Sei sulla pagina 1di 10

The HKIE Geotechnical Division Annual Seminar 2017

Finite Element Analysis of Deep Excavations


in Soft Ground
Leslie H. Swann & Ci Wang
Jacobs China Ltd.

ABSTRACT

Excavations in reclaimed ground in urbans need to be designed to control ground settlement


to limit damage to adjacent structures. As 3D finite element models become more user
friendly, the benefits of using 3D models to design excavation and lateral support can be
significant. This paper describes how 3D analysis can be used to produce economic design
solutions which are impossible with conventional 2D analysis.

1 INTRODUCTION

Evaluating the magnitude and distribution of ground movement adjacent to a deep excavation is an
important part of the design process when excavating in an urban environment. The magnitude
of ground movement introduced by a deep excavation depend on numerous factors, such as the soil
and groundwater conditions, excavation geometry, surcharge loads, existence of adjacent structures,
construction method, stiffness and penetration of the retaining walls, type and installation method
of the props, etc. Furthermore, in certain excavations, the distance from the evaluated section to the
corner of the excavation also affects the magnitude and distribution of ground movement. The effect
of the section’s position on the displacement is known as the “three-dimensional” effect or the “corner
effect”, whereby ground movement decreases near the corners of the excavation due to the stiffening
effects of the corners. This corner effect of ground displacement may not correspond to that found in
conventional plane stain (2D) simulation of excavation.
This paper presents a design development of a particular section of a cut and cover tunnel at tender
design stage from the numerical modeling aspect, concentrating on demonstrating the benefit of adopting
3D finite element analysis to mobilizing the most of corner effect of ground movement assessment, and
comparing it to the 2D finite element analysis proving the benefits of the 3D modeling for potential
economical engineering design.

2 Project Background

The Government of the Hong Kong SAR was proposing to build a new highway, known as Route 6, to
connect West Kowloon with Tseung Kwan O via a series of tunnels, viaducts and at-grade roads. A short
section of Route 6 will be constructed as advance works as an underground supporting structure (SUS)
in Shing Cheong Road in the former South Apron of the old Kai Tak airport, this short section as part of
the Route 6 is known as Trunk Road T2. The proposed Trunk Road T2 alignment is presented in Figure 1.
The Conforming scheme of the Supporting Underground Structure (SUS) is approximately 409m in
length, 30m in width and 13m in height. It is formed as a twin cell box structure founded on external
diaphragm walls and central barrettes bearing on rock. The diaphragm walls also form the permanent
external walls of the box structure and the barrettes also form, in part, the central dividing wall and roof
support. The thickness of diaphragm walls varies from 1.2m to 1.5m. The thickness of cast in situ base

299
The HKIE Geotechnical Division Annual Seminar 2017

slabs, roof slabs, and central walls were up to 1.5m. The proposed tunnel box was found on diaphragm
walls and H piles beneath the center wall. A typical section of the proposed SUS structure is presented
in Figure 2.
The excavation and lateral support system for the proposed T2 tunnel box includes propped
permanent diaphragm walls with water cut-off. The Public Works Central Laboratory (Lab) is located
adjacent to the proposed East end of the tunnel. To confine the settlement limit to 6mm, ground treatment
is proposed. The location of the proposed T2 tunnel to the Lab can also be found in Figure 1.

Figure 1 – Proposed Trunk Road T2 Tunnel Alignment

Figure 2 – Typical Cross Sections of the Proposed T2 Tunnel Box

300
The HKIE Geotechnical Division Annual Seminar 2017

This paper is focused on design development of the temporary lateral support system and ground
treatment at the east end of the proposed T2 tunnel alignment, in order to control the maximum design
settlement at within the Lab within the settlement limit.

3 Site Description and Geology

3.1 Site location and topography

The site was formed by reclamation in the 1970s for the South Apron of the former Kai Tak Airport. It is
generally flat with ground level at approximately +5.0mPD. It is bounded to the north and east by Kwun
Tong Bypass; south by Cheung Yip Street and Public Works Central Laboratory Building; and west by Kai
Tak approach channel and former Kai Tak runway.

