and shorter duration. In contrast to ENSO and Local Air–Sea Interaction”, Journal of Climate,
a higher initial suction (Figure 6), it takes longer to form a wetting 16, pp. 1775-1790.
zone. During infiltration, rainwater displaced the air in unsaturated Lee, H.S. (2015) “General Rainfall Patterns in Indonesia and the
zone and delaying infiltration was induced. In contrast at a lower Potential Impacts of Local Season Rainfall Intensity”,
initial suction, the soil is nearly saturated. Thus, the rainfall Water, 7, pp. 1751-1768
infiltration rate is slower. At the observed point at the mid slope, Lee, L. M., Gofar, N. and Rahardjo, H. (2009), "A simple model for
infiltration rate tends to increase with increasing initial suction preliminary evaluation of rainfall-induced slope
(Figure 8a). This condition generally explains that slope surfaces with instability." Engineering Geology 108(3-4), pp: 272-285
a higher suction values tend to be near dry or dry. Thus, rainwater can Muntohar, A. S. and Soebowo, E. (2013), "Mechanism of rainfall
rapidly flow into the slope subsurface. However, the depth of the triggering landslides in Kulonprogo, Indonesia." Geo-
wetting zone achieved is relatively shallow depending the rainfall Congress 2013. Ed. C. L. Meehan, D. Pradel, M. A. Pando
intensities. and J. F. Labuzs. Reston, Virginia: American Society of
Civil Engineers, pp: 452-461.
It can been seen in Figure 8b that the factor of safety decreases as the Muntohar, A. S., Ikhsan, J. and Liao, H. J. (2013), "Influence of
rainfall intensity increases. The decrease of the factor of safety was Rainfall Patterns on the Instability of Slopes." Civil
attributed to the increase in pore water pressure (see Figure 6 and 7). Engineering Dimension 15(2), pp: 120-128
Ng and Shi (1998) explained that increase in pore water pressure Ng, C. W. W. and Shi, Q. (1998), "A Numerical Investigation of the
reduce the shear strength of the soil according to the Mohr-Coulomb Stability of Unsaturated Soil Slopes Subjected to Transient
failure criterion for unsaturated soils. Slope with a higher initial Seepage." Computers and Geotechnics 22(1), pp: 1-28
suction result in delayed slope failure because of delayed infiltration Rahardjo, H., Ong, T. H., Rezaur, R. B. and Leong, E. C. (2007),
and build up pore water pressure. The result was also explained by "Factors controlling instability of homogeneous soil slopes
Rahardjo et al. (2010). A shallower initial groundwater table under rainfall." Journal of Geotechnical and
generated a lower initial suction. As a result, the combination Geoenvironmental Engineering 133(12), pp: 1532-1543
constituted to a worst factor of safety of slope (Rahardjo et al., 2007). Rahardjo, H., Satyanaga, A. and Leong, E.-C. (2013), "Effects of flux
Thus, the determination of the initial conditions of suction greatly boundary conditions on pore-water pressure distribution in
affects the stability of the evaluated slopes. slope." Engineering Geology 165, pp: 133-142
Rahardjo, H., Satyanaga, A., Leong, E. C. and Song, N. Y. (2010),
"Effects of Groundwater Table Position and Soil Properties
4. CONCLUSION on Stability of Slope during Rainfall." Journal of
A numerical analysis has been successfully performed to investigate Geotechnical and Geoenvironmental Engineering 136(11),
the effect of limiting initial suction on the rainfall-induced slope pp: 1555–1564
failure. A remarkable results and discussion conclude that the initial Tsaparas, I., Rahardjo, H., Toll, D. G. and Leong, E. C. (2002),
conditions of suction greatly affect the stability of the evaluated "Controlling parameters for rainfall-induced landslides."
slopes. In general, a higher initial suction induces a delayed rainwater Computers and Geotechnics 29, pp: 1-27
infiltration and result in delayed slope failure. When the initial suction
378
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This paper explores a new method of in situ debris-cement mixtures for laboratory experiments in order for the reuse of
dredging debris for the ground improvement of bank in wild creeks. The paper undertook laboratory testing to determine the optimum ratios
of water, cement and in situ debris by weight which can be suitable for the field work and compressive strength of ground improvement in
wild creeks. Taking into account an economical and efficient mixing method of recycling debris in wild creeks, the optimal cement-
aggregate ratio is 1:12, slump test is 15 ± 3.8 cm, sand content is less than 50%, and unit weight of test specimen is greater than 2.07 t/m3 as
compressive strength of the admixture can be as high as 400 psi, which is suitable for erosion control and bank protection of ground
improvement in wild creeks.
Keywords: Debris-cement Mixtures, ground improvement, wild-creek remediation.
379
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
380
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
aggragate ratios. The results illustrate that compression strengths of ISDCM ground improvement. It is economically and practicably
and cement-aggregate ratios are positive correlations which means considerable for the ground improvement that the admixtures of a
the more cement contents, the stronger compressive strength of test ratio of 1:12 cement to aggregate, slump test of 15 ± 3.8 cm and
specimens are. As can be see in Figure 3, the compressive strenght sand content less than 50% can be designed to achieve 28-day
dramatically drops when sand content is approximately 60% and the compressive strength as high as 28 kgf/cm2 (400 psi) as a standard
soil class of the USCS classification is ‘SP’. It means that if sand of ISDCM method.
contents are greater than 60% and soil class is poorer than ‘SP’ in
field sites, there will be in need of increasing cememt amount to
cement-aggregate ratio of 1:8 or 1:9, for desired strength of ISDCM
method.
381
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
day test specimens which is in line of the permanent structural fill System). ASTM International, West Conshohocken, PA
designed by CLSM are a ratio of 1:9 cement to aggregate, a water- www. ASTM. org.
cement ratio of 1.25. Meanwhile, it is economically considerable for Bell, F. G. (1993). Engineering Treatment of Soils. London: CRC
the aim of river erosion control and bank protection of ISDCM Press.
ground improvement that the admixtures of a ratio of 1:12 cement to Chang, C. F. and Chen, J. W., 2006, “Development and Production
aggregate can be designed to achieve 28-day compressive strength of Ready-Mixed Soil Material”, Journal of Materials in Civil
as high as 28 kgf/cm2 (400 psi) of CLSM standards; however, it is Engineering, Vol. 18, No. 6, pp. 792-799.
not suitable for soil class ‘SP’ of the USCS classification because Das, B. M. (2016). Principles of foundation engineering (Eighth
there is a significant drop in the strength of the soil. The paper edition). Cengage learning.
suggests that if the in-situ soil or debris belongs to Class S of the Gabr, M. A., and Bowders, J. J., 2000, “Controlled Low-strength
USCS classification or sand content greater than 50%, the cement- material using fly ash and AMD sludge " , Journal of
aggregate ratio of ISDCM should be increased to 1:9 or more for the Hazardous Materials, Vol. 76, No. 2-3, pp. 251-263.
desired strength of the method. Kantoush, S. A., & Sumi, T. (2010). River morphology and
Overall, two types of ISDCM designed standards are: 1) type I sediment management strategies for sustainable reservoir in
is the compressive strength of up to 84 kgf /cm2 (1200 psi) which is Japan and European Alps.
suitable for the permanent structural fill of embankments, berms or Knighton, D. (2014). Fluvial forms and processes: a new
dams in wild creeks. The optimal design is a ratio of 1:9 cement to perspective. Routledge. London and New York.
aggregate , slump test of 15 ± 3.8 cm (corresponding to a water- Kondolf, G. M., Gao, Y., Annandale, G. W., Morris, G. L., Jiang,
cement ratio of 1.25), sand content of 20% to 40% and unit weight E., Zhang, J., ... & Hotchkiss, R. (2014). Sustainable sediment
of greater than 2.15 t/m3. 2) Type II is the compressive strength management in reservoirs and regulated rivers: Experiences
between 28 kgf /cm2 (400 psi) to 84 kgf /cm2 (1200 psi) is used for from five continents. Earth's Future, 2(5), 256-280.
erosion control or bank protection in wild creeks. The optimal Krell, W. C. (1989). Flowable fly ash. Concrete International,
design is a ratio of 1:12 cement to aggregate, slump test of 15 ± 3.8 11(11), 54-58.
cm and sand content of less than 50% and unit weight of higher than Siddique, R. (2009). Utilization of waste materials and by-products
2.07 t/m3. In conclusion, type II of ISDM is an economical, efficient in producing controlled low-strength materials. Resources,
and useful method of recycling debris and ground improvement in conservation and recycling, 54(1), 1-8.
wild creeks. In the future, the method could be applied to filed Su, N., & Miao, B. (2003). A new method for the mix design of
experiments in order to understand the practices of the optimal medium strength flowing concrete with low cement content.
relative ratio of ISDM. Cement and Concrete Composites, 25(2), 215-222.
Watanabe, F., Fukumoto, K. and Horiguchi, T., 1999, Landslide
5. REFERENCES Disaster Countermeasures Aiming at Twenty-first Century.
Construction of Debris Barrier by ISM (in-situ Mixing
ACTEC(Advanced Construction Technology Center) & ISM
Solidification) Method. Cost Reduction by Foundation
method Association Office (2017), ISM (in situ mixing)
Excavation Labor Saving, Civil Engineering Journal, 40(6):9-
Method. Retrieved from http://www.ism-method.jp/
15.
Adaska, W. S., 1997, “Controlled Low-Strength Materials”, A
Yan, D. Y., Tang, I. Y., & Lo, I. M. (2014). Development of
Report From ACI Committee 229, Concrete International,
controlled low-strength material derived from beneficial reuse
Vol. 19, No. 4, pp. 41-43.
of bottom ash and sediment for green construction.
ASTM D2487 (2011). Standard practice for classification of soils
Construction and Building Materials, 64, 201-207.
for engineering purposes (Unified Soil Classification
382
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Dynamic compaction is a ground improvement technique to densify soils and fill materials by dropping a falling weight called
pounder typically made of steel via a crawler crane. The method is to densify the compacted material so that its bearing capacity can be
increased, and the settlement can be reduced when additional loads are introduced over the compacted material. Its application has been
associated mainly to the compaction of sand deposit or sand fill, construction debris and even sanitary landfill. In this paper, the compaction
of lateritic fill using dynamic compaction is presented via two case studies. The first project is the filling of a 23 m deep valley using lateritic
soil compacted with dynamic compaction for a heavy industry factory. The second project is the application of dynamic compaction to
recompact up to 6 m deep of a filled ground where proper compaction was not done during filling for a housing development project.
Keywords: Dynamic compaction, Menard pressuremeter, lateritic soil, ground improvement.
383
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
CCM DC
Thickness of
0.3 m thick 3 to 5.15 m thick
compaction layer
Number of
> 70 layers 5 layers
compaction layer
Duration of
> 3 months 2 months
compaction
Depth (m)
These methods rely very much on the accuracy of the soil modulus
and limit pressure obtained from in-situ soil test. Therefore, Menard
pressuremeter has been specified as the pre-treatment and post- 3 3
treatment tests for this project. Based on the calculations conducted,
to comply with the above performance criteria, the compacted
laterite fill for each layer should have the following properties: 4 4
Harmonic mean pressuremeter modulus, EM ≥ 6,000 kPa;
Geometric mean limit pressure, PL ≥ 600 kPa
In this project, the spacing of the dynamic compaction (DC)
5 5
prints is 5 to 6 m. The DC works is carried out in 2 phases, the main
phase and the ironing or overlapping phase. The steel pounder
Pre-PMT-02 Pre-PMT-02
weight is 15 tonnes and the drop height is 20 m. The DC works has
been carried out using two lattice-boom cranes with capacity to lift Post-PMT-B (outside) Post-PMT-B (outside)
the 15 tonnes steel pounder up to 20 m. Figure 3 shows the DC rig
Figure 4 Pressuremeter test results at RL +50.15 m
setup for the compaction of the valley. This photograph is taken
during the DC works at RL +45.00 m, where the water ponding area
has been filled with 3 m thick of crusher-run and to be improved
with DC.
384
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The PMT test results clearly show that the DC works has Change the foundation system from pad footings to piles
significantly increase the EM and PL of the lateritic soil. The EM and foundation;
PL after DC compaction are as follows: Excavate the fill material and recompact;
Harmonic mean pressuremeter modulus, EM = 27,180 kPa; Deep compaction using dynamic compaction (DC).
Geometric mean limit pressure, PL = 3,150 kPa Eventually, DC is selected as it is the cheapest and fastest solution
These values exceed the required values (i.e., acceptance criteria) by among the options available. The deep compaction up to 6 m thick
more than 4.5 times. This means the DC has been successfully using DC is possible for this project because the filled ground is
implemented on the lateritic fill. unsaturated.
area so that the pad footings for the double storey houses and shop
houses can be constructed at any location within the building
4 footprint. The typical layout of DC prints adopted for the project is
shown in Figure 6. The spacing adopted for the main phase DC
compaction is 5.2 m centre-to-centre on a square grid and this is
5 followed by the ironing phase. A lattice-boom crane capable of
lifting 15 tonnes pounder with 20 m drop height is used for this
project to give maximum compaction energy of 300 tonnem per
6 blow.
A3 B14 C11
3.3 Pressuremeter Test Results
E62a E64
Thirteen Menard pressuremeter tests (PMT) have been carried out in
Figure 5 Mackintosh probe test results this project. The tests are carried up to 6 m or refusal depth. Figure 7
shows the 3 pre-treatment PMT conducted before dynamic
From the Mackintosh probe (MP) test results, it is concluded compaction (DC) works. The results show that the thickness of the
that the ground is non-uniformly compacted. Consequently, the fill is not uniform and ranging from about 2 to 6 m throughout the
construction of the buildings cannot be started. The developer has development area. In terms of degree of compaction, it is shown that
considered a few options to solve this issue as follows: there are localised layers within the lateritic fill which are not
385
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
properly compacted (i.e., pressuremeter modulus, EM less than 5,000 Pressuremeter modulus, EM (bar) Limit pressure, PL (bar)
kPa). For instance, at depth of 2 m for Pre-PMT-01, at depth of 1 m 0 100 200 300 400 0 10 20 30 40
0 0
for Pre-PMT-02, and at depth of 3 to 4 m for Pre-PMT-03.
1 1
2 2
Depth (m)
Depth (m)
3 3
4 4
5 5
6 6
Post-PMT-1 Post-PMT-2 Post-PMT-1 Post-PMT-2
Post-PMT-3 Post-PMT-3
4. DISCUSSION
Figure 6 Layout for dynamic compaction (DC) works Notwithstanding the uncertainness of dynamic compaction (DC) to
treat soil with high fines content, this paper has presented two case
studies of the application of DC to compact lateritic fill for property
Pressuremeter modulus, EM (bar) Limit pressure, PL (bar)
development projects. The main reason for the two projects to be
0 100 200 300 400 0 10 20 30 40
0 0 successfully implemented, despite the fill material is mainly silty
soil with fines content more than 50%, is because the lateritic fill is
unsaturated or partially saturated. Furthermore, for the case of the
1 1 deep valley compaction, additional construction measure has been
taken to minimise the effect of rainwater from wetting the lateritic
fill by placing a layer of crusher run on top of each layer of fill to be
2 2
compacted. This increases the effectiveness of the DC works
tremendously. Indeed, DC will not be the go-to solution for the
Depth (m)
Depth (m)
386
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
387
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRAK: Analisis untuk memprediksi penurunan konsolidasi umumnya dilakukan dengan menggunakan metode Terzaghi 1 dimensi.
Berdasarkan beberapa data pengamatan lapangan (settlement plate) yang dilakukan, prediksi penurunan konsolidasi dengan metode tersebut
memberikan nilai terlalu konservatif. Penyebabnya kemungkinan besar adalah masalah konsepnya yang terlalu sederhana dan nilainya
cukup jauh dengan yang terjadi di lapangan. Penurunan berdasarkan analisis 2D memberikan hasil yang lebih baik, hal itu dapat
dilihat pada perbandingan prediksi nilai penurunan berdasarkan analisis 2D elemen hingga dengan data hasil instrumentasi lapangan.
Penggunaan analisis 2D untuk sebagian para praktisi juga merupakan masalah tersendiri, selain perlu pemahaman yang cukup mendalam
juga membutuhkan software yang harganya tentu saja tidak murah. Dalam penelitian ini secara sederhana disampaikan suatu nilai rasio
penurunan, yaitu perbandingan antara nilai penurunan yang diperoleh berdasarkan analisis 1D dan nilai penurunan yang diperoleh dari hasil
pengamatan lapangan. Penelitian sebelumnya berdasarkan data laboratorium, secara tidak langsung menyebutkan nilai rasio ini antara 0.81-
1.1, untuk kondisi tanah di lokasi studi dan pendekatan yang berbeda tentu saja nilainya dapat berbeda. Dengan demikian seharusnya untuk
memperoleh nilai penurunan yang lebih sesuai dengan data pengamatan lapangan, maka para praktisi dapat menggunakan analisis 1D yang
lebih sederhana kemudian tinggal mengkalikannya dengan rasio penurunan berdasarkan penelitian ini untuk memperoleh nilai prediksi
penurunan yang lebih mendekati data pengukuran lapangan, dengan batasan-batasan tertentu. Data penyelidikan tanah (uji lapangan dan
laboratorium) dalam penelitian ini tersedia cukup baik, demikian juga data hasil pengamatan lapangan seperti settlement plate dan
piezzometer. Area yang menjadi studi kasus untuk penelitian ini adalah area Proyek Jalan Tol Palindra seksi 1. Nilai rasio penurunan juga
telah dibandingkan dengan penelitian sebelumnya untuk meningkatkan kualitas penelitian yang dilakukan.
Kata Kunci: Penurunan konsolidasi, settlement rasio, settlement plate
1. PENDAHULUAN
Mekanisme konsolidasi secara sederhana telah jelaskan oleh
Analisis dan pemodelan vacuum, terutama masalah tekanan vacuum Terzaghi (1940), dimana dalam teorinya diasumsikan struktur atau
dan distribusi tekanan vacuum, merupakan tantangan tersendiri partikel tanah saling terikat satu sama lain dengan tegangan
dalam perencanaan perbaikan tanah dengan metode vacuum molekuler tertentu dan merupakan material yang porous dengan
preloading. Pendekatan konvensional dalam analisis penurunan properties yang elastik. Pori tanah dianggap dipenuhi oleh air.
timbunan adalah dengan menggunakan analisis penurunan Beban yang diberikan kepada tanah akan mengakibatkan penurunan
konsolidasi 1D Terzaghi, yang dapat saja diterapkan dalam bertahap yang tergantung kepada rate keluarnya air dari pori tanah.
perencanaan vacuum preloading, yaitu dengan mengalikannya Terzaghi menerapkan konsep ini dalam penurunan tanah pada suatu
dengan suatu faktor yang disebut settlement ratio. kolom tanah di bawah beban konstan dan dijaga dari deformasi
lateral. Persamaan aliran air transient pada tanah butir halus adalah
sebagai berikut:
2u 2u 2u e S 0 w 0 u a v u
kx ky 2 kz 2 0 C 2 (1)
x 2
y z 1 e0 100 t e0 t t
k z (1 e 0 ) 2 u u
(2)
w0 av z 2 t t
2u u
cv (3)
z 2 t
Figure 1 Proyek Tol Palembang Indralaya
388
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Teori ini umum digunakan dalam analisis penurunan Area Tol Palindra merupakan area genangan dan rawa. Pada
mengingat metode yang diusulkan mudah dan sangat sederhana. musim hujan, kondisi muka air dapat naik hingga 1 m di atas
Namun demikian dari asumsi-asumsi yang digunakan dan permukaaan tanah. Pada kondisi banjir, area yang tergenang dapat
penyederhanaan yang dilakukan, terdapat kondisi tertentu dimana mencapai 80% dari total rencana panjang jalan tol.
metode ini tidak bisa digunakan. Seringkali terdapat perbedaan yang
cukup besar antara hasil analisis dengan metode ini dengan hasil 3. KONDISI TANAH SEKSI 1 TOL PALINDRA
pengukuran di lapangan.
Untuk settlement ratio, dalam Stapelfeldt, T (2006) disajikan Untuk Seksi 1 (STA 0+000 s.d STA 7+100) terdapat 9 titik
hubungan antara settlement ratio terhadap stress ratio, diketahui pemboran teknik. Nilai NSPT & sampel inti menunjukkan ketebalan
bahwa untuk area sekitar permukaan tanah maka nilai settlement tanah lempung lunak berkisar antara 15.5 m hingga 33.5 m.
ratio adalah sebesar 0.81, lalu nilai settlement ratio ini meningkat
seiring dengan penurunan nilai stress ratio, hingga dicapai nilai
maksimum settlement ratio sebesar 1.1 pada stress ratio 0.5.
389
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Settlement (cm)
Ratio
Zona 1-D Terzaghi Settlement Plate (b/a)
Consolidation (a) Monitoring (b)
B 115.98 108.500 0.94
9 69.69 43.300 0.62
19 66.46 63.100 0.95
12 86.68 77.600 0.90
17 132.26 84.900 0.64
10 63.97 46.600 0.73
Figure 6 Stratifikasi Tanah Seksi 1 Tol Palindra 10 48.28 46.600 0.97
6 128.35 100.025 0.78
Monitoring dengan instrumentasi dilakukan setiap hari hingga 25 85.42 72.225 0.85
semua persyaratan penerimaan penghentian vacuum dihentikan, 25 97.54 72.225 0.74
yaitu; derajat konsolidasi sudah mencapai minimum 90%, load ratio
13 113.06 116.425 1.03
juga sudah mencapai minimum 1.2 sesuai dengan persyaratan yang
13 163.21 116.425 0.71
ada dalam Speksifikasi Khusus untuk Metode Pra-beban dan
Vacuum.
390
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
391
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Soil replacement technique is the easiest and oldest method to improve the soft soils underneath the shallow foundations by
reducing the settlement by replacing weak soil (e.g. medium or soft clay and organic soils) with more competent materials such as sand,
gravel or other suitable granular materials. This paper presents the research on soft soil improvement underlain a large embankment using
Lightweight Expanded Clay Aggregate (LECA) as a replacement material which is substitute of normal aggregate. LECA has been used
successfully in geotechnical application where weight is an issue because the materials can help reduce dead loads by more than half. The
analysis of performance of LECA as a replacement material was conducted through finite elements methods by using Plaxis 3D software and
the recorded settlement magnitude then were compared to Tarzaghi’s equations for evaluating the average settlement of uniform loading. The
obtained results demonstrated that the settlement decreases with increasing of LECA replacement thickness.
Keywords: Soft clay, settlement, numerical modelling, soil improvement, replacement technique.
392
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Many studies have been conducted on soil replacement works. Most 3.1 Constitutive model
of the studies were focused on settlement of soft soil under square or
circular footings. However, a number of researchers are also It is very important to choose the right constitutive model to
interested to evaluate the settlement behaviour of soil with column represents the soft soil behaviour during conducting the numerical
raft installation (soil replacement and stone column). analysis. Numerous models have been introduced and used by
A. K. Gabr (2012) conducted a study on the effect of researchers to analyse the behaviour of soft soil stabilized by various
replacement soil depth in controlling the settlement of square methods. In order to select the right constitutive model, several
footing through centrifuge test and numerical modelling. The three famous models have been selected for analysis in this study. The
dimensional numerical analysis was carried out using the software selection of model made based on availability of soft soil and LECA
Plaxis (2005) adopting the Mohr-Coulomb soil behaviour mode properties. While the non-existent parameters will be calculated
(Gabr, 2012). The study found that, the use of replacement soil can based on correlation or indirectly calculated using others suitable
reduce the settlement magnitude. It was observed that during the parameters, and this particular matters readily available in Plaxis
experiment, when a uniform load exceeding 100kN/m2, the soil 3D. The results obtained from the analysis will be compared to
suffered complete failure if the soil replacement was less than 0.2B analytical calculation.
(B is footing width). However, the numerical model results
indicated no signs of failure for the whole range of loads and soil 3.1.1 Comparison of soil constitutive models
replacement thicknesses, where the settlement values recorded Consolidation settlement magnitude can be predicted using
increase almost linearly and do not trend towards a constant value analytical method One-dimensional Terzahgy’s Equation.
which is does not seem logical, as there is a minimum limit to the Immediate Settlements and settlements due to primary consolidation
void ratio of any soil. This may be due to the fact that no lower limit occur during construction, while settlements due to secondary
condition is assigned for the void ratio in the Mohr-Coulomb failure consolidation and creep occur after the end of construction.
surface. However, only primary consolidation settlement will be calculated
The performance of partial replacement for soft clay under since, the immediate settlement is insignificant in clay and can be
square footing has been analysed using finite elements methods ignored, while secondary consolidation settlement is more important
(PLAXIS 3D program), where clay soil and sand material were in highly compressibility clay and organic soils such as peat. In over
simulated by the standards of linear elastic-perfectly plastic Mohr- consolidated inorganic clay, the secondary compression index is
Coulomb theory (Abbas, 2016). Sand was used as a replacement very small, of less practical significance (Das, 2015).
materials. The settlement values recorded were compared to Janbu The models selected for analysis are; Mohr Columb Model
equations which is adopted to calculate the average settlement of (MC), Linear Elastic Model (LE), Soft Soil Model (SSM), Soil
flexible foundations on saturated clay soils. The obtained results Hardening Model (SHM) and Soft Soil Creep Model (SSC). From
demonstrated that the stress-settlement behaviour increases with the sensitivity analysis conducted, Soil Hardening Model and Soft
increasing the thickness of sand layer. Similar findings were found Soil Creep Model (SSC) gave a closed result to analytical
by Abdel Salam through the study on the effect of using different calculation. However, since creep behaviour is only significant in
types and thickness of replacement layer on increasing bearing organic soft soil and peat, therefor, Soil Hardening Model will be
capacity and reducing consolidation settlement of soft clayey soil adopted to analysis the behaviour of soft soil in this study.
experimentally, where the study concluded that the vertical For untreated soil, the settlement magnitude calculated is
settlement decreased with increasing of the replacement layer 0.3153m (for 50kN/m2 uniform load) Plaxis 3D provided a result
thickness (S. Abdel Salam, 2007). slightly higher, as predicted by SS model, while MC model
The artificial neural networks (ANNs), and the multi-linear predicted lower value up to 40%. Compare to other models, SSC
regression model (MLR) have been utilised to predict the bearing and SHM estimated closed value, however SHM model provided
capacity of circular shallow footings supported by layers of granular almost equal to analytical value (0.3209m) where the percentage
replacement over natural clay soil (Ornek, Laman, Demir, & Yildiz, was less than 2%. However, since the sensitivity analysis is based
2012). The data used in running the network models have been on analytical calculations, the physical model should be carried out
obtained from an extensive series of field tests, including large-scale to see the accuracy of the settlement values estimated by each
footing diameters. The results indicate that the use of granular fill model.
layers over natural clay soil has a considerable effect on the bearing
capacity characteristics and that the ANN model serves as a simple 3.2 Methodology of numerical analysis for LECA
and reliable tool for predicting the bearing capacity of circular replacement
footings in stabilized natural clay soil. In addition, the field test
results indicate that the use of partially replaced granular-fill layers The numerical analysis of LECA replacement has been performed
over natural clay soil has considerable effects on the bearing based on following conclusions and limitations;
capacity and the settlement characteristics. For a given value of Mesh refined used in the analysis: Fine Mesh.
replacement thickness to footing diameter ratio, H/D, the ultimate Since the replacement of soft soil by LECA will be
bearing capacity increases in a nonlinear manner with the footing constructed on very large area with uniform distributed load,
diameter. On the other hand, for a given value of D, the magnitude the boundary distance is not critical, the replacement was
of qu increased with the increase in the compacted granular-fill layer modelled using three dimensional.
thickness, H (Ornek, Laman, Yildiz, & Demir, 2012). Type of analysis: Drained analysis was selected in this study
to reduce the time consumed and to simulate the soft soil
3. METHODOLOGY behaviour in long-term, and also to allow for a greater
number of sensitivity and parametric analyse to be done.
The objective of the study is to evaluate the effects of a LECA Type of constitutive model used in the study; MC (drained)
aggregates thickness, H as soil replacement substitute the normal model for LECA and SHM to represent the soft clay
aggregate in reducing compressibility, S of soft clay soil through behaviour.
numerical modelling. Three dimensional numerical analysis will be
Construction effect was neglected in the analysis.
performed using commercial software Plaxis 3D (2017). To permit
The total load applied on the structural system is assumed to
timely analysis in this research and to simulate long term soil
be uniformly distributed over the entire surface area of the
behaviour, drained analysis is adopted to allow for a greater number
LECA replacement surface and plate will be assigned as a
of sensitivity and parametric analysis to be done. Sensitivity analysis
medium for load transfer to the above ground. The uniform
needs to be conducted before numerical modelling is carried out.
load imposed used in this study are 50, 100 and 150 kN/m2.
393
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The analysis of settlement was focused to the centre of load, replacement technique. The results show that the increasing of depth
therefor settlement behaviour at the edge of surface loading of replacement will contribute to decreasing of settlement magnitude
was not include in the study. for all unit weight of LECA as well as normal aggregates. LECA 3,
Unit weight of LECA: from the review of the previous study LECA 5, LECA 7 and LECA 9 represent the LECA with unit
and numerous LECA manufacturer through the world, the weight of 3kN/m3, 5kN/m3, 7kN/m3 and 9kN/m3, respectively, while
unit weight of LECA was found to be various between 3 to NA is normal aggregate. This finding is in good agreement with
10 kN/m3, which is lower than water density. Therefore, the previous study conducted utilizing other granular materials such as
settlement of soft clay will be evaluated based on the sand and gravel as a replacement material.
different value of LECA density (3kN/m3, 5kN/m3, 7kN/m3
and 9kN/m3).
The analysis of normal aggregate used as a replacement
material was also performed as comparison.
Depth of soft soil is 10 meter.
Settlement analysis was analysed based on different depth of
replacement; which are 1.5m, 2.5m and 3.5m.
Two stage of calculation involved; initial and loading phase
as shows in Table 1;
394
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
395
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
6. ACKNOWLEDGEMENT
The authors wish to thanks to Universiti Teknologi Mara,Pulau
Pinang and Universiti Teknologi Malaysia for providing facilities
during the research work and Ministry of Higher Education for
granting the financial for this research work.
7. REFERENCES
Figure 10 Stress-settlement behaviour of improved soft clay Abbas, B. J. (2016). The Settlement Evaluation of Improved Soft
Clay using Sand Columns and Partial Replacement
In addition, stress ratio was found to be increases with Technique. International Journal of Engineering Research &
increasing of replacement depth as presented in Figure 11. Stress Technology (IJERT), 5(7), 348–355. Retrieved from internal-
ratio is defined as the ratio of the maximum shear stress of an pdf://79.23.57.127/The Settlement Evaluation of Improved
untreated soil to the shear stress of a treated soil. This finding were Soft Cla.pdf
contrast with normal aggregate replacement, where the stress Das, B. M. (2010). Principle of Foundation Engineering. USA:
resistance will increased up to 12% with increasing of replacement Cengage Learning.
layer under 50kN/m2 uniform load. However, NA replacement Das, B. M. (2015). Principles of Foundation Engineering, SI. (7th
indicates a slight increase (less than 5%) of shear stress when 100 Edition, Ed.). USA: Cengage Learning.
and 150kN/m2 load are applied. Gaafer, M., Bassioni, H., & Mostafa, T. (2015). Soil Improvement
Techniques. International Journal of Scientific &
Engineering Research, 6(12), 217–222.
Gabr, A. K. (2012). The Uncertainties of Using Replacement Soil in
Controlling Settlemen. Journal of American Science, 8(12),
662–666.
Han, J. (2015). Principles and Practice of Ground Improvement.
John Wiley & Sons, Inc. Hoboken, New Jersey.
Holm, T. A., & Valsangkar, A. J. (1993). Lightweight Aggregate
Soil mechanics: Properties and Applications. Transportation
396
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
397
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Performance of Helix Piled Raft Foundation in Tropical Fibrious Peat Soil Under
Traffic Loads
A. Ardy1, S. Lawalenna1, B.M. Ahmad1, B. Eko1, B.F. Andi1
1
Department of Civil Engineering, Faculty of Engineering, Hasanuddin University, Makassar, Indonesia
E-mail: ardy.arsyad@unhas.ac.id
ABSTRACT: This paper presents the investigation of settlement based performance of helix piled raft foundation under traffic load through
analytical and numerical methods. The analytical method is based on the concept of PDR analysis of piled raft where the stiffness of helix
piled raft is computed from the stiffness of helix pile and raft with a certain helix piled – raft interaction factor. In this study, the stiffness of
helix pile to peat soil is estimated by using a modified Randolph and Wroth equation, incorporating the mechanical behaviour of helix pile.
On the other hand, the stiffness of raft is determined by using Richart et al equation. In the numerical model, 3-dimension FEM model of
helix piled raft is undertaken in which helix pile is modeled as fixed end anchor, and raft as a plate above soft peat soil. The effect of helix
pile number on the bearing capacity and settlement of helix piled raft foundation system is investigated. The results reveal that the stiffness
of helix pile is too small compared to the stiffness of raft with the ratio from 1/6 to 1/4. The helix pile stifness is much influenced by the
number of helice and the radius of helice in a single helix pile. Since the stiffness of helix pile is small, such critical number of helix piles is
needed to have effect on the foundation system. Below this number, helix pile is insignificant to have effect in the helix piled raft foundation
system, and the elasticity of the raft plays important role. In general, however, it is obvious that the increasing number of helix piles can
decrease the elastic settlement of helix piled raft foundation when it is subjected with traffic load. In the stiffness ratio of helix pile to raft at
1/6, the increasing number of helix pile can reduce the settlement up to 80%, whereas that in the stiffness ratio helix pile to raft at ¼ can
reduce the settlement by 65%. These finding s would beneficial for development of alternative helix piled raft foundation for road
infrastructure in tropical fibrious peat soil.
Keywords: tropical fibrious peat soil, helix piled raft, stiffness, elastic settlement.
398
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3. RESULT
Analysis of bearing capacity and settlement of helix piled raft in
peat soil under axial loading was conducted by incorporating
modified Randolph and Wroth (1978) method into Randolph (1994),
Poulos and Davis (1980). FEM analysis with PLAXIS 3D was also
undertaken to comprehend the result.
K
1 0.6 r
K
K pr hp K (4)
Kr
hp
1 0.64
K
hp
0.2 Kr (5)
a
K r K hp
1 0.8
K
hp
Php hp P Pup (7) Raft stiffness was estimated in this study. For the raft with area
of 100 m2, in peat soil with shear modulus of 65.217 kN/m2 and v of
0.15, is 1,687 kN/m. The raft stiffness increases as the dimension of
Pr P Php (8) the raft and the shear modulus of peat soil aslo increase. Fot the raft
where Php is load on helix piles, Pr is load on the raft, Pup is the with area of 400 m2 and the shear modulus at 173.913 kN/m2, the
ultimate bearing capacity of helix pile group. raft stiffness would be at 8,184 kN/m, and that with area of 900 m2,
the stiffness is 12,276 kN/m.
399
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
400
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5. REFERENCES
Arsyad, A., Hamid, W., Yusmin, A. (2014). Case study of
reinforced embankment over a deep softsoil. Proceedings of
Softsoils, 21(23rd).
Arup Geotechnics, (2005). Design of Screw Piles: Assessment of
Pile Design Methodology. Ove Arup & Partners Ltd, London.
Edil T.B., Dhowian A.W. (1981). At-rest lateral pressure of peat
soils. Proc. Conf. Sediment. Consolidation Model, ASCE,
San Fransisco, pp. 411-424.
Huat, B.B.K. Prasad, A., Asadi, A., Kazemian,S. (2014).
Geotechnics of organic soils and peat. CRC Press, p.250.
Islam S, Hashim R (2008b). Engineering properties of peat soils in
Figure 8 The effect of number of helix piles on the elastic peninsular Malaysia. J. Appl. Sci., 8 (22).
settlement of helix piled raft foundation obtained from FEM model, Kazemian, S., Huat, B., Prasad, A. and Barghchi, M., (2011). “A
in very soft peat soil. State of Art Review of Peat: Geotechnical Engineering
Perspective”, International Journal of Physical Sciences, Vol.
6(8), 2011, pp. 1974-1981.
Munro, N. (2005). Dealing With Bearing Capacity Problems on
Low Volume Roads Constructed on Peat. ROADEX II
Northern Periphery.
Mesri G., Ajlouni M. (2007). Engineering properties of fibrous
peats. J. Geotech. Geoenv. Eng., 133(7): 851-866.
Mooney, J.S., Adamczak, S., Clemence, S.P., (1985). Uplift
Capacity of Helix Anchors in Clay and Silt. American
Society of Civil Engineers48–72.
Mohajerani, A., Bosnjak, D., Bromwich, D. (2016), Analysis and
design methods of screw piles: A review, Soils and
Foundations, Volume 56, Issue 1, Pages 115-128.
Narasimha Rao, S., Prasad, Y.V.S.N., Veeresh, C., (1993).
Behaviour of embedded screw anchors in soft clays.
Figure 9 The effect of number of helix piles on the elastic
Geotechnique 43, 605–614.
settlement of helix piled raft foundation obtained from FEM model,
Nasr, M.H., 2009. Performance-based design for helical piles. In:
in stiff peat soil.
Contemporary Topics in Deep Foundations. American
Society of Civil Engineers, USA, pp. 496–503.
4. CONCLUSION
Poulos, H.G., (2001) Piled raft foundations: design and applications,
- The increase of number of helix piles under raft foundation in Géotechnique 51:2, 95-113
peat soil can reduce the elastic settlement of the foundation from Poulos,H.G., and Davis, E.H. (1980). Pile Foundation, Analysis and
65% to 80%. Design. University of Sidney.
- Critical number of helix pile is found at about 10 piles to 21 Richart, F.E., Jr., J. R. Hall, Jr. R. D. Woods.Richart, F. E. (Frank
piles, depending of the stiffness of helix pile. Beyond this Edwin), (1970). Vibrations of soils and foundations. Prentice
number, the reduction of elastic settlement is less significant. Hall, p. 414.
From FEM Model, the critical number of helix pile is from 9 to Randolph, M.F., (1994). Design methods for pile groups and piled
15 piles, depending helix pile stiffness. rafts Proc. 13th International Conf. on Soil Mechanics and
- Load sharing carried by helix pile varies from 10% (1 pile) to Foundation Engineering, New Delhi (5):61-82.
65% (25 piles). This is influenced by the ratio stiffness of helix Randolph, M. F., & Wroth, C. P. (1978). Analysis of deformation of
pile to raft. The larger stiffness ratio of helix pile to raft, the vertically loaded piles. Journal of Geotechnical and
larger load sharing carried by helix pile will be. Geoenvironmental Engineering, 104(ASCE 14262).
- Further study is needed to investigate the mechanical behaviour Schmidt, R., Nasr, M., (2004). Screw piles: uses and considerations.
of helix pile raft through experimental model and FEM with Struct. Mag., 29–31.
SSC model.
401
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Haiphong city is situated on the north-eastern coast of Vietnam. This city has a large coastal and sea area, which is advantage
for marine economic development. However, in order to reclamation the coastal area, river sand material is much demand which lead to face
some environmental problems due to exploitation of river sand. Thus, the intensive laboratory experiments were conducted including
physical, chemical and mechanical tests of river sand and sea sand to evaluate the potential use of sea sand in road embankment in Haiphong
city. The samples were prepared by compacting to value of 90% and 95% of maximum density defined by the Standard Proctor test.
Consolidated undrained cyclic triaxial tests were conducted for sea sand samples and river sand samples. The cyclic triaxial test results
showed that all samples are not liquefied under normal vehicle conditions which corresponding to cyclic deviator stress about of 8 kPa and
frequency of 1Hz. In addition, sea sand samples were not liquefied when increasing the compaction effort to 95% of maximum density.
Therefore, it is possible to use sea sand instead of river sand in some cases for reclamation land in Haiphong area.
Keywords: dynamic properties, road embankment, CU triaxial test, liquefaction
Sea sand
Test (N1 or (N1)60) , Cone Penetration Test (CPT), and relied on 14 S05 B14 0.90 1.597 2.675
experience from the earthquake occurred during the past (Ishihara, 15 S01 B15 0.95 1.682 2.675
1977; Bolton, M. and Ignacio, 1983; P. and J., 1995; Jakka, Datta 16 S03 B16 0.95 1.635 2.670
and Ramana, 2010; Y. and J., 2014). 17 S04 B17 0.95 1.676 2.674
The primary aim of this study is to determine the liquefaction 18 S05 B18 0.95 1.676 2.658
capacity of the sea sand and river sand samples at density 19 S06 B19 0.95 1.645 2.669
corresponding to 90% and 95 % of maximum dry density. 20 S07 B20 0.95 1.649 2.670
A series of undrained cyclic triaxial tests were conducted for sea RS-
21 0.95 1.651 2.590
River sand
sand and river sand samples which were taken in Hai Phong city S01
area. These tests were considered the effect of compaction, RS-
frequency, amplitude and stress ratio to study the dynamic 22 0.90 1.713 2.570
S02
characteristic of sand fill in loading condition due to vehicle loading. RS-
23 0.90 1.681 2.590
S03
2. MATERIALS AND METHODS
2.1 Material 2.2 Method of undrained cyclic triaxial test
The sea sand and river sand material are taken from Hai Phong city. 2.2.1. Cyclic Triaxial equipment
The characteristic of materials is shown in Table 1. The sea sand
and river sand samples are prepared by compacting to reach the The cyclic Triaxial test was conducted at Laboratory of
compaction density of 90% and 95% of maximum dry density. Geotechnical Engineering Department, Hanoi University of Mining
Compaction ratio is defined of ratio between dry density for and Geology. The name of equipment for cyclic triaxial test is
prepared sample and maximum dry density which were conducted Tritech 100 from Controls – Group Italia. The advanced of the
by Standard Proctor test. The maximum dry density was from 1.584 machine is following:
to 1.644 g/cm3 and from 1.651 to 1.713 g/cm3 for sea sand and river This equipment can determine in both the static condition
sand, respectively. (effective stress and stress line) and cyclic condition.
Table 1 Initial characteristic of sand sample It remedies automatically the back pressure/ cell pressure
No. Group Sample Compaction ᵞd max ᵞs Note during cyclic triaxial test
of Ratio (R) g/cm3 g/cm3 Maximum control frequency is 10 Hz (based on test condition)
Sample
1 B01 0.95 1.644 2.58 Two load option/ dynamic deformation
2 S04 B02 0.90 1.644 2.58 - Maximum is ±25 mm with ±5 kN double-effect actuator
3 B03 0.95 1.644 2.58
Sea sand
4 B04 0.90 1.631 2.58 - Maximum is ±15 mm with ±14 kN double-effect actuator
5 S01 B05 0.90 1.631 2.58 2.3.1. Testing preparation.
6 B06 0.90 1.631 2.58
7 B07 0.90 1.584 2.57 Samples with 70 mm in diameter and 140 mm in height, were
8 S02 B08 0.90 1.584 2.57 compacted with optimal moisture content to ensure that the sample
9 B09 0.90 1.584 2.57 were reached the density ratio corresponding with 90% and 95% of
402
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
maximum dry density. All samples were conducted in undrained a = .Px = 0.5 x16 = 8 kPa
condition accordance to ASTM D 5311-92-1996.
where:
2.3.2. Testing procedure.
- : loss dynamic stress factor from road surface to the sand layer
The samples were tested by the cyclic triaxial test and listed in Table at 1.5m depth with = 0.5
2. The sample goes through into 3 steps: saturation process,
consolidation process and after finish consolidation steps, the vale is - Px: calculated vehicle loading, Px is taken as 16 kN/m3
closed to ensure that no water dissipated during the loading process. Thus, the effect of stress amplitude and frequency were
In the final step, the sample was subjected dynamic loading by considered with the following conditions:
acting the cyclic load with different amplitude of shear strain and
changing frequency. During the test, pore water pressure, cell Effect of stress amplitude on liquefaction:
pressure, axial strain and axial stress will be read and noted
continuously. The testing procedure is illustrated as below: Samples B1-B6 and Frequency f= 1Hz
Change of Stress amplitude range: 8± 32(kPa)
- Saturation method step: Using back pressure (back) of 90 kPa
and cell pressure cell of 100 kPa Effect of frequency on liquefaction:
To determine the stability of material when the vehicles moving
- Consolidated step: The sample is consolidated with cell pressure
on the road with difference speed with change of frequency but
(cell ) of 170 kPa; back pressure (back) of 130 kPa and the
stress amplitude remains constant.
effective stress (3’) of 40 kPa.
Frequency f (Hz) = 1-8 Hz
- Shearing step: Dynamic loading will be applied
Change of Stress amplitude a = ± 8(kPa)
Table 2 The cyclic triaxial test for samples
The shear strength of sand when liquefied is evaluated by
Sample Freq Cell Back Eff Stress Cyclic
(f) pressure pressure stress
conducting cyclic triaxial test in undrained test condition. In this
amplitude stress
3 (kPa) b ’c (kPa) Ratio test, the sand sample is saturated and consolidated under constant
(kPa) (kPa) CSR cell pressure and amplitude repeated stresses until the sample is
B01 1 170 130 40 0.4 deformed, or the pore water pressure reaches similar value with
32 initial cell pressure. In this state, the sample can be considered at
B02 1 170 130 40 24 0.3 unstable dynamic state, and then the liquefaction occurred.
B03 1 170 130 40 16 0.2 Normally, the liquefaction phenomena can be determined by 2 ways
B04 1 170 130 40 8 0.1 (Figure 1):
B05 1 170 130 40 16 0.2
(1) The increase of pore water pressure reaches 95 % of initial
B06 1 170 130 40 24 0.3
stress,
B07 1 170 130 40 8 0.1
B08 2 170 130 40 8 0.1 (2) or (2) The development of large deformation, are described
B09 3 170 130 40 8 0.1 with amplitude corresponding with 5% of axial strain.
B10 1 170 130 40 8 0.1
For establishing the liquefaction of specific sample (up to 95%
B11 2 170 130 40 8 0.1 pore water pressure or 5% amplitude axial strain), the number of
B12 3 170 130 40 8 0.1 cycles must be indicated in one specify case with the uniform
B13 1 170 130 40 16 0.13 dynamic loading amplitude. It is noted that to reach the liquefaction
B14 1 170 130 40 16 0.1 stage corresponding to the 95 % of pore water pressure ratio and 5%
B15 1 170 130 40 16 0.2 of axial shear strain could not occur at the same time, then two
B16 3 170 130 40 16 0.13 separate curves can be drawn.
B17 8 170 130 40 16 0.1
B18 1 170 130 40 24 0.16
B19 3 170 130 40 24 0.2
B20 8 170 130 40 24 0.2
RS- 1 170 130 40 8 0.1
S01
RS- 1 170 130 40 8 0.1
S02
RS- 1 170 130 40 16 0.13
S03
403
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
100
kPa
100 Comment [ln1]: Đường mầu
90 90 xanh bên dưới bị lỗi
80 80 Ngoài ra đường ký hiệu em
70 70 cho thưa ra (khoảng cách giữa
60 Ru (%) 60 dấu tròn hoặc dấu tam giác
cho thưa ra)
Excess Pore Pressure (kPa)
Ru (%)
50 Compression Stress (kPa) 50
Extension Stress (kPa)
40 40
30 30
20 20
10 10
Figure 1 Liquefaction evaluation method 0 0
0 40 80 120 160 200
Cycle
3. RESULTS AND DISCUSSION Figure 3: Test results of sample B01, SR = 0.4; f = 1HZ, stress
3.1 Physical properties of sand amplitude a =32 kPa
Figure 2 shows particle size distribution of sea sand and river sand.
It is observed that the particle sizes of river sand samples are larger 100
kPa
100 Comment [ln2]: ĐƯờng màu
than those of sea sand samples. 90 90 xanh bị lỗi
Ru (%)
80 Excess Pore Pressure (kPa) 80
100
Compression Stress (kPa)
90 70 70
Extensive Stress (kPa)
80 60 60
Ru (%)
Percent Passing (%)
70 50 50
60 River sand sample RS-S01 40 40
River sand sample RS-S02
50 River sand sample RS-S03 30 30
Sea sand sample S01
40 Sea sand sample S02 20 20
Sea sand sample S03
30 Sea sand sample S04 10 10
Sea sand sample S05
20 Sea sand sample S06 0 0
Sea sand sample S07
10 0 20 40 60 80 100
Cycle
0 Figure 4: Test results of sample B02, SR = 0.3; f = 1HZ, stress
0.1 1 10
Seivesize (mm) amplitude a =24 kPa
Figure 2 Particle size distribution of sea sand and river sand samples
(S01 to S07 are group name of sea sand samples, see in Table 1) 100
kPa
100
3.2 Effect of stress amplitude on liquefaction capacity 90 Ru (%) 90
Excess Pore Pressure (kPa)
80 Compressive Stress (kPa) 80
To study the failure mechanism of sea sand sample, the number of Extensive stress (kPa)
70
70
dynamic loading tests were conducted with difference amplitude (8-
60 60
12kPa), corresponding to the maximum value that can be reached
50
Ru (%)
for fill embankment location at 1.5 depth which is below pavement 50
surface. 40 40
In case of amplitude value increasing from 8 to 32 kPa with the 30 30
frequency (f) of 1Hz, the results are presented in Fig. 3 to 8. From 20 20
Table 3, conclusions are given following: 10 10
0 0
Samples with R95 are not liquefied even the sample B01 0 200 400 600 800 1000
Cycle
having the large vibration amplitude of 32 kPa.
Figure 5: Test results of sample B03, SR = 0.2; f = 1HZ, stress
Samples with R90 are not liquefied even the sample B04
amplitude a = 16kPa
having the vibration amplitude of 8 kPa.
However, it is observed that the sea sand samples were
liquefied for R90 samples at amplitude of 16 and 24 kPa (see kPa Comment [ln3]: Vẽ lại theo
100 100
Figure 7 and 8). phần mềm
90 Excess pore pressure (kPa) 90
Table 3: Test result of various amplitude 80 Ru (%)
80
Extensive Stress (kPa)
70 Compressive Stress (kPa) 70
Sample Density f (Hz) a Ru Comment 60 60
Ru (%)
index (kPa) 50 50
B01 0.95 1 32 77.5 Not 40 40
liquefied 30 30
B02 0.90 1 24 95.0 Liquefied 20 20
B03 0.95 1 16 21.0 Not 10 10
liquefied 0 0
B04 0.90 1 8 13.5 Not 0 30 60
Cycle
90 120 150
liquefied
Figure 6: Test results of sample B04, SR = 0.1; f = 1HZ, stress
B05 0.90 1 16 100.0 Liquefied
B06 0.90 1 24 100.0 Liquefied amplitude a = 8 kPa
404
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
100
kPa Table 5: Test result for river sand Comment [ln4]: Đường màu
100
90 Ru (%) 90 f a Stress xanh bị lỗi
Excess Pore Pressure (kPa)
Sample Density Ru Comment
80 Compressive Stress (kPa)
Extensive Stress (kPa)
80 (Hz) (kPa) ratio
70 70 RS-01 0.95 1 8 0.1 32 Not liquefied
60 60
8.2
RS-02 0.9 1 8 0.1 Not liquefied
Ru (%)
50 50
5
40 40
30
RS-03 0.90 1 16 0.2 60 Not liquefied
30
20 20
kPa
10 10 100 100
0 0 90 90
Ru (%)
0 50 100 150 200
Cycle 80 Excess Pore Pressure (kPa) 80
Compressive Stress (kPa)
Figure 7: Test results of sample B05, SR = 0.2; f = 1HZ, stress 70 Extensive Stress (kPa) 70
amplitude a = 16 kPa 60 60
Ru (%)
50 50
40 40
100
kPa
100 30 30 Comment [ln5]: Đường xanh
90 90 20 20 lá cây bị lỗi
80 80 10 10
70 70 0 0
0 200 400 600 800 1000
60 60 Cycle
Ru (%)
Ru (%)
50 50
amplitude a = 24 kPa 40 40
30 30
3.2 Effect of frequency on liquefaction capacity 20 20
10 10
The vibration amplitude is remained constant in this case, to study 0 0
vehicles capacity moving on the road with difference velocity, the 0 200 400 Cycle 600 800 1000
frequency f will be changed to check the liquefaction capacity
Figure 3: Test results of sample B08, f = 2 HZ, R90
corresponding to the samples having R90 and R95. The test results
in this case are shown from Figure 9 to Figure 24.
kPa
From the test results shown in Table 3 and 4, it is clear that in 100 100
the case of the amplitude remained constant to 8 kPa, the frequency 90 Ru (%)
Excess Pore Pressure (kPa)
90
varies from 1 to 3 Hz, the pore pressure ratio is less than 95%, all 80 Compressive Stress (kPa)
Extensive Stress (kPa)
80
the sea sand samples are not liquefied. 70 70
60 60
Table 4: Test result in case of different frequency of vehicles
Ru (%)
50 50
Sample Density f (Hz) a Ru Comment 40 40
index (kPa) 30 30
B07 0.90 1 8 4.5 Not liquefied 20 20
10 10
B08 0.90 2 8 4.0 Not liquefied 0 0
0 200 400 600 800 1000
B09 0.90 3 8 74.3 Not liquefied Cycle
B10 0.95 1 8 13.0 Not liquefied Figure 4: Test results of sample B09, f = 3 HZ, R90
40 40
B16 0.95 3 16 100.0 Liquefied
30 30
B17 0.95 8 16 41.5 Not liquefied 20 20
10 10
B18 0.95 1 24 100 Liquefied
0 0
B19 0.95 3 24 68 Not liquefied 0 200 400 Cycle 600 800 1000
405
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
kPa kPa
100 100 100 100
90 Ru (%) 90 90 90
Excess Pore Pressure (kPa) Ru (%)
80 Compressive Stress (kPa) 80 80 Excess Pore Pressure (kPa) 80
Extensive Stress (kPa) Compressive Stress (kPa)
70 70 70 Extensive Stress (kPa) 70
60 60 60 60
Ru (%)
Ru (%)
50 50 50 50
40 40 40 40
30 30 30 30
20 20 20 20
10 10 10 10
0 0 0 0
0 50 100 150 200 250 300 350 400 0 200 400 Cycle 600 800 1000
Cycle
Figure 6: Test results of sample B11, f = 2 HZ, R90 Figure 10: Test results of sample B15, f = 1 HZ, R95, Amplitude 16
kPa
kPa 100
100
90 90 100
kPa
100 Comment [ln7]: ĐƯờng màu
80
Ru (%)
Excess Pore Pressure (kPa) 80 90 90 trắng và xanh nước biển bị lỗi
Compressive Stress (kPa) Ru (%)
70 Extensive Stress (kPa) 70 80 Excess Pore Pressure (kPa) 80
Compressive Stress (kPa)
60 60 70 Extensive Stress (kPa) 70
Ru (%)
50 50 60 60
Ru (%)
40 40 50 50
30 30 40 40
20 20 30 30
10 20 20
10
10 10
0 0
0 200 400 600 800 1000 0 0
Cycle 0 50 100 150 200 250 300 350
Cycle
Figure 7: Test results of sample B12, f = 3 HZ, R95
Figure 11: Test results of sample B16, f = 3 HZ, R95, Amplitude 16
kPa
100
kPa
100 Comment [ln6]: Vẽ lại theo
90
phần mềm
90 kPa
100 100
80 80
90 Ru (%) 90
70 70 Excess Pore Pressure (kPa)
80 Compressive stress (kPa) 80
Extensive Stress (kPa)
60 60
Ru (%)
70 70
50 50 60 60
Ru (%)
40 40 50 50
30 Excess pore pressure (kPa) 30 40 40
20 Ru (%) 20 30
30
Extensive Stress (kPa)
10 Compressive Stress (kPa) 10 20 20
0 0 10 10
0 200 400 Cycle 600 800 1000 0 0
0 150 300 450 600 750 900
Figure 8: Test results of sample B13, f = 1 HZ, R95 Cycle
Figure 12: Test results of sample B17, f = 8 HZ, R95, Amplitude 16
kPa
kPa
100 100
90 90 100
kPa
100 Comment [ln8]: Lỗi đường
90 90 màu trắng, xanh
80 80
70 70 80 80
60 70 70
60
Ru (%)
60 60
Ru (%)
50 50
50 50
40 40
40 40
30 30
Excess pore pressure (kPa) 30 30
20 Ru (%) 20
20 Ru (%) 20
Extensive Stress (kPa) 10 Excess Pore Pressure (kPa)
10 Compressive Stress (kPa) 10
Compressive Stress (kPa) 10 Extensive Stress (kPa)
0 0 0 0
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60
Cycle Cycle
Figure 9: Test results of sample B14, f = 1 HZ, R95 Figure 13: Test results of sample B18, f = 1 HZ, R95, Amplitude 24
kPa
406
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
100
kPa
100 100
kPa Comment [ln9]: Lỗi hình màu
100
90 Ru (%) 90 xanh
Excess Pore Pressure (kPa) 90 90
Comment [ln11]: Dung phần
Compressive Stress (kPa) 80
80 Extensive Stress (kPa) 80 80
70 70
70 70 mềm
60 60
Ru (%)
60 60
Ru (%)
50 50
40 50 50
40
30 30 40 40
20 20 30 30
Excess pore pressure (kPa)
10 10 20 Ru (%) 20
Extensive Stress (kPa)
0 0 10 Compressive Stress (kPa) 10
0 100 200 300 400 500 600 700 800
Cycle 0 0
Figure 14: Test results of sample B19, f = 3 HZ, R95, Amplitude 24 0 30 60 90 120 150
Cycle
kPa Figure 18: Test results of sample S3, f = 1 HZ, R90, Amplitude 16
kPa
100
kPa
100 Comment [ln10]: Lỗi hình
Ru (%) 90
90 Excess Pore Pressure(kPa)
80
Compressive stress(kPa)
80 3.3 Comparison of test results between sea sand and river
Extensive stress(kPa)
70 70 sand.
