GeorgeGazetas
YozoGoto
TakashiTazoh
Hellenic Society for Earthquake Engineering & Laboratory of Soil Mechanics NTUA
Earthquake Engineering Committee of the J apanese Society of Civil Engineers
Proceedingsofthe
3
rd
GreeceJapanWorkshop
Santorini2223September2009
Seismic
Design, Observation, Retrofit
of Foundations
Special Theme :
Seismic Protection of Cultural Heritage
3
rd
GreeceJ apan Workshop : Seismic Design, Observation, and Retrofit of
Foundations.
Special Theme : Seismic Protection of Cultural Heritage
Proceedings
Organized by:
the Hellenic Society for Earthquake Engineering and
the Laboratory of Soil Mechanics of the National Technical University of Athens
with the
Earthquake Engineering Committee of the J apanese Society of Civil Engineering
Sponsored by :
ATTIKO METRO, NEOTEK
2009 Laboratory of Soil Mechanics, National Technical University of Athens
All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means
electronic or mechanical, including photocopying, recording or any information storages and retrieval
system now known or to be invented, without written permission from the publisher.
Scientific & Organizing Committee
CoChairmen :
G. Gazetas (NTUA, Greece)
Y. Goto (ERI, J apan)
T. Tazoh (Shimizu Co., J apan)
Members :
N. Yoshida
M. Saitoh
K. Ohtomo
F. Miura
M. Sato
H. Kimata
K. Pitilakis
G. Boukovalas
P. Dakoulas
N. Gerolymos
I. Anastasopoulos
Local Organizing Team
M. Georgiopoulou
Th. Zafeirakos
D. Fyselia
Preface
The 3
rd
GreeceJapan Workshop on Seismic Design of Foundations follows the very successful
previousWorkshops(WS):inAthens(October2005)andTokyo(April2007).Acommonfeature
of these WS was the active participation of leading researcherspractitioners (mostly from
Japan) and academics (mostly from Greece). The topics of these WS covered a wide range of
foundation problems, as well as a few topics of more general interest in earthquake
geotechnics.TheGreekauthorshadmostlyemphasizednumericalandanalyticalsolutions;the
Japanese researchers utilised largescale experimental testing along with numerical
simulations, mainly in support of innovative design concepts. As a result, valuable cross
fertilisation of methods, topics, and ideas has been developing between our two engineering
communities.
WhilemanyofthecharacteristicsofthefirsttwoWShavebeenpreserved,this3
rd
WSinitiates
adoubledeparture:
Thetopicsbroadentoencompassearthquakeandgeotechnicalproblemsnotstrictlyrelated
to foundations, including seismic retrofitting and analysis of bridges, as well analysis of
dams,slopes,retainingsystems.
A Special Theme Session on the Seismic protection of cultural heritage is established as a
significantpartoftheWS.Thetopicsofthissessionrangefromseismicanalysisofancient
monuments,protectionofartifacts,restorationofearthquakedamagedhistoricstructures,
to policy aspects of restoration and prioritisation. Seven of the related lectures have been
submittedinwritingtoappearintheseProceedings.
As always, in addition to Japanese and Greek authors, there are articles and/or oral
presentationsfromleadingengineersfromFrance,USA,U.K.,India,China,Algeria,Austria.
WehopethatthereadersoftheProceedingswillbesatisfiedwith(ordespiteof)suchabroad
varietyofsubjects.
G.Gazetas
1
,Y.Goto
2
,T.Tazoh
3
September2009
1
NationalTechnicalUniversityofAthens(NTUA),Greece
2
EarthquakeResearchInstitute,TheUniversityofTokyo,Japan
3
InstituteofTechnology,ShimizuCorporation,Japan
Proceedings:ListofPapers
1.PileFoundations,Liquefaction
Astudyonthebehaviorofrakedpilesinseismicallyliquefiablesoils
S.Bhattacharya,T.Tazoh,J.Jang,M.Sato........................................................................................1
KinematicResponseofBatterPileFoundation:CentrifugeTests
T.Tazoh,M.Sato,J.Jang,Y.Taji,G.Gazetas.....................................................................................20
KinematicandInertialBehaviorofBatterPiles
A.Giannakou,N.Gerolymos,G.Gazetas...........................................................................................36
SeismicResponseofBridgePileColumns
V.A.Drosos,N.Gerolymos,G.Gazetas........................................................................................51
EfficiencyofEqualEnergyAssumptionforEvaluatingDuctilityFactorsofaPile
S.Mori..............................................................................................................................................71
SinkageofaPileFoundationduringtheNiigatakenChuetsuokiEarthquakein2007
Y.Goto.................................................................................................................................................82
LargescaleShakeTableTestsonLateralSpreadingofSheetpileQuayWallandPileFoundation
M.Sato,K.Tabata...............................................................................................................................89
EDefenseShakingTableTestonLiquefactionInducedLateralSpreadingofLargeScaleModelGround
withQuayWallandPileSupportedStructure
K.Tabata,M.Sato...............................................................................................................................96
Designchartsforsinglepilesunderlateralspreadingofliquefiedsoil
A.Valsamis,G.Bouckovalas,E.Drakopoulos....................................................................................104
PilesinLiquefactionInducedSoilFlowbehindQuayWall:ASimplePhysicalMethodVersusCentrifuge
Experiments
P.Tasiopoulou,N.Gerolymos,T.Tazoh,G.Gazetas.........................................................................114
FiniteElementAnalysisofPileSoilInteractionSystembyOverlayingMeshMethod
A.Ohta,F.Miura,Y.Ono,J.Kiyono....................................................................................................128
UniquedescriptionofliquefactionbehaviorofToyourasandswithdifferentdensities
F.Zhang,Y.Jin.....................................................................................................................................140
2.EarthquakeObservations,GroundMotions
ImpactofModerateEarthquakesinPostBhujEra:CaseStudyofSikkim2006andDurgapur2008
Earthquakes,India
S.C.Dutta,P.Mukhopadhyay,S.Bhattacharya.........................................................................154
StatisticalEvaluationofEmbedmentEffectonDamagetoRCBuildingStructuresduringthe1995
HyogokenNanbuEarthquake
A.Mikami,Y.Nariyuki,T.Matsuda...............................................................................................171
GlobalIncreaseofNaturalDisastersandInternationalCooperationforDisasterMitigation
MasanoriHamada,XuWu..............................................................................................................177
TsunamiDamageStudiesandConstructionoftheMemorialPolesinBandaAceh
H.Iemura,M.H.Pradono,M.Sugimoto.......................................................................................186
TheAlgerianExperienceofCoveringEarthquakeDamagesandAppliedTechniquesof
Reinforcement
A.Moulay..........................................................................................................................................193
NumericalAnalysisofNearFieldAsymmetricVerticalMotion
T.Tobita,S.Iai,T.Iwata..................................................................................................................206
EffectofPreYieldingElasticityonSlidingTriggeredbyNearFaultMotionsModeledasIdealized
Wavelets
E.Garini,G.Gazetas,N.Gerolymos..............................................................................................219
TheIrrationalityofCurrentSeismicCodeSpectraforSoftSoils:ProposedRemedy
A.Ziotopoulou,G.Gazetas.............................................................................................................229
3.SoilStructureInteraction
CalculationofseismicresponseofbuildingbasedonpushoveranalysisofSSImodel
M.Iiba,Y.Umemura,O.Kurimoto,T.Akita,M.Teshigawara,K.Watanabe..238
AftershockEffectsonDamageEvaluationsforSoilStructureInteractionSystem
K.Kawano,Y.Kimura.258
SoilStructureInteractionProblemofaSchoolBuildingBasedonEarthquakeRecordsand
DynamicSubstructureMethod
M.Nakamura,Y.Kitamura,J.Suzumura,K.Hanada...........................267
EffectofFoundationSoilInterventionstotheSeismicResponseofMdofStructures
K.Pitilakis,E.Kirtas,E.Rovithis.....276
Shakingtabletestonseismicbehaviorofgravestomesimilartostonepillarswithandwithout
reinforcement
S.Miwa,A.Furukawa,J.Kiyono...............................287
MechanicalRepresentationofDynamicStiffnessofSoilFoundationSystems
M.Saitoh.....................................................................................................296
NonlinearDynamicAnalysisofPartiallySupportedBeamColumnsonNonlinearElastic
FoundationIncludingShearDeformationEffect
E.J.Sapountzakis,A.E.Kampitsis............................302
Seismicriskoftheundergroundstructureconsideringtheuncertaintyofseismicaction
S.Nakamura,S.Nishiyama,T.Matsumoto,Y.Miyagawa..........................................330
Topographicirregularitiesandsoilfoundationstructureinteraction
D.Pitilakis.............................................335
Numericalmodellingofshallowfoundationundercyclicoverturningmomentandvalidation
throughcentrifugeexperiments
M.Apostolou.............................................344
4.Bridges:Analysis,Design,andRetrofit
RetrofittechnologiesoflongspanbridgesinHanshinExpressway
Y.Adachi,H.Kanaji,T.Nishioka......................................................................................355
PileFoundationsinImprovedSoilforHighwayBridgefacingLiquefactionInducedSoilFlow
N.Gerolymos,G.Gazetas.................................................................................................362
MethodologyforDesignAnalysisofBridgesagainstanEmergingFaultRupture
I.Anastasopoulos,R.Kourkoulis,V.Drosos,T.Georgarakos,G.Gazetas..375
TowardsaReversalofSeismicCapacityDesign.PartA:AnalysisofBridgePierFoundation
System
I.Anastasopoulos,M.Loli,N.Gerolymos,M.Apostolou,G.Gazetas...............................393
TowardsaReversalofSeismicCapacityDesign:PartB.ShakingTableTestingofBridgePier
FoundationSystem
I.Anastasopoulos,T.Georgarakos,V.Drosos,S.Giannakos,G.Gazetas...407
SimplifiedMethodfortheDesignofRaftFoundationsagainstaDirectHitbyThrustFaulting
I.Anastasopoulos....419
PreliminarySFSIStudiesfortheMessinaBridgeFoundations
E.Stavropoulou,I.Anastasopoulos,G.Gazetas.....438
5.Dams,Slopes,RetainingWalls
FieldObservationsofWenchuanEarthquakeofMay122008
H.Wang..449
LateralandLongitudinalSeismicVibrationsofConcreteFaceRockfillDams
P.Dakoulas....470
DynamicCrackPropagationAnalysisofConcreteGravityDamswithJointedRockFoundation
H.Kimata,Y.Fujita,H.Horii,M.Yazdani.................................486
EngineeringaspectsofsiteandtopographyeffectsatAegion,Greece
O.J.Ktenidou,D.Raptakis,K.Pitilakis,F.J.ChvezGarca.....495
PilesforStabilisingSeismicallyPrecariousSlopes.PartA:DevelopmentandValidation
R.Kourkoulis,F.Gelagoti,I.Anastasopoulos,G.Gazetas......506
PilesforStabilisingSeismicallyPrecariousSlopes.PartB:ParametricAnalysisandDesignCharts
R.Kourkoulis,F.Gelagoti,I.Anastasopoulos,G.Gazetas......520
Onseismicresponseofretainingstructures
N.Sitar,L.AlAtik.........................................................................534
TheNikawa(1995)andHigashiTakezawa(2004):ModellingandNumericalAnalysis
N.Gerolymos......545
EffectofhystereticdampingandstiffnessatunloadingonresponseofgroundduringEarthquake
N.Yoshida......573
6.Monuments
Structuralrestorationprocesstoprotectaworldheritagemonumentafteranearthquake
disaster:thecaseoftheKatholikonofDafniMonasteryinAtticaGreece
A.MiltiadouFezans...584
OntheDynamicBehaviorofaLightweightIsolatorforMuseumArtifacts
V.Koumousis......605
LessonslearntfromtheImpactofSikkim2006EarthquakeonHeritageStructures
P.Mukhopadhyay,S.C.Dutta,S.Bhattacharya.....615
SeismicResponseAnalysisofUndergroundMonumentalStructures.TheCatacombsofKom
ElShoqafa,Alexandria,Egypt.
K.Pitilakis,S.Hemeda.....629
Estimationoftheseismicresponseofhistoricalstructuresandevaluationofinterventions
I.N.Psycharis,H.P.Mouzakis,A.Miltiadou,P.Pavlopoulou,I.M.Taflampas644
InPlaneResponseofoldMasonryunderSeismicLoads
T.Zimmermann,A.Strauss....658
ThereconstructionofthenortheastcornerofthetempleofZeusatNemea,Greece
N.Makris..665
DynamicAnalysisofRuptureProcessofMasonryBuildingsusingNewSimulationMethod
FurukawaA.,KiyonoJ.,TokiK....678
1.
PileFoundations,Liquefaction
+ =
(
+ = u u u u
2 2
2
3
2
3
2
sin cos
12
12
4 sin
12
cos 4
I
AL
L
EI
L
EI
L
EA
K
raked
(2) where
EA = Axial stiffness of the pile where A = Cross sectional area of the pile
u = Angle made by the pile with the horizontal as shown in the Figure 3
If it is noted that the second moment of area I is given by:
2
min
Ar I = (3) where
r
min
= Minimum radius of gyration
Combining equations 2 and 3 we get:
(
+


.

\

=
(
+ = 1 cos 1
12
12
4 sin cos
12
12
4
2
2
3
2 2
2
min
2
3
u
u u
L
EI
Ar
AL
L
EI
K
raked
(4) where
min
r
L
= known as Slenderness ratio of the pile in the likely unsupported zone.
Bhattacharya et al (2004) analyzed 14 case studies of pile performance in seismic zones. The
analysis showed that piles that survived the earthquake has slenderness ratio less than 50. Com
paring equation 1 and 4 we define an improvement factor Z
i
as:
1 cos 1
12
2
2
+


.

\

= = u
K
K
Z
raked
i
(5)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
4
Equation 5 suggests that raked piles are always stiffer that vertical piles. Equation 5 can be re
casted using angle of batter ( ) as Equation 6.
1 sin 1
12
2
2
+


.

\

= = 
K
K
Z
raked
i
(6)
Equation 6 is plotted in Figure 3. The factor Z
i
represents the ratio of the stiffness of a raked
pile with respect to the vertical piles and can be coined as Improvement Factor. In the Figure
Z
i
is plotted against the angle of batter for various slenderness ra io ( t ).
0 2 4 6 8 10 12 14 16 18 20
1
10
20
30
40
50
60
70
80
90
100
Angle of Batter ()
I
m
p
r
o
v
e
m
e
n
t
F
a
c
t
o
r
(
Z
i
)
=40
=60
=80
=100
Figure 3: Normalized stiffness ratio of a raked pile plotted against the angle of batter
The next section describes a series of centrifuge tests that were carried out at Institute of Tech
nology (Shimizu Corporation) to investigate various aspects of raked piles and vertical piles.
4 CENTRIFUGE TESTS
4.1 Description of the test
A series of centrifuge tests has been carried out in Institute of Technology (Shimizu Corpora
tion) to study the various aspects of raked piles. Two types of pile group were tested as shown
in Figure 4. Table 2 summarizes the different centrifuge tests carried out. However, this paper
concentrates on one particular test (Case 32). In some of the tests, a small scaled superstructure
was used: see Figure 5 for the photograph of the superstructure and Figures 6 and 7 for sche
matic diagram of the pilesupported structures. The input motion applied at the base of the
model is given by Figure 8 and the FFT of the signal is given by Figure 9.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
5
Table 2: Schedule of the tests
ID of the test Details of the test Remarks
SERIES 1
Case 11
Two vertical pile groups near a quay wall were tested.
In one case the quay wall collapsed and in the other the
wall did not collapse. The piled foundations are placed
200mm from quay wall.
SERIES 1
Case 12
Two vertical pile groups near a quay wall were tested.
In one case the quay wall collapsed and in the other the
wall did not collapse. The piled foundations are placed
100mm from quay wall.
SERIES1
Case 13
Two vertical pile groups near a quay wall were tested.
In one case the quay wall collapsed and in the other the
wall did not collapse. The piled foundations are placed
50mm from quay wall.
In the three tests of SERIES 1, the
main aim was to study the effect of
quay wall collapse on the piled
foundations. The distance between
the quay wall and the piled founda
tions were varied. All the piled
foundations had a superstructure.
SERIES2
Case 21
Two vertical pile groups were tested. In one case, there
was a superstructure and in the other there is no super
structure.
SERIES2
Case 21
Two vertical pile groups were tested. In one case, the
foundation was 50mm from the quay wall and in the
other the foundation was 100mm from quay wall.
In SERIES 2 tests, the effect of in
ertia on the superstructure was in
vestigated
SERIES3
Case 31
Two pile groups were tested. One had a vertical pile
and the other had a raked pile. The angle of batter is
. The distance from the quay wall is 200mm.
10
SERIES3
Case 32
Two pile groups were tested. One had a vertical pile
and the other had a raked pile. The angle of batter is
. The distance from the quay wall is 100mm.
10
SERIES3
Case 33
Two pile groups were tested. One had a vertical pile
and the other had a raked pile. The angle of batter is
. The distance from the quay wall is 50mm.
10
SERIES3
Case 34
Two pile groups were tested. One had a vertical pile
and the other had a raked pile. The angle of batter is
. The distance from the quay wall is 100mm.
5
In SERIES 3 tests, comparison is
made between vertical piles and
raked piled.
Figure 4: Photograph of the vertical pile and raked piles used in the centrifuge tests
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
6
Figure 5: Centrifuge setup of the raked piles
Liquefiable
soil
ACCFooting
ACCBase AG(0)
ACCStructure
APP2
APP1
100mm
130mm
30mm
588.75gm
392.5gm
80mm
60mm
Quay wall
Columns consists of 4
nos of 6mmx2mmx60
mm steel section
A4
A2
A1
A5
A6
A7
A9
A10
Figure 6: Schematic diagram of the vertical pile showing the instrumentation. [{Ai denotes Strain
Gauges where i varies from 1 to 10 except 3 and 8}, {ACC is accelerometer}, {PP denotes Pore Pres
sure}].
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
7
Liquefiable
soil
ACCFooting
ACCBase AG(0)
ACCStructure
BPP2
APP1
100mm
130mm
30mm
588.75gm
392.5gm
80mm
60mm
Quay wall
Columns consists of 4
nos of 6mmx2mmx60
mm steel section
B4
B6
B7
B9
B10
Figure 7: Schematic diagram of the vertical pile showing the instrumentation [{Ai denotes Strain Gauges
where i varies from 1 to 10 except 3 and 8}, {ACC is accelerometer}, {PP denotes Pore Pressure}].
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
10
8
6
4
2
0
2
4
6
8
10
Time(s)
B
e
d
r
o
c
k
a
c
c
e
l
e
r
a
t
i
o
n
(
g
)
AG(0)
Figure 8: Input motion used in Test Case 32
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
8
0 30 60 90 120 150 180 210 240
20
0
20
40
60
80
100
Frequency (Hz)
L
o
g
a
m
p
l
i
t
u
d
e
(
d
B
)
o
f
i
n
p
u
t
Figure 9: Input motion used in the tests
4.2 Measured responses from the tests
Figure 10 shows the time history of the measured responses for the vertical pile. The input ac
celeration, the pore pressure responses and the acceleration of the footing and structure are plot
ted. The responses indicate that as the soil liquefies i.e. at about 0.15sec and after the responses
of the footing and structures alter. The instrumentation layout can be found in Figure 6. Figure
11 shows the similar data for the raked piles and the corresponding instrumentation. Figure 12
compares the acceleration of structure for two types of foundations. Figure 13 shows the corre
sponding acceleration for footing for two types of foundations.
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
10
0
10
A
c
c
e
l
e
r
a
t
i
o
n
(
g
)
Vertical pileStructureAS1
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
10
0
10
A
c
c
e
l
e
r
a
t
i
o
n
(
g
)
Vertical pileFootingAS1
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
0
20
40
45
P
o
r
e
p
r
e
s
s
u
r
e
(
k
P
a
)
APP1
APP2
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
10
0
10
Time(s)
B
a
s
e
a
c
c
e
l
e
r
a
t
i
o
n
(
g
)
AG(0)
Figure 10: Time history of results for vertical piles
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
9
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
20
0
20
A
c
c
e
l
e
r
a
t
i
o
n
(
g
)
Raked pileStructureAS1
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
20
0
20
A
c
c
e
l
e
r
a
t
i
o
n
(
g
)
Raked pileFootingAS1
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
0
20
40
P
o
r
e
p
r
e
s
s
u
r
e
(
k
P
a
)
BPP1
BPP2
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
20
0
20
Time(s)
B
a
s
e
a
c
c
e
l
e
r
a
t
i
o
n
(
g
)
AG(0)
Figure 11: Time history of results for raked piles
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55
30
20
10
0
10
20
30
Time (s)
A
c
c
e
l
e
r
a
t
i
o
n
(
g
)
Input
Structurevertical piles
Structureraked piles
Figure 12: Time history of acceleration for structures for two types of foundations
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
10
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55
10
8
6
4
2
0
2
4
6
8
10
Time (s)
A
c
c
e
l
e
r
a
t
i
o
n
(
g
)
Input
Footingvertical piles
Footingraked piles
Figure 13: Time history of acceleration of footing for two types of foundations
4.3 Frequency response analysis of the foundations
Before we analyse the experimental data, it is necessary to estimate various parameters. Table 3
characteristics of the pile.
Table 3: Characteristics of the pile
Material Steel
Pile outside diameter (d
o
) 10mm
Wall thickness (t) 0.2mm
Second Moment of Area (I) 73.95mm
4
Radius of gyration ( mm
A
I
r 46 . 3
min
= = )
3.46mm
Pile length (L) 270mm
Youngs Modulus (E) 210GPa
EI (Bending rigidity of the pile) 15.5310
6
Nmm
2
4.3.1 Stiffness of vertical piles
Based on Figure 6, the boundary condition of the vertical pile is assumed as free to translate at
the pile head and free to rotate. The length of unsupported length is given by the summation of
top nonliquefiable zone (as the zone moves and 60mm), the liquefiable zone (130mm) and
50mm (depth of fixity). Therefore L is about 240mm. The lateral stiffness of a group of 4 piles
is given by equation 1 is given by equation 7.
mm
N
mm
Nmm
L
EI
K 92 . 53
240
10 53 . 15 48 12
4
3 3
2 6
3
=
= = (7)
The section of the column is (2mm x 6mm) and 60mm high. This can be considered to be
very rigid compared to the pile. As a first approximation, the time period of the structure can be
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
11
computed by considering the stiffness of the pile and the mass of the footing, columns and
structure. The total mass of the structure is given by equation 8.
kg gm gm gm gm M 1 85 . 10003 6 . 22 75 . 588 5 . 392
I
= = + + = (
8)
he frequency is given by: T
Hz
m kg
N
f
n
9 . 36
10 1
92 . 53
2
1
3
=
=
t
(9)
Figures 14 and 15 shows the frequency domain analysis of the acceleration data of the foot
ing and structure. This is essentially, plotting the acceleration time history of the footing data
and structure in the normalized frequency domain. The analysis indicates that there is a peak at
about 36 Hz which indicates that this is natural frequency of the system.
0 10 20 30 40 50 60 70 80 90 100
0
0.5
1
1.5
2
2.5
3
3.5
4
Frequency (Hz)
D
y
n
a
m
i
c
a
m
p
l
i
f
i
c
a
t
i
o
n
f
a
c
t
o
r
(
F
o
o
t
i
n
g
)
Figure 14: Frequency response function of e footing for the vertical pile foundation. th
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
12
0 10 20 30 40 50 60 70 80 90 100
0
0.5
1
1.5
2
2.5
3
3.5
4
Frequency (Hz)
D
y
n
a
m
i
c
a
m
p
l
i
f
i
c
a
t
i
o
n
f
a
c
t
o
r
(
S
t
r
u
c
t
u
r
e
)
Figure 15: Frequency response function of the structure for the vertical pile foundation
4.3.2 Stiffness of raked piles
The angle of batter of the raked pile is . Based on Figure 7, and equation 6 we may obtain
the stiffness of the raked pile. The slenderness ratio of the pile required is given by equation 10
10
36 . 69
46 . 3
240
min
= = =
r
L
(10)
The improvement factor is given by equation 11.
13 1 ) 10 sin 400 ( 1 sin 1
12
2 2
2
~ + = +


.