3.2 Ground condition

The site is underlain by Fill, marine deposits, alluvium, and localized colluvium overlying completely
weathered granite. Fill is encountered in every borehole with thickness varying from 6m to 25m. It is
generally composed of medium dense to dense, occasionally loose, silty gravelly SAND with occasional
cobbles and small boulders. Marine clay deposits underlying fill are encountered in some boreholes
with approximately 1m to 10m in thickness. Lenses of marine sand are observed in few boreholes. The
stratum beneath is alluvium, which is encountered in all boreholes and blankets the underlying in-
situ material. Its thickness varies between 5m and 27m. The alluvial deposits on site consist mainly
of medium dense to dense, clayey silty medium to coarse SAND. In some locations, it has lenses or/
and is interbedded with alluvial clay and occasionally alluvial gravel/cobble of various thickness.
Colluvium, underlying alluvium, is observed in 2 boreholes at the SE end of the east bound of tunnel
alignment. Beneath is a stratum of completely decomposed granite (CDG) and highly decomposed
granite (HDG) with moderately strong to strong corestones. The CDG/HDG layer is approximately 5m to
35m thick. Bedrock is typically medium-grained granite with undulating rock head level in the range of
approximately -40mPD to -60mPD, which is 45m to 65m below ground level.
Groundwater levels were measured in piezometers and standpipes that were installed in boreholes
during previous GI at or near the site. It is noted that the available monitoring data is limited for
analysing seasonal or anomalous variations at the site. According to the limited monitoring records, the
groundwater level surrounding the site varies from 2.2m to 4.5m below existing ground level.
Based on the existing borehole information and the interpreted geology, the adopted geological
profile at the east end of the T2 tunnel alignment (focus area for this study) is presented in Figure 3.

Figure 3 – Interpreted Typical Geological Profile at East End of the Proposed Trunk Road T2 Tunnel Alignment

301
The HKIE Geotechnical Division Annual Seminar 2017

4 Adopted Design Parameters

4.1 Adopted Geotechnical Parameters

Based on the borehole information in the vicinity along the T2 tunnel alignment, the adopted design
parameters are presented in Table 1. The selected values of E’ are determined based on the derived
correlations of SPT N values with depths. The values are also considered as representative values
typically used in Hong Kong. It should be noted that a constant SPT N value has been adopted for
increasing depth of CDG although the SPT N values for CDG generally increase with depths based on the
available data.
Due to lack of sufficient monitoring data, the design groundwater level is taken to coincide with the
existing ground level for the worst case condition consideration.

Table 1 – Adopted Geotechnical Design Parameters for Each Strata Layer

Unit Correlated Young’s Undrained Effective Shear Strength


Permeability Poisson’s
Material Weight SPT Result Modulus E’ Strength Cu Cohesion Shear Angle K (m/s) Ratio (ν’)
(kN/m³) (N values) (kN/m²) (kN/m²) C’ (kPa) ϕ’ (deg)

Fill 18.5 10 15,000 – 0 33 2E-05 0.3

Marine
18.0 – – 2.5x D* – – 1E-06 0.3
Deposits

Alluvium 18.5 25 37,500 – 0 35 5E-06 0.3

N=2.7x D*–33
2x(2.7 x D* –33)
(D* <50m)
1E-06 to
CDG 19.0 – 5 36 0.3
1E-07
N = 100
200,000
(D*>50m)

Note: * D – depth below existing ground level (m).

4.2 Adopted Structural material properties

4.2.1 Concrete

Grade 45D/20 concrete will be used for the tunnel external walls, centre wall, base slab and top slab
structures of the proposed tunnel box. This concrete grade is used to meet the 120 years design life and
durability requirement. The properties of concrete are listed below.

• Concrete Grade (fcu) 45 MPa


• Young’s Modulus, E short term 26.4 kN/mm2
• Young’s Modulus, E long term 13.2 kN/mm²
• Poisson’s Ratio 0.2
• Unit Weight of Structural Concrete 24.5 kN/m³

4.2.1 Structural Steel

Structural steel properties for piles, props, walers and king posts are listed below.

• Steel Grade (fy) 430MPa


• Young’s Modulus 205 KN/mm2
• Poisson’s Ratio 0.3

302
The HKIE Geotechnical Division Annual Seminar 2017

5 Excavation and Lateral Support (ELS) System Design Development

Based on the proposed tunnel design and the existing ground condition, the proposed tunnel box at the
East end of the T2 tunnel alignment is founded approximately 30m below the existing ground level. The
majority of the tunnel box is located within the Alluvium layer, above the CDG layer.
Three rounds of the design assessment were undertaken for getting the settlement within the
required limit at the foot print of the Lab. The logistic of the developing the design assessment is
presented in Figure 4.