60 60 Sea sand and river sand samples are compared based on the cylic
Ru (%)
100
kPa
100
Sea sand samples are liquefied when the vibration amplitude is
90
twice higher than those in normal condition of vehicle moving
90
on the road.
80 80
70 70 Liquefaction state are occurred corresponding to some
60 60 conditions:
Ru (%)
kPa 4. CONCLUSION
100 100
90 Ru (%) 90 Based on the test results, conclusions are drawn following:
Excess Pore Pressure (kPa)
80 Compressive Stress (kPa) 80
Extensive Stress (kPa)
70 70 All of sea and river sand samples are not liquefied under normal
60 60 vehicle moving condition (f-1Hz; stress amplitude a = 8 kPa).
Ru (%)
50 50
Sea sand samples are liquefied under the vibration amplitude is
40 40
twice higher than normal condition of vehicle moving
30 30
20 20 It is recommended that in case of embankment having
10 10 compaction ratio R90 and R95, the embankment will not be
0 0
liquefied under normal amplitude value (a = 8 kPa).
0 200 400 Cycle 600 800 1000
Thus, the sea sand can be used to replace river sand for
Figure 17: Test results of sample S2, f = 1 HZ, R90, Amplitude 8 embankment fill.
kPa
407
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5. REFERENCES Jakka, R. S., Datta, M. and Ramana, G. V (2010) ‘Liquefaction
behaviour of loose and compacted pond ash’, Soil Dynamics
Bolton, S. H., M., I. I. and Ignacio, A. (1983) ‘Evaluation of and Earthquake Engineering, 30(7), pp. 580–590.
Liquefaction Potential Using Field Performance Data’, P., V. Y. and J., T. (1995) ‘Liquefaction and Postliquefaction
Journal of Geotechnical Engineering. American Society of Behavior of Sand’, Journal of Geotechnical Engineering.
Civil Engineers, 109(3), pp. 458–482. American Society of Civil Engineers, 121(2), pp. 163–173.
Ishihara, K. (1977) ‘Simple Method Of Analysis For Liquefaction Y., S. H. and J., Y. (2014) ‘Failure Modes of Sand in Undrained
Of Sand Deposits During Earthquakes’, Soils And Cyclic Loading: Impact of Sample Preparation’, Journal of
Foundations, 17(3), Pp. 1–17. Geotechnical and Geoenvironmental Engineering. American
Society of Civil Engineers, 140(1), pp. 152–169.
408
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Large earthquakes have occurred and caused liquefaction damage in urban areas in Japan every few years. Recently, the
applicability of suspension grout has been expanded because of the development of micronized technology of particles. Permeability of
suspension grout has been conventionally evaluated using groutability ratios obtained from sizes of micro-particles contained suspension
grout and particle size of soil. The problem is that permeability is affected not only by grain size of micro-particles of grout but also pore
structure. In this study, in order to evaluate permeability of micro-particles more accurately, the pore index was proposed, which can be
calculated from the Kozeny-Carman equation. And from the results of one-dimensional permeation experiments of micro-particles, it was
confirmed that permeability of suspension grout can be evaluated accurately by the pore index.
Keywords: suspension grout, coefficient of permeability, pore index.
409
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
In this paper, in order to establish a new evaluation method, the its applicability was confirmed through one-dimensional permeation
pore index was formulated from the Kozeny-Carman equation and experiments of micro-particles for mixed sand
where l is a length of soil specimen (cm), lr is length of bending- 4.1 Summary of samples for permeation experiments
related extension of pipe in pore of the soil skeleton (cm) and n is
porosity of the soil. Then, lr expresses the actual length of the flow The samples used for the experiments were mixed sand with
0.075mm to 2.000mm in diameter. The conditions and the grain size
channel. Eq. (4) is obtained by substituting Eq. (3) for Eq. (2).
410
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
distribution of the samples are shown in Table. 1 and Figure 1. In 4.3 ONE-DIMENSIONAL PERMEATION EXPERIMENTS
order to consider the effect of uniformity coefficient; Uc and the OF MICRO-PARTICLES
grain size of soil at 15%; D15 on permeability of micro-particles,
mixed sand was prepared by mixing classified sands. The D15 of 4.3.1 Experimental conditions
samples No.1 to No.4 were made to be same. In order to confirm the One-dimensional permeability experiments of micro-particles were
effect of uniformity coefficient, Uc on permeability of micro- carried out using the sand samples in Figure 1. The flow chart of
particles, the Uc of samples No.5 to No.13 were set at 3.0 to 5.5. permeability evaluation of micro-particles is shown in Figure 4.
Then, Uc of samples No.14 to 17 were made to be same to compare Even though at least 50cm in radius is required for the ground
the change of the coefficients of permeability depending on the improvement in chemical grout injection methods, in order to save
particle sizes. time and effort, micro-particles were permeated to the specimens
with 5.0cm in diameter and 10.0cm and 15.0cm tall (hereinafter
4.2 Permeability tests of mixed sand referred to as short specimens) in first screening. From these result,
In order to calculate the pore index, permeability tests were carried the samples for second screening were decided. In second screening,
out on each sand sample shown in Figure 2. The specimens for specimens with 50.0cm or 90.0cm tall (hereinafter referred to as
permeability tests were prepared to be 5.0cm in diameter and long specimens) were prepared. When the volume of permeated
10.0cm tall and their void ratio e was set at 0.82 by air-pluviation micro-particles is greater than that of pore in a long specimen, the
method. Only in the sample No.16, the void ratio was set at e 0.70. micro-particles are determined to be permeable.
Then, CO2 gas and de-aired water were infiltrated into them to The samples of specimens are shown in Figure 1. Both the grain
heighten the saturated rates and permeation tests were conducted. size distribution of micro-particles measured by the laser diffraction
Here, the coefficients of permeability of specimens were the average particle size analyser (SALD-3100 produced by Shimadzu
of three values measured within 0.5 of hydraulic gradient. Corporation) and the grain size of micro particles at 85% G85 are
The results of permeability tests were shown in Figure 2. Since shown in Figure 5. The weight concentration C of micro-particle
the void ratios e of specimens were almost equal, the dispersion in was set by mixing with water as follows; WC-1(G50=18.5μm), WC-
the permeability seen in Figure 2 was likely to be caused by the 2(G50=14.9μm) and WC-3 (G50=8.9μm). Then, they were permeated
differences in uniformity coefficient; Uc and grain size of sand into the specimens with stirring to avoid deposition of micro-
particles. Therelationship between the calculated pore index d and particles. The concentrations of mixture C were set at 10.0% in
the grain size of soil at 15% D15 is shown in Figure 3. From the every case, which was calculated from Eq. (7).
figure, the pore index d cannot be expressed by D15 uniquely.
411
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Mm (7)
C 100(%)
Mw
412
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(c) Difference in grain size of soil at 15% D15 As observed above, the permeability of micro-particles cannot
In Figure 9(c), the effect of difference in grain size of soil at 15% be evaluated accurately only by Gr. Figure 11 shows the relationship
D15 which is used for Gr was considered. All the cases in Figure 9(c) between uniformity coefficient Uc and the pore index d normalized
showed same trend, that is, the injection velocities decreased to by the grain size of micro-particles at 85% G85. It is clear whether
almost 0 by clogging. From this result, it was confirmed that the size the micro-particles are permeable or not in Figure 11. The
of micro-particles was close to those of pore diameter even in Case5 experimental discriminant was obtained from uniformity coefficient
where the particle size was bigger than other cases. As seen from Uc, a pore index d and the grain size of micro-particles at 85% G85
Figure 9(a), on the other hand, even though the D15 of Case13 was using the following equation; Eq. (8).
smaller than that of Case5, satisfactory permeability was obtained.
From these results, it was confirmed that it is difficult to evaluate d
0.887 U c 8.125 (8)
permeability of micro-particles only by Gr.
G85
4.4 Permeability evaluation of micro-particles
If calculated value from Eq. (8); d/G85 is plotted upper than the
From the results of experiments, the permeability of micro-particle approximate line, permeation of micro-particles could be determined
was evaluated. Figure 10 shows the relationship between Gr and to be possible. For obtaining an equation with high-precision, it is
coefficient of permeability k. A red frame indicates a range of Gr necessary to accumulate more experimental data since Eq. (8) was
less than 11, which means the permeation of micro-particles is estimated from the measured values during the experiments.
estimated to be impossible. Figure 10 indicates that permeation of
micro-particles succeeded in some cases even in a red frame.
Figure 9 Experimental result regarding index (H=10.0cm) Figure 10 Relationship of Gr and coefficient permeability k
413
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
414
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: An investigation into behaviour and potential applications of a polymer-enhanced bentonite-sand mixture (PEBSM) in its
compacted state had been carried out in an environmental controlled laboratory. The experiments undertaken included microscopic
visualization of the mixture, drying-wetting processes, consolidation and shear strength tests. The results obtained indicate that depending on
the overburden pressure acting on the compacted mixture and its water content, settlement in combination with swelling and shrinkage may
result in cracks or on contrary may close the initial cracks induced by the mechanical compaction in the field. The conclusions have been
drawn based on an analytical assessment performed on the field and laboratory data gathered. A discussion is also brought up on the
difference between two approaches used – single-valued effective stress and two-independent stress state variable approaches. The two
approaches predict different values of suction at which crack starts to occur and different time-settlement behaviour.
Keywords: Bentonite, swell/shrink, settlement, shear strength, unsaturated soils, cracks.
415
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
f (S r ) Sr (2)
v ' ( v u a ) S r s (3)
h ' ( h u a ) S r s K o ( v u a ) S r s (4)
Figure 3 Measured matric suction and computed degree of
saturation of the compacted PEBSM layer As suction in the field changes, the effective stress of the liner
will also change, which follows the following equation:
60
Figure 4 Soil-water characteristic curve for the compacted PEBSM vertical effective stress
speciemen (from Schanz et al., 2004) 50 horizontal effective stress
s'v, s'h (kPa)
40
On site, the PEBSM was laid approximately 1 meter deep from
the final ground surface leading to a surcharge of approximately 20 30
kPa or slightly lesser. The initial conditions of the PEBMS in the
20
field test are as following:
10
Initial dry density (d) : 1.75 Mg/m3
Initial void ratio (e) : 0.52 0
28.07.01
05.11.01
13.02.02
24.05.02
01.09.02
10.12.02
20.03.03
28.06.03
06.10.03
14.01.04
23.04.04
01.08.04
09.11.04
17.02.05
28.05.05
05.09.05
14.12.05
With a 1-m thick recultivation and drainage layer, the total Date
vertical overburden stress acting in the middle of the compacted Figure 5 Changes in vertical and horizontal effective stresses with
PEBSM of 7 cm can be computed to be equal to 20.6 kPa. time
416
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The Ko condition only holds when soil blocks do not develop as 4.3 Modulus of Elasticity with Respect to Change in Net
a result of cracks that develop due to desiccation. In the case of Stress (E)
crack, the horizontal net stress follows:
The E value has been derived using the elastic theory from the
h u a v u a 1 2 S r s
saturated oedometer test data for the unloading-reloading, which is
(7) deemed to be within the elastic range:
1 1
e 1 1 1 1 2
mv (15)
The changes in vertical and horizontal strains in the event of
crack can be expressed as:
1 e0 ' E (1 )
where mv is the coefficient of volume compressibility of soil. The
v u a S r s 2 S r s
1
v (8) average mv value obtained from the saturated oedometer test results
E of the PEBSM specimen compacted at dry side as shown in Figure 6
is 1.798x10-5 /kPa. Thus, the corresponding E value is 36,134 kPa.
1 S r s v u a S r s
1
h (9)
E 0.65
Void ratio, e
Rahardjo (1993), where changes in two state variables used (i.e. net
stress and suction) and the resulting effects are considered 0.50
separately.
In the case of the Ko condition (Δεv = Δεvol), the volumetric 0.45 Swelling at 7 kPa
strain is given as: Loading
0.40 Unloading
1 1 2 u 1 s (10) Reloading
E 1 H 1
v a 0.35
1 10 100 1000
In the above equation, H represents the modulus of elasticity of Vertical stress, s v (kPa)
the soil with respect to suction change (the 2nd stress state variable), Figure 6 Oedometer test result on a saturated compacted
while as described before, E signifies the modulus of elasticity for PEBSM specimen
the effect of net stress (the 1st stress state variable) change.
During crack, the horizontal net stress is defined as: 4.4 Modulus of Elasticity with Respect to Change in Suction
(H)
h u a v u a E
s (11) The H value was computed from the e-log suction plot that had been
1 H (1 ) obtained while determining the SWCC curve (Schanz et al., 2004).
The SWCC had been determined at three-dimensional free
The changes in vertical strain (Δεv) and horizontal strains (Δεh) swell/shrink conditions. Therefore, the following relationships hold:
after crack are formulated as:
v vol Ct v u a C a s (16)
v u a 2 ( h u a ) s
1 1 (12)
E H e 1 e 1
Ct ; Ca (17)
1 e0 ua 1 e0 s
( h u a ) ( v u a ) ( h u a )
1 (13)
E Analogous to the unloading-reloading path, from which the E
value has been derived, the corresponding drying-wetting cycles of
4. PARAMETERS USED IN COMPUTATION the PEBSM should also be used to determine Ca value. However,
The following parameters have been used in the computation and due to the unavailability of the wetting data, only the drying path
the values have been derived as outlined in this section. has been considered in the derivation of Ca value. The following
1) Coefficient of earth pressure at rest (Ko) figure shows the drying test on the material.
2) Poisson ratio (υ) 0.58
3) Modulus of elasticity with respect to change in net stress (E) Experimental (drying)
4) Modulus of elasticity with respect to change in suction (H)
0.56
4.1 Coefficient of Earth Pressure at Rest (Ko)
Void ratio, e
0.54
The Ko value was determined from the experiment carried out in
Geodelft, Netherlands. The value is equal to 0.52.
0.52
4.2 Poisson Ratio (υ)
The Poisson ratio (υ) is calculated from the elastic theory as follows 0.50
(with Ko equal to 0.52): 10 100 1000 10000 100000 1000000
Suction, st (kPa)
KO
0.34 (14) Figure 7 Change in void ratio with increasing suction from a
1 KO drying test on compacted PEBMS
417
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The void ratio-suction curve exhibits a bimodal characteristic The subsequent changes in vertical and horizontal strains were
with negligible changes for suction above 500 kPa. The linear computed based on an assumption that the changes in vertical and
portion of the curve up to 500 kPa suction is within the suction horizontal net stresses are equal to zero. This is adopted for both
range, at which the material was subject to in the field and thus can approaches when computing the evolution of deformation and
be used to determine the Ca value. From the curve, the Ca value is settlement due to changes in suction of the PEBSM layer on site.
1.017x10-4 /kPa, and thus a H value of 29,500 kPa as for this case Ca
= 3/H. 5.3 Comparison of Computed Strain and Settlement
It is shown in the two tables that the use of two different approaches
5. DEFORMATION AND SETTLEMENT
affects the initial strains and deformation computation for the case
Computation of settlement and deformation as a result of the change with cracks with the two-independent stress-state variable approach
in suction, which was measured by the tensiometer, has been carried estimating half the values compared with the single-valued effective
out using the parameters derived as described in the preceding stress concept.
section. The following figure illustrates that, for the case of Ko condition,
the use of H value from the drying curve result in much higher
5.1 Computation using Single-Valued Effective Stress settlement estimated using the two-independent stress-state variable
Concept approach. This is inconsistent since at the Ko condition, cracks do
not occur and both approaches should give approximately the same
As described in Section 2, a change in the net vertical stress ((v- result.
ua)) of 20.6 kPa was applied on the compacted PEBSM layer after
placement. The corresponding change in the net horizontal stress -0.15
((h-ua)) is therefore equal to 10.7 kPa. The initial suction of the
PEBSM was determined from the SWCC given the placement water -0.10
Settlement, D H (cm)
content and dry density of the compacted layer and the value was
4.4 MPa. Immediately after placement, the tensiometer measured -0.05
20.1 kPa and therefore the material was subjected to a significant
suction change once placed in-situ. The corresponding change in 0.00
degree of saturation of 59% can be computed from the SWCC.
The initial strain and deformation due to placement can be 0.05
computed and the values are summarised in the following table. Two-independent stress state variable
0.10
Single-valued effective stress
Table 1 Initial Strain and Deformation Computed using the Single-
0.15
valued Effective Stress Concept
28.07.01
05.11.01
13.02.02
24.05.02
01.09.02
10.12.02
20.03.03
28.06.03
06.10.03
14.01.04
23.04.04
01.08.04
09.11.04
17.02.05
28.05.05
05.09.05
14.12.05
Without cracks With cracks
Vert. strain, v (%) -3.3 -3.2
Hor. strain, h (%) 0 -3.3 Date
Figure 8 Comparison of settlement with time from two different
Vert. deform. ,dv (cm) -2.3 -2.3 approaches
Hor. deform. ,dh (cm) 0 -3.3
The variation of time-settlement plot for the two-independent
Note: -ve indicates expansion
stress state variable approach is shown in the following figure using
the realistic H value equal to 8 and 10 times that determined from
5.2 Computation using Two-Independent Stress-State the drying test as described earlier. It is demonstrated that an
Variable Approach average value of 9H provides the best fit to the settlement plot from
As mentioned earlier, the H value for the compacted PEBSM layer the single-valued effective stress concept.
was derived from the drying path. Realistically, the value for the
wetting path is much higher than that from the drying cycle. The -0.020
following table summarises the initial strain and deformation due to -0.015
Settlement, DH (cm)
In this computation, the realistic H value used in the The evolution of vertical strain assuming Ko condition is shown
computation was first approximated by matching the settlement plot in Figure 10 for both methods. The single-valued effective stress
from both approaches. It is later demonstrated that taking the concept consistently gives lesser vertical strain compared with that
realistic H value of 8 to 10 that of the drying is reasonable and an given by the two-independent stress-state variable approach,
average value of 9H (i.e. 265,500 kPa) was subsequently used. although the magnitude does not differ significantly.
418
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 11 and Figure 12 demonstrate that the two-independent sufficiently deep to facilitate full development of preferential flow
stress-state variable approach produce lower values of vertical and paths.
horizontal strains for the case when crack occurs. The suction value, at which cracks first occur is defined as the
critical suction (scr) and the value can be computed by equating the
-3.36 net horizontal stress with the maximum tensile strength (tmax).
-3.34 Schanz et al. (2004) reported that tmax value of the PEBSM is 20 kPa
(%)
on average.
-3.32 The following equations represent the critical suction for the
v
Vertical strain, e
v ua tmax 1 1
-3.26
scr (18)
-3.24
independent stress state variables 1 1 2 S r
-3.22
single effective stress
v ua tmax H 1
-3.20
scr (19)
28.07.01
05.11.01
13.02.02
24.05.02
01.09.02
10.12.02
20.03.03
28.06.03
06.10.03
14.01.04
23.04.04
01.08.04
09.11.04
17.02.05
28.05.05
05.09.05
14.12.05
1 E
-3.34
The critical suction can also be plotted as a function of the total
-3.32 vertical stress using both the above equations. The following figure
shows such plot for the PEBSM studied.
-3.30
-3.28 100
Single effective stress
Total vertical stress, s v (kPa)
-3.26 approach
80 Independent stress-state
-3.24
variables approach
-3.22 horizontal strain
vertical strain 60 No crack Crack
-3.20
No crack Crack
28.07.01
05.11.01
13.02.02
24.05.02
01.09.02
10.12.02
20.03.03
28.06.03
06.10.03
14.01.04
23.04.04
01.08.04
09.11.04
17.02.05
28.05.05
05.09.05
14.12.05
40
Date 20
Figure 11 Vertical and horizontal strains computed using the single-
valued effective stress concept (after crack occurs) 85 kPa 147 kPa
0
-1.68 0 50 100 150 200 250 300
Vertical, horizontal strain, ev, e h(%)
-1.64 The above figure reveals that at the applied total vertical stress
of 20.6 kPa (the air pressure is atmospheric), the scr value is 85 kPa
computed using the single-valued effective stress concept and 147
-1.62
horizontal strain kPa using the two-independent stress-state variable approach. At
vertical strain any given applied total vertical stress, the two-independent stress-
-1.60 state variable approach gives a higher value of scr compared with the
other approach. Further detailed investigation will warrant clarity at
28.07.01
05.11.01
13.02.02
24.05.02
01.09.02
10.12.02
20.03.03
28.06.03
06.10.03
14.01.04
23.04.04
01.08.04
09.11.04
17.02.05
28.05.05
05.09.05
14.12.05
419
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
kPa), which is practically even higher than the maximum value that 2) Strain and deformation computed using the two-independent
can occur in the field when the compacted PEBSM layer is used in stress-state variable approach are approximately half those
landfill. derived using the single-valued effective stress concept. The
former computation is realistic when adopting a realistic
modulus of elasticity with respect to change in suction
(realistic H) of 8 to 10 times H derived from the drying test.
3) Critical suction at which cracks start to occur (scr) has been
computed and the value is consistently lower for the two-
independent stress-state variable approach at any given net or
total vertical stress (with atmospheric air pressure).
8. REFERENCES
Agus, S.S. (2005) An Experimental Study on Hydro-Mechanical
Characteristics of Compacted Bentonite-Sand Mixtures.
Doktor-Ingenieur Dissertation, Bauhaus-University Weimar,
Germany.
Agus, S.S., Arifin, Y.A., and Schanz, T. (2007) “Settlement and
swells/shrink behaviour of a compacted polymer-enhanced
Figure 14 Possible representation of a compacted bentonite-sand bentonite-sand mixture as induced by field mechanical
mixture (from Agus, 2005) loading and wetting-drying cycles”, Proceeding of Brazilian
Symposium of Unsaturated Soils, Salvador, Brazil, 2007.
Arifin Y.F., Agus S.S., and Schanz T. (2016) “Swelling pressure
6. POSSIBLE APPLICATIONS and hydraulic conductivity of compacted polymer-enhanced
Due to its characteristics as described, the compacted PEBSM is bentonite-sand mixture”, Proceeding of Geotechnische
suitable for landfill application. The material is better placed under a Aspekte in Umweltschutz 2006, Stahlman and Rosenberg
drainage layer to ensure that suction remains low and deep cracks Edition. Brauschweig, 87-98.
are not present during the landfill design life. The effectiveness of Fredlund D.G., and Rahardjo H. (1993) Soil Mechanics for
placing the compacted PEBSM below the waste is questionable Unsaturated Soils. John Willey & Sons. Canada.
since interaction between the bentonite-polymer with leachate from Schanz, T., Agus, S.S., and Tscheschlok, G. (2004) “Hydraulisch-
the landfill may degrade the beneficial properties of the mixture. mechanische Eigenschaften einer polymerverbesserten Sand-
Bentonit- Mischung beim Einsatz im Deponiebau”,
Geotechnik, 27, 4: pp344-355.
7. CONCLUSION
An investigation into the behaviour of a compacted PEBSM has
been conducted by means of a field case study. The following ACKNOWLEDGEMENT
conclusions can be drawn from the outcome of the study. The authors would like to express sincere gratitude to the late Prof.
1) A maximum suction value of 50 kPa was measured during an Dr.-Ing. habil. Tom Schanz for his guidance when this research
approximately 4-year period in the case study presented in this work was carried out in Bauhaus-University Weimar, Germany. He
paper. The corresponding drop in degree of saturation as was also the authors’ PhD supervisor in the same university. He
computed from the soil-water characteristic curve (SWCC) is passed away in 2017 in Bochum, Germany, where he and his family
hardly below 80%, which indicate a good water retention lived.
capability of the material.
420
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Disasters are increasingly difficult to predict due to severe climate change. In addition to disaster prevention, it is an urgent
task to prepare for post-disaster restoration and prevent secondary damage to the environment to reduce the impact of disasters. The
construction site in this paper is located on the primary trunk road which connects eastern and western Taiwan. Due to major typhoon
disasters, the slope below the road severely collapsed. It needs to be repaired in a short period to reduce the impact on transportation and
economy. Considering the issue of time-urgency and cost-saving, geosynthetics reinforced structure in accordance with concrete retaining
wall and pile foundation was constructed as road embankment. In addition to achieve a rapid recovery after disaster, the construction method
also lowers carbon emission environmental impact. In this paper, geosynthetic reinforced structure has been proved as a sustainable solution
for disaster recovery.
Keywords: disaster recovery, geosynthetics reinforced structure, carbon emissions, green engineering
1. INTRODUCTION and made of sand, silt and gravel. The SPT-N value of the
foundation soil is equal to 20 – 50 ((1) Chen, et al. 2017).
Since the discovery of the greenhouse effect as one of the causes of Taiwan is located in the orogenic belt from obduction of micro-
climate change, green engineering for energy conservation and oceanic plate into Eurasian continental plate. The orogeny
carbon reduction have gradually been introduced into the design and movement resulting in a fractured basement rock with several
construction of engineering projects (Heerten, 2012). In recent geologic structure such as fault and fold (Sibuet and Hsu, 2004). In
years, Reinforced Concrete (RC) retaining structures have been this case, only a fractured basement rock is observed, without any
partially replaced by Geosynthetics Reinforced Structures (GRS), occurrence of fault and fold nearby.
which excludes the casting and curing time of concrete, to shorten The site is located adjacent to Fenggang River. It is one of the
the construction time and disaster recovery process. Compared to major river contributed to a high sediment erosion in Taiwan,
GRS, RC structure also demands higher material quantity, especially during the typhoon season ((2) Chen, et al. 2017). The
transportation cost, and environmental impact (WRAP, 2010). main river has 41 km length with average elevation of 698.3 m and
This paper presents a case study of slope rehabilitation project average slope of 28.4o.
using a composite structure made of pile foundation, RC wall, and
GRS. The case is located in the southern part of Taiwan, specifically
3. DESIGN
in Pingtung County, Provincial Highway No. 9 470K+500, adjacent
to Fenggang River. In 2013, Typhoon Kong-Rey was attacking In the design part, several factor need to be considered upon
Taiwan with heavy rainfall, in which the concentrated rainfall rise (FHWA-NHI-00-043). First, the main cause of collapsed slope is
upon 250 to 300 mm within a single day (JTWC, 2013). Rapid scouring phenomenon in the toe of the slope. Therefore, the new
change of precipitation rate caused a fluctuation in the adjacent river constructed slope should be made to prevent any damage from the
water level. The impact of high energy and high velocity in the river scouring. The occurrence of typhoon and earthquake are also a main
stream was an accelerated scouring of soil. Thus, a slope failure was concern in this structure. Typhoon will results in river stream
occurred beneath Provincial Highway No. 9 (Figure 1). fluctuation, whereas earthquake will impact the overall stability of
the structure.
The collapsed slope is 150 m in length and 22 m in height. At
the bottom, all-casing pile foundations were built to deal with the
scouring phenomenon from Fenggang River. On the top of pile
foundations, a semigravity RC wall was designed to deal with the
fluctuation of river stream due to typhoon impact. For the upper
structure, two-tiered GRS consists of geogrids, soil bags, and water
culvert was installed (Figure 2).
2. SITE DESCRIPTION
Pingtung County of Taiwan was geologically located above
Chaochow Formation, Quaternary Terrace Deposit, and Quaternary
Alluvium Deposit. Chaochow Formation consists of slate, a
metamorphic rock, with thinly bedded sandstone. Terrace Deposit
and Alluvium Deposit are younger sediment from Quaternary period Figure 2 Design Drawing of Composite Structure
421
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3.1 All-casing pile foundation pressure inside the geogrid reinforced slope. The facing system of
the GRS was piled up soil bags as the vegetation medium. Above
The foundation was set with double-row cased piles with a diameter the GRS, a soil slope with a gradient of 1:0.3 and 2 meter height was
of 1.2 meters (Figure 3). The outside piles were set at intervals of 2 backfilled and compacted as the road base.
meters in a row, and the interval of rear piles was set as 5 meters.
The depth of each pile is about 10 m with an initial drill penetration
of at least 2 meters into the rock sublayer. A total of 155 piles were
installed to transfer the vertical load into metamorphic basement
rock and improve the bearing capacity of the whole structure.
Furthermore, a reduced influence of riverbed scouring was expected,
and the sliding drive force raised by the earth pressure of slope was
resisted.
3.5 Hydroseeding
Hydroseeding, a plantation work using a mixture of seed and mulch,
was performed after the completion of GRS installation (Figure 7).
The main function of hydroseeding is to control the surface erosion
using vegetation growth, while it also provides an aesthetical
function to enhance the appearance of the whole GRS structure.
Overall, hydroseeding has a beneficial impact on the environment
Figure 4 Semigravity RC Retaining Wall Construction aspect by lowering the carbon footprint from the construction work.
422
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
423
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
424
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The use of the Observational Method in Deep Excavations for the Realization of
a Residential Compound in a Complex Hydrogeological Context
M. Carassini1, F. Bucci2, A. Antiga3
1
Technical Director, Bonatti S.p.A. BU Civile & Infrastrutture, Parma, Italy
2
Engineering Manager, Bonatti S.p.A. BU Civile & Infrastrutture, Parma, Italy
3
Technical Director, Soil S.r.l., Milan – Italy
E-mail: andrea.antiga@gmail.com
ABSTRACT: The urban redevelopment works of the "Railway Station Area ex-Boschi”, was developed on an area of nearly 17.0000 m2, it
entailed the execution of deep excavations in a difficult geotechnical and hydrogeological context. In order to overcome these critical issues a
design based on an observational approach was developed. This design considered several possible hydrogeological scenarios and staged
excavations over small areas.
The Observational Method has been successfully extended from the more traditional case of tunneling works to the less usual case of deep
excavations in large construction areas.
Keywords: Observational Method, deep excavations, diaphragms
1. INTRODUCTION Real estate surfaces and the remaining € 42.3 mln delivered
according to periodic progress payment certificates, as per initial
The urban redevelopment project of the "Railway Station Area ex- contract data. The Project contract duration was 57 months. From a
Boschi” area in Parma (Italy), involved the creation of a large road technical and organizational point of view, the Project was
link, under-passing the existing eight-track railway line Station and characterized from the outset of the basic design phase, by a strong
the construction of five new buildings. interaction with the existing urban context and with the local
The works, developed on an area of nearly 17.0000 square meters, infrastructures. In fact, within the Train Station or in the immediate
entailed the execution of considerable excavations and supporting vicinity, there are:
structures.
public and private buildings both in the north and south areas,
The hydrogeological context was complex because of soils with
the existing railway station, whose outer masonry envelope and
poor mechanical characteristics and of two different aquifers. The
steel decorations had to be preserved according to the constraints
presence of an important railway line in operation and several
imposed by the local Superintendence of historic buildings,
existing buildings next to the excavations, completed the project
framework. bus stops and terminals,
In order to overcome these critical issues and assure greater road and pedestrian traffic ways,
safety and continuity of the works, the Contractor developed a the Milan-Bologna railroad tracks, with the need to maintain at
detailed design proposal based on an observational approach, which least four lines always running at the same time, being this the
considered several possible hydrogeological scenarios and staged main Italian railway line.
excavations over small areas. As a result, the contract basic design solution was conceived to
In this case, the Observational Method has been successfully minimize as much as possible the impacts and risks associated with
extended from the case of tunnels works, more traditional, to the the execution of works.
less usual case of deep excavations in large construction areas, For example, the need to carry out deep excavations required the
showing its own peculiarities and giving important indications for execution of perimeter r.c. diaphragms along the entire project area.
future applications. This technique permitted the use of machinery and workmanship
from the top ground level before proceeding to the excavation and
2. RIQUALIFICATION PROJECT underground works.
In 2007, the Municipality of Parma (Italy), through its subsidiary 3. GEOTECHNICAL WORKS IN THE NORTH AREA
company "STU Società di Trasformazione Urbana", launched a
tender for the public works of the urban renovation of the Parma The renovation project envisaged the creation of a pedestrian and
Railway Station Area. traffic road under-passing of the rail tracks from the North to South
The contract basic design was developed by an association of areas of the Railway Station, from which the new access to the
design firms with MBM Arquitectes from Barcelona as a leader and station would be created, and two levels of underground parking in
it essentially entailed the renovation and upgrading of the existing the north area.
Railway Station of fascist epoch and of both the North and South These works covered a very large area of approximately 17,000
Areas of the Station, as identified in figures 1 and 2. This was square meters and resulted in significant excavations of varying
achieved through the following interventions: depths ranging up to approx. 15m below the existing ground level.
redesigning the respective squares with their green areas, the The excavations were carried out in a complex and variable
traffic roads and pedestrian walkways, hydrogeological context because of the different excavated soil
linking the two areas with the creation of an underground layers and their variable characteristics, due to the extension of the
roadway underneath the existing station and the entire railway area.
tracks park, Specifically, the geotechnical context of this area is
characterized by an intermediate silty-clay layer with poor
providing in the northern area, formerly dedicated to railway
mechanical characteristics, while the hydrological context is
storage and maintenance, two floors of public and private
characterized by the presence of a first phreatic water table in direct
underground car parks, a newly constructed bus station and
contact with the Parma river – called A0 – and a second artesian
buildings, including residences, offices and a hotel.
water table under pressure – called A1 – both characterized by
The works were awarded to the JV Bonatti SpA - Di Vincenzo
seasonal level variations (fluctuation above 5 m annually).
under a Public-Private partnership procedure for a work amount of:
€ 98.3 mln, of which € 56 mln as compensation for the value of the
425
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
426
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
levels to consider in the design. The first aquifer (A0) is present Figure 5 shows a geotechnical section (layers A, B, C -
within level A; it is a phreatic aquifer whose levels are directly boreholes (BH and CPT).
correlated with the Parma river; during summer time the The most significant geotechnical parameters are shown below:
groundwater is absent, as it reaches the ground level in the river
flood stages. The second aquifer (A1) is present within level C. This
is an important artesian aquifer of considerable hydro-potable Ɣ Փ c' Cu
interest confined by the silt-clayey layer B. (kN/m3) (°) (kPa) (kPa)
Polygenic gravel (layer A) 20.5 39 0 -
Silty clay/clayey silt (layer B) 19 21 15 80
Polygenic gravel (layer C) 20.5 39 0 -
427
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
2. they were intended to improve the functional aspects, as well as probable" resistance characteristics of the jet-grouting and strand
the individual technological elements or components of the anchor bulbs to be used in the project.
project, Afterwards, during the works, specific testing and control fields
3. they did not entail a reduction in the performance of the Project were established and carried out to verify the actual anchor bulbs
itself and kept the work execution time and the worker's safety tensile strength and jet grouting mechanical resistance in areas
conditions unaltered. where, even after preliminary investigations, the greatest variability
In addition, for the Italian and European standards in force remained.
(Eurocode n.7 1997, NTC2008), the adoption of the so-called The strand anchors tensile test field has determined the
Observational Method in the design of provisional and final minimum design resistances of the ground layers interested by the
geotechnical works was innovative, if compared with the previous anchor bulbs (superficial deposits and gravel, clayey limestone-clay)
technical regulations, under the following main conditions: according to the drilling diameters, see figure 6. The jet-grouting
that the particular complexity of the geotechnical situation and test field has determined the technique (Bi-fluid), cement type and
the importance of the work was established; quantity, execution parameters (feed rate and rotation speed,
that the acceptability limits were established of the values of injection pressure, etc.), see figure 7.
those parameters values which represent the ground-structures The introduction of local variability and execution phases in the
system’s overall behavior; detailed design related to the geotechnical and foundational works,
that other alternative solutions congruent with the project were together with the execution of such tests during the works, allowed
verified as possible and envisaged; to satisfy loading hypothesis of more conservative and wider
that an appropriate monitoring system was set in place during variance and to absorb the uncertainties of geotechnical aspects of
works and control plans aimed to confirm the assumptions made direct relevance on ground support interventions (strands anchors,
and the validity of the design solution adopted or the adoption of jet grouting base r.c. foundation slab excavation and casting).
one of the proposed alternatives if the acceptability limits were Furthermore, as a result of the various studies, it was found that the
reached. basic design assumptions on material resistance could be improved
As a result of the comparison between all the various available and opening the finale excavation level (foundation slab) for limited
data and design proposals developed in a preliminary form, the phases and dimensions would have created improved conditions for
choice was made in favor of an Observational Method of design the perimeter diaphragm stability, as further explained below.
(Peck 1969, Ciria 1999).
For the choice of the design parameters, the approach
recommended by Peck was used with the adoption of "most
probable" values, the evaluation of "most unfavorable scenarios”
and the definition of mitigation measures for such situations.
Schematically, the solution adopted, according to the national
regulation requirements, was developed according to the following
macro steps:
1. design and calculation of works by adopting "most probable"
water levels;
2. analysis of possible scenarios with rising water levels up to "most
unfavorable" values and definition of mitigation measures for
such situations;
3. active monitoring during works of the two water table levels, to
identify the real scenario.
We want to point out that the unavoidable geotechnical
uncertainties identified in the previous chapter (with two possible
scenarios: "more probable" and "most unfavorable") made it
absolutely unsatisfactory to use a classic design approach based on a
single "predefined design".
Indeed, in this case, the generalized use of "most unfavorable" Figure 6 Strand anchor tests
values would have resulted in excessive costs. On the other hand,
less prudent values, such as the generalized use of "most probable"
values would have led to a risk level considered unacceptable.
With regard to the maximum piezometric load, detailed analyses
and simulations were carried out on the basis of the historical data
available from the hydro-potable wells in the Municipality of Parma,
which showed a strong variability in the aquifer levels with annual
oscillations up to 5m.
At the end of the analysis, for the detailed design it was decided
to consider:
as “most unfavourable” occurrence: the value of 48 m a.s.l. for
the artesian aquifers (A1) and 51.50 m a.s.l. for the phreatic
aquifer (A0),
as “most probable” occurrence: the value of 45.50 m a.s.l. for
the artesian aquifer (A1) and 49.00 m a.s.l. for the phreatic
aquifers (A0).
The contractual basic design considered a single value equal to
the value of 45.50m a.s.l. for the artesian aquifer (A1) and 49.00m
a.s.l. for the phreatic aquifer (A0).
Regarding the aforementioned influence of ground composition
and plasticity values variability, field tests have been performed
before the detailed design phase in order to evaluate the "most Figure 7 Jet grouting tests
428
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
429
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
This also allowed the beneficial 3D effects to be developed, as 8.1 Variability in the sizing of the supporting works
described in the following paragraph.
The design improvement proposal establishes two “project limit
configurations”: a “minimum configuration” and a “maximum
8. DESIGN DETAILS OF THE CONTRACTORS
configuration”, as a result of the approach described in the geology
IMPROVEMENT PROPOSAL
section referring to the levels of aquifers (they identify the “most
Following the "open and close" mode of execution, we could rely on probable” occurrence and the “most unfavorable” occurrence of
a set of beneficial static effects otherwise unavailable with a general water levels for aquifers A0 and A1).
excavation: Typical sections of the excavation support works have been defined,
1. three dimensional effect of contrast in the foot of the diaphragm where a variability has been hypothesized for ground anchor
wall due to the presence of the unexcavated perimetral berm characteristics (free strand length, bond length and drilling
and/or the foundation slabs; diameter), for jet grouting (geometries and spacing of columns,
2. plate behavior of the jet grouting band (the whole jet grouting execution quotas) and for the execution procedures of the
volume is affected by a limited section of diaphragm than would excavation steps below height 43.50 m a.s.l. as per figure 11.
be with a general excavation); By varying these parameters, it was possible to cover the
3. overload effect behind the diaphragm due to the weight of the cast combined effect of:
foundation slab that allows the increase of the passive earth 1. possible variations in the aquifers (A0, A1) level (i.e. load) within
pressure by increasing the effective stresses; the considered range;
4. undrained behaviour for the cohesive layer B during the 2. possible variations in the mechanical properties of jet grouting
construction phases below the excavation quota of 43.5 m. a.s.l. and anchor bond pull-out strength (due to the possible variations
Three-dimensional numerical analysis, using 3D Flac software, in layer B characteristics).
was performed to evaluate the beneficial effects on the stress-strain The method set out that the support works are dimensioned and
behaviour of the reinforced concrete diaphragm wall produced by verified in detail within the defined range (minimum to maximum),
the excavation and concrete pouring stages of small size segments depending on actual site conditions encountered during works.
(see figure 10). Two-dimensional analyses were not suitable to The definition of support works within the range defined in the
capture the confinement effects provided by the unexcavated berm detailed design occurred at each single excavation stage and in the
and the diffusion of the stresses due to the jet-grouting band; indeed, presence of significant increases in the aquifers level, for which “in
in a traditional two-dimensional calculation, the excavation of the progress detailed design” documentations were issued.
perimetral berms, even of a single segment, is equivalent to the
complete removal of the berm through the whole longitudinal
development of the diaphragm wall, situation this which does not
correspond to reality.
430
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
8.3 Field Testing of jet-grouting soil improvement and tieback In these occasions activities proceeded either on less critical
stranded anchors work fronts, or excavations proceeded according to the above
strategies 2.ii or 2.iii.
After the execution of the preliminary tests, by proceeding with the
excavations, the execution of field tests during construction were
also carried on according to an observational approach.
This was done along the perimeter, where there were expected
particular situations, in order to assess in detail the influence of
geotechnical and stratigraphic characteristics.
Detailed notes for the interpretation of the results of individual test
fields and calibration of project support works have been issued.
The execution of the field tests and controls during construction,
followed by the issuance of the specific “in progress detailed
design” reports, in conjunction with the "key" moments and
execution phases, allowed to eliminate the geotechnical and
stratigraphic uncertainties.
431
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
This situation has shown its own peculiarities, which may affect
the definition of mean values, most probable values and the limits of
acceptability of project hypotheses as well as the alternative
remedial solutions.
The Observational Method has also proved to be a mostly
suitable design criterion when dealing with situations in which the
construction methods, operating sequences and the statics of the
works are particularly related and the quantities are to be
rationalized.
The potential optimization of the quantities that the Method
offers, must necessarily be supported by detailed studies and
verifications, which ultimately lead to greater reliability of the
project, both in terms of cost and execution times.
In this sense, the Observational Method may also be framed into
Projects according to "cost-plus fee" or “fixed-rate” contractual
schemes, and also according to "lump sum" logic as the present case,
for the main benefit of the Client, especially where the limits of
variability of the design parameters can be clearly established.
9. CONCLUSIONS
The design improvement proposal based on the Observational
Method as developed by the Contractor together with its design
consultant Soil srl and approved by the Client, allowed to finish the
excavation works on time and with no significant interruptions.
Under the more specific point of view of the Design process, it
allowed for the main following benefits:
1. to include in the design possible "most unfavorable" groundwater
conditions and absorb the uncertainties about the geotechnical
aspects, extending the calculation hypothesis for provisional
works, which have direct effects on the quality of the support
structures, specially the maximum water table levels, in order to
ensure a greater continuity of the site activities (excavations and
casting of the foundation slab);
2. to rationalize the sizing and the dimensioning of the provisional
works (strands number, diameter and barrel length and jet Figure 14 Works near completion in the north area
grouting extension) as a partial compensation of the execution
burden related to the extension of the calculation hypothesis; 10. REFERENCES
3. to offer greater guarantees and reliability in general terms of risk
Ciria - Report 185 (1999) - Observational Method
management related to the execution of the project (timing of the
Eurocode 7 (BS EN 1997-1:2004) “Geotechnical design - Part 1:
work and execution costs) both for the Client and for the
General rules”
Company.
Decreto legislativo 12 aprile 2006, n. 163, Codice dei contratti
Therefore, in this case, the Observational Method is a valid
pubblici
design support for the execution of complex geotechnical works and
NTC2008 - Norme tecniche per le costruzioni - D.M. 14 gennaio
projects.
2008
Its application has been successfully extended from the case of
Peck, R. B. (1969). Advantages and limitations of the observational
underground works, more traditional, to the less usual case of large
method in applied soil mechanics. Geotechnique, 19: 171-
“open-sky” excavation works.
187.
432
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Due to rapid urbanization, the land available for construction becomes more and more scarce. Within a built-up environment,
the construction safety of a deep excavation becomes more crucial with the ever-increasing building density. For deep excavations in
mountain cities, the areas of the foundation pit to be excavated are generally the passive soil pressure zones for the upper existing slope.
Construction disturbance, weakening of the passive area, as well as the formation of even higher slope through superposition of foundation
pit to the upper slope, will result in more deformation and even failure of the slope. This study numerically investigates the influences of
excavation geometries, the system stiffness and the distance between the excavation and slope and develops simplified ultimate and
serviceability limit state models with regard to the overall factor of safety and the maximum lateral wall deflection of the supporting system,
respectively. Considering the uncertainties of the design parameters, a probabilistic framework combining the estimation models with First-
Order Reliability Method (FORM) is proposed to determine the probability that a threshold factor of safety or the maximum wall deflection
is exceeded. The study presents preliminary guidelines for reliability assessment of ultimate and serviceability limit state designs for deep
braced excavations adjacent to high slopes in mountain cities.
Keywords: braced excavation, upper slope, wall deflection, factor of safety, reliability assessment.
1. INTRODUCTION
2. FINITE ELEMENT ANALYSIS
With promotion of the Belt and Road Initiatives in China, the fast
development of urbanization progress has brought the great demand 2.1 Numerical modeling
of commodity housing and public transport facilities. Thus it is
inevitable that more deep excavations to be constructed for The PLAXIS2D software was utilized for the numerical simulations.
residence, commercial buildings as well as the sky scrapers designed The Mohr-Coulomb constitutive model was selected for the soil. A
aside the existing slopes in densely populated mountainous cities typical cross-section of the excavation and slope system, the
such as Chongqing and Guiyang. It is well known that for deep geometries as well as the properties of the soil and the supporting
excavations in mountain cities, the areas of the foundation pit to be elements are shown in Figure 1.
excavated are generally also the passive soil pressure zones for the
upper existing slopes. Consequently, the construction disturbances,
weakening of the passive area, as well as the formation of the even
higher slope through superposition of foundation pit to the upper
slope, will result in more deformation and even collapse of the
slope. Nevertheless, there are few investigations of the interaction
between braced excavation and the adjacent slope and the influence
of such interaction on the overall stability. Li et al. (2011)
investigated the stability of supporting system and the safety of deep
braced excavation adjacent to slope, through analyzing the influence
of excavation of Shangshuijing station Shenzhen Metro Line 5 on
side slope using FLAC3D. Wang et al. (2011) examined the
deformation characteristics and behaviors of retaining structures for Figure 1 Cross-sectional soil and wall profile
a complex geotechnical system comprising of a high building slope
and a nearby deep excavation, based on field instrumentations. The analyses considered a plane strain excavation supported by
Varzaghani and Ghanbari (2014) presented a new analytical model a retaining wall system near an unreinforced slope. The soil was
to determine the seismic displacements of the shallow foundations modeled by 15-noded triangular elements. The structural elements
adjacent to slopes. However, there is still a lack of systematic were assumed to be linear elastic with the wall represented by 5-
investigation of the key influential factors and the effects on the noded beam elements and 3-noded bar elements were used for the 6
ultimate limit state and serviceability limit state of the excavation levels of struts located at depths of 1 m, 4 m, 7 m, 10 m, 13 m and
and slope system. 16 m below the original ground surface. The nodes along the side
In this study, the global factor of safety FS obtained via the boundaries of the mesh were constrained from displacing
shear strength reduction (SSR) technique (also called c/φ reduction horizontally while the nodes along the bottom boundary were
method) is used as the criterion for the ultimate limit state and the constrained from moving horizontally and vertically. The left and
calculated maximum lateral wall deflection is adopted as the right vertical boundary extend far from the excavation to minimize
serviceability limit state criterion. It then numerically investigates the effects of the boundary restraints. The ranges of the design
the influences of the excavation geometries, the supporting system parameters varied are shown in Table 1.
stiffness, the distance between the new excavation and the existing The strut stiffness per meter EA is assumed as a constant at
slope on excavation responses including the global FS and the wall 3.0×105kN/m since the influence of strut stiffness on wall deflection
deflection using PLAXIS software. Estimation models with regard is not very significant when the strut is stiff (Poh and Wong 1997).
to both the ultimate and serviceability limit states are developed. A total of 162 hypothetical cases were analysed.
Probabilistic framework combining the proposed estimation models The construction sequence comprised the following steps:
with the First-Order Reliability Method (FORM) is adopted to the wall is installed (“wished into place”) without any
determine the probability that a threshold factor of safety or the disturbance in the surrounding soil;
maximum wall deflection is exceeded. This proposed approach
enables a cost-effective analysis to be conducted for a rational
design of excavation system adjacent to an existing high slope.
433
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
434
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
435
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Function (PDF) over the failure domain. Considering that the PDFs 2007). For nonnormals, the nonnormal distributions are replaced by
of the random variables are not known in most geotechnical an equivalent normal ellipsoid, centred at the equivalent normal
applications and the integration is computationally demanding when mean. Cells D3:E5 are parameters which are set corresponding to
multi-variables are involved, an approximate method, known as the the normal distribution in this study. The correlation matrix R in
First-Order Reliability Method (FORM) (Hasofer and Lind, 1974), cells G3:I5 are used to define the correlations between B, He and S.
is commonly used to assess the probability failure Pf. (Low 1996) The ni vector in cells J3:J5 contains equations for (xi ‒ uiN) / iN.
has shown that Microsoft EXCEL spreadsheet can be used to The design point (x* values) was obtained by using the
perform the minimization and determine reliability index. spreadsheet’s built-in optimization routine SOLVER to minimize
The reliability index and the probability of failure Pf for both the cell, by changing the x* values, under the constraint that the
the ultimate and the serviceability limit states can be performed performance function G(x*) = 0. Prior to invoking the SOLVER
using FORM based on the built PR and LR models. The ULS model search algorithm, the x* values were set equal to the mean values
Eq. (1) is incorporated into an EXCEL spreadsheet environment (30, 17, 4.5) of the original random variables. Iterative numerical
based on the approach by (Low and Tang 2007), from which the derivatives and directional search for the design point x* were
reliability index can be determined. Figure 7 shows a sample automatically carried out in the spreadsheet environment.
spreadsheet for computing the factor of safety FS where the Probabilistic assessment of SLS in Figure 8 is almost the same
statistics of the design parameters are the same as those used in the as Figure 7 except the G(x) formulations. For the detailed
previous section. The spreadsheet cells B3:B5 allows the selection procedures in performing the FORM spreadsheet framework to
of various distribution types for the input variables, including derive and the corresponding Pf, the paper published by Zhang
normal, lognormal, triangular etc. as explained in (Low and Tang and Goh (2012 can be referred to.