\

= = 
K
K
Z
raked
i
(11)
The stiffness of the raked pile foundation is given by equation 12.
mm
N
mm
N
K
raked
710 92 . 53 13 = = (12)
As a first approximation, the frequency of the raked pile foundation supported structure is
given by:
Hz
m kg
N
f
n
134
10 1
710
2
1
3
=
=
t
Figures 16 and 17 plots the frequency response function of the acceleration data of the foot
ing and structure. The results indicate that there is a peak at about 82Hz (Figure 17).
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
13
0 20 40 60 80 100 120 140 160 180 200
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Frequency (Hz)
D
y
n
a
m
i
c
a
m
p
l
i
f
i
c
a
t
i
o
n
f
a
c
t
o
r
(
F
o
o
t
i
n
g
)
Raked pile
Figure 16: Dynamic amplification factor measured from the structure for the raked piles (Figure 7).
0 20 40 60 80 100 120 140 160 180 200
0
0.5
1
1.5
Frequency (Hz)
D
y
n
a
m
i
c
a
m
p
l
i
f
i
c
a
t
i
o
n
f
a
c
t
o
r
(
S
t
r
u
c
t
u
r
e
)
Raked pile
Figure 17: Dynamic amplification factor measured from the structure for the raked piles (Figure 7).
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
14
4.4 Bending moment at the front and back side of the vertical pile
Strain measurements were taken at various locations of the vertical pile (see Figure 6 for the in
strumentation layout). Figure 18 shows the bending strain at various locations of the pile near to
the quay wall. Figure 19 shows the bending strain at various locations in the back side. The tests
indicate that there are large moments at the interface possible due to kinematic interactions.
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
2000
1500
1000
500
0
500
1000
Time (s)
B
e
n
d
i
n
g
s
t
r
a
i
n

v
e
r
t
i
c
a
l
p
i
l
e
(
)
A1(1d)
A2(6d)
A4(19d)
A5(27d)
Figure 18: Bending strain along the vertical pile in the front of the quay wall
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
2000
1500
1000
500
0
500
1000
1500
Time (s)
B
e
n
d
i
n
g
s
t
r
a
i
n

v
e
r
t
i
c
a
l
p
i
l
e
(
)
A6(1d)
A7(6d)
A9(19d)
A10(27d)
Figure 19: Bending strain along the vertical pile in the back pile
Figures 20 and 21 shows the bending strain for the raked piles (see Figure 7 for instrumenta
tion layout). Figure 22 compares the bending strain for the raked pile and the vertical pile which
is possible due to inertia. Figure 23 shows the kinematic bending moment at the interface for
vertical pile and raked pile.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
15
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
1500
1000
500
0
500
1000
1500
Time (s)
B
e
n
d
i
n
g
s
t
r
a
i
n

r
a
k
e
d
p
i
l
e
(
)
B1(1d)
B2(6d)
B4(19d)
B5(27d)
Figure 20: Bending strain along the raked pile in the front of the quay wall
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
1600
1200
800
400
0
400
800
1200
1600
Time (s)
B
e
n
d
i
n
g
s
t
r
a
i
n

r
a
k
e
d
p
i
l
e
(
)
B6(1d)
B7(6d)
B9(19d)
B10(27d)
Figure 21: Bending strain along the vertical pile in the back pile
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
16
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
600
400
200
0
200
400
600
800
1000
1200
Time (s)
I
n
e
r
t
i
a
l
b
e
n
d
i
n
g
s
t
r
a
i
n
(
)
Vertical Pile (A1)
Raked Pile (B1)
Figure 22: Comparison with vertical pile and raked pile
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
600
400
200
0
200
400
600
800
1000
1200
Time (s)
K
i
n
e
m
a
t
i
c
b
e
n
d
i
n
g
s
t
r
a
i
n
a
t
t
h
e
i
n
t
e
r
f
a
c
e
(
)
Vertical Pile (A4)
Raked Pile (B4)
Figure 23: Vertical pile and raked pile
4.5 Bending strain in the front and back pile
Figures 24 and 25 shows the bending strain for the front pile and the back pile.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
17
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
500
400
300
200
100
0
100
200
300
400
Time (s)
B
e
n
d
i
n
g
s
t
r
a
i
n

v
e
r
t
i
c
a
l
p
i
l
e
(
)
(Front pile A4)
(Back pile A9)
Figure 24: Strain measurements taken at the interface (A4 and A9) of the raked pile configuration.
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
600
400
200
0
200
400
600
800
1000
1200
Time (s)
B
e
n
d
i
n
g
s
t
r
a
i
n