Figure 4 – Overall ELS Design Development for Settlement Control

5.1 1st Round Design Assessment

First round assessment was focused on the natural ground’s response based on the proposed ELS design
scheme, using PLAXIS 2D software to assess sections along and across the proposed T2 tunnel alignment
at the East end. The sections indication is shown in Figure 5.
The proposed ELS system for this round of the design assessment comprised 1.5m thick diaphragm
walls, with eight layers of corner struts. The sizes and levels of the proposed H section props are can be
found in Table 2.

Figure 5 – Section Indication for 1st and 2nd round Design Assessment

303
The HKIE Geotechnical Division Annual Seminar 2017

Table 2 – 1st Round Design Assessment – Details of Props

Level Pre-stress load Twin Section Propping


Strut ID Size
(mPD) (KN/m) (Y/N)

S1 3.4 356x406x551 0 N

S2 -0.6 914x305x253 300 Y

S3 -4.6 914x305x289 500 Y

S4 -8.6 914x305x289 600 Y

S5 -12.6 914x419x388 800 Y

S6 -16.6 914x419x343 800 Y

S7 -20.6 914x419x343 1000 Y

S8 -23.6 914x419x343 1000 Y

The material properties for each soil layers used in the PLAXIS 2D analysis for are presented in
Table 3. The Mohr-Coulomb Soil Model is used for all strata layers.

Table 3 – PLAXIS 2D Soil Model and Drainage Condition

Soil Layer FEM Material Model Drainage Type

Fill Mohr-Coulomb Drained

Marine Deposits Mohr-Coulomb Undrained Type B

Alluvium Mohr-Coulomb Drained

CDG Mohr-Coulomb Drained

The proposed construction sequence at this stage of the design is comprised 30 stages to model each
stage of excavation, strut installation, tunnel construction and back filling / reinstatement.
Based on the PLAXIS 2D analysis result, settlement increases as construction stages increases.
The maximum ground settlement occurred at the final stage (back filling to the road level), and the
magnitude of the maximum settlement outside the tunnel box footprint is approximately 75mm. The
ground movement results from the PLAXIS 2D analysis far exceeded the design limit of 6mm. The
maximum settlement resulted from Section G is presented in Figure 6.

Figure 6 – Settlement result from 1st round of the design assessment

304
The HKIE Geotechnical Division Annual Seminar 2017

5.2 2nd Round Design Assessment

Based on the result from the 1st round design assessment, and four ground treatment options with
an additional layer of strut at -13.6mPD was proposed and tested by PLAXIS 2D analysis at the same
sections showing in Figure 5. The indications of each ground treatment options are shown in Figure
7 to Figure 10. The Results of maximum ground settlement 10m to 20m outside of the excavation are
summarized in Table 4.
The adopted stiffness parameters for the jet grouping were E = 200MPa and Su = 2000kPa. The
proposed ground treatment options are summarized below:

Option 1 – Jet grout the entire MD layer inside excavation with 3m wide jet grouting strips, 5m apart
center to center (2m apart edge to edge); leave the struts in when back filling to the top;

Option 2 – Jet grouting the entire MD layer inside excavation with 3m wide jet grouting strips, 5m
apart center to center (2m apart edge to edge); 3 rows of 1.2m jet grout at MD layer outside excavation
(jet grouting properties same as inside excavation, 3m wide, 2m edge to edge distance);

Option 3 – Jet grouting the entire MD layer inside excavation with 3m wide jet grouting strips, 5m
apart center to center (2m apart edge to edge); 5m wide jet grout through MD layer from -5mPD to
-20mPD (outside excavation, jet grouting properties same as inside excavation, 3m wide, 2m edge to
edge distance). An additional layer of struts was required at -13.6mRL; and

Option 4 – Jet grout the entire MD layer inside excavation with 3m wide jet grouting strips, 5m
apart center to center (2m apart edge to edge); 10m wide jet grout through MD layer from -5mPD to
-20mPD (outside excavation, jet grouting properties same as inside excavation, 3m wide, 2m edge to
edge distance)

An additional layer of struts was required at -13.6mRL.