Figure 7 Calculation on β and Pf for ultimate limit state using FORM spreadsheet
Figure 8 Calculation on β and Pf for serviceability limit state using FORM spreadsheet
4.1 Probabilistic assessment of the ultimate limit state Figure 9 plots the influence of the various design parameters on the
and Pf of ULS. It is clear that both the coefficient of variation of
For either the braced excavation or the slope, there are design
the system stiffness COVS and the critical factor of safety FS_cr
guidance with regard to the choice of the critical factor of safety.
significantly influence the and Pf. In addition, the influence of
However, for the excavation and slope system, there are no
guidelines for the determination of such critical safety factor values. COVS on and Pf is also as significant as that for FS_cr. The plots
Thus the influence of the critical factor of safety FS_cr on and Pf in Figure 10 indicate that the influence of either B1 or COVS on
of ULS is examined in this study. and Pf is also obvious when different excavation widths B of 20, 30,
40 m are considered.
436
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(a) (b)
Figure 9 Influence of COVS and FS_cr on (a) and (b) Pf for B1=5m, B=30m, He=17m, S=4.5
437
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 10 also compares the influence of both the COVS and B1 Regression models for the ultimate and serviceability limit states
on and Pf for B=20, 30, 40 m, respectively, for He=17m, S=4.5, are developed respectively. Through the use of the automated
and chosen critical factor of safety FS_cr=2.0. It is obvious that spreadsheet search algorithm to determine the design point, to meet
becomes greater with increase of the excavation width B1 while the different target performance levels, the critical FS or the
decreases with the increase of excavation width B. Meanwhile, Pf threshold max. lateral wall deflection can be obtained. The
decreases as the excavation becomes further away from the slope. A influences of the key parameters, as well as the design uncertainties
greater excavation width B generally results in a larger Pf. Generally on the reliability index and the probability failure are examined. The
Pf converges to 0 when the separation is sufficient. However, procedures outlined in this paper can be used to obtain a rational
different B causes different convergence speeds. design of braced excavation adjacent to high slope and a cost-
effective analysis.
4.2 Probabilistic assessment of the serviceability limit state
There are also discussions as for the choice of the threshold lateral 6. REFERENCES
wall deflections for serviceability considerations. Figure 11 plots the
influence of COVS and the critical max. wall deflection δhm_cr on Brinkgreve, R.B.J., Bakker, H.L.(1991). "Non-linear finite element
and Pf for B=30 m, He=17m, S=4.5 and B1=20, 20, 40 m, analysis of safety factors. " In: Proc. 7th Int. Conf. on
respectively, indicating that both COVS and δhm_cr significantly Computer Methods and Advances in Geomechanics, vol. 23,
influence the and Pf. However, the influence of COVS on and Pf pp. 1117–1122.
is not as significant as that for δhm_cr, especially when COVS is Cheng, Y. M., et al. (2007). "Two-dimensional slope stability
greater than 0.20. has a tendency to grow with the critical analysis by limit equilibrium and strength reduction
maximum lateral wall deflections δhm_cr since the probability that a methods." Computers and Geotechnics 34(3): pp 137-150.
greater threshold is exceeded is much lower. decreases with the Dawson, E. M., et al. (1999). "Slope stability analysis by strength
increase of COVS. In addition, it can be observed that the influence reduction." Geotechnique 49(6): pp 835-840.
Hasofer, A.M., Lind, N.(1974). "An exact and invariant first-order
of B1 on is also significant since increases substantially with the
reliability format." J. Eng. Mech. ASCE 100 (1), pp 111–121.
separation B1.
Li, Y. H., et al. (2011). "Stability and Safety Analysis of Braced
Figure 12 shows the influence of COVS on for He=17m, S=4.5,
Excavation for Subway Station during Construction under the
δhm_cr=23 mm, B=20,30, 40 m and B1=10,15m respectively. It is
Condition of Side Slope." Applied Mechanics and Materials
clear that decreases as the variation of the system stiffness 99-100: pp 1166-1170.
becomes greater. It is logical that increases when the excavation is Lian, Z.-y., et al. (2001). "Stability analysis of excavation by
becoming further away from the slope. strength reduction FEM." Chinese Journal of Geotechnical
Engineering 23(4):p 5.
Low,B.K.(1996)."Practical probabilistic approach using spreadsheet.
" In: Shackelford, C.D., Nelson, P.P., Roth, M.J.S. (Eds.),
Uncertainty in the Geologic Environment, GSP 58. ASCE,
Reston, pp. 1284–1302.
Low, B.K. and Tang, W.H. (2004), "Reliability analysis using
object-oriented constrained optimization."Struct. Saf., 26(1),
pp 69-89.
Low, B.K. and Tang, W.H. (2007), "Efficient spreadsheet algorithm
for first-order reliability method." J.Eng. Mech. ASCE,
133(12), pp 1378-1387.
Poh TY, Wong IH, Chandrasekaran B, (1997) "Performance of two
propped diaphragm walls in stiff residual soils. "Journal of
Performance of Constructed Facilities, 11(4): pp 190–199.
Varzaghani, M. I., & Ghanbari, A. (2014). "A new analytical model
to determine dynamic displacement of foundations adjacent
to slope. " Geomechanics and Engineering. pp 561-575.
Figure 12 Influence of COVS on for He=17m, S=4.5, δhm_cr=23 mm, Wang, Q., et al. (2011). "analysis of slip-risk and dynamic
B=20,30, 40 m and B1=10,15m monitoring of a high building slope fringed a deep foundation
pit." journal of safety and environment .11(2):p 6.(in Chinese)
5. SUMMARY AND CONCLUSIONS Zhang, W. and A. T. C. Goh (2012). "Reliability assessment on
ultimate and serviceability limit states and determination of
This paper presents numerical investigations about influence of critical factor of safety for underground rock caverns."
braced excavation on the existing slope, from perspectives of the Tunnelling and Underground Space Technology 32: pp 221-
global factor of safety and the maximum lateral wall deflections. It 230.
also proposed probabilistic framework for quantitative assessment Zienkiewicz, O.C., et al.(1975). "Associated and non-associated
of both the ultimate and the serviceability limit states in view of visco-plasticity in soil mechanics. "Geotechnique 25 (4), pp
some design and construction uncertainties. 671–689.
438
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Most previous studies focused on the performance of braced excavations in clays, where limited publications involved braced
excavations in sands. In this study, to better understand the performance of braced excavations in sand, a series of two-dimensional (2D) and
three-dimensional (3D) finite element analyses using the Hardening Soil (HS) model of PLAXIS software were performed to investigate the
influences of soil properties, wall stiffness and excavation geometries, on the maximum wall deflection induced by braced-excavation in
sand.. Results show that the maximum wall deflections calculated by 3D analysis are greater than those from 2D. Based on the numerical
results, regression models were developed for estimating the maximum wall deflections induced by braced excavation in dense sand and
medium dense sand, respectively.
Keywords: Braced excavation, wall deflection, sand, finite element analysis, Hardening Soil (HS) model, regression models
439
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
440
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Hardening soil parameter Medium dense sand Dense sand Stiff clay
Parameters Unit (drained) (drained) (undrained)
γunsat kN/m3 drained
17 18.2 20
γsat kN/m3 19.8 20.3 20
kx=ky=kz m/day 110-8 110-8 110-8
E50ref kN/m2 30000 48000 14847
Eoedref kN/m2 30000 48000 14847
Eurref kN/m2 90000 144000 44540
c kN/m2 0 0 0.05
° 34.3 38 33
Ψ ° 4.3 8 0
νur [-] 0.3 0.35 0.2
pref kN/m2 100 100 100
m [-] 0.544 0.45 1
K0nc [-] 0.436 0.384 0.455
Rf [-] 0.938 0.9 0.96
Rinter [-] 0.8 0.8 1
A series of FE simulations using Hardening Soil (HS) model different excavation stages from 2D analyses in medium dense sand
were carried out to investigate the influences of soil properties, for wall stiffness α = 0.06 and 1.0. For the 3D rectangular braced
excavation geometries, wall stiffness on the wall deflections excavation, only the profiles of the horizontal wall deflection at the
induced by excavation. For brevity, the numerical results of the centre of excavation are presented because the maximum lateral
study are presented with the main findings, as described in the next displacement occurs at this particular centreline location for
sections. symmetry. For brevity, only wall deflections caused by braced
excavations in medium dense sand with L/B = 2.8 for α = 0.06 and
3. FINITE ELEMENT ANALYSES 1.0 are plotted in Figure 4.
Figure 3 shows the typical wall deflection profiles corresponding to
441
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
442
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Soil type a0 a1 a2 a3
Medium dense sand 0.1896 0.36193 -0.18585 1.757
Dense sand 0.1067 0.2847 -0.09736 1.8802
4. ESTIMATION MODELS
Based on the numerical results, Logarithmic Regression(LR) models
and Polynomial Regression (PR) models have been developed for
estimating the maximum wall deflection δhm induced by braced
excavations in dense sand and medium dense sand, respectively.
Four input parameters (Sv, L/B, α and He) are considered. For 2D
numerical results, the optimal regression equation for wall
Figure 7 Predicted maximum wall deflection δh* versus δhm by 2D
deflection takes the following form:
FEM
h* = a0S va1(L/ B ) a2 α a3 He a3 (1)
For 3D results, the optimal regression equation takes the form of:
443
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
444
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: The opening of Downtown Line 3 (DTL3) in October 2017 marked the longest underground and driverless MRT line in
Singapore. The complicated technical challenges in DTL3, especially Package A, were associated with the nature of its geological strata and
construction of underground infrastructure below existing buildings and structures which were conserved by the Authority to some extent of
the alignment. Due to its vertical alignment, tunnelling works in DTL3-A was carried out within the Kallang Formation which was recent
deposits consisting of soft/loose soils. At some extent of the alignment, tunnelling works was significantly close to conserved shophouses
which were very sensitive to induced settlements. Settlement predictions, risk assessments and mitigation measures were therefore carried
out to ensure safety during TBM tunnelling. The monitoring regime and protection measures were successfully implemented with actual
ground movements to be well within the design predicted movements. This paper presents some of the technical challenges in DTL3 Package
A in relation to tunnelling works in soft soil condition and under conserved shophouses at some extent of the alignment.
Keywords: Downtown Line Stage 3, control tunnelling, face pressure, damage assessment, building protection measures.
445
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
soil condition. The target TBM face pressures for each tunnel bound
and drive are shown in Figure 4.
446
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ensure that no loss of face pressure will be encountered resulting in damage classification table first put forward by Burland et al. (1977)
excessive ground movement during TBM excavation. as shown in Table 4.
447
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
448
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
6. REFERENCES
Burland, J.B., et. al. (1977) “Behaviour of foundations and
structures”. Proceedings of the 9th International Conference of
Soil Mechanics and Foundation Engineering, pp495-546.
New, B.M., and O’Reilly, M.P. (1991) “Tunnelling induced ground
movements: Predicting their Magnitude and Effects”, 4th
International Conference on Ground Movements and
Structures, pp671-697.
Shirlaw, J.N., et. al. (2003) “Local settlements and sinkholes due to
EPB tunnelling”, Proceedings of the Institution of Civil
Engineers, Geotechnical Engineering 156, pp193-211.
Land Transport Authority (LTA) (2010) Civil Design Criteria for
Road and Rail Transit Systems, February 2010, Rev A1.
Arup. (2011) C933 Geotechnical Interpretative Baseline Report
(Tender).
Figure 12 Five-foot way protection system
449
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
HKIE Geotechnical Division. (2014) Ground Control for EPB TBM Ong, C.W., et. al. (2015) “A case study of twin bored tunnelling
Tunnelling, GEO Report No. 298. under mixed-face soil – Bendemeer MRT station project
Poea, J., et.al. (2014) “Protection of old shophouses due to MRT (Downtown Line 3), Singapore”, Proceedings of the 15th
excavation & tunnelling works on Downtown Line Stage 3 Asian Regional Conference on Soil Mechanics and
C933”, Proceedings of Underground Singapore 2014, Session Geotechnical Engineering, pp176-181.
6.5, pp33. Urban Redevelopment Authority (URA). (2017) Conservation
Guidelines, December 2017 edition.
450
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This paper presents soil water content and pore-water pressure distribution in a vegetated residual soil slope in response to a
one-year seasonal variation in Singapore. The measurements were taken by tensiometers and soil moisture sensors TM1-SM1, TM2-SM2,
TM3-SM3 and TM4-SM4 installed at vertical distances of 2.0, 2.0, 1.37 and 0.23 m from slope surface. Seasonal variation was represented
by rainfall and actual evaporation calculated based on data collected by a weather station installed at the site. The field monitoring shows
that the soil water content and pore-water pressure measurements by TM1-SM1, TM2-SM2, TM3-SM3 were representative of soil-water
characteristic curve (SWCC) of the residual soil. On the other hand, the measurements by TM4-SM4 shows a lower pore-water pressure
during dry period and a lower volumetric water content during wet period as compared to SWCC of residual soil. The study showed that the
response recorded by TM4-SM4 was representative of top soil used as media for vegetation planting. The study also showed that the larger
variation of pore-water pressure recorded by TM4 in December 2016 and January 2017 was due to the long dry period with high temperature
which is not normal in Singapore during these months.
Keywords: Field monitoring, soil water content, pore-water pressure, soil-water characteristic curve, vegetated soil slope.
451
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ET C K C ET O (1) protection rod and the pyranometer were installed on top of the
where ETo is the grass reference evapotranspiration in mm/day mast. The rain-gauge was installed at a height of 1.6 m from the
which can be calculated as follows: ground surface together with the data logger. Soil temperature (ST)
was measured at depths of 0.10 m, 0.15 m, 0.20 m, and 0.25 m.
900
0 , 408 ( R n G ) u 2 (e S ea ) The tensiometers (TM) and soil moisture sensors (SM) were
(T 273 ) (2)
ET O installed within residual soil layer at a depth of 2 m from crest and at
(1 0 ,34 u 2 ) distances 0.4 m; 2.4 m; 3.8 m and 4.6 m from the slope face. The
jet-fill tensiometers were calibrated properly prior to installation and
where T = mean air temperature (oC); u2 = wind speed (m/s) at 2 m then inserted into a tube to the specified depth. The reading from
above the ground; Rn = net radiation flux (MJ/m2/day; G = sensible transducer was verified by a bourdon gauge attached to it. The
heat flux into the soil (MJ/m2/day); es = saturation vapour pressure response of tensiometers to pore-water pressure change must be
(kPa); ea = actual vapor pressure (kPa); ∆ = slope of saturation checked regularly to ensure the quality and physical performance of
vapour pressure curve; and γ = psychrometric constant. The detailed high air-entry ceramic tips. In this study, regular maintenance of the
calculation of this method can be found in Zotarelli et al. (2009). tensiometers was conducted twice a week by refilling the jet-fill
Typical Kc values are given in Lazzara and Rana (2010), for grass, reservoir with de-aired water and flushing the tensiometers to
Kc= 1. remove the accumulated or trapped air in the tubes caused by
The water flow through an isotropic unsaturated soil is cavitation of water and air diffusion through the ceramic tip. Soil
formulated using Darcy’s law as follows (Fredlund and Rahardjo, moisture sensor used in this study was of TDR type which was
1993): capable of measuring soil moisture up to saturation (0–100%) with
accuracy of 1% and response time of 0.5 second. All soil moisture
hw hw hw sensors were tested in water and air environment and their values
k w m 2w w g (3)
kw corresponded to the values specified by manufacturer which was
x x y y t about 70 to 90% in pure water and zero in air.
with,
uw (4)
hw z
wg
2. METHODOLOGY
2.1 Field Instrumentation
The investigated slope in this study is located at the central part of
Singapore within residual soil from Bukit Timah Granite. Figure 1
shows diagram of residual soil slope and the field instrumentation
carried out in this study. The instrumentations can be divided into
two parts; the first part was weather instrumentation for climatic
data measurements while the second part was instrumentation for
quantifying soil response to the flux boundary conditions. The
monitoring period was one year (1st July 2016 to 30th June 2017).
The weather instrumentation includes tipping bucket rain gauge
for rainfall, pyranometer for solar radiation, wind monitor for wind
speed and direction, as well as temperature and relative humidity
probes. The weather station for climatic data measurement was
located near the toe. A 10 m high galvanized mast was erected on a
concrete footing. The wind monitor, the solar panel, the lightning Figure 2. Data acquisition system used at the study site
452
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
2.2 Soil Properties than the mean monthly rainfall for the months in Singapore.
November 2016 was the wettest month during the monitoring period
The slope was formed by residual soil with a thin layer of top soil as with 21 rainfall days accumulating to 311 mm rain, while August
media for growing vegetative cover. The soil properties required for 2016 was the driest month with only 8 days of rainfall accumulating
this study were investigated by performing laboratory tests on to 98.6 mm of monthly rainfall. Past data (1981 – 2010) showed that
samples taken from the field. The grain size distribution shows that the wettest month in Singapore was November while the driest
about 50% of the material is clay. The unified soil classification month were June and February. Figure 5 shows that the maximum
system USCS (ASTM D2487-00) categorized the soil to be highly daily rainfall occurred on 23rd January 2017 i.e. 103.8 mm. The
plastic clay (CH). The bulk density of the residual soil was 1.8 month of January 2017 represented extreme condition because it
Mg/m3. started with a two-week dry period followed by very wet period
The saturated permeability of the soil was determined using the towards the end of the month. The cumulative rainfall in January
flexible-wall saturated permeability test (ASTM D5084-10). The 2017 was 298.4 mm. Even though the number of days with rainfall
saturated permeability of the soil is 6×10-7m/s. In order to perform in January was 17, much of the rainfall (217.8 mm) was actually
analyses related to water flow in unsaturated soil, permeability occurred on 18th to 24th January.
functions of the soil need to be obtained. Permeability function can
be determined from the soil-water characteristic curve (SWCC)
using a statistical method proposed by Childs and Collis-George
(1950). The procedure for the prediction can be found in Fredlund
and Rahardjo (1993). The SWCC of the soil from the site was
obtained using Tempe cell and pressure plate according to ASTM
D6838-02. The wetting and drying SWCCs of the residual soil are
shown in Figure 3. The permeability functions obtained using the
prediction method based on wetting and drying SWCCs are shown
in Figure 4.
3. RESULTS AND DISCUSSION Figure 6. Mean monthly and number of rainfall days
453
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
both rainfall and air temperature, the month of January 2017 3.2 Evapotranspiration
represent an extreme condition. Past data indicated that the
maximum temperature usually occurred in March to May while Figure 8 shows the ETc calculated using FAO-Penman or FAO-56
minimum temperature occurred in December and January. In this Method for ETo and coefficient of 1 for Kc, plotted together with
study, the trend of higher temperature in March to May was rainfall data. The minimum, mean and maximum daily potential
overcome by heavy downpours (Figure 6). evaporations in the study area were 0.66, 1.92, and 3.56 mm,
Soil temperature was measured at depths of 0.10, 0.15, 0.20 and respectively. The highest daily ETc occurred in June 2017 while the
0.25 m from ground surface. Field measurements performed in this lowest was in November 2016. The highest monthly ETc in June
study indicated that the soil temperature varied with depth and the was 65.29 mm) while the lowest monthly ETc in November was
time of the day. The least variation with the time of the day was 47.77 mm. The ETc obtained in this study were slightly less than
obtained in the measurement of soil temperature at depth of 0.20 m those reported by Rahardjo et al. (2017) based on their studies in
from the ground surface. Figure 7a shows the variation of soil different parts of Singapore. The minimum and maximum ETc
temperature measured at depth 0.20 m as compared to the maximum calculated in their studies were 0.55 and 4.3 mm/day, respectively.
and minimum daily temperatures. The soil temperature was In their study, the highest daily ETc occurred in February while the
relatively low in November, December and January, and relatively lowest was in June 2017.
high in May, June and July.
Relative humidity plays an important role in the prediction of
ET. The minimum and maximum relative humidity recorded in the
study site were 67.4% to 99% with a mean value of 80%. The
recorded RH was higher than the typical range recorded in
Singapore i.e. 60 to 95%. The maximum relative humidity occurred
in November due to the high number of days with rainfall. The
lowest relative humidity was recorded in August due to the least
number of rainfall days. Figure 7b shows the maximum and
minimum daily RHs during the monitoring period. Observation of
Figure 7b indicates that low relative humidity was maintained
Figure 8. Evapotranspiration on slope surface at the study site from
during the first two weeks in both December 2016 and January 2017
July 2016 to June 2017
due to lack of rainfall.
The minimum and maximum solar radiations during the
3.3 Pore-water pressure measurements
monitoring period were 1.12 and 14.76 MJ/m2/day, respectively
with a mean value of 7.93 MJ/m2/day. These values are the typical The response of soil slope to rainfall and evapotranspiration from
values for Singapore. The maximum solar radiation was recorded in vegetated surface in terms of pore-water pressure is presented in
the dry month of August 2016. The variation of solar radiation at the Figure 9. The figure shows that the measurements at TM1 – TM3
study site during the monitoring period is shown in Figure 7c. were quite consistent, while the measurements at TM4 varied
Observation of Figure 7c also indicated that high solar ration was significantly especially from November 2016 to January 2017.
maintained during the first two weeks in both December 2016 and Based on the evaluation of rainfall and climatic data, this period
January 2017. represented extreme changes from dry to wet conditions.
The maximum wind speed (4.92 m/s) was recorded in February As shown in Figure 1, TM1 and TM2 have vertical distances of
2017. The recorded wind speed was lower than the range of wind 2 m from crest while TM3 was 1.37 m below the slope face. TM4
speed in Singapore i.e. 0–13.375 m/s with a mean value of 2.65 m/s. was very close to the slope face i.e. 0.4 m horizontal distance or
This may be because the location of the study site was surrounded only 0.23 m vertical distance. Thus, TM4 was more affected by flux
by high rise buildings. The variation of wind speed throughout the boundary conditions as compared to the other tensiometers.
year is also presented in Figure 7c.
454
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
by TM4 were below the SWCC wetting curve of the residual soil. It water pressure to be recorded by TM4 which was located closest to
was deduced that the soil response at TM4-SM4 was more the ground surface (Figure 9).
representative of the top soil instead of the residual soil. Readings of pore-water pressure and soil water content recorded
Soil samples were collected from the slope surface to a depth of by TM1-SM1, TM2-SM2, TM3-SM3 are representative of the
20 cm. SWCC and saturated permeability tests were carried out on SWCC of the residual soil. Thus, the readings are representative of
the samples using the same procedure as for the residual soil. It was the soil where the instrumentations were installed. However, the plot
found that the saturated volumetric water content of the top soil was for TM4-SM4 located at a vertical distance of 23 cm from the
lower than that of the residual soil. The top soil had a lower density ground surface was more representative of the top soil which has a
as compared to that of the residual soil, thus the higher void ratio lower density and a higher coefficient of permeability as compared
shifted the SWCC to the left, resulting in a lower air-entry value. to the residual soil. This showed that the readings of instrumentation
The saturated coefficient of permeability of the top soil was 1 × 10-5 closer to the ground surface was affected by the flux boundary
m/s which was higher than that of the residual soil (6 × 10-7 m/s). conditions and the consistency of surface soil as well as vegetative
cover.
4. REFERENCES
ASTM D5084-10 Standard Test Method for Measurement of
Hydraulic Conductivity of Saturated Porous Material using a
Flexible Wall Permeameter, ASTM International, West
Conshohoken. PA.
ASTM D6838-02 Standard Test Method for Determination of
SWCC for Desorption using Hanging Column, Pressure
Extractor, Chilled Mirror Hygrometer or Centrifuge, ASTM
International, West Conshohoken. PA.
ASTM D 2487-00 Standard Practice for Classification of Soils for
Engineering Purposes (Unified Soil Classification System) ,
ASTM International, West Conshohoken. PA.
Figure 10. Plot of pore-water pressure and volumetric water content
Allen R.G., Pereira L.S., Raes D., Smith M. (1998) Crop
measured at the study site from July 2016 to June 2017 in SWCC of
evapotranspiration: guidelines for computing crop water
residual soil.
requirements Food and Agriculture Organization of the
United Nation, Rome.
The field data from TM4-SM4 were plotted on the SWCC
Blight G.E. (1997) The “Active” Zone in Unsaturated Soil
curves of the top soil as shown in Figure 11. The figure shows that
Mechanics. 1st GRC Lecture, Nanyang Technological
the field data could be plotted well inside the hysteresis of the
University, Singapore.
SWCC of the top soil. The range of suction was from 8 to 90 kPa,
Childs E.C., Collis-George G.N. (1950) “The permeability of porous
while the range of volumetric water content was from 13 to 30%.
materials”. Proc. Royal Society of London 210A. pp. 392-
The maximum volumetric water content was less than the saturated
405.
volumetric water content of the surface soil. It may be attributed to
Chipp, P.N., Henkel, D.J., Clare, D.G. & Pope, R.G. (1982) Field
some water intercepted by vegetation (grass) on the ground surface.
Measurement of Suction in Colluvium Covered Slopes in
Hong Kong. Proc. 7th Southeast Asian Geotechnical Conf.,
Hong Kong, pp.49-62.
Fredlund D.G., & Xing A. (1994) “Equations for the Soil-Water
Characteristic Curve”. Canadian Geotech. Journal 31 pp. 521-
532.
Fredlund, D. G. & Rahardjo, H. (1993). Soil Mechanics for Un-
saturated Soils. New York: John Wiley & Sons, Inc.
Joel, A., Messing, I., Seguel, O., & Casanova, M. (2002) “Meas-
urement of surface water runoff from plots of two different
sizes”. Hydrological Processes. 16(7) pp. 1467-1478.
Lazzara P. & Rana G. (2010) “The use of crop coefficient approach
to estimate actual evapotranspiration: a critical review for
major crops under Mediterranean climate”. Italian Journal of
Figure 11. Plot of pore-water pressure and volumetric water content Agrometeorology 2 pp. 25-39.
measured by TM4 from July 2016 to June 2017 in SWCC of top Lee, M.L. Gofar N. & Rahardjo H. (2009) “A Simple Model for
soil. Preliminary Evaluation of Rainfall-Induced Slope Instability.
Engineering Geology”. 108(3-4) pp. 272-282.
4. CONCLUSIONS Li, A.G., Yue, Tham, L.G. and Lee C.F., (2005) “Field-monitored
variations of soil moisture and matric suction in a saprolite
Field measurements of flux boundary conditions and soil response slope”. Canadian Geotechnical Journal. 42, pp. 13–26.
were carried out in a residual soil slope at the central part of Gasmo, J.M., Rahardjo, H. and Leong, E.C. (2000) “Infiltration
Singapore. The measurements presented in Figures 5 and 7 specified Effects on Stability of a Residual Soil Slope”. Computer and
that the seasonal variation of rainfall was quite different from the Geotechnics, 26, April, pp.145–165.
typical trend in Singapore (Figure 6). The dissimilar trend also Gitirana Jr., G., Fredlund, D.G., Fredlund, M. (2006) “Numerical
observed for the other flux boundary conditions such as air modeling of soil-atmosphere interaction for unsaturated
temperature and relative humidity. Highly variable conditions surfaces”. Proc. 4th Intl. Conf. Unsaturated soils, US., pp.658-
occurred during the months of December 2016 and January 2017. 669.
This could be identified by unusually high air temperature and long Gitirana Jr., G., Fredlund, D.G., Fredlund, M. (2005) “Infiltration-
period of dry condition which are different from the normal runoff boundary conditions in seepage analysis”. Proc. 58th
conditions in Singapore. The long dry period in combination with Canadian Geotech. Conf. Canada, pp.516-523.
low relative humidity and high solar radiation allowed high pore-
455
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Gofar N. Lee, M.L. & Kassim, A. (2008) “Response of Suction Rahardjo, H., Satyanaga, A. and Leong, E.C. (2013). “Effects of
Distribution to Rainfall Infiltration in Soil Slope”. Electronic Flux Boundary Conditions on Pore-water Pressure
J. of Geotechnical Engineering. 13 (E), pp.1-13 Distribution in Slope”. Engineering Geology, 165, October,
National Environment Agency, 2010.Meteorological Services Data. pp. 133–142.
National Environment Agency, Singapore. Rahardjo, H., Lee, T.T. Leong, E.C. and Rezaur, R.B. (2005).
Ng, C.W.W., Springman, S.M. and Alonso E.E. (2008). “Response of a Residual Soil Slope to Rainfall”. Canadian
“Monitoring the Performance of Unsaturated Soil Slopes”. Geotechnical Journal, 42(2), pp. 340–351.
Geotechnical and Geological Engineering. 26(6), pp.799-816. Rahardjo, H., Leong, E.C., Gasmo, J.M. and Tang, S.K. (1998)
Ng. C.W.W., Zhan, L.T., Bao, C.G., Fredlund, D.G. and Gong, “Assessment of rainfall effects on stability of residual soil
B.W., (2003) “Performance of an Unsaturated Expansive Soil slopes”. Proc. 2nd Intl. Conf. Unsaturated Soils, China, pp.
Slope Subjected to Artificial Rainfall Infiltration”. 280-285.
Géotechnique. 53(2), pp. 143-157. Sattler, P.J. & Fredlund, D.G. (1991). “Modelling vertical ground
Penman, H.L (1948) “Natural evapotranspiration from open water, movements using surface climatic flux”. Proc. Geotechnical
bare soil and grass”. Proc. Royal Society London Ser. A.193, Engineering Congress, US., pp. 1292-1306.
pp.120-145. Yunusa, G.H. Kassim, A. & Gofar, N. (2014) “Effect of Surface
Rahardjo, H., Amalia, N., Leong, E.C. Harnas, F.R. Lee, T.T. and Flux Boundary Conditions on Transient Suction Distribution
Fong, Y.K. (2017) “Flux Boundary Measurements for the in Homogeneous Slope”. Indian J. Science & Technology,
study of Tree Stability”. Landscape Ecol Eng. 13:81–92 7(12), 2064–2075
Rahardjo, H., Satyanaga, A. Harnas, F.R. and Leong, E.C. (2014) Zotarelli L, Dukes MD, Romero CC, Migliaccio KW, Morgan KT
“Comprehensive Instrumentation for Real Time Monitoring (2009) Step by step calculation of Penman-Monteith
of Flux Boundary Conditions in Slope” Proc. 3rd Italian evapotranspiration (FAO-56 Method), University of Florida.
Workshop on Landslides, Italy.
456
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: When constructing road embankment over peat, ground improvement is needed to avoid problems such as bearing capacity
failure and excess post-construction settlement. The cement stabilization technique is one of the new proven ground improvement which can
greatly improve ground in a short period of time. In this study a series of laboratory testing using unconfined compression test on peat mixed
with some type of cements with different chemical compositions, including a special cement which contains large amount of SO 3 than other
type of cement, was conducted. The tested peat was collected in a site at Dumai in Sumatera Island, Indonesia. As a result, the cement
stabilization has considerable potential to improve strength for the peat from very soft to stiff and hard consistency.
Keywords: ground improvement, cement stabilization, peat, unconfined compression strength, modulus of elasticity.
2. METHODOLOGY
A laboratory mixing test was performed on Dumai peat with
ordinary and special cement commercially available in Japan for
cement stabilization in order to clarify their effect on the peat and
confirm the suitability of implementation this technology in
Indonesia, especially with similar condition with Dumai peat.
2.1 Soil Property of Peat tested Figure 1 Location of sampling site
Peat sample for the laboratory test was collected from Dumai,
Sumatera Island, Indonesia. (Figure 1). Table 1 shows engineering Table 1 Engineering Soil Properties of Peat Sample
soil properties of the peat sample which resulted high water content Specific gravity Gs 1.51
of over 700%, ignition loss of over 95% and low pH value of 3.4. Natural water content Wn (%) 769
The peat included a lot of organic materials such as Bitumen and Iginition loss Li (%) 95.2
Humic acid that known to hinder the solidification of cement (e.g., PH 3.4
Okada et al, 1983; Noto, 1991; Hayashi and Nishimoto, 2005). The Degree of decomposition (von Post) H3-H4
physical and chemical properties of the peat are difficult soil
Content of organic material (%) Bitumen 8.3
condition for cement stabilization.
Humid acid 56.2
Unconfined compression strength UCS (kPa) No. 1 7.3
(undisturbed sample before stabilization) No.2 12.8
457
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
458
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
459
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
4. CONCLUSIONS
In this study, a series of laboratory mixing testing on peat collected
at a site of Dumai in Sumatera Island, Indonesia using different
types of cement was conducted. The main results can be
summarized as follows:
1. The tested peat has high water content of over 700%, ignition
loss of over 95% and low pH value of 3.4. The physical and
chemical properties are difficult soil condition for cement
stabilization.
2. In cases in which Portland cement or blast furnace cement were
used, small strength enhancement from unconfined compression
strength (UCS) resulted from curing time of 7 and 28 days.
Stabilized peat can only achieve stiff consistency.
3. When special cement (product name: ET201) containing large
amounts of sulfur trioxide or aluminium oxide was used,
enhancement of high UCS result was obtained. Stabilized peat
can achieve hard consistency.
4. For most of the data except for some cases in the ET201, the
UCS after 28 days curing ranged from 1.5 to 3.0 times the UCS
after 7 days curing.
5. Without depending on the type of cement, the modulus of
elasticity (E50) increased linearly with the increase of the UCS.
Based on the result, an experimental correlation between the E50
and the UCS is presented.
This study was collaboratively carried out based on
“Agreement on Research Exchange and Cooperation between Civil
Engineering Research Institute for Cold Region (CERI), Japan and
Institute of Road Engineering (IRE), Indonesia”, in order to
technically support national projects of highway construction over
peat in Indonesia.
5. REFERENCES
Civil Engineering Research Institute for Cold Region (CERI). (2017)
Manual for Countermeasure for Peat Soft Ground, pp.1-6. (in
Japanese)
Hayashi, H. and Nishimoto S. (2005) Strength Characteristic of
stabilized peat using different types of binders, Proceedings
of The International conference on Deep Mixing Best
Practice and Recent Advances (CD-R), Stockholm.
Huat, B. B. K., Prasad, A., Asadi, A. and Kazemian, S. (2014)
Geotechnics of Organic Soils and Peat, CRC press, pp.81-95.
Institution of Road Engineering (IRE). (2001) Guideline of
Indonesian Road Construction over Peat and Organic Soils
Figure 7 Relationship between UCS and modulus elasticity Japanese Geotechnical Society (JGS). (2016) Laboratory Testing
Standards of Geomaterials, Vol.2.
Japanese Geotechnical Society (JGS). (2015) Laboratory Testing
Standards of Geomaterials, Vol.1.
Kitazume, M. and Terashi, M. (2013) The Deep Mixing Method,
CRC press, pp.83-84.
Noto, S. (1991) Peat Engineering Handbook, CERI, pp.102-110.
Okada, Y., Kutara, K. and Miki, H. (1983) Effect of humic acid on
soil stabilization, Proceedings of the 53rd Annual Conference
of Japan Society of Civil Engineers, pp.467-468. (in
Japanese)
Terzaghi. K., and R.B. Peck. (1967). Soil Mechanics in engineering
practice, Wiley New York p.729.
460
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Thick deposits of very soft marine clay can be found in Klang which lies along the West Coast of Malaysia. Basement excavation
up to 3.5m below the existing ground level has been carried out in Bukit Tinggi, Klang using three ground retention methods namely propped
sheet pile wall, cantilever sheet pile wall and cantilever contiguous bored pile wall. This paper elaborates on the implemented ground retention
methods and explores the ground’s response to the excavation in soft marine clay through the recorded instrumentation data.
The estimated wall deflections from the 2D FE analysis were compared against the measured field data to verify the adopted soil model and
the FE analysis. The results show that the adopted soil parameters are realistic and the sequence of work adopted in the FE analysis and
subsequently in the basement excavation work has minimized the wall deflections.
Keywords: Soft clay, propped excavation, cantilever excavation,
1. INTRODUCTION
The coastal areas of Peninsular Malaysia are typically composed of
3. SOIL CONDITION
thick deposits of very soft marine clay generally deposited during the
Holocene period. The soft marine clay can extend to depths exceeding The subsoil at this site based on the boreholes that were carried out
50m below the existing ground level close to the coast. indicates that the site is underlain by very soft clay with depths
Bukit Tinggi in Klang, Selangor is a fast developing varying from 24m to 30m below the existing ground level. Thin sand
township which is predominantly composed of low-rise residential lenses were found at selected locations across the site. The hard clay
and commercial development. Basements are rare in this area due to layer is typically found at depths ranging from 32m to 40m below the
the difficulties and high cost usually associated with deep excavation existing ground level.
in very soft marine clay. Table 1 shows the summary of the soil description and
Mixed commercial and residential development has been properties where γ is the bulk density of the soil (kN/m3), Su is the
proposed in this area in two (2) adjoining plots i.e. Plots A and B with undrained shear strength of the soil (kPa), N is the SPT N of the soil
one (1) level of basement. The plots are flanked by existing (blows/ft) and Eu is the Undrained Young’s Modulus of the soil (kPa).
commercial and residential development on either sides. The
basement excavation works were carefully planned to minimise the Table 1 Summary of the soil description and properties
settlement of the surrounding and to avoid distress to the adjacent
structures. Depth Soil γ Su Eu
A soil model was developed using the available soil (m)
investigation (SI) data and a Finite Element (FE) analysis was carried 0–8 Very Soft Clay 16 10 1200
out by simulating the full excavation and construction process. 8 - 20 Very Soft Clay 16 15-25 2200-3500
Instrumentation was installed to monitor the ground movement. The
field behaviour of the retention system is compared with the FE 4. PROJECT DESCRIPTION
analysis and discussed.
The proposed development comprises of two (2) plots which share a
2. SITE GEOLOGY boundary as shown in Figure 2.
The proposed site is underlain by Quaternary Deposit as shown in
Figure 1.
This formation consists of extensive deposits of
unconsolidated/semi-consolidated marine deposit consisting of
mainly CLAY deposited under marine environment. This is likely to
be underlain by the Pre-Quaternary sedimentary formation.
461
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5.1 Plot A
Propped temporary sheet pile wall was proposed for Plot A. The
sequence of works for Plot A is shown in Figures 3 - 8 and is
elaborated below.
Stage 1 : Install temporary sheet piles along project boundary. Allow Figure 5 Stage 3 Construction Sequence
1m working space from temporary sheet pile to permanent RC wall.
Install first two (2) rows of foundation piles to set using jack-in Stage 4 : Excavate temporary slope and cast pile cap.
method from existing platform level
462
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Stage 5 : Cast remaining slab and RC wall. Allow for opening in the
wall for the strut. Repeat Stages 4 and 5 between alternate struts
without removing the soil berm beside these struts until the full slab
and RC wall has been cast.
Stage 6 : Fill the gap between the RC Wall and the sheet pile wall up
to maximum 2m using suitable fill.
The sequence of works for the cantilever temporary sheet pile wall is
shown in Figure 10 and is elaborated below.
Stage 7 : Remove the temporary strut, waler and sheet pile wall.
Backfill to the existing ground level.
5.2 Plot B
Along the site boundary adjoining Access Road 1, cantilever CBP
wall was installed taking the functioning water pipe into
consideration. Along the site boundary adjoining Access Road 2,
cantilever sheet pile wall was installed.
The sequence of works for the cantilever CBP wall is shown in Figure
9 and is elaborated below.
463
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Stage 4 : Install foundation spun piles from base of excavation and pipe was monitored during the excavation works and no distress was
cast pile cap, base slab and RC wall. observed on the pipe at the end of the construction works. The
Stage 5 : Fill the gap between the RC Wall and the sheet pile wall cantilever CBP wall and sheet pile wall also allowed for speedier
with suitable granular fill. construction.
Stage 6 : Remove the temporary sheet pile. Reinstate the ground 8. CONCLUSION
behind the RC wall to the finished platform level.
Basement excavation up to 3.5m below the existing ground level has
6. INSTRUMENTATION been carried out in Bukit Tinggi, Klang using three ground retention
methods namely propped sheet pile wall, cantilever sheet pile wall
Instrumentation in the form of ground settlement markers and and cantilever contiguous bored pile wall.
inclinometers were installed behind the temporary sheet pile walls A FE analysis was carried out to confirm the details of the
and CBP wall for both Plots A and B. retention system. The estimated wall deflections from the 2D FE
The monitoring was carried out on a daily basis during the analysis were found to be comparable to the measured field
excavation works to monitor the wall deflection and ground instrumentation data thereby validating the adopted soil model, the
settlement. The monitoring was terminated when the ground floor FE analysis and the construction sequence.
slab was fully constructed. The ground retention systems elaborated in this paper
provide simple and economical methods of retaining very soft clay
7. FIELD PERFORMANCE OF RETENTION SYSTEM where excavation depths do not exceed 3.5m below the existing
ground level.
7.1 Plot A
9. REFERENCES
The performance of the temporary sheet pile wall was evaluated by
studying the recorded instrumentation data and a comparison has O’Rourke, T. D., (1993) “Base stability and ground movement
been made between the recorded wall deflection and the data prediction for excavations in soft clay”. Retaining structures,
extracted from the FEM analysis, as shown in Table 2 below. Thomas Telford, London, pp. 131-139.
Ukritchon B., Whittle, A. J., and Sloan, S. W., (2003) “Undrained
Table 2 Summary of Estimated & Recorded Wall Deflection for stability of braced excavation in clay”. ASCE J. Geotech.
Plot 1 Geoenviron. Eng. Vol. 129, No. 8, pp. 738-755.
Bjerrum, L. & Eide, O. (1956) “Stability of strutted excavations in
Location Estimated Deflection Actual clay”. Geotechnique, 6, pp. 32-47.
from FEM Deflection Chang, M. F. (2000) “Basal stability analysis of braced cuts in clay”.
(mm) (mm) J. Geotechnical and Geoenvironmental Engineering. ASCE,
Along Access Road 1 15 7 126(3), pp. 273-275.
CIRIA C580 (2003). Embedded retaining walls – guidance for
Along Access Road 2 15 15 Economic design. Graba A.R., Simpson, B., Powrie, W. &
Beadman,
As can be observed from Table 2, the estimated wall deflections are Goh, A. T. C. (1994) “Estimating basal-heave stability for braced
quite comparable to the actual wall deflections recorded on site. The excavations in soft clay”. J. Geotech. Engrg. Div., ASCE,
actual wall deflections were lower along Access Road 1 as the 120(GT8) pp. 1430-1436.
excavation for the temporary slope restraining the temporary sheet Karlsrud, K. (1986) “Performance monitoring of deep supported
pile wall was carried out in very short intervals of 6m as compared to excavations in soft clay”. Proc. 4th Int. Geot. Seminar, Field
the excavation along Access Road 2 which was carried out at 12m instrumentation and in-situ measurement’, NTU, Singapore,
intervals. pp. 187-202.
Lim, P. C. & Tan, T. S. (2003) “A Floating-Type braced excavation
7.2 Plot B in soft marine clay”. Proceedings of Underground Singapore
2003, Singapore, pp. 326-337.
Mana, A. I. & Clough, W. G. (1981) “Prediction of movements for
The performance of the temporary sheet pile wall and the permanent
braced cuts in clay”. J. Geotech. Engrg. Div., ASCE, 107(GT6)
CBP wall was evaluated by studying the recorded instrumentation
pp. 759-778.
data and a comparison has been made between the recorded wall Terzaghi, K. (1943). Theoretical soil mechanics. John Wiley, New
deflection and the data extracted from the FEM analysis, as shown in York.
Table 3 below.
464
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: In order to advantages the Rapid Impact Compaction (RIC) test method, a number of Unconfined Compression Tests (UCT)
were performed on the RIC compaction samples. The UCT sample size is 6 cm in diameter and 12 cm in height, which was drilled out from
the mould compaction that has been previously compacted with RIC Method. The compaction process is carried out in a standard modified
mould (15 cm in diameter and 25 cm in height) by varying number of 5 - 35 blows for each compacted sample with frequency 30 - 40 blows
per minute. All compaction processes are controlled by electro-mechanical new development system of repeated RIC method. The test
results represented that at the same energy level with larger number of blow variation of RIC results produce higher unconfined compressive
strength value in comparing to that of the energy level for larger masses with smaller number of blows. It is possible to make relative
compaction relation which is the result of the achievement of UCS value using Rapid Impact Compaction method to UCS value of proctor
method that fulfill a linear line equation. Based on these results, it could be created a simple relation of energy compaction in determining
unconfined compressive strength value with appropriate process and lower RIC energy consumption.
465
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
70
3.3 Test carried out
60
Percent Passing (%)
50 The soil in the compaction mould was obtained from the remoulded
40 sample, the required amount of soil was mixed with water and left
30 under a plastic for air drying to obtain a water content in accordance
20
with the standard proctor compaction of 15-35% and 0.046 kN of
10
soil was filled in the compaction mould without compacting by
0
controlling the loose density of 10.4 kN/m3. The compaction process
10,0000 1,0000 0,1000 0,0100
Diameter (mm)
0,0010 0,0001 with rapid impact compaction was performed using various energy
levels by applying the various impact blows.
Figure 1 Grain size distribution
In the compaction process, the mass of compactor was
periodically dropped over the anvil with ranging of frequency
3.2 Apparatus of Testing 30 - 40 blows per minute and adjustable falling height were set in
In this experimental test, the equipment of repeated load impact 15 cm, 20 cm and 25 cm for each sample tested, the impact process
used air-pneumatic piston. The main components of the equipment being carried out by varying blow ranging from 5 to 35 amount of
consisted of air compressor, piston double-acting air pneumatic- the blows for each mould compaction at specified height.
pressure, valve solenoid, panel controller and mass of compactor as The UCT sample size is 6 cm in diameter and 12 cm in height,
shown in Fig 2. Repeated load impact equipment has 150 kg vertical which was drilled out from the mould compaction that has been
lift capacity of mass compactor and falling height can be set 10 cm previously compacted. In this research, all compaction process were
to 30 cm, the time period of the impact compaction is being 2 - 3 controlled by a new developed electro-mechanic system of rapid
seconds per blow. For testing stability, the apparatus was attached impact compaction.
and positioned on a steel frame.
466
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
360
these three variables, in this case the relationship is expressed as a
comparison to the energy of standard proctor method which is 310
Table 3. Dry Density at Different Energy and Water Content Levels 110
Type of Hammer Falling Blows Cummulative Energy Water Dry 60 intercept energi line
Compaction Weight Height Energy Ratio to Content Density
(kg) (m) (kJ/m3) Proctor (%) (kN/m3)
10
15 10.52
0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0
20 10.55
RIC-01 70 0.15 5 117.051 0.20 25 10.66 Energy Ratio to Proctor Standar Energy
30 10.32
35 9.31 Figure 6 Linearization of UCS values from the relation of the
15 11.05
20 11.05
proctor compaction energy line to the RIC energy line
RIC-02 70 0.15 15 315.153 0.59 25 11.33
30 10.86
35 9.80 The energy line compaction Rapid Impact Compaction which
15 11.30 in this case according to Eq. (1), the other line is the energy line
20 11.53
RIC-03 70 9.15 25 585.256 0.98 25 11.74 connecting the standard energy level of the proctor and the modified
30 10.90 proctor which has previously been examined the relationship of its
35 10.00
15 11.60 unconfined compressive strength value which has been linearly and
20 12.09 satisfies the Eq. (2).
RIC-04 70 0.15 35 819.358 1.38 25 12.16
30 11.27
35 10.20 UCSRIC = 61.175(ER) + 82.477 (kPa) (1)
UCSProctor = 13.25(ER) + 189.99 (kPa) (2)
Based on Table 3 it can be known and stated that the tendency
of dry density relation is influenced by energy level of compaction From the Eq. (1) and Eq. (2) in as the Figure 6, there are three
effort and moisture content, at least this can be seen in Figure 4. The important parts are visible: first is for lower energy levels of RIC
characteristics of the relation are illustrated by the prediction line of energy than the proctor energy in this case showing that the value of
the optimum moisture content corresponding to the energy level of UCS obtained is also lower than the UCS of the proctor energy, then
the compaction effort given. The results show that the addition of the next section is on the energy level zone greater than the standard
energy tends to reduce the value of the optimum water content. energy of the proctor but still smaller than the proctor compaction
energy line in this section shows the level of attainment of UCS
values in general equal to the attainment of UCS at the proctor
compaction energy line to a point of intersection, while the last part
is the comparison of energy rapid impact compaction greater than
the proctor energy line, where in general the results of achievement
of its UCS value is also higher than the energy of compaction
proctor.
467
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
468
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Nowadays, costs for soft soil improvement account for a large proportion of the total construction expenses. According to
statistics of construction cost in Viet Nam, these costs were about 20% - 35% of the total expenses. As a result, advanced technologies have
been applied to reduce the costs; one of them was mass stabilization technology. This paper introduces a study on soft soil stabilization using
mass stabilization method experimented via a full-scale physical model, and comparison between settlement results computed by analytical
formulas and the physical model.
Keywords: Mass stabilization, Deep mixing, Soft soil improvement, Binder, Dike
embankment
mass stabilization
T
ALLU Systems uses compressed air to transfer dry binder from Figure 2 Diagram for settlement
the container into the soil throughout a tube that connected to power
2.1.2 Selection the appropriate thickness of reinforced mass (HT) in
mixer. Through the DAC device, it can control both the PF function
order to make the ground behaviour to be in the linear phase and
and adjust the volume of the binder pump to the ground. This helps
weak soil under the reinforced mass ensured load bearing capacity.
the soft soil improvement processing obtain the good quality and
Then the following conditions are described:
economical. To create the reinforcing mass, the blades cut soil
slowly from the top and mix with the binder. The Allu's technology
is especially effective in muddy areas because it does not need any P Rtc (1.2)
additional solutions. Thus, this technology has the advantage of fast in which:
construction, economic efficiency, material, and energy savings. In P = load acting on the ground under reinforced mass (kN);
addition, this method is easy to compatible with surrounding Rtc = standard strength of soil below reinforced mass (kN),
structures and environment (with no settlement differences). was determined by the following formula:
Besides, because of using soil in place, it makes transportation costs
reduced as well as no vibration, no noisy and non-polluted Rtc m A b B H Dc (1.3)
environment during construction. The disadvantage is that the initial in which:
investment cost of equipment is quite high. m = coefficient of working condition, m=1;
BT = width of reinforced mass (m);
2. CACULATION METHODS FOR EMBANKMENT AND HT = depth of reinforced mass (m);
MASS STABILIZATION c = cohesion of soft soil (kN/m2);
For creating a legal basis to apply mass stabilization technology in = the unit weight of the ground (=wđ if above
Vietnam, The Ministry of Agriculture and Rural Development groundwater level, and =đnđ if below groundwater
(MARD) has developed the based standard TCVN 04: 2014/VTC level);
based on the experience of applying deep-mixing technology in A, B, D = coefficients depend on the friction angle of the soil,
Vietnam (Phung et al., 2016) combined with several research results selected in the table 1.1.
469
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The load acting on the reinforcing mass was defined as follows: in which :
lE = the length of circle sliding in embankment (m);
P w H qxe B (1.4) lI = the length of circle sliding in reinforced mass (m);
in which: lS = the length of circle sliding in soft soil (m);
RR = radius of circle sliding (m);
w = the weight of the embankment (kN/m3);
WE = the weight of the embankment (kN);
HT = depth of reinforced mass (m);
XE = the horizontal arm of the embankment;
qxe = live-load distribution of the vehicle on the embankment
(kN/m2); E = shear strength of the embankment (kN);
B = width at top of embankment (m). I = shear strength of weak ground (kN);
S = shear strength of soft soil (kN).
Table 1.1 Value of A, B, D coefficients
At present, the facto of safety can be quickly estimated by using
Friction Coefficients numerical simulations in Geotech such as Geo-slope, Plaxis, etc.
angle (o) A B D 2.3 Calculation by the second limiting state
0 0 1,00 3,14
Total of settlement S were determined by settlement of reinforced
2 0,03 1,12 3,32 mass and the settlement of soft ground below the reinforced mass:
4 0,06 1,25 3,51
6 0,10 1,39 3,71 S h1 h2 (1.6)
8 0,14 1,55 3,93 in which:
10 0,18 1,73 4,17 S = total settlement of reinforced mass (m);
12 0,23 1,94 4,42 h1 = settlement of reinforcement mass (m);
14 0,29 2,17 4,69 h2 = settlement of soft ground under stabilization mass (m).
16 0,36 2,43 5,00
18 0,43 2,72 5,31 load q (kN/m2)
T
2.2.1 Stability examination of circular sliding surface Mass stabilization
2.2.1.1 Embankment material: There are 3 cases for the
embankment, including: (1) Embankment was bulk material with a
friction angle > 30o. For this case, according to British Standard BS
s
8003, it was not essential to examine the overall stability of the
slide; (2) Embankment was cohesive soil, weak in-place that
embank in multiplied stages or one stage using geotextile. In this
case, as embankment met the stability requirement or geotextile met Figure 4 Calculation diagram for reinforcement mass
the loading capacity, no overall stability is required. However, it is
difficult to determine when it was enough and when it is not enough. h1 was determined as as follows:
Therefore, it was necessary to monitor the overall stability of the
slope in this case. qH T
h1 (1.7)
ET
2.2.1.2 Reinforced mass: This mass was considered as a
homogeneous, isotropic material. This mass directly transfers the
load from the embankment to the ground and played the most wH Q
q (1.8)
important role in stabilizing of the embankment. BT
in which:
2.2.1.3 Depending on the case, the circular sliding can be cut only q = distributed load on the reinforced mass (kN/m2);
on the embankment or cut through the reinforced mass. Stable Q = concentrated load on the reinforced mass (kN);
condition k [k]. HT = depth of reinforced mass (m);
ET = elastic modulus of reinforced mass (kN/m2).
h2 was determined as follows:
CC ' q'
h2 H S 1g 0 (1.9)
1 e0 '0
qBT (1.10)
q'
H
BT S
2
Figure 3 Calculation schema of overall stability in which:
h2 = calculated settlement of ground (m);
2.2.1.4 The overall sliding stability coefficient FS was defined by: q’ = loading on soft soil below reinforced mass (kN/m2);
Hs = the thickness of soft soil below reinforced mass (m);
RR (1E E 1I I 1S S Cc = compression index of soft soil under pile tip;
Fs (1.5) e0 = porosity ratio of soft soil under pile tip;
X EWE 0’ = effective stress (kN/m2).