r
a
k
e
d
p
i
l
e
(
)
(Front pile B4)
(Back pile B9)
Figure 25: Strain measurements taken at the interface (B4 and B9) of the raked pile configuration.
5 CONCLUSIONS
Raked piles are often used in seismic areas to support structures. However, confusion still exists
on their performance i.e. whether or not they are beneficial. This paper presents a study in this
regard. Following conclusions are drawn:
1. Raked piles are always stiffer than the vertical piles. The stiffness increase depends
on the angle of batter and also on the slenderness ratio of the pile. An expression of
this increase i.e. the ratio of lateral stiffness of raked piles to the lateral stiffness of
the vertical pile denoted by Improvement factor is derived. This enhanced stiffness
affects the performance in various ways and is explained in the next two points.
2. As the stiffness of a raked pile system is comparatively high, the frequency of the
structurefoundation system is also high. Therefore, the raked pile foundation is dy
namically sensitive and the performance depends on the type of earthquake.
3. Due to relatively high horizontal stiffness of the raked pile foundation, the horizontal
displacement of the pile cap or the structure is comparatively less making its per
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
18
formance satisfactory from serviceability point of view. This further reduces the P
delta effect of the superstructure.
4. Due to this high stiffness, raked piles attract significant bending moments near the
pile head. As a result, proper connection details are necessary.
5. Raked pile requires careful design consideration and can prove beneficial to most en
gineering structures.
6 REFERENCES
AFPS (1990), Recommendations AFPS 90, 3 Volumes, Association franaise de gnie parasismique,
France
Bhattacharya, S., S.Adhikari and N.A.Alexander (2009): A simplified method for unified buckling and
free vibration analysis of pilesupported structures in liquefiable soils, Journal of Soil Dynamics and
Earthquake Engineering, August 2009 Issue
Bhattacharya, S., Bolton, M.D. and Madabhushi, S.P.G. (2005): A reconsideration of the safety of the
piled bridge foundations in liquefiable soils, Soils and Foundations, Volume 45, August 2005 issue,
No 4, pp 1326
Bhattacharya, S., Madabhushi, S.P.G., and Bolton, M.D. (2004): An alternative mechanism of pile fail
ure in liquefiable deposits during earthquakes, Geotechnique 54, April issue, No.3, pp 203213.
Berril J. B., Christensen S.A., Keenan R. P., Okada W. and Pettinga J. R. (2001), Case study of lateral
spreading forces on a piled foundation, Geotechnique, Vol. 51, No 6, pp. 501517
EERI (1991), Costa Rica Earthquake of April 22, 1991, Reconnaissance Report, Earthquake Spectra,
Vol. 7, Supplement B.
Giannakou, A., Gerolymos,N. and Gazetas,G. (2007): Influence of Batter Piles on the Seismic Response
of Pile Groups, 2
nd
Japan Greece workshop on seismic design, observation and retrofit of foundations,
3
rd
4
th
April 2007, Tokyo.
Gerelymos, N., Giannakou, A., Anastasopoulos, I., and Gazetas, G. (2008): Evidence of beneficial role of
inclined piles: Observation and summary of numerical analysis, Bulletin of Earthquake Engineering,
Vol 6, pp 705722.
Escoffer,S., Chazelas, JL., and Garnier, J. (2008): Centrifuge modeling of raked piles, Bulletin of Earth
quake Engineering, Vol 6, pp 689704.
Kastranta G. & Gazetas G, Tazoh T. (1998), Performance of three quay walls in Maya Wharf: Kobe
1995, Proceedings of the 11th European Conference on Earthquake Engineering
JRA (1996,1980,1972): Japanese Road Association, Specification for Highway Bridges, Part V, Seismic
Design.
Mylonakis G., 2001, Simplified model for seismic pile bending at soil layer interfaces, Soils and
Foundations, 41, 4758.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
19
Kinematic Response of Batter Pile Foundation: Centrifuge
Tests
T. Tazoh
Institute of Technology, Shimizu Corporation, Japan
M. Sato
National Research Institute for Earth Science & Disaster Prevention, Japan
J. Jang & Y. Taji
Institute of Technology, Shimizu Corporation, Japan
G. Gazetas
National Technical University of Athens, Greece
ABSTRACT: We carried out centrifuge tests to clarify the seismic behavior of batterpile
foundations. A verticalpile foundation and a batterpile foundation without the presence of a
superstructure were installed parallel to each other in a soil container filled with dry sand, and
were excited simultaneously. Through a comparison of the acceleration and displacement
response of the footing, as well as the axial and bending strain of the piles for the two pile
foundations, the kinematic response of the seismic behavior of the batterpile foundation was
experimentally investigated.
1 INTRODUCTION
The lateral stiffness of a pile foundation can be increased by adopting batter piles, which is why
they are commonly used in landing piers that are subject to large lateral forces. However, batter
piles are seldom used for buildings or civil engineering structures even in the case of large
lateral forces. The reasons are as follows:
1) When soil settlement occurs, not only the safety of the pile foundation but also that of the
structure as a whole system may be threatened by settlementinduced vertical loads acting
on the batter piles.
2) During an earthquake, the piles in a batterpile foundation may be subject to excessive
axial compression and pullout forces, which are not generated in a verticalpile foundation.
3) The strength of concrete piles is reduced by decreasing the compressive force acting on the
piles due to rocking motions induced by the adopted batter piles.
4) Since infinite lateral ground planes cannot be assumed for batter piles, they cannot be
expected to have the same horizontal subgrade reaction as that of vertical piles.
5) In urban areas, the use of batter piles is constrained by the boundary lines of adjacent land.
The 1995 Great Hanshin Earthquake in Japan has increased the demand for pile foundations
with high seismic performance, as well as lower cost and easier construction. Batter piles can be
used with little additional expense, no special design, and hardly any difficulty in construction.
Therefore, the seismic behavior of batter piles has recently attracted much research interest, as
has research and development related to easy and accurate methods of installing batter piles
(Gerolymos, N., et. al., 2008, Giannakou, A., et. al., 2007, Poulos., N., 2006).
In this study, we carried out centrifuge shaking table tests to clarify the seismic behavior of
batterpile foundations. A verticalpile foundation and a batterpile foundation were installed
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
20
parallel to each other in a soil container filled with dry sand, and were excited simultaneously
(Tazoh, T., et. al., 2005, Tazoh, T., et. al., 2007). As our objective was to investigate the
fundamental characteristics of the seismic behavior of batter piles, none of the pilefoundation
models had a superstructure. This study focused on the kinematic interaction of batter piles
(Fan, K., et. al., 1991, Mylonakis, G., et. al., 1997, Mylonakis, G., 2001, Nikolaou, S., et. al.,
2001, Sica, S., et., al., 2007, Tazoh, T., et. al., 1987). Through a comparison of the acceleration
and displacement response of the footing, as well as the axial and bending strain of the piles for
the two pile foundations, the kinematic nature of the seismic behavior of the batterpile
foundation was experimentally studied.
2 CENTRIFUGE TESTS
The most direct and effective way to quantitatively and qualitatively investigate the seismic
behavior of batter piles is to compare the seismic behavior between a verticalpile foundation
and a batterpile foundation under the same input motions. Each test for each model must be
carried out under nearly identical conditions with respect to input motions, soil conditions, and
soil behavior. Note, however, that it is impossible to achieve complete similarity between
shaking table tests due to the difficulty of reproducing the input motion and nonlinear behavior
of the soil.
Therefore, a verticalpile foundation and a batterpile foundation without the presence of a
superstructure were installed parallel to each other in a soil container, as shown in Figure 1, and
were excited simultaneously.
A laminar box was used as the soil container to allow shear deformation of the soil deposit as
in the free field. Actually, installing two models that behave differently in a laminar box is not
an appropriate testing method because the behavior of the models might influence each other.
However, considering the inconsistency of the input motion and the difficulty of reproducing
the soil conditions and nonlinearity, we believe that this method is more reasonable than
individually testing the verticalpile foundation and batterpile foundation separately.
The interior of the soil container is 805 mm in length, 474 mm in width, and 324 mm in
height. All tests were conducted at centrifugal acceleration of 30 g on a 1/30scale model.
Table 1 shows the scaling ratios of the models.
The verticalpile foundation and the batterpile foundation each have four piles, and the pile
heads and pile tips are rigidly connected to the footing and the base of the soil container,
respectively. The batter piles are identically inclined at a 10 angle. The soil deposit is a
uniform layer consisting of dry silica sand No. 7 (Mean particle diameter D
50
= 0.15 mm; Soil
density
s
= 2.635 g/cm
3
; Maximum dry density
max
= 1.539 g/cm
3
; Minimum dry density
min
= 1.206 g/cm
3
). Thickness and relative density of the soil deposit is 300 mm (prototype:
9 m) and Dr = 60%, respectively.
Figure 2 shows the grain size accumulation curve of silica sand No. 7. Table 2 shows the
materials and size of the experimental model used in the tests and Photograph 1 shows the test
model. Sixtytwo monitoring channels in total were installed, with the sensors comprising
seventeen accelerometers, five noncontact displacement meters, and forty strain gauges
(Table 3). The test was conducted a total of nine times, varying the input motion and maximum
acceleration as shown in Table 4.
While the purpose of this study was to clarify the kinematic interaction of the batter piles,
consideration must also be given to effects from the mass of the footing (made of steel, size:
355 cm). The inertial interaction caused by the inertial force of the footing might be included
in the results, which consequently may not represent the perfect kinematic interaction.
3 KINEMATIC NATURE OF SEISMIC BEHAVIOR OF BATTER PILE
Figure 3 shows the frequency transfer function calculated by the acceleration records between
the soil surface and the input motion of the sweep test. The predominant frequency of the
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
21
ground is 3.5 Hz in the case of maximum acceleration of input motion at 5 Gal. The
predominant frequencies are 3.23.3 Hz and 3.03.1 Hz, and also the peak acceleration
amplification factors decrease corresponding to the increase in maximum acceleration of the
input motion to 15 Gal and 30 Gal.
Figure 4 shows the frequency transfer function between the ground surface and input motion
obtained from El Centro record excitation. The predominant frequency of the ground is 3.4 Hz
in the case of maximum acceleration of input motion at 50 Gal. The predominant frequencies
are 2.82.9 Hz and 2.42.5 Hz, and also the peak acceleration amplification factors decrease
according to the increase in maximum acceleration of the input motion to 100 Gal and 200 Gal.
These phenomena were obviously produced by the nonlinearity of the soil.
Figure 5 shows the relationship between horizontal displacement and rotational angle of the
footing based on the data from sinusoidal excitation of 3.5 Hz, in order to investigate the
rotational characteristics of the footing of the verticalpile foundation and the batterpile
foundation. The rotational angle is calculated by dividing the difference in the vertical
displacement based on the data of the accelerometers installed at both sides of the footing by the
distance between the two accelerometers.
The fact that there is no phase difference between the sway and the rocking motion indicates
that the response of the footing to motion to the right is counterclockwise rotation, as shown in
Figure 7. There is no phase difference between the sway and the rocking motion of the
verticalpile foundation; on the other hand, antiphase behavior can be seen in the data for the
batterpile foundation.
Figure 6 shows the data obtained from El Centro record excitation at the maximum
acceleration of 200 Gal. The same trend as seen in the case of sinusoidal excitation can also be
found in Figure 6. The phenomena of the opposite phase between the sway and the rocking
motions of the verticalpile foundation and the batterpile foundation can be found in all of the
other test data. From Figures 5 and 6, it can also be seen that the rotation angles of the
batterpile foundation are almost two times larger than those of the verticalpile foundation.
Figure 8 shows the maximumvalue distribution of the bending and axial strains of the piles
in the verticalpile foundation (pileVA1) and the batterpile foundation (pileBA1) obtained
from sinusoidal excitation of 3.5 Hz. The frequency of 3.5 Hz closely corresponds to that of the
predominant frequency of the ground as shown in Figure 3. The largest values were obtained at
the pile heads, and the bending and axial strains of the batterpile foundation are larger than
those of the verticalpile foundation in all cases, as shown in Figure 8.
Figure 9 shows the maximumvalue distribution of the bending and axial strains of the piles
in the verticalpile foundation (pileVA1) and the batterpile foundation (pileBA1) obtained
from El Centro record excitation. The largest values were obtained at the pile heads, and the
bending and axial strains of the batterpile foundation are larger than those of the verticalpile
foundation, likely due to the sinusoidal excitation.
Figures 10 and 11 show the maximum values for acceleration of the footings and the ground
surface, and the bending and axial strains at the pile heads corresponding to the increments in
maximum acceleration of the input motion. From the figures, it can be seen that the maximum
acceleration of the footing of the verticalpile foundation is larger than that of the batterpile
foundation and that both the bending and axial pile strain of the batterpile foundation are larger
than those of the verticalpile foundation in both the sinusoidal and El Centro record excitation.
4 ASEISMICITY OF BATTER PILE
Figures 12 and 13 compare the frequency transfer functions of the horizontal acceleration of the
footing and input motion between the verticalpile foundation and the batterpile foundation
obtained from sweep test and El Centro record excitation. The difference between the frequency
transfer functions of the two pile foundations represents the aseismicity of the batterpile
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
22
foundation. From these figures, it can be elucidated that the batterpile foundation has a certain
level of aseismicity in all of the frequency ranges.
Figures 14 and 15 compare the frequency transfer functions of the bending and axial strains
of the piles and input motion between the verticalpile foundation and the batterpile foundation
calculated using the data from the sweep test and El Centro record excitation. From these
figures, it can be seen that the strain of the batter piles is larger than that of the vertical piles.
Therefore, it is considered that the compensation for the aseismicity of batter piles seeks large
crosssectional efficiency for the batter piles.
5 CONCLUSIONS
The main conclusions of the study are as follows:
1) The response of the footing of the verticalpile foundation to motion to the right is
counterclockwise rotation. On the other hand, that of the batterpile foundation is rotation
in the opposite direction to that of the verticalpile foundation.
2) Bending and axial strains attain the largest values at the pile heads in both the verticalpile
foundation and batterpile foundation.
3) Improved aseismicity by adopting batter piles can be gained in almost all frequency ranges.
4) Bending and axial strains of the batterpile foundation are larger than those of the
verticalpile foundation. In other words, the compensation for the aseismicity of batter piles
seeks large crosssectional efficiency for the batter piles.
6 ACNOLEDGEMENT
We would like to express our sincerest gratitude to Mr. Katsumi Yoshinari for his invaluable
help to carry out the centrifuge tests.
7 REFERENCES
Fan, K., Gazetas, G., Kaynia, A., Kausel, E., & Ahmad, S. 1991. Kinematic seismic response of single
piles and pile groups, Journal of Geotechnical Engineering, 117, 18601879.
Gerolymos, N., Giannakou, A., Anastasopoulos, I. & Gazetas, G. 2008. Evidence of beneficial role of
inclined piles: observations and summary of numerical analyses, Bulletin of Earthquake Engineering:
Vol. 6, No. 4, 705722: Springer.
Giannakou, A., Gerolymos, N., & Gazetas, G. 2007. Kinematic response of groups with inclined piles,
Proceedings of the 4
th
International Conference on Earthquake and Geotechnical Engineering
(ICEGE), Thessaloniki, Greece, on CDRom.
Mylonakis, G., Nikolau, S., & Gazetas, G. 1997. Soilpile bridge seismic interaction: kinematic and
inertial effects. Part I: Soft soil, Earthquake Engineering and Structural Dynamics, 26, 337359.
Mylonakis, G. 2001. Simplified model for seismic pile bending at soil layer interfaces, Soils and
Foundations, 41, 4, 4758.
Nikolaou, S., Mylonakis, G., Gazetas, G., & Tazoh, T. 2001. Kinematic pile bending during earthquakes:
analysis and field measurements, Geotechnique, 51, 5, 425440.
Poulos, N. 2006. Raked pilesvirtues and drawbacks, Journal of Geotech Geoenviron Engineering,
132(6), 795803. Doi:10.1061/(ASCE) 10900241(2006)132:6(795).
Sica, S., Mylonakis, G., & Simonelli, A. L., 2007. Kinematic bending of piles: Analysis vs. code
provisions, Proceedings of the 4
th
International Conference on Earthquake and Geotechnical
Engineering (ICEGE), Thessaloniki, Greece, on CDRom.
Tazoh, T., Shimizu, K., & Wakahara, T. 1987. Seismic observations and analysis of grouped piles.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
23
Geotechnical Special Publication No. 11, Dynamic Response of Pile Foundations, ASCE, 120.
Tazoh, T., Sato, M., & Gazetas, G. 2005. Centrifuge tests on pilefoundation structure systems affected
by liquefactioninduced flow due to quaywall collapse, Proceedings of the 1
st
GreeceJapan
Workshop on Seismic Design, Observation and Retrofit of Foundations, Athens, Greece, 79106.
Tazoh, T., Sato, M., Jang, J., & Gazetas, G. 2007. Centrifuge tests on remedial measure using batter piles
against liquefactioninduced soil flow after quay wall failure, Proceedings of the 2
nd
GreeceJapan
Workshop on Seismic Design, Observation and Retrofit of Foundations, Tokyo, Japan, 431439.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
24
Figure 1 Longitudinal sections and plan of the 1/30scale centrifuge model
(scale unit: mm, for the prototype dimensions: multiply by 30. A verticalpile foundation and a
batterpile foundation without the presence of a superstructure were set parallel to each other in a soil
container which was filled with dry sand, and were excited simultaneously.)
Photograph 1 Testing Model
(The pile foundations have four piles.)
807
4
7
5
3
2
0
2
3
7
.
5
3
0
0
2
0
3
2
0
3
0
0
2
0
2
3
7
.
5
330
DVFX
DBFX
AGX1
10 10
AGX55
AGZ1
AVFZ
330
DTX1
AGX2
AGX3
AGX4
AVF+Z
AVFX
ABFX
ABF+Z ABFZ
AGX6 AGX7
AGX8 AGX9
SVA130 SVA230
SVA2260
SVA2190
SVA2130
SVA270
SVA1260
SVA1190
SVA1130
SVA170
SBA230 SBA130
SBA1260
SBA1190
SBA1130
SBA170
SBA2260
SBA2190
SBA2130
SBA270
DGX
AGX1
AGX55 AGZ1
DGZ
DVFX
DBFX
AGX6
AGX7
AVF+Z AVFX
ABSX
ABFZ ABF+Z
DTX1
DGZ
DGX
AGX4
X
Z
X
Z
X
Y
25 50 125 130
3
0
0
1
1
8
EVP
EPB
Accelerometer
Strain gauge
Noncontact displacement meter
dmax
=1.539
dmin
=1.206
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
26
Table 2 Materials and dimensions of the test model
Parts Material & size
Laminar box 805mm, 475mm, 324mmInner size: length, width, depth
Soil deposit Dry sand: Silica No.7Dr 60Thickness: 300 mm
Vertical pile
Stainless steel
No. of piles422, Inclination angle: 0
Length: 270 mmDiameter:10 mmThickness: 0.2 mm
Batter pile
Stainless steel
No. of piles422, Inclination angle: 10
Length: 274 mmDiameter: 10 mmThickness: 0.2 mm
Footing
Steel
Thickness: 30 mmPlan size: 50 mm50 mm
Table 3 Installed sensors
(62 monitoring channels were installed, with the sensors comprising 17 accelerometers,
5 noncontact displacement meters, and 40 strain gauges.)
Transducer Location Direction Number Subtotal Total
Accelerometer
Batter pile
X 2
17
62
Z 2
Vertical pile
X 2
Z 2
Ground X 6
Base X 1
Table control X 1
Centrifugal acc. Z 1
Noncontact
displacement
meter
Batter pile X 1
5
Vertical pile X 1
Ground
X 1
Z 1
Base X 1
Strain gauge
Batter pile
PileBA1 10
40
PileBA2 10
Vertical pile
PileVA1 10
PileVA2 10
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
27
Table 4 Test cases
Input Motion
Freq. of input
motion (Hz)
Max. acc. of input
motion (Gal)
Test case No.
Sweep test motion 1.710 Hz
5 11
15 12
30 13
Sinusoidal
excitation
3.5 Hz
50 21
100 22
200 23
El Centro record
El Centro record
NS component
50 31
100 32
200 33
Figure 3 Frequency transfer function of the ground surface obtained from sweep test
(5 Gal, 15 Gal, 30 Gal)
0
5
10
15
0 2 4 6 8 10
Sweep 5 Gal
Sweep 15 Gal
Sweep 30 Gal
A
m
p
l
i
f
i
c
a
t
i
o
n
Frequency (Hz)
Ground surface
180
90
0
90
180
0 2 4 6 8 10
Sweep 5 Gal
Sweep 15 Gal
Sweep 30 Gal
P
h
a
s
e
(
d
e
g
r
e
e
)
Frequency (Hz)
Ground surface
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
28
Figure 4 Frequency transfer function of the ground surface obtained from the El Centro record excitations
(50 Gal, 100 Gal, 200 Gal)
Figure 5 Comparisons of horizontal displacement and rotational angle of the footings between the
verticalpile foundation and the batterpile foundation (Sinusoidal excitation: 3.5 Hz, 200 Gal)
180
90
0
90
180
0 2 4 6 8 10
El Centro 50 Gal
El Centro 100 Gal
El Centro 200 Gal
P
h
a
s
e
(
d
e
g
r
e
e
)
Frequency (Hz)
Ground surface
15
0
15
0.8
0
0.8
Footing displacement
Footing rotation
D
i
s
p
.
(
m
m
)
R
o
t
a
t
i
o
n
(
d
e
g
.
)
Vertical pile
10
0
10
1.5
0
1.5
0 2 4 6 8 10
Footing displacement
Footing rotation
D
i
s
p
.
(
m
m
)
R
o
t
a
t
i
o
n
(
d
e
g
.
)
Time(sec)
Batter pile
0
10
20
30
0 2 4 6 8 10
El Centro 50 Gal
El Centro 100 Gal
El Centro 200 Gal
A
m
p
l
i
f
i
c
a
t
i
o
n
Frequency (Hz)
Ground surface
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
29
Figure 6 Comparisons of horizontal displacement and rotational angle of the footings between the
verticalpile foundation and the batterpile foundation (El Centro record: 200 Gal)
Figure 7 Kinematic responses of footings
(a) 50 Gal (b) 100 Gal (c) 200 Gal
Figure 8.1 Bending strain distributions of the verticalpile foundation (pileVA1) and the batterpile
foundation (pileBA1) obtained from the sinusoidal excitation of 3.5 Hz (50 Gal, 100 Gal, 200 Gal)
30
0
30
1
0
1
Footing displacement Footing rotation
D
i
s
p
.
(
m
m
)
R
o
t
a
t
i
o
n
(
d
e
g
.
)
Vertical pile
30
0
30
2
0
2
0 2 4 6 8 10
Footing displacement Footing rotation
D
i
s
p
.
(
m
m
)
R
o
t
a
t
i
o
n
(
d
e
g
.
)
Time(sec)
Batter pile
()
()
AV_FZ() AV_FZ()
AV_FX
()
AB_FZ () AB_FZ()
AB_FX
()
0 50 100
9
6
3
0
Vertical pile
Batter pile
Bending strain (x10
6
)
D
e
p
t
h
o
f
p
i
l
e
h
e
a
d
(
m
)
0 100 200
9
6
3
0
Vertical pile
Batter pile
Bending strain (x10
6
)
D
e
p
t
h
o
f
p
i
l
e
h
e
a
d
(
m
)
0 200 400
9
6
3
0
Vertical pile
Batter pile
Bending strain (x10
6
)
D
e
p
t
h
o
f
p
i
l
e
h
e
a
d
(
m
)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
30
(a) 50 Gal (b) 100 Gal (c) 200 Gal
Figure 8.2 Axial strain distributions of the verticalpile foundation (pileVA1) and the batterpile
foundation (pileBA1) obtained from the sinusoidal excitation of 3.5 Hz (50 Gal, 100 Gal, 200 Gal)
(a) 50 Gal (b) 100 Gal (c) 200 Gal
Figure 9.1 Bending strain distributions of the verticalpile foundation (pileVA1) and the batterpile
foundation (pileBA1) obtained from El Centro record (50 Gal, 100 Gal, 200 Gal)
0 25 50
9
6
3
0
Vertical pile
Batter pile
Bending strain (x10
6
)
D
e
p
t
h
o
f
p
i
l
e
h
e
a
d
(
m
)
0 50 100
9
6
3
0
Vertical pile
Batter pile
Bending strain (x10
6
)
D
e
p
t
h
o
f
p
i
l
e
h
e
a
d
(
m
)
0 75 150
9
6
3
0
Vertical pile
Batter pile
Bending strain (x10
6
)
D
e
p
t
h
o
f
p
i
l
e
h
e
a
d
(
m
)
0 50 100
9
6
3
0
Vertical pile
Batter pile
Bending strain (x10
6
)
D
e
p
t
h
o
f
p
i
l
e
h
e
a
d
(
m
)
0 150 300
9
6
3
0
Vertical pile
Batter pile
Bending strain (x10
6
)
D
e
p
t
h
o
f
p
i
l
e
h
e
a
d
(
m
)
0 200 400
9
6
3
0
Vertical pile
Batter pile
Bending strain (x10
6
)
D
e
p
t
h
o
f
p
i
l
e
h
e
a
d
(
m
)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
31
(a) 50 Gal (b) 100 Gal (c) 200 Gal
Figure 9.2 Axial strains distributions of the verticalpile foundation (pileVA1) and the batterpile
foundation (pileBA1) obtained from El Centro record (50 Gal, 100 Gal, 200 Gal)
Figure 10 Maximum values of the accelerations of the footings and the ground surfaces, and the
bending and axial strains at the pileheads (Sinusoidal excitation: 3.5 Hz)
0 25 50
9
6
3
0
Vertical pile
Batter pile
Bending strain (x10
6
)
D
e
p
t
h
o
f
p
i
l
e
h
e
a
d
(
m
)
0 50 100
9
6
3
0
Vertical pile
Batter pile
Bending strain (x10
6
)
D
e
p
t
h
o
f
p
i
l
e
h
e
a
d
(
m
)
0 75 150
9
6
3
0
Vertical pile
Batter pile
Bending strain (x10
6
)
D
e
p
t
h
o
f
p
i
l
e
h
e
a
d
(
m
)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
32
Figure 11 Maximum values of the accelerations of the footings and the ground surfaces, and the
bending and axial strains at the pileheads (El Centro record)
(a) 5 Gal (b) 15 Gal (c) 30 Gal
Figure 12 Aseismicity of the batterpile foundation: Comparison of the frequency transfer function
between the horizontal acceleration and input motion of the footing in the verticalpile foundation and the
batterpile foundation obtained from sweep tests (5 Gal, 15 Gal, 30 Gal)
0
2.5
5
7.5
10
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
Frequency (Hz)
Footing
(Sweep 5 Gal)
0
2.5
5
7.5
10
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
Frequency (Hz)
Footing
(Sweep 15 Gal)
0
2.5
5
7.5
10
0 2 4 6 8 10
Verticalpile
Batterpile
A
m
p
l
i
f
i
c
a
t
i
o
n
Frequency (Hz)
Footing
(Sweep 30 Gal)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
33
(a) 50 Gal (b) 100 Gal (c) 200 Gal
Figure 13 Aseismicity of the batterpile foundation: Comparison of the frequency transfer function
between the horizontal acceleration and input motion of the footing in the verticalpile foundation and the
batterpile foundation obtained from El Centro record (50 Gal, 100 Gal, 200 Gal)
(a) Bending strains
(b) Axial strains
Figure 14 Comparisons of the frequency transfer functions of the bending and axial strains of the piles