Table 4 – 2nd Round Design Assessment Settlement Results

Maximum Settlement (10m to 20m outside of


Ground Treatment Options
the excavation) at Back Filling Stage (mm)

Option 1 49.2

Option 2 45.0

Option 3 39.0

Option 4 34.0

Based on the settlement results from this round of PLAXIS 2D analysis, the settlement was controlled
significantly by adopting ground treatment and an additional layer of strut. Understanding the ground
should have greater stiffness due to the corner effect at the proposed design target location. However,
due to the limit of 2D analysis, the amount of ground moment due to excavation at the corner was still
unknown at this stage of the project. Furthermore, adopting jet grouting ground treatment was an
expensive construction activity, whether it was necessary to grout this much (as in option 3 and option
4) to control the ground movement was unknown at this stage.

5.3 3rd Round Design Assessment-3D FEM analysis

In order to access the settlement at outside at the corner of the T2 trunk road excavation, incorporate
with the proposed ground treatment options, 3D FEM analysis was undertaken by using PLAXIS 3D,
2013 version.

305
The HKIE Geotechnical Division Annual Seminar 2017

Figure 7 – Ground Treatment Option 1 Figure 8 – Ground Treatment Option 2

Figure 9 – Ground Treatment Option 3 Figure 10 – Ground Treatment Option 4

The 3D model was developed in a 150m long x 90m wide x 100m deep rectangular block, representing
a 90m long x 28m wide x 32m deep tunnel excavation section, supported by 9 layers of struts (including
corner struts) for the temporary stage. Tunnel permanent stage (i.e., Tunnel box and back filling to
the road level) is also simulated in the 3D model. The 3D model comprised approximately 96,000 soil
elements. A general view of the 3D model is presented in Figure 11 and Figure 12.

Figure 11 – General View of T2 Tunnel 3D Model Figure 12 – General View of T2 Tunnel 3D Model
Structural Elements

306
The HKIE Geotechnical Division Annual Seminar 2017

The soil parameters, material properties and drainage types in the 3D model remain the same from
2D models. In PLAXIS 3D model, the diaphragm wall and the tunnel boxes (base slab, roof slab and the
middle wall) were modeled as “plates”; struts (including corner struts, and minor struts) were modeled
as “anchors”; the walings were modeled as “beams”; tension piles were modeled as “embedded piles”.
Groundwater and groundwater boundary condition in 3D model remain the same from 2D models.
Construction sequence in 3D model kept the same as the 2D model.
Ground treatment in the 3D model only considered grouting inside of the excavation from -6.0mPD
to -15.0mPD (the marine clay layer), similar to Option 1 from 2nd round design assessment.
The result of the PLAXIS 3D analysis showed the maximum ground settlement at the Lab footprint
was within the design limit of 6mm. From the 3D analysis, a trend of the corner effect was indicated from
the ground movement plot. The settlement results from PLAXIS 3D analysis are presented in Figure 13
and Figure 14. An indication of the trend of corner effect is presented in Figure 15.

Figure 13 – PLAXIS 3D Result – Settlement due to T2 Figure 14 – PLAXIS 3D Result – Settlement due to
Tunnel Excavation (Plan View) T2 Tunnel Excavation (3D View)

Figure 15 – PLAXIS 3D Result Indication of Corner Effect


in terms of Ground Movement due to Excavation

6 Conclusion

Numerical analyses of deep excavations are commonly simplified by assuming plane strain conditions
when designing support systems and making estimates of associated ground movements. However, in
reality excavations behave in a manner more accurately described as three-dimensional, as a result of
the differences in boundary conditions due to the stiffening effects at corners. This paper was using a
design development as an example by comparing three rounds of design assessment; ground settlement

307
The HKIE Geotechnical Division Annual Seminar 2017

was reduced by ground treatment and corner stiffening effects. It is appears engineers can adopt a more
economical solution by reviewing the design development from plane strain FEM analysis to 3D analysis
as it accounts for the corner effect benefits.

References

PLAXIS 2D 2015, Reference Manual.


PLAXIS 3D 2013, Reference Manual.
Bin-Chen Benson Hsiung, Kuo-Hsin Yang, Wahyuning Aila, Ching Hung, 2016. Three-dimensional effect of a deep
excavation on wall deflection in loose to medium dens sand, Computers and Geotechnics.

308

Potrebbero piacerti anche