470
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3. EXPERIMENT TO DETERMINE SETTLEMENT BY Table 1.3 Shear resistance of the soil model
PHYSICAL MODEL
Shear resistance Su
3.1 Experimental purposes TT Depth (m) Remark
(kg/cm2)
The experiment purposes: (1) evaluate effectiveness of mass
stabilization method; (2) verify the results calculated by analytical 1 0,2 0,11 The results are
resolution. comparable to soil
2 0,4 0,09 tests at the sampling
site and in the
3.2 Site of Tests
3 0,6 0,10 laboratory
The experiments were built at Thanh Oai industry zone, near the Ba
La street, Ha Dong town, 10km from Ha Noi capital. This location
was selected because of its wide layout for carrying out the tests. Before carrying out the compression tests, the shear test was
carried out for monitoring and evaluation the shear resistance Su to
3.3 Test design compare between two experiments. If the Su was not suitable, the
moisture of soil could be adjusted. The location of shear test as
Digging soil to the depth of 40cm to build test models. The test shown in figure 5 and the test results were shown in the Table 1.3.
models were constructed by brick and cement M100. The thickness
of model wall was 35 cm and the height was 2.0 m. The container
dimensions were 1.0 x 8.0 x 2.0 m. Front of container was made by
glass thickness 15 mm and laid on steel frame system I20 by 5 mm
rubber bearings in thickness and 12 mm in width. Loading system
was placed inside the container. There were two I10 shaped steel
bars connected two sides of steel frame.
33
33
150
mass at different levels based on the experimental design. The
33 33 jack could be loaded 20 tons. A gauge with the accuracy of 0.2
103
103
bullet-pfoof glass 15mm cutting position A ton was attached on the jack to read data that produced by China
270 as shown in Figure 7.1.
- Stress sensor (sensor): To measure the stress at the bottom of
33
33
33
33
45
45
steel I 10 steel frame I 20 steel I 10 reinforced mass in each load level. The sensor has a diameter of
45
33 227 260 45 227 33 28 mm, a thickness of 6.5 mm, could be measured stress in range
Figure 5 Shear test location
471
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
of 0 ÷ 0.2 MPa, made by China, as shown in Figure 7.2. The 3.8 Experiment design
layout of sensors S1, S2, S3, S4 and S5 were shown in Figure 8.
- The Data sensor (Data Taker DT80) was used to record the data After installing experiment model as shown in Figure 10, carrying
from a pressure transmitter then transfer to the computer, made out the experiment with load level in the Table 3 for two cases:
in Australia, which can be connected and recorded - case 1: compression soil without mass stabilization;
simultaneously with five sensors, as shown in Figure 7.3. - case 2: compression soil with mass stabilization.
- Specific Gravity Test: To measure the vertical displacement of
the compression table and the ground surface at different load
levels. The range of specific gravity test was up to 50 mm and
have a precision of 0.01 mm, was made in China and was
arranged in both symmetric sides through the centre, far 10 cm
from the edge of the table.
- Steel compression table was used to transfer the concentrated
load from the hydraulic jack to the foundation. This compression
table was made of steel, 1cm in thickness, 1m x 1m in
dimension, was reinforced at both side by I10 shaped I10 steel
bars to increase the hardness, as shown in Figure 7.4.
- Laptop: storing and analysing data from the reader.
- Digital Cameras: A digital camera with a projection light and a
cover to capture the displacement of points in the background.
S1 S5,S3 S4
Table 4 Settlement measurement of two experimental cases
0,25 0,25
Single-point Average settlement of Average settlement
TT
0,5 0,5 load (T) foundation MS (mm) of ground (mm)
472
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The settlement results were graphed as follows: h1 was calculated according to Equation 1.7:
qH T
h1 0.36 (mm)
ET
in which:
q = distributed load on mass stabilization, q= 4,2 (T/m2);
HT = thickness of mass stabilization, HT=0,3 (m);
ET = elastic modulus of mass stabilization, was determined by
taking samples and conducting tests in the laboratory, ET=
3500(T/m2).
In case of ground was fully consolidated, h2 was calculated by
Equation 1.9 :
CC ' q'
h2 H S 1g 0 144,08 (mm)
Figure 11 Stress-displacement relations of two experimental cases: 1 e0 0'
in which:
It is can be seen that: for natural ground with the intensity limit
q’ = loading on soft soil below mass stabilization, q’=2,27
of 3,6 (T/m2), the settlement was 163,25 mm, and for ground with
(T/m2);
30 cm thickness of mass stabilization (MS) and intensity limit of 4,2 Hs = thickness of soft soil below mass stabilization, Hs =1,7(m);
(T/m2), the settlement was 133,53 mm. Thus, with the mass Cc = compression index of soft soil under pile tip, Cc=0,29;
stabilization (MS), the bearing capacity of ground was increased by
e0 = porosity ratio of soft soil under pile tip, e0= 1,52;
15% and the settlement was decreased by 46.64% in compared with
0’ = effective stress, 0’ =0,51(T/m2).
those of natural ground.
The results showed that the error between the analytical formula
and the experiment was 7.56%. However, analytical formula was
applied for the case of full consolidation. In fact, the experiment
maintained the compression level at 30 minutes for each load level.
Therefore, the soft soil was under full consolidation. Hence, the
error between test results and analytical equation could be less than
7.56%.
4. CONCLUSIONS
Mass stabilization technique is an effective method for soft soil
improvement. Especially for works such as dikes (embankments),
roads in soft soil conditions and difficulties in finding construction
materials. Because of several advantages such as construction
without additional equipment, using in-place soft soil, and fast
construction speed. Mass stabilization technology is a potential
Figure 12 Displacement of ground without reinforcement that
method in Vietnam in general and the Mekong Delta in particular.
marked with sawdust
The calculation method was relatively simple, can use analytical
formulas as presented to design. In addition, it is possible to use
geotechnical software to calculate.
The results showed that the error between the analytical formula
and the experiment was less than 7.56%. This value was acceptable.
5. ACKNOWLEDGEMENT
The work presented in this article was supported by the Viet Nam
Ministry of Agriculture and Rural Development and following two
studies:
1. Research on solutions of soft soil treatment technology by
mixing in-place soil with inorganic binder for construction of
dikes in the Mekong River Delta.
2. Research on dike crack phenomenon and solutions
improvement to protect dike combined rural road.
473
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Japan standards. Experimental method for field bench press for Allu and Rambol. Mass stabilisation manual. Allu Finland Company.
ground JSF: T25-80T. EuroSoilStab. Design Guide Soft Soil Stabilisation. CT97-
Vietnam standard TCVN 9354-2012. Construction soil - Method for 0351 Project No: BE 96-3177.
determining deformation modulus at field by flat plate. Le V. T., Yonglai Z., ShuXin D., Phung V. A., Ha T. L. (2015): “An
Phung V. A. et al., (2015) "Research on solutions for soft soil Numerical Analysis of the Influence of Replacement Area
treatment by soil mixing equipment with inorganic binder for Ratio in Foundation Reinforced with Soil Cement Columns”.
hydraulic construction". Report on the studies of the Ministry The Electronic Journal of Geotechnical Engineering ISSN
of Agriculture and Rural Development. 1809-3032. Vol.20.8. pp 3821-3828. 5.2015 (EI)
Phung V. A. (2014) “Research on solutions of soft soil treatment Phung Vinh An, Vu Ba Thao, Nguyen Quoc Dung (2016) “An
technology by mixing in-place soil with inorganic binder for Analytical Approach for Determining the Bearing Capacity of
construction of dykes in the Mekong River Delta”. Report on Soil Cement Column Using Jet Grouting Technology”.
Water Resources Science and Technology for the Mekong 19th Southeast Asian Geotechnical Conference & 2nd
Delta and the South East Provinces. AGSSEA Conference (19SEAGC & 2AGSSEA) Kuala
Nguyen Q. D., Phung V. A., Vu N. B. (2014) “Research on Ground Lumpur 31 May – 3 June 2016. pp 937-941.
Improvement in Ca Mau with cement and cement additives
combined”. Journal of Water Resources Science and
Technology 2009-2014, ISBN 978-604-67-0392-1.
474
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Study for Structural Performance of Steel Sheet Piles Used for Retaining Wall
P.S. Eka1, M. Nobuyuki2, T. Tomoya3
1
PT Nippon Steel and Sumitomo Metal Indonesia, Indonesia
2
Nippon Steel & Sumitomo Metal Southeast Asia Pte.Ltd., Singapore
3 Nippon Steel & Sumitomo Metal Vietnam Co., Ltd., Ho Chi Minh City, Vietnam
E-mail: eka.2tb.susanto@id.nssmc.com1; matsui.c4d.nobuyuki@sg.nssmc.com2; tominaga.5xg.tomoya@vn.nssmc.com3
ABSTRACT: This paper describes structural performance of steel sheet pile walls based on laboratory and full-scaled field test. The steel
sheet piles are widely used as a permanent and temporary retaining wall. In order to realize reliable structure of the walls, the actual
performances should be evaluated properly in design especially for in case height of walls are deeper and close proximately to existing
structures. From this point of view, a reduction of sectional stiffness of U-type due to insufficient transmission of shear force in interlocks are
regulated in some of design standards. Depending on soil condition and restraint against slippage of interlocks, U-type should be considered
reduction factor for section modulus and moment of inertia in case of composite section. On the other hand, Hat-type sheet piles fully
perform the stiffness because interlocks are located on outermost surface where the shear force is negligible. The test results show the actual
stiffness performed according to sheet pile shapes.
Keywords: Steel sheet pile, retaining wall, stiffness
475
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
H-section
Hat-type steel
Figure 3 Comparison between Hat and U type sheet pile
Fillet 900 mm
weld
Table 1 Sectional Properties of HAT Type Sheet Pile
Figure 4 Outline of the Hat+H pile
Type Dimension Property per meter of wall
Effective Effective Thick- Sectional Moment Section Unit
width height ness area of Inertia Modulus mass 2.2 Installation methods
2 4 3
mm mm mm cm cm cm kg/m
NS-SP-10H 900 230 10.8 122.2 10,500 902 96
Hat type can be installed by ordinary piling methods such as
vibratory driving method and the press-in method. The vibratory
NS-SP-25H 900 300 13.2 160.4 24,400 1,610 126
diving method is the major method to reduce ground friction by
NS-SP-45H 900 368 15.0 207.8 45,000 2,450 163 vibration as shown in Figure 5 This method can shorten the driving
NS-SP-50H 900 370 17.0 236.3 51,100 2,760 186 time and drive SSP into hard ground. This method is useful not only
driving but also pulling out the pile. The press-in method shown is
Hat type is applicable for both temporary and permanent Figure 6 is using hydraulic mechanism by grasping the middle place
retaining walls, and able to contribute to enhance structural safety of sheet pile while taking reaction by holding driven piles. Driving
and productivity in view of aspects below. machine is compact. It is applicable for the narrow and low
<Efficient pile-driving work> clearance construction site. This method provide low noise and low
Since Hat-type is more rigid than U type in terms of 1 pile, pile vibration of construction. For hard soil layer, additional method
deformation while the pile is driven into the ground is effectively such as water jet cutter or augering is applicable.
restrained, and hence even a longer pile can be driven in efficiently
<Higher structural reliability>
Since the pile joints are at the outermost part of the wall
structure, the neutral axis of each of the piles coincides with that of
the wall structure formed. Therefore, Hat type is able to display its
structural performance under various construction conditions.
<Applicable to deep excavation and high retaining wall >
Hat type is able to provide higher section modulus and moment
of inertia by combining with H shape by simple fillet welding as
shown in Figure 4. The fabrication is easily done at the site. Thanks
to high stiffness, it is possible to simplify additional supports such as
strut and tie rod, which leads to enhance construction productivity.
In addition, rapid installation is possible by using percussion
hammer or vibratory hammer, compared to existing piles i.e
concrete pile or steel pipe.
HAT+H is composite section, so moment of inertia is calculated
as follows.
I Is As ys2 IH AH yH2
I I /w (1)
I = moment of inertia of the Hat-type and H-shape per pile Figure 5 Vibro hammer (In Indonesia)
Is = moment of inertia of the Hat-type per pile
As = section area of the H-shape section
IH = moment of inertia of the H-shape section
AH = section area of the H-shape section
yS = distance from the neutral axis of the Hat-type and H-shape to
the neutral axis of the Hat-type per pile
yH = distance from the neutral axis of the Hat-type and H-shape to
the neutral axis of the Hat-shape
I’ = moment of inertia of the Hat-type and H-shape per pile wall
width
w = effective width of the Hat+H pile (900 mm)
476
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3 STRUCTURAL PERFORMANCE OF SSP The Flexural tests for Hat+H were conducted in order to verify
3.1 Influence due to joint slippage due to shear force the sectional performance under bending force as well. The sheet
pile and H-shape were combined by fillet welding with a leg length
Steel sheet pile must be integrated with each other with its joint
of 6 mm. The flexural moment was generated by having vertical
interlocking for showing stiffness performance. When SSP resist the
soil and water pressure, bending moment and shear force occur. forces acting on the flange portion of the H-shape as shown in
These force cause joint slippage, so that SSP cannot perform actual Figures 9. The vertical load acted on the upper flange of the H-
nominal section properties i.e. section modulus and moment of shape. Figure 10 shows the relationship between the bending
inertia. This situation, so-called “lack of joint integrity” depends on moment and the curvature obtained by the loading test. The
the shape and joint location of SSP wall. The percentage of curvature, was calculated by the measured displacements at the
reduction is designated as the reduction factor. For that reason,
loading point and center of the specimen. It shows that the bending
consideration of reduction factor in the SSP design is very important
to keep the structure safety and stability. In this paper, theory of moment of the Hat+H pile is almost the same as theoretical curve up
reduction factor between Hat and U type is proved through to the yield bending moment of the composite section. This large
structural experiment and numerical analysis and its reduction factor section did not cause local buckling before yielding. In addition, the
is confirmed. strain distribution in the section along the vertical direction kept the
As conventional SSP, when U type is subjected to bend due to plane, and it was not discontinuous like a built-up section as shown
soil pressure or other horizontal load, a large bending shear force in Figure 11. In addition, the strain curve is also equivalent to that of
occurs in the its interlocking joints because its interlocks are located
the calculated value of a composite wall. These results shows Hat+H
at the center of the wall, defined as neutral axis position. In these
cases, the shear force does not transmit sufficiently between the can be treated as a composite section by simple fillet welding for the
adjacent piles because the interlocking joints slip from each other, Hat-type and H-shape.
and its possessed sectional properties, the moment of inertia and
section modulus of the wall, are reduced by the reduction factor.
Hat as relatively new steel sheet pile has many advantage.
Beside of its 900mm width perform less number of piling than U
type for rapid construction, Hat also has been proved able to
perform 100% joint integrity as discussed in Eurocode. It means that
the actual nominal value of stiffness i.e section modulus and
moment inertia can be inputted without considering the reduction
factor due to the shape factor and site condition. On the other hand,
Hat-type sheet piles does not slip because interlock located on the
outermost side of the wall. So, Hat-type wall is no need to consider
reduction factor. Figure 7 illustrates interlocking joint behaviours
both U-type sheet piles and Hat-type sheet piles.
Figure 8 Bending resistance of HAT
477
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
eAM
T
2I 1 y 0 eA
(4)
M 2Ty 0
w y1 2T (5)
I1 A
y1 = the distance between the edge of the web side and the neutral t
M
y0 e (9)
I 1 y 0 eA
axis of a single sheet pile,
y0 = the distance between the neutral axis of a single sheet pile and
Thus, the section modulus zt at the interlock is:
the edge of the interlock side,
e = the distance between the neutral axis of a single sheet pile and
I 1 y 0 eA
the neutral axis of the wall. zt (10)
y0 e
The actual sectional strain distribution depends on the
transmission of the shear force T in the interlock. In the full share In the full shear transfer condition, the neutral axis of the wall
transfer condition, the neutral axis of the wall (e=y0)(Figure 10.a). located the center of the wall (e = y0). Then, the stress w0, the
On the other hand, in the zero shear transfer condition, the neutral section modulus zw0 at the web and the moment of inertia is:
axis of the wall corresponded to that of a single sheet pile
(e=0)(Figure 10.b). When the partial shear force transfers in the
interlock, the neutral axis of the wall located within the region w0
M
y1 y 0 (11)
I 1 Ay 0
2
bounded by the full and zero shear transfer conditions (0<e<y0)
(Figure 10.c).
The bending moment M is caused by the soil pressure acts on I 1 Ay 0
2
the wall, the stress σy at the distance y from the neutral axis of the z w0 (12)
single sheet pile is y1 y 0
y
M 2Ty 0
y
2T (2) I 0 I 1 Ay 0
2
(13)
I1 A
Thus, the factor accounting for the possible reduction of the
where I1 is the moment of inertia of a single sheet pile, A is the moment of inertia due to lack of shear force transmission in the
sectional area of a single sheet pile. interlocks D is defined as:
Since the stress y is zero at the center of the wall (y = e):
I I y 0 eA
D 1
M 2Ty 0
y
2T I 0 I 1 Ay 0 2
(14)
I1 A (3)
478
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Regarding the factor accounting for the possible reduction of the 4. STRUTURAL TEST
section modulus due to lack of shear force transmission in the
4.1 Test purpose
interlocks B, there are two equations, i.e., w at the web and t at
the interlock: The purpose of this test is to verify the lack of interlock integrity of
U-type sheet pile by comparing its deformation in actual situation.
Zw y y1 I 1 y 0 eA y 0 y1
w 0 D (14)
Z w0 y1 e I 1 Ay 0 2 y1 e 4.2 Test method
3.3 Reduction factors stipulated in design codes Figure 15 Excavation inside the cofferdam
In the table 3, it is shown that the reduction factors stipulated in 4.3 SSP properties
design codes in Japan and Eurocode. In Japan, D as the reduction
factor for the moment of inertia is set as 0.45, with no welding or In the Table 3 it is shown about SSP type, length, and sectional
crimped at each sheet pile. In Eurocode, it is set as 0.30 to 0.40, with properties of this test. Both IV and 25H has same 12 m length with
cantilever, uncrimped, and favorable or unfavorable soil conditions. embedded 11.1 m into the ground. Inside excavation depth is 5.8m.
Possessed moment of inertia of IV is approx. 1.58 times higher than
its 25H.
Table 4 SSP properties
Sheet pile Embedded Excavatio Moment
Type length length n depth of inertia
(m) (m) (m) (cm4/m)
U-type NS-SP-IV 12.0 11.1 5.8 38,600
Hat-Type NS-SP-25H 12.0 11.1 5.8 24,400
479
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5.2 Results
As the results, in the Figure 19 it is shown about the deformation
view in the -5.8 m excavation depth from G.L. The deformation of
IV is larger than 25H. In the Figure 20 it is compared about
displacement from PLAXIS with the actual test data. In -5.8m from
G.L, the displacement simulated by PLAXIS is close to the test data
rested 16 hours. The moment of inertia of IV is adjusted 0.4 times
from its nominal value which means that the reduction factor for
moment inertia is 0.4.
Step1 Step2
480
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
6. CONCLUSION
Through the excavation test, the stiffness between Hat-type sheet
pile and U-type sheet pile is compared in actual situation. The
horizontal displacement of the sheet piles wall in increasing in
parallel with the incremental of excavation depth. This excavation
test verified that 25H (HAT) show higher stiffness than U type IV in
the same actual situation.
It is proved that U type IV sheet pile experience lack of joint
integrity because displacement of the U type IV wall is larger than
25H (HAT). This lack of joint integrity condition is stated in the
reduction factor for the moment inertia which is estimated 0.4 by
FEM analysis (PLAXIS); this value is approximately stay in range
of several standards value. For that reason, it is necessary to
consider the reduction factor for U type sheet pile to ensure the
Step3 Step4 structural stability and safety in steel sheet pile design stage. In other
Figure 19 Excavation protocol hand, it is not necessary to consider reduction factor for stiffness in
case of HAT type sheet pile design.
7. REFERENCES
Harata, N. (2008). Development of Hat-Type Sheet Pile 900,edited
by Tatsuta M, Kurosawa T, Nishiumi K, Taenaka
S,Wakatsuki T, Miura Y, and Eda K. Nippon Steel Technical
Report, Japan.
Otsushi K. et al. (2016) “Line-up Expansion of Hat-shaped Steel
Sheet Pile (NS-SP-45H, 50H)”, Nippon Steel & Sumitomo
Figure 20 Comparison the PLAXIS result and Test result Metal Technical Report, No.113, pp57-63.
Matsui et al. (2015) “Novel Compound Steel Sheet Pile for Earth
Retaining Works” The IES Journal Part A : Civil & Structural
Engineering
Nippon Steel & Sumitomo Metal Corporation. (2016). Product
brochure, STEEL SHEET PILES.
Japan Road Association. (1999) Design guideline of temporary
structure.
Eurocode3. (2007) Design of steel structures, Part5, Pilling
Susanto,Eka et al. (2017). “Study on Seismic Reinforcement
Method for Coastal Dyke Using Hat Steel Sheet Pile”.
HATTI (21th Annual National Conference on Geotechnical
Engineering Jakarta, 7-8 November 2017). Indonesia
Vietnam National Standard. (2013): TCVN 9860:2013 Steel sheet
Figure 21 Deformation view of the PLAXIS result pile structures in transportation engineering – Design
(Left: 25H, Right: IV) requirement.
481
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: With the booming of urban development, construction of new infrastructures faces challenges on rapidly disappearing space.
More and more underground constructions are inevitably conducted adjacent to existing buildings, and the associated safety concerns
become an important issue. This paper presents a particular case where a deep excavation is conducted in the proximity of operating public
transportation facilities. The underground construction is a joint-development for major substation of the Taipei MRT Circular Line, intercity
bus terminal, and commercial purposes. The close-by operating facilities include tunnels for Taipei MRT Blue Line, Taiwan Railway (TRA),
and Taiwan High Speed Rail (THSR) and none of them can be suspended during the excavation. The paper first briefs the environment
constraints. It then discusses the design considerations of protection measures against excavation-induced risks to the operation and safety of
the public transportation, including demolition of existing underground structure, trench protection of diaphragm wall (i.e., D-wall), strutting
and groundwater control, and monitoring. Conclusions are provided at the end of the paper
Keywords: Deep excavation, shield tunnels, underground structure, pumping, protection measures.
1. INTRODUCTION could thus be restricted within a relatively small region that would
make the distribution and change of groundwater easy to control.
The project is located at the north side of Banqiao Train Station in This paper first introduces the background the project, including the
New Taipei City. The site was an intercity bus terminal and is adjacent existing railway and MRT tunnels, and geological and
planned to merge with the adjacent lands to construct a joint- groundwater condition. It then illustrates the design consideration of
development building for the original bus terminal, main substation the protection measures for excavation. Preliminary monitoring
of the Taipei MRT Circular Line and commercial purposes. The results are presented followed by the concluding remarks.
scope of the project includes the construction of the building from
the basement to the fifth floor, above which will be constructed by
2. BACKGROUND OF THE PROJECT
the future developer. Since there are quite a few important
infrastructures around the site, least construction effect on these 2.1 Site Layout
facilities will be the first priority. Furthermore, the bus terminal
shall be relocated to not only remain its functions but also allow the To meet the requirement of the structure layout for the joint
project to proceed. development and the substation, the depth of the excavation is
The deep excavation of the building is a critical and key issue to designed to be approximately 33.3 m and 29.8 m, respectively.
the success of the project. The surrounding facilities include the Figure 1 shows that the site is a rectangular shape with about 157 m
tunnels of the MRT Blue Line, joint-tunnels of Taiwan Railway in the east-west direction, 27 m in the north-south, and the total area
(TRA) and Taiwan High Speed Rail (THSR), and three underpasses of about 3,750 m2. The basement of the building is a single wall
connecting the bus terminal and Banqiao Train Station. The system constructed using the bottom-up method. The D-wall is
excavation therefore is required not only to keep the facilities in adopted for both retaining and permanent structure purposes with
operation but also to remain the bolts on the top slab of the joint- 1.5 m thick and 52 m deep into the Gravels. Besides cross-wall and
tunnels that were pre-installed for the piers of the elevated Circular buttress, the inner strutting is adopted to reduce wall deformation
Line in a tolerant position. Further, since part of the piers of the during excavation. A ten- and nine-level strutting system is planned
elevated station that situates right next to the joint-development for the joint-development building and substation, respectively.
building is overlapped with the existing underpasses it is of 2.2 Overview of Adjacent Railway and MRT Tunnels
particular concern to the construction interaction between the
excavation and the foundation. In response to such complicated As depicted by Figure 1, on the west side of the site is the operating
construction constraints, the retaining structure adopted for the Blue Line of Taipei MRT; the south side is Banqiao Train Station
project include diaphragm wall (i.e., D-wall), cross-wall, and which is connected to the site through three underpasses; the east
buttress. In addition, deep guide wall and ground improvement were side is the operating tunnels for TRA and THSR; the north side is a
conducted at some regions to reduce ground disturbance by D-wall temporary bus terminal substituted for the one situated on the site.
excavation. Among the surrounding facilities, the tunnels for MRT, TRA, and
Since the joint-development building is constructed on the THSR are of most importance. There is literally no substitution of
existing bus terminal, conflicts of the new and existing structure transportation for the tunnels and protecting them from excavation-
would be inevitable. The D-wall of the building is especially induced effects is the only feasible way.
designed for both retaining and permanent use and thus its quality The up- and down-track of the Blue Line are constructed by two
shall be strictly controlled during the removal process of conflicted single shield tunnels that operate daily for as long as 18 hours and
structure and backfill. serve the trains as short as in every 3 minutes during the rush hours.
In addition, the building is seated on the Songshan Formation, a The outer diameter of the tunnels is 6.1 m with clearance of about 7
sand-clay interbedded alluvium that is underlain by the highly m and overburden of about 18 m within the region of the site. The
permeable Jingmei Gravels with abundant groundwater. In view that tunnels are aligned roughly in the south-north direction passing by
the project only covers the building up to the fifth floor, the control the site with the closest distance of about 6.4m on the southwest
of hydraulic pressure on the bottom of the excavation becomes a corner.
particular concern for both excavation stage and the complete stage On the other hand, the joint-tunnels for TRA and THSR are box
of the project before the developer finishes the upper stories of the structure with one and two story below ground surface, respectively.
building. Based on the deep excavation experience accumulated on Both tunnels adopted D-wall as retaining structure with 0.8-m-thick,
the Taipei Basin (e.g., Young et al., 1996; Huang et al., 1997; Chen 30-m-deep for TRA and 1-m-thick, 40-m-deep for THSR. The
et al., 1999; Moh, 2004), it was suggested except for base grouting tunnels pass by the site with the closest distance of about 6 m on the
to introduce cross-wall to panel the excavation site and dewatering southeast corner. There are 9 places on the top slab of the tunnels
482
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
THSRConference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Tunnel Banqiao Train Station
TRA Foundation of MRT Elevated Station
Tunnel
Underpass
Existing Underground Structure
24m
27m
Joint-Development Building
& Major Substation
157m
Banqiao Temporary
Intercity Bus Terminal
and concourse level for the Banqiao Train Station that were distribution with depths. It is especially noted that there is no sign of
embedded bolts for the future piers of the elevated Circular Line. hydraulic pressure in the Gravels that is below the hydro-static state
(e.g., Chao et al., 2015; Hu et al., 1996). The hydraulic pressure has
3. GROUND AND GROUNDWATER CONDITION been recovered in the Gravels after deep well pumping was banned
for domestic and industrial water use in late 1960s.
According to the results of geological investigations, the depth of
the Gravels on the site is about 50 m below ground surface. Except
4. PROTECTION MEASURES FOR CONSTRUCTION
for the fills, the Gravels are overlain by 5 alternating sand and clay
layers. It is identified as the sublayer I to V in the Songshan In the deep excavation where D-wall is adopted as the retaining
Formation but the sublayer VI is not clearly shown on the site (Hu et structure, the induced ground deformation generally originates from
al., 1996). The soil profile for the site is generally divided into 7 trench excavation of D-wall, base excavation and strutting, and
layers that from top to the bottom are: dewatering. The project is particularly challenging in view of the
(1) Fills – distributed from ground surface to 4.0 m below with existing underground structure and foundation for bus terminal and
an average thickness of about 4.3 m. underpasses on the site that shall be demolished during excavation.
(2) Silty Sand – distributed from 4.0 m to 11.0 m below ground In response to such environmental constraints, any possible
surface with about 7.3 m in average. construction risks were evaluated and the corresponding protection
(3) Silty Clay – distributed from 11.0 m to 20.0 m with a measures were planned to meet the regulated requirements. They are
significantly varying thickness and about 6.7 m in average. illustrated in the following sections and summarized in Table 2.
(4) Silty Sand – distributed from 20.0 m to 28.0 m in depths
with an average thickness of about 6.7 m. Table 2 Critical Issues and Strategies for Excavation
(5) Silty Clay – distributed from 28.0 to 36.0 m with about 6.8
m thick in average. (a) Removal of existing underground structure
(6) Silty Sand – distributed from 36.0 to 50.0 m with about 11.3 Existing Piled Wall Removal of existing piled wall
m thick in average. and backfill with grouting
(7) Gravels – distributed from 50.0 m deep and below at least materials.
till the bottom of the borehole at the greatest depth of 59.1 Partly destruction of existing
m with the largest grain size of about 10 cm. Major
underground structure to
Substation build deep guide wall for D-
A simplified soil profile and its engineering properties adopted wall excavation
for excavation analyses are presented in Table 1. Site excavation and demolish
Existing Underground the remaining underground
Structure
Table 1 Simplified Soil Profile and its Engineering Properties structure
Soil
SPT- (b) Retaining system
Layer Depth Classification t su '
N D-Wall
USCS
1 4.0 SF 8 19.7 - 28
2 11.0 SM 9 20.0 - 30 D-wall
3 20.0 CL 9 19.4 62 32 Cross-wall
4 28.0 SM 15 19.7 - 31.5 Cross-Wall Buttress
Buttress
5 36.0 CL 13 19.5 115 33 Steel strutting
6 50.0 SM 32 20.0 - 32.5
7 > 50.0 GM 50 22.0 - 38
Note: Unit for depth, unit weight, undrained shear strength, and effective (c) Trench protection and construction monitoring during D-wall
friction angle is m, kN/m3, kN/m2, and degree, respectively. excavation
In addition, based on the recordings obtained from the
observation wells and piezometers during the investigation period
between May and August of year 2009, the groundwater level is
located at about 5 m below ground surface with a hydro-static
483
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
MRT Blue Line piled wall within the site and cavity was backfilled with grouting
Shield Tunnels
Depth of micro-piles: 30m materials. For those in conflict with the new D-wall, the separation
Low-pressure grouting: panels and retaining structure were constructed, respectively, within
depth 4 m ~ 20 m and outside the existing basement as the guide wall for the D-wall.
Automatic monitoring: total They were demolished after the D-wall was complete.
Major Substation
station, reflector prism, beam
sensor, and convergence 4.2 Trench Protection During D-Wall Excavation
point for the shield tunnels of
Low-Pressure The project is in close proximity of the tunnels for MRT, TRA, and
Grouting for MRT Blue Line
Trench Protection Micro-Piles THSR with minimum clearance of about 6 m. Since D-wall was
used for the excavation of the TRA/THSR box tunnels, it is
estimated that the effect of ground deformation induced by trench
(d) Groundwater control
excavation on the site would be relatively slight on the box tunnels.
Pumping For the MRT’s shield tunnels that exhibit smaller volume and have
D-Wall
Strut
no retaining structure as a protection, the effect would be significant.
Base grouting (existing Figure 2 shows that a series of low-pressure grouting with depths
Excavation Base basement floors must be varying from 4.0 m to 20.0 m was therefore conducted accompanied
penetrated in partial area) by 30-m-deep, steel-tube micro-piles to strengthen the soils outside
Pumping within D-wall the trench. The ground deformation could be reduced as well as the
Buttress Cross-Wall
Section dewatering in effect on the surrounding soils.
Base Grouting excavation site
4.3 Strutting and Groundwater Control During Excavation
Pumping Well Within D-Wall
Strutting and groundwater control are the two major issues for the
stability of excavation. As depicted by Figure 2, except for steel
4.1 Removal of Existing Underground Structure struts used in the retaining system the inner buttress and cross-wall
If the removal of existing underground structure is not fully are additionally introduced to the site to reduce the excavation-
complete, the remaining part would result in potential risks in induced wall deflection. The inner buttress is a pure concrete mass
excavation. It would create a passage for groundwater leakage with depths of 50 m on the side close to the MRT shield tunnels and
11.5m 14.5m
Joint-Development Major
Building Substation
CL/ML
SM
Shield Tunnels for
CL Blue Line
Cross-Wall
4m D-Wall (L=52m)
SM
3m
5m 3m
2m
GM
Base Grouting
(b) Section
Figure 2 Schematic layout of protection measures for excavation
resulting in additional ground deformation around the site. 40 m for the rest of the site. The cross-wall is 50 m deep with pure
Therefore, the casing method was adopted to remove the existing and reinforced concrete for depths above and below excavation
484
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
levels, respectively. The cross-wall is used not only as part of mass in the Gravels. To effectively control the dewatering efficiency,
retaining system but also as panels for dewatering in the project. the site was panelled by six pumping area through cross-walls.
Since the groundwater level is about 5 m deep and the site is
excavated up to 33.3 m below ground surface, it was estimated that 4.4 Construction Monitoring
piping and uplift would be the issues against the stability of
excavation. Furthermore, the pumping output would increase as the To keep the site and adjacent structure in a safe condition over the
excavation depth closes to the underlain aquifer (i.e., the Gravels). It excavation period, a construction monitoring plan was implemented
is thus particularly crucial to maintain the stability on the excavation and focused on the significant factors. During the excavation, the
base. The base grouting was adopted to block the recharge of monitoring targets, instruments, and the associated frequency and
groundwater from the Gravels to the bottom of the excavation. The management levels involved are summarized in Table 3.
permeability of the improved soils is required to be below 1×10-5 Figure 3 shows the monitoring results from the inclinometer
cm/sec. Pumping wells were installed on the site to lower the installed in the D-wall close by the shield tunnels and in between the
groundwater table below the excavation level. Besides, additional buttresses. Due to the support of ground improvement, micro-piles,
pumping wells were installed within the D-wall and penetrated and barrettes, the resulting wall deflection was significantly reduced
down to 90 m deep with inlets opened at depths greater than 60 m to with a maximum of about 16.6 mm in comparison of the
control the hydraulic pressure on the bottom of the improved soil corresponding calculated free-field value of about 35 mm. The
485
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Deflection (mm)
0 10 20 30 40 50
0
EXC. 1 (2015.10.06)
5 EXC. 2 (2015.10.20)
EXC. 3 (2015.11.03)
10 EXC. 4 (2015.11.13)
EXC. 5 (2015.11.20)
15
EXC. 6 (2015.12.22)
EXC. 7 (2015.12.25)
20
EXC. 8 (2016.01.05)
EXC. 9 (2016.01.22)
25
dhmax =
Depth (m)
FINAL EXC. (2016.01.29)
35mm 30 d max=16.6mm
35
40
45
50
55
Figure 3 Comparison of measured and calculated wall deflection closest to MRT tunnels
effect on the shield tunnels was therefore greatly reduced as well.
According to the readings of instruments mounted on the tunnels,
the deformation and differential deflection were remained within the 7. REFERENCES
management levels over the excavation period. Chao, H.C., Yang, G.R., Su, T.C., and Chen C.H., (2015) "Joing
Dewatering for MRT Deep Excavations – A Case Study".
5. CONCLUSION Sino-Geotechnics, 146, pp65-74.
In the early stage of design for the project, the deep excavation Chen, Y.W., Duann, S.W., Hwang, R.N, and Kong, S.K., (1999)
experience on the Taipei Basin was collected and studied. The "Influence of Excavation on Adjacent Piles", Proceedings of
protection measures, such as cross-wall, inner barrette, and base the 8th Conference on Current Researches in Geotechnical
grouting, were proposed based on the past experience, site Engineering, Pingtung, Taiwan, pp1425-1432.
characteristics, and environmental constraints. Up to the first half of Hu, I.C., Chin, C.T., and Liu, C.J., (1996) "Review of the
year 2018, the construction of basement is almost complete and the Geotechnical Characteristics of the Soil Deposits in Taipei".
effects of D-wall and base excavation were fairly controlled with no Sino-Geotechnics, 54, pp5-14.
interference to the tunnel operation. The relevant monitoring results Moh, Z.C., (2004) "Protection System to the Neighbouring
are still under summary and study. It is hoped to be shared in the Buildings in Taipei Metro", Proceedings of the Symposium
near future. on City and Underground Space and Underground
Construction, Beijing, China.
Wong, L.W., and Chin, C.T., (1997) "Groundwater Control for
6. ACKNOWLEDGEMENTS
Deep Excavation". Sino-Geotechnics, 63, pp15-22.
Permission for publishing this paper from Moh and Associates, Inc. Yang, G.R., Wong, L.W., and Wang, S.N., (1995) "Groundwater
and Department of Taipei Rapid Transit Systems of Taipei City Problems for Deep Excavations in Gravel Deposits in Taipei
Government is gratefully appreciated. Special thanks to Dr. R.N. Basin", Proceedings of the International Symposium on
Hwang and Mr. C.R. Chou who helped to prepare the manuscript. Underground Construction in Gravel Formations, Taipei,
Taiwan, pp4-21~4-28.
486
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: The 14 November 2016 Mw 7.8 Kaik�ura Earthquake triggered a number of landslips along the Kaik�ura coastline. Ohau
Point is located approximately 26 km north of Kaik� ura and was the site of massive landslide which destroyed the road and rail transport
corridors. This paper describes the bored pile retaining wall solutions that were developed as part of the North Canterbury Transport
Infrastructure Recovery (NCTIR) earthquake recovery program of works for the New Zealand Transport Agency (NZTA) and KiwiRail.
The subject site is named as NCTIR Site 6 and the subject bored pile retaining walls are named RTW 3.
Keywords: Bored pile retaining wall, Kaikoura earthquake and bored pile construction
487
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 4 below provides an oblique drone photograph of the Maximum Wall Deflection
Site 6- RTW 3 location, which was taken during late 2017. The new 100mm
Site 6 seawall (under construction), old rail and road alignments are (ULS Seismic)
clearly visible in Photograph 4.
488
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
4. LOADING CRITERIA
4.1 Static and Live Loads
In general accordance with the wider project design philosophy
4.1.1 Rail Loading and Rockfall Protection Structure (based on records and codes), a vertical seismic acceleration (PGAV)
(Tunnel Extension) value of 0.31g (approximately 40% of the horizontal PGAH value)
A rail live load of 90kPa was adopted for the design of this project was used in the pseudo-static models in conjunction with the
in accordance with the recommendations which are published in the horizontal acceleration.
KiwiRail Specification. A load of 100kPa has also been applied,
where appropriate, to account for the weight of the rockfall The design value for the vertical acceleration is in general
protection structure (tunnel extension) which is to be constructed at accordance with the recommendation published in Eurocode 8.
the southern approach to Tunnel 19. This allowance was selected in
collaboration with the designers of the tunnel extension structure. 5. GEOLOGY AND GROUND MODEL
4.1.2 Live Vehicle Loading 5.1 Ground Conditions
A live load of 30kPa has been adopted, where appropriate, during The road and rail corridor is located between the South Pacific
design of the retaining walls, to account for heavy construction Ocean and the foothills of the Kaik�ura mountain range. The
traffic and equipment. following sub-sections provide a general description of the
geological, hydro – geological and geotechnical conditions which
4.1.3 Rock Fall Protection Bund and Attenuation Energy are inferred to underlie the subject site.
Load 5.2 Geological Setting
The following load allowances have been made for the rock fall At the location of Site 6 the major geological unit is Pahau Terrane
protection bund and rock fall energy attenuation loads. These loads (Ktp). This unit is described in the relevant published geological
were provided by the designers of the Rock fall bund. map as well bedded and poorly bedded sandstone. These units are of
Rock fall bund dead load = 55kPa. early cretaceous age and are approximately 145 million years old.
The site-specific geotechnical investigation results indicate a thin
Rock fall energy attenuation energy load = 80kPa to
layer of beach gravels and/or colluvium overlies the sandstone
340kPa.
bedrock beneath most locations along the proposed Site 6 RTW 3
4.2 Seismic Loads footprint.
Immediately following the Mw Kaik�ura Earthquake, the damage at 5.3 Ground and Surface Water Conditions
Site 6 and to the existing road and rail infrastructure was observed to Fifteen machine drilled borehole investigations have been
be predominantly due to landslip debris impact that originated in undertaken within the general area around Site 6. The bedrock level
sandstone/mudstone materials above these assets. The peak beneath the Site 6 retaining walls varied between 11 m and 14 m
horizontal ground acceleration at the site during the Mw 7.8 below the existing ground surface level (bgl) and around 1m bgl in
earthquake of 14 November 2016 is inferred from data published by the Shag Rock area (refer to Figure 4 for approximate location). The
the University of Canterbury as being in the order of 0.50g. As material overlying the bedrock is generally described as colluvium
such, the site is considered to have been recently tested to between a comprising medium dense silty sandy gravel and occasionally Silt
1 in 500 and 1 in 1000 year return period seismic event. with some gravel and boulders. The ground water level was
In terms of AS/NZS1170.5 the site is assessed to be a Class B – generally observed to be sitting on the colluvium rock interface and
Rock site as it meets the following criteria: at a depth of between 11 and 14m bgl.
The subgrade material comprises rock with a compressive 5.3 Geotechnical Parameters
strength of between 1 and 50 MPa, and, an average shear wave
The geotechnical design parameters which were adopted during the
velocity greater than 360 m/s. Further, this site is underlain by
analysis of RTW 3 are summarised in Table 2 below.
materials having a compressive strength greater than 0.8 MPa and
shear wave velocity greater than 300 m/s as per Class B Rock site. Table 2: RTW 3 Geotechnical Design Parameters
489
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Existing and
Proposed
Final Rail Proposed Retaining Wall RTW 3C
maintenance
Location (The primary subject of this paper)
Access track
490
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The preliminary options for the Site 6 RTW 3 were assessed and the 6.2 Selection of Preferred Option
preferred solution selected based on due consideration of the
following key criteria: The conventional bored pile retaining wall option (Option 7 above)
The likely performance of the retaining wall. was identified as the preferred option primarily because it was
The constructability of the retaining walls within the judged to have the lowest level of construction risk, could be
timeframes stipulated by the New Zealand Government to constructed within the available corridor width, and could achieve
most if not all of the design requirements in particular rapid
open the road and rail transportation corridors. reconnection of this lifeline route post- earthquake. All of the other
Consideration of the temporary live operation of the rail options were judged to have some technical fatal flaw.
on the old alignment and,
Cost. A reinforced concrete bored pile wall with capping beam was
identified from the options assessment process as the best available
solution to retain the ground adjacent to the proposed road. The
The following retaining wall options were considered during the
ground between the piles was designed to activate an arching
options assessment phase of the Site 6 RTW 3 design. mechanism. Megaflow drainage and concrete shotcrete facing was
also provided between the piles to provide long-term durability,
Option 1 - No retaining walls and provide stable batters. control ground water pressures, and, ensure an appropriate level of
This option was found to be infeasible due to the in-service performance was obtained. For aesthetic reasons an
unavailability of space between the new road architectural facing is proposed to dress the exposed section of the
and rail alignment. wall, however, this architectural facing was not finalised at the time
of writing this paper. Due allowance for an architectural dressing
Option 2 - Cantilever pole retaining wall designs was made during the detailed design of the wall super structure.
comprising steel or timber poles. These wall
designs were found to be unsuitable for reasons 7. GEOTECHNICAL DESIGN METHODOLOGY
of durability, strength and/or deflection.
The geotechnical stability and structural demands on the proposed
Option 3 - Gabion basket retaining wall. As with Option 2, retaining walls were analysed using the specialized software
preliminary analysis indicated that this solution package FLAC version 8.00.448 and Mohr-Coulomb soil models.
was unlikely to achieve a satisfactory level of Figure 8 provides an example typical output from the FLAC Model,
performance with respect to deflection, in showing total wall displacements.
particular as insufficient space was available to
provide geogrid ‘tie backs’ and enable the The structural demands on wall RTW 3 which were obtained
fastest possible reinstatement of this critical from the FLAC analysis were cross checked using the geotechnical
lifeline route (i.e. similar issues to options 4 software package Wallap V6.06. For the purposes of the Wallap
below). assessment, the cantilever walls were assessed to act as a stiff
system and the anchored wall were assessed to act as a rigid system
Option 4 - Reinforced earth wall (soil nail or MSE). This
option was assessed to have significant Table 3 above summarises the key aspects of the final design
constructability issues as it required a significant which was developed for each of the eight sub sections of RTW 3.
excavation to be made behind the wall and the
placement of geogrids and engineered fill which
would have required the relocation or removal
of the rail line and ballast. This in turn had the
RTW3D1
potential to cause significant disruption to the RTW3C
rail service and delay reconnecting of this
critical lifeline route. RTW3B
Option 5 - Secant bored piles (SBP). This option was
assessed to be infeasible due to construction Shag Rock
constraints such as available cutters for
secondary piles and duration. RTW3D2& D3
Option 6 - Contiguous flight auger piles (CFA). The
primary advantage of CFA piles was the method RTW3A1 RTW3A2
does not require temporary casing to support the
pile holes. However, this option was assessed to Debris Flow Structure N6
be infeasible due to constructability issues in the
colluvium and rock materials. Figure 5: Aerial Photograph of complete portion of RTW 3 which is
located adjacent to Shag Rock and Debris flow Structure N6
Option 7 – Conventional bored pile wall with shotcrete
infill panels. This was selected as the preferred
491
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
492
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
11. CONCLUSION
NCTIR Site 6 was severely damaged by landslides originating
above the road surface during the November 2016 Kaik�ura
Earthquake.
493
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: In urban area with soft clay deposit, it is required that the diaphragm wall deflection of a deep excavation be limited to
minimize damages to adjacent buildings. The auxiliary measures such as buttress wall, cross wall and ground improvement have been
applied to deep excavation. The so-called three-dimensional effect accounts for the presence of buttress wall and cross wall as well as the
size of the excavation area and depth. When considering the three-dimensional effect, the diaphragm wall deflection can be much reduced
compared with its plane strain condition. This study uses three-dimensional finite element analysis, by means of PLAXIS 3D, to predict the
wall deflection. The numerical results are compared with the outcome from a one-dimensional analysis and field measurements. In this study,
the effect of auxiliary measures is incorporated into the system stiffness of the excavation, where the presence of auxiliary measures is given
in form of increasing the undrained shear strength of the soil mass within the excavation zone. In turn, it leads to an increase of the factor of
safety against basal heave and a reduction of the wall deflection. One case history in soft clay are presented in this paper to evaluate the
effectiveness of the three-dimensional finite element analysis considering the effect of auxiliary measures.
Keywords: deep excavation, three-dimensional effect, wall deflection
2. BACKGROUND
2.2 System stiffness
Clough et al. (1989) first presented the relationship between wall
displacement, system stiffness and factor of safety against basal Figure 1 Relationship between maximum wall movement and
heave (Fb). As shown in Figure 1, wall displacement is controlled by system stiffness (after Clough et al. 1989)
system stiffness and Fb. If the system stiffness is constant, as the
value of Fb increases, the maximum wall displacement would 2.4 PLAXIS 3D program
significantly reduce. The factor of safety against basal heave, Fb, is
calculated based on the equations defined by Terzaghi (1967). The PLAXIS 3D is a finite element program intended for three-
Y-axis in Figure 1 is the ratio of the maximum wall displacement dimensional analysis of deformation and stability in geotechnical
(δhm) to the excavation depth (He), while the X-axis is the system engineering. In three-dimensional model, it can consider three-
stiffness (S) defined by Clough et al. (1989) as follows. dimensional effect, like buttress wall and cross wall. A 3D FEM
model of excavation site was used and the Mohr-Coulomb soil
EI (1) model was applied in analysis. The length, width and depth of the
S model is 120 m, 120 m and 59.9 m, respectively, as presented in
rw xh 4 avg
Figure 2.
Where EI is the flexural stiffness of the retaining wall, rw is the 3. CASE HISTORY
unit weight of water, and havg is the average vertical spacing of
bracing system. It could be seen from Figure 1 that induced 3.1 Site description
maximum wall movement could be interpreted by excavation depth, Case A is a 17-story office building with 4 basement levels. The
S and Fb. Further, having the same S and excavation depth, the project site is about 40 m in length 38 m in width and 59.9 in depth,
greater Fb leads to smaller wall movement. and the excavation depth of the basement is 17.1 m. Diaphragm wall
is 0.8 m in thickness and 34 m in depth together with 5 levels of H-
2.3 TORSA program steel bracing as the retaining system. The typical horizontal spacing
of H-steel bracing is 6 m, and each level of bracing was preloaded to
TORSA is a one-dimensional numerical program which considers
the behavior of retaining wall as a beam resting upon an 50% of its allowable axial capacity. The depth of the cross wall
494
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
extends from GL.-16 m to GL.-34 m. There are six inclinometer maximum wall displacement were G.L.-19.04 m and G.L.-1.63 m.
casings installed in the perimeter diaphragm wall as shown in Figure In other words, the wall behavior at SID-1 and SID-4 were similar,
3. The ground water level was at 2 m below surface level. and SID-3 was coincided with SID-5. As shown in Figure 3, both of
SID-1 and SID-4 were installed in the wall which attached to cross
wall. This might be the reason that the induced displacement could
become very small but unlikely to move outward at final excavation
stage. In contrast, the location of SID-3 and SID-5 were not close to
the cross wall so the wall displacements were larger than SID-1 and
SID-4.
495
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Table 3 Material properties of the diaphragm walls, cross walls and basement floors
Diaphragm Cross Cross
Parameters Unit F1 B1F B4F FS
Wall Wall (a) Wall (b)
m 0.8 0.8 0.8 0.25 0.4 0.2 0.6
5.65 5.65 5.65 0 0 0 0
2.75E+07 1.37E+07 2.40E+07 2.75E+07 2.75E+07 2.75E+07 2.75E+07
2.46E+07 1.74E+07 2.30E+07 2.46E+07 2.46E+07 2.46E+07 2.45E+07
0.2 0.2 0.2 0.2 0.2 0.2 0.2
Table 4 Material properties of the anchors stress analysis, as the clay layers are under undrained type C of total
stress analysis due to limitation of available data from site
Parameters Unit H300x300 H400x400 2H400x408
investigation. As shown in Table 1, the effective Young’s modulus
-7.3, -10.5, of sand layers was determined by the following equation after
Depth m -1 -4.1
-13.7 Hsiung (2009).
E kgf/cm2 2.04E+06 2.04E+06 2.04E+06
cm2 E 2000 N (kPa) (2)
A 118.4 218.7 501.4
EA kN 2.37E+06 4.38E+06 1.00E+07 where N is the number of standard penetration teat (SPT). The
kN undrained Young’s modulus of clay layers was obtained by the
Preload 490 980 980 following empirical equation as reported by Bowles (1996), Lim et
al. (2010), Likitlersuang et al. (2013), Khoiri and Ou (2013).
Details of dimension of analytical model used are described The results of wall displacement of real 3D numerical analysis at
previously and the excavation is approximately 40 m by 38 m. Soil final excavation stage at various locations are shown in Figure 6. It
properties of each layer are listed in Table 1 and elastic-perfect is realized that the wall displacement of above excavation level is
plastic Mohr-Coulomb model is chosen for ground behavior larger than the wall displacement of below excavation surface. This
simulation. The sand layers are under drained condition of effective might be connected with depth of cross wall as the cross wall was
496
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
constructed from one meter above final excavation level down to the
same depth of the wall toe which expects to limit wall displacement
at same depth. In general, the maximum wall displacement is less
than 20 mm at final excavation stage which is insignificant.
Additional corner effect formed by construction of cross wall also
reduces the overall wall displacement, no matter the wall is attached
by the cross wall directly or not.
497
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5. DISCUSSIONS
As described before, only data at final excavation stage are
available, it is thus difficult to have a full discussion for an
excavation in clay with additional cross wall. Through conclusions
in Section 4.4, it is agreed that simulation using TORSA still has a
room to be improved. Therefore, stage by stage deformations from
real 3D analyses are selected for further discussions herein. The wall
deformations of PLAXIS 3D analysis for each excavation stage are
shown in Figure 9, 10 and 11, where the dash line is the depth of
excavation surface. As further excavated, the wall deformation
would become larger. The results of wall deformation of SID-3 and
SID-5 are similar, with maximum wall displacements in the vicinity
of 15 mm. However, it is obvious that the results of SID-1 are
smaller than SID-3 and SID-5, and the maximum wall displacement
is 4 mm. It might be attributed to an apparent corner effect at the
location of SID-1. Figure 10 Wall displacement of SID-3
498
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
7. REFERENCE
Terzaghi, K. (1967). Theoretical Soil Mechanics, John Wiley &
Sons, New York.
Clough, G.W., Smith, E.M., and Sweeney, B.P. (1989). "Movement
control of excavation support systems by iterative design".
Proceedings, ASCE Foundation Engineering: Current
Principles and Practices, 2, 869-884.