and input motion between the verticalpile foundation and the batterpile foundation (sweep tests)
0
2
4
6
8
10
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
Frequency (Hz)
Footing
(El Centro 50 Gal)
0
2
4
6
8
10
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
Frequency (Hz)
Footing
(El Centro 100 Gal)
0
2
4
6
8
10
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
Frequency (Hz)
Footing
(El Centro 200 Gal)
0
1
2
3
4
5
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
(
m
i
c
r
o
/
G
a
l
)
Frequency (Hz)
Sweep 5 Gal
PileA1 head
(Bending strain)
0
1
2
3
4
5
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
(
m
i
c
r
o
/
G
a
l
)
Frequency (Hz)
Sweep 15 Gal
PileA1 head
(Bending strain)
0
1
2
3
4
5
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
(
m
i
c
r
o
/
G
a
l
)
Frequency (Hz)
Sweep 30 Gal
PileA1 head
(Bending strain)
0
0.5
1
1.5
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
(
m
i
c
r
o
/
G
a
l
)
Frequency (Hz)
Sweep 5 Gal
PileA1 head
(Axial strain)
0
0.5
1
1.5
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
(
m
i
c
r
o
/
G
a
l
)
Frequency (Hz)
Sweep 15 Gal
PileA1 head
(Axial strain)
0
0.5
1
1.5
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
(
m
i
c
r
o
/
G
a
l
)
Frequency (Hz)
Sweep 30 Gal
PileA1 head
(Axial strain)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
34
(a) Bending strains
(b) Axial strains
Figure 15 Comparisons of the frequency transfer functions of the bending and axial strains of the piles
and input motion between the verticalpile foundation and the batterpile foundation (El Centro record)
0
1
2
3
4
5
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
(
m
i
c
r
o
/
G
a
l
)
Frequency (Hz)
El Centro 50 Gal
PileA1 head
(Bending strain)
0
1
2
3
4
5
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
(
m
i
c
r
o
/
G
a
l
)
Frequency (Hz)
El Centro 100 Gal
PileA1 head
(Bending strain)
0
1
2
3
4
5
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
(
m
i
c
r
o
/
G
a
l
)
Frequency (Hz)
El Centro 200 Gal
PileA1 head
(Bending strain)
0
0.5
1
1.5
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
(
m
i
c
r
o
/
G
a
l
)
Frequency (Hz)
El Centro 50 Gal
PileA1 head
(Axial strain)
0
0.5
1
1.5
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
(
m
i
c
r
o
/
G
a
l
)
Frequency (Hz)
El Centro 100 Gal
PileA1 head
(Axial strain)
0
0.5
1
1.5
0 2 4 6 8 10
Vertical pile
Batter pile
A
m
p
l
i
f
i
c
a
t
i
o
n
(
m
i
c
r
o
/
G
a
l
)
Frequency (Hz)
El Centro 200 Gal
PileA1 head
(Axial strain)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
35
1 INTRODUCTION
Batter piles have been used for a long time to resist large lateral loads from winds, water waves,
soil pressures, and impacts. Their distinct advantage over vertical piles is that they transmit the
applied loads partly in axial compression, rather than only through shear and bending. Thus,
batter piles offer larger stiffness and bearing capacity than samediameteranddepth vertical
piles a superiority of particular importance when the nearsurface soils are soft and/or the
lateral load is large.
Despite these advantages, they do not enjoy a good reputation for seismic resistance.
Following the poor performance of batter piles in a series of earthquakes, the seismic behaviour
of inclined piles has been considered detrimental, and many codes require that such piles be
avoided. For instance, the French Seismic Code (AFPS 90) states flatly that Inclined piles
should not be used to resist seismic loads. The seismic Eurocode EC8 / Part 5, dealing with
geotechnics and foundations, is a little less restrictive, stating: It is recommended that no in
clined piles be used for transmitting lateral loads to the soil. If, in any case, such piles are used,
they must be designed to carry safely axial as well as bending loading.
Case histories that have recently confirmed the potential for unsatisfactory performance of
improperlydesigned batter piles include the wharf in the Port of Oakland in the 1989 Loma
Kinematic and Inertial Behavior of Raked Piles
A. Giannakou
1
, N. Gerolymos
2
, G. Gazetas
2
1
Fugro West, Oakland, California, USA
2
National Technical University of Athens, Greece
ABSTRACT: Several aspects of the seismic response of groups containing nonvertical piles are
studied, including the lateral pile head stiffnesses, the kinematic pile deformation, and the
inertial soilpilestructure response. A key goal is to explore the conditions under which the
presence of batter piles is beneficial, indifferent, or detrimental. Parametric analyses are carried
out using 3D finite element modeling, assuming elastic behavior of soil, piles, and superstruc
ture. The model is first utilized to obtain the lateral stiffnesses of single batter piles and to show
that its results converge to the available solutions from the literature. Then, real accelerograms
covering a broad range of frequency characteristics are employed as base excitation of simple
fixedhead 2pile group configurations, embedded in homogeneous, inhomogeneous, and
layered soil profiles, while supporting very tall or very short structures. Five pile inclinations are
considered while the corresponding verticalpile group results serve as reference. It is found
that in purely kinematic seismic loading, batter piles tend to confirm their negative reputation,
as had also been found recently for a group subjected to static horizontal ground deformation.
However, the total (kinematic plus inertial) response of structural systems founded on groups
of batter piles offers many reasons for optimism. Batter piles may indeed be beneficial (or
detrimental) depending on, among other parameters, the relative size of the overturming
moment versus the shear force transmitted onto them from the superstructure.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
36
Prieta earthquake (M
s
= 7.1), the Port of Los Angeles in the 1994 Northridge earthquake (M
s
=
6.8), and the Rio Banano and the Rio Vizcaya Bridges in the 1991 Costa Rica earthquake (M
s
=
7.5). The bad reputation of batter piles has been reinforced by these incidents. The culmination
was the following statement in the ASCE monograph on Seismic Design of Port and Harbor
Facilities : The use of batter piles in ports is typically not encouraged because of their poor
seismic performance during past earthquakes. [Technical Council on Lifeline Earthquake En
gineering 2001, Kavazanjian, 2006].
However, a more thorough investigation on the causes of these failures, showed that the in
adequate reinforcement in the top of the piles and also the improper connection of piles to their
caps were the culprits of the observed damage (Mitchell et al, 1991; Priestley et al, 1991) a
result of the early isostatic method of analysis which assumed that batter piles transmit only
axial load.
But what if batter piles were properly designed to resist the developed moment and shear
loads at their head ? And if, furthermore, they were designed to posses sufficient ductility at the
head and the connection to the cap ? Would their seismic behavior still remain poor ?
A goal of this paper is to give at least a partial answer to these questions. Indeed, in recent
years, evidence has been accumulating that welldesigned batter piles may not only have a satis
factory performance themselves, but may also be beneficial for the structure they support. Re
cent research on the seismic response of batter piles and micropiles (Guin, 1997; Lam and Mar
tin, 1987; Sadek and Shahrour, 2004 and 2006) has shown that the seismic response of a
structure may improve in many respects when supported by inclined piles. Moreover, case his
tories referring to the Maya Warf in the Kobe 1995 earthquake and the Landing Road Bridge in
the Edgecumbe New Zealand 1987 earthquake have highlighted the potential help provided by
inclined piles (Berril et al 2001, Gazetas & Mylonakis 1998).
As a result of the improved understanding of the source of the observed poor performance,
batter piles in recent years seem to have been reestablished in their traditional role of with
standing large horizontal loads applied to deep foundations (as pointed out in an enlightening
professional article by Kavazanjian, 2006). The piers for the new San Francisco Bay Bridge
East Span present a characteristic example of this trust in batter piles to carry huge lateral seis
mic loads in very soft soil.
Presently, research on the seismic response of batter piles has been rather limited (Juran et al,
2001; Sadek and Shahrour, 2004 and 2006; Okawa et al, 2005; Poulos, 2006 ; Deng et al, 2007 ;
and Ravazi et al, 2007). Aiming at filling part of this gap, we study several aspects of the
seismic response of batter piles through parametric 3D analyses employing the FE method. Only
the idealized case of linear viscoelastic soil response is treated here the shortcomings of line
arity will be explored in a forthcoming companion paper, but the conclusions drawn in this
study remain qualitatively valid even in the presence of soil nonlinearities.
The first part of the paper outlines the numerical model and shows its consistency with avail
able analytical results.
2 PROBLEM DEFINITION AND FINITE ELEMENT MODELING
The seismic behavior of symmetric 2 x 1 group configurations with piles battered at various
angles is investigated using ABAQUS. Batter angles commonly encountered in practice are
considered, such as 5
o
15
o
, in addition to the less usual cases of 20
o
and 25
o
. The vertical fixed
head pile group is used as a reference for delineating the role of pile inclination. Figure 1 depicts
the finite element meshes of the six configurations. All piles are of Youngs modulus E
P
= 30
GPa, diameter d = 1 m, and the depth to their tip is L = 15 m. The centertocenter distance, s,
between the piles at pilehead elevation is three pile diameters (S = 3d). The piles are rigidly
connected to a perfectlyrigid pile cap. The massandcolumn superstructure is modeled as a
single degree of freedom oscillator. The concentrated superstructure mass M
str
is such that the
load per pile in each configuration is 1.0 MN typical of actual pile designs. In all cases
studied, the fundamental period of the superstructure is 0.44 sec and of the soil 0.29 sec.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
37
= 15
= 20
= 25
= 0
= 5
= 10
= 15
= 20
= 25
= 0
= 5
= 10
s1
s2
s1
s2
3.5 d
I II III IV
s
z / d
T
s
= 0.29 s
Soil Profiles
11.5 d
3d
L
=
1
5
m
d = 1 m
s
z
s1
s2
s1
s2
3.5 d
I II III IV
s
z / d
T
s
= 0.29 s
Soil Profiles
11.5 d
3d
L
=
1
5
m
d = 1 m
s
Figure 1.The six studied pile group configurations and the corresponding 3D finite element models.
Four idealized soil profiles are considered : (a) a homogenous, (b) a nonhomogenous Gib
son soil, (c) a twolayer profile with a bottom stiffer layer, and (d) a twolayer profile with a
top stiffer layer [crust] (Figure 2).
Both pile and soil are linear viscoelastic. Soil is modeled with eightnoded brick elements
while the piles are represented with a series of beam elements rigidly linked to the peripheral
nodes in order to properly model the pile geometry. The performed mesh sensitivity study
revealed that an element dimension of 0.5 m (i.e., one pile radius) leads to nearly accurate
results. Appropriate kinematic constraints are imposed to the lateral edges of the model, allow
ing it to move as the free field, while elementary transmitting boundaries (V dashpots) ab
sorb much of the wave energy emitted from the oscillating piles.
Three real acceleration time histories, covering a wide range of frequencies, are used as base
excitation: (i) The record of the 2003 Lefkada M
s
6.4 earthquake: PGA = 0.42 g, dominant pe
riod range T
p
0.20.65 sec, [Gazetas et al., 2005]; (ii) the rock outcrop motion of the 1995 Ae
gion M
s
6.2 earthquake : PGA = 0.39 g, dominant period range T
p
0.140.6 sec [Gazetas,
1996]; and (iii) the JMA record of the 1995 M
JMA
= 7.2 Kobe earthquake PGA = 0.83 g, T
p
0.251.0 sec [Architectural Institute of Japan, 1995].
Figure 2.Problem geometry and the four soil profiles.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
38
3 STATIC STIFFNESSES, CONVERGENCE TO PUBLISHED SOLUTIONS
As a starting point, the static stiffnesses of a single batter pile are computed and contrasted with
available solutions for vertical (Poulos & Davis 1980, Gazetas 1991) and batter (Poulos 1980)
piles in Figures 3 and 4. For a homogeneous soil stratum of thickness H, which exceeds the
depth L to the pile tip (Soil Model I), our numerical (FE) results are compared with those com
puted : (a) on the basis of Poulos (1980) and Poulos & Daviss (1980) approximation for angles
of batter = 0
and 25
o
, and (b) the approximate closedform expressions of Gazetas (1991) for
a vertical pile. The latter were developed using the Blaney & Roesset (1976) innovative
dynamic finiteelement which incorporates perfect transmitting boundaries. Figure 3 depicts the
variation of the normalized lateral static stiffnesses, K
HH
, K
RR
, K
HR
, as functions of the ratio
E
p
/E
s
of the pile and soil Youngs moduli, and the angle of batter as a parameter. We draw the
following conclusions :
The rocking K
RR
and, especially, the crosscoupled horizontalrocking K
HR
stiffnesses are
essentially independent of : vertical and battered piles have nearly identical response, as
expected by Poulos 1980 simplification (Poulos & Madhav 1971, Poulos & Davis 1980).
Therefore the closedform expressions developed for vertical piles (Gazetas 1991),
( )
0 75
3
0 15
RR s p s
K E d E E
.
. (1)
( )
0 50
2
0 22
HR s p s
K E d E E
.
. (2)
are in excellent accord with both the FE results of this study and Poulos approximation.
The above expressions can therefore be used even with batter piles.
The horizontal stiffness K
HH
exhibits a small dependence on . Both the FE analysis and
Poulos approximation show that stiffness increases by about 30% on average when
increases from 0
o
to 25
o
. In percentage, the difference declines with increasing E
P
/E
s
ratio.
The results of the Poulos (1980) and Gazetas (1991) approximations only slightly underes
timate the finite results. The following simple expression :
( ) ( )
2
0.21 tan
2
1.08 1 4 tan +
HH s p s
K E d E E
(3)
has been developed by fitting the FE results for batter piles. It will suffice in practical
applications for any angle .
For the linearlyinhomogeneous stratum (Gibson Soil) with Youngs modulus of the form :
( ) ( ) =
s
E z E z d (4)
in which apparently E
S
= is the modulus at onediameter depth, similar conclusions can be
drawn from Figure 4 :
The rocking and cross swayingrocking stiffnesses, K
RR
and K
HR
, are again practically
unaffected by the inclination of the pile, while they are sensitive to the E
P
/E
S
ratio. The
Poulos (1980) and Gazetas (1991) approximations are again in excellent accord with the
present FE results. Thus the expressions developed for vertical piles
( )
0 80
3
0 15
RR s p s
K E d E E
.
. (5)
( )
0 60
2
0 17
HR s p s
K E d E E
.
. (6)
provide very good estimates for all values of . (he reader should notice that E
S
, the
modulus at depth z = d, in the above Eqns has a different meaning from the constant E
S
modulus of Eqns 13.)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
39
0
0
0
0
0
0
E
d
L
L
(z)
E
s
100 1000 10000 100000
E
p
/ E
s
= 0 [Poulos, 1980]
= 25
o
[Poulos, 1980]
= 0 [Gazetas, 1991]
= 0 [FE Analysis]
= 25
o
[FE Analysis]
12
10
8
6
4
5
0
200
160
120
80
40
0
30
25
20
15
10
5
0
K
H
H
/
E
s
d
K
R
R
/
E
s
d
3
K
H
R
/
E
s
d
2
=25
o
=0
o
all
0
0
0
0
0
0
E
d
L
L
(z)
E
s
E
d
L
L
(z)
E
s
100 1000 10000 100000
E
p
/ E
s
= 0 [Poulos, 1980]
= 25
o
[Poulos, 1980]
= 0 [Gazetas, 1991]
= 0 [FE Analysis]
= 25
o
[FE Analysis]
12
10
8
6
4
5
0
200
160
120
80
40
0
30
25
20
15
10
5
0
K
H
H
/
E
s
d
K
R
R
/
E
s
d
3
K
H
R
/
E
s
d
2
=25
o
=0
o
all
Figure 3. Normalized static stiffnesses of single pile : (a) swaying, (b) rocking, and (c) cross swaying
rocking for batter angle = 0
o
and 25
o
as a function of pilesoil stiffness ratio E
p
/ E
s
(L/d = 15, homoge
nous soil). Comparison with solution for vertical and inclined piles from the literature.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
40
100 1000 10000 100000
E
p
/ E
s
= 0 [Poulos, 1980]
= 25
o
[Poulos, 1980]
= 0 [Gazetas, 1991]
= 0 [FE Analysis]
= 25
o
[FE Analysis]
E
d
L
L
(z)
E
s
z/d
60
50
40
30
20
10
0
300
250
200
150
100
50
0
54
45
36
27
18
9
0
=25
o
=0
o
all
K
H
H
/
E
s
d
K
R
R
/
E
s
d
3
K
H
R
/
E
s
d
2
100 1000 10000 100000
E
p
/ E
s
= 0 [Poulos, 1980]
= 25
o
[Poulos, 1980]
= 0 [Gazetas, 1991]
= 0 [FE Analysis]
= 25
o
[FE Analysis]
E
d
L
L
(z)
E
s
z/d
E
d
L
L
(z)
E
s
z/d
60
50
40
30
20
10
0
300
250
200
150
100
50
0
54
45
36
27
18
9
0
=25
o
=0
o
all
K
H
H
/
E
s
d
K
R
R
/
E
s
d
3
K
H
R
/
E
s
d
2
Figure 4. Normalized static stiffnesses of single pile : (a) swaying, (b) rocking, and (c) cross swaying
rocking for batter angles = 0
o
and 25
o
as a function of pilesoil stiffness ratio E
p
/ E
s
(L/d = 15, inho
mogenous Gibson soil). Comparison with solution for vertical and inclined piles from the literature.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
41
On the contrary, the horizontal stiffness K
HH
is very sensitive to the batter angle . ncreasing
25
1
Q
1
2
Q
2
Q
(a)
(b)
Q
Q / 2
N 0
1
Q
1
2
Q
2
Q
(a)
(b)
Figure 8. Exaggerated snapshots of the deformed shapes of groups with vertical piles (lefthand side) and
inclined (righthand side) supporting: (a) short (H
str
= 1 m) and (b) tall (H
str
= 12 m) superstructures in
Gibson soil
Figure 9. Mechanisms for undertaking the inertial forces of a short superstructure for : (a) vertical, and
(b) inclined pile groups. The vectors indicate the forces imposed from the on the piles. The dashed lines
correspond to the location of the cap if the axial displacements of the piles are completely ignored.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
46
M
q
f
M
q
f
M
q
f
With the tall structure, where the overturning moment dominates, the vertical piles are sub
jected to a pair of axial forces that undertake most of this load (Figure 10a). Secondarily, head
moments develop due to pile fixity to the cap. In stark contrast, batter piles may undertake this
large moment mainly largely by flexure [bending] (Figure 10b). As a result, substantial cap
rotations take place. Indeed, the capped batter piles are now rotationally more flexible than the
capped vertical piles! To elucidate why, imagine two piles inclined at very large (certainly un
realistic) angles (Figure 11) with a large overturning moment applied on the cap : replacing
conceptually the soil with Winkler springs, one can realize that the piles will unavoidably bend,
activating these springs alternately in tension and compression, near the cap, q. The (remaining)
portion of the load, undertaken by the framing action of the piles activates the friction, f, at the
pilesoil interface, would be very small (equal to zero in the extreme [unrealistic, of course] case
of two piles inclined at 90
o
).
Figure 11. Mechanisms for undertaking the inertial forces of a short superstructure for : (a) vertical, and
(b) inclined pile groups. The vectors indicate the forces imposed from the on the piles. The dashed lines
correspond to the location of the cap if the axial displacements of the piles are completely ignored.
Having analysed the mechanics of the pile distress, let us examine the structural distress. Di
mensionless diagrams for the drift of the superstructure are presented in Figure 12. The
horizontal drift when the structure is supported on a group of batter piles is generally smaller
than with the exclusivelyvertical pilegroup . This reduction of structural distress is
appreciable only with the tall structure, and is attributed to the observed simultaneous
occurrence of smaller horizontal cap displacement, and larger cap rotation with the group of
batter piles. Increased pile batter relates to smaller structural distress and smaller horizontal
displacement of the mass, at the cost of larger cap rotation.
6 CONCLUSIONS
This paper has tried to contribute towards a better understanding of the seismic behavior of bat
ter piles, which may under certain circumstances be beneficial rather than detrimental, for both
the structure they support and the piles themselves. Admittedly, the linear approximation of the
SSI phenomena is not without shortcomings. Phenomena created by strong nonlinearities of the
soil, such as permanent soil deformations due to extensive soil plastification, residual bending
moments on the piles etc, cannot be captured by linear (or equivalent linear) FE analyses. None
theless, valuable insight is gained into the description, understanding, and explanation of the be
havior of batter piles. The main conclusions are as follows :
1) The purely kinematic response of batter piles tends to confirm their negative reputation :
the parametric analyses show that they experience larger bending moments than vertical
piles. Moreover, batter piles exhibit significantly larger axial forces than vertical piles for
all four idealized profiles, due to exclusively horizontal shaking of the soil. In fact both of
these internal forces increase as the angle increases. This conclusion is in full accord with
the conclusion of a recent study by Poulos (2006) who imposed on the piles a lateral static
ground displacement linearly decreasing with depth, and thus not very different in shape
from the first mode freefield displacements that is the main source of kinematic pile
straining in our study.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
47
0
o
5
o
10
o
15
o
20
o
25
o
1.4
1.2
1
0.8
0.6
0.4
0.2
0
H
str
= 1 m
H
str
= 12 m m
a
x
U
s
t
r
m
a
x
U
s
t
r
,
f
o
r
v
e
r
t
i
c
a
l
g
r
o
u
p
0
o
5
o
10
o
15
o
20
o
25
o
1.4
1.2
1
0.8
0.6
0.4
0.2
0
H
str
= 1 m
H
str
= 12 m m
a
x
U
s
t
r
m
a
x
U
s
t
r
,
f
o
r
v
e
r
t
i
c
a
l
g
r
o
u
p
0
o
5
o
10
o
15
o
20
o
25
o
1.4
1.2
1
0.8
0.6
0.4
0.2
0
H
str
= 1 m
H
str
= 12 m m
a
x
U
s
t
r
m
a
x
U
s
t
r
,
f
o
r
v
e
r
t
i
c
a
l
g
r
o
u
p
m
a
x
U
s
t
r
m
a
x
U
s
t
r
,
f
o
r
v
e
r
t
i
c
a
l
g
r
o
u
p
Figure 12. Total (kinematic + inertial) response: normalized peak maximum horizontal drift of a short
(H
str
= 1 m) and a tall (H
str
= 12 m) structure. Normalization with respect to the response of the group of
vertical piles (maximum, minimum, and average values from the three accelerograms [Lefkada, Aegion,
JMA]). (Gibson soil, T
soil
= 0.29 s, E
p
/ E
s
= 1000, T
str
= 0.44 s).
2) However, the total (kinematic plus inertial) response of structural systems founded on
groups of batter piles offers many reasons for optimism. It has been shown that the role of
batter may be quite beneficial or detrimental depending on the relation between shear force
and overturning moment. Specifically, a tall (slender) structure and a short (squatty) struc
ture have been selected, as two extreme cases : large moment and small shear charac
terizes the tall structure; large shear and small moment the short one.
3) For the batter piles supporting a tall slender structure we conclude that :
(a) Configurations with batter piles undergo smaller horizontal displacements than the
vertical group, but at the same time they develop larger cap rotations, often outof
phase with lateral displacements.
(b) The bending moment in batter piles is larger than in vertical piles. In fact, the bending
moment increases as the pile rake increases. Proper reinforcement of the pilecap
connection is necessary for undertaking safely this bending moment and securing
adequate inelastic deformation in case of an unpredictably large (exceeding the
design) earthquake motion.
(c) Perhaps surprisingly, the symmetric group of batter piles attracts smaller axial forces
than a group of exclusively vertical piles! This is attributed to the disproportionately
large share of the overturning moment resisted by axial reactions in the vertical piles 
 not the case with batter piles which undertake this moment partially with flexure
(bending).
(d) The lateral distortion (and drift) of the structure on batter piles is significantly smaller
than of on vertical piles.
4) For batter piles supporting a short squatty structure :
(a) Embedded in Gibson soil they develop smaller bending moments than vertical piles,
within the range of the considered excitations.
(b) Now batter piles sustain larger axial forces than the vertical group, for two reasons : (i)
kinematic loading, which constitutes an important component of the total loading,
produces larger head moment in the batter piles ; (ii) the inertial loading induces
mostly a dynamic shear force which, while being resisted by lateral loads in vertical
piles, it loads axially (and laterally) the batter piles.
(c) The horizontal drift of the superstructure is less sensitive to pile batter.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
48
ACKNOWLEDGEMENT
This work formed part of the EU research project QUAKER which is funded through the EU
Fifth Framework Programme: Environment, Energy, and Sustainable Development, Research
and Technological Development Activity of Generic Nature: the Fight against Natural and
Technological Hazards, under contract number : EVG1CT200200064. We also thank Evan
gelia Garini for her kind contribution in the preparation of the figures.
REFERENCES
AFPS (1990). Association franaise de Gnie Parasismique. Recommandations AFPS 90. Presses des
Ponts et Chausses, Paris.
Architectural Institute of Japan (1995). Strong motion records of the Hyogoken Nanbu Earthquake (in
Japanese).
Banerjee P.K. and Driscoll R.M.C. (1975). A program for the analysis of pile groups of any geometry
subjected to any loading conditions, Department of the Environment, London, HECB/B/7/.
Banerjee P.K. and Driscoll R.M.C. (1976). Three dimensional analysis of raked pile groups, Proceed
ings of Institution Civil Engineers, Part 2, Vol. 61, 653671.
Banerjee P.K. and Davies T.G. (1977). Analysis of pile groups embedded in Gibson soil, Proceedings of
9
th
International Conference Soil Mechanics, Tokyo.
Banerjee P.K. and Davies T.G. (1978). The behavior of axially and laterally loaded piles embedded in
nonhomogeneous soils, Geotechnique, Vol. 28 (3), 309326.
Banerjee P.K. and Davies T.G. (1980). Analysis of some reported case histories of laterally loaded pile
groups, Numerical Methdods in Offshore piling. Institution of Civil Engineers, ICE, London, 101
108.
Bentley K.J. and El Naggar H. M. (2000). Numerical analysis of kinematic response of single piles,
Canadian Geotechnical Journal , Vol. 37, 13681382.
Berrill J. B., Chris Tensen S. A., Keenan R. R., Okada W., & Pettiga R. J (2001), Case study of lateral
spreading forces on a piled foundation, Gotechnique, Vol. 51(6), 501517.
Deng N, Kulesza R., and Ostadan Farhang. (2007). "Seismic soilpile group interaction analysis of a bat
tered pile group", 4
th
International Conference on Earthquake Geotechnical Engineering, Thessalo
niki, June 2528.
Dobry R. & O Rourke M. J. (1983). Discussion on "Seismic response of endbearing piles" by Flores
Berrones R. & Whitman R. V., Journal of Geotechnical Engineering Division, ASCE, Vol. 109.