Ou, C.Y., Chiou, D.C., and Wu, T.S. (1996). "Three-dimensional
finite element analysis of deep excavations". Journal of
Geotechnical Engineering, ASCE, 122(5), 337-345.
Finno, R.J., Blackburn, J.T., and Roboski, J.F. (2007). "Three-
dimensional effects for supported excavations in clay".
Journal of Geotechnical and Geoenvironmental Engineering,
ASCE, 133(1), 30-36.
Hsieh, H.S. and Lu, F.C. (1999). "A note on the analysis and design
of diaphragm wall with buttresses". Sino-Geotechnics, 76, 39-
50 (in Chinese).
Hsieh, H.S. and Huang Y.H., Hsu, W.T. and Ge, L. (2017). "On the
system stiffness of deep excavation in soft clay". Journal of
GeoEngineering, Vol. 12, No. 1, pp. 21-34.
Hsiung, B.C. (2009). "A case study on the behavior of a deep
excavation in sand". Computers and Geotechnics, 36, pp.
665-675
Figure 11 Wall displacement of SID-5 Bowles, J.E. (1996). Foundation analysis and design, 5th Edition,
McGraw-Hill Book Company, New York, USA.
6. CONCLUSIONS Khoiri, M. and Ou, C.Y. (2013). "Evaluation of deformation
Through this study, following conclusions are made. parameter for deep excavation in sand through case histories".
(1) System stiffness (S) in excavation is an index which includes Computers and Geotechnics, Vol. 47, pp.57-67.
flexural stiffness of the retaining wall, the unit weight of water, Lim, A., Ou, C.Y., and Hsieh, P.G. (2010). "Evaluation of clay
and the average vertical spacing of bracing system. constitutive models for analysis of deep excavation under
(2) By considering S, safety factor against basal heave (Fb) and undrained conditions". Journal of GeoEngineering, TGS, Vol.
excavation depth, the maximum wall displacement could be 5, No. 1, 9-20.
predicted. In addition, having the same S and excavation depth, Likitlersuang, S., Surarak, C., Wanatowski, D., Oh, E., and
larger Fb should lead to smaller displacement. Balasubramaniam, A. (2013). "Finite element analysis of a
(3) It is aware that wall displacement could be much smaller at the deep excavation: a case study from the Bangkok MRT". Soil
place once cross wall directly attached to it. However, the depth and Foundations, Vol. 53, No. 5, pp. 756-773.
and height of cross wall also affect the magnitude and shape of
wall displacement. The wall displacement might still be
generated if there is no cross wall at said depth.
499
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: In construction practices, diaphragm walls are a cast-in-situ reinforced concrete retaining wall that are constructed using a
slurry supported trench method. The installation process includes slurry supported trench excavation, placing the reinforcement cage,
concrete casting and curing. This installation process would modify the in-situ stress state in the soil close to the trench and generate ground
surface settlements, which might be significant compared to those induced by the main excavation. Also, the construction of buttress walls, a
concrete wall that perpendicular to diaphragm walls, might generate additional ground surface settlement, and this issue has not been
investigated. For clarify this issue, a series of three-dimensional finite element analysis was performed to quantify the amount of ground
surface settlement induced by the diaphragm and buttress walls installation process using the Wall Installation Modeling (WIM) method.
Results show that the installation of buttress walls inside or outside the excavation zone did not yield significant additional ground surface
settlement outside the excavation zone because the diaphragm wall was completed first before the construction of the buttress wall. But, the
construction of outer buttress walls could widen the settlement zone.
Keywords: Excavation, Settlement, Installation, Buttress wall, Diaphragm wall.
1. INTRODUCTION restrained while they were restrained in all directions at the bottom
of the geometry.
In construction practices, diaphragm/buttress/cross walls are a cast-
in-situ reinforced concrete retaining wall constructed using a slurry
supported trench method. The installation process includes slurry
supported trench excavation, placing the reinforcement cage,
concrete casting and curing. This installation process would modify CL (1)
the in-situ stress state in the soil close to the trench and generate
ground surface settlements, which might be significant compared to Diaphragm wall
those induced by the main excavation. panel
500
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
X=0 m
L dw =4m L bw =6m
Figure 2 Construction stages of the diaphragm wall and buttress wall panels
For each stage, three additional steps should be conducted to model not necessary to be considered because the soil would be excavated
the WIM method such as: soon after the retaining wall system was constructed. Moreover, the
outer buttress wall trench excavation widened the ground settlement
(1). The excavation under slurry support was modeled by zone but the additional ground surface settlement induced by the
deactivating the respective finite elements inside the trench and outer buttress walls trench excavation was insignificant. Thus, it
applying the distributed loads on the surface of the trench could be concluded that the installation of buttress walls has no
walls. The magnitude of the loads corresponds to the significant effect on the additional ground surface settlement
hydrostatic slurry pressure with a bulk unit weight of b=10.3 induced by the buttress walls trench excavation because the
kN/m3. diaphragm wall was completed first before the construction of the
(2). On the subsequent process of concrete pouring, the distributed buttress wall.
loads were increased from the slurry to the fresh concrete
Uz (mm)
pressure (c). The pouring process was modeled following the
10
bilinear approximation by Lings et al. (1994), which adopts a 8
hydrostatic pressure distribution up to a critical depth (hcrit) of
6
20–30% of the panel depth. Below hcrit, the pressure gradient
4
corresponds to bulk unit weight b of the bentonite slurry:
2
c z , z hcrit 0
c (1)
b z c b hcrit , z hcrit
-2
-4
-6
where c is the bulk unit weight of the concrete, b the bulk unit
(a) (b) -8
weight of the slurry, z the depth below surface ground level.
-10
(3). The finite elements representing the fresh concrete were -12
activated inside the trench, and the distributed loads are -14
removed. The increased stiffness of the concrete due to aging is -16
considered by a suitable evolution of Young’s modulus, E, and -18
the Poisson ratio, in the course of 28 days. -20
-22
-24
3. RESULTS AND DISCUSSIONS -26
-28
Fig 3 shows the contour of ground surface settlements induced by
(c) (d) -30
the diaphragm and buttress walls installation. As shown in Fig 3, the
installation of buttress walls inside the excavation zone did not yield Figure 3 The contour of ground surface settlements induced by: (a)
significant ground surface settlement outside the excavation zone. diaphragm wall installation, (b) diaphragm wall with single buttress
However, it could generate the ground surface settlement inside the wall installation, (c) diaphragm wall with three buttress walls
excavation zone. Furthermore, buttress walls also modeled outside installation, (d) diaphragm wall with seven buttress walls
the excavation zone. Fig 4 shows the contour of ground surface installation
settlements induced by the diaphragm and outer buttress walls
installation. As shown in Fig 4, based on the maximum settlement Furthermore, the ground surface settlements at each cross
point of view, the installation of outer buttress walls yielded section of the diaphragm wall with inner and outer buttress walls
insignificant additional ground surface settlements outside the were plotted in Fig 5 and Fig 6, respectively. The ground surface
excavation zone. But, the construction of outer buttress walls settlement (vw) and the distance behind the diaphragm wall (d) are
widened the settlement zone. normalized with the depth of the trench (Ht). The main influence
Table 2 summarizes the maximum ground surface settlement range of settlement was 0.3 to 0.5Ht from the diaphragm wall trench
induced by the diaphragm and buttress walls installation inside and panel, and small settlement occurred beyond 1.0Ht from the panel.
outside the excavation zone. The maximum ground surface This observed settlement characteristic was also reported by Ou and
settlement occurred at the center of the diaphragm wall section Yang (2000) in which they monitored the settlement induced by the
(x=0). The maximum ground surface settlement induced by construction of the diaphragm walls for the excavations in the Taipei
diaphragm wall installation was 16 mm. It was apparent that the Rapid Transit System. In addition, at some cross-sections, small
installation of seven inner and outer buttress walls only increased amount of ground surface heave was detected from the computation
1.1 mm and 0.6 mm of ground surface settlements, respectively, and results, especially at the location between 0.5 to 1.0 d/Ht. According
they were minimal. The inner buttress wall trench excavation only to author experiences, it seems that the ground heave was unlikely to
induced ground surface settlement in the excavated zone but it was be occurred in the field. The possible reason might due to the
501
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
D-wall only 16.0 11.3 14.8 11.8 12.4 11.8 15.5 d/Ht
0 1 2 3 4
D-wall + 1 Inner
16.5 11.4 15.0 11.9 12.5 11.9 15.5 -0.02
B-wall
0
/Ht (%)
D-wall + 3 Inner
16.6 11.3 15.3 12.2 12.9 12.3 15.8 0.02
B-walls
vw
0.04
D-wall + 7 Inner
17.1 11.8 15.9 12.8 13.4 12.7 16.3 (e) 0.06 (f)
B-walls
D-wall + 1 Outer
16.2 13.5 15.2 11.4 13.7 11.3 15.3
B-wall
D-wall + 3 Outer Legend:
16.3 13.3 15.5 11.8 14.3 11.9 15.6
B-walls
D-wall + 7 Outer
16.6 13.6 16.0 12.10 14.7 12.1 16.5
B-walls
Note: x indicates the distance away from the center section of the diaphragm (g)
wall
Figure 5 Profile of ground surface settlements at each cross section
of the diaphragm wall and inner buttress walls: (a) x = 0 m, (b) x = 4
Uz (mm) m, (c) x = 8 m, (d) x = 12 m, (e) x = 16 m, (f) x = 20 m, (g) x = 24 m
10
8
6
4
2
0
-2
-4
-6
-8 (a) (b)
(a) (b) -10
-12
-14
-16
-18
-20
-22
-24 (c) (d)
-26
-28 d/Ht
0 1 2 3 4
-30
-0.02
(c) (d)
0
/Ht (%)
Figure 4 The contour of ground surface settlements induced by: (a) 0.02
diaphragm wall installation, (b) diaphragm wall with single outer
vw
0.04
buttress wall installation, (c) diaphragm wall with three outer
0.06 (e) (f)
buttress walls installation, (d) diaphragm wall with seven outer
buttress walls installation d/Ht
0 1 2 3 4
Legend:
-0.02
4. CONCLUSION
0
502
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
503
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Shield tunnelling construction in urban area could influence surrounding structures. Several empirical equations and
laboratory testing using centrifuge model already published to determine effect of tunnel excavation, especially due to the shield loss and tail
loss effects. Gap between shield machine and tunnel segment lining is the main subject that determine the amount of surface settlement and
lateral displacement. In this study, case study for evaluating ground deformation during tunnel construction is located in Mass Rapid Transit
Jakarta project around diaphragm-wall Bendungan Hilir station. Evaluation consists of comparison between prediction analysis using 3-
dimensional finite element numerical model and actual deformation from surface settlement and inclinometer are conducted to evaluate the
modelling effectiveness.
Keywords: shield tunneling, ground loss, surface settlement, lateral displacement and mass rapid transit Jakarta
1. INTRODUCTION then conducted to evaluate best fit result between predicted and
actual measured deformation.
Mass Rapid Transit Jakarta is the first shield tunnelling using Earth
Pressure Balance (EPB) method constructed in Indonesia. EPB
method are commonly used for fine grained soils (<0,06 mm) such
as clay-silt and silt-sand soft to stiff consistency and active support
pressure control conducted to ensure low influence to settlement and
lateral deformation (B. Maidl, et al., 2012). Anticipation of possible
problems especially to surrounding structure due to excessive
settlement, cave in and building damage should be considered
during design and construction phase. According to Mair and Taylor
(1997) and Nagen Loganathan, et al (2011) deformation due to
tunnel construction mainly caused by face loss, over cutting, shield
loss and tail loss. Even though gap usually determined by
effectiveness of shield machine, soil condition could also influence
displacement that happened. In this study, case study for evaluating
ground deformation during tunnel construction is located in Mass
Rapid Transit Jakarta project around diaphragm-wall Bendungan
Hilir station. Surface settlement marker and inclinometer were
placed above tunnel alignment and near the tunnel cross section to
measure surface settlement and lateral displacement. Evaluation
consists of comparison between prediction analysis using 3-
dimensional finite element numerical model and actual deformation
from surface settlement and inclinometer are conducted. Since
tunnel in this location constructed into hard silty sand material, two
constitutive models using Mohr-Coulomb and hardening-soil model
Figure 1 Location of monitoring site
were evaluated to determine which of the following constitutive
model best suit the actual deformation occur on the field during
construction. 2.1 Soil Parameter
2. METHODOLOGY There were 3 borings conducted in the tunnel section at Bendungan
Hilir station, with distance range from 50 m to 92 m from
Numerical modelling 3-dimension finite element method using
diaphragm-wall (D-wall). First boring was conducted during design
Midas Gts Nx software is conducted to predict and simulated
stage on 2010, while the other boring conducted in 2016 and 2017.
excavation phase in stage construction. Deformation acting around
According to the evaluation from all of the field testing, tunnel
tunnel excavation then compared with surface settlement and lateral
construction was predicted going through hard silty sand and below
deformation measured using digital waterpass combined with
ground water level. According to ASTM D2487-17, soil at project
robotic total station and inclinometer that placed 5 m from outer
location mostly have liquid limit more than 50% and plasticity index
tunnel cross section. Ground condition that modelled in this paper
plots below “A”-line which can be defined as elastic silt (MH) as
based on primary borehole, in-situ test and laboratory test also
seen in Figure 2, while prediction of Young’s modulus according to
secondary data collected by MRT project. All of the data then
pressuremeter test can be seen in Figure 3. Soil parameter used in
review to gain soil parameter that suitable for Mohr-Coulomb and
the modelling can be defined in Table 1 and Table 2.
hardening-soil modelling. Actual and numerical matching method
504
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
100 70,000
MRTJ (2010) MRTJ (2010)
90
BH-1 (2016) 60,000 Bowles (1997)
80 BH-2 (2017) BH-2 (2017)
P r e s u r e m e t e r M o d u lu s, ( k P a )
70 50,000
Plasticity Index (PI)
60
40,000
50
30,000
40
y = 756.91x + 2301.8
30 20,000 R² = 0.649
MH or OH
20
10,000
10
CL or ML
ML or OL 0
0
0 10 20 30 40 50 60 70 80 90 100 110 0 10 20 30 40 50 60
N-SPT
Liquid Limit (LL)
Figure 3 Elastic modulus from Pressuremeter test
Figure 2 Plasticity chart data
2.2 Numerical modelling Surface elevation in tunnel modelling are according to topography at
site to evaluate influence of overburden height (H), diameter of
Numerical modelling using finite element method is usually used for tunnel (D) is 6.55 m and advance of excavation (d) is 1.50 according
evaluating soil structure interaction between tunnel boring machine to lining segment placement as illustrated in Figure 5.
and the surrounding soil, complex model also to estimate
deformation around tunnel on design stage. German Tunneling
Committee (ITA-AITES, 2016), recommended 3 dimensional (3D)
numerical modelling with stage construction to evaluate tunnel
excavation phase. 3D model can give a more thorough information
not only deformation around tunnel excavation, also face stability
each step of excavation progress. In this paper, thickness of ground
loss is according to site project information as seen in Figure 4 and
modelled as soil interface that reduced in term of strength.
505
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3. RESULTS AND DISCUSSION displacement from all of 10 sections show that the trend is displaced
and the value decline along with the distance away from tunnel
3.1 Surface settlement measurement cross section. This trend is relatively similar with previous
According to average value of surface settlement monitoring for 10 researcher (Sven Moller, 2006), and (Loganathan, et al, 2011).
cross section, daily measured starting from November 21, 2016 until Surface settlement that occurred are relatively small. This is
April, 9 2017 as seen in Figure 6, it is shown that at every section possibly due to the high overburden above tunnel construction and
trend of surface settlement curve is quite different. 0 number in x- less disturbance of ground loss from shield and tail loss. Empirical
axis is denote centreline of tunnel section position which shown approach as mentioned in (JSCE, 2006) describe that minimum
largest surface settlement (until -1.39 cm) occurred during overburden that could cause large surface settlement is less than 1.5
monitoring time frame. In some position of surface settlement times tunnel diameter.
marker section, heaving occurred, nevertheless average
-9 15
3.45 m
-8 16
-7 17
-6 18
Tunnel direction
-5 19
-4
20
-3 6.65 m 21
Displacment
direction -2
22
10 9 8 7 6 5 4 3 2 1 0 -1
-1 -2 -3 -4 -5 -6 -7 -8 -9 -10
D (Direction C (Direction 23
0
to D-Wall) to Setiabudi
1 24
Station)
2 Tunnel position 25
3 (Not Scaled) 26
4 27
5 28
6 29
November 28, 2016 (08:00 AM) 7
December 08, 2016 (01:00 AM) 30
December 09, 2016 (01:00 AM) 8 November 28, 2016 (08:00 AM)
December 10, 2016 (01:00 AM) December 08, 2016 (01:00 AM)
December 11, 2016 (04:00 PM) 9
April 06, 2017 (01:00 PM)
December 09, 2016 (01:00 AM)
10 December 10, 2016 (01:00 AM)
December 11, 2016 (04:00 PM)
B (Direction to HSBC Building) April 06, 2017 (01:00 PM)
Figure 7. Result of top view cumulative displacement Figure 8. Result of lateral displacement during tunnel construction
506
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
D e fo rm a tio n (c m )
HARDENING SOIL MOEL
at ground surface while hardening soil model give the opposite 1.00
result.
0.00
-1.00
-2.00
-3.00
Distance (m)
Figure 11. Comparison between numerical and actual surface
settlement
A
(Direction to Meridien B
Building) (Direction to HSBC
Building)
Displacement (mm)
-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10
0
1
2
3
4
Figure 9. Numerical simulation result using Hardening soil Model 5
6
7
8
9
Ground water level (-10.10m) 10
11
12
13
14
D ep th (m )
15
3.45 m
16
17
18
19
20
6.65 m 21
22
23
24
Tunnel position 25
Figure 10. Numerical simulation result using Mohr-Coulomb model (Not Scaled) 26
27
28
Comparison between numerical and actual deformation
29
according to measured instrumentation result are describe in Figure
30
11 and Figure 12. Mohr-Coulomb model fit in great correlation with
November 28, 2016 (08:00 AM) December 08, 2016 (01:00 AM)
average surface deformation result, while hardening soil close match December 09, 2016 (01:00 AM) December 10, 2016 (01:00 AM)
with the maximum surface ground deformation. Opposite result December 11, 2016 (04:00 PM) April 06, 2017 (01:00 PM)
shown in the lateral displacement, i.e. Mohr-Coulomb model fit in HARDENING SOIL MODEL MOHR-COULOMB MODEL
great correlation with maximum lateral displacement, while Figure 12. Comparison between numerical and actual lateral
hardening soil model give a smaller result. displacement
507
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
4. CONCLUSIONS 5. REFERENCES
In this study, numerical analysis compared with actual measurement ASTM D2487-17. Standard Practise for Classification of Soils for
data from instrumentation of surface ground settlement and lateral Engineering Purposes (Unified Soil Classification System).
displacement. The main results can be summarized as follows: B. Maidl, M. Herrenknecht, U. Maidl, G. Wehrmeyer. (2012)
(1) Mohr-Coulomb model give a good result to evaluate lateral Mechanised Shield Tunnelling, 2nd Edition, Wilhelm Ernst
displacement on tunnel construction but give a less and Sohn. Berlin, Germany.
conservative result on the surface ground settlement. ITA-AITES. (2016) Recommendations for face support calculations
(2) Hardening soil model give a smaller lateral displacement on for shield tunneling in soft ground, German Tunneling
tunnel construction but give a more conservative result on the Committee.
surface ground settlement JSCE. (2007): Standard specifications fortTunneling: Shield tunnels.
Japan Society of Civil Engineers.
This study was collaboratively carried out based on Mair, R. J., Taylor, R. N. (1997): Bored tunneling in urban
“Agreement on Research Exchange and Cooperation between environment, Plenary Session 4, Proc. 14th Int, Conf. on
Institute of Road Engineering and MRT Jakarta in order to develop SMFE, Hamburg, Vol. 4, pp. 2353-2385.
technical guideline for shield tunnelling using earth pressure balance Nagen Longanathan. (2011): An Innovative Method for Assessing
method. Tunnel-Induced Risks to Adjacent Structures. Parsons
Brickerhoff Inc. New York.
Sven Moller. (2006) Tunnel induced settlements and structural
forces in linings. Phd thesis. Universitat Stuttgart. ISBN-10:
3-921837-54-5.
508
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Aplikasi Metode Elemen Hingga dan Kriteria Runtuh Papaliangas Batuan Berkekar:
Problem-Problem Analisa Terowongan, Kestabilan Lereng dan Wellbore Stability
dalam Mekanika Batuan
G.L. Wilham1
1
PT Signifikan Solusi Integrasi and Lecturer at Postgraduate Program of Civil Engineering Dept.,
National Institute of Science and Technology (ISTN), Jakarta, Indonesia.
1
E-mail: wilham@signtegra.com and wilham.george@gmail.com
ABSTRACT: On the early days of Finite Element Method research, Prof. Zienkiewicz et.al. (1968) in his monumental article, implemented
a no-tension material for jointed rock mass (material that cannot withstand tensile stress) into existing Mohr-Coulomb Joint Rock criterion
for shear stress analysis. With background from the geologists / geophysicists research between 1960s to1990s, Papaliangas et.al. (1996)
proposed a new breakthrough in joint rock failulre criterion, that is the Brittle-Ductile Transition. The old Mohr-Coulomb equation had been
used in various rock mechanic problems (Zienkiewicz et.al. 1971, Locat et.al. 2000) and in petroleum engineering / engineering geologists
(Hetcher 1995, Aoki et.al. 1993, McLean 1998, Ramsay & 2000), and it does not include the Brittle Ductile Transition element. In this paper,
the author tries to implement a new algorithm of Papaliangas into the older Finite Element Method code of Zienkiewicz-Pande (1977), so
this assistance can help to solve the rock mechanics problems (in Slope Stability, Tunneling, wellbore stability, etc.). Some examples have
been presented. Failure Indices and Failure Algorithm have been proposed and explained.
Keywords: joint rock, Tunneling, Failure Indices, Britle-Ductile Transition (Papaliangas), Slope Stability, Wellbore Stability
ABSTRAK: Di masa permulaan riset Metode Elemen Hingga Prof. Zienkiewicz dkk. (1968) dalam artikel monumentalnya mengusulkan
kriteria kekar/batuan sebagai material yang tidak dapat menahan tegangan tarik (no-tension material) kedalam kriteria runtuh kekar Mohr-
Coulomb untuk analisa tegangan gesernya. Dengan memakai latar belakang riset dari para ahli geologi / geofisik antara masa 1960-an s/d
1990-an, Papaliangas dkk. (1996) mengusulkan suatu terobosan baru dalam rumus dasar keruntuhan kekar dengan mengimplementasikan
pengaruh Transisi Getas-Daktail (Brittle-Ductile Transition) pada kekar/batuan. Kriteria runtuh lama (Mohr-Coulomb) yang dipergunakan
para pakar mekanika batuan (Zienkiewicz et.al. 1968, Locat dkk. 2000) dan ahli geologi teknik / teknik perminyakan (Hetcher 1995, Aoki
dkk. 1993, McLean 1998, Ramsay & 2000) tidak memasukan unsur transisi Getas-Daktail. Penulis mencoba mengimplementasikan
algoritma Papaliangas yang baru ini kedalam code (listing program Fortran) Metode Elemen Hingga lama dari Zienkiewicz-Pande (1977),
sehingga hasilnya diharapkan dapat bermanfaat dalam pemecahan problem-problem mekanika batuan, perencanaan terowongan (tunneling),
kestabilan lereng dan stabilitas lubang bor (wellbore stability). Beberapa contoh telah ditampilkan dan dibandingkan disini. Pengenalan akan
Indeks keruntuhan dan algoritma keruntuhan juga diterangkan disini.
Keywords: Kekar/batuan, Penerowongan, Indeks Keruntuhan, kiteria runtuh Britle-Ductile (Papaliangas), Kestabilan Lereng, Kestabilan
Lubang Bor
1. PENDAHULUAN ahli Geologi Teknik / Teknik Permiyakan (Hetcher 1995, Aoki dkk.
1993, McLean 1987, Ramsay 2000) belum bisa menerapkan
Banyak pekerjaan teknik sipil dan pertambangan melibatkan fenomena Transisi Getas-Daktail (brittle-ductile transition).
pembuatan perkuatan lereng, terowongan dibatuan, seperti Umumnya tinjauan dipisahkan dalam masing-masing masalah.
terowongan untuk “spillway” bendungan, ruang penyimpanan mesin Seperti McLean (1987), hanya meninjau keruntuhan batuan utuh
pembangkit listrik pada bendungan, terowongan pada tambang- (intact rock) karena umumnya hal itu terjadi dalam problem
tambang. Demikian pula untuk dunia perminyakan, analisa wellbore rekayasa perminyakan (simulasi lubang bor / “wellbore”). Problem
stability juga sangat penting. Umumnya batuan yang ditemukan “wellbore” menyangkut kedalaman yang besar, maka umumnya
adalah batuan utuh (intact rock) dan batuan berkekar (joint rock). dijumpai keruntuhan pada batuan utuh (intact rock); walaupun
Perencanaan pembangunan proyek terowongan batuan melibatkan keruntuhan kekar (joint rock) juga terjadi.
berbagai bidang ilmu, dan satu diantaranya adalah mekanika batuan Sedangkan Locat dkk. (2000) sudah menyuguhkan grafik
dan pemodelan konstitutip (kriteria runtuh) dari material batuan utuh keruntuhan batuan utuh dan kekar batuan dalam satu gambar, namun
/ berkekar tsb. karena maslah yang dihadapi adalah kestabilan lereng dengan
Era teknologi informasi telah membawa kita/ para insinyur, kedalaman rendah maka yang dijumpai umumnya keruntuhan pada
kepada era penyelesaian dengan dibantu komputer dan metode kekar batuan (joint rock).
numerik dengan berbagai algoritma perhitungan yang kompleks. Berkaitan dengan itu penulis, dalam software Newmo3962_2000,
Prof. Zienkiewicz et al. (1968) mengusulkan kriteria batuan mencoba mengimplementasikan algoritma baru (Louhenapessy
sebagai material yang tidak dapat menahan tarik (no-tension 2000), yang diharapkan dapat bermanfaat dalam memecahan
material) dan memakai kriteria runtuh Mohr-Coulomb untuk problem-problem mekanika batuan. Beberapa contoh numerik akan
tegangan gesernya. Papaliangas (1996) melakukan riset-percobaan ditampilkan dan dibandingkan dalam sub-bab sub-bab berikut ini.
kekar batuan dan mengusulkan suatu terobosan baru dalam rumus Adapun sistematika pembahasan makalah ini dapat digambarkan
dasar keruntuhan batuan. Tentu saja Papaliangas didukung riset-riset seperti pada bagan dibawah ini (Gambar 1).
para pakar Geologi dan Geofisik sebelumnya (Greenwood &
Williamson 1966, Byerlee 1978, Scholz 1990). Model konstitutip
baru Papaliangas segera disambut dengan pemodelan numeric
2. KRITERIA RUNTUH LAMA (MOHR COULOMB)
konstitutip dari batuan utuh dan kekar batuan oleh penulis
(Papaliangas 1999, Louhenapessy 2000). 2.1 Kriteria runtuh Mohr-Coulomb
Paper ini memperkenalkan algoritma yang diusulkan penulis
sebelumnya (Louhenapessy 2000) dan juga memberikan beberapa Pemodelan kriteria runtuh Mohr-Coulomb unuk kekar batuan
hasil dari contoh simulasi numerik. Dari studi literatur, kriteria adalah:
runtuh Mohr-Coulomb yang dipergunakan para pakar mekanika
batuan sebelumnya (Zienkiewicz et.al. 1971, Locat dkk. 2000) dan F = |j | + n tan j = 0 (1)
509
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
dimana, dimana,
j = sudut geser kekar batuan (joint rock friction angle),
j = tegangan geser pada bidang kekar, dan
log 10 nT (5)
n = tegangan normal pada permukaan kekar. n
tan tan o
log 10 nT
no
j = n tan ( r + i) (2a)
dimana,
r = sudut geser yang tersisa (residual) pada kekar batuan,
j = tegangan geser pada bidang kekar, dan
n = tegangan normal pada permukaan kekar.
Gambar 2 Grafik garis runtuh dan sudut geser dalam, m
Sehingga kriteria runtuhnya menjadi : (Papaliangas dkk. 1996)
_
F = |j | + n tan (m+) = 0 (4) dimana , dan m adalah komponen-komponen invariant.
510
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
7. CONTOH NUMERIK
7.1 Terowongan (Tunnels)
Dalam contoh ini sebuah terowongan lingkaran digali pada sebuah
kedalaman didalam massa batuan berkear (jointed rock mass) yang (b)
memiliki satu set kekar (one sets of joints). Hal ini ditampilkan Gambar 5 a) Algoritma tahap-tahap pengambilan keputusan untuk
untuk contoh perhitungan Indeks Keruntuhan (Failure Indices). menghasilkan Indeks Keruntuhan Kekar Batuan; b) Algoritma
Gambar 6 dan Tabel 2 menunjukan geometri dan material tahap-tahap pengambilan keputusan untuk menghasilkan Indeks
properties dari contoh numerik ini. Keruntuhan Global (Louhenapessy 2000).
511
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(a)
(b)
512
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
513
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
514
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This paper is elaborating the application of active lime and cement as soil stabilization material, especially in remoulded
residual soil. The main focus of this paper is measuring the mixing effects of original soil + active lime + cement in soil strength and
workability perspective. The original soil, which is used as filing materials, consists of sandy clay (high water content) with gravel. The
active lime is supposed to have reaction with the clay material and also to reduce the water content, while cement is supposed to have
reaction with sand/granular material. Mixing formula for the materials are: original soil (Sandy silt/clay with sand & gravel) + 10% Active
lime + 3% Cement. The criterion for active lime is it has more than 90% of CaO which will react with the clay minerals such as Sodium or
Potasium.
Keywords: Active lime, cement, clay, sand, stabilization.
1. PREFACE
Weather condition in mountainous area is quite unpredictable. The
morning sky is clear, but heavy rain could come suddenly in the
noon, and can last until evening. This event can happened even
between April to September in the same year, when normally
considered as dry season. Obviously this case was a disruption for
the construction works, which has tight schedule. The original soil,
that'll be used as filling materials, is always wet due to the heavy rain
and couldn’t be compacted by hand or Vibro-Compactor. Several Figure 1.2 Original Soil Compaction Curve
methods have been proposed to overcome this issue, such as; (Source: BPJN 3 Padang Soil Lab)
Maximum Optimum Moisture Content (OMC): 31%
1. Importing granular material or red clay Maximum Dry Unit Weight (γd): 1.34 t/m3
2. Waiting the original soil to drained naturally
3. Speeding up the drying process by mixing the original soil with
As mentioned in abstract section above, the active lime contains
something that makes it quickly dry and well compacted.
CaO more than 90% and was made of limestone which is burnt in the
The last afore mentioned method has been chosen, based on +1000°C stove. After burning process in stove, the burned limestone
consideration of the weather, schedule and cost. The Original become active lime and has different characteristic from its original
material is the residual soil where is located at the mountainous area form as follows:
(+1410 MASL) and has characteristic as follows: Chemical compostition:
Soil Description : Sandy Silt 1. Minimum CaO and MgO (%): 90
USCS Chart : MH 2. Maximum CO2 (%): 5
Specific Gravity (GS) : 2.59 3. Maximum water content (%): 5
Wet Unit Weight (γm) : 1.34 t/m3 Physical properties:
Dry Unit Weight (γd) : 0.58 t/m3 1. 100% passing 1” sieve.
Water Content (ωn) : 131% 2. Unit weight (kN/m3): 13
Liquid Limit (LL) : 157.3% The active lime-soil reaction can be explained as follow; active
Plastic Limit (PL) : 46.75 lime addition in to the soil gives the calcium (Ca2+) and magnesium
Plasticity Index (PI) : 110.6 (Mg2+) to the soil. These ions tend to change the existing ions in the
Passing no.200 : 66% soil such as sodium (Na+) or potasium (K+). This changing will
reduces the plasticity index significantly and also adsorbs the water
which indicates the chemical reaction between lime & soil. Beside
reducing the plasticity index and water content, the grain size
distribution also changed. The new grain size, which is comprises of
slightly cemented clay aggregate, mostly floculated and become to
sand size particle.
2. METHODOLOGY
Several samples of original soils, active lime and cement were
prepared prior to mixing and testing in laboratory. These materials
have its percentage composition (By weight) in the mixing process
based on its particle size distribution. The percentage of active lime
is 10% and for cement is 3%. Before mixing process, the original
Figure 1.1 Grain Size Distribution soil, active lime & cement have been separated and measured by
(source: PT. SOILENS) weight based on defined percentage. The active lime was broken by
hammer until meet the grain size criteria (max 5 cm). After breakage
of active lime; original soil, water & cement with certain percentage
were poured together in bucket and mixed manually by hand (with
To check the performance of original soil if being compacted, the gloves). The mixing process has been taken carefully since the active
original soil also was tested using standard proctor test as follow: lime is very reactive and feels so hot in the hand. The mixing process
515
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
stopped after the materials mixed well and could be indicated by 3.3 Unsoaked & Soaked CBR Test
colour and texture changing. The mixed materials were put into the
closed bucket and left for 1 day for curing time. After curing process, Since internal plant road laid on the mixed material fill, the CBR
the mixed materials are ready to be tested. There are several samples value of this materials also was tested as part of the road design. The
of mixed material with different water content added. mixed materials were checked by soaked CBR test using SNI 1744-
The laboratory tests for mixed materials are: 2012 with curing time 24 hours for unsoaked CBR and 96 hours for
1. Modified Proctor soaked CBR. The result as follow:
2. UU Triaxial
3. Unsoaked & Soaked CBR Laboratory
4. Consolidation
5. Grain Size Distribution
From the tests above (Figure 3.3 and 3.4), it can be seen that the
soaked CBR value is higher than unsoaked CBR, this is mostly
related with the curing time process, which the longer curing time
will give the higher result of CBR value. Based on the road design
requirement (Rigid pavement), the minimum CBR at the subgrade
must be at least 6%. Thus we can conclude that the designed CBR of
this mixed materials has strong enough to fulfill the requirement
whether it’s unsoaked or soaked CBR.
Figure 3.2 UU Triaxial Test (Source: BPJN 3 Padang Soil Lab) 3.4 Consolidation
Undrained Cohesion (Cu): 141.7 kN/m2
Internal Friction Angle (φ): 13.08° A constitutive sample of mixed material also was made for
consolidation test (ASTM D 2435), and the result as follow (Figure
The cohesion value which was obtained by UU Test is 141.7 3.5):
kN/m2. Regarding this value, the mixed materials could be Coefficient of Consolidation (Cv) : 0.744 m2/year
considered as very stiff material. Coefficient of Compressibility (Mv) : 0.004 m2/MN
Permeability (k) : 8.927x10-9 m/s
516
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
m v H ' (1) There are significant differences between original and improved
characteristics of the soil. The remarkable characteristic differences
where: are water content, grain size distribution and also shear strength.
δ = Settlement These advantages obviously help the construction works which was
mv = Coefficient of Compressibility supported by this method, regarding the weather and construction
H = Layer thickness schedule. The mixing process at field also has been done with tight
Δσ’ = Effective stress increases QC supervision, especially focusing on water content, uniformity of
materials, maximum lift thickness and also actual field CBR and
3.5 Grain Size Analysis density after compaction. Beside the percentage of lime and cement,
the curing time is also important to increases soil shear strength. 1
The grain size distribution also was tested to identify the particle size day has been taken as curing time since water content has reached the
distribution after mixing (Figure 3.6). optimum, even more days is better (see Figure 3.4 soaked CBR), the
Table 2.1 Grain Size Distribution increases of the strength also have been contributed by cement
contain in the mixed material. The mixed materials at site have been
Sieve Size (mm) Retained (%) Passing (%) treated carefully to maintain its water content during the curing time,
30.48 0.00 100.00 such as providing the tarpaulin sheet on the banks and also providing
the sufficient drainage around the banks. These mixed material in
25.40 0.00 100.00
this paper are purposed as filling material underneath the power plant
19.05 5.16 94.84 facilities such as cooling tower, water treatment plant, warehouse,
12.70 13.95 86.05 road, etc. which required the good bearing subgrade. Since there is
9.53 28.83 71.17 limitation of this paper scope, author suggests the further
4.76 48.99 51.01 investigation/experiments of this mixed material should be done to
get better understanding of active lime & cement mixing behaviour
2.00 54.23 45.77 on each case.
0.42 64.01 35.99
0.01 74.29 25.71 5. REFERENCES
Pan 100.00 0.00
(Source: BPJN 3 Padang Soil Lab) Hardiyatmo, Hary Christady, Dr, Ir., 2013, ”Stabilisasi Tanah untuk
Gravel : 48.99% Perkerasan Jalan” Gadjah Mada University Press.
Sand : 25.3% Wesley, D. Laurence., 2012, “Mekanika Tanah untuk Tanah
Silt & clay : 25.71% Endapan & Residu” Penerbit Andi Yogyakarta.
Balai Pelaksana Jalan 3 Soil Laboratory, 2017 “Laporan Hasil
Already known that the majority of original soil is consist of Pengujian Tanah Timbunan”, Padang.
silt/clay, but after mixing with active lime and cement, it’s altered PT. Soilens Soil Laboratory, 2017, “Soil Investigation Report”,
became granular material (mostly) and it’s aligned with explanation Bandung.
in section I (The Preface) above. Bowles, J. E., 1996, Foundation Analysis and Design, McGraw
HillBook Company, New York, pp. 66–68.
517
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
518
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Dalam kegiatan pelaksanaan pekejaan dilapangan ada hal-hal yang dapat memberikan kelancaran dalam pelaksanaan yaitu
pengetahuan praktis yang di lakukan perhitungan secara manual di bantu dengan program excel untuk dapat mendampingi perhitungan
dengan Software (Program Computer) Slide3D. Pada kasus ini dicoba melakukan kajian kestabilan lereng dilakukan perkuatan dengan
jangkar,dalam kajiannya tahap awal dilakukan analisa dengan program oleh perencana, kemuadian dilakukan analisa dengan cara Fellenius
dengan didang longsor yang dihasilkan dari perhitungan program kindisi statis dan gempa . Dari hasil analisa di peroleh hampir dekat antara
hasil program software (Rocscience’s SLIDE) dan manual (Microsoft Exce)l,untuk kondisi statis: SFsoftware = 1,728 (Fillenius) dan 1,339
(Bishop dan GLE), sedangkan SF untuk hitungan manual, SFmanual = 1,450. Untuk kondisi dinamis, SFsoftware-DYN = 0.861 (Fillenius)
dan 0,998 dan 0.999 (Bishop dan GLE), sedangkan SF untuk hitungan manual, SFmanual-DYN = 1.036. Sementara, besar gaya jangkar
untuuk mencapai harga safety factor, SFsoftware-DYN = 1,095 (Fillenius) dan 1.125 (Bishop) dengan kondisi dinamis gaya jangkar
(software dan manual) adalah sebesar 15 ton. Telah dibandingkan pula hasil analisa 2D dan 3D.
Keywords: Jangkar, Limit Equilibrium, Safety Factor (Faktor Keamanan), Kestabilan Lereng
1. PENDAHULUAN bidang longsor dari analisa ini seperti gambar dibawah ini dengan
bidang longsor terjadi pada lereng.
Analisa pekuatan lereng ini di dasarkan atas hasil analisa yang telah
dilakukan oleh kami dengan penyelidikan Tanah oleh Tarumenegara Table 1 Resume Data Lapisan Tanah
Bumiyasa dan Jurusan Teknik Sipil Universitas Cendrawasih.
Analisa ke satabilan lereng atas pendekatan data hasil uji Tanah Berat isi Sudut ges Kohesi
oleh Tarumenegara Bumiyasa dan hasil uji tanah Jurusan Sipil Lapis Tebal (m) Level
3
T/m dalam t/m3
universitas local, dengan model lereng seperti gambar dibawah ini,
kimiringan lereng berkisar atas perbandingan H = 2 dan V = 5. 1
Sesuai
1.62 14° 1.750
Sesuai
(Gambar 1). gambar gambar
2 2.00 1.68 38° 1.610 +(21-23)
+(15.2-
3 5.80 1.68 36° 2.078
21)
+(14.2-
4 1.00 1.72 39° 2.152
15.2)
+(4.6-
5 10.60 1.68 33° 4.865
14.2)
-7.8 s-
6 12.40 1.68 32° 6.034
+4.6
519
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Atas nilai berat dan sudut kelongsoran pada table diatas dan nilai
parameter kekutan geser tanah sesuai dengan data lapisan tanah
(kohesi C t/m2 dan sudut geser φ o), maka dapat di hitung nilai
ketsabilan seperti dengan sistim table dibawah ini.
520
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Dari table diatas maka dapat di hitung nilai kestabilan lereng maka besarnya gaya penahannya seperti dari penjumlahan dibawah
dari bidang longsor pendekatan di atas tanpa beban gempa yang ini.
nilainya adalah sebesar.
Fpenahan = ∑(Fjkr SinØ + Fjkr*CosØ*Tanφ) (2)
SF = 145 ,963 132 .616 1,450
192 .091 Dengan memasukkan nilai gaya Fjkr secara coba-coba seperti
pada table dibawah ini, maka di peroleh nilai Fjkr adalah sebesar 15
(lebih besar dari hasil program computer SF = 1,301) ton.
Melihat nilai FS = 1,45 diatas lereng cukup aman, namun jika Tabel 5 Perhitungan Gaya-gaya Penahan
bidang longsor ini di kenakan beban gempa dengan kegempaan 0,2* dari Sistim Jangkar Dalam Ton.
W (20% dari berat, W), maka akan di peroleh SF menjadi sangat No. Ø sin Ø cos Ø Fjkr*sinØ Fjkr*cosØ*tanφ
kecil dengan gempa ini akan menambah gaya pelongsor seperti pada 1 73 0.956 0.292 14.34 3.5458
Table 4 dibawah ini. 2 68 0.927 0.375 13.905 4.556
Tabel 4 Gaya Pelongsor Akibat Gempa. 3 63 0.891 0.454 13.365 4.42
0.2*W 0.28W*Cosα 0.2*W*sinα 0.2*W*cosα*tanφ 4 59 0.857 0.515 12.855 5.014
4.72 2.720 3.860 3.860 54.465 17.537
2.93 1.760 2.350 2.350 Total : 72.002
13.62 8.900 10.310 10.320
3.17 2.240 2.240 2.240 Nilai tambahan gaya penahan dari sistim penjangkatran adalah
57,624 ton yang membuat sedemukian sehinga nilai SF menjadi
30.26 23.670 18.880 18.890
1,25 dari yang sebelumnya 1,036.
1.53 1.320 0.770 0.770
40.610 38.420 38.430 5. KESIMPULAN
Besar keamanan longsor akaibar gempa ini menjadi seperti Dari hasil analisa perhitungan terlihat bahwa perbedaan nilai factor
rumus di bawah ini keamanan (SF) antara hasil software dan analisa manual hampir
145 ,963 132 .616 mendekati (Tabel 6), yaitu:
SF = 1,036 1. Kondisi statis SFsoftware = 1,339 (Bishop 2D), hitungan manual
192 .091 38, 42 38, 43
1,450
(sedikit lebih besar dari SF=0.998) . 2. Kondisi dinamis SFsoftware = 0. 861 (Bishop 2D) hitungan
manual 1,036.
Nilai SF = 1,036 suatu suatu nilai yang sangat kecil perlu di
3. Besar gaya jangkar untuuk mencapai SFmanual = 1,25 dalam
lakukan perbaikan dengan minimal SF menjadi 1,25 dan sehinga
kondisi dinamis besar gaya jangkar manual 15 ton , gaya ini
penambahan gaya untuk menahan kelongsoran adalah sebesar :
Tambahan Gaya = 1,25*(192.091+38.433+38.418) sama dengan dari Software, 15 ton, namun SFsoftware = 1.095
– ( 132.616+145,963) (Fillenius) dan 1.125 (Bishop).
4.
= 57,624 ton. Tabel 6 Summary Hasil Perhitungan
SFAnali SF Analisa Program
SF Syarat
Jangkar
4.4 Perkuatan Dengan Sistim Jangkar. sa Komputer
Gaya
Manual
Prinsip Pekuatan dengan sistim jangkar adalah seperti sket gambar Fillenius Bishop GLE
Fillenius
di bawah ini. 2D 2D,3D 2D, 3D
(Excel)
Tanpa Jangkar
Static 1.45 1.728 1.339 1.339 1.5 -
1.488 1.446
Gempa 1.036 0.860 0.875 0.873 1.0 -
0.998 0.999
Dengan Jangkar
Static 1.825 1.71 1.729 1.731 1.5 15
1.811 1.800
Gempa 1.25 1.095 1.155 1.146 1.0 15
1.125 1.126
6. REFERENSI
RocNews (Spring 2017) “Slide3 - Introducing 3D Limit Equilibrium
Slope Stability Analysis , Rocscience, Toronto, Canada”
technology.pitt.edu (2013) “Microsoft Excel 2013 Fundamentals
Workshop”
Gambar 6 Sket gaya-gaya pada sistim Jangkar.
521
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
522
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
G.L. Wilham1
1
PT Signifikan Solusi Integrasi and Lecturer at Civil Engineering Dept.,
Postgraduate Program, National Institute of Science and Technology (ISTN), Jakarta, Indonesia
1
E-mail: wilham@signtegra.com and wilham.george@gmail.com
ABSTRACT: The natural rock slope of Weak Limestone has been studied, modelled analysis and reviewed. This paper describes the Rock
Slope Stability Analysis for the construction of a Generating Station. Four scenarios were analysed i.e.: original proposal of 75o; continuous
slope of 60o; and double step slope of 60o and 60o with a bench; double Steep slope of 65o and 75o with a bench. Slope reinforcements (End
Anchored and High Tensile Mesh) were recommended. The analysis is with Limit equilibrium method, LEM and Finite element method,
FEM.
Keywords: Limestone, Rock Slope Stability, LEM, FEM.
1. INTRODUCTION
This paper describes the Slope Stability Analysis for the
construction of a Generating Station. This paper aims to analyse the
rock slope stability: A steep slope early proposed 75o, and then
propose the following;
a. Steep slope of 60o
b. Double Steep slope of 60o and 60o with a bench
c. Double Steep slope of 65o and 75o with a bench
d.Recommendation of slope reinforcement (End Anchored and
High Tensile Mesh) if needed.
The main analysis conducted is aimed at finding the Factor of
Safety (FS), through the Limit Equilibrium method using
Rocscience’s slope stability Analysis Software Slide 3D and with
Finite Element Method (Griffiths 2015), and a Microsoft Excel
Software.
2. CASE STUDY
2.1 Location
The drilling hole point at the nearest future Slope is at BHX
depicted in Figure 1. Figure 2 showed the Borelog BH-X
523
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3. METHODS OF ANALYSIS
3.1 Limit Equilibrium Method (LEM)
LEM is a method that uses the principle of force equilibrium. This
method of analysis first assumes the field of sliding that can occur,
the field of assumption that is assumed to be circular and non- Figure 9. Algorithm for Failure (Fi) Calculation of the Intact Rock
circular (Figure 7) (Limestone).
LEM is the most popular approach in slope stability analysis.
This method is well known to be a statically indeterminate problem, b. Generalized Hoek-Brown Failure Criterion (constitutive Model)
and assumptions on the interslice shear forces are required to render For a constitutive Model of the Generalized Hoek-Brown (Hoek
the problem statically determinate. Based on the assumptions of the 1995; Eberhardt 2012), the equation is as follows:
internal forces and force and/or moment equilibrium, there are
more than 10 methods developed for slope stability analysis. a
The famous methods include those by Janbu (1957, 1973),
Spencer (1967) and Morgenstern and Price (1965). 1 3 ci mb 3 s 0 (1)
ci
524
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
GSI 100 and Franklin (1984) showed that the Safety Factor for earthquake
mb mi exp (2) loading should be no less than 1.0. See (Right Column Above).
28 14 D
5. RECOMMENDATION AND CONSLUSION
GSI 100
s exp (3) It is recommended that using The Slope Angle of 60os without
9 3D Shotcrete (but instead will use High Tensile Mesh) and using end-
anchor of 5 m x 5 m, with 15 m long and 100 kN tensile capacity.
Table 1 and Table 2 below details the results of the slope
1 1 GSI 20
a exp exp (4) stability analysis (72 Run) along with the loadings (static and
2 6 15 3 seismic) and some of the reinforcements that were provided.
Table 3 showed that from the Matrix of Percentage for Factor of
mi is a material constant for the intact rock, GSI (the Geological Safety compared with Required Factor of Safety, the best option for
Strength Index) relates the failure criterion to geological design is using the double slope of 60o dip angle. While a Table in
observations in the field, and D is a "disturbance factor" which Appendix D, is the FEM Summary Results.
depends upon theo of disturbance to which the rock mass has been The above recommendation is still a bit higher than the one
subjected by blast damage and/or stress relaxation. It varies from 0 recommended by Rodrigues (Appendix B), where in his Table 6.5,
for undisturbed in situ rock masses to 1 for very disturbed rock Rodrigues et.al. recommend that (for broken Limestone) a double
masses. slope with first slope angle is 53o and a higher level slope is 45o
angle. The chosen of broken Limestone is due to the Rock
3.3 The Finite Element Method (FEM) Mechanics Lab. Results and the broken/low values of RQD (Figure
2).
The FEM is a numerical method for solving problems of It is recommended to check the tension crack and make sure if
engineering and mathematical physics. Typical problem areas of there are joint rock appears in the weak limestone the dip angle of
interest in engineering and mathematical physics that are solvable the joint rock should be lower than 20o (Appendix E).
by use of the finite element method include structural analysis,
geotechnical engineering (inc’l Slope Stability, Tunneling etc.), heat Table 1 Summary of Without Shotcrete (Instead, now with: High-
transfer, fluid flow, mass transport, and electromagnetic potential. Tensile steel wire and Erosion Control)
The majority of slope stability analyses performed in practice
still use traditional limit equilibrium approaches involving methods
of slices that have remained essentially unchanged for decades.
Then, in 1967, Whitman & Bailey (1967) set criteria for the then
emerging methods to become readily accessible to all engineers
(Louhenapessy 1995). The FEM represents a powerful alternative
approach for slope stability analysis which is accurate,versatile and
requires fewer a priori assumptions, especially, regarding the failure
mechanism. The author will use Griffiths’ FEM Software, name:
SLOPE64 (Griffiths 2004, Griffiths & Lane 1999) to check the
Limestone Rock Slope Stability.
4. ASSUMPTIONS
The calculation of the factor of safety of the given slope was
conducted, which is done depending on the rock material properties
(based on Mohr-Coulomb and/or Hoek-Brown Constitutive Model),
the slope geometry, static (live) load (10 kPa on top of the slope),
Seismic Load (0.255 g) and if necessary the slope reinforcement.
The two loadings that were used were Non-Earthquake (NE)
loading, and Earthquake (E) loading. Earthquake loading was
provided from Figure 4, The Peak Ground Acceleration Map in
Bedrock (Sb) for Probability Exceeding 2% in 50 Years (Peta
Percepatan Puncak di Batuan Dasar (Sb) Untuk Probabilitas
Terlampaui 2% Dalam 50 Tahun) by The Public Work Department
Republic of Indonesia.
Recommended values of safety factors for rock slopes use, SF >
1.5 reference RSNI “Persyaratan Perancangan Geoteknik BSNI
2017. Recommendation of earthquake values of safety factors using,
SF > 1.0 Hynes-Griffin and Franklin (1984). (Figure 10).
525
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Table 2 Summary (with Shotcrete: Not Recommended, due to sShotcrete is Prone to Groundwater and Can Crack)
526
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
527
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
APPENDIX C – Rock Classification with Mohr Coulomb APPENDIX D –Finite Element Method: SLOPE64
and Hoek-Brown Intact Rock Properties
528
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
529
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
530
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Naturally, soil parameters are uncertain random phenomena in a certain range. This uncertainty is also caused by the process of
soil testing. In a slope stability analyses, the probabilistic approach may become an alternative method to take into account this uncertainty.
The safety factor is optimized with the probability of failure (Pf) that is matching with the soil parameter distribution function using Chi-
Square, Kolmogorov-Smirnov and Anderson-Darling methods. In this study, a number of slopes in Grobogan, Central Java, were analyzed
using Monte Carlo simulation and simplified Bishop with the available soil parameters from laboratory tests. The result indicates that the
greater the safety factor, the smaller the probability of failure. In this study, category of high risk impact (maximum Pf = 25%) produces a
safe slope angle recommendation for maximum slope reaches 5 m with a slope angle up to 55°. The category of low risk impact (maximum
Pf = 50%) produces a safe slope angle recommendation for maximum slope reaches 6 m with a slope angle of less than 60°.
Keywords: Monte Carlo, parameter uncertainty, probability of slope failure.
Legend:Borehole
531
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Frequency
maximum 18,79 26 19
Frequency
3
Unit weight (kN/m )
532
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3.3 Geometry model Based on figure 7, the slope have probability failure (Pf) about
28,8%. Its means that 28,8% x 2000 = 576 have SF < 1, and then the
In this study used 36 variation geometry models.
2000-576 = 1424 have SF > 1. Reliability index () produced (1,176
Slope height (H) : 2.5 m, 4.5 m, 5 m and 6 m, - 1) / 0,28 = 0,62.
Slope angle () : 45o to 85o. (with range 5 o) The probability of failure in all model variations is then plotted
with low risk (Pf = 50%) and high risk (Pf = 25%) in the SRK 2010
3.4 Slope Stability Analysis with Probabilistic Approach standard, that shown in figure 8.
The example result of slip surface obtained from slope stability
analyses is shown in Figure 6. In the example is the result for slope
with height is 6 m and angle is 45°.
Low risk
Height (m)
High risk
Length (m)
Figure 6. Slip surface deterministic analysis (H = 6 m, = 45°)
Figure 8. Probability failure for variations slope model that
Based on figure 6, the safety factor based on deterministic compared with SRK 2010
analysis with mean value of each parameter is 1,185. Then, the
probability of failure from probabilistic analysis using 2000 trial of Figure 8 shows the variation of slope model based on the high
Monte Carlo simulation is shown in probability density function risk impact category with allowed probability of failure up to 25%.
(PDF) and cumulative density function (CDF) in figure 7. Slopes with height 2.5 m are safe to an angle of slope of 85°. Slope
with height 4 m are safe up to 70° slope angle. Slopes with height 5
m are safe up to of 55° slope angle. Slopes with height 6 m are not
safe until the slope angle is 45°. From this result, it is indicated that
the slope angle must be determined to be less than 45°.
For low risk impact category of slope with allowed probability of
failure up to 50%, slopes with height 2.5 m and 4 m, are safe up to
85° slope angle. Slopes with height 5 m are safe up to 70° slope
angle. Slopes with height 6 m are safe with slope angles of less than
60°.
The relationship between the mean safety factor and the
probability of failure is shown in Figure 9 below:
Figure 9 shows that the greater the safety factor, the smaller the
probability of failure. According to Shen (2012), the relationship
between safety factor and probability of failure (Pf) is influenced by
the overlap that occurs between probability density function of
resisting and forces in slope failure. The relationship between safety
factor and Pf is inversely proportional. Figure 9 indicate that the
higher the slope, the variation of the probability failure is greater
Figure 7. Probability of failure (H = 6 m, = 45°)
than the variation in safety factor. The frequency of safety factor of
less than 1.0 for 6 m slopes is greater than 5 m, 4 m and 2.5 m slope.
533
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The relationship between the probability of failure and the Arif, Irwandy (2016). Geoteknik Tambang. Jakarta: PT. Gramedia
reliability index on variations in research slopes is shown in figure Pustaka Utama.
10 below. Azizi. (2011). Aplikasi Probabilistik untuk Analisis Kestabilan
Lereng Tunggal (Studi Kasus di PT. Tambang Batubara Bukit
Asam Tbk. Tanjung Enim, Sumatera Selatan). TPT XX
PERHAPI. Mataram: ResearchGate.
Azizi, M., & Handayani, H. (2011). Karakterisasi Parameter
Masukan untuk Analisis Kestabilan Lereng Tunggal (Studi
Kasus di PT. Tambang Batubara Bukit Asam Tbk. Tanjung
Enim, Sumatera Selatan). Prosiding Seminar Nasional AVoER
ke-3. Palembang: ISBN : 979-587-395-4.
Begemann. (1965). The maximum pulling force on a single tension
pile calculated on basis of results of the adhesion jacket cone.
proceedings of the 6th international conference on soil
mechanics and foundation engineering. Montreal.
Bowles, J.E. (1997). Analisa dan Desain Pondasi, Edisi keempat
Jilid 1, Erlangga, Jakarta.
Christian, J., Ladd, C., & Baecher, G. (1994). Reliability applied to
slope stability analysis. Jounal of Geotechnical Engineering,
2180-2207.
Dirjen Pertambangan Umum (1999). Keputusan Menteri
Figure 10. The relationship between reliability index and probability Pertambangan dan Energi Nomor : 555.K/26/M.PE/1995.
failure Tentang Keselamatan dan Kesehatan Kerja Pertambangan
Umum.
Figure 10 is a scatter diagram between the reliability index as the x GeoStudio. (2004). Stability Modelling with GeoStudio. Canada:
axis and the probability of failure as the y axis. Reliability and GeoSlope International, Ltd.
probability indexes are obtained from the results of slope stability Gibson, W. (2011). Probabilistic Methods For Slope Analysis and
analysis with a probabilistic approach on all variations of the slope. Design. Australian Geomechanics Vol 46 No.3.
Then all the points are made trend line so that the trend line can be Khan, F.S, & Malik, A.A. (2013). Probability and Sensitivity
determined from the point of the plot of the reliability index and Analysis of the Slope Stability of Naulong Dam. Pak. J. Engg
probability of failure of all variations. It is indicated that the & Appl. Sci. Vol 13, 54-64.
reliability index and probability of failure have a relationship; that is Laboratory, S. M. (1983). Buku Data Laporan Penyelidikan Tanah
the smaller the probability of failure, the greater the reliability index. Tiang Transmisi 500 KVA Ungaran-Krian Bagian Jawa
This result is consistent with the theory of the relationship index of Tengah. Semarang: Fakultas Teknik, Universitas Diponegoro.
reliability index and probability of failure in Christian (1994). Trend Palisade. (2016). Risk User Guide. NY USA: Palisade Corporation.
line in the Figure 10 produces an empirical formula of the Santoso, E., Arif, I., & Wattimena, K. (2013). Aplikasi Pendekatan
relationship between reliability index (x) and probability of slope (y) Probabilistik dalam Analisis kestabilan Lereng pada Daerah
as y = 3.2418 x2 - 29.967 x + 47.2226 with a value of R2 = 0.9872. Keidakstabilan Dinding Utara di PT. Newmont Nusa
Tenggara. Prosiding TPT XXII PERHAPI 2013, (pp. 283-
4. CONCLUSION 291).
Shen, H. (2012). Non-Deterministic Analysis of Slope Stability
Based on this case study, it can concluded that:
Based on Numerical Simulation. Faculty of Geosciences,
1. The greater the safety factor, the probability of failure is getting
Geoengineering and Mining of The Technische Universitat
smaller. The smaller the probability of failure, the greater the
Bergakademie Freiberg.
reliability index.
Standar Nasional Indonesia. (1995). SNI 03-3997-1995 : Klasifikasi
2. High risk slope category (SRK, 2010) with a permissible
Tingkat Kecuraman Lereng. Indonesia: Badan Standardisasi
probability slope limit of 25% has a maximum slope
Nasional (BSN).
recommendation up to a height of 5 m with a slope angle of 55°.
Tang, W., & Ang, A.-S. (1984). Probability Concepts in
3. Low risk slope category (SRK, 2010) with allowable slope
Engineering Planning and Design. New York: Wiley:
probability limits of 50% has a maximum slope
Decision, Risk, and Reliability Vol 2.
recommendation up to 6 m high with a safe slope angle of less
U.S. Army Corps of Engineer. (1999). ETL 1110-2-556, Risk-based
than 60°.
analysis in geotechnical engineering for support of planning
studies. http://www.usace.army.mil/publications/eng-tech-
5. REFERENCES ltrs/etl-cw.html, Appendix A, page A11 & A12.
Ang, A., Yucemen, M. S., & Tang, W. (1976). Probability Based Whilliam T dan Whitman, Robert V. (1962). Soil Mechanics.
Short Term Design of Soil Slopes. Canadian Geotechnical Whitman, R. (1984). Evaluating Calculating Risk in Geotechnical
Journal, 201-215. Engineering. J. Geotech Engrg, 110: 143-188.
534
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Perancangan Proteksi Galian Dalam Dengan Soldier Pile. Studi Kasus: Underpass
Pada Pusat Perbelanjaan Di Kawasan Jakarta Selatan
R. Siska1 dan A.L. Ryan2
1
Fakultas Teknik Program Studi Teknik Sipil, Universitas Katolik Parahyangan, Bandung, Indonesia
2
Fakultas Teknik Program Studi Teknik Sipil, Universitas Katolik Parahyangan, Bandung, Indonesia
E-mail: 1siska@unpar.ac.id, 2alexanderlyman95@gmail.com
ABSTRAK: Pekerjaan galian tanah merupakan pekerjaan dalam ilmu geoteknik yang memerlukan perhatian dan penanganan yang cukup
matang dalam perencanaan maupun pelaksanaanya. Pada makalah ini, peneliti ingin mengetahui kondisi kritis dari suatu sistem galian dalam
sebelum terjadi keruntuhan. Galian dalam pada penelitian ini diperkuat dengan struktur penahan tanah berupa soldier pile. Analisis dilakukan
menggunakan metode elemen hingga yaitu program komputer PLAXIS 2D. Gaya dalam dan deformasi merupakan komponen yang
dianalisis dalam makalah ini. Dari hasil analisis, pada kedalaman pemancangan 7,5 m untuk kondisi short term stability dan long term
stability memberikan momen lentur maksimum masing-masing 397.37 kNm dan 662.34 kNm.
Keywords: soldier pile, metode elemen hingga, metode konvensional, underpass, galian dalam.
1. PENDAHULUAN
Ketersediaan lahan parkir yang cukup pada gedung bertingkat tinggi
(high rise building) menjadi hal yang harus diperhatikan dengan
baik oleh pihak developer. Semakin kompleks struktur bangunan
yang akan dikonstruksi, maka semakin besar juga lahan parkir yang
diperlukan untuk menampung kendaraan yang ada. Secara umum,
lahan parkir dapat berada diatas permukaan tanah dan dibawah
permukaan tanah (basement). Pada makalah ini, pekerjaan galian
yang akan dibahas yaitu galian konstruksi underpass pada akses
menuju lahan parkir di salah satu pusat perbelanjaan di Jakarta
Selatan.
Konstruksi underpass tersebut dimaksudkan untuk
mempermudah akses menuju fasilitas lahan parkir pada basement
serta memperlancar arus lalu lintas di sekitar pusat perbelanjaan di
Gambar 1 Konstruksi Underpass dan Titik Bor NBH-22 Pada
Jakarta Selatan. Sistem perkuatan dan proteksi galian tanah sedalam
Lokasi Pembangunan Proyek
kurang lebih 15 meter menggunakan struktur penahan tanah yang
dalam perancangannya turut memperhatikan jenis tanah, sifat tanah,
lokasi sekitar proyek, kedalaman penggalian tanah yang cukup,
3. METODE ELEMEN HINGGA
metode pelaksanaan penggalian, metode analisis serta pemodelan
yang sesuai dan menyerupai kondisi pada lapangan. 3.1 Model Material Tanah
Struktur penahan tanah yang digunakan pada konstruksi
underpass adalah dengan soldier pile. Pemilihan jenis soldier pile Dalam analisis dengan metode elemen hingga, analisis kondisi
dengan sistem pengeboran menjadi pilihan yang baik sebagai sistem parameter tanah secara umum terbagi atas dua, yaitu kondisi total
proteksi galian pada tanah dasar yang relative padat. Disamping itu, stress analysis dan kondisi effective stress analysis yang masing-
pelaksanaan konstruksi soldier pile akan lebih praktis, cepat, dan masing kondisi berdasarkan tipe material dapat dibagi atas Drained
pengontrolan terhadap mutu beton pada tiang akan lebih mudah dan UnDrained.
(GEC, 2013). Pada total stress analysis (TSA), efek dari tekanan air pori atau
Adapaun tujuan dari penulisan makalah ini yaitu: muka air tanah dapat diabaikan dalam analisis. Hal tersebut
1. Memperoleh besaran gaya – gaya dalam pada soldier pile yang disebabkan oleh efek dari air telah tercakup kedalam parameter total
muncul akibat beban kerja arah lateral; stress dimana parameter air dan tanah ditinjau sebagai suatu
2. Memperoleh dan mengontrol defleksi pada soldier pile akibat material tunggal. Sedangkan pada effective stress analysis (ESA),
beban kerja arah lateral; analisis dilakukan dengan memperhitungkan efek dari tekanan air
3. Memperoleh penurunan muka tanah akibat penggalian. pori atau muka air tanah. Hal tersebut disebabkan oleh efek dari air
belum tercakup kedalam parameter effective stress dimana
2. LOKASI PROYEK parameter air dan tanah ditinjau sebagai dua buah material yang
berbeda.
Data yang dipergunakan dalam penulisan makalah ini berasal dari Secara umum, terdapat beberapa kondisi parameter tanah yang
pembangunan konstruksi underpass pada pusat sentra bisnis di dapat digunakan dalam analisis yaitu: UnDrained A, UnDrained B,
kawasan Jakarta Selatan. Pembangunan konstruksi underpass ini UnDrained C dan Drained.
dilakukan setelah bangunan utama yang terdiri atas pusat
perbelanjaan, hotel dan office tower telah selesai dikonstruksi. Tabel 1 Penentuan Parameter Tanah Berdasarkan Tipe Material
Pembangunan konsturksi underpass ini dimaksudkan untuk (Tjie Liong, 2011)
mengurangi kemacetan lalu lintas pada Jl. Kyai Moh. Syafii
Hadzami (Jl. Terusan Gandaria – Jakarta Selatan) yang disebabkan
oleh antrian kendaraan yang hendak masuk menuju fasilitas lahan
parkir basement. Pada penelitian ini, penulis menggunakan data
hasil uji bor dengan kode titik yaitu NBH-22 yang memiliki
kedalaman bor hingga 29.5 m dengan lokasi muka air tanah pada
keadalaman 8 m.
535
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3.2 Penentuan Parameter Tanah Kondisi Total Sterss Analysis 3.4 Model Elemen Hingga dan Tahapan Penggalian
dan Effective Stress Analysis
Proses analisis sistem proteksi galian dalam dilakukan dengan
Berdasarkan tabel 1 desain pada kondisi long term stability, memodelkan sistem galian pada program komputer PLAXIS 2D
parameter tanah pada umumnya akan ditentukan dengan sedemikian rupa mendekati proses konstruksi secara nyata
menggunakan kondisi effective stress analysis. dilapangan yakni dengan membagi pekerjaan konstruksi menjadi
Pada kondisi desain short term stability, parameter tanah pada beberapa tahapan yang dirangkum dalam tabel berikut, yaitu:
umumnya akan ditentukan dengan menggunakan kondisi total stress
analysis. Namun pada penelitian ini, parameter tanah lempung Tabel 4 Pemodelan Tahapan Konstruksi
ditentukan dengan menggunakan kondisi effective stress undrained Tahapan
analysis (ESUA) dengan pertimbangan bahwa terdapat pengaruh Deskirpsi Tahapan Konstruksi
Konstruksi
muka air tanah dan tekanan air pori.
Tahap 1 konstruksi soldier pile dan aplikasi beban kerja
Tahap 2 penggalian pertama hingga kedalaman 1 m
Tahap 3 penggalian kedua hingga kedalaman 3 m
Tahap 4 penggalian ketiga hingga kedalaman 5 m
Tahap 5 penggalian keempat hingga kedalaman 7 m
Tahap 6 penggalian kelima hingga kedalaman 7.5 m
Tahap 7 disipasi tekanan air pori ekses (long term stability)
Adapun model akhir dari tahapan penggalian dapat dilihat pada
gambar berikut:
536
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
537
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
6. DAFTAR PUSTAKA
Geothechinal Engineering Center (GEC), (2013). Manual Pondasi
Tiang. 4th ed. Deep Foundation Research Institute,
Parahyangan Catholic University, Bandung, Indonesia.
Ou, Chang Yu, (2006). Deep Excvation: Theory and Practice.
Taylor & Francis Group, London, U.K.
Rahardjo, Paulus P. (2016). In Situ Testing and Soil Properties
Correlation 2nd Edition. Universitas Katolik Parahyangan,
Bandung, Indonesia.
Tjie-Liong, Gouw. (2011). Deep Excavation Failures, Can They Be
Gambar 8 Bidang Momen Soldier Pile Berbagai Variasi Prevented ?, International Symposium On Sustainable
Kedalaman Pada Kondisi: (a) Short Term Analysis dan (b) Long Geosynthetics and Green Technology for Climate Change 2011
Term Analysis (SGCC 2011), 7-8 December, 2011, Bangkok.
538
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Study of Slope Failure and Reinforcement Analysis to Restore and Increase the Slope
Stability in Flores, East Nusa Tenggara Using Plaxis 2D and 3D
Rendra Priatno1, Ali Iskandar2, and Jo Lian Huat3.
1
Geotechnic Engineer, PT. Tarumanegara Bumiyasa, Jakarta, Indonesia.
2
Director of Engineering, PT. Tarumanegara Bumiyasa, Jakarta, Indonesia.
3
President Director, PT. Tarumanegara Bumiyasa, Jakarta, Indonesia.
E-mail: engineering@tarumanegara.co.id
ABSTRACT: The steeper modification of a natural slope due to construction demand required additional reinforcement to minimize the
slope failure potential. The excavation and removing of sloping soil create new internal instability which means slope failure potential
increase and must anticipated with additional reinforcement. The reinforcement as internally stabilized systems intended to decrease the
stress release during and after excavation activity. Steeper modification of natural slope does need proper geotechnical design and
construction method as well post construction monitoring. In Flores, East Nusa Tenggara, the slope fails after end of excavation with a height
of about 20m and a slope angle of 60o without reinforcement. Internally stress release of slopes increases gradually and causing slope
failures and cracks in the around of collapsing area. To restore the modification slope is required the reinforcement of the slope such as
ground anchor, soil nailing and crib wall as additional support the slope stability. Comprehensive analysis of slope reinforcement is done
using finite element method in Plaxis 2D and 3D program. The design parameters were obtained from the results of the soil investigation in
the study area and field observation. According to slope analysis model for slope failure condition with crack around the collapsing area,
safety factor is less than 1. The process of analysis is done by back fill slope which serves as counter weight then done excavating gradually
and follow by installing the combination of ground anchor, soil nailing, and crib wall. The length of ground anchor and soil nailing is
designed to exceed the slope failure pattern that occurs and ends in the rock layer so that the final SF obtained is more than 1.5 or is in a safe
condition. The stage of construction should be implemented based on the on engineering design of slope stabilization to avoid failure.
Keywords: Slope stability, finite element method, ground anchor, soil nailing.
539
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Slope Failure
Cracks
540
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
With the local collapsing of the slope in the field, then the back
analysis of the soil parameters is reduced on the deformation plane
by adding an interface line on the layer based on the slope failure
and crown patterns that occur around it. The existing safety factor
obtained will be lower than the equilibrium condition (SF = 1).
The reinforcement alternatives used in slope back analysis are
fill material, Ground Anchor, Soil Nailing, Crib Wall, and Gabion.
L Fmax, Fmax,
Length EA
Structure Spasi Type Tens Comp
Layer 1: Gravelly
Sand, N = 36-50
Layer 2: Medium
Weathered – Fresh
Dacite, N = >50
541
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3. The Slope reinforcement phase 1. The installation of ground 5. Slope reinforcement phase 3, the installation of soil nailing and
anchor and crib wall for slopes close to the main road. crib wall down to +9.8m elevation and external load (traffic on
the main road has been running normally).
Figure 10: Slope Reinforcement Phase 2, Top: Plaxis 2D, Bottom: Figure 12: Earthquake load phase, Top: Plaxis 2D, Bottom: 3D.
3D.
542
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
543
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Pile driving using hydraulic hammer will induce large vibration so that the effects of pile driving needs to be analyzed to
avoid building damage and malfunction of the sensitive device. This paper discusses a case study in South Sumatera, Indonesia regarding the
effect of pile driving on a sensitive gas plant instrument and building surrounding the project area. The effect of vibration due to pile driving
is calculated using empirical formulas (Attewel & Farmer, 1973 and J.M. Ko et al, 1990). However, there is correction factor of the empirical
formulas that are dependent to soil types. Continuous monitoring has been carried out in the project area by using geophone sensors to
measure vibrations and PDA (Pile Driving Analyzer) to measure actual energy of driving. Peak particle velocity (PPV) from empirical
formulas are approaching PPV value from monitoring test at the same energy of pile driving. The results will be used to back analyse to
modify the correction factor of the empirical formulas.
Keywords: Peak Particle Velocity (PPV), Pile Driving, Vibration Monitoring.
1. INTRODUCTION
When pile driving operations are carried out near other structures,
several technical investigations must be carried out to evaluate the A
vibration impact of the pile driving. This investigation must take
into account the ground vibration due to pile driving, the BH-08
Sensitive Device
transmission through the soil media to the nearest structure and how
the vibration will affect the structure or sensitive devices. BH-10
In general, the type of vibration propagation through the soil is
more predictable than its effect on the structure. This is because it
BH-01 B’ A’
depends on the quality of the structure, foundation system and B BH-07
configuration. Ground vibrations that can affect the structure include: BH-02 BH-04
Soil settlement and foundation due to compaction. BH-03
Vibration in the structure due to impact vibration.
Vibration in sensitive devices. BH-05 BH-06
Determination of whether the structure will experience a
decrease or horizontal vibration due to pile driving can be estimated. Figure 1. Geotechnical Investigation Location
Identification of local soil conditions can provide an indication of
the potential decline of buildings. The movement of buildings or Soil stratification on this project is represented by Section A-A’
sensitive devices near pile driving location is a function of soil type, and Section B-B’ as follows.
vibration frequency, and amplitude.
Therefore, when the pile is placed near the existing structure or
BH-08 BH-10
sensitive devices, it is the responsibility of the engineer to evaluate
N-SPT N-SPT
its potential for damage to the structure and the devices. For this 0 10 20 30 40 50 60 0 10 20 30 40 50 60
reason the engineer must understand how ground vibrations are 0
15
0
3
caused by the pile-up process and how soil parameters such as 4
2
5 5 5
damping can give effect. 11
9
4 Silty Sand 11
On this paper, case study is taken from a project located in South 10 13 10 9
22
Sumatera, Indonesia. This project uses a type of pile foundation with 9
12
Depth [m]
Depth [m]
15 13 15
a 400 mm diameter spun pile. The hydraulic hammer for driving that 10
46
27
will be carried out has vibration issues that have the potential to 20
11
20 17
20
interfere with the performance of a sensitive instrument that is near 60
60 60
that location and existing structure. The closest distance between the 25 Coal 60
60 25 60
60
60 60
pile and the instrument is about 186 m. Vibration monitoring is Carbonaceous 60 51
30 60 30
carried out to estimate the amount of vibration arising from the pile Clay 60 54
Clay
60
driving of a 400 mm spun pile using the hydraulic hammer. Besides 35 35
that, PDA monitoring was conducted to determine the actual energy
40 40
of the hammer that entered the pile foundation during pile driving.
Figure 2. Section A-A’
2. PROJECT DATA
Soil stratification on section A-A’, which are represented by
2.1 Soil Condition BH-08 and BH-10, shows that around 22 m of the top layer is loose
to medium silty sand. Layer 2 is coal with 6 m thickness. Layer 3 is
Before analyze the vibration effect due to pile driving, it is a dark brown brittle carbonaceous clay. Layer 4 is a layer of clay
important to understand the soil conditions at the project site. with very hard consistency. In this section, it is shown that coal
Geotechnical investigation is consists of drill testing and SPT layer tends to be uniform even though there are different elevation.
(Standard Penetration Test). The location of the test is shown on
Figure 1.
544
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Depth [m]
Depth [m]
Depth [m]
Depth [m]
45 9
15 33 15 15 14 15 15 15
60 12
52 Coal 60 12 9
46 11
20 20 20 14 20 20 22
60 25 20
55 14
49 Clay 41 Coal 52 19
46
25 35 25 30 25 50 25 17 25
55 27
60 50 26
Carbonaceous Clay 59 45 60
60 Clay 51 60 Coal 60
30 30 30 30 60 30 7
60 36
60 60 30
60 60 60 60
35 35 51 35 41 35 35
Clay 60
54 51 60 Clay
Clay 52
60 52
40 40 40 40 40
Section B-B’ shows different soil stratification from Section A- Figure 4. Correlation between Distance from Piling Operation and
A’ where the coal depth is vary. In the upper layer there is sandy silt Peak Particle Velocity (Eurocode 3)
and silty sand. The layer below is a coal layer that is brittle with
thicknesses ranging from 6-10 m at various depths. The layer below The other empirical formulas derived by J.M. Ko et al. (1990)
is carbonaceous clay. The layer is found between the layers of hard are as follows:
clay. Hard soil layers are found at different depths in each area.
exp α (r 1)
1
v = 70 (2)
2.2 Piling Method r
where :
The foundation system on this project is 400 mm diameter spun pile. v = particle velocity (mm/sec)
The pile driving will be done with a hydraulic hammer with hammer r = distance from the vibration source to the location
specifications as follows. observed (m)
Hammer type : NH-70 α = damping coefficient (depends on the soil), about 5-20%
Hammer weight : 7 tons
Energy efficiency : 60-70%
Blow speed : 25 blows / minute 4. STUDY ON THE EFFECTS OF VIBRATION BY
Hydraulic pressure : 210 kg/cm3 CONTINUOUS MONITORING
Piling with a hydraulic hammer will induce great vibration so
that the vibration impact of the piling done needs to be analyzed so 4.1 Vibration Monitoring
that the pile driving does not cause damage to the building, Vibration monitoring is carried out by a geophone sensor as an
malfunctioning of the existing gas plant instrument, and failure of accelerometer. The geophone is used to recording particle velocity
concrete settings after pouring concrete on the need for casting the (v) and frequency (f). The data is displayed against time so that
shallow foundation in the surrounding area. variations in particle velocity or frequency with time can be
obtained along piling. The reading of particle velocity and frequency
is done in 3 axis directions (x, y, and z) so that analysis can be
3. STUDY ON THE EFFECTS OF VIBRATION BY carried out on the three directions. From the measurement of the
EMPIRICAL FORMULAS three directions, the resultant magnitude of the three directions will
Peak particle velocity (PPV) values can be calculated by empirical be calculated.
formula from Attewel & Farmer (1973) or J.M Ko (1990). Attewell Vibration monitoring is carried out using 2 sensors. The first
and Farmer (1973) conducted an analysis of the results of tool is used to measure vibration in the sensitive device (existing
measurements of vibration due to the pile driving on several GTC). The second tool measures in 4 different directions with
variations of soil and pile type. They suggested that the peak particle varying distances to obtain a vibration relation to the distance from
velocity (v) due to pile pile driving could be predicted with the the point of piling. From the test at each point, the particle
following formula. displacement, particle velocity, acceleration, and frequency will be
obtained. The monitoring results will be concentrated on the final
Wo set condition as a critical condition.
v= C (1)
r Pile driving is carried out in 2 stages. In the first stage the piling
where : is located at 200 m from the existing GTC where the vibration
v = peak particle velocity (mm/s) measurements were carried out at distances varying from the point
C = constant value of pile driving (30 m, 40 m, 60 m, 90 m, 115 m, 120 m, and 200 m).
Wo = hammer energy (Joule/blow) The measurements at various distances is used to evaluate safe
r = distance of the point of purging to the location being distance. Pile driving in the second stage was carried out at New
reviewed (m) GTC Unit 1 which was as far as 186.19 m from the existing GTC.
The equation states that the partivle velocity at any point is The pile that was placed on the New GTC Unit 1 was the location
proportional to the square root of hammer energy and inversely with the closest distance between the location of the existing GTC.
proportional to the distance of a point at the stake location. This In other words, the measurement of vibration due to pile driving at
vibration reduction has a linear tendency to distance in the the New GTC Unit 1 will obtain vibration data in critical conditions.
logarithmic scale. The proposed equation is then adopted by The data to be used in this paper only refers to the data taken in the
Eurocode 3. The constant C is used to express the variation in soil final set condition.
conditions and type of pile type. The value of C ranges from 0.25- The location of stage 1 piling point (loc pilling) and stage 2
1.50. Eurocode 3 recommends a value of 0.7 for piling vibratory. At piling point (New GTC Unit 1) to the existing GTC location can be
impact piling, the value of C is around 1.0 for solid or firm soil and seen in Figure 5 and Figure 6. The value of peak particle
0.5 for loose or soft soil. Here is the relationship of the distance of displacement, peak particle velocity, peak particle acceleration, and
one point from the location of the location of the peak particle the maximum frequency measured from monitoring is shown in
velocity according to Eurocode 3. Table 1 and Table 2.
545
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 5. Piling and Existing GTC Location The pile driving was carried out with a hydraulic hammer where
7 tons ram weight and 0.32 m ram stroke, so that theoretical energy
was obtained at 2.24 t.m. PDA test equipment shows that with the
same hammer weight and falling height the actual energy is in the
range of 1.2 - 2.01 (see Figure 7) and is taken on an average of
1.501. Energy efficiency was taken based on the actual average
energy divided by theoretical energy, so that the energy efficiency
E D A
obtained in the hydraulic hammer was obtained by 67%.
(105 meter) (90 meter) (30 meter)
PILING LOC
1.5
Figure 6. Vibration Monitoring Location
1.0
546
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
included in the safe category. Nevertheless, these criteria are only The amount of energy is determined based on theoretical energy,
used to determine the structural conditions of the building so that it which is 2.24 t.m. With a large energy of 2.24 t.m or equal to 22400
cannot be used for evaluation of other needs such as the effect of Joule, a modification of the Attewel & Farmer equation is obtained
pile driving on the gas plant instrument. for this project which can be seen in the following equation.
10
SERIOUS
240
CRACKING
9
220
v = 15 Wo (5)
MAJOR r 1.75
8 DAMAGE 200
where :
7
DAMAGE 180
100.00 Energy = 2.24 t.m criteria. Radius of vibration influence due to pile driving on this case
= 22400 Joule
is 300m.
10.00 7. CONCLUSION
The case study is taken from South Sumatera, Indonesia
1.00
regarding the effect of pile driving on a sensitive gas plant
instrument and building surrounding the project area. Soil
Hydraulic condition in the project area is consists of silty sands in the
Hammer
upper layer, coal and carbonaceous clay in the middle, and the
0.10
1 10 100 1000
underlying layer is hard clay.
Distance [m]
The effect of vibration due to pile driving is calculated using
Attewel & Farmer (1973) J.M.Ko (1990) Data Monitoring
empirical formulas (Attewel & Farmer, 1973 and J.M. Ko et al,
Figure 9. Relationship Between Distance from Piling Operation and 1990). However, there is correction factor of the empirical
Peak Particle Velocity from Empirical Formulas and Continuous formulas that are dependent to soil types. To overcome the
Monitoring uncertainty of the correction factor, continuous monitoring by
geophone sensors to measure vibrations and PDA (Pile Driving
Based on Figure 9, it can be seen that the relationship between Analyzer) to measure actual energy of driving.
peak distance and particle velocity (PPV) from the results of The pile driving is done using a hydraulic hammer which uses
monitoring in the field forms a linear line in the logarithmic scale. NH-70 with 7 tons ram weight and 0.32 m ram stroke. The
The relationship of distance and PPV obtained by empirical theoretical energy obtained is 2.24 t.m while the average actual
formulas from the Attewel and Farmer (1973) forms a linear line energy from the PDA test is 1,501 t.m, so that the energy
whereas from the J.M. Ko (1990) does not form a linear line on a efficiency of the hydraulic hammer is 67%.
logarithmic scale. With the similarity of the linear function form, it In existing GTC, the maximum vibration acceleration allowed is
is necessary to do a back analyze the empirical formula from 0.4 mm/s2 and the maximum allowable frequency is 1.2 Hz. The
Attewel and Farmer (1973) based on the results of monitoring as a maximum allowable particle velocity is 0.33 mm/s. Based on the
reference for determining the PPV value for various distances on monitoring data, it was found that the pile driving was carried
certain energy quantities. out at the closest distance, peak acceleration, and particle
547
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
velocity that occurred in the GTC area was smaller than Eurocode 3, (1992). “Design of steel structures, chapter 5, piling.”
required. The frequency that occurs exceeds the required DD ENV 1993-1-1:1992
frequency. However, due to small particle switching, the FHWA, 2016, “Design and Construction of Driven Pile Foundations
existing GTC is in a safe condition. – Volume I” FHWA-NHI-16-009
Radius of vibration influence due to pile driving on this case is Ko, J.M., Luk, S.T., and Cheng, C.Y., (1990), “Vibration and Noise-
300m. Measurement Prediction and Control”, Proceedings of
Australian Vibration and Noise Conference 1990.
PT GEC, 2017, “Analisis Vibrasi Akibat Pemancangan, Sumatera
8. REFERENCES Selatan”, Final Report
PT GEC, 2017, “Laporan Hasil Monitoring Vibrasi dan PDA Test”,
Attewell, P. B. & Farmer, I. W., (1973) “Attenuation of Ground Final Report
Vibrations from Pile Driving”, Ground Engineering, vol. 3, Wiss, J.F., 1967.“Damage Effects of Pile Driving Vibration”,
No. 7, July. Highway Research Record, Number 155, P.14-20
Deep Foundation Research Institute (DFRI). (2017) “Manual Wiss, J.F., 1981, “Construction Vibrations: State-of-the-Art”,
Pondasi Tiang Edisi 5”. Universitas Katolik Parahyangan. Journal of the Geotechnical Engineering Division, ASCE,
Centre of Excellence for Geotechnical Engineering Vol. 107, No. 6T2, pp. 167-181, February,.
(GEOCENTER)
548
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Pengaruh Tahanan Ujung Tiang dan Faktor Aman Terhadap Lendutan Pelat
Terpaku Tiang Tunggal Pada Lempung Lunak
P. Anas1, O. Ferly2, M. Roza3
1
Pengajar Program Studi Teknik Sipil,Universitas Islam Riau, Riau, Indonesia
2
Mahasiswa Program Sarjana Teknik Sipil, Universitas Islam Riau, Riau, Indonesia
E-mail: anaspuri@eng.uir.ac.id1, ferlyoktavia@student.uir.ac.id2, rozamildawati@eng.uir.ac.id3
ABSTRAK: Perkerasan Sistem Pelat Terpaku merupakan salah satu alternatif solusi untuk mengatasi permasalahan konstruksi jalan yang
melalui tanah lunak. Kenaikan kapasitas dukung tanah dasar akibat pengaruh dukungan tiang-tiang pada pelat akan mengurangi kebutuhan
tebal perkerasan beton dan memperkaku sistem perkerasan. Kapasitas dukung tiang tersebut kontribusi dari tahanan ujung dan tahanan gesek.
Lazimnya tahanan ujung diabaikan untuk tiang berada pada tanah lunak. Penelitian ini bertujuan untuk mempelajari pengaruh tahanan ujung
tiang dan faktor aman (SF) terhadap lendutan pelat terpaku tiang tunggal. Dimensi struktur pelat terpaku dan data material mengacu pada
Puri (2017). Pelat terpaku tiang tunggal mempunyai ukuran pelat 120 cm x 120 cm, tebal 15 cm, diperkuat dengan tiang bulat beton masif
tunggal diameter 20 cm, dan panjang tiang 150 cm. Sistem ini dibebani dengan pembebanan tekan sentris. Perhitungan lendutan dan gaya
dalam pelat menggunakan metode BoEF (Beam on Elastic Foundation), berdasarkan tambahan modulus reaksi tanah dasar ekivalen yang
dibandingkan dengan lendutan hasil pengamatan. Hasil analisis menunjukkan bahwa besaran faktor aman mempengaruhi nilai lendutan
hitungan. Peningkatan SF cenderung meningkatkan lendutan hitungan. Penggunaan SF = 1,0 sudah memadai. Perhitungan lendutan dengan
memperhitungkan tahanan ujung tiang (Qb> 0) lebih baik dibanding tanpa memperhitungkan tahananan ujung tiang (Qb = 0). Momen pelat
memberikan perbedaan tidak signifikan untuk hitungan tanpa atau dengan memperhitungkan tahanan ujung tiang.
Keywords: perkerasan kaku, pelat terpaku, diameter tiang, reaksi subgrade, lendutan pelat.
1. PENDAHULUAN tanah dan tiang, karena penurunan relatif tanah dan tiang kecil maka
tahanan gesek tiang belum sepenuhnya optimum, sehingga
Perkerasan jalan di atas tanah lunak sering dibangun menggunakan digunakan pendekatan faktor perpindahan tiang (α) dalam
perkerasan kaku. Lapis perkerasan tersebut langsung menumpu pada menentukan modulus reaksi subgrade (∆k). Menurut Puri, dkk
tanah dasar lunak atau di atas lapis fondasi jalan. Perkerasan jalan (2012), sulit untuk menentukan faktor perpindahan tiang (α), serta
terletak pada tanah dasar yang tidak stabil seperti lempung lunak mengusulkan sebuah pendekatan dengan mempertimbangkan
atau tanah ekspansif, banyak menimbulkan masalah pada kinerja penurunan toleransi perkerasan kaku. Dalam menentukan kapasitas
jangka panjang perkerasan (Hardiyatmo, 2014). Perkerasan jalan dukung tiang diizinkan biasanya diambil 1/2,5 kapasitas ultimit.
yang terletak pada tanah dasar lunak atau ekspansif, sering Nilai 2,5 adalah faktor aman, dengan mengambil α = 1/SF = 1/2,5
mengalami deformasi yang berlebihan oleh beban kendaraan berat, dengan asumsi perpindahan kepala tiang sama dengan penurunan
sedangkan menurut Hardiyatmo (2012), metode perancangan toleransi perkerasan pelat kaku (δp = δa; maka β = 1 / δa).
perkerasan jalan yang telah ada, umumnya diasumsikan bahwa Analisis variasi faktor aman pada modulus reaksi subgrade
tanah-dasar dalam kondisi stabil, sehingga tebal komponen struktur ekivalen untuk model skala penuh Pelat Terpaku satu baris tiang
perkerasan hanya didasarkan pada daya dukung tanah-dasar yang dengan pembebanan tekan sentris telah dilakukan Puri (2017).
dinyatakan oleh nilai CBR atau modulus reaksi subgrade vertikal. Sistem tersebut berada pada media tanah lempung lunak. Belum
Perkerasan jalan tidak hanya menerima beban siklik oleh kendaraan, dilakukan analisis penerapan variasi faktor aman pada Pelat Terpaku
namun juga menerima beban akibat temperatur yang membuat tiang tunggal dengan memperhitungkan tahanan ujung tiang.
perkerasan jalan mengalami momen lentur bolak-balik (Puri, 2015). Penelitian ini akan menerapkan nilai modulus reaksi subgrade
Sistem Pelat Terpaku (Nailed Slab System) yang ditemukan oleh ekivalen (k’) dengan pendekatan variasi faktor aman pada nilai
Prof. Dr. Ir. Hary Christady Hardiyatmo, M. Eng., DEA. (2008) tambahan modulus reaksi subgrade (∆k) dan pengaruh tahanan
adalah suatu perkerasan beton bertulang (tebal antara 12 - 20 cm) ujung tiang.
yang didukung oleh tiang-tiang mini (panjang 150 - 200 cm dan
diameter 15 – 20 cm). Tiang-tiang dan pelat beton dihubungkan
2. LANDASAN TEORI
secara monolit dengan bantuan tulangan-tulangan. Interaksi antara
pelat beton-tiang-tanah disekitarnya menciptakan suatu perkerasan 2.1 Modulus Reaksi Subgrade Ekivalen (k’)
yang lebih kaku, yang lebih tahan terhadap deformasi tanah-dasar.
Sistem ini merupakan salah satu alternatif solusi untuk mengatasi Modulus reaksi subgrade ekivalen merupakan salah satu cara
permasalahan konstruksi jalan yang melalui tanah lunak. Sistem analisis yang digunakan pada perkerasan sistem pelat terpaku.
Pelat Terpaku cocok digunakan untuk perkerasan yang tanah Modulus reaksi subgrade ekivalen adalah modulus reaksi akibat
dasarnya dipengaruhi oleh penurunan tidak seragam, karena adanya tiang-tiang beserta pelat. Nilai modulus reaksi subgrade
interaksi tanah-tiang-pelat membuat pelat lebih kaku, sehingga ekivalen (k’) diperoleh dengan menjumlahkan nilai modulus reaksi
mengurangi terjadinya beda penurunan permukaan perkerasan subgrade tanah (k) dan nilai tambahan dari modulus reaksi dari tiang
(menciptakan permukaan perkerasan beton yang selalu rata) tunggal (Dk). Hardiyatmo (2011) mengusulkan metode penentuan
(Hardiyatmo, (2016). Pelat berfungsi ganda yaitu sebagai struktur tambahan modulus Dk. Puri, dkk, (2012b) menyederhanakan metode
perkerasan sekaligus sebagai pile cap, sedangkan tiang-tiang selain yang diusulkan oleh Hardiyatmo tersebut dengan
berguna untuk menaikkan daya dukung tanah-dasar, juga menjaga mempertimbangkan penurunan izin dari perkerasan. Istilah modulus
agar pelat beton tetap kontak dengan baik pada lapis pondasi bawah. reaksi subgrade ekivalen (k’) digunakan untuk membedakan
Perancangan Pelat Terpaku dapat menggunakan Metode modulus reaksi subgrade klasik (yaitu dari uji beban pelat standar)
Hardiyatmo Modifikasi (Puri, dkk., 2012). Metode ini merupakan dan modulus subgrade pelat yang sudah dipengaruhi oleh dimensi
salah satu metode dalam menentukan nilai modulus reaksi subgrade pelat yang lebih lebar dan perlawanan kelompok tiang yang
ekivalen, dan merupakan modifikasi dari metode Hardiyatmo mendukung pelat.
(2011). Oleh karena tiang berada pada tanah lunak dan dimensi Pada sistem pelat terpaku, pendekatan penentuan modulus reaksi
tiang kecil maka tahanan ujung dibaikan (Hardiyatmo, 2011; Puri, subgrade ekivalen (k’) didefinisikan sebagai modulus reaksi
dkk. 2012, Puri, 2015). Menurut Hardiyatmo (2011), penurunan subgrade yang memperhatikan konstribusi tiang dalam menahan
pelat beban lebih besar dari pada penurunan atau perpindahan relatif gesekan perpindahan vertikal pelat ditentukan sebagai berikut
549
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
(Hardiyatmo, 2011; Dewi, 2009; Puri, et.al., 2011b, 2012a dalam modulus reaksi subgrade karena memperhitungkan tahanan ujung
Puri, dkk, 2013): tiang (Qb) dinyatakan oleh persamaan berikut:
k ' k k (1) 0 , 4 (Q S Q b )
k (5)
dengan : a A PS
k’ = modulus reaksi subgrade ekivalen pelat terpaku (kN/m3)
k = modulus reaksi subgrade dari tanah (kN/m3) atau,
Δk = tambahan modulus reaksi subgrade karena adanya tiang
(kN/m3) 0 ,4 ( f S A S f b Ab )
k (6)
a A PS
2.2 Reaksi Subgrade
Koefisien reaksi subgrade adalah hubungan konseptual antara Sehingga modulus reaksi subgrade ekivalen dengan
tekanan tanah dengan defleksi (Bowles, 1982 dalam Muhu, 2007). memperhitungkan tahanan ujung tiang dapat dihitung dengan:
Salah satu besaran yang dapat digunakan dalam analisis lendutan
pelat adalah koefisien reaksi tanah dasar. Rumus dasar perhitungan 0 , 4 ( f S A S f b Ab )
k' k (7)
nilai koefisien subgrade (kv) untuk pelat kaku adalah: a A PS
Untuk pelat terpaku yang berada di atas tanah lunak, maka
q
kv (2) tahanan ujung tiang diabaikan. Dalam hal ini tambahan modulus
reaksi subgrade karena adanya tiang di bawah pelat ditentukan
dengan: (Puri, dkk, 2012a; Puri, 2015) sebagai berikut:
q = tekanan tanah (kN/m2)
δ = lendutan yang terjadi (m) 0 ,4 a d c u A S
k (8)
a A PS
Menurut Puri, dkk (2013), koefisien reaksi tanah dasar arah
vertikal (kv) dapat digunakan dalam hitungan lendutan. Koefisien ini Sehingga modulus reaksi subgrade ekivalen dapat dihitung dengan :
ditentukan sebagai tekanan fondasi (q) yang dibagi dengan
penurunan yang bersesuaian (δ) dari tanah di bawahnya. Dengan 0,4a d cu AS
kata lain, reaksi tanah dasar tidak lain adalah distribusi reaksi tanah k' k (9)
(q) di bawah struktur fondasi rakit guna melawan beban fondasi. a APS
Reaksi subgrade terdistribusi tidak linier akibat beban merata Faktor adhesi ad dapat ditentukan berdasarkan kurva Tomlinson
fondasi. Pada lempung, distribusi reaksi tanah berbentuk cembung pada Grafik McClelland (1974) dengan menggunakan nilai cu
dengan reaksi maksimum di sekitar pinggir fondasi dan reaksi yang (Gambar 1).
lebih kecil pada tengah-tengah fondasi.
Modulus reaksi subgrade dari uji beban pelat (k) biasanya
menggunakan pelat lingkaran dan harus dikoreksi terhadap ukuran
dan bentuk pelat dari pelat terpaku (Puri, dkk, 2012). Koreksi
ukuran dan bentuk pelat mengacu pada metode Das (2011) seperti
pada persamaan berikut:
Koreksi terhadap ukuran pelat dinyatakan sebagai :
0,3
k v k 0, 3 (3)
B
Koreksi terhadap bentuk pelat dinyatakan sebagai :
0,5B
k ( B B ) 1
kv L (4)
1,5
dengan: Gambar 1. Faktor adhesi untuk tiang pancang pada tanah lempung
k = modulus reaksi tanah dasar dengan lebar B (kN/m3), (McClelland, 1974)
k0,3 = modulus reaksi tanah dasar pelat beban 0,3 m x 0,3 m
(kN/m3), Mobilisasi perlawanan gesek tiang masih dalam zona elastis
B = lebar pelat (m), (Puri, dkk (2012). Menurut aturan praktis dalam menentukan
kapasitas dukung tiang diizinkan biasanya diambil pada 1/2,5
= modulus reaksi tanah dasar pelat beban B x B (kN/m3),
kapasitas ultimit. Dimana 2,5 adalah faktor aman (SF). Dengan
L = panjang pelat (m). mengambil α = 1/SF = 1 /2,5 dengan asumsi perpindahan kepala
tiang sama dengan penurunan toleransi perkerasan pelat kaku (δp =
δa; maka β = 1 / δa). Oleh karena itu, tambahan modulus reaksi
2.3 Modulus Reaksi Subgrade Ekivalen Metode Hardiyatmo
subgrade karena adanya tiang dibawah pelat dimodifikasi dengan
Modifikasi
(Puri, dkk. 2012) :
Metode Hardiyatmo (2011) menggunakan pendekatan reduksi
tahanan tiang dalam penentuan Dk. Perpindahan relatif antara tanah f S AS (10)
dan tiang, dan reduksi tahanan tiang telah diperhitungkan. Untuk k
2,5 a APS
keperluan desain praktis, penentuan reduksi tahanan tiang sulit
diperhitungkan. Puri, dkk, (2012a) mengusulkan suatu pendekatan Dan persamaan (3.10) tersebut ditulis menjadi (Puri, 2015) :
dalam penentuan tambahan modulus reaksi subgrade dengan
f S AS (11)
menggunakan penurunan izin/toleransi pelat perkerasan. Tambahan k
SF a APS
550
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
f S AS (12)
k' k
SF a APS
Dengan fs : tahanan gesek satuan tiang (kN/m2), fs = ad.cu; ad :
faktor adhesi, c : kohesi (kN/m2), δa : toleransi penurunan pelat
perkerasan kaku sebesar 5 (mm), As : luas selimut tiang (m2), Aps:
luasan zona pelat yang didukung oleh satu tiang (m2), Aps= s2, s :
jarak antar tiang (m), dan 2,5 : nilai faktor aman (SF).
Di lapangan sistem ini diaplikasikan untuk perkerasan kaku
bertulang menerus (CRCP). Untuk menjaga agar pelat beton tidak
mengalami regangan tarik berlebihan maka lendutan toleransi (δa)
maksimum 5 mm atau 1/256 bentangnya (Puri, 2015). Balok lantai
biasanya dibatasi hingga lendutan maksimum 1/360 bentangnya
agar menghindari keretakan pada beton menurut Singer (1985).
3. Metode Penelitian
3.1 Metode Pengumpulan Data
Penelitian dilakukan menggunakan data sekunder hasil uji skala
penuh Sistem Pelat Terpaku tiang tunggal pada lempung lunak
(Puri, 2017). Gambar 2 adalah skema uji dengan sifat-sifat tanah dan
beton ditunjukkan pada Tabel 1. Kolam uji berukuran 7 m x 3,54 m
x 2,5 m yang diisi lempung lunak setebal 2,15 m. Pelat terpaku tiang
tunggal terdiri atas satu buah tiang masif beton bertulang
berdiameter 0,20 m dan panjang 1,5 m. Tiang ini terhubung secara
monolit pada pelat beton bertulang di atasnya. Pelat berukuran 1,2 m
x 1,2 m dengan tebal 0,15 m dimana di bawah pelat juga diberi
lantai kerja setebal 0,05 m. Penjelasan konstruksi dan metode
pelaksanaan uji Pelat Terpaku dapat merujuk pada Puri (2017). Gambar 2. Perkerasan Sistem Pelat Terpaku Tiang Tunggal (Puri,
2017).
3.2 Tahapan Pelaksanaan Penelitian
3.3 Cara Analisis
Penelitian dimulai dengan pengumpulan data sekunder hasil uji
skala penuh Sistem Pelat Terpaku tiang tunggal pada lempung lunak Analisis dilakukan semi-manual yang didasari pada perkerasan pelat
(Puri, 2017), kemudian perhitungan modulus reaksi subgrade terpaku tiang tunggal, perhitungan nilai modulus reaksi subgrade
ekivalen (k’) menggunakan Metode Hardiyatmo Modifikasi dengan ekivalen (k’) dengan tambahan modulus reaksi subgrade (∆k)
variasi faktor aman. Selanjutnya, perhitungan berdasarkan lendutan Metode Hardiyatmo Modifikasi sebagaimana Persamaan (1) dan (8),
pengamatan (Puri, 2015) sebagai lendutan izin (δa = δs). dan perhitungan lendutan dan gaya dalam dilakukan dengan bantuan
Perhitungan lendutan dan gaya dalam menggunakan metode program “BoEF.xls” Program Versi 1.4.
BoEF (Beam on Elastic Foundation) dengan bantuan program
“BoEF.xls” Program Version 1.4 dengan meng-input nilai modulus Table 1 Sifat-sifat Lempung dan Beton (Puri, 2017)
reaksi subgrade ekivalen (k’). Hasil analisis berupa lendutan pelat
dan gaya dalam berupa momen. Dari pembahasan dapat No. Parameter Unit Rerata
A. Lempung Lunak
disimpulkan pola lendutan akibat penerapan nilai modulus reaksi
1. Berat Spesifikasi, Gs - 2,55
subgrade ekivalen (k’) Metode Hardiyatmo Modifikasi dengan
2. Batas-batas Konsistensi :
pendekatan variasi faktor aman, serta membandingkan pola lendutan - Batas Cair, LL % 88,46
perhitungan saat Qb = 0 dan Qb > 0 berdasarkan lendutan - Batas Plastis, PL % 28,48
pengamatan (Puri, 2017) sebagai lendutan izin (δa = δs) dan lendutan - Batas Susut, SL % 9,34
izin maksimum (δa = 5 mm). Lebih jelasnya tahapan-tahapan - Indeks Plastisitas, PI % 59,98
penelitian dirangkum ke dalam sebuah bagan alir sebagaimana - Indeks Cair, LI % 0,36
Gambar 3. 3. Kadar Air, w % 54,87
4. Kadar Lempung % 92,93
5. Kandungan Pasir % 6,89
6. Berat Volume Basah, kN/m3 16,32
7. Berat Volume Kering, d kN/m3 10,90
8. Kuat Geser Undrained, Su
- Tak Terganggu kN/m3 20,14
- Remolded kN/m3 11,74
9. CBR % 0,83
10. Klasifikasi Tanah :
- AASHTO - A-7-6
- USCS - CH
B. Beton
11. Pelat beton bertulang :
- Kuat tekan MPa 29,21
- Kuat tarik kPa 3.13
12. Kuat tekan beton tiang MPa 17,4
13. Kuat tekan lantai kerja MPa 14,5
551
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
552
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
553
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This paper describes a case study of a bi-directional load test on a working pile located at limestone formation area. The test
pile was instrumented with Distributed Fibre Optics Strain Sensors (DFOS) to measure the change in strain and to determine the pile shaft
friction and end bearing. This paper highlights the advantages and limitations of DFOS in measuring the continuous strain profile of a test
pile. Interpretation on the anomalies detected through the DFOS results is discussed. The paper aims to introduce to the industry, the superior
information obtained using the innovative fibre optic technology for geotechnical testing and monitoring.
Keywords: Bidirectional Static Load Test, Distributed Fibre Optic Strain Sensor, pile load test.
2. GENERAL PRINCIPLES
This section describes the general principles of bidirectional static
load test (BDSLT) and distributed fibre optic sensor (DFOS).
2.1 Bidirectional Static Load Test (BDSLT)
Figure 1 Typical Schematic Diagram of Bi-Directional Static Load
Rapid urbanization process is now a global phenomenon in response Test
to a fast-growing population and development. Construction of
skyscrapers and mega infrastructure projects can be seen in many In a conventional MLT, sometimes it is not easy to determine
developing countries to serve specific functions in meeting the the ultimate shaft friction and end bearing of rock socket in a bored
urban need and to improve the quality of life. The construction of pile, as the magnitude of top down load significantly reduces before
these gigantic structures in urban areas with space constraints often it reaches the pile toe to fully mobilizing the bored pile. In order to
adopts very long bored piles of large diameter, in order to limit the optimize the design of a rock socketed bored pile, designers are keen
number of piles within a pile group. Therefore, tremendous loads to determine/verify the ultimate shaft friction and end bearing of the
will be imposed on these piles. rock layer. This can be achieved relatively easier in a BDSLT as the
554
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
jack can be placed in the rock layer. The load will be directly measured for many years, hence the DFOS is also very good for
exerted on the rock layer before it is transferred to the soil layer. long term monitoring. Finally, as compared to conventional sensors,
Another advantage of BDSLT is the improved safety as the the DFOS system provides a continuous strain measurement along
sacrificial jack is embedded within the pile body located deep down the pile rather than discrete measurements at several designated
in the ground. There is no need for platform preparation or stacking points.
of concrete blocks on the bored pile and this can eliminate the risk The limitations of the DFOS system is that it is not suitable for
of collapse due to soft ground. Besides, for BDSLT, there is no need dynamic measurements as the current technology requires several
to extend the pile up to ground level (especially for pile with deep minutes to obtain a complete measurement. Also, the cost of the
cut off level), hence there is a saving in time and cost. The bored analyzer is high as compared to conventional loggers. However, the
pile can be cast up to cut off level only as the jack exert forces price for DFOS system will surely be more competitive when the
within the pile body. system is more commonly adopted.