EC8 (2000). Structures in seismic regions, Part 5: Foundations, retaining structures, and geotechnical
aspects. Seismic Eurocode.
Fan K., Gazetas G., Kaynia A., Kausel E. and Ahmad S. (1991). "Kinematic seismic response of single
piles and pile groups," Journal of Geotechnical Engineering, ASCE, Vol. 117, 18601879
FloresBerrones R. & Whitman R. V. (1982). "Seismic response of endbearing piles," Journal of Geo
technical Engineering Division, ASCE, Vol. 108, 554569.
Gazetas G. (1984). "Seismic response of endbearing piles Soil Dynamics and Earthquake Engineering,
Vol. 3, 8993.
Gazetas G. (1991). Foundation Vibrations, Foundation Engineering Handbook 2
nd
Edition, H.Y. Fang
(Ed), Kluwer/Springer Publishers, Chapter 15, 5535938993.
Gazetas G. and Mylonakis G. (1998). "Seismic soilstructure interaction: new evidence and emerging is
sues," Geotechnical Earthquake Engineering and Soil Dynamics III, ASCE, Vol. 2, 11191174.
Gazetas, G., Anastasopoulos I., and Dakoulas P. (2005). Failure of harbor quaywall in the Lefkada 2003
earthquake, Performance Based Design in Earthquake Geotechnical Engineering : Concepts and Re
search, Proceedings of Geotechnical Earthquake Engineering Satellite Conference, Osaka, 6269.
Giannakou A. (2007). Seismic behavior of inclined piles. Ph.D Thesis, National Technical University of
Athens.
Guin J. (1997). Advances in soilpilestructure interaction and nonlinear pile behavior, PhD thesis, State
University of New York at Buffalo.
Harada T., Kubo K., and Katayama T. (1981). Dynamic soilstructure interaction by continuum formula
tion method, Report, Institute of Industrial Science, University of Tokyo, Vol. 29 (5).
Juran I., Benslimane A., and Hanna S (2001). Engineering analysis of dynamic behavior of micropile
systems, Transportation Research Record, No. 1772, 91106
Kavazanjian E. (2006). A drivenpile andvantage: Batter piles, Piledriver, Q4, 2125
Kavvadas M. and Gazetas G. (1993). "Kinematic seismic response of freehead piles in layered soil,"
Geotechnique, Vol. 43, 207222
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
49
Kaynia A. M. & Kausel E. (1982). Dynamic stiffness and seismic response of pile groups, Research re
port R8203, Massachusetts Institute of Technology.
Lam, P. I and Martin, G.R. (1986). Seismic design of highway bridge foundations, Vol. II, Design
procedures and guidelines. Report No. FHWA/RD86/102, Federal Highway Administration, Vir
ginia.
Mitchell, D., Tinawi, R., and Sexsmith, R.G. (1991). Performance of bridges in the 1989 Loma Prieta
earthquake : Lessons for Canadian designers. Canadian Journal of Civil Engineering, Vol. 18 (4),
711734.
Nikolaou S., Mylonakis G., Gazetas G., and Tazoh T. (2001). Kinematic pile bending during earth
quakes: analysis and field measurements, Geotechnique Vol. 51, 425440.
Okawa K., Kamei H., Zhang F., and Kimura M. (2005). Seismic performance of grouppile foundation
with inclined steel piles, Proceedings of the 1
st
GreeceJapan Workshop on Seismic Design, Obser
vation, and Retrofit of Foundations, Athens, 5360
Poulos H.G. and Madhav M.R. (1971). Analysis of the movement of battered piles, Proceedings of the
1
st
AustraliaNew Zealand Conference on Geomechanics, Institution of Australian Engineers, Mel
bourne, 1, 268275.
Poulos H.G. (1974). Analysis of pile groups subjected to vertical and horizontal loads, Australian Ge
omechanics Journal, G4, No. 1, 2632.
Poulos H.G. (1979). Settlement of single piles in nonhomogeneous soil, Journal of GeotechnicalEngi
neering Division, ASCE, Vol. 105(5), 627641.
Poulos H.G. & Davis E.H. (1980). Pile Foundation Analysis and Design. John Wiley & Sons.
Poulos H.G. (1980). An approach for the analysis of offshore pile groups, Numerical Methods in
Offshore Piling, Institution of civil Engineers. ICE, London, 119126.
Poulos H.G. (1999). Approximate computer analysis of pile groups subjected to loads and ground
movements. International Journal Numerical Analytical Methods in Geomechanics, Vol. 23, 1021
1041,
Poulos H. G. (2006). Raked pilesVirtues and drawbacks Journal of Geotechnical and Geoenviron
mental Engineering, ASCE, Vol. 132, 795803.
Priestley, N., Singh, J., Youd, T., and Rollins, K. (1991). Costa Rica Earthquake of April 22, 1991
Reconnaissance Report, Pub. 9102, Earthquake Eng. Research Inst., 5991.
Randolph M.F. (1981). The response of flexible piles to lateral loading Geotechnique, Vol. 31, 247
259.
Ravazi S.A, Fahker A., and Mirghaderi S.R. (2007). An insight into the bad reputation of batter piles in
seismic performance of wharves, 4
th
International Conference on Earthquake Geotechnical Engi
neering, Thessaloniki, K. Pitilakis editor, June 2528, 2007
Sadek M. & Shahrour I. (2004). "Threedimensional finite element analysis of the seismic behavior of in
clined micropiles", Soil Dynamics and Earthquake Engineering, Vol. 24, 473485.
Sadek M. & Shahrour I. (2006). "Influence of the head and tip connection on the seismic performance of
micropiles," Soil Dynamics and Earthquake Engineering, Vol. 26, 461468.
Takewaki I. & Kishida A. (2005). "Efficient analysis of pilegroup effect on seismic stiffness and strength
design of buildings," Soil Dynamics and Earthquake Engineering, Vol. 25, 355367
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
50
1 INTRODUCTION
Current seismic design of bridge structures is based on a presumed ductile response. A capacity
design methodology ensures that regions of inelastic deformation are carefully detailed to pro
vide adequate structural ductility, without transforming the structure into a mechanism. Brittle
failure modes are suppressed by providing a higher level of strength compared to the corre
sponding to ductile failure modes. For most bridges, the foundation system may be strategically
designed to remain structurally elastic while the pier is detailed for inelastic deformation and
energy dissipation. Essentiallyelastic response of the foundation is usually ensured by increas
ing the strength of the foundation above that of the bridge pier base so that plastic hinging oc
curs in the pier instead of the foundation.
The concept of ductility design for foundation elements is still new in earthquake engineering
practice. The potential development of a plastic hinge in the pile is forbidden in existing regula
tions, codes and specifications. The main reasons are: (i) the location of plastic hinges is not ap
proachable for postseismic inspection and repair, (ii) the high cost associated with repair of a
severely damaged foundation, and (iii) failure due to yielding in the pile prior to exceeding soil
capacity is an undesirable failure mechanism, by contrast to that in which soil capacity is mobi
lized first.
However, several casehistories (especially from the Kobe 1995 earthquake) have shown that:
(a) pile yielding under strong shaking cannot be avoided, especially for piles embedded in soft
soils; and (b) pile integrity checking after an earthquake is a cumbersome, yet feasible task. Fur
thermore, there are structures where plastic hinging cannot be avoided in members of the foun
dation during a severe earthquake. A good example of such structure is the pilecolumn (also
Seismic Response of Bridge PileColumns
V.A. Drosos, N. Gerolymos & G. Gazetas
National Technical University of Athens, Greece
ABSTRACT: While seismic codes do not allow plastic deformation of piles, the Kobe earth
quake has shown that limited structural yielding and cracking of piles may not be always detri
mental. As a first attempt to investigate the consequences of pile yielding in the response of a
pilecolumn supported bridge structure, this paper explores the soilpilebridge pier interaction
to seismic loading, with emphasis on structural nonlinearity. The pilesoil interaction is modeled
through distributed nonlinear Winklertype springs and dashpots. Numerical analysis is per
formed with a constitutive model (Gerolymos and Gazetas, 2005a; 2005b; 2006a) materialized
in the OpenSees finite element code (Mazzoni et al. 2005) which can simulate: the nonlinear
behaviour of both pile and soil ; the possible separation and gapping between pile and soil; ra
diation damping ; loss of stiffness and strength in pile and soil. The model is applied to the
analysis of pilecolumn supported bridge structures, focusing on the influence of soil compli
ance, intensity of seismic excitation, pile diameter, aboveground height of the pile, and above
or below ground development of plastic hinge, on key performance measures of the pier as is:
the displacement (global) and curvature (local) ductility demands and the maximum drift ratio.
It is shown that kinematic expressions for performance measure parameters may lead to errone
ous results when soilstructure interaction is considered.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
51
known in the American practice as extended pileshaft), where the column is continued below
the ground level as a pile of the same or somewhat larger diameter. Obviously, the design of
such foundation requires careful consideration of the flexural strength and ductility capacity of
the pile.
An advantage of supporting a column bent on drilled pile is the cost savings associated with the
construction of large castindrilledhole (CIDH) piles instead of multiple piles of smaller di
ameter, which must later be integrated into a structural unit using a pile cap. Another advantage
of such a design is that localized damage that could otherwise develop at the columnpile cap
joint is avoided by the pilecolumn combination, since there is no structural distinction between
the pile and the column other than the presence of a construction joint at the pilecolumn inter
face.
While the design of piled footings favors forcing plastic hinging into the superstructure, with
the piles remaining elastic, pilecolumns can be designed with overall ductile performance in
mind. In case of a single pilecolumn, formation of a plastic hinge in the pile shaft is the only
mechanism by which ductile performance can be attained. A pilecolumn bent may first tend to
plastify at the columnbeam joint, but the full flexural capacity of the system can only be ob
tained through the formation of a secondary plastic hinge, below ground surface (at least
slightly below). Bending moment distribution varies with height, but diminishes after attaining a
maximum bending moment below the ground level. A typical depth for maximum bending
moment, and possibly the location of the plastic hinge, ranges from one to three or four pile di
ameters below ground surface, depending on the aboveground height and soil stiffness.
Damage below ground in the form of cracking or spalling of concrete, fracture of transverse re
inforcement, or buckling of longitudinal reinforcement is generally difficult to assess after an
earthquake. This, coupled with the potential high cost of repair, resulted in the current use of a
design displacement ductility factor that is smaller than that of columns in order to limit the
amount of yielding in the pile below the ground level. For example, in the United States, ATC
32 (ATC 1996) prescribes a displacement ductility factor of 3 for pilecolumns compared to a
displacement ductility factor of 4 for wellconfined fixedbase reinforced concrete columns. A
similar approach of prescribing higher lateral strength for piles has been adopted for seismic de
sign of highway bridges in New Zealand. For plastic hinging that may develop at a depth less
than 2 m below the ground level, but not below the mean water level, the design displacement
ductility factor is limited to no more than 4. For plastic hinging at a depth greater than 2 m be
low the ground level or below the mean water level, the design displacement ductility factor is
reduced to no more than 3 (Chapman 1995, Park 1998).
In this paper, a parametric investigation of the nonlinear inelastic response of pilecolumn
bridge systems is conducted, and the influence of pile inelastic behavior and soilstructure in
teraction on structure ductility demand is identified. The role of various key parameters are ex
amined, such as: (a) soil compliance, (b) aboveground height of the column shaft, (c) pile di
ameter, (d) intensity of the input seismic motion, and (e) location of the plastic hinge, on
characteristic performance measures of the soilstructure system response, such as: the dis
placement (global),
=
(3)
where E
p
I
p
is the initial (elastic) bending stiffness (also called flexural rigidity),
p
is a parame
ter controlling the post yielding bending stiffness, My is the value of bending moment that initi
ates structural yielding in the pile, and
p
is the hysteretic dimensionless parameter which con
trols the nonlinear structural response of the pile. The latter is governed by
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
54
( )
{ }
0
1
p
n
p p p
p p p p
d h
b g sign d
d
= +
(4)
where is the pile curvature, and b
p
, g
p
, n
p
,
p
, r
p
, and h
p
, are dimensionless quantities that con
trol the shape of the hysteretic bending momentcurvature loop in the same manner as n
s
, b
s
, g
s
,
p
, r
p
, and h
p
, control the shape of the lateral soil reactiondeflection loop.
0
is the value of pile
curvature at initiation of yielding in the pile.
Evidently, Equations 3 4 are of the same form as Equations 1 2, except that no viscous term
(radiation damping) is included in the structural pile response.
The seismic response of the soilpilestructure system is investigated herein via a beamon
nonlinearWinklerfoundation (BNWF) finite element model developed in OpenSees (Fig. 3).
The pilecolumn is discretized into nonlinear beam elements with length 0.5 to 1.0 m, whose
bending behavior is governed by the macroscopic constitutive BWGG model. The mass of the
deck is simulated as a concentrated mass at the top node of the pilecolumn, whereas the dis
tributed mass of the extended pile is simulated by lumped masses on beamelement nodes.
The nearfield soilpile interface is simulated with nonlinear py spring elements, the behavior
of which is described also by the BWGG model. Model parameters were appropriately cali
brated to match the py curves of Reese et al. (1974) and Matlock (1970). The free extremities
of the soil springs were excited by the acceleration time histories obtained at each depth from
the freefield seismic response analysis (Banerjee et al. 1987).
Although the developed finite element model has the capability to reproduce higher order phe
nomena (e.g. P effects), such phenomena were ignored, considering that their strong depend
ence on the mass of the structure and the geometry would obscure the role of other parameters
(e.g. structural inelasticity and soil compliance).
Figure 3. Schematic illustration of the model used for the analyses
2.3 Soil parameters
The influence of nearfield soil compliance on the seismic response of the soilpilestructure
system is investigated parametrically considering four different homogeneous soil profiles (Fig.
1): (a) sand with friction angle = 30
o
, (b) sand with friction angle = 40
o
, (c) clay with
undrained shear strength S
u
= 40 kPa, and (d) clay with undrained shear strength S
u
= 200 kPa.
The smallamplitude stiffness k (= p
y
/ y
0
) was obtained from the available beamondynamic
WinklerFoundation solutions (e.g. Gazetas & Dobry 1984, Makris & Gazetas 1992) in terms of
the Youngs modulus of the soil.
For piles in cohesive soils the ultimate soil reaction per unit length of pile can be approximated
by the well known expression
d S P
u y 1
=
(5)
where S
u
is the soil undrained shear strength, and
1
varies from 9 to 12, depending on the fric
tion ratio f
s
/ S
u
at the pilesoil interface. A value of
1
= 9 is often used for a soft clay, while
1
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
55
= 11 is more appropriate for a stiff clay. At shallow depths, the plane strain assumption of Equa
tion 8 is inappropriate because of the nonzero vertical deformation of the soil during lateral
motion of the pile. The following formulation has been proposed for P
y
near the surface
(Matlock 1970)
( )
J
S
d
d
z d S
d
z
J
S
P
u
s
u
u
v
y
+
<
+ + =
,
2 1
,
2
,
(6)
where
'
v
is the vertical effective stress, and
s
the effective specific weight of the soil, and
2
and J are a dimensionless quantities. Broms (1964) proposed a value of
2
= 2, whereas Matlock
(1970) used
2
=3. Matlock (1970) stated that the value of J was determined experimentally to
be 0.5 for a soft clay and about 0.25 for a medium clay, whereas Reese (1975) suggested a value
of J = 2.83 for every type of clay. For piles embedded in cohesionless soils, Broms (1964) pro
posed an analytical expression for the ultimate soil reaction :
z d P
s y
+ =
2
45 tan 3
2 ,
(7)
where is the angle of friction. Equation 7 is very often preferred in practice among other more
rigorous expressions for its simplicity and compatibility with experimental results.
For the description of the nonlinear behavior of the nearfield soil the wellknown py relations
of Reese et al. (1974) and Matlock (1970) are used for sand and clay, respectively.
2.4 Soil Profiles, Seismic Excitations, and Site Response Analysis
The influence of soil amplification on the seismic response of the soilpilestructure system is
not examined, mainly for two reasons: (a) a thorough investigation of seismic ground response
is out of scope of this paper, and (b) the unavoidable differences in freefield motions from the
soil response analysis of the four different soil profiles, would complicate the comprehension of
the related phenomena. Therefore, a single soil profile was selected for ground response analy
sis (Fig. 4): a category C profile, according to NEHRP (1994). Bedrock was assumed to be at 50
m depth.
Figure 4. Shear wave velocity distribution of the adopted soil deposit used for the wave propagation
analysis
The influence of shaking on the seismic response is investigated by selecting three real accelera
tion records as seismic excitations:
the record from Aegion earthquake (1995),
the record from Lefkada earthquake (2003), and
the JMA record from Kobe earthquake (1995).
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
56
The first two records are representative strong motions of the seismic environment of Greece,
with one and many cycles, respectively. JMA record is used to investigate the dynamic response
of the soilpilestructure system to a quite unfavorable incident. The dominant periods of the
acceleration time histories for the aforementioned three earthquake records range from 0.2 to
0.8 s, resulting in a fixed base fundamental period ratio (designated as the fixed base fundamen
tal period of the superstructure divided by the predominant period of the freefield surface ac
celeration time history) which ranges from 0.66 to 2.67. This is a wide range of values which
ensures generalization of the results presented herein. Nearfault effects such as rupture
directivity and fling (Gerolymos et al. 2007) are also captured by the utilized accelerograms.
All the records were first scaled to a PGA of 0.5 g and 0.8 g at the ground surface; then through
deconvolution analyses conducted with SHAKE (Schnabel et al. 1972), the bedrock motion as
well as the motion at various depths along the pile, were estimated. The ground motion profiles
obtained from SHAKE analyses are then used as input motion in the developed BNWF model.
The acceleration time histories at the surface and the corresponding elastic response spectra
scaled to a SA = 0.8 g (T = 0 s) for 5 % damping, are presented in Figure 5.
It should be stated here in that from a seismological point of view, simply scaling an accelera
tion time history to a large PGA value for representing the severity of an earthquake might not
be always correct. It is well known from the literature that high peak ground accelerations are
usually accompanied by a large number of predominant cycles. Obviously, this is not the case
for Aegion record which can be satisfactorily approximated by a single sinusoidal pulse.
Figure 5. Real acceleration time histories used as seismic excitation, after scaling to a peak ground accel
eration of a
g
= 0.5 and 0.8 g, and corresponding ( = 5 %) response spectra scaled to S
a
(T = 0 s) = 0.8 g.
2.5 Analysis Methodology and Performance Measure Parameters
Besides the fundamental response amounts (acceleration, displacement, moments, etc.) that de
scribe the behavior of a structure under dynamic loading, other important seismic performance
measures are the local and global ductility demand
and
max
=
(8)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
57
For bridge structures supported on extended piles, the local ductility demand imposed on the
pile shaft might govern the design of the system, because damage to the pile (such as spalling of
cover concrete, crack widths, potential for buckling or fracture of longitudinal reinforcement) is
related to the local curvature ductility.
The following procedure is followed for the assessment of local curvature ductility demand in
the analyses conducted. The momentcurvature curve of each pilecolumn crosssection is ap
proximated by a bilinear elasticperfectly plastic relation, in which the first (linear) section is
defined as the secant stiffness through the firstyield point
f y
(yielding of first longitudinal rein
forcement bar) and the second section by the tangent line on the postyielding section of the ac
tual momentcurvature curve. The intersection of these two lines defines the crosssection yield
curvature
y
(Fig. 6).
Figure 6. Definition of yield curvature of the soilpilestructure system
Similarly, the global (displacement) ductility demand
(9)
The yield displacement u
y
is assessed through static nonlinear analyses (pushover analyses) ac
cording to the following procedure:
At the center of mass of the superstructure, a horizontal force is gradually applied. The maxi
mum displacement and the curvature along the pilecolumn are continuously monitored. The
displacement measured, when the pile curvature reaches the firstyield point
f y
, is defined as
the firstyield displacement u
f y
. Then, similarly to the procedure followed for the determination
of yield curvature, the loaddisplacement curve is approximated by an equivalent bilinear elas
ticperfectly plastic curve, in which the first (linear) section is defined as the secant stiffness
through the firstyield point u
f y
and the second section by the tangent line on the postyielding
section of the loaddisplacement curve. The intersection of these two lines defines the yield dis
placement u
y
(Fig. 7).
Figure 7. Definition of yield displacement of the soilpilestructure system
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
58
It has to be noticed, that for the estimation of pile curvature, we did not use the FEM original
curvature results as these showed mesh sensitivity. Instead, plastic rotation results which are
mesh insensitive, were used and divided by the plastic hinge length L
p
to derive pile curvature.
The length of plastic hinge L
p
for the pilecolumns was estimated according to Budek et al.
(2000) approximation:
H d L
p
+ = 06 . 0
(10)
where d is the pile diameter and H the aboveground height. Similar expressions, based however
on different assumptions, have also been provided in Caltrans (1986,1990), Dowrick (1987),
Priestley et al. (1996), Chai (2002), and Chai & Hutchinson (2002).
The drift ratio is defined as the maximum displacement of the deck imposed by an earthquake
relative to pier base displacement divided by the height of the pier:
H
u u
base pier deck
=
max max
(11)
3 ANALYSIS RESULTS AND DISCUSSION
In this paragraph, typical results of the nonlinear analyses are presented in terms of acceleration
timehistories; peak bending moment, curvature and displacement distributions.
Results of the seismic response of the examined structural systems are presented in Figures 8
26. The acceleration time histories calculated for a pilecolumn of diameter d = 3 m and height
H = 5 m are presented in Figure 8a for every soil profile examined. The response of the deck is
quite smaller in case of soft clay. On the contrary, stiff foundation soils lead to increased re
sponse of deck. An exception to this is the increased structural (deck) response for the case of
loose sand. A possible explanation is that the lateral confinement providing by the soil might be
considerable even for small values of internal friction angles of the soil, thus stiffening the re
sponse of the pile.
Figure 8. Acceleration time histories of the deck for freehead pilecolumns with aboveground height H
= 5 m. Left diagram: pile with diameter b = d = 3.0 m. Right diagram: pile founded in soft clay (excitation
at ground surface: JMA, Kobe 1995 a
g
= 0.5 g)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
59
Similarly, differences are observed in Figure 8b, where the acceleration time histories of a 5 m
high pier embedded in soft clay are illustrated for different pile diameters. Smaller pilecolumn
diameter leads to higher deck acceleration. The small pilecolumn diameters (d = 1.5, 2.0 m)
correspond to a deck mass which is 16 times smaller than that for the larger pilecolumn diame
ters (d = 3.0, 4.0 m). This substantial difference in deck masses is responsible for an also large
discrepancy in the fundamental natural periods of the pierfoundationsoil system (effective pe
riods). Indeed, as can be hardly seen in the response acceleration time histories after the input
motion has subsided, at t = 15 s (free response), the effective period of the smalldiameter
bridge columns is approximately 1.0 s while that of the largediameter bridge columns is about
2.0 s. It is therefore shown that the effective period of the pier systems increases with increasing
mass of deck, despite that the fixedbase period is held constant. A similar trend in the response
of bridgepiers supported either by a single pile or by a group of piles has been also shown in
Gerolymos et al. (1998).
Furthermore, a small decrease in peak acceleration values is observed in constantdiameter ex
tended pile (d = 1.5, 3.0 m) compared to the response of variablediameter systems (d = 2.0, 4.0
m). The response of the constantdiameter systems is associated with more intense pile and soil
inelasticity compared to the response of variablediameter systems. The ample mobilization of
soil plastification and structural yielding mechanisms in the case of constantdiameter piers re
sults in an increase of both hysteretic and radiation damping, which in turn slows down the deck
response.
The distributions of maximum displacements with depth for several cases are presented in Fig
ure 9 for the JMA record as seismic excitation at ground surface. The influence of soil in this
distribution is depicted in Figure 9a: the softer the soil, the larger the maximum displacement
within the soil. The increased compliance of the soft soils implies significant deflection of the
pile. In Figure 9b, the role of pile diameter on the system response is illustrated. Larger pile
diameter piers (d = 3.0 and 4.0 m) are associated with larger effective periods compared to those