Without the requirements for massive reaction systems, BDSLT
can be performed with minimal space and is able to significantly
reduce time, cost and risk.
3. METHODOLOGY
This section describes the geological information, pile information
and the procedures of testing. Criteria for the termination of load test
is also included in this section.
Figure 2 Principle of BOTDA sensor system (Lan et al, 2012) 3.1 Geological Information
The main strengths of fibre optic sensors are the nature of glass This paper describes the case study of an instrumented pile load test
which is immune to electromagnetic interference, corrosion resistant using DFOS in a residential building project. The test pile is a
and inert to chemical reactions. The FO sensors are very versatile working pile located at limestone formation. The bidirectional static
and can be safely used in many harsh environments such as high- load test aims to verify the geotechnical design parameters for the
voltage, marine or explosive environments. Besides, the DFOS highly fractured limestone formation.
cables are thin and therefore simplifies the installation on the The soil profile consists of loose to medium dense silty clay,
reinforcement cage without disrupting the concreting procedures. with SPT-N values range from 0 to 28 as shown in Figure 5. The
The DFOS cables embedded in concrete piles are durable and can be limestone layer was found at approximately 61 m below ground
555
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
level. The Rock Quality Designation (RQD) of the limestone the reinforcement cage at four sides, at 90-degree interval around
samples are all 0 %, indicating highly fractured limestone, with total the circumference of the reinforcement cage.
core recovery (TCR) ranging from 47 % to 100 % as illustrated in
Figure 5. 3.3 Testing Procedures and Criteria
The BDSLT was conducted in accordance to ASTM
D8691/D8691M-18 (Standard Test Methods for Deep Foundations
Under Bi-Directional Static Axial Compressive Load) and in
accordance to load schedule and requirements provided by the
design engineer. The test pile is a working pile designed to a factor
safety of 2. Therefore, the test load of the pile is twice the working
load (WL=12,800 kN and TL=25,600 kN). The loading sequence is
summarized in Table 1.
The increment or decrement of loading can only be executed if
the rate of the pile settlement is not more than 0.0625 mm per 15
minutes, or 0.25 mm per hour. If this criterion is not met, the
maintaining time of the loading stage shall be extended.
4. DATA ANALYSIS
This section describes the strain profile, pile shaft friction and the
pile movement. The potential causes of the low shaft friction is also
discussed in this section.
4.1 Strain Profile
Figure 7 shows the continuous axial strain profile at several loading
stages, i.e. 60%, 100%, and 150% of the working load (WL). Along
the pile, the DFOS measurement indicates consistent strain
increment as the load increased. It is interesting to observe that at
the depth between 16 m to 43 m, spikes were recorded at the same
localise depth for the four sensing cables located at the four different
corners of the pile. They are visible at the same depth in all cables
Figure 6 Schematic diagram of the BDSLT set up despite each having an independent baseline profile. This spike is
likely due to the variation of pile stiffness (EcAc+EsAs) (caused for
Prior to the construction of bored pile, a bidirectional jack was example by features such as local change in concrete, Ec and steel
attached to the reinforcement cage. The DFOS cables were tied to modulus, Es; cross sectional area of concrete, Ac and steel, As or
556
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
isolated stiffening bands in the reinforcement cage) (Soga et al, Other possible reason is the concrete may have collapsed during
2015). the construction process when the machinery unintentionally hit the
If the spike is due to the inhomogeneity of concrete, the concrete surface of the bored hole. For such a long pile, the concrete might
modulus can be back calculated by assuming consistent pile be slightly contaminated and compromised the quality of concrete
diameter and steel area along the pile. Figure 8 shows the back (e.g. Young’s Modulus, E) at these localized areas. Hence, much
calculated concrete modulus at selected depth along the pile. At higher strain measurements were recorded.
depth without spikes, etc. 30.0 m, 35.0 m and 46.5 m, the concrete Generally, the back calculated concrete modulus, Ec is lower
modulus is generally at range of 20 to 21 kN/mm2. At depth with than usual 28 days concrete modulus as suggested in some of the
spikes, etc. 27.7 m, 36.5 m and 41.65 m, the concrete modulus was literature. It could be due to the pile was tested at only 14 days after
much lower, at about 10 kN/mm2. These could indicate that the concreting. The concrete had achieved required strength but the
concrete strength at those pile depths with spike are likely lower concrete modulus may have not fully gained yet. Local change in
than the average strength. It is likely due to muck from pile toe that steel cross sectional area, reinforcing links or isolated stiffening
was not completely flushed above the cut-off-level and caused bands in the reinforcement cage may also cause the recorded spikes.
contamination to concrete on certain depths of the pile. If the muck Those localise depth with spike is excluded from the
was completely flushed, localise spike on strain profile should had interpretation of load transfer and mobilised shaft friction analysis
been recorded at the 1 to 2 m overcast concrete above the cut-off- as the pile stiffness are uncertain.
level.
4.2 Pile Shaft Friction
The strain data in this paper is analysed to obtain the axial forces. It
is important to note that uniform axial stiffness along a bored pile is
nearly impossible due to the generally non-uniform shape of bored
pile and the minor inconsistency in the concrete stiffness. However,
the pile stiffness, EA is assumed as constant along the pile to
facilitate the interpretation of pile load transfer, mobilised shaft
friction and mobilised end bearing mechanism.
Figure 9 illustrates the load distribution along the test pile. The
jack was embedded within the limestone bedrock with 0.8 m rock
cover above the jack. However, from Figure 10 and Figure 11, it is
observed that the 0.8 m rock layer above the jack did not contribute
to an expected high shaft friction to the pile. Maximum mobilised
ultimate shaft friction at the rock layer above the pile is only 39
kN/m2, which is lower than the soil layer above it that has mobilised
shaft friction up to 62 kN/m2. The ultimate soil friction value was
estimated based on SPT-N method. Initially, the ultimate soil
friction was estimated based on a constant ultimate shaft resistance
factor, Ksu of 4 throughout the soil layer. However, the denser soil
layer only achieved Ksu value of 2.5; while, the looser soil layer
achieved Ksu value of approximately 3.5.
Figure 8 Back Calculated Concrete Modulus Figure 9 Load distribution along test pile
557
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
40
30
10
The ultimate pile shaft friction of the rock socket below the jack
(Figure 11) was also lower than initial proposed friction value but is -20 Average Pile Top Settlement (Upward) 1st Cycle
near to proposed limestone friction value in the literature (Neoh, Average Pile Top Settlement (Upward) 2nd Cycle
Average Pile Top Settlement (Downward) 1st Cycle
1998) based on the RQD value of the rock. From 52.5 m to 54.0 m,
Average Pile Top Settlement (Downward) 2nd Cycle
the mobilised shaft friction is 176.5 kN/m2 and 260.5 kN/m2 at 54.0
m to 56.5 m. These two layers have mobilised ultimate shaft friction Figure 13 Upward and downward movement against applied load
near to the suggested value in literature which suggested 300 kN/m2
for rock with RQD <25%. However, at the last rock socket layer,
56.5 m to 57.8 m, a mobilised shaft friction of 667.3 kN/m2 (>300
kN/m2) was achieved.
558
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
1 2
equivalent
W / Y 2 F L
(4)
procedure. Therefore, the lower RQD, lower TCR and inclined
bedrock is likely the biggest contributing factor of the resulting low
2 EC A p shaft friction of this test pile.
The equivalent top load, Ftop can be calculated:
F equivalent
W 5. CONCLUSION
Ftop F (5)
A bidirectional static load test on a working pile was carried out
Y
Therefore, the equivalent pile top settlement, Stop is: with instrumented Distributed Fiber Optic Strain Sensor cables. The
test pile was constructed in Kuala Lumpur limestone formation.
S top s (6)
According to Neoh (1998), for limestone with RQD<25%, the
where, suggested ultimate rock friction value is 300 kN/m2. The mobilised
stop = Equivalent pile top settlement ultimate rock friction for this pile are generally ranged from 175 to
s↓ = Downward movement of jack 260 kN/m2 for first 4 m rock socket, which very close to the value
suggested. The highest mobilised ultimate shaft friction achieved
Figure 15 shows the upward and downward curves obtained was at segment near the pile toe (last 1.8 m), approximately 700
from BDSLT and Figure 16 shows the equivalent top loaded kN/m2, which also with RQD< 25% but with generally higher
settlement curve (after conversion). (>50%) TCR value.
The novel way of instrumented load test using Distributed Fibre
Equi. Top Loaded Load (MN) Optics Strain Sensing indicated significant advantage by obtaining
ࡲ࢚ the continuous strain profile of the entire pile length easily as
E q u i. To p S e ttle m e n t (m m )
559
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
6. REFERENCES
Annamdas, V.G.M. (2011) "Review on Developments in Fiber Soga, K; Kwan, V; Pelecanos, L; Rui1, Y; Schwamb, T; Seo, H and
Optical Sensors and Applications". International Journal Wilcock, M. (2015) "The Role of Distributed Sensing in
of Materials Engineering, vol.1(1), pp1-16. Understanding the Engineering Performance of
Federation of Piling Specialists (FPS). (2006) "Handbook on Pile Geotechnical Structures". Proceedings of the XVI
Load Testing". ECSMGE Geotechnical Engineering for Infrastructure and
Lan, C.G.; Zhou, Z. and Ou, J.P. (2012) "Full-Scale Prestress Loss Development, Keynote Paper, pp13-48.
Monitoring of Damaged RC Structures Using Distributed Transportation Industry Specification of China. (2009) "JT/T 738-
Optical Fiber Sensing Technology". Sensor 2012, vol. 12, 2009 - Specification for Static Loading Test of Foundation
pp5380-5394. Pile – Self-balanced Method", pp1-12.
Neoh, C.A. (1998) "Design & Construction of Pile Foundation in Tucker, K.L. and Reese, L.C. (1984) "Research Report Number
Limestone Formation". Journal of Institution of Engineers, 351-1F - The Effect of Bentonic Slurry on Drilled Shafts",
Malaysia, Vol. 59, No. 1, pp23–29. Federal Highway Administration (FHWA), pp1-85.
560
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Design of Simple Drapery Systems for Rock Cuts and Natural Slopes
Matteo Lelli1, Alberto Grimod2, Riccardo Laneri1, Deny1, Vincent Setiawan2, Dinda Mutiara Savitri1
1
PT Maccaferri Indonesia, Indonesia,
2
Maccaferri France,
E-mail: m.lelli@id.maccaferri.com , agrimod@maccaferri.fr2, r.laneri@id.maccaferri.com3, deny@id.maccaferri.com4,
1
v.setiawan@id.maccaferri.com5, d.safitri@id.maccaferri.com6
ABSTRACT: Simple drapery mesh (or simple drapery system) is commonly installed to provide rockfall protection in open pit mines or
along the highways and railways. It is generally used for rock cuts or natural rock faces. The system is a fast and cost-effective rockfall haz-
ards mitigation measure, and it consists of a steel mesh fixed at the top of the slope with anchors and a longitudinal suspension cable. This
paper describes a new design approach to define all the components of the system. A case study of a drapery system recently designed and
installed in Indonesia is presented.
Keywords: rockfall, remediation, metallic mesh, drapery, Indonesia
561
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
In comparison to other types of rockfall interventions (i.e. se- According to the literature (Muhunthan et al., 2005; Sasiharan et
cured drapery systems, rockfall barriers, etc.), the simple drapery is al. 2006), the inclusion of vertical ropes reduces the stress concen-
more cost-effective, and its maintenance is easier, but on the other tration on the mesh, only if they are woven and not simply applied at
hand it cannot be considered the remedy for shallow instabilities, the job site. The mesh coupled with interwoven cables is fit to trans-
because it can only mitigate the effect of the falling trajectories. fer the loads directly to the top anchor system, thus the tension on
This system is usually installed on high rocky slopes, where pin the wires is reduced. The figure below describes the load effect on a
drapery systems are not too costly, or where rockfall fences or rock- simple drapery system in the case of a mesh without cables (cases
fall embankments cannot be installed due to the uneven or high A, above), and in the case of a mesh with woven edge cables (cases
gradient of the slope. B, below).
The figure defines 3 different load conditions on the system:
theoretical case (left images), considering only the weight of the
mesh (center images), and considering also the debris accumulation
at the toe (right images). It is possible to observe that in case B the
presence of the vertical cables minimizes the deformation in the
center of the panel, and the ropes transfer directly the forces to the
top anchor system. Therefore, the real condition (bottom-right image
in figure 6) looks like the theoretical one (left images on figure 6).
For this reason, SteelGrid, which is a woven composite mesh made
of steel wire and metallic ropes woven together during the produc-
tion of the hexagonal double twist wire mesh, is effective.
Figure 6 illustrates that SteelGrid mesh is ideal for use on high
rock faces and slopes with a long drop or where large volumes of
debris are expected. The longitudinal steel ropes enable the efficient
transmission of loads to the crestline ropes and anchors, with mini-
mal mesh deformation. This aspect allows the mesh to be stressed
with larger loads or be less maintained, thanks to the increased ca-
pability of the system to retain debris at the toe.
3. TYPE OF MESH
During the design, the first question that consultants must solve is
“Does the mesh have a quasi-static or a dynamic behaviour?”. The
market offers a wide portfolio of meshes, such as single twist (chain
link) or double twist wire meshes, steel geocomposites with cables
and wires (SteelGrid), cable meshes, cable panels and ring net pan-
els. The graph in figure 5 recaps, in a semi quantitative way, the
performances of the main meshes in the case of static and dynamic
applications. The graph shows a non-direct proportion between the
tensile resistance (quasi-static), and the dynamic resistance, which
depends on the mesh deformability.
For most applications, the dynamic resistance is useless, and the
required tensile resistance is pretty low. Other interventions require
steel geocomposite (SteelGrid) to reduce the stresses on the suspen-
sion system. If high dynamic stresses are predicted (“dynamic Figure 6. Comparison between a mesh without interwoven ca-
shield”) High Energy Absorption (HEA) cable panels should be bles (case A, above) and with interwoven cables: SteelGrid (case B
applied, because of their high performance in dynamic conditions, below), considering 3 different load conditions
such as for the attenuator systems (Arndt et al., 2009). If the dy-
namic impacts are extremely severe, rings nets are required. Another important aspect to be evaluated during the mesh de-
sign is the capacity of the mesh to avoid the unravelling. This is an
intrinsic property of the mesh fabric. The mesh must be able to in-
hibit the propagation of the tears between the wires. In order to
avoid this issue, double twist wire meshes are usually preferred to
simple twist meshes (i.e. chain link). Researches and laboratory tests
show that damage to a double twist wire mesh remains localized and
the mesh does not unravel (figure 7) (Agostini et al., 1988).
562
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
4. DESIGN
4.1 Factors influencing the design
The main factors affecting the proper choice of the mesh aimed at
the simple drapery system are the following:
1. Slope morphology: the probability to have dynamic impacts
against the drapery increases if the slope is uneven and/or
steep. For instance, in case of very uneven slopes, the drapery
can only push on the crest lines and convexities, whereas the
debris can freely run down on gullies and concavities. In this
situation drapery has a negligible capacity in erosion control,
and the falling stones can reach high velocities. Therefore, the Figure 8. Example of a simple drapery system with cable panels and
installation of the drapery requires a particular care in order to double twist mesh. The proper weight of the meshes (approx. 5.0
maximise the contact between ground and mesh, or the slope kg/m2) can avoid the rolling of single boulders (red arrows) along
must be preventively regularized. the slope.
2. Top anchor spacing: this factor is strictly related to the previ-
ous one, in fact the crest anchors shall be placed in order to 4.2 Design goals
guarantee a good contact mesh-ground. Moreover, the resis-
tance of the crest supporting system must be increased if the The main design goal of simple drapery systems is getting a propor-
space between the top anchors rises. In order to generate over tioned protection system that gives the possibility to the single com-
stress on the anchoring system (anchors and cable), the anchor ponents to work all together properly. Only the top anchor system
spacing should be generally lower than 3 - 4 m. could be slightly oversized in order to guarantee the safety of the
3. Prevalent instability: if the erosion represents the main prob- foundations in case of the collapse of the lower part. Another goal is
lem, the most suitable mesh should have small opening size recognizing the limit load of the drapery system. This allows to
and enough weight to push down and stabilize the loose predict the maximum height of the drapery, and when the cleaning
ground surface. As soon as the contact between mesh and of the debris pocket is required.
ground is reached, the drapery becomes quite effective as ero-
sion control. In fact, under this condition, it is possible to have 5. CALCULATION APPROACH
the growth of the vegetation (to obtain a better result the mesh The design of simple drapery depends on different variables related
can be coupled with an erosion control mat) and the retention to the geometry of the slope, the type of mesh and the hypothetical
of debris and boulders. If the slope is vertical, the drapery debris accumulation at the bottom of the system. Nowadays, the
must be quite strong to absorb the impacts and guide the de- only research carried out to give a design guideline for these appli-
bris towards the bottom. In case of large blocks, a “dynamic” cations was done by the Washington State Department of Transpor-
drapery, like cable panels or ring nets, should be foreseen. tation (Muhunthan et al. 2005).
Whereas in case of small blocks (i.e. thin layered limestone Using these studies, and the results obtained from several labo-
cliff) lighter draperies, like geocomposite SteelGrid or double ratories and field tests, Maccaferri has developed a new calculation
twist wire mesh, might be sufficient. approach (implemented in the software MacRo 2) which give the
4. Expected life span of the drapery and maintenance costs: re- possibility to design the type of mesh, the up-slope longitudinal
garding the life span, the design shall consider the exposure of suspension cable and the characteristic of the crest anchors (diame-
the mesh to aggressive environments (i.e. salted winds, water ter, type of steel and length). This tool allows designer to have a
etc.) and the possible abrasion due to the debris movements. If quick and easy, yet reliable solution to the problem. Sometimes, a
the drapery is applied for temporary protection (i.e. mining complex numerical analysis has to be done, but this is not practical
industry) light corrosion protections (i.e. simple galvanisation for all projects, especially if the intervention has a modest size and
with zinc) can be applied. Whereas, heavier corrosion protec- has to be done in a limited period of time (emergency protection).
tions (i.e. zinc-aluminum alloys, polymeric coatings) may be The equations and the procedures at the base of this design ap-
required for permanent infrastructures applications. In term of proach are quite simple, but they give reliable and fast results.
maintenance, designers shall predict the maximum size of the
debris pocket at the toe of the mesh or the dimension of the 5.1 Mesh design
ditch at the base of the system. Once the designed accumula-
tion limit is reached, the debris shall be removed (from the The simple drapery system is a passive system capable to contain
mesh pocket or from the trench). the debris at the bottom of the slope. It has to be designed by taking
5. Weight: the weight affects the capacity of the system to stabi- into account all the loads able to generate a stress on the mesh:
lize single boulders. If the mesh is too light, its capacity to 1. The weight of the chosen mesh;
push down and stabilize large blocks is reduced. Heavy 2. The weight of the debris accumulated at the toe of the mesh;
meshes (i.e. cable panels) may avoid the rolling/falling of 3. External weights, like the snow or ice accumulation on the
block along gentle slopes (figure 8). drapery.
563
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Where:
s = snow (or ice) unit weight;
Ts = snow (or ice) thickness, considered homogenous along the
entire area to be protected;
s = friction angle between soil and snow (or ice) (typical values
are between 30° and 40°).
Figure 9. Geometrical input data used to calculate the stresses on the
To design the drapery system at the limit equilibrium state, three
mesh due to the debris accumulation
partial coefficients have to be introduced in the calculation to in-
crease the acting forces and decrease the resisting ones:
These three loads may be described by the following formulas,
1. mts = partial coefficient which reduces the tensile strength of
based on the researched of the U.S. Department of Transportation
the mesh (≥ 1.0; from the in-situ evidences and in-situ and
FHA (Muhunthan et al. 2005).
laboratory tests, this factor should not be lower than 2.0).
Firstly, the total load due to the mesh (Wm) has to be defined:
2. vl = load partial coefficient which increases the variable
loads, like the snow thickness and the debris accumulation (≥
Wm m Hs / sin (sin cos tan ) (1)
1.0; suggested value according to the Eurocode 7, 1997 =
1.5).
Where: 3. pl = load partial coefficient which increases the permanent
m = steel mesh unit weight; loads, like the weight of the mesh (≥ 1.0; suggested value ac-
Hs = total height of the slope (figure 9); cording to the Eurocode 7, 1997 = 1.3).
= inclination of the slope (figure 9); The total stress acting on the netting (Sw) is therefore:
= friction angle between mesh and slope, typical values are
shown in Table 1. Sw Wd Ws vl Wm pl (5)
Table 1. Typical Values of The Friction Angle Mesh-ground (Sasi- The Serviceability tensile strength of the mesh (Rm) is calculated as:
haran et al., 2006).
564
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Where TCABLE is ultimate tensile strength of the designed rope (it Where, Sbar(j) is the working shear resistance of the anchor j, N(j)
depends on the steel grade, the type of rope core and the diameter of is force that the cable and the mesh transfer to the anchor j (calcu-
the rope, i.e. see ASTM A1023/A 1023M, 2002 or UNI EN 12385- lated with the catenary solution), and j is position of the anchor (in-
4:2008), and CABLE is safety coefficient which reduces TCABLE (≥ termediate or lateral).
1.0).
The safety coefficient of the cable is therefore:
Sbar ( j ) (Ybar ( j ) / st ) / 3 2
1
(11)
FSCABLE TWLC / FCABLE 1 (8.a)
Where, Ybar(j) is the yield load of the steel bar j, and st is safety
Moreover, using the catenary model it is possible to define the coefficient for the steel strength of the bar (> 1.0).
maximum length of the rope and its maximum sag between two
consecutive anchors. Ybar( j ) ESS ( j ) ADM ( j ) (12)
ESS ( j ) / 4 e( j ) 2tc ( j ) i ( j )2
2
(13)
Where:
e(j) = external diameter of the steel bar j;
tc(j) = thickness of corrosion on the external crown of the steel
bar j;
i(j) = internal diameter of the steel bar j.
Thus, the safety coefficient of the different anchors may be calcu-
lated as follows:
Figure 10. Example of the deformation of the up-slope cable
(red line) between two anchors (A and B) calculated using the cate-
nary model. FS ANCHOR ( j ) Sbar ( j ) / N ( j ) 1 (10.a)
565
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 12. Example of a pull-out test to verify the designed length of Figure 14. Sketch representing the slope cross section and the type
of rockfall protection to be implemented.
the top anchors.
To accommodate the height of the rock cut the mesh was manu-
6. CASE STUDY: MAROS ELEVATED ROAD factured in customized rolls with variable length (from 15m to 50m)
long. In this way it was possible to decrease the installation time and
The construction of a new road corridor in Maros needed some mas- minimize transversal connections between rolls of mesh along the
sive cutting of the existing rock slopes to gain room for the elevated slope. In this way, the overall cost of the intervention was consid-
road (Figure 13). During the period of 2017 – 2018, the local Public erably reduced.
Works department issued a tender to design and install rockfall
drapery mesh along the exposed cut slopes in order to mitigate the
potential of rockfall hazards to vehicles and people the base of the
slopes, some of which exceed 30 m in height.
Figure 15. Detail of the cable woven in the double twist mesh to
form the SteelGrid (the 8mm cables are interwoven in the mesh with
30 cm spacing in order to reach very high tensile resistance, up to
177 kN/m).
566
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
7. CONCLUSIONS
Figure 18. The supporting system is composed by 32 mm steel bar
A simple drapery system is a cost-effective measure to mitigate
anchors, spaced 3.0 m, and a 22 mm up-slope cable (type
rockfall hazards. This system has been successfully and extensively
6x19+IWRC – 1770 MPa).
used in several scenarios (e.g. mining, highway and railway protec-
tions), and in complex conditions (slope height up to 150 m). The
The table below summarize the safety factors of all the system
aim of this kind of intervention is to control the falls of the unstable
components obtained with MacRo 2:
rocks and debris, which are driven toward the bottom with reduced
velocity and therefore reduced energy. In comparison to other types
Table 2. Output from MacRo 2 calculation.
of rockfall interventions such as secured drapery systems or rock
Element Factor of safety or value anchors, they are simpler to install, and their maintenance is easier.
Nevertheless, they cannot be considered as a remediation meas-
Mesh Capacity: FoS = 2.8 ≥ 1.0 - Satisfied
ure for shallow instabilities. Based on the researches carried out by
Supporting up-slope cable: FoS = 1.09 ≥ 1.0 - Satisfied Muhunthan et al. 2005 and in-situ and laboratory tests, Maccaferri
Intermediate crest anchors: FoS = 3.25 ≥ 1.0 - Satisfied has developed a calculation approach (implemented in the software
Lateral crest anchors: FoS = 1.39 ≥ 1.0 - Satisfied MacRo 2) able to optimize all the components of a drapery system
Min. length of the crest anchors: Lmin = 4.74 m (drill=40mm)
(mesh, supporting cable and crest anchors), and estimate the re-
quired maintenance. Moreover, the case of the simple drapery sys-
(adopted: L=5.0 m,(drill =
76mm) tem installed for the protection of the new Maros elevated road in
South Sulawesi (Indonesia) has been presented. References
567
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
568
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Landslides like any other natural disaster are unpredictable. These events may compromise the function of some infrastruc-
tures or structures around the affected area. Mitigation measures or emergency works must be performed as soon as possible to regain the
function of the affected infrastructure. Reinforced soil structure is one of the most suitable solution to be applied to restore the landslide slope
and prevent future landslide disaster. Reinforced soil structures built using geosynthetics have shown many advantages for emergency works
or landslide remediation if compared to conventional retaining structures such as mass gravity and cantilever walls. This paper aims to give
explanation about the advantages of performing remediation works using reinforced soil structure with polymer as the primary reinforce-
ment. Two completed projects are presented to provide author’s experience regarding the application of the reinforced soil structure and
proper construction process for emergency works and landslide remediation in Indonesia.
Keywords: landslide, remediation, polymer, reinforcement
569
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
While, the most commonly used for reinforcement materials are information, the project site is classified as the loose and medium
steel meshes, steel strips and geosynthetics in the form of woven dense material.
geotextiles, geogrids and geocomposites. But, in cases of the land-
slide remediation or emergency works, steel strips are not suitable
due to its stiff and heavy form that require a lot of effort and time to
be installed. Therefore, geosynthetic considered as the most suitable
material for this occasion because of its form and weight that easy to
handle and install.
Geosynthetic have been known a long time ago and well ac-
cepted in this construction industry. For the retaining soil structure,
the primary function of the geosynthetics needed is reinforcement.
This reinforcement works thanks to its tensile strength that can resist
the stress and deformation in the structure. Some type of geosyn-
thetic that can be applied for the reinforcement function are geogrid, Figure 4 Soil stratigraphy based on bore hole test.
geotextile, or geocomposite.
Reinforced soil structure offers more flexibility that the structure 3.2 Solution
can accommodate differential settlement and give the structure bet-
The proposed solution is a 8 m reinforced soil structure, using
ter capability in terms of drainage. Recently, this system is becom-
woven geogrids with the commercial name MacGrid as a primary
ing more often applied in private and government projects even for
reinforcement and anchored gabions (Terramesh System) will act as
landslide remediation or emergency works. The main factors that
a front wall fascia unit to maintain stability in the front of the struc-
driving the choice of this systems are their cost-effectiveness, flexi-
ture and protect the structure from surface erosion.
bility and most importantly the construction period to construct this
Terramesh System are reinforcing elements made up of hexago-
structure that is relatively faster in terms of implementation than the
nal double twisted steel wire mesh type 8x10, forming a facing ga-
other permanent solutions such as concrete and masonry walls.
bion and a variable length reinforcement according to the project.
The steel wire used for manufacturing Terramesh System units is
heavily galvanized. A polymer coating in compliance with EN
10245-1 having a nominal thickness of 0,50 mm is then applied to
provide additional protection against environmental agents (UV
rays, rain, etc.). The reinforcement base panel, front panel and ga-
bion box lid of the Terramesh System are manufactured from one
continuous panel of mesh; the back panel is factory connected to the
base panel. The facing section of the unit is formed by connecting
the back panel and diaphragms to the main unit; the gabion facing is
divided into uniformly partitioned cells by internal diaphragms.
Figure 2 Typical section of a reinforced soil slope combining poly- Steel ring fasteners or coated lacing wire are used to connect units
meric reinforcements (geogrids) and steel wire mesh reinforcements and to close and secure them.
(anchored gabions and steel wrap around units)
570
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
571
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
4.2 Solution
The Public Works decided to repair the road and improve the drain-
age system, so that this problem can be avoided in the future. The Figure 15 Backfill and Compaction Procedures
original proposal from the consultant was a gravity mass retaining
structure made from gabion with woven geotextile as the primary
reinforcement and separately placed behind the gabion wall as the
separator. Due to its massive structure, the remediation act shall be
difficult to realize due to limitation of funds and time, so that the
client asked for an alternative solution in order to generate both cost
and construction schedule optimization. It is important to note that
the client must restore the function of the road as soon as possible to
reduce the risk of loss of the local community.
Therefore, a retaining soil structure with the combination of an-
chored gabion facing and geogrid reinforcement was adopted. The
structure itself can reach 14 m maximum height with the combina- Figure 16 View of the Completed Access Road Remediation
tion between anchored gabion facing and polymer reinforcement.
The proposed system was the combination of Maccaferri’s Ter- 5 Conclusions
ramesh System and High Strength Geogrids Paralink. The geogrid Remediation works after landslide needs to be performed immedi-
ultimate tensile strength (UTS) is 300 kN/m. The system was pro- ately to minimize the risk or losses that can be caused. Reinforced
vided with an internal drainage using a 30 cm thick rock layer soil structure is one of the most suitable solution to restore the
wrapped with geotextile. slopes after landslide and any other emergency works related. This
This new technical proposal can be concluded that the retaining method proved to be more effective in terms of construction time
soil structure was found to be more cost and time effective com- and cost compared to the other solutions. This paper presented two
pared to the original one. The retaining soil structure proposed re- cases that applied reinforced soil structures with polymer reinforce-
quires only one layer of anchored gabion as the facing instead using ment as the remediation method after landslide occurrence in Indo-
the massive gabion structure as the mass gravity wall. nesia.
The first case history is the remediation of 45 m access road for
Leuwigoong Dam in Garut regency, West Java. Landslide happen
after heavy rain all night long and allegedly collapsed due to poor
drainage system of the existing retaining structure. Reinforced soil
structure was applied using woven geogrid as the primary rein-
forcement and anchored gabion as a structural facing and secondary
reinforcement. The internal drainage system inside the structure has
been done by placing available permeable material, namely, gravel.
The thickness of the gravel material is 30 cm and wrapped with
geotextile as separator. Furthermore, the horizontal drainage applied
Figure 13 Typical Cross Section of Reinforced Soil Structure Bukit by using drainage pipe with proper inclination to collect and channel
Kemuning Road in Batam the water out.
The second case is the remediation of 100 meters road in Bukit
Static and seismic analysis had been performed, checking global Kemuning road in Batam. The existing slope collapsed under con-
stability, internal stability and stability as retaining wall (sliding, tinuous exposure to high intensity rainfall and as a result, adjacent
overturning and bearing capacity) using Maccaferri internal soft- residential areas were damaged. Consequently, the Reinforced soil
ware Macstars W. Maccaferri provided support to the contractor structure was chosen as the alternative solution to optimize cost and
with product installation training and supervision. construction schedule from the original solution which was mass
The geogrids are supplied in rolls with 4.5 m width and 200 m gravity wall in the form of gabions. This choice allowed the owner
long. The total weight of each roll geogrid (Paralink) approximately to deliver the project on time and within the target budget. The se-
around 900 kg. Geogrids are arranged before the Terramesh facing lection of the system was also guided by the system fascia perme-
units, from the front of facing up to the length required by the de- ability, flexibility and construction ease. Reinforced soil structure
sign. The laying direction will be in the rolling direction that having proposed was anchored gabion and high strength geogrid 300 kN/m.
greater resistance has to be placed perpendicularly to the facing. The construction itself was completed in approximately 4 months.
572
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
6 References
Koerner, R.M. (2005). Designing with Geosynthethics, Pearson and Metallic Reinforcements using Limit Equilibrium Meth-
Prentice Hall, New Jersey, 5th Edition. ods, Indonesian Geosynthetics Conference 2016, Indonesian
Lelli, M., Cerro, M., Scotto. M., and Rimoldi. P. (2017). Geosyn- Chapter of International Geosynthetics Society, Jakarta.
thetic-reinforced soil structures, recent Indonesian experi- Lelli, M., Gouw,T.L., and Laneri, R. (2017). The Important Role of
ences, Indonesian Chapter of International Geosynthetics So- Execution and CQA/CQC Plans in Building with Geosynthet-
ciety, Jakarta. ics, Geosintetik Indonesia 2017 seminar on Quality in Con-
Lelli, M., Gouw, T.L., Laneri, R., Cerro. M., and Meinata L.E. struction using Geosynthetics, by Indonesian Chapter of In-
(2016). Design of Earth Structures Reinforced with Polymeric ternational Geosynthetics Society, Century Hotel, Jakarta.
573
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Department of Civil Engineering, College of Engineering, Central Mindanao University, Musuan, Bukidnon, Philippines
E-mail: einstineopiso@cmu.edu.ph
ABSTRACT: Improving the geotechnical engineering properties of mine tailings using cement stabilization process can be adopted to
reduce the volume of tailings and consequently the risk of toxicity it poses to people and the environment. Geotechnical properties of
cement-stabilized mine tailings from Brgy Gango, Libona, Bukidnon, Philippines were investigated to assess its applicability for reuse as a
construction material. Mine tailings were stabilized using ordinary Portland cement in four different formulations—0%, 6%, 8% and 10% by
weight of soil and were subjected to compaction test, California bearing ratio (CBR) test, permeability test, unconfined compressive strength
(UCS) test and direct shear test in accordance with ASTM standards. Based on the AASHTO Soil Classification System, the soil was
classified as clay of low plasticity. The maximum dry density and optimum moisture content showed a decreasing trend from the varied
mixture which resulted that adding cement did not improve the compaction characteristics of the stabilized mine tailings. Addition of OPC
up to 6% and 8% had a significant effect on the increase of CBR and UCS values, respectively. In permeability test, it was observed that the
rate of flow of water increases as the percent of OPC increases. The increase of the coefficient of permeability was influenced by the
cementitious bond which forms new solid material within the treated sample. The soil particle becomes bigger in size and the rate of flow of
water increases. Up to 6% addition of OPC is the optimum cement percentage to stabilized the mine tailings based on direct shear testing. In
overall, the cement-stabilized mine tailings from Brgy. Gango, Libona, Bukidnon, Philippines can be safely used in sub-base road
construction and as typically stabilized wastes for land disposal.
574
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
2. METHODOLOGY
T2 8% additional OPC
2.4. Determination of Atterberg Limits
T3 10% additional OPC
The Atterberg Limits were determined in accordance with ASTM
D4318 - Standard Test Method for Fall Cone Test. The soil sample
is placed in a 55 mm diameter, 40 mm deep metal cup. The cone is
2.2. Sample Collection
released for 5 seconds so that it may penetrate the soil. Liquid limit
Mine tailings were collected in the common mine tailing pond at is defined as the water content of the soil which allows the cone to
Gango, Libona, Bukidnon with a latitude of 80 20’ 25” N and penetrate exactly 20 mm during that period of time. The sample was
longitude 1240 43’58” E (Figure 1). Actual collection of soil allowed to dry on the glass plate until it become plastic enough to be
samples were shown in Figure 2. The soil samples were air-dried at shaped into a ball. The sample-ball was molded and rolled between
the laboratory. The dry disturbed soil samples were pulverized using the palms of the hands until cracks on the surface of the sample
rubber mallet and undergone sieve analysis through U.S sieve no. appeared. Then, sample was divided into small pieces and rolled to
40. form thread having 3.2 mm diameter under enough pressure. The
first crack appeared on the thread surface was determined as plastic
limit. Plasticity index was calculated as the difference between the
liquid limit and the plastic limit which measures the plasticity of the
soil.
575
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The stabilized samples were cured for 28 days to let the cement was attached to the syringe to the stand pipe. The valves were first
hydrate and react with the soil. The samples were zipped to prevent locked to ensure that no water may pass through the cylinder. Water
the moisture to evaporate during the curing period and stored away was poured above the stand pipe through a funnel. The safety valve
from direct sunlight. All the samples after 28 days curing were was removed leaving a hole above the cylinder and the upper valve
subjected to testing. was turned on. Bubbles came gushing out on the hole at the safety
valve as water flowed through the cylinder. The lock was then
attached back to the safety valve. The upper valve was also turned
off when there were no more bubbles which were signs of air voids.
2.7. Compaction Test After it was turned off, a timer was set together when the upper and
lower valve handle attached above and below the cylinder was
The modified proctor test was conducted to all the samples based on
turned on and was observed until it penetrates the soil. When it
ASTM 1557 and AASHTO T180 in order to determine
penetrated the soil, a drop of water was be noticed from the faucet
experimentally the optimum moisture content at which the tailings
attached below the cylinder. The timer was stopped and recorded
became most dense and achieved its maximum dry density. The
and the upper faucet was turned off to record the initial height of
proctor compaction test consists of compacting soil samples at a
water at the stand pipe. Another timer was set one (1) minute after
given water content in a standard mould with standard compaction
the faucet was turned on again for three (3) trials to record initial
energy. It uses a 4-inch-diameter mould with the compaction of five
and final heights. The coefficient of permeability was then
separate layers of soil using 56 blows by a 4.5 kg hammer falling 18
determined by using the formula given below.
inches.
Each of the soil samples were first air dried and then separated
into 5 samples. The water content of each sample were adjusted by
(1)
adding water (3% - 5% increments or more depending on the type of
the soil). The soil samples were then placed and compacted in the
Proctor compaction mould in five different layers where each layer
received 56 blows of the standard hammer. Before placing each new
Where
layer, the surface of the previous layers were scratched in order to
ensure a uniform distribution of the compaction effects. At the end
K = coefficient of permeability in cm/s
of the test, after removing and drying of the sample, the dry density
and the water content of the samples were determined for each a = inside cross-sectional area of stand pipe in cm2
Proctor compaction test.
L = length of specimen in cm
h1 = initial head in cm
2.8. California Bearing Ratio Test
A = cross-sectional area of specimen in cm3
The California Bearing Ratio was determined in accordance with
ASTM D1883 (Standard Test Method for California Bearing Ratio t = time interval in seconds in which the head drops from h1 to
of Laboratory –Compacted Soils). The soil specimens were h2
compacted so that their compacted densities ranges from 95% to
100%. The samples were cured for about four days. Penetration h2 = final head in cm
pistons were applied to the samples. The optimum mixture of soil
and cement was determined based on the maximum CBR observed
from the several mixtures that was used.
2.10.Unconfined Compressive Strength Test
576
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
cylinder of soil without lateral support was tested to failure in 3.2. Soil Classification
simple compression, in a constant rate of strain.
Table 3 shows the soil classification of the samples based on
American Association of State Highway and Transportation Official
(AASHTO) and Unified Soil Classification System (USCS). T0 and
2.11.Direct Shear Strength Test T3 sample were both classified under the group A-7-6, while T1 and
T2 sample were under group A-6. All samples belong to fine-grained
The test was performed in the laboratory according to the standard soil category for having 50% or more percent passing through sieve
test procedure ASTM D3080 – Direct Shear Test of Soils for no. 200, and were classified as CL or clay of low plasticity/lean clay
Consolidated Drained Conditions. Treated soil samples was mixed based on USCS soil classification system.
with water according to the moisture content result from compaction
test. It was compacted progressively by 1/3 of its total volume
having 25 blows. The compacted soil was carefully removed from
the sampler and placed inside a clean wrapper. Table 3. Soil Classification of the Samples
The dimension of each specimen was measured. Area and Treatment AASHTO USCS Soil Plasticity
volumes were calculated. The specimen was placed in smooth
layers. The upper grating on stone and loading block on top of soil T0 A-6(13) CL Medium Plasticity
were positioned. Desired normal load was applied before shear pin
was removed. The dial gauge which measures the change of volume T1 A-7-6(19) Cl Medium Plasticity
was then attached. Initial reading of the dial gauge and calibration
values were recorded. As the motor was operated, the reading of the T2 A-7-6(19) CL Medium Plasticity
shear force was recorded as well as the volume change readings
T3 A-6(14) CL Medium Plasticity
until failure.
35
T2 0.16 2.36 97.48
0% 2% 4% 6% 8% 10%
% of OPC
T3 0.38 3.74 95.88
Figure 3. Liquid Limit of the Soil Samples.
577
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
30
Plastic Limits (%)
26,1
25,31
25
23,83 23,83 14
20
Plasticity Index (%)
17,18 17,18
15
13,01 3.5. California Bearing Ratio
12,14
10
The addition of OPC in samples has increased the CBR values as
5 observed Figure 7. Bowles (1992) stated that the subgrades having
CBR value of 3-7% are poor to fair quality of subgrade and greater
0
0 2 4 6 8 10 12 than 50% CBR value can be used as base material . T0 is under poor
to fair subgrade with the CBR value of 4.87%. T1, T2 and T3 have
% of OPC 105.67%, 79.84% and 93.81% CBR values, respectively. All percent
additives are excellent subgrade that can be used as base.
Figure 5. Plasticity Index of the Soil Samples.
3.4. Compaction
120
The cement-stabilized soil samples showed an increasing trend of 105,67
California Bearing Ratio, %
20
4,87
0
Table 4. Compaction Test Result. 0% 2% 4% 6% 8% 10% 12%
% of OPC
Treatment Optimum Moisture Maximum Dry Density
Content (%) (kg/m3)
T0 13.57 1751.39
Figure 7. California bearing ratio of cement stabilized mine tailings.
T1 11.53 1587.51
T2 11.56 1524.29
3.6. Permeability Test Result
T3 12.72 1475.77
578
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Unconfined Compressive
T0 and T2. Setting aside the soil sample reduces its moisture content
250
due to evaporation and causes the optimum moisture content to
Strength, KPa)
239,92
shrink. 200
130,04
150
The decrease or increase of K was influenced by the formation 100
of cementitious bonds which creates new solid material within the
50
treated mine tailings. These tailings are considered to have low to
very low permeability typical of silty sand. This factor increases the 0
0% 2% 4% 6% 8% 10%
average pore size to create a macro-pore structure which causes an
% of OPC
increase in K (Tay & Goh 1991).
0,00065497
3.7. Direct Shear Stress
0,0007 0,0006491
0,0006 It was observed in Figure 10 the relationship of shear stress with
Coeffiicient of Permeability , K
0,0005 time increases until failure occurs which is at its peak. Compared to
0,00035289 all other treatments, T2 has the maximum shear stress which is
0,0004
166.11 kPa that was observed at 30 seconds of the testing and
(cm/sec)
0
0 2 4 6 8 10 12
% of OPC 180
160 T0
120
T2
100
T3
80
3.6. Unconfined Compressive Strength 60
40
Unconfined compressive strength is a control parameter for road fill 20
design. Achieving a compressive strength greater than 345kPa can 0
help to significantly reduce the potential for settlement in deep fills 0 30 60 90 120 150 180 210
(Ferguson and Levorson, 1999). Sample added with 8% of OPC has Time, s
the highest value of UCS with 320.8 kPa, not reaching the desired
value for deep fill purposes. However, the stabilized soil were Figure 10. Time vs. Average direct shear stress graph of soil
classified as very stiff (200kPa- 400KPa) clay consistencies based samples.
on the classification of Das (2010). According to Amini et al.
(2011), the effect of longer curing time (from 14 to 28 days) has
greater effect to the mixture.Thus, the sudden decrease of strength
may be also caused by the number of curing days. The UCS of soil 4. CONCLUSION
significantly increases with the curing period and it is more
Based on the experimental results, the following conclusions were
consistently increasing up to the first 14 days according to Saha et
drawn:
al. (2013).
i. The physical property of the artisanal gold mine tailings found
in Gango, Libona, Bukidnon, Philippines were classified as olive
brown and identified as silts with traces of sandy clays by the
results in the grain size distribution. It has the value of 37.45%,
25.31% and 12.14% for liquid limit, plastic limit and plasticity,
579
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
respectively. The soil was classified as A-6 (13) soil according to ASTM D6913 Standard test methods for particle-size distribution
ASSHTO classification and CL soil based on USCS classification. (gradiation) of soils using sieve analysis.
The soil was low plasticity that implies low compressibility and
lower tendency in swelling which was a good attribute for the ASTM D2487-11 Standard practice for classification of soils for
acceptability of earthen materials in road construction applications; engineering purposes (Unified Soil Classification).
ii. Cement causes the decrease in both MDD and OMC of the mine ASTM D2166 -Standard Test Method for unconfined compressive
waste. Controlled (T0) sample acquired the largest maximum dry strength of cohesive soil.
density and optimum moisture content value of 1751.39 kg/m3 and
13.57%, respectively. ; ASTM D3080 – Direct Shear Test of Soils for Consolidated Drained
Conditions
iii. Stabilized soil samples have the CBR values of greater than
50% classified as excellent quality for sub-grade that can be used as Adajar, M.A. and Zarco, M.A.. 2014. An empirical model for
base materials; predicting hydraulic conductivity of mine tailings, Dec.,
2014, Vol. 7, No. 2 (Sl. No. 14), pp. 1054-1061.
iv. As the percent of OPC increases, the coefficient of permeability
also increases. Stabilization using OPC gave a great significance in Appel, P., Na-Oy, L., Hatsukawa, Y., Osawa, T., Kystol, J., and
the results yet it produced higher permeability results than that of Sorensen, L. (2011) “Cleaning mercury polluted mine
the controlled treatment because OPC formed bonds between tailing tailings in the Philippines”.
particles that creates new solid material and caused faster discharge
Alhomiar, S. (2016). Thesis-Hydraulic Conductivity of Fly Ash-
of water through the sample.
Amended Mine Tailings. Retrieved:
v. Sample added with 8% of OPC has the highest value of UCS https://dspace.library.colostate.edu/bitstream/handle/1021
with 320.8 kPa, not reaching the desired value for deep fill purposes. 7/176619/Alhomair_colostate_0053N_13648.pdf?sequenc
However, the stabilized soil were classified as very stiff (200kPa- e=1.
400kPa) clay consistencies;
Bowles, J. (1992). Engineering properties of soil and their
vi. As the cement content for each treatment increases, there is an measurements, 4th Edition.
evident decrease in soil strength but values were still higher than of
Canencia, O., Palmes, N., Ibonia, S., Talampas, W., Lammawin, V.,
the controlled sample; and
Kawaling, G., Yañes, S., and Savaña, MJ. (2015) “Minin
vii. The study indicates that cement can be very effective in Industry in Northern Mindanao: Its Environmntal, Social
stabilizing the ASGM tailings up to 6% addition of cement that can and Health Impact Toward Responsible Mining. 2015”.
be safely used in sub-base and base road construction and in typical
Das, B.M. 2010. “Principles of geotechnical engineering”. 7th Ed.
stabilized wastes for land disposal.
Stamfold: Cengage Learning”.
5. REFERENCES Ferguson G. and Levorson S.M. (1999). “Soil and Pavement Base
Stabilization with Self-Cementing Coal Fly Ash.
American Coal Fly Ash Association International,
Alexandria, VA
ASTM D421 – Standard Method for Sieve Analysis of Fine and
Coarse Aggregates. Israel, C. and Asirot J., 2000. Mercury pollution due to small-scale
gold mining in the philippines: an economic analysis,
ASTM D3080 - Direct Shear Test of Soils for Consolidated Drained 2000-06, 1-62.
Conditions.
Sivakugan, N.,Rankine, R.M., Rankine, K.J., and Rankine, K.S.,
ASTM D4318 – Standard Test Method for Fall Cone Test. (2006). Geotechnical considerations in mine backfilling in
Australia. Journal of Cleaner Production, 1168-1175.
ASTM C136. “Standard method for sieve analysis of fine and coarse
aggregates”. ASTM International. Supriya Saha and Sujit Kumar Pal. (2013). “Influence of Fly Ash on
Unconfined Compressive Strength of Soil and Fly Ash
ASTM D1883. “Standard test method for California Bearing Ratio Layers Placed Successively”.
of laboratory – compacted soils”. ASTM International.
580
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Acknowledgment:
The authors would like to the Central Mindanao University for the
support in the implementation of the study as well as the
Department of Science and Technology – Philippine Council for
Industry, Energy, and Emerging Technology Research and
Development (DOST-PCIEERD) which funded this research.
581
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT:
Pemalang-Batang (Pematang) Toll Road as a part of Trans Jawa Toll Road is currently being constructed. The Pematang Toll Road which has
39.2 km length with 13 embankments before bridges (bridges approach / oprit) will be constructed above very soft river sedimentation soil.
The construction of embankment on the soft soil leads to excessive settlement and high susceptibility toward slope stability failure. Ground
improvement using Controlled Modulus Columns (CMC) is adopted to reinforce the soil, increase the strength thus reduce the settlement and
stabilize the embankment. The application of CMC at Pemalang-Batang Toll Road is indeed a breakthrough in history of toll road construction
as a new construction technology in Indonesia, where majority the settlement/stability issue commonly solved by structural or conventional
solution such as: slab on pile, cerucuk matras beton, PVD/Vacuum Preloading etc. Control Modulus Column is one of ground improvement
technique installed by forming a regular grid of semi-rigid inclusion to enhance global stiffness of soil mass so that the problem such as:
settlement, bearing capacity can be controlled. The installation of CMC itself performed using special design auger element that is able to force
the soil displace laterally induce volumetric expansion resulted a lateral compaction of surrounding soil. In Indonesia, Menard has been
successfully implementing CMC in several projects for different sector. The CMC’s design technology makes them uniquely efficient for the
immediate support of road for public transport, large liquid or bulk solid storage tanks warehouse, and other infrastructures building or facilities.
This paper gives a brief description of application of CMC for supporting embankment to achieve settlement criteria, special design concept,
and recommendation are briefly discussed.
Keywords: control modulus column, very soft soil, ground improvement, embankment, toll road
582
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
583
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3.3 Wiremesh
The thickness of LTP shall be designed to transfer the load Figure 9. Configuration examined at the edge of the
efficiently. The load transfer’s efficiency is increasing with the embankment (Asiri National Project, 2012)
thickness until optimum thickness. If the optimum thickness is
exceeded, the thickness will become additional load. The efficiency Beneath the embankment slopes, the tensile force Td;thrust
is also influenced by the spacing between CMC and shear strength mobilized in the geosynthetic by the embankment lateral active
characteristic of LTP material (elastic modulus and friction angle). pressure decreases with distance from the slope crest; this force may
Typical thickness of LTP is commonly ranged from 0.3m to 1.0m. be calculated at any point based on the embankment slope and
Figure 8 illustrates load distribution mechanism within LTP. thickness above this given point.
It is also important to note that this configuration may also
induce lateral displacements of the compressible soil capable of
creating bending moments in the inclusions that also require
verification.
584
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
1. Calculate the axial design load and design moment that work with
4. PROJECT CASE STUDY: UP BOJONG
the following equation
N ed N plaxis spacing G 4.1 Project Description
M ed M plaxis spacing G Pemalang-Batang Toll Road with 39.2-kilometer highway connects
Pemalang with Batang Area, located in Central Java Province,
Where: Indonesia (Figure 13). The toll road is part of Trans-Java
Ned, Med : axial and moment acting on CMC Expressway (Figure 14 that connect Merak Port to Banyuwangi. The
NPLAXIS, MPLAXIS : axial and moment obtained from Plaxis Pemalang-Batang Toll Road has 13 bridge approach, embankment
Spacing : Spacing of CMC near bridge with height ±6.0m, where 7 bridge approach is treated
γG : safety factor; 1.2 - 1.4 with CMC. One of the bridge approach treated using CMC methods
are located at UP Bojong (STA 354+870) with estimated area 4,200
The axial load and bending moment are obtained from inner forces m2.
for every CMC columns as described at chapter 5.4 The Pemalang-Batang Toll Road is planned to be used during
Eid Al-Fitr, hence the construction of embankment need to be
2. Calculate resistance CMC (Nrd): accelerated using CMC method. The design criteria to be achieved
for ground improvement are listed below:
a. Eccentricity (e) with following equation = ,
The maximum allowable residual settlement shall be 100
illustrated as below: mm after the 10 years of ground improvement
The factor of safety against slope failure shall be equal or
more than 1.5 static condition.
The loading component shall be calculated from fill
embankment soil from existing ground level to the final
platform plus the 15 kPa traffic loading.
585
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Additional soil investigation campaign consist of two boreholes and Elevation Elevation Fill thickness
In situ test
one CPT was conducted in 2017 by Menard to verify data consistency EGL (+m) FPL (+m) to FPL (m)
and provide more detail information for design purposes. Figure 15 BH-23 +7.2 +15.2 ± 8.0 m
shows the location of additional soil investigation. BH-24 +7.0 +15.0 ± 8.0 m
CPT-07 N/A N/A N/A
Elevation Elevation
Depth of
In situ test EGL (m) GWL
GWL (m)
(+m)
BH-23 +7.2 3.95 +3.25
BH-24 +7.0 5.6 +1.4
CPT-07 N/A N/A N/A
According to soil investigation results, a soft silt layer was
observed at 6.0m depth from existing level; while stiff silt layer was
Figure 15. Location of Additional Soil Investigation. found at ± 20.0m depth. The NSPT result, soil layering are shown
The following table briefly summarized the elevation reference below figure:
and ground water level for design purposes. For design, 8.0m fill
thickness was used, while ground water level was conservatively
assumed at 1.0m below existing level.