of smaller pilediameter (d = 1.5 and 2.0 m). Given the response spectra of the JMA accelera
tion time history, this means that the larger pilediameter systems exhibit larger displacements
and smaller response accelerations compared to the piers with smaller pile diameters (see also
Fig. 8).
Figure 9. Maximum displacement distributions for pilecolumns with aboveground height H = 10 m. Left
diagram: pile with diameter b = d = 3.0 m. Right diagram: pile founded in soft clay (excitation at ground
surface: JMA, Kobe 1995 a
g
= 0.5 g)
The distributions of maximum bending moments with depth for several cases are presented in
Figures 1011. The influence of soil type in this distribution is depicted in Figure 10. Like in
displacement distributions, soft soils result to increased pile effective lengths. With a maximum
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
60
moment of 1200 kNm (yield moment), the depth where this value becomes half increases from
4 m, in case of stiff clay and dense sand, to 6 m for loose sand, and to 10 m in case of soft clay.
It is interesting to observe in Figure 12 that the maximum bending moments among the four
cases are almost equal, indirectly revealing the formation of plastic hinges at certain locations of
the piles. Given the severity of the ground motion (the JMA record scaled to a peak ground ac
celeration of 0.8g), mobilization of the full bending moment resistance of the piles is unavoid
able irrespectively of the type of foundation soil.
As shown in Figure 11, the increase of pile diameter results to shifting of the maximummoment
depth to greater depths. It is noticed, that the maximummoment depth does not always coincide
with plastic hinge position, due to difference in pile and pier diameters.
Figure 10. Maximum bending moment distributions for freehead pilecolumns with aboveground height
H = 10 m and diameters b = d = 1.5 m (excitation at ground surface: JMA, Kobe 1995 a
g
= 0.8 g)
Figure 11. Maximum bending moments distributions for freehead pilecolumns with aboveground
height H = 10 m founded in stiff clay (excitation at ground surface: JMA, Kobe 1995 a
g
= 0.8 g)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
61
The position of plastic hinge is easily assessed via distributions of pile curvature with depth. In
Figure 12a, the distributions of curvature are presented for a 3m diameter pilecolumn with
aboveground height H = 10 m, embedded in different soils. It is observed that the plastic hinge
is developed within 2 diameters below ground surface. The amplitude of curvature increases
significantly in stiffer soils. The effect of pile diameter is shown in Figure 12b. In case of pile
columns with pier of smaller diameter than that of the pile (d = 2.0, 4.0 m), the pier is highly
stressed and the plastic hinge is formed at its base. Below the ground surface, curvature values
decrease rapidly. On the other hand, constantdiameter pilecolumns (d = 1.5, 3.0 m) may de
velop plastic hinge below surface. In every case, however, plastic rotations are distributed in
greater length with consequent decrease of the maximum curvature.
Figure 12. Maximum curvature distributions for freehead pilecolumns with aboveground height H = 10
m. Left diagram: pile with diameter d = 3.0 m (excitation at ground surface: JMA, Kobe 1995 a
g
= 0.5
g). Right diagram: pile founded in soft clay (excitation at ground surface: JMA, Kobe 1995 a
g
= 0.8 g)
In Figure 13, the correlation of the local curvature ductility demand to the global displacement
ductility demand is presented. All the analyses resulted to nonlinear behavior of the extended
pile shaft (
> 1) are depicted categorized according to the foundation soil. The mean ratio (
1) / (
1) equals to 5.4 for soft clay, 3.4 for loose sand, 2.6 for dense sand, and 2.7 for stiff
clay. Similar results have been also obtained by Hutchinson et al. (2004). At first sight, it seems
that founding pilecolumns in soft soils is unfavorable: for a given earthquake imposed global
displacement ductility, the local curvature ductility demand is higher than the one corresponds
to stiffer soils. This impression, as will be revealed later on, may be deceptive.
Figure 13. Correlation of local and global ductility demands for different soil types
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
62
A similar trend appears in Figure 14 where analyses results have been categorized according to
the potential location of plastic hinge. For constantdiameter pilecolumns the plastic hinge is
likely developed below the ground surface (on pile) whereas for variablediameter pile
columns, plastic hinges are developed at the base of pier. The average ratio (
1) / (
1)
takes a value of 3.5 for plastic hinge on the pile, and 2.7 for plastic hinge on the pier. The results
discourage the inelastic design of pile, however, the picture is yet to be cleared.
In the next figures (Figs. 15 16), analyses results have been grouped according to pier diame
ter and shaking intensity, in respect. A slight predominance of the larger pier (d = 3.0 m) is ob
served as the average value of (
1) / (
/
max
ratio (Figs. 20 and 21). The influence of aboveground height H on the correlation between local
ductility demand and maximum drift ratio is illustrated in Figure 22. Taller piers exhibit greater
local curvature ductility demand for a given drift ratio. Indeed, for a given drift ratio the differ
ential horizontal displacement between deck and pier base decreases with decreasing above
ground height of the pier, leading to smaller pier distress and thus to smaller ductility demand.
Figure 18. Correlation of local ductility demand and maximum drift ratio for different soil types
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
64
Figure 19. Correlation of local ductility demand and maximum drift ratio for different plastic hinge loca
tions
Figure 20. Correlation of local ductility demand and maximum drift ratio for different diameters
Figure 21. Correlation of local ductility demand and maximum drift ratio for different seismic motion
amplitudes
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
65
Figure 22. Correlation of local ductility demand and maximum drift ratio for different aboveground
heights
In Figures 2326, the mean and peak values of the factors
, and
max
are illustrated for vari
ous parameters examined. It is clearly observed that the mean and maximum values of both
and
factors are lower for soft soils and plasticized piles. This phenomenon discredits the trend
appeared in Figures 13 14 and reveals the beneficial influence of soil compliance and pile ine
lasticity on the response of the structure examined. The apparent paradox stems from the fact
that kinematic expressions do not distinguish between capacity and demand, as also stated in
Mylonakis et al. (2000). For example, according to Figure 13, for a given displacement ductility
demand the curvature ductility capacity of a pilecolumn embedded in soft soil needs to be lar
ger than that of a pilecolumn embedded in stiff soil. However, this does not mean that for a
given seismic excitation both pilecolumns would exhibit the same displacement ductility.
Figure 23. Variation of local curvature ductility (
1) / (
1) may take higher values for soft soils, the absolute values of
max
, whereas mean
value slightly increases, if not remain constant.
4 CONJECTURES
From the analysis of the results of the exploratory parametric analyses conducted herein, the fol
lowing conclusions could be drawn:
For a given global (displacement) ductility demand
(Mu),
the local (curvature) ductility demand
(M).
the curvature ductility demand slightly decreases with increasing pile diameter.
the curvature ductility demand increases in case of columnpiles with smaller aboveground
height ratios (d / H).
The opposite trends for the local ductility demand
1) / (
+ + =
e
c
c c c eq
E
K
M
K M M
K
4
1
1
2
2
2
2
(
y e c
E E E < < )(2)
+ + + =
e
c y
c
y y y eq
E
K
M M
K
M
K M M
K
4
1
2
2 2
1
2
3
2
3
(
u e y
E E E < < )(3)
Figure 1. Equal energy assumption in a bending moment curvature relation in the case of trilinear model
M
(
e
,M
e
)
(
p
,M
p
)
K
1
K
2
K
3
M
u
u
M
y
(
eq
,M
eq
)
M
(
e
,M
e
)
(
p
,M
p
)
K
1
K
2
K
3
M
u
u
M
y
(
eq
,M
eq
)
Curvature,
M
o
m
e
n
t
,
M
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
72
where
2 /
e e e
M E = ,
2 /
c c c
M E = ,
2 / ) )( (
c y c y c y
M M E E + + = , and
2 / ) )( (
y u y u y u
M M E E + + =
3 METHOD AND MODEL OF ANALYSIS
3.1 Description of object for study
The object of this study is an actual bridge pier recently constructed at a soft soil site shown in
Figure 2a. The pier is Tshaped and its column has a rectangular section having 3.0 m width and
3.5 m length. The mass of the pier is 336 ton and the design horizontally inertial force according
to the JRA (Japan Road Association) design specification (JRA,1996) is 11.76MN (1200 tf).
The footing of the pier was constructed 4 m below the ground surface and is supported by 12
castinplace reinforced concrete endbearing piles of lengths of 30 m. The pile arrangement is
in a form of 4 rows of 3 piles, and Figure 2a shows the elevation view in a longitudinal direc
tion. The piles were placed through a 10 m thick soft clayey soil layer characterized by the av
eraged SPT Nvalue of 2 underlain by a 14 m thick mediumhard clayey soil layer characterized
by that of 8 in the bearing stratum of sandy soil with SPT Nvalues of more than 50 with 2 m
embedded lengths. Therefore, stress concentrations along the depth of the piles raised by seis
mic soil displacements are anticipated to be generated at or around boundaries of certain two
soil layers in between which ratios of rigidity show high contrast. Such boundaries would be lo
cated at relative depths of 10 m and 24 to 28 m to pile heads. Additionally, the depth of pile
heads is also another point of stress concentration due mainly to horizontal inertial forces from
the superstructure, the pier and the footing and to the seismic soil displacements.
Figure 2. Object of study and its schematic model of analysis
Fill
N=4
Clayey soil
N=2
Clayey soil
N=8
Sandy soil
N=16
Sandy soil N=50
10 m
3.5 m
RC piles
Bridge pier
Fill
N=4
Clayey soil
N=2
Clayey soil
N=8
Sandy soil
N=16
Sandy soil N=50
10 m
3.5 m
RC piles
Bridge pier
1
2
.
5
m
3
0
m
3
0
m
1
0
m
4
m
2
m
4
mFill
N=4
Clayey soil
N=2
Clayey soil
N=8
Sandy soil
N=16
Sandy soil N=50
10 m
3.5 m
RC piles
Bridge pier
Fill
N=4
Clayey soil
N=2
Clayey soil
N=8
Sandy soil
N=16
Sandy soil N=50
10 m
3.5 m
RC piles
Bridge pier
1
2
.
5
m
3
0
m
3
0
m
1
0
m
4
m
2
m
4
m
Pilestructure system
Free field soil
system
Pairs of spring and dashpot
for interaction
Superstructure mass
Input earthquake motion
Pilestructure system
Free field soil
system
Pairs of spring and dashpot
for interaction
Superstructure mass
Input earthquake motion
(a) Soilpilestructure for study (b) Schematic model of analysis
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
73
3.2 Method of analysis
The earthquake response analyses in this study is carried out in time domain adopting a lumped
mass and nonlinear element model for a structurepilesoil coupled system proposed by Penzien
et al. (1964). The schematic diagram of the model shown in Figure 2b illustrates that the model
consists of both a structurepile system and a soil system representing a free field, and that the
springs and dashpots between the two systems transmit the displacements of the free field to the
structurepile system. Such springs and dashpots, therefore, is hereafter called interaction
springs. The method proposed by Penzien et al. has two features for considering a pilesoil in
teraction; firstly, a set of vertically distributed interaction springs realizes a spatial or kine
matic interaction effect, and secondly, newly introduced effective masses realize an inertial in
teraction effect and a dissipation damping effect at the same time. The effective mass at a node
corresponding to a unit thickness of a soil layer is evaluated to be equivalent with regard to an
integral of kinematic energy over soil elements surrounding a pile in the soil layer. The inertial
interaction is incorporated into the equation of motion as a product of the effective mass and the
relative displacement of a pile element to a soil element in the free field. It is here necessary to
be noticed that the rigidity of a vertical column of the effective masses is not taken into account.
(Mori, 2000)
Eventually in this study, earthquake response analyses are carried out by two kinds of mod
els; one model has a nonlinear pile, and the other has an elastic linear pile. In the two models,
shear springs connecting lumped masses in the column of the soil system are fully nonlinear, but
the interaction springs are elastic linear. The condition that the interaction springs are elastic is
expected to be conservative for evaluating the response of pile moment and curvature in deeper
portions according to the previous research results obtained by Mori & Hirata (2002b).
3.3 Models of analyses
The structurepile system is modeled to have 43 nodes with lumped masses that are basically ar
ranged in an equal interval of 1 m. The soil system is also modeled to be arranged in an equal
interval of 1 m, accordingly it has 34 nodes with lumped masses. The boundary conditions of
the model are as follows; the end of the deepest pile element is pined to the base for the model,
and a rotation spring, which represents the resistance of a spatiallydistributedpile group, is
fixed to the base. The rotation spring has a constant calculated by only considering axial rigidity
of piles and neglecting resistance of surrounding soils. Further, it is modeled as an elastic spring
fixed to the base, because that the axial force in the most outer pile is expected not to reach its
ultimate bearing capacity. The interaction springs are evaluated according to Sugimura (1972).
The input earthquake motions for the model are defined as a total motion in the base. The
damping for the model is numerically Rayleigh damping, whose parameters and are deter
mined for making damping ratios at the 1st and 2nd eigenfrequencies of the soil system be 2 %.
3.4 Constitutive model of soils and its parameters
The nonlinear model of shear springs for soil elements in the column for the free field adopts
the model proposed by RambergOsgood (abbreviated as RO model.) Determination of pa
rameters of the RO model is basically using soil properties adopted for the design of the bridge
is shown in Table 1. Shear wave velocity, Vs for each layer is determined by Equations 4 and
5 empirically based on SPT Nvalue, N according to JRA design specification (JRA, 1996). Ri
gidity of soil at a strain of 10
4
%, G
0
is determined by the Equation 6 using density of soil,
theoretically.
3 / 1
100N V
s
= for finegrained soils (4)
3 / 1
80N V
s
= for sandy soils (5)
2
0 s
V G = (6)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
74
Table 1. Soil properties for design and this study
Rigidity of soil, G and damping constant, h are dependant on a shear strain in soil element, .
For specifying such dependency as a target, Equations 7 and 8 proposed by Imazu & Fukutake
(1986) through compilation of related literatures. These equations can be applied up to a strain
of 0.5 %. Parameters a, b, c, and d in these equations are provided in cases of sand, clay, and
gravel for the average.
( )
b
a G G + = 1 1
0
(7)
d
c h = (8)
In order to fit the target dependency curves, G/G
0
 and h, the two parameters of RO
model. One of the two parameters is reference strain,
0.5
, which is strain of an element when the
rigidity of the element decreased to half of rigidity at a strain of 10
4
%, and the other is damp
ing ratio when the strain of element is infinity, h
max
. Eventually, the parameters adopted in this
study are shown in Table 2, and G/G
0
degradation curves of the model and target for sands, as
an example, is shown in Figure 3. Using the two parameters, the loading curve, or skeleton
curve in the plane of shear stress, and shear strain is expressed in Equation 9, where sub
parameters and (These and are along the same notations as the two parameters for
Rayleigh damping, but these have different meanings and different values.) shown in Equations
10 and 11.
( )
+ = 1
0
G
(9)
max
max
2
2
h
h
= (10)
=
0 5 . 0
2
G
(11)
Table 2. Parameters of RambergOsgood model for analysis
No. Soil type
Thickness
of layer
Averaged
SPT N
Unit mass Cohesion
Shear wave
velocity
Rigidity
(m) (ton/m
3
) (kN/m
2
) (m/s) (kN/m
2
)
1 Fill 4 4 1.8 0 127 29028
2 Clayey soil 10 2 1.7 19.6 126 26989
3 Clayey soil 14 8 1.8 78.4 200 72001
4 Sandy soil 4 16 1.9 0 202 77528
5 Sandy soil >= 2 50 2.0 0 295 174048
Note
Sandy soil layer of No.5 was adopted as the bearing stratum for the design.
Reference strain
*1
Maximum damping ratio
*2
0.5
(%) h
max
(%)
Sand 0.0607 24.9
Clay 0.1406 14.5
Gravel 0.0348 14.4
Soil type
*1 Strain of an element when the rigidity of the element
*2 Damping ratio when the strain of element is infinity
decreased to half of that at a strain of 10
4
%
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
75
Figure 3. G/G0 degradation curves of the model and target for sands
3.5 Constitutive model of RC piles
Properties of the RC pile are shown in Table 3. A trilinear model is adopted as a nonlinear
model for the moment and curvature relationship. The trilinear model adopted in this study fol
lows Takeda model defining a loading path along a curve consisting of three line segments with
two bents, an unloading path directing a past maximum response point, and a reloading along
the unloading path toward the unloading point. Characteristic points of the trilinear model for
RC piles are shown in Table 4. Using the points in this table, the skeleton curves of the RC piles
are shown in Figure 4. On the other hand, an elastic linear model with an initial flexural stiff
ness is adopted in the model to be applied EEA to.
Table 3. Properties of pile
Table 4. Characteristic points of trilinear model for RC piles
Depth (m)
c
M
c
y
M
y
u
M
u
Condition
(m
1
)
(kNm)
(m
1
)
(kNm)
(m
1
)
(kNm)
Cracking 2.3*10
4
675 2.3*10
4
675 2.3*10
4
648
Yielding 2.4*10
3
2199 2.4*10
3
2225 2.2*10
3
1556
Ultimate 2.1*10
2
3190 1.2*10
2
3190 1.4*10
2
2105
0  2.4 m 2.4  10 m 10  30 m
0
0.2
0.4
0.6
0.8
1
1.2
0.0001 0.001 0.01 0.1 1
)
G
/
G
0
RO
ImazuFukutake (Target)
RambergOsgood model
Shear strain, (%)
0
0.2
0.4
0.6
0.8
1
1.2
0.0001 0.001 0.01 0.1 1
)
G
/
G
0
RO
ImazuFukutake (Target)
RambergOsgood model
Shear strain, (%)
Item
Type
Diameter
Length
Strength of concrete
Young's modulous
Specification of Rebar
Depth from pile head major bar stirrup
0 to 2.4 m 28D29 D16 ctc150
2.4 to 10 m 28D29 D16 ctc300
10 to 30 m 14D29 D16 ctc300
24 N/mm
2
2.5*104 N/mm
2
SD295A
Specification
Castinplace reinforced concrete pile
1.2 m
30.0 m
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
76
Figure 4. Skeleton curves of the RC piles by trilinear model
3.6 Input earthquake motions
Two types of design earthquake motion waves are defined by and provided by JRA specifica
tions. Type I is classified as a typical strong ground motion recorded at site a large distance
away during a large earthquake occurred in a subduction zone. On the other hand, Type II is
classified to a typical severely strong ground motions recorded at a site within a small distance
during a moderatesized earthquake occurred in an inland seismic region. Input earthquake mo
tion waves adopted in this study are show in Table 5. KPI wave, which is the NS component of
accelerogram of earthquake motion recorded at a depth of 83 m of Kobe Port Island (KPI) site
during HyogokenNambu earthquake, is shown in Figure 5 as an example.
Table 5. Input earthquake motions
Figure 5. Acceleration time history of KPI wave for an input earthquake motion
0
500
1000
1500
2000
2500
3000
3500
0 0.005 0.01 0.015 0.02 0.025
Curvature (m
1
)
M
o
m
e
n
t
(
k
N
m
)
0  2.4 m
2.4  10 m
10  30 m
Depth (m)
Cracking
Ultimate
Yielding
No. Type Earthquake motion wave Earthquake PGA (cm/s
2
)
1 I Kaihoku Bridge (Longi.) 1978 MiyagikenOki 319
2 I Itajima Bridge (Longt.) 1968 TokachiOki 363
3 II Kobe Port Island (NS) 1995 HyogokenNambu 679
4 II JR Takatori Station (NS) 1995 HyogokenNambu 687
5 II Osaka Gas Fukiai (N27W) 1995 HyogokenNambu 736
6 II JMA Kobe Obs. (NS) 1995 HyogokenNambu 812
Note
Type I: Typical strong ground motion recorded at site a large distance away during a large
earthquake occurred in a subduction zone
Type II: Typical strong ground motions recorded at a site within a small distance during a
moderatesized earthquake occurred in an inland seismic region
Types of design earthquake motion waves are defined by and provided by JRA.
800
400
0
400
800
0 5 10 15 20 25 30
Time (s)
A
c
c
.
(
c
m
/
s
2
)
KPI83 m, NS
Abs.Max 679cm/s
2
, 8.06
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
77
4 RESULTS AND DISCUSSIONS
4.1 Effects of nonlinearity of interaction springs
In this study, the interaction springs are, for simplicity, elastic linear in both models with the lin
ear pile and with the nonlinear pile. However, effects of nonlinearity of the interaction springs
on the response of pile need to be discussed, for an example, on the basis of the results of a pre
vious research by Mori & Hirata (2002b), prior to discussion on effectiveness of EEA for esti
mation of nonlinear pile curvature. In their research, the object and basic model of theirs are the
same of this research, but the difference is additional adoption of a nonlinear model of interac
tion springs, which used RO model adjusted to characteristics of interaction springs. Due to
space limitation, the detail of determination of parameters for this RO model is omitted here.
Input earthquake motion wave is only KPI wave, which has a severe intensity. Therefore, actual
inputs, however, are four differently scaled waves; 0.1, 0.2 and 0.5 times and no scaled or origi
nal for studying the progress of degree of nonlinear response.
The observations on difference of responses of reactions and deformations of the interaction
springs under different magnitudes of input motions in terms of comparison between the linear
and nonlinear interaction springs will be basis for coming discussion. As the magnitude of input
motion from 0.1 times to 1.0 times scale, nonlinear responses of the interaction springs appear
firstly at and around the pile head, secondly at depths of 28 m and 32 m corresponding to
boundaries of high contrast of impedance, and finally appear at a depth of 14 m corresponding
to boundaries between soft and hard clayey soil layers. This suggests that an inertial force from
superstructure is most influential to the pile head, and that soil displacements are more influen
tial to deep portions having higher contrast of impedance of soils at a boundary.
Distributions of maximum moment of pile in cases of differently scaled input motions are
shown in Figure 6a to Figure 6d for comparing linear elastic and nonlinear models of interaction
springs. We cannot find any obvious difference in the case of 0.1 times scaling in spite of ap
pearance of clearly nonlinear responses of the interaction springs and the piles at and around
pile head. Even in the case of 0.2 times scaling, differences are very slight except depths of 28
m and 32 m showing clear predominance of the linear spring case. Taking into account the re
sponse of the nonlinear spring case under cracking moment while the response of the linear case
go beyond it, we can understand a mechanism that nonlinear behavior of the interaction spring
reduce flexural responses of pile at deeper layer boundaries with relatively high impedance.
Quantifying the effect of such mechanism, a new index, nonlinear effect ratio (NER) with re
gard to interaction spring is defined as the ratio of response of nonlinear case to that of linear
case. Inverse of this index means degree of overestimating flexural response of pile by elastic
interaction springs.
Figure 6. Distributions of maximum moment of pile in cases of differently scaled input motions for com
paring linear elastic and nonlinear models of interaction springs
0
5
10
15
20
25
30
35
(
m
)
N
4
2
8
16
50
0
5
10
15
20
25
30
35
(
m
)
N
4
2
8
16
50
Interaction springs
Nonlinear
Linear
(c)
0.5
0 2000 4000
(kNm)
M
c
M
y
M
u
(d)
1.0
0 2000 4000
(kNm)
M
c
M
y
M
u
(b)
0.2
(kNm)
M
c
M
y
(a)
0.1
(kNm)
M
c
Moment (kNm) Moment (kNm) Moment (kNm) Moment (kNm)
(a) 0.1 times scaling (b) 0.2 times scaling (c) 0.5 times scaling (d) No scaling
Soil profile
and SPT N
value
0
D
e
p
t
h
(
m
)
0
5
10
15
20
25
30
35
(
m
)
N
4
2
8
16
50
0
5
10
15
20
25
30
35
(
m
)
N
4
2
8
16
50
Interaction springs
Nonlinear
Linear
(c)
0.5
0 2000 4000
(kNm)
M
c
M
y
M
u
(d)
1.0
0 2000 4000
(kNm)
M
c
M
y
M
u
(b)
0.2
(kNm)
M
c
M
y
(a)
0.1
(kNm)
M
c
Moment (kNm) Moment (kNm) Moment (kNm) Moment (kNm)
(a) 0.1 times scaling (b) 0.2 times scaling (c) 0.5 times scaling (d) No scaling
Soil profile
and SPT N
value
0
D
e
p
t
h
(
m
)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
78
In the case of 0.2 times scaled input, NERs of moment varies from 0.9 to 0.95 at most of pre
dominantly responding portions except a depth of 32 m showing a NER of 0.6. In the case of
0.5 times scaled input, NERs of moment varies from 0.85 to 1.03 at predominantly responding
portions except a depth of 32 m showing a NER of 0.45. This indicates that a model with linear
elastic interaction springs overestimate flexural responses of pile at deeper layer boundaries
with relatively high impedance, especially a boundary of the upper end of a bearing layer.
Furthermore, in the case of no scaled input, a NER of moment is 1.25 at pile head but it is
0.95, 0.75 and 0.3 at depths of 15, 28 and 32 m, respectively. Observing the tendency in change
of NER distributions, we can understand that the increase of NER at pile head is induced from
deepening a point of resultant reaction of the interaction springs. On the other hand, the de
crease of NERs at deeper potions can be understood that the deformation of nonlinear pile due
to larger shear strain of soil becomes more and more plastic and that the growth of moment of it
is reduced as well.
4.2 Influences of nonlinearity in flexural rigidity of pile and effects of EEA application
In the previous section, we understand about modeling of interaction springs that the linear elas
tic model tends to overestimate the moment of pile at deeper portions rather than the nonlinear
(or trilinear) model when the entire model has a nonlinear soil system and nonlinear pile ele
ments. This means linear elastic modeling of interaction springs leads us conservative or safer
judgment. The results to be discussed hereafter are according to the analyses described in Chap
ters 2 and 3 of this paper.
For discussing applicability of EEA to estimation of nonlinear flexural responses of a pile,
distributions of the responses are studied. The distributions of maximum moment of pile along
the depth in the case of Kaihoku wave input is shown in Figure 7a as an example. We can find
that there is much difference in the distributions between nonlinear and linear elastic models in
terms of magnitude and shape. Further discussions focus the depths of 4 m, 15 m, and 32 m. The
depth of 4 m corresponds to pile heads and that of 32 m to the upper end of the bearing stratum.
We should reconsider the tendencies that a lateral force on a pile head produces concentration of
moment there, and on the other hand, lateral soil displacements acting on the pile along the
depth produces concentrations of moment at boundaries with high contrast of impedance such
as the upper end of a bearing stratum. In addition, previously defined new index, nonlinearity
effect ratio (NER) is also applied with regard to a pile element as the ratio of response of
nonlinear case to that of linear case. On the basis of Figure 7a, NERs of moment at depths of 4
m, 15 m and 32 m are 3.6, 3.6 and 6.1, respectively. On the other hand, NERs of moment with
EEA application at the same depths are 0.60, 0.80 and 0.63, respectively. Therefore, EEA appli
cation remarkably can revise estimation of moment response because of NER becoming closer
to one.
(a) Moment for Kaihoku (b) Curvature for Kaihoku (c) Curvature for KPI
Figure 7. Distributions of maximum flexural response of pile in cases of Kaihoku & KPI waves
100 1000 10000 100000
(kNm)
Mc My Mu
0.0001 0.001 0.01 0.1
(m
1
)
u
0.0001 0.001 0.01 0.1
(m
1
)
u
0
5
10
15
20
25
30
35
(
m
)
N
4
2
8
16
50
0
5
10
15
20
25
30
35
(
m
)
N
4
2
8
16
50
Linear
Nonlinear
EEA estimation
Moment (kNm) Curvature (m
1
) Curvature (m
1
)
D
e
p
t
h
(
m
)
SPT N
100 1000 10000 100000
(kNm)
Mc My Mu
0.0001 0.001 0.01 0.1
(m
1
)
u
0.0001 0.001 0.01 0.1
(m
1
)
u
0
5
10
15
20
25
30
35
(
m
)
N
4
2
8
16
50
0
5
10
15
20
25
30
35
(
m
)
N
4
2
8
16
50
Linear
Nonlinear
EEA estimation
Moment (kNm) Curvature (m
1
) Curvature (m
1
)
D
e
p
t
h
(
m
)
SPT N
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
79
Hereafter, the curvature of pile, which is of importance in performancebased approach, is
discussed. The distributions of maximum curvature of pile along the depth in the case of Kai
hoku wave input is shown in Figure 7b. We find that there is also much difference in the distri
butions between nonlinear and linear elastic models in terms of magnitude and shape, while
EEA application revise a distribution shape of linear elastic response much better than moment.
NERs of curvature at depths of 4 m, 15 m and 32 m are 0.16, 0.39 and 0.11, respectively. On the
other hand, NERs of curvature with EEA application at the same depths are 0.56, 1.02 and 0.66,
respectively. Moreover the situation in the depth of 28 m is very similar to that in the depth of
32 m. Therefore, EEA application considerably can revise estimation of curvature response in
this case. In particular, it is understood that EEA is practically very effective to estimation of
plastic response of pile curvature in deeper portions. For confirming this understanding, let us
see other example. The distributions of maximum curvature of pile along the depth in the case
of KPI wave input is shown in Figure 7c. The shape of each models and the tendencies of
nonlinear effects that can be evaluated by NER are almost the same as described previously.
Thus, the understanding can be confirmed.
4.3 Influences of nonlinearity in flexural rigidity of pile and effects of EEA application
In the previous section, influences of nonlinearity in flexural rigidity of pile was discussed on
the basis of distribution of flexural responses of pile, and accordingly effects of EEA application
for estimation of nonlinear curvature of pile were confirmed, especially in its deeper portions.
Then, the applicability of EEA is discussed through comparison among predominant curvature
responses at three specific depths in cases of various types of input waves. Figure 8 shows
variation of NERs of curvature at depths of 4 m, 15 m and 32 m subjected to all six input mo
tions. For linear models of pile, NERs vary approximately from 0.1 to 0.4. On the other hand,
NERs vary approximately from 0.5 to 1.2 for EEA application. In particular, the average of
NERs in deeper portions is roughly 0.8. Considering the degree of the effect of overestimate by
using linear interaction springs, which was discussed in 4.1, I can conclude that EEA applica
tion to an entire model with linear elastic pile elements and linear elastic interaction springs is
practically very efficient for evaluating plastic curvature of pile subjected to severe earthquake
motions.
Figure 8. Variation of nonlinearity effect ratio (NER) of curvature at depths of 4 m, 15 m and 32 m sub
jected to six input motions
5 CONCLUSIONS
The applicability of equal energy assumption (EEA) to linear elastic pile response for estimation
of ductility factor response of a nonlinear pile was studied on the basis of fully nonlinear dy
namic response analyses on an actual bridge pier. It is concluded as follows;
0.01
0.1
1
10
/
4 m
15 m
32 m
/
4 m
15 m
32 m
1 2 3 4 5 6
EEA application/Nonlinear
4 m
15 m
32 m
Linear/Nonliner
4 m
15 m
32 m
Number of input motion
R
a
t
i
o
o
f
c
u
r
v
a
t
u
r
e
s
0.01
0.1
1
10
/
4 m
15 m
32 m
/
4 m
15 m
32 m
1 2 3 4 5 6
EEA application/Nonlinear
4 m
15 m
32 m
Linear/Nonliner
4 m
15 m
32 m
Number of input motion
R
a
t
i
o
o
f
c
u
r
v
a
t
u
r
e
s
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
80
(1) On the basis of fully nonlinear response analyses of soilpile structure system subjected to
four differently scaled input waves for studying the progress of degree of nonlinear behavior in
each element of the system, the nonlinear behavior of the interaction spring reduces flexural re
sponses of pile at deeper layer boundaries with relatively high contrast of soil impedance. De
fining a nonlinear effect ratio (NER) as the ratio of response of nonlinear case to that of linear
case with regard to interaction spring for quantifying the effect of such mechanism, it was found
that NERs at pile head and deeper potions become lesser than one as the input wave increases.
This means elastic modeling of interaction spring overestimates flexural response of pile rather
than nonlinear modeling of it.
(2) EEA application to linear elastic pile modeling considerably can revise the estimation of
nonlinear flexural responses of pile such as moment and especially curvature in terms of pre
dominant response in their distribution along the depth. In particular, it is understood that EEA
is practically very effective to estimation of plastic response of pile curvature in deeper portions.
NERs of curvature of pile subjected to all six input motions vary approximately from 0.1 to 0.4
for a linear elastic pile, on the other hand, vary approximately from 0.5 to 1.2 for EEA applica
tion. In particular, the average of NERs in deeper portions for EEA application is roughly 0.8.
(3) EEA application to an entire model with linear elastic pile elements and linear elastic inter
action springs is practically very efficient for evaluating plastic curvature of pile subjected to
severe earthquake motions.
REFERENCES
Imazu, M. & Fukutake, K. 1986. Dynamic deformation properties of gravels. Proc. 21st Ann. Conf. Jap.
Soc. Soil Mech. Found. Eng., JSSMFE: 509512. (in Japanese)
Japan Road Association (JRA) 1996. Design Specification of Highway Bridge, Part V Seismic Design:
Maruzen. (in Japanese)
Mori, S. 2000. Proposal of springmass model for pilefoundation structure and its application to really
damaged structures, Journal of Applied Mechanics, JSCE, 3: 609620. (in Japanese with English ab
stract)
Mori, S. & Hirata, A. 2002a. Evaluation of nonlinear seismic response of pile by constant energy rule.
Proc. 37th Ann. Conf. Jap. Geotech. Soc., JGS, 14951496. (in Japanese).
Mori, S. & Hirata, A. 2002b. Effects of nonlinearity of members on seismic response of pilefoundation
structure. Journal of Structural Engineering, JSCE 48A: 469478. (in Japanese with English abstracts)
Mori, S, Suga, K. & Akaishi, T. 2009. Evaluation chart of existing pile foundation against seismic soil
displacements. PerformanceBased Design in Earthquake Geotechnical Engineering, Proc. Intern.
Works. Kokusho, Tsukamoto & Yoshimine (eds) , London: Taylor & Francis Group: 667674.
Newmark, N.M. & Rosenblueth, E. 1971. Fundamentals of Earthquake Engineering, PrenticeHall Inc.
Penzien, J., Scheffey, C.F. & Parmelee, R.A. 1964. Seismic Analysis of Bridges on Long Piles, Journal
of the Eng. Mech. Div., Proc. of ASCE 90(EM3): 223254.
Sugimura Y. 1972. Research on Seismic Vibration Characteristics of Foundations Supported by Long
Piles in Soft Soils, Dissertation to Waseda University. (in Japanese)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
81
Sinkage of a Pile Foundation during the Niigataken Chuetsuoki
Earthquake in 2007
Y. Goto
Earthquake Research Institute, the University of Tokyo, Bunkyoku, Tokyo, Japan
1 INSTRUCTION
ABSTRACT: A new building supported by pile foundation in the Kashiwazaki sewage treat
ment plant sank by a hundred and several tens millimeter during the strong shaking (more than
X of EMS98) of the Niigataken Chuetsuoki Earthquake in 2007. The building was designed
according to the new seismic design code revised after the 1995 Kobe Earthquake but sank
more than the adjacent old building of mat foundation did. This paper reports an overview of the
damage, analyses the mechanism of the sinking and discusses some issues about the earthquake
proof design of foundations against high level of seismic load.
1.1 General overview of the earthquake
Kashiwazakicity, one of the middle scale cities on
the northerncentral shore of Japan main island, hav
ing ninety thousand population, was hit by a very
strong shaking more than X in EMS98 scale due to
the earthquake of July 16, 2007 having Magnitude
6.8. The world largest scale nuclear power plant is
located in Kashiwazakicity. Very many professionals
focused on the effect of so strong shaking to the nu
clear power plant and considerable lessons have been
extracted. However, there are some useful lessons ex
tracted from damages of ordinary structures.
Epicenter
Aftershock
Area
Nuclear
Power Plant
The author made damage reconnaissance of a sew
agetreatment plant of Kashiwazakicity focusing his
interest on a pile foundation. Sewagetreatment plant
is often located on a shore or a river side, where
ground condition is commonly poor. So it is unavoid
ably easy to suffer earthquake damages.
1.2 Kashiwazakicity sewagetreatment plant
The Kashiwazaki sewagetreatment plant (purifica
tion capacity is 40,700 kiloliter a day) is located on
the mouth of Sabaisi River also (Fig. 1). It is about
15 km southsouthwest of the epicenter.
Kashiwazaki
City Office
Railway
Station
Sewage
Treatment
Plant
2km
200km
Figure 1. Location of Epicenter and
Kashiwazakicity
(detached to Google
)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
82
Although there was not any seismometer in the plant, the shaking intensity is estimated to have
exceeded X (EMS98) from an interpolation of seismometer records around the site, namely,
JMA Intensity 6upper (about EMS98 X) at the city office of Kashiwazaki and Kashiwazaki
railway station and JMA Intensity 7 in the nuclear power plant site (maximum among the many
seismometers in that site).
Main building
Purification facilities
Digester tanks
Sabaisi River
50 m
Main sludgetreatment building
Sludge Inspecttion
Building
Dense sand dune
Gas facility
Figure 3. Total layout of Kashi
wazaki sewage treatment plant
Figure 2. Typical boring logs
Figure 2 is a typical borehole exploration data of the site. The surface layer is loose buried
sandy soil of 3 meters depth overlying on loose sand sediment layer of 4 meters. These layers
are followed by dense sand dune layer. The depth of bedrock of this area is supposed to be sev
eral hundred meters.
1.3 Damage overview of the sewagetreatment plant
Figure 3 shows total layout of the plant. Although some sections of the sewage line in Kashi
wazakicity were heavily damaged (mainly by liquefaction), influent wastewater to the plant
was increased 1.3 times because of inflow of underground water, and drift sand increased con
siderably.
The function to flow the sewage through the plant and to run it out was not damaged. Main fa
cilities for sewagetreatment were not damaged too, because they were series of massive rein
forced concrete pools based on improved sand layer, where sand compaction pile method was
used for the improvement. Power outage was only one minute and purification function of the
plant was recovered soon after the earthquake.
While, the sludge treatment function of the plant went down because of some damages on the
treatment facilities, such as digester tanks, sludge squeeze transportation line, and gas facilities
for heating sludge. These damages did not affect the purification function of sewage directly. A
temporary sludge transportation line was equipped within 5 days, and the sludge was stored un
digested, then dewatered and carried out from the plant.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
83
2 DAMAGE OF SLUDGE INSPECTION BUILDING
2.1 Profile of the Sludge Inspection Building
As mentioned above, main facilities for purification were not damaged, but a large vibrational
and residual displacement of an attached building, which had been designed according to the
new seismic design code revised after the 1995 Kobe Earthquake, made trouble. The building,
called as Sludge Inspection Building, had pile foundation and was constructed on a side of main
sludgetreatment building which had deep mat foundation based on dense sand layer. Figure 4 is
a side view of the main building and the Sludge Inspection Building. The latter is reinforced
concrete building having three stories and had been supported by fiftyone PHC piles (shear
strengthimproved type) as shown in Figure 5. The pile was eight meters in length and 700mm
in diameter and supported 900kN static load.
Sludge Inspection
Building
Main Sludge
Treatment Building
Figure 4. Side view of Main Sludge Treatment Building (left) and Sludge Inspection Building (right)
Future
extension
Figure 5. Profile of pile and plan of pile layout
Pile top
reinforcement
Footing
Concrete
filled in
4
5
0
0
4
5
0
0
5
6
0
0
5
4
0
0
6000 6000
5
4
0
0
PHC pile
shear strength improved type
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
84
2.2 Sinking and relative displacement
Due to the earthquake, the Sludge Inspection Building settled down 90mm and displaced later
ally 200mm relative to the main building (Fig. 6 & 7). As the large interrrspaces, more than one
hundred millimeter, were left between the sides of the base slab and the ground, relative instan
taneous displacement between the two buildings during the earthquake was supposed to be lar
ger than 200 mm. This large gap was one of the causes of the breakage of the sludge squeeze
transportation line (Fig. 8). The main sludgetreatment building, itself, was estimated to have
sunken down several tens millimeter by a leveling after the earthquake. Accordingly, the Sludge
Inspection Building sank down one hundred and several ten millimeter absolutely.
2.3 Damage of the piles
Figures 9 & 10 show the upper ends of the piles, which were cracked by bending clearly. How
ever, this type of failure is unexplainable for the
sinking more than several ten millimeters unless
the extent of the failure causes detaching of
concrete from the pile. The used piles were
typical shearstrengthimproved type which was
developed after the 1995 Kobe earthquake. Al
though there were many evidences of liquefac
tion of the surface layer around the building,
large scale lateral spreading or flowing were not
observed in the site. So shear failure of the piles
seemed hard to occur at underground. Finally, it
is concluded that the piles sank down from their
tips.
Figure 6. Right side is pile foundation
Figure 8. Breakage of sludge transportation line
(Kume 2007)
Figure 7. Left side is pile foundation
Figure 10. Crack at a pile top Figure 9. Crack at a pile top
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
85
2.4 Design and construction method of the piles
The applied design formula for estimating bearing capacity of the piles was commonly used
one;
(1)
here, RaL ; long term allowable bearing
capacity (kN)
R l =
3
1
( N A 1.5 L ) a
_
p + f
; end bearing capacity factor
10
D
L
90 ; = 25 (L 5m)
90 <
D
L
110 ; = 25
4
1
(
D
L
 90)
L ; length of pile
D ; external diameter of pile
; average standard penetration number value of ground between pile tip  4D depth
and pile tip +1D depth.
Ap ; cross section area of pile defined by outer diameter
Lf ; length (in meter) of pile where surface friction can be countable.
N
_
PWPEast2
PWPEast3
PileA
Figure 1 Schematic illustration of the shake table test model (CASE2)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
90
while 350 gal and 37 cm/s with 1.5 Hz were used in CASE2. The maximum velocity in
CASE2 was about twice of CASE1. In consequence, the table input energy of Case2 was
about four times of CASE1.
2.3 Soil material
Clean sand from Kasumigaura area in Ibaraki prefecture in Japan was sampled and used in this
experiment. Index tests were performed on this sample, and Figure 3 shows the grain size distri
bution and physical properties of the material.
2.4 Sample preparation method
Four reinforced concrete piles in CASE1 were installed in the center of the laminar box by pin
connection to the base. Diameter of the piles was 15 cm and their length was 4.5 m. Center to
center space between the piles was 0.9 m and a steel top cap provided a rigid connection on top
of the piles. Similarly, four steel piles in CASE2 were installed. A sheet pile in CASE1 and
CASE2 was installed in the east side of the piles with 30 cm space from the laminar box wall.
This sheet pile was used to provide lateral soil pressure on the liquefied soil to keep it moving
toward the water side. The box was partially filled with water, and dry sand was pluviated in the
water. As the hydraulic sediment was filled in the box, the water level was also increased. Based
on the past experience, it was proved that this method gives a uniform and relatively loose satu
rated sample (Dr= 45%). In the next step, the sheet pile installed, and the filling procedure con
tinued until the soil level in east and west side of the sheet pile reached to 4.0 and 3.2 m in
CASE1, respectively. The water level at this stage was 0.5 m from ground surface. Then, in the
east side, a layer of unsaturated soil was placed to raise the ground level to 4.5 m. Soil, piles and
sheet piles were heavily instrumented with pore water pressure and displacement transducers, as
well as accelerometers and strain gauges. Instruments were fixed on a thin net and installed in
the box, before sand pluviation. Total number of 256 channels was used for data acquisition of
500
0
500
A
c
c
.
(
g
a
l
)
(a)Input Acc.
50
0
50
0 1 2 3 4 5 6 7 8
V
e
l
.
(
c
m
/
s
)
Time (s)
(b)Input Velocity
500
0
500
A
c
c
.
(
g
a
l
)
(a)Input Acc.
50
0
50
0 2 4 6 8 10 12 14
V
e
l
.
(
c
m
/
s
)
Time (s)
(b)Input Velocity
(CASE1Sine 4.0Hz, a
max
=450Gal) (CASE2Sine 1.5Hz, a
max
=350Gal)
Figure 2 Measured input motion time histories on shake table
0
20
40
60
80
100
0.01 0.1 1 10
P
e
r
c
e
n
t
f
i
n
e
r
b
y
w
e
i
g
h
t
(
%
)
Grain size (mm)
Kasumigaura Sand
D50 (mm) 0.31
Uc 3.0
Gs 2.718
Fc (%) 5.4
emax 0.961
emin 0.570
Figure 3 Grain size distribution curve and physical
properties of Kasumigaura sand
Photo 2 Structure and sheetpile model on largescale
test
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
91
outputs of the instruments. Then, heavy plates were mounted and fixed on top cap until its
weight reached to about 10 tons, as a model for a massive super structure.
3 TEST RESULTS
3.1 Acceleration responses and excess pore water pressures of sand deposits
Figure 4 shows the time histories of response accelerations observed at the backyard of the pile
foundation and seaside in front of the quay wall. Similarly, Figure 5 shows the time histories of
excess pore water pressures. The comparison of these time histories of the accelerations and
excess pore water pressures in Figures 4 and 5 indicates that soil liquefaction occurred after 2 or
3 cycles and then vibrations did not transfer to upper layers due to soil liquefaction. The test re
sults reveal that after a few pulses of shaking, excess pore water pressure in loose backyard in
creased.
3.2 Displacements of sheetpiles and structures
The large displacement caused by soil liquefaction was measured on the top of the pilegroup
structure and sheet pile. The displacement time histories in CASE1 and CASE2 are shown in
Figure 6. The lateral displacements of the sheet pile in these cases cyclically accumulated dur
ing the shaking and slowly grew up after the shaking. On the other hand, the lateral displace
ments of the structure accumulated during the shaking, while these did not grow up remarkably
after the shaking.
The displacement of the top of the sheet pile at the end of the shaking (about t=5 second) was
about 25 mm in CASE1. The figure indicates that at t=50 second the lateral displacement in
creased only about 5 mm, and after this moment a rapid change in rate of lateral displacement
was observed. The flow continued until t=200 second and the maximum displacement reached
to about 100 mm at the top of the sheet pile. In CASE2, the displacement of the top of the sheet
500
250
0
250
500
East1(GL1140mm)
East2(GL2400mm)
A
c
c
.
(
G
a
l
)
(a) Acc. Ground East (backyard)
500
250
0
250
500
West1(GL1100mm)
West2(GL2360mm)
0 1 2 3 4 5 6 7 8
A
c
c
.
(
G
a
l
)
(b) Acc. Ground West (seaside)
Time (s)
500
250
0
250
500
East1(GL1220mm)
East2(GL2480mm)
A
c
c
.
(
G
a
l
)
(a) Acc. Ground East (backyard)
500
250
0
250
500
West1(GL1180mm)
West2(GL2440mm)
0 2 4 6 8 10 12 14
A
c
c
.
(
G
a
l
)
(b) Acc. Ground West (seaside)
Time (s)
(CASE1Sine 4.0Hz, a
max
=450Gal) (CASE2Sine 1.5Hz, a
max
=350Gal)
Figure 4 Acceleration time history measured in ground of backyard and seaside
10
0
10
20
30
East1(GL1140mm)
East2(GL2400mm) P
W
P
(
k
P
a
)
(a) PWPEast
(backyard)
10
0
10
20
30
West1(GL1100mm)
West2(GL2360mm)
0 1 2 3 4 5 6 7 8
P
W
P
(
k
P
a
)
(b) PWP_West
(seaside)
Time (s)
10
0
10
20
30
East1(GL1220mm)
East2(GL2480mm)
P
W
P
(
k
P
a
)
(a) PWPEast
(backyard)
10
0
10
20
30
West1(GL1180mm)
West2(GL2440mm)
0 2 4 6 8 10 12 14
P
W
P
(
k
P
a
)(b) PWPWest
(seaside)
Time (s)
(CASE1Sine 4.0Hz, a
max
=450Gal) (CASE2Sine 1.5Hz, a
max
=350Gal)
Figure 5 Excess pore water pressure generation and dissipation time histories in ground of backyard and seaside
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
92
pile at the end of the shaking (about t=9 second) was about 300 mm. The flow continued until
t=180 second and the maximum displacement reached to about 440 mm.
Based on the results, it can be clearly said that these tests were able to reproduce the phenome
non of postliquefaction behavior of the liquefied backfill sand that slowly flew toward seaside.
The displacements of the sheet pile and structure dividing with during and after the shaking
show in Table 1. The displacements of the sheet pile and structure in CASE1 and CASE2 were
almost same after the shaking, but these were large different during the shaking. The reason is
that these displacements depended on the input motion energy and the liquefaction response of
the sand deposit.
The influence on the displacement of the structure during the shaking was larger than that after
the shaking because the ratio of the displacement of the structure and sheet pile quay wall was
0.80 during the shaking and 0.20 after the shaking in CASE1, and 0.47 during the shaking and
0.19 after the shaking in CASE2.
These facts indicate that, when predicting displacement, it is not sufficient to take into account
the liquefactioninduced large ground deformation only after the shaking.
3.3 Bending Strains of piles
Strain gauges were attached on ten different height levels of the piles to measure the bending
strain time history during and after the shaking. Figure 7 shows bending strain time histories on
each depth of the sheet pile in CASE1 and CASE2. The distributions of bending strain about
the altitude shows in Figure 8.
The bending strain of pile top in CASE1 was too small to release bending moment by the struc
ture rocking. In addition, the large bending strain concentrated at the local point of GL3000
mm to break out cracks of the concrete pile. A local point of concentration of the bending strain
of the steel pile in CASE2 could not be found, but the value at around GL3000 mm was large.
This depth was near the bottom of the sheet pile because the stress of the pile became large in
the depth to cause ground deformation due to lateral spreading.
In CASE1, the distributions of bending strain at t=2.39 and t=2.58 seconds during the shaking
was small due to rotation of the pile top. The distribution at t=3.33 second during the shaking
showed peak at GL3,000 mm, and that bending strain caused by ground deformation due to lat
eral spreading. When the test model ground was demolished after the test, the damage of piles
50
0
50
100
150
Sheetpile
Structure
1 10 100 1000
D
i
s
p
.
(
m
m
)
Time (s)
250
0
250
500
/ /
Sheetpile
Structure
1 10 100 1000
D
i
s
p
.
(
m
m
)
Time (s)
(CASE1Sine 4.0Hz, a
max
=450Gal) (CASE2Sine 1.5Hz, a
max
=350Gal)
Figure 6 Displacement time histories on top of the sheet pile and structure
Table 1 Displacements of sheetpile and structure for during shaking and after shaking
During shaking
(CASE1:25sec)
(CASE2:29sec)
After shaking
(CASE1:5650sec)
(CASE2:91000sec)
CASE1
Sine4.0Hz,
a
max
=450Gal
Sheetpile 25mm 75mm
Structure 20mm 15mm
Structure/Sheetpile 0.80 0.20
CASE2
Sine 1.5Hz,
a
max
=350Gal
Sheetpile 300mm 85mm
Structure 140mm 15mm
Structure/Sheetpile 0.47 0.19
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
93
was investigated. As the results, several large cracks on the reinforced concrete pile caused near
GL3000 mm and remained large bend.
The distribution shape of bending strain in CASE2 is smooth because of the steel pile. The dis
tributions at t=3.11 and t=4.49 seconds during the shaking are small similarly CASE1, but
show large value near bottom of the sheet pile. The distributions at t=50 and t=665 seconds after
the shaking show a steady shape except near the pile tip.
4 CONCLUDINS
Two largescale shake table tests were conducted to study the seismic response of sheet pile
wall system and the liquefaction and deformation characteristics of saturated sand backfill, as
well as the response of its neighboring pile foundations.
The test results revealed that after a few pulses of shake, excess pore water pressure in saturated
and relatively loose backfill increased, and the consequent loss of effective stress resulted in lat
eral spread of the liquefied sand.
Postliquefaction behavior of the liquefied sand was quite remarkable. The results clearly dem
onstrate that the tests can reproduce the phenomenon of the postliquefaction behavior of the li
quefied backfill sand that slowly flows toward seaside.
200
0
200
400
S
t
r
a
i
n
(
x
1
0