586
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
k c’
No SOIL TYPE ‘
(m/day) (kN/m³) (kPa)
The concrete strength used for CMC were 20 Mpa. While for
material design used in PLAXIS calculate using following equation
and summarized in table below:
Eref(E50) = 2700 ∛ 28
Where:
Eref(E50) = modulus elasticity of CMC, kPa
F28 = grout srength after 28 days, MPa
587
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
5.1 Ground Improvement Design Steps 5.2 Settlement Analysis-Without Ground Improvement
The elevation reference and loading calculation is shown in Figure Analysis Settle3D analysis was performed using information from
18. The calculation was started from settlement analysis without Figure 18 and parameter from Table 3 and Table 4. The settlement
ground improvement due to design load using Settle3D software. In of bridge approach due to embankment load and traffic load without
this condition, the settlement due to embankment and traffic load ground improvement ranged from 600 to 800 mm which exceed
exceed 100 mm. Ground improvement was required at bridge settlement criteria 100 mm. The analysis result is illustrated at below
approach area, hence first step of CMC analysis shall be performed figure:
using Finite Element Method-axisymmetric to predict CMC single
column behavior such as settlement and verification of compression
capacity of CMC.
588
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
589
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
The bending moment and axial force inside the KGM is limited to
FoS – Short Term: 1.543 certain value to ensure limited strain envelope inside the CMC. The
procedure to check limit of axial and bending is using BS EN 1992-
FoS – Long 1-1:2004 §12. The axial forces and bending moment forces for each
Term: 1.703 CMC column were obtained from FEM model (Figure 25 and
Figure 29).
From the rebar checking, we conclude that rebar is required to
be placed inside the CMC for two outer most CMC columns.
Where:
fe = stress yield stress, kPa
n = number of steel bars per meter
dy = bar diameter in longitudinal section, m = 0.008m
spacing = 0.15m
590
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
6. QUALITY CONTROL
7. CONCLUSION
ACKNOWLEDGEMENT
The author would like to thank Mr. Abi Hakim and Ms. Febrini
Hartianty Adinda for helping to improve the document presentation
and collection of references, as well as their continuous support for
the development of this paper.
Reference:
591
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRAK: Tanah merupakan bagian penting pada konstruksi jalan sehingga kekuatan dan stabilitas subgrade
sangatlah diperlukan untuk mendukung beban jalan raya tersebut. Pada umumnya penentuan kekuatan,
kekakuan dan stabilitas bahan subgrade pada konstruksi jalan dinyatakan dengan nilai CBR (California
Bearing Ratio) dan Modulus Resilient (Mr). Modulus resilien merupakan perbandingan antara tegangan siklis
dengan regangan. Saat ini pengujian modulus resilien sangat jarang digunakan untuk menentukan kekakuan
tanah dasar. Hal tersebut dikarenakan ketersediaan alat triaxial siklik sangat kurang, pengujian yang rumit dan
biaya yang besar, sehingga untuk meminimalisir kekurangan tersebut, maka diusulkan suatu prediksi nilai
modulus resilien (Mr) dengan nilai CBR. Prediksi nilai modulus resilen dan CBR bertujuan sebagai acuan
pada perencanaan nilai Mr pada tanah ekspansif yang distabilisasi dengan abu vulkanis dan kapur.
hasil pengujian triaksial siklis dan CBR pada penelitian ini diperoleh korelasi antara Mr dan CBR
dengan menggunakan regresi linier berganda.
Kata Kunci: Modulus resilien, stabilisasi tanah, abu vulkanis
ABSTRACT: Soil is an important part of road construction, so subgrade strength and stability are needed to
support the load of the highway. In general, the determination of the strength, stiffness and stability of subgrade
materials in road construction is stated by the value of CBR (California Bearing Ratio) and Resilient Modulus
(Mr). Resilient modulus is a comparison between cyclic stress and recovery strain. Currently testing resilient
modulus is very rarely used to determine subgrade stiffness. This is because the availability of cyclic triaxial
equipment is very lacking, complicated testing and large costs, so as to minimize these deficiency, a prediction
of resilient modulus (Mr) is proposed with the CBR value. The prediction of resilient modulus and CBR values
is intended as a reference for planning the value of Mr. on expansive soil stabilized with volcanic ash and lime.
The results of the cyclical triaxial test and CBR in this study obtained the correlation between Mr and CBR
using multiple linear regression.
Keywords: Resilien modulus, soil stabilization, volcanic ash.
1 PENDAHULUAN alat triaxial siklik sangat kurang, pengujian
yang rumit dan biaya yang besar, sehingga
Tanah merupakan bagian penting pada untuk meminimalisir kekurangan tersebut,
konstruksi jalan sehingga kekuatan dan maka diusulkan suatu prediksi nilai modulus
stabilitas subgrade sangatlah diperlukan untuk resilien (Mr) dengan nilai CBR.
mendukung beban jalan raya tersebut. Pada Prediksi nilai modulus resilien dengan CBR
umumnya penentuan kekuatan, kekakuan dan sebenarnya telah banyak dilakukan oleh
stabilitas bahan subgrade pada konstruksi jalan beberapa peneliti. Namun korelasi tersebut
dinyatakan dengan nilai CBR (California hanya dapat digunakan untuk jenis tanah
Bearing Ratio) dan Modulus Resilient (Mr). tertentu. Sedangkan untuk jenis tanah lempung
Modulus resilien merupakan perbandingan ekspansif yang distabilisasi dengan abu
antara tegangan siklis dengan regangan. Saat ini vulkanis dan kapur tidak dapat menggunakan
pengujian modulus resilien sangat jarang korelasi tersebut. Sebelum melalukan korelasi
digunakan untuk menentukan kekakuan tanah antara Mr dan CBR, dilakukan beberapa tahap
dasar. Hal tersebut dikarenakan ketersediaan penelitian yaitu tahapan penentuan kadar abu
592
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
593
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
pada perancangan perkerasan. Hasil dari Mr 1. South African Council on Scientific and
memberi pengaruh yang sangat besar terhadap Industrial Research (CSIR)
perancangan ketebalan lapis fondasi dan M r ( psi) 3000 CBR 0.65 .......................(4)
lapisan aspal.
Modulus Resilien digambarkan sebagai fase 2. U.S. Army Waterway Experiment station
yang diperoleh dari kemiringan tegangan dan M r ( psi) 5409 CBR 0.71 ………….......(5)
regangan yang terjadi selama pembeban
berlangsung yaitu beban kendaraan melewati 3. Transport and road research laboratory
perkerasan. Uji modulus resilien mengukur M r ( psi) 2555 CBR 0.64 …..................(6)
kekakuan pada benda uji berbentuk silinder 4. Heukolom and Fuster (1960)
dengan menerapkan beban berulang. Untuk
tegangan deviatorik yang diberikan modulus
M r ( psi) 1565 CBR ............................(7)
resilien didefinisikan sebagai kemiringan kurva 5. Hopkins (1991)
tegangan deviatorik dan regangan aksial M r ( psi) 2596 CBR 0.874 ......................(8)
(Gambar 2)
Persamaan tersebut diatas digunakan untuk
jenis tanah berbutir halus sehingga persamaan
diatas tidak dapat digunakan untuk jenis tanah
yang distabilisasi dengan abu vulkanis dan
kapur. dengan demikian perlu dilakukan
pengembangan model persamaan yang dapat
digunakan untuk tanah lempung ekspansif yang
distabilisasi dengan abu vulkanis dan kapur.
3 METODE PENGUJIAN
Uji modulus resilien dilakukan sesuai
standar AASHTO T307 menggunakan alat
triaxial siklis pada Gambar 3. Prosedur
Gambar 2 Definisi modulus resilien pengujian untuk tanah dasar terdiri dari dua
����
tahap. Tahap pertama adalah tahap kondisi
��= .................................................(2) 1000 siklus. Tahap kedua terdiri dari 15 beban
��
dengan, kompresi siklik urutan dengan berbagai tekanan
���� : tegangan maksimum siklis (kN/m2),
kekang. Tekanan aksial maksimum
� : regangan elastis (%) menekankan selama fase ini berkisar dari
�
sekitar 14 kPa sampai 70 kPa, dan tekanan
kekang berkisar dari sekitar 14 kPa sampai 42
2.3 Model korelasi Modulus resilien dan CBR kPa.
Nilai modulus resilien diperoleh dengan
pengukuran dilaboratorium menggunakan alat
triaxial siklis. Namun, ketersediaan alat triaxial
siklis terbatas sehingga nilai modulus resilien
ditentukan dengan korelasi empirik baik
dengan nilai CBR maupun nilai-R. AASHTO
mengusulkan korelasi Mr dan CBR untuk untuk
tanah berbutir halus :
M r ( psi) 1500 CBR .................................(3) Gambar 3. Alat triaxial siklis
Korelasi ini efektif untuk nilai CBR
unsoaked < 20 % dan CBR Soaked <10% yang 4 HASIL DAN PEMBAHASAN
dapat digunakan untuk perancangan Korelasi nilai CBR dan modulus resilien
perkerasan. Selain itu, ada juga berbagai bertujuan untuk memprediksi nilai modulus
persamaan lain yang digunakan untuk resilien yang dapat digunakan sebagai acuan
memperkirakan nilai Modulus resilien perencanaan dilapangan untuk jenis tanah
berdasarkan hasil uji CBR yaitu : ekspansif distabilisasi menggunakan abu
vulkanis dan kapur. Untuk memperoleh nilai
594
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
tersebut maka dilakukan analisis regresi yang Analisa statistik diperlukan untuk
terdiri dari satu variabel bebas yaitu variabel mengembangkan pengembangan model
nilai CBR laboratorium. Setiap data pengujian prediksi yang dapat diandalkan sebagai
di plotkan kedalam grafik dan dianalisis dengan alternatif korelasi antara nilai modulus resilien,
cara regresi untuk mendapatkan nilai secara dan CBR dengan faktor pengaruh tegangan
umum pada model konstitutif yang disajikan utama mayor, tegangan minor atau tegangan
pada Gambar 4 dan Gambar 5. Berdasarkan kekang.
hasil dari prosedur regresi nonlinier dengan Analisa statistik diperlukan untuk
model fungsi berpangkat, analisis dilakukan mengembangkan pengembangan model
untuk menentukan model terbaik untuk prediksi yang dapat diandalkan sebagai
memprediksi nilai Mr dari regresi koefisien alternatif korelasi antara nilai modulus resilien,
berdasarkan Nilai CBR yang disajikan pada tegangan utama mayor, tegangan minor atau
Tabel 2. tegangan kekang dan CBR. Pada hasil analisa
hubungan antara modulus resilien dengan CBR,
d dan diperoleh persamaan sebagai berikut
……………………………………………..(9)
Untuk mengetahui variabel b,d,f maka
dilakukan analisa regresi berganda dengan
bantuan excel. Variabel dependen dan
independen disajikan pada Tabel 2.
595
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
596
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: This paper presents soil water content and pore-water pressure distribution in a vegetated residual soil slope in response to a
one-year seasonal variation in Singapore. The measurements were taken by tensiometers and soil moisture sensors TM1-SM1, TM2-SM2,
TM3-SM3 and TM4-SM4 installed at vertical distances of 2.0, 2.0, 1.37 and 0.23 m from slope surface. Seasonal variation was represented by
rainfall and actual evaporation calculated based on data collected by a weather station installed at the site. The field monitoring shows that the
soil water content and pore-water pressure measurements by TM1-SM1, TM2-SM2, TM3-SM3 were representative of soil-water characteristic
curve (SWCC) of the residual soil. On the other hand, the measurements by TM4-SM4 shows a lower pore-water pressure during dry period
and a lower volumetric water content during wet period as compared to SWCC of residual soil. The study showed that the response recorded
by TM4-SM4 was representative of top soil used as media for vegetation planting. The study also showed that bigger variation of pore-water
pressure recorded by TM4 in December 2016 and January 2017 was due to long dry period with high temperature which is not normal in
Singapore during these months.
Keywords: Field monitoring, soil water content, pore-water pressure, soil-water characteristic curve, vegetated soil slope.
597
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
2. METHODOLOGY
2.1 Field Instrumentation
The investigated slope in this study is located at the central part of
Singapore within residual soil from Bukit Timah Granite. Figure 1
shows diagram of residual soil slope and the field instrumentation
carried out in this study. The instrumentations can be divided into two
parts; the first part was weather instrumentation for climatic data
measurements while the second part was instrumentation for
quantifying soil response to the flux boundary conditions. The
monitoring period was one year (1st July 2016 to 30th June 2017).
The weather instrumentation includes tipping bucket rain gauge Figure 2 Data acquisition system used at the study site
for rainfall, pyranometer for solar radiation, wind monitor for wind
speed and direction, as well as temperature and relative humidity 2.2 Soil Properties
probes. The weather station for climatic data measurement was
The slope was formed by residual soil with a thin layer of top soil as
located near the toe. A 10 m high galvanized mast was erected on a
media for growing vegetative cover. The soil properties required for
concrete footing. The wind monitor, the solar panel, the lightning
this study were investigated by performing laboratory tests on
protection rod and the pyranometer were installed on top of the mast.
samples taken from the field. The grain size distribution shows that
The other instruments were installed at a height of 1 m from the
about 50% of the material is clay. The unified soil classification
ground surface together with the data logger. Soil temperature (ST)
system USCS (ASTM D2487-00) categorized the soil to be highly
was measured at depths of 0.10m, 0.15m, 0.20m, and 0.25m.
plastic clay (CH). The bulk density of the residual soil was 1.8 Mg/m3.
The tensiometers (TM) and soil moisture sensors (SM) were
The saturated permeability of the soil was determined using the
installed within residual soil layer at depth of 2 m from crest and at
flexible-wall saturated permeability test (ASTM D5084-10). The
distances 0.4 m; 2.4 m; 3.8 m and 4.6 m from slope face. The jet-fill
saturated permeability of the soil is 6×10-7m/s. In order to perform
tensiometers were calibrated properly prior to installation and then
analyses related to water flow in unsaturated soil, permeability
inserted into a tube to the specified depth. The reading from
functions of the soil need to be obtained. Permeability function can
transducer was verified by a bourdon gauge attached to it. The
598
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
3. RESULTS AND DISCUSSION Figure 6 Mean monthly and number of rainfall days
3.1 Rainfall
3.2 Climatic data
Figure 5 shows daily rainfall recorded from 1st July 2016 to 30th June
2017. The cumulative yearly rainfall was 2819 mm which is higher The climatic data observed from the weather station from July
than the average annual rainfall in Singapore based on long term 2016 to June 2017 including air temperature (Ta), relative humidity
record from NEA Singapore (1981–2010) i.e. 2166 mm. The number (RH), solar radiation (SR) as well as wind speed (WS) are shown in
of rainfall days during the monitoring period was 178 days, which Figures 7. The climatic data obtained from the study site were also
was higher than the average annual number of rainfall days in compared with the typical climatic data reported by NEA Singapore
Singapore (167 days). Rainfall monitoring at the study site indicates based on data from 1981 to 2010.
that the monthly rainfalls are quite different from the typical trend in The minimum and maximum air temperature during the
Singapore as shown in Figure 6. monitoring period was 22.9 and 36.9oC respectively. The range of air
For example, the monthly rainfall in December 2016 (214.3 mm) temperature was wider and higher than the mean range of air
was too low as compared to the mean monthly rainfall in Singapore temperature in Singapore i.e. 23–33oC. As for the rainfall, the
for December 318 mm). On the other hand, the monthly rainfall in variation of air temperature during the monitoring period was
May 2017 (319 mm) was very high as compared to the average different from the typical values recorded in Singapore. Both
rainfall in Singapore for May (171 mm). The monthly rainfall in maximum and minimum temperature measured in this study occurred
September, October and November 2016 were higher than the mean in January 2017. This is in accordance with the rainfall condition in
monthly rainfall for the months in Singapore. November 2016 was which January 2017 represent an extreme condition. Past data
the wettest month during the monitoring period with 21 rainfall days indicated that the maximum temperature usually occur in March to
accumulating to 311 mm rain and while August 2016 was the driest May. In this study, the trend of higher temperature in March to May
month with cumulative monthly rainfall of 98.6 mm. Figure 5 shows was overcome with heavy downpour (Figure 6).
that the maximum daily rainfall occurred on 23rd January 2017 i.e. Soil temperature was measured at depths of 0.10, 0.15, 0.20 and
103.8 mm. The month of January 2017 represented extreme condition 0.25 m from ground surface. Field measurement performed in this
because it started with a two-week dry period followed by very wet study indicated that the soil temperature varies with depth and the
period towards the end of the month. time of the day. The least variation with the time of the day was
obtained in the measurement of soil temperature at depth of 0.20 m
from ground surface. Figure 7a shows the variation of soil
temperature measured at depth 0.20 m as compared to maximum and
minimum daily temperature. The soil temperature was lower in
November, December and January, and higher in May, June and July.
Relative humidity plays an important role in the prediction of ET.
The minimum and maximum relative humidity recorded in the study
site were 67.4% to 99% with mean value of 80%. The recorded RH
599
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
was higher than the typical range recorded in Singapore i.e. 60 to 3.3 Pore-water pressure measurements
95%. Maximum relative humidity occurred in November due to high
number of days with rainfall. The lowest relative humidity was The response of soil slope to rainfall and evapotranspiration from
recorded in August due to the least number of rainfall days. Figure 7b vegetated surface in terms of pore-water pressure is presented in
shows the maximum and minimum daily RH during the monitoring Figure 9. The figure shows that the measurements at TM1 – TM3
period. were quite consistent, while measurements at TM4 varied
The minimum and maximum solar radiation during the significantly especially from November 2016 to January 2017. Based
monitoring period was 1.12 and 14.76 MJ/m2/day respectively with on the evaluation of rainfall and climatic data, this period represents
mean value of 7.93 MJ/m2/day. These values are the typical of extreme changes from dry to wet condition.
Singapore. The maximum solar radiation was recorded in the dry As shown in Figure 1, TM1 and TM2 have vertical distances of 2
month of August 2016. The variation of solar radiation at the study m from crest while TM3 was 1.37 m below slope face. TM4 was very
site during the monitoring period is shown in Figure 7c. close to slope face i.e. 0.4 m horizontal distance or only 0.23 m
The maximum wind speed (4.92 m/s) was recorded in February vertical distance. Thus, TM4 is more affected by flux boundary
2017. The recorded wind speed was lower than the range of wind conditions as compared to the other tensiometers.
speed in Singapore i.e. 0–13.375 m/s with mean value of 2.65 m/s.
This may be because the location of the study site was surrounded by
high rise buildings. The variation of wind speed throughout the year
was also presented in Figure 7c.
600
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
coefficient of permeability of the top soil was 1 × 10-5 m/s which is Allen R.G., Pereira L.S., Raes D., Smith M. (1998) Crop
higher than that of the residual soil (6 × 10-7 m/s). evapotranspiration: guidelines for computing crop water
The field data from TM4-SM4 were plotted on the SWCC curves requirements Food and Agriculture Organization of the United
of the top soil as shown in Figure 11. The figure shows that the field Nation, Rome.
data could be plotted well inside the hysteresis of the SWCC of the Blight G.E. (1997) The “Active” Zone in Unsaturated Soil
top soil. The range of suction was i.e. 8 – 90 kPa, while the range of Mechanics. 1st GRC Lecture, Nanyang Technological
volumetric water content was from 13 to 30%. The maximum University, Singapore.
volumetric water content was less than the saturated volumetric water Childs E.C., Collis-George G.N. (1950) “The permeability of porous
content of the surface soil. It may be attributed to some water materials”. Proc. Royal Society of London 210A. pp. 392-405.
intercepted by vegetation (grass) at the surface. The same reason that Chipp, P.N., Henkel, D.J., Clare, D.G. & Pope, R.G. (1982) Field
the evapotranspiration (ET) is less than the actual evaporation (AE) Measurement of Suction in Colluvium Covered Slopes in
from base soil surface. Hong Kong. Proc. 7th Southeast Asian Geotechnical Conf.,
Hong Kong, pp.49-62.
Fredlund D.G., & Xing A. (1994) “Equations for the Soil-Water
Characteristic Curve”. Canadian Geotech. Journal 31 pp. 521-
532.
Fredlund, D. G. & Rahardjo, H. (1993). Soil Mechanics for Un-
saturated Soils. New York: John Wiley & Sons, Inc.
Joel, A., Messing, I., Seguel, O., & Casanova, M. (2002) “Meas-
urement of surface water runoff from plots of two different
sizes”. Hydrological Processes. 16(7) pp. 1467-1478.
Lazzara P. & Rana G. (2010) “The use of crop coefficient approach
to estimate actual evapotranspiration: a critical review for
major crops under Mediterranean climate”. Italian Journal of
Agrometeorology 2 pp. 25-39.
Figure 11 Plot of pore-water pressure and volumetric water content Lee, M.L. Gofar N. & Rahardjo H. (2009) “A Simple Model for
measured by TM4 from July 2016 to June 2017 in SWCC of top Preliminary Evaluation of Rainfall-Induced Slope Instability.
soil. Engineering Geology”. 108(3-4) pp. 272-282.
Li, A.G., Yue, Tham, L.G. and Lee C.F., (2005) “Field-monitored
variations of soil moisture and matric suction in a saprolite
4. CONCLUSIONS slope”. Canadian Geotechnical Journal. 42, pp. 13–26.
Field measurements of flux boundary conditions and soil response Gasmo, J.M., Rahardjo, H. and Leong, E.C. (2000) “Infiltration
were carried out in a residual soil slope at the central part of Effects on Stability of a Residual Soil Slope”. Computer and
Singapore. The measurements presented in Figures 5 and 7 specified Geotechnics, 26, April, pp.145–165.
that the seasonal variation of rainfall was quite different from the Gitirana Jr., G., Fredlund, D.G., Fredlund, M. (2006) “Numerical
typical trend in Singapore (Figure 6). The dissimilar trend also modeling of soil-atmosphere interaction for unsaturated
observed for the other flux boundary conditions such as air surfaces”. Proc. 4th Intl. Conf. Unsaturated soils, US., pp.658-
temperature and relative humidity. Highly variable conditions 669.
occurred during the months of December 2016 and January 2017. Gitirana Jr., G., Fredlund, D.G., Fredlund, M. (2005) “Infiltration-
This could be identified by unusually high air temperature and long runoff boundary conditions in seepage analysis”. Proc. 58th
period of dry condition which are different from the normal Canadian Geotech. Conf. Canada, pp.516-523.
conditions in Singapore. The long dry period allowed high pore-water Gofar N. Lee, M.L. & Kassim, A. (2008) “Response of Suction
pressure to be recorded by TM4 which is located closest to the ground Distribution to Rainfall Infiltration in Soil Slope”. Electronic
surface (Figure 9). J. of Geotechnical Engineering. 13 (E), pp.1-13
Readings of pore-water pressure and soil water content recorded Ng, C.W.W., Springman, S.M. and Alonso E.E. (2008). “Monitoring
by TM1-SM1, TM2-SM2, TM3-SM3 are representative of the SWCC the Performance of Unsaturated Soil Slopes”. Geotechnical
of the residual soil. Thus, the readings are representative of the soil and Geological Engineering. 26(6), pp.799-816.
where the instrumentations were installed. However, the plot for Ng. C.W.W., Zhan, L.T., Bao, C.G., Fredlund, D.G. and Gong, B.W.,
TM4-SM4 located at vertical distance of 23 cm from ground surface (2003) “Performance of an Unsaturated Expansive Soil Slope
was more representative of top soil which has lower density and Subjected to Artificial Rainfall Infiltration”. Géotechnique.
higher coefficient of permeability. This showed that the readings of 53(2), pp. 143-157.
instrumentation closer to the ground surface was affected by the flux Penman, H.L (1948) “Natural evapotranspiration from open water,
boundary conditions and the consistency of surface soil as well as bare soil and grass”. Proc. Royal Society London Ser. A.193,
vegetative cover. pp.120-145.
Rahardjo, H., Amalia, N., Leong, E.C. Harnas, F.R. Lee, T.T. and
4. REFERENCES Fong, Y.K. (2017) “Flux Boundary Measurements for the
study of Tree Stability”. Landscape Ecol Eng. 13:81–92
Rahardjo, H., Satyanaga, A. Harnas, F.R. and Leong, E.C. (2014)
ASTM D5084-10 Standard Test Method for Measurement of “Comprehensive Instrumentation for Real Time Monitoring of
Hydraulic Conductivity of Saturated Porous Material using a Flux Boundary Conditions in Slope” Proc. 3rd Italian
Flexible Wall Permeameter, ASTM International, West Workshop on Landslides, Italy.
Conshohoken. PA. Rahardjo, H., Satyanaga, A. and Leong, E.C. (2013). “Effects of Flux
ASTM D6838-02 Standard Test Method for Determination of SWCC Boundary Conditions on Pore-water Pressure Distribution in
for Desorption using Hanging Column, Pressure Extractor, Slope”. Engineering Geology, 165, October, pp. 133–142.
Chilled Mirror Hygrometer or Centrifuge, ASTM Rahardjo, H., Lee, T.T. Leong, E.C. and Rezaur, R.B. (2005).
International, West Conshohoken. PA. “Response of a Residual Soil Slope to Rainfall”. Canadian
Geotechnical Journal, 42(2), pp. 340–351.
ASTM D 2487-00 Standard Practice for Classification of Soils for Rahardjo, H., Leong, E.C., Gasmo, J.M. and Tang, S.K. (1998)
Engineering Purposes (Unified Soil Classification System) , “Assessment of rainfall effects on stability of residual soil
ASTM International, West Conshohoken. PA.
601
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
slopes”. Proc. 2nd Intl. Conf. Unsaturated Soils, China, pp. 280-
285.
Sattler, P.J. & Fredlund, D.G. (1991). “Modelling vertical ground
movements using surface climatic flux”. Proc. Geotechnical
Engineering Congress, US., pp. 1292-1306.
Yunusa, G.H. Kassim, A. & Gofar, N. (2014) “Effect of Surface Flux
Boundary Conditions on Transient Suction Distribution in
Homogeneous Slope”. Indian J. Science & Technology, 7(12),
2064–2075
Zotarelli L, Dukes MD, Romero CC, Migliaccio KW, Morgan KT
(2009) Step by step calculation of Penman-Monteith
evapotranspiration (FAO-56 Method), University of Florida.
602
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Expanded polystyrene (EPS) yang disebut Geofoam merupakan material pengisi yang sangat ringan dengan berat kurang dari
satu persen dari berat tanah. Pada daerah yang mengalami kesulitan untuk tanah timbunan dengan volume cukup besar, maka material
pengisi dapat digunakan sebagai alternatif tanah timbunan. Penggunaan Geofoam sebagai material pengisi yang ringan perlu dikaji
kesetaraanya sebagai material pengganti tanah timbunan, dalam hal ini akan diteliti efek Geofoam sebagai material pengisi terhadap nilai
CBR Laboratorium pada tanah kohesif di Bandung. Pada Penelitian ini dilakukan uji California Bearing Ratio (CBR) dengan rendaman
(soaked) dan tanpa rendaman (unsoaked) pada lokasi di Rancaengkek, Gedebage dan Setiabudi sebagai sampel uji. EPS 29 digunakan
sebagai material pengisi dalam penelitian ini dengan spesifikasi kepadatan 28.8 kg/m3. Dari hasil Uji CBR dengan standard Proctor nilai
CBR tanah asli kondisi unsoak menggunakan material pengisi EPS 29 mengindikasikan peningkatan 50 – 70% dan peningkatan 15 – 150%
pada kondisi soak, hal ini terjadi pada tanah kohesif dengan plastisitas tinggi. Tanah dengan plastisitas sedang dan berpasir tidak memberikan
peningkatan nilai CBR, tapi dapat disimpulkan bahan EPS Geofoam dapat digunakan sebagai bahan pengisi .
Keywords: EPS Geofoam, CBR , material pengisi.
603
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Di lapangan geofoam tidak langsung menerima beban, karena Berdasarkan dari grafik plastisitas Casagrande dan ukuran butir
hanya merupakan material pengisi, umumnya diatas geofoam (gambar 4), seluruh sampel termasuk katagori tanah lanau dengan
diberikan lapisan perkerasan berupa beton atau aspal dan plastisitas tinggi.
dibawahnya merupakan lapisan pasir dan geotextile.
3 LOKASI SAMPEL
Ketiga sampel diambil dari daerah Bandung yaitu dari Setiabudi,
Gedebage dan Rancaengkek (Gambar 3). Ketiga sampel merupakan
tanah kohesif. Klasifikasi berdasarkan ukuran butir sampel tanah
dapat dilihat pada tabel 3.
604
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Hasil kompaksi dari tiga sampel dengan nilai berat isi kering Berdasarkan persamaan 2 untuk energi pada kondisi tanah dan
maksimum dan kadar air optimum pada ketiga sampel tanah kondisi tanah dan geofoam sebagai material pengisi, maka:
tersebut dapat dilihat pada gambar 5, 6, 7. tumbukan hammer perlapis pada tanah saja 10 tumbukan
menjadi 15 tumbukan tanah dan geofoam;
Tabel 4 Hasil Kompaksi Tiga Sampel Dengan Standard Proctor tumbukan hammer perlapis pada tanah saja 25 tumbukan
menjadi 38 tumbukan tanah dan geofoam;
Sampel tanah ɣd maximum (gr/cm3) Woptimum (%)
tumbukan hammer perlapis pada tanah saja 56 tumbukan
Setiabudi 1.295 36.4 menjadi 84 tumbukan tanah dan geofoam.
Rancaekek 1.360 29.8
Gedebage 1.275 34.8 Pada kondisi A dengan meletakkan geofoam pada lapis tengah,
nilai CBR meningkat namun tetap jauh lebih rendah dibandingkan
dengan nilai CBR tanah asli. Oleh karena itu dilakukan revisi
5.4 CBR (California Bearing Ratio) peletakan EPS Geofoam diletakkan pada lapis bawah kondisi B.
Dengan acuan kadar air optimum dari hasil kompaksi tanah Geofoam dapat merendam energy tumbukan yang diberikan pada
dilakukan uji CBR untuk masing masing sampel tanpa geofoam saat proses pemadatan dan hasil lebih tinggi dari nilai CBR tanah
dengan kondisi unsoak dan soak. tanpa geofoam. Pada kondisi B lokasi geofoam berada di lapis
bawah, karena geofoam tidak bisa langsung menerima beban
sebuah ketentuan Schmertmann yang sering dijadikan pegangan
adalah bahwa distribusi beban pada tanah sampai kedalaman 2 x
diameter bidang tekan yang mengalami tegangan. Jika luas bidang
tekan (piston CBR) 3 in2, maka diameter piston sebesar 1.732 in ≈
4.4 cm. Sehingga sampel tanah dalam mold CBR yang mengalami
tegangan sampai pada kedalaman 8.8 cm, posisi lapis 2 terdalam
pada 8.6 cm pada mold. Perletakan EPS Geofoam pada lapisan ke-3
hanya bersinggungan 2 mm dari zona tegangan.
Gambar 8 Uji CBR tanah metode Standard Proctor 5.4.1 Uji CBR dengan geofoam kondisi B
5.4.1.a Tanah Setiabudi
Pada uji tanah tanpa geofoam, metode Standard Proctor adalah
sebagai berikut :
- Tumbukan dengan 10 kali tumbukan /lapis,
- Tumbukan dengan 25 kali tumbukan/lapis,
- Tumbukan dengan 56 kali tumbukan/lapis.
Energi Standard atau Modified Proctor tergantung pada berat
hammer, tinggi jatuh Hammer, jumlah lapis pemadatan dan jumlah
tumbukan tiap lapis.
N n W H
Energy (1)
V
Dimana :
N = jumlah tumbukan hammer perlapis dalam mold;
(a) (b)
n = jumlah lapis dalam mold;
W = berat hammer;
Gambar 10 (a) Nilai CBR tanah tanpa geofoam (Setiabudi), (b)
H = tinggi jatuh hammer;
Nilai CBR tanah + Geofoam (Setiabudi)
V = volume mold = volume sample.
Nilai CBR sampel tanah Setiabudi tanpa geofoam kondisi unsoak
Karena Geofoam sebagai material pengisi, sehingga geofoam
pada 95% ɣ dry maximum dari hasil uji kompaksi diperoleh 15.8%
tidak terkena beban langsung, jadi geofoam diletakkan coba coba
dan nilai CBR soak diperoleh 7.6%. Nilai CBR tanah Setiabudi +
pada lapis kedua dan ketiga.
Geofoam kondisi unsoak pada 95% ɣ dry maximum diperoleh 14.2
% dan nilai CBR kondisi soak 7.3 %.
Nta n W H Ntag n W H
(2)
V V 5.4.1.b Tanah Rancaengkek
Dimana :
Nta = jumlah tumbukan hammer pada tanah asli;
Ntag= jumlah tumbukan hammer pada tanah dan geofoam.
605
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Nilai CBR Tanah Rancaengkek unsoak dengan kondisi 95% ɣ dry Hidrostatis uplift :
maximum dari uji kompaksi sebesar 7.9% dan kondisi CBR soak Karena EPS geofoam mempunyai berat sangat ringan dan
adalah 1.1 %. Nilai CBR tanah Rancaengkek + Geofoam unsoak kepadatan rendah. Timbunan tanah harus dihitung untuk gaya
sebesar 11.8 % dan nilai CBR kondisi soak adalah 2.75 %. hidrostatis uplift.
7 KESIMPULAN
EPS Geofoam dengan berat dan kepadatan kecil tidak bisa
langsung menerima beban,utama, hanya sebagai material
(a) (b) pengisi.
Gambar 11 (a) Nilai CBR tanah (Gedebage), (b) Nilai CBR tanah + EPS Geofoam akan effektif jika diletakkan pada kedalaman
Geofoam (Gedebage) kurang lebih dua kali zona tegangan akibat beban utama.
Dari hasil nilai CBR unsoak dan soak, EPS Geofoam dapat
Tabel 5 Nilai CBR Tanah Tanpa Geofoam dan Tanah + Geofoam digunakan sebagai alternatif material pengisi pada tanah
kohesif, hal ini sesuai nilai CBR dengan EPS Geofoam
Sampel Nilai CBR mendekati atau lebih besar nilai CBR tanah tanpa geofoam.
Tanah Unsoaked Soaked Pada kondisi muka air tinggi, harus diperhitungkan tinggi
Tanah Tanah + Tanah Tanah + tanah diatas geofoam agar dapat mengimbangi adanya gaya
Geofoam Geofoam uplift akibat tekanan hidrostatik.
Setiabudi 15.8 14.2 7.6 7.3
Rancaengkek 7.9 11.8 1.1 2.75
Gedebage 7 12.4 1.9 2.2
Dari Table 5, terlihat Tanah Setiabudi nilai CBR tanah + 8 DAFTAR PUSTAKA
Geofoam kondisi unsoak dan soak agak mengecil, jika diperhatikan
Tanah Setiabudi mempunyai persentase pasir lebih banyak diantara American Society for Testing Materials, “ Standard Test Method for
dua sampel yang lain. CBR ( California Bearing Ratio) of Laboratory-compacted
Soils”, ASTM designation: D1883-07e2. Philadelphia, Pa.,
6 APLIKASI GEOFOAM MATERIAL PENGISI 309-310, 2007.
TIMBUNAN DI LAPANGAN Beton Elemindo Perkasa ( BEP ), “Calculation B-Foam Geofoam-
Grade 29 as fill material Embankment“ , Bandung Indonesia,
2015.
Reynaud, David, ”Guidelines for Geofoam Applications in
Embankment Projects”, NCHRP Project 24-11(02), January
2013.
Stark, Timothy, “Design Procedure for Geofoam Applications in
Embankment Projects”, Illinois , August 2003.
Stark, Timothy, “Guidelines and Recommended Standard for
Gambar 12 Aplikasi Geofoam sebagai timbunan Geofoam Applications in Highway Embankments”, NCHRP
(BEP, Bandung-Indonesia) Report 529, July 2004.
606
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: Piezocone test (CPTu) is one of the most versatile tools for geotechnical investigation. It is commonly used for soil type
identification, soil stratification, and geotechnical parameters determination. In soft soils, the CPTu test is very important because of its
capability of pore pressure measurement. This paper presents the result of CPTu test in the soft soil after pile driving works. The dissipation
test results are evaluated to distinguish between the excess pore pressure induced by cone penetration and by pile driving works. The results
show that the generated excess pore pressure induced by pile driving is up to 1.27 to 1.40 times the effective overburden pressure. In
addition, the dissipation of the excess pore water pressure is observed by conducting another CPTu test five months later after the first test.
The result shows that the excess pore pressure due to pile driving still exists, with the remaining value of 0.25 times the effective overburden
pressure.
KEYWORDS: CPTu, pile driving, soft soil, excess pore water pressure
607
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 3 Summary of Measured Pore Pressure (Poulos and Davis, Figure 4 Extrapolation of Dissipation Test Curve (Rahardjo et al.,
1980) 2008)
608
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
Figure 6 Comparison of Pore Pressure Measurement and Normalized Excess Pore Pressure (Recalculated from Rahardjo et al, 2014)
The value of excess pore pressure due to pile driving then can be From the results, it can be concluded that the excess pore
calculated by subtracting with the value of hydrostatic pressure. pressure ratio in October 2012 (CPTu PC-01 and PC-02) is as high
Furthermore, the ratio between existing pore pressure and effective as 1.40 σ’v, while in March 2013 (CPTu-01 – CPTu-03) is much
overburden pressure is also calculated. Table 2 summarizes the less, with the value of 0.25 σ’v. These obtained values agree well
calculated value of excess pore pressure ratio after pile driving. with the previous publication from Poulos and Davis (1980), where
the generated excess pore pressure due to pile driving can be as high
as 2.0 σ’v.
609
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
where: 8. CONCLUSIONS
uf = residual excess pore pressure (obtained from
CPTu test has proven its ability to measure the pore pressure in the
extrapolation dissipation test)
soil strata, as well as the generated excess pore pressure due to pile
uo = hydrostatic pressure
Δu2/σ’v = normalized excess pore pressure driving by conducting the dissipation test. The dissipation test
during CPTu can be conducted easily and generally fast. In addition,
the dissipation test is so flexible that can be conducted at several
Figure 6 shows the pore pressure measurement of CPTu test in
depths during a test.
October 2012 and in March 2013, along with the calculated
normalized excess pore pressure. Furthermore, the extrapolation of the dissipation curve can be
done easily to calculate the proportion of the excess pore pressure
From the results, it can be seen that one day after the pile driving
due to cone penetration and due to pile driving.
works finished (October 2012), the value of the excess pore pressure
due to pile driving is as high as 2.0 times the effective overburden
pressure. While from the CPTu test conducted five months later 9. REFERENCES
(March 2013), it can be seen that the excess pore pressure due to
pile driving is not fully dissipated. The remaining excess pore Geotechnical Engineering Consultant. 2012. Laporan Hasil Uji
pressure due to pile driving is relatively low, with the value of 0.20 Geoteknik Lapangan: Uji Piezocone/CPTu Proyek Tunjungan
– 0.25 σv’. Plaza 5 Surabaya (In-situ Testing Result Report: Piezocone
Test/CPTu of Project Tunjungan Plaza 5 Surabaya).
6.3 Degree of Dissipation Bandung: GEC (In Indonesian)
Lo, K.Y., and Stermac, A.G. 1965. “Induced Pore Pressures During
The degree of excess pore pressure dissipation (U) can be estimated Pile-Driving Operations”. Proceeding of the International
using the dissipation test results, following the equation as follows: Conference of Soil Mechanics and Foundation Engineering.
߂ݐݑ (2) Poulos, H. G., and Davis, E. H. (1980). Pile Foundation Analysis
ܷ =1− and Design. John Wiley & Sons, Inc.
߂݅ݑ
Rahardjo, P.P., Anggoro, B.W., Yakin, A.Y., and Darmawan, H.
2008. “Determination of the Degree of Consolidation of
where: Δut = excess pore pressure due to pile driving at time t Reclaimed Site on Deep Soft Mahakam Deltaic Soils using
Δui = excess pore pressure due to pile driving at the initial CPTu”. Proceeding of the Fourth International Symposium on
time after pile driving finished Deformation Characteristics of Geomaterials, Atlanta, GA,
883-889.
In this particular study, the CPTu test is not conducted directly Rahardjo, P.P., Wirawan, A., and Setiawan, R. 2014. “An Analytical
after the pile driving works finished, but conducted one day after. and Empirical Study on Ground Movement and Excess Pore
Therefore, the calculated degree of dissipation will be based on this Pressure Generation Due to Pile Driving in Soft Soils”.
available data, which will produce an overestimated value. Figure 7 Proceeding of the International Conference on the Soft Soils
shows the calculation steps to determine the degree of dissipation, and Ground Improvement, Bandung, Indonesia, A8-1 – A8-
obtaining 80.6% dissipation degree in October 2013. 13.
610
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ABSTRACT: The 3rd phase MRT underground project of Bangkok that is being implemented has faced increased difficulties in design and
construction owing to the congested corridor along the alignment and the phenomenon of groundwater rebound in Bangkok subsoils in the
aftermath of the end of land subsidence era from deep well pumping. The rebound is more than 10 meters from the situation during the
initial phase construction and thus it necessitates careful soil investigation and design approach to prevent instability of groundwater ingress
and uplift in bore tunnelling and excavation of station boxes. The tight space along the MRT underground alignment resulted in the twin
bored tunnels being placed at very small clearance in a few locations. It creates high risk in the shield driving, particularly in the saturated
sand layer. The works also involve tunnel driving at close proximity or under a number of existing buildings or road viaducts. In such cases,
various schemes of underpinning are employed, and one of them is strengthening of founding soil layers by cement jet grouting. An
intensive instrumentation program according to the observational design method is employed to safe guard against occurrence of any
undesirable instability in the excavations and impacts to the third party properties.
Keywords: MRT Underground, shield tunneling, deep excavation, underpinning, groundwater rebound, the observational method
611
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
ideal condition for tunnelling. However, the soft soil condition Variation of Piezometric Pressure Profile with
necessitates the use of shield method for tunnel construction. The Elevation
(Standpipe Piezometer)
shallow zone consists of a 12-15 m thick layer of soft to medium
stiff clay (Bangkok clay) that is followed by a layer of stiff to hard Short Term Design Piezometric Profile
clay and a layer of sand (the First Stiff Clay and the First Sand). 100 Long Term Design Piezometric Profile
Below them, alternating layers of stiff to hard clay and dense to very Existing Tender Orange Line (2005)
dense sand exist to a great depth. The sand layers are aquifers of Additional Tender Orange Line (2015)
Contract E1 Contract E2
70
Elevation, (m asl)
60
50
40
30
It is well known that Bangkok had suffered land subsidence Figure 1 - Variation of Piezometric Profile with Depth
problem owing to excessive deep well pumping from the sand Figure 3 Piezometric pressure in the subsoils at different times
aquifers for water supply. The decline in groundwater head in
shallow aquifer layers was as much as 25-35 m from the ground
surface during the peak of the crisis in early 1980s, when the
subsidence occurred at a rate as high as 120 mm/year. After the
surface water supply was made adequate in most of the city area
around the turn of the century, the situation improved as deep well
pumping diminished. Since then the rebound of groundwater started
to develop and the subsidence ceased in most city areas. The
rebound continues till present and the accumulative amount is about
10-12 m during the last 15 years. At present the piezometric water
head in the 1st and 2nd sand layers of Bangkok where the MRT
underground works are involved with is about 13-14 m depth below
the ground surface that is 10-12 m higher than the condition when
the 1st Phase MRT Underground line was constructed (Figure 3). It
has been risen at a rate of approximately 0.8-1.0 m per year (Figure
4). Because the groundwater law that has been in effect since 2004
totally bans deep well pumping in the entire area of Bangkok and Figure 4 Changes in piezometric heads in Bangkok sand layers
surrounding suburbs, it is foreseen that the ground water heads in
the sand layers might rebound to the hydrostatic level in the near 4. INCIDENTS IN PAST MRT UNDERGROUND WORKS
future.
In the past the subsidence situation created complication of long Problematic incidents related to the impact of groundwater rebound
term settlement and differential settlement in the design of MRT in the underground works of the past MRT project are presented.
underground structures. On the other hand, the condition of the The recently completed underground structures of the Blue Line
lowered groundwater head in the sand layers resulted in ease in deep Extension were placed in the stiff clay and sand layers as shown in
excavations for station construction and shield tunnelling. Incidents Figure 5. The deepest bored tunnel section is at 37 m depth at tunnel
of hydraulic instability were rare in past deep excavations and invert at a location near the river crossing.
tunnelling. However, at present the situation has changed since the
groundwater head started to show significant rebound about 15
years ago. Difficult and problematic incidents were experienced in
a number of recent deep excavations of building construction as well
as in the works of the Blue Line Extension Project. In addition, the
rise of groundwater head was linked to incidents of difficulties in
EPB shield tunnelling in the city including a severe incident in the
2nd phase MRT Underground project that led to a six-month delay of
the stall and repair of the TBM.
612
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
of the soil with the side of the diaphragm wall. The U-shaped
Facing the situation of higher water pressure in the 1st and 2nd grouted zone as shown in Figure 7 was adopted. The grouting work
sand layers than that indicated by early investigation, the detailed was launched underneath the ground surface from a platform at the
design of two of the three station boxes that were placed at large mid-height of the station height. The scheme was proven to be an
depths needed to be modified from those planned in the pre-tendered effective method to prevent the hydraulic uplift failure of the
design. This is to avoid problems of base instability during excavation base and it allowed a normal construction process of the
excavation from hydraulic uplift of the stiff clay at the excavation concrete base slab casting under a dry condition.
base. The excavation procedure adopted for the third station
remained the normal approach as the designer considered it was still
adequate. Unfortunately, after a portion of the base area was
excavated and concrete casting, the hydraulic uplift occurred in the
next stage of the base excavation that resulted in flooding and
difficulty in control of water ingress in the remaining stages of the
excavation. The work suffered almost 20 months delay in the
construction to fight with the persistent problem of groundwater
ingress despite extensive remedial grouting exercise. A brief outline
on the situations of hydraulic instability involved with the base
excavation design and construction of these stations is given in the
following sections.
The second station was Station BS13 which was 181 m long, 23.4
m wide and 30 m deep. The diaphragm wall was 1.2 m thick and 40
m deep. Cross section and subsoil is shown Figure 9. The
+101.5
piezometric head was observed at approximately 16.0 m below the
ground surface.
Soft Clay
GWL+86.3 +88.0
st
1 Stiff Clay
+77.5
1st Sand
+65.0
2nd Stiff Clay +63.0
2nd Sand
613
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
stages of a small sized excavation zone bounded by high soil BLE project that the EPB shield tunnelling had to be made through
surcharge + cast concrete slab around the perimeter of the the sand layer as deep as 37 m below the ground surface with the
excavation, the stability of the excavated base against the hydraulic piezometric head 10 m higher than that existed during the time of
uplift will be significantly assisted by the side shear resistance of the the first MRT shield tunnelling, an incident of serious water ingress
soil along the perimeter as compared with that of large excavation occurred in August 2013 during the early drive of an EPB shield
area. The Outline Design Specification stipulated at that in case through the interface between the 1st stiff clay and the 1st sand layer.
shear strength of base soil was considered, a minimum factor of The consequence was flooding and severe damage to the TBM that
safety was required at 1.20. This requirement could be satisfied for led to an interruption of the work for over 7 months. The incident
the soil and groundwater condition below the excavation base of the was an uncontrolled water ingress into the tunnel in the tail area of
station. The accurate determination of the density of the soil layers the EPB shield machine. Flooding and excessive movements of
and the spatial variation of the thickness of the stiff clay layer were segmental rings in vicinity of the shield were experienced. It was
a must in the adaptation of this method. The staged excavation reported the ingress initially occurred through a 20 mm wide gap
sequence adopted for BS13 excavation is shown in Figure 10. The formed at the circumferential joint of the key segment placed near
work was successfully completed as planned. the invert of the ring. The groundwater ingress from the 1st Sand
layer shortly led to piping of fine sand into the tunnel, and
subsequently excessive movements and distortion of several
segmental rings behind the shield. During the incident various
remedial grouting measures were urgently carried out to control the
ingress and avoid excessive distortion of the concrete segmental
rings that might ultimately result in a tunnel collapse.
It was obvious that the incident was caused by difficulty of the
EPB shield in controlling stability of the excavated tunnel face while
the tunnel was being excavated in the mixed face stiff clay and sand
layers. At that position the groundwater head in the sand layer was
17 m. The sand was in the lower part of the tunnel and as the
tunnel was progressively advanced, the portion of sand in face area
progressively increased as the TBM drive pitched down. This
condition was one of the most difficult tunnel grounds to deal with
in soft ground tunnelling. It was apparently that with high ground
water pressure in the sand layer in the excavated tunnel face, the
shield machine experienced difficulty in maintaining adequate face
Fig. 10 Sequence of staged base excavation of BS13 pressures. Thus over excavation would significantly occur. After a
careful examination on the TBM operation data it was believed that
Successful base excavation of the BLE Underground station was the triggering factor of the incident was an excessive displacement
not the case at BS11 Station which was the deepest excavation of the key segment of the segmental ring resulted from the over
among all MRT stations in Bangkok. The base of the excavation excavation of the mixed face soil by the TBM under the high
was at 32 m below the ground surface. Prior to the start of the groundwater condition.
excavation it was judged by the designer that the subsoil and
piezometric head in the sand layers data indicated that the base 5. ORANGE LINE EAST
excavation could be with adequate factor of safety against hydraulic
uplift. Unfortunately, after the excavation was made to the lower The 3rd Phase MRT Underground work that is under construction at
level of the bracing of the D-wall and base excavation and concrete this stage is a part of the Orange Line East project (Cultural Center
base slab casting was progressed from one end of the station box to to Minburi) that is implemented as design-build contracts. The
about 15 percent of the length of station, significant water ingress underground portion constitutes 12.1 km length out of the total 21.1
occurred that seriously affected the work. The whole station was km length of the line. There are 10 underground stations along the
flooded by 2-m-depth of water within 12 hours. Extensive underground section that the works are administered in 3 contracts
dewatering pump system was installed and a comprehensive (E1-E3). The underground line is mainly constructed by bored
program of chemical and cement grouting subsequently employed. tunneling using 3 EPB shields except for a short cut-and-cover
The amount of ingress was as high as 1400 cubic meter/hour at one startup section. Owing to the tight space along the corridor of the
stage. Groundwater ingress repeatedly occurred in the subsequent alignment of the line that traverse along two most congested roads
stages of excavation for concrete casting that altogether delayed the of the city that also have foundation of elevated roadway along the
work by 20 months. The investigation pointed that there might be a right or way, the MRT underground twin tunnels are mostly placed
condition of lateral non-continuity of the 2nd stiff clay below the in the vertically stacked configurations and in many places they
excavation base that in the initial design was considered sufficiently need to be placed at very close proximity to each other, particularly
thick and covering the entire area of the excavation base. This when leaving or entering the station boxes (Figures 2 and 11). The
incident of severe and prolonged groundwater ingress also led to tunnels also pass very near to or cross foundations of existing
subsidence problem of the surrounding ground and buildings that buildings or other structures, some of which need underpinning
necessitated extensive stabilizing grouting works throughout the works. The close proximity shield tunneling works for the deeper
period. section of alignment in the sand layer of high groundwater pressure
for the shallow section in soft clay layer necessitate special
4.2 Water ingress in EPB shield tunnelling considerations in the design and construction planning to avert
instability problems during construction and to ensure stability of
Prior to the tunnelling in the Blue Line Extension project, bored the completed structures in long term. Because of the adoption of
tunnelling by means of EPB shields and concrete segmented ring the vertically stacked configuration of the running tunnels most of
support had been commonly used with great success including those the underground stations are excavated at large depth. And with the
in the Blue Line Initial System project, in which the tunnelling was very extraordinary thickness of the 1st sand layer in the subsoils in
made deep in the sand layer at depth as much as 25 m below the the considerable part of the alignment, careful considerations are
ground surface. There had been no incidents of groundwater needed for the design and planning of the station excavation to avert
problem in EPB shield driving in Bangkok sands. However, in the potential basal instability problems as earlier experienced in the
previous project.
614
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
OR14
OR15
Tunnel Profile-Contract E1
Figure 11 Vertically stacked alignment of tunnels in Contract E1
The design of the bored tunnelling and concrete segment lining has
to deal with the situation of very close proximity placement of the
twin tunnels in various places. The clear distance between the
tunnels is as close as 1.6 m which is only 25% of the tunnel
diameter (e.g. Figure 12). For such conditions where the tunnels are
to be excavated in the 1st sand layer, special measures are provided
to avert potential instability problems to the lining of the firstly
excavated tunnel from the TBM weight and any unexpected
excessive over-excavation during the excavation of the second
tunnel. The protective measures generally consist of ground
improvement by jet cement grouting. Where the ground space does
not permit the launching of ground improvement, other measures are
used. They include installation of steel ring bracing inside the
concrete segmental lining of the first driven tunnel, the careful
driving of the shield, and the close monitoring of the movements of
the lining and ground.
1.6 m
Close proximity tunneling of the works also need to deal with the
situation of potential impacts to the piled foundation of the existing
structures and vice versa the impact of the load to the stability of the
tunnel lining. Extensive numerical analyses are employed to check
the situation of all cases in details and necessary preventive and
corrective measures are provided in the design. In some cases, the
excavation may be made without the need for auxiliary measures,
but in some cases, ground improvement by jet cement grouting is
adopted (Figure 13). In the worst case, underpinning with pile
replacement needed adopted.
Besides the cases of close proximity tunneling, there is a case of
the station box excavation placed very close to the piers of the
elevated expressway at a clear distance between the diaphragm wall
and the foundation piles of the pier as small as 2.7 m. In this case
the impact assessment with the aid of rigorous numerical analysis
(PLAXIS) suggests implementation of the mitigation measure to Figure 14 Pre-installed struts for station box excavation next to
control the lateral movement of the viaduct pier and induced flexural expressway viaducts
stress in the piles within the allowable limits by using preinstalled
struts for the station excavation at two levels by means of jet cement
615
Proceeding 20th SEAGC - 3rd AGSSEA Conference in conjunction with 22nd Annual Indonesian National
Conference on Geotechnical Engineering. Jakarta - INDONESIA, 6 -7 November 2018. ISBN No. 978-602-17221-6-9
616