6
)
(a) Bending strain (PileA)
GL250mm
200
0
200
400
S
t
r
a
i
n
(
x
1
0

6
)
(b) Bending strain (PileA)
GL500mm
300
0
300
600
S
t
r
a
i
n
(
x
1
0

6
)
(c) Bending strain (PileA)
GL1250mm
400
0
400
800
S
t
r
a
i
n
(
x
1
0

6
)
(d) Bending strain (PileA)
GL2000mm
1200
600
0
600 S
t
r
a
i
n
(
x
1
0

6
)
(a) Bending Strain (PileA)
GL380mm
1200
600
0
600
S
t
r
a
i
n
(
x
1
0

6
)
(b) Bending Strain (PileA)
GL1220mm
1200
600
0
600
S
t
r
a
i
n
(
x
1
0

6
)
(c) Bending Strain (PileA)
GL2060mm
1200
600
0
600 S
t
r
a
i
n
(
x
1
0

6
)
(d) Bending Strain (PileA)
GL2900mm
800
0
800
1600
1 10 100 1000
S
t
r
a
i
n
(
x
1
0

6
)
(e) Bending strain (PileA)
GL3000mm
Time (s)
1600
800
0
800
1 10 100 1000
S
t
r
a
i
n
(
x
1
0

6
)
(e) Bending Strain (PileA)
GL3320mm
Time (s)
(CASE1Sine 4.0Hz, a
max
=450Gal) (CASE2Sine 1.5Hz, a
max
=350Gal)
Figure 7 Bending strain time histories on pileA
1000 0 1000 2000
5000
4000
3000
2000
1000
0
t=2.39(s)
t=2.58(s)
t=3.33(s)
t=5.00(s)
t=50. 0(s)
t=665(s)
Bending strain (x10
6
)
D
e
p
t
h
(
m
m
)
2000 1000 0 1000
5000
4000
3000
2000
1000
0
t=3.13(s)
t=3.50(s)
t=3.85(s)
t=4.94(s)
t=50. 0(s)
t=665(s)
Bending strain (x10
6
)
D
e
p
t
h
(
m
m
)
(CASE1Sine 4.0Hz, a
max
=450Gal) (CASE2Sine 1.5Hz, a
max
=350Gal)
Figure 8 Bending strain distributions on pileA
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
94
The influence of during the shaking on the displacement of structure is larger than that after the
shaking. The displacements of the sheet pile and structure is dominated in the input motion
energy and the liquefaction response of sand deposit.
It is important to note that this study is an ongoing research process and more data analyses,
shake table tests, centrifuge tests and numerical analyses are being conducted and the results
will be published.
5 REFERENCES
Editorial Committee for the Report on the HanshinAwaji Earthquake Disaster, 1997, Report on the Han
shinAwaji Earthquake Disaster, Damage to Civil Engineering Structure, Port and Coastal Facilities,
River Management Facilities and Sediment and Slope Movement Control Facilities, JSCE
Inatomi, T., et al., 1997, Technical Note of Damage to Port and Portrelated Facilities by the 1995 Hyo
goken nanbu Earthquake, The Port and Harbor Research Institute, Ministry of Transport, Japan
Hanshin Expressway Public Corporation, 1996, Investigation on the Seismic Damages of Bridge Founda
tion in the Reclaimed Land
Sato, M., Ogasawara, M., Tazoh, T., 2001, Reproduction of lateral ground displacements and lateralflow
earth pressures acting on pile foundations using centrifuge modeling, Fourth International Conference
on Recent Advance in Geotechnical Earthquake Engineering and Sail Dynamics, San Diego, USA,
6pp.(CDRom), Paper No.9.33.
Tazoh, T., Sato, M., Gazetas, G., 2005, Centrifuge tests on pilefoundationstructure systems affected by
liquefactioninduced flow due to quaywall collapse, Proceedings of the 1
st
GreeceJapan Workshop
on Seismic Design, Observation and Retrofit of Foundations, pp.79106.
Tokimatsu, K.., Mizuno H., Kakurai M., 1996, Building Damage Associated with Geotechnical Prob
lems, Special Issue on the 1995 Hyogokennanbu Earthquake, Soils and Foundations, p.p. 219234.
Yasuda, S., Ishihara, K., Morimoto, I., Orense, R, Ikeda, M., and Tamura, S., 2000, Largescale shakingt
able tests on pile foundations in liquefied ground, Proceedings, 12th World Conference on Earth
quake Engineering, Auckland, New Zealand, Paper No. 1474, 8pp.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
95
1 INTRODUCTION
Many large earthquakes have caused severe damage to various structures. Especially in the
1995 HyogokenNambu earthquake, a lot of pilesupported structures behind quay walls in port
areas collapsed due to widespread liquefaction and its resulting lateral spreading of ground.
Such structures are necessary as facilities for rescue supplies to disaster areas as well as eco
nomical distribution systems during rehabilitation process after the disaster. Therefore, under
standing the lateral spreading behavior of liquefied ground and its influence so as to mitigate
earthquake disaster for port structures is one of the very important problems in geotechnical
earthquake engineering discipline. In spite of its importance, reproduction of a lateral spread
ing phenomenon in the model ground on a centrifuge apparatus or shaking table is almost im
possible because mainly of the small model size, so that the influences of lateral spreading on
structures still have not been completely revealed. In other words, it is necessary to conduct
shaking tests of a largescale model ground with a structure subjected to actual magnitude
earthquake motions to understand the behavior and influences of lateral spreading.
For this reason, the authors carried out a series of tests of a largescale model ground with a
quay wall and grouppilesupported structure on lateral spreading due to liquefaction by the E
Defense shaking table (MEXT 2007). The objective of the testing series is reproduction of li
quefactioninduced lateral spreading to observe the phenomena of the model ground and struc
tures in detail. This experimental study involves revealing the failure mechanism of a quay
wall and pilesupported structure behind the wall. The paper here describes the results of one
of the tests of the model ground with a caissontype quay wall and pilesupported structure, and
explains their behaviors.
EDefense Shaking Table Test on LiquefactionInduced Lateral
Spreading of LargeScale Model Ground with Quay Wall and
PileSupported Structure
K. Tabata & M. Sato
National Research Institute for Earth Science and Disaster Prevention, Tsukuba, Japan
ABSTRACT: A shaking table test of a largescale model was performed at the EDefense
threedimensional shaking table facility in order to observe the behavior of lateral spreading of
liquefiable ground and to evaluate the mechanism of its influence on the failure of structures.
The model ground was prepared in a large rectangular container with a caissontype quay wall,
pilesupported structure and about nine hundred sensors. To this model, horizontal and vertic
al input motions based on one of the 1995 HyogokenNambu earthquake records were applied.
The motions induced liquefaction of the ground and horizontal displacement of the caisson to
waterside, causing deformation of the structure due to the piles bent and collapsed. Observa
tion of the test explains contributions of structural inertia and ground deformation to the beha
vior of the caisson and structure.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
96
2 OUTLINE OF THE TEST
2.1 EDefense shaking table
The test was carried out by the EDefense shaking table. EDefense is the name of a fullscale
threedimensional earthquake testing facility operated by the National Research Institute for
Earth Science and Disaster Prevention (Ohtani et al. 2003). After the start of its operation in
2005, a lot of shaking tests on soilstructure interaction and collapse behavior of various struc
tures have been successfully performed (Kajiwara et al. 2007, Tabata & Kajiwara 2009). The
key of the EDefense is a shaking table, which is the worlds largest threedimensional table 20
m long and 15 m wide as shown in Figure 1. The remarkable feature of the table presented in
Table 1 is the capability to reproduce ground motions recorded in the 1995 Kobe earthquake for
a 12MN structure by 10 horizontal and 14 vertical actuators.
Table
14 vertical
actuators
10 horizontal
actuators
2
0
m
lo
n
g
15m
w
ide
Figure 1. Illustration and photo of the EDefense shaking table.
Table 1. Specifications of the EDefense shaking table.
Horizontal (x and y) Vertical (z)
acceleration 900gal 1500gal
velocity 2m/s 0.7m/s
displacement 1m 0.5m
allowable moment 150MNm 40MNm
* at the maximum load
Table size
Loading capacity
Maximum performance*
20m x 15m
12MN
2.2 Specimen of the model ground
Figure 2 and Photo 1 show the specimen of the model prepared in a rectangular container 16 m
long, 4 m wide and 5 m high. The specimen was a liquefiable ground with a caissontype quay
wall and structure supported by a 3by2 pile group. The liquefiable ground was made of Al
bany silica sand compacted to 60percent relative density and saturated by deaired water before
testing. The properties and indices of the sand are presented in Table 2, and its deformation
characteristics are described by Yasuda et al. (2006). The grain size distribution of the sand is
also shown in Figure 3, indicating that the distribution is similar to that of Toyoura sand. The
deposit was divided by the caisson into 2.5meterthick waterside and 4.5meterthick land
side deposits. The water table was 0.5 m below the landside ground surface, i.e. 1.5 m above
the waterside ground surface. In the landside deposit, the pilesupported structure was in
stalled behind the caisson, consisting of six hollow steel piles, footing and the weight modeling
a superstructure. The six piles were aligned with three parallel to the caisson in two rows and
fixed to the footing and pinned at the containers bottom. The 10ton footing penetrated into
the landside deposit to a depth of 0.5 m. The weight was placed on the footing, inducing iner
tial force by a 12ton weight.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
97
1
6
0
0
WATER TABLE
4
0
0
0
5
0
0
2
0
0
0
7
5
0
16000
4
5
0
0
4
0
0
0
1300 4400 9300
1
7
5
0
CAISSON
MOUND
4
0
0
1600
PILE
LAYOUT
1000
2
0
0
0
9
6
0
9
6
0
CAISSON
1600
3550
600 250 800 600
Dr = 60%
Dr = 90%
SUPER
STRUCTURE
D F
D F
PILE
FOOTING
UNIT: mm
(a) Plan view
(b) Section view
PILE A2
PILE B2
WATERSIDE
WATERSIDE
LANDSIDE
LANDSIDE
y
z
y z
Figure 2. Illustrations of the specimen and the locations of the measurement points D and F: (a) Plan
and (b) section views.
Photo 1. Specimen on the table before shaking.
Table 2. Properties and indices of Albany silica sand.
Figure 3. Grain size distributions of Albany silica sand and Toyoura sand.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
98
2.3 Measurement installation
To achieve the objective , measurement sensors listed in Table 3 were installed with the speci
men. The table presents a total of 836 sensors that were mounted to observe the behavior in
detail. Additionally, a unique threedimensional displacement measuring system was em
ployed to investigate the motions of ground surface and structures (Tokuyama et al. 2007). In
this system, displacement is determined by digital video cameras that observed reflective, spher
ical markers attached to an object.
Table 3. Sensors installed with the specimen.
Type of sensor
G
r
o
u
n
d
C
a
i
s
s
o
n
P
i
l
e
s
S
u
p
e
r
s
t
r
u
c
t
u
r
e
a
n
d
f
o
o
t
i
n
g
C
o
n
t
a
i
n
e
r
T
o
t
a
l
Strain gauge 234 40 274
Accelerometer 89 8 22 24 24 167
Veolcity transducer 3 3 6
Displacement transducer 11 14 11 36
Earth pressure transducer 17 104 16 137
Water pressure transducer 119 7 72 198
Load cell 18 18
2.4 Testing program
The specimen set on the table was shaken under twodimensional, horizontal and vertical mo
tions based on the northsouth and updown motions recorded at the JR Takatori station during
the 1995 Kobe earthquake. In the test, the northsouth component was applied to the speci
mens long direction, and the updown component to the vertical direction. The peak table ac
celerations were approximately 6.0 and 1.7 m/s
2
in the horizontal and vertical directions respec
tively, and the shaking duration was about 42 seconds. Figure 4a shows the acceleration time
histories of the target input motion and actual, observed table motion from zero (data acquisition
start time) to 30 seconds. Because the nature of the specimen was significantly changed due to
liquefaction caused by strong motions, the tables control system could hardly follow such
changes and reproduce table motions that were identical to the target. However, as shown in
Figure 4b, both shapes of Fourier spectrum are very similar, especially in the domain of lower
frequency.
Figure 4. Comparison between the target input motion signaled to the table and actual, observed table
motion: (a) Acceleration time histories and (b) acceleration Fourier amplitude spectra.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
99
3 TEST RESULTS
3.1 Observation of the specimen after test completion
Photo 2 shows the specimen before and after the test, and Figure 5 illustrates the section of the
specimen that demonstrates its change due to shaking. The caisson overturned toward the wa
terside (left of the specimen in the photo) with the horizontal displacement at the top of about 2
m and 25degree decline, accompanying its mound with horizontal deformation and small set
tlement. Three waterside piles bent at almost the same level of the mound. This, in turn,
caused horizontal deformation of the pilesupported structure toward the caisson with 47degree
decline of the weight and footing.
As shown in Figure 5, horizontal displacement of the landside ground surface ranged up to 2
m and decreased with the distance from the caisson. Such trend is similar to some cases ob
served in the 1995 Kobe earthquake (Ishihara et al. 1996). The liquefaction and resulting lat
eral spreading also caused settlement of the landside deposit of about 20 cm, while relatively
small settlement was observed at the waterside deposit.
Photo 2. Caisson, pilesupported structure and their surroundings (a) before shaking and (b) after test
completion.
Dr = 60%
Dr = 90%
WATERSIDE LANDSIDE
y
z
Figure 5. Section of the change of the specimen due to shaking. Gray and black colors indicate the
structures and measurement points before and after the test respectively.
3.2 Excess pore water pressure change during and after shaking
Figure 6 shows the changes of excess pore water pressure, u, at four different levels (0.6, 1.7,
2.9 and 3.7 m from the landside surface) under the measurement point F. As shown in the fig
ure, u buildup occurred at all levels when the excitation started, and then reached their over
burden pressure in 7 to 10 seconds after the start. In consequence, it can be said that all layers
of the deposit were liquefied due to the applied input motions. Following this u increasing
process, u dissipation started from the bottom layer to the surface. This observation implies
that liquefied deposit became dense from the bottom to the surface during the dissipation
process of u. In addition, u began to dissipate when its pressure level reached that of the
lower layer.
(a) (b)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
100
Figure 6. Excess pore water pressure changes at different levels under the point F.
3.3 Behavior of the caisson during shaking
Just after start of the shaking, the caisson moved to the waterside and landside alternately a few
times, and finally overturned toward the waterside at 10.6 seconds after the test start. To eva
luate the effects of its inertial force and backfill earth pressure change on the caissons behavior,
the time histories of the displacements and effective earth pressure changes are introduced as
following. Figure 7 shows the vertical displacements, z, of the waterside and landside tops of
the caisson, presenting decline of the caisson began at 6 seconds. Figure 8 shows the horizon
tal displacements, y, of the caisson, the measurement point D on its backfill, and the shaking ta
ble. In the figure, a positive y value means the displacement to the waterside. As shown in
the figure, the horizontal displacement of the caisson is always larger than that of the point D
with almost opposite phase of the table. Additionally, the effective earth pressure change on
the caissons landside, p', at different levels from the surface shown in Figure 9 is constantly
negative from 5 seconds to 10.6 seconds at which the caisson overturned toward the waterside.
These facts suggest that the behavior of the caisson during shaking was principally dominated
by its inertial force rather than the p' change of the backfill. Indeed, many cracks on the
ground surface and gaps between the caisson and its backfill are observed in Photo 3, which is
the close view of the caisson and its surroundings including the footing and the point D around
at 8 seconds. Therefore, it can be said that the influence of caissons inertial forces is more
dominant to the stability of the caisson under earthquake motions than that of the p' change of
its backfill.
Figure 7. Vertical displacements determined at the waterside and landside tops of the caisson.
Figure 8. Horizontal displacements of the caisson, point D and shaking table.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
101
Figure 9. Effective earth pressure changes on the backfill side of the caisson.
Waterside Landside
Caisson
Point D Footing
Photo 3. Close view of the caisson, backfill surface and footing during shaking.
3.4 Behavior of the pilesupported structure during shaking
Figure 10 shows horizontal displacements of the caisson, weight of the pilesupported structure
and their surroundings including the measurement points D and F. As already explained, the
caisson overturned toward the waterside at 10.6 seconds after the test started. Until the caisson
overturned, the horizontal displacement of the point D on the backfill of the caisson was always
smaller than that of the caisson, while the displacement of the weight was smaller than those of
the caisson and point D and almost same behavior of the point F on the ground behind the pile
supported structure. It is also observed that these displacements change with nearly same
phase. Just after the caisson overturned at 10.6 seconds, the displacement of the weight
reached the exactly same of the point D and increased, and finally declined to the waterside.
In this process, the displacement of the point F was smaller than those of the weight and point
D. This observation implies that one of the triggers to decline the pilesupported structure can
be the large deformation of the ground between its footing and the caisson due to the caisson
overturned.
Figure 10. Horizontal displacements of the caisson, weight, and points D and F.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
102
4 CONCLUSIONS
In order to investigate the behavior of liquefactioninduced lateral spreading and the mechanism
of its influence on structures, a shaking table test of a largescale model ground with a caisson
type quay wall and grouppilesupported structure assumed as a situation of port areas was con
ducted at the EDefense shaking table facility. In the test, about nine hundred sensors moni
tored the behavior in detail and the threedimensional displacement measuring system was em
ployed to observe large displacements.
The specimen of the model ground was shaken under two dimensional, horizontal and vertic
al motions based on one of the 1995 Kobe earthquake records. Such motions induced lique
faction in all layers of the deposit, causing overturn of the caisson toward the waterside and ho
rizontal deformation of the landside ground. According to the displacement and effective earth
pressure change of the caisson, the overturn was mainly caused by its inertial force, not the in
fluence of its backfill ground deformation. After the caisson overturned and the following de
formation of its backfill ground caused, the pilesupported structure declined to the waterside.
It can be considered that this phenomenon was triggered by large deformation of the ground be
hind the caisson. Note that these observations and considerations are based on the model test
ing results, meaning that the observed area is still limited compared to the actual field. Hence,
to apply the knowledge from this study to practice, other research procedures, such as parame
tric testing of small models, field investigations and computational analyses, as well as more
largescale model testing are needed.
ACKNOWLEDGEMENT
The testing program described here is a part of the Special project for mitigation of earthquake
disaster in urban areas funded by the Japanese Ministry of Education, Culture, Sports, Science
and Technology. The motions applied to the shaking table were based on the record obtained
at the JR Takatori station during the 1995 Kobe earthquake that is originally from the Railway
Technical Research Institute in Japan. These supports and contributions are gratefully ac
knowledged.
REFERENCES
Ishihara, K., Yasuda, S. & Nagase, H. 1996. Soil characteristics and ground damage. Special issue, Soils
and Foundations: 109118.
Kajiwara, K., Sato, M. & Nakashima, M. 2006. Shaking table and activities at EDefense. Proc. 1st Eu
ropean Conference on Earthquake Engineering and Seismology: Paper no.733.
Ministry of Education, Culture, Sports, Science and Technology (MEXT). 2007. Summary Report, Spe
cial project for mitigation of earthquake disaster in urban areas: 6675 (in Japanese).
Ohtani, K., Kagawa, N., Katayama, T. & Shibata, H. 2003. Construction of EDefense (3D fullscale
earthquake testing facility). Proc. 2nd Intern. Symp. New Technologies for Urban Safety of Mega Ci
ties in Asia: 6976.
Tabata, K. & Kajiwara, K. 2009. Experimental research on behavior of various structures under earth
quake motions at EDefense. Proc. 2nd ChinaJapan Science Forum, Beijing, China: 205206.
Tokuyama, H., Tabata, K., Nakazawa, H. & Sato, M. 2007. Applicability of threedimensional displace
ment measurement to model ground in largescale shaking table testing at EDefense. Proc. 42nd An
nual Symp. Japanese Geotechnical Society: 16431644 (in Japanese).
Yasuda, S., Saito, S. & Suzuki, S. 2006. Effect of confining pressure on liquefaction behavior of sand.
Proc. 61st Annual Symp. Japanese Society of Civil Engr.: 547548 (in Japanese).
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
103
Design charts for single piles under lateral spreading of liquefied
soil
A. Valsamis
Dr Civil Engineer, Civil Engineering School, National Technical University of Athens, Greece
G. Bouckovalas
Professor, Civil Engineering School, National Technical University of Athens, Greece
E. Drakopoulos
Civil Engineer Msc, National Technical University of Athens, Greece
1 INTRODUCTION
ABSTRACT: Multivariable design charts and relationships are presented for the preliminary
computation of maximum pile displacement and bending moment in the case of earthquake
induced liquefaction and lateral spreading of the natural ground. The charts were initially devel
oped on the basis of results from a large number of parametric analyses which were performed
with the pseudo static (Py) method, and consequently calibrated with the aid of 2D and 3D
dynamic analyses of the liquefied soil and the pile. Three different combinations of pile and
soil conditions were considered, which are commonly encountered in practice.
One of the most damaging effects of earthquakeinduced soil liquefaction is the lateral spread
ing of soils, where large areas of ground move laterally to lengths ranging from some centime
ters to a few meters. This phenomenon may occur in the case of even small surface inclination
(e.g. 24%) or small topographic irregularities (e.g. 23m) such as those near river and lake
banks.
In such cases, the kinematic interaction of single piles and pile groups with the lateral
spreading ground may induce significant additional residual horizontal loads and bending mo
ments to the pile, which cannot be predicted by common design methods for superstructure
loading.
2 PSEUDOSTATIC PREDICTION METHODS
The detailed analysis of piles against lateral spreading is a rather complicated soilstructure in
teraction problem which, strictly speaking, requires a sophisticated numerical simulation, well
beyond the limits of common applications. Thus, for simplified computations, a number of
pseudostatic methodologies have been developed, where the loads or displacements applied by
the laterally spreading ground are being estimated independently and subsequently applied as
external loads to the pile. Existing pseudostatic methodologies may be divided in two catego
ries:
a) The Py method, which relies upon the substitution of the ground with Winkler
type springs that are governed by a nonlinear loaddisplacement (Py) law. Ac
cording to this methodology an independent estimation of the ground displacement
is made and the resulting displacements are applied to the base of the springs in or
der to evaluate the pile deflection and the corresponding shear forces and moments
(e.g. Tokimatsu 1999, Boulanger et al 2003).
b) The limit equilibrium method, which is based on a pseudostatic estimation of the ul
timate pressure that the laterally spreading ground applies to the pile. Pile displace
ments and bending moments can be consequently evaluated (e.g. JRA 1996, Dobry
et al 2003) from beam theory.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
104
Recently, Ashford & Juirnarongrit (2004) concluded that the Py method is the most reliable
method of the two, after comparing the two most commonly used limit equilibrium methods
(JRA, 1996 and Dobry et al., 2003) with a simple Py method that used the curves proposed
from Reese et al. (1974) for sands, degraded with a factor = 0.1 in order to take into account
the soil liquefaction. Bhattacharya et al. (2003) also concluded that the limit equilibrium method
of JRA (1996) is systematically nonconservative. Thus, on the ground of these independent
findings, the Py method has been chosen to derive the design charts in this paper.
More specifically, the method proposed by Branderberg (2002) has been selected, according
to which the Py curves of (1995) for the nonliquefied sands should be used, after being
degraded with a loading factor . This factor represents the effect of liquefaction on the me
chanical characteristics (soil strength and deformation) of the natural soil and can be computed
according to Table 1, in terms of the corrected blow count of the Standard Penetration Test
(N
1
)
60CS
.
The aforementioned methodology has been chosen among seven (7) compatible methodolo
gies (Ishihara & Cubrinovski, 1998, Cubrinovski et al., 2006, Rollins et al., 2005 & 2007, To
kimatsu, 1999, High Pressure Gas Safety Institute of Japan, 2000, Railway Technical Research
Institute of Japan, 1999, and Matlock, 1970) following an extensive evaluation through com
parison to three centrifuge experiments (Abdoun 1998) and one large shaking table experiment
(Cubrinovski et al. 2004).
Table 1. Proposed degradation factors after Branderberg (2000)
(N
1
)
60CS
<8 0 to 0.1
816 0.1 to 0.2
1624 0.2 to 0.3
>24 0.3 to 0.5
3 PARAMETRIC ANALYSES
The numerical analyses have been performed with the help of the finite elements program
NASTRAN (MacNealSchwendler Corp. 1994). Simulation of the liquefied soil layers was
based on the Py methodology outlined in the previous paragraph. The nonliquefied soil layers
have been simulated with the Py curves proposed by API (1995, 2002) without the use of a
degradation factor. It should be mentioned that, as long as the nonliquefiable base layer does
not fail, the exact Py curve used for its simulation does not affect significantly the results, as its
stiffness is almost 100 times larger than that of the liquefied soil above it.
Based on a previous study of the lateral spreading phenomenon (Valsamis et al., 2007), the
variation with depth of the lateral displacements of the liquefied soil was assumed as a quarter
sine, with the maximum displacement developing near the top of the layer and zero displace
ment at the bottom of the layer. On the other hand, the displacement of the nonliquefied soil
layers was assumed to remain constant with the depth.
In total, one hundred sixty two (162) parametric analyses have been performed, concerning
three different combinations of piles and ground layer profile (Figure 1):
o A 2layered soil profile, where there is a surface liquefied soil layer which spreads lat
erally, while the pile rests inside a nonliquefiable bottom layer. This is a common case
near river or lake banks, where there are loose surface alluvial deposits.
o A 2layered soil profile, same as above, where the pile head is fixed in position, due to
superstructure constraints. This case is often encountered in piles supporting bridges or
other large structures, where the superstructure restrains the pile head from moving.
o A 3layered soil profile, similar with the 2layered profile with the addition of a non
liquefiable soil crust. This is the most common case in small distances from water
fronts, where the water table is some meters below the soil surface, leaving a non
liquefiable surface soil crust.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
105
This categorization is justified on the grounds that any possible constraints on the free pile
head displacement and rotation, either due to the nonliquefiable soil crust or a superstructure,
proved to be among the the most important factors controlling the pile response. Note that, Ishi
hara & Cubrinovski (1998), Brandenberg (2002), Rollins et al. (2005) have also reached similar
conclusions for the effect of the pile head constraint enforced by a nonliquefiable soil crust.
Sixty six (66) parametric analyses have been performed for the 2layerd case and the follow
ing pile & soil input parameters:
o Relative Density Dr=35 90 % (degradation factor = 0.05 0.4 and friction angle
= 32
o
42
)
o Thickness of liquefied soil layer
liq
= 6 to 10m
o Elasticity Modulus for the pile ( = 30 to 210 GPa)
o Pile diameter (D = 0.15m to 0.6m),
o Pile stiffness = 16 to 1336 Mm2, and
o Ground surface displaments D
h
= 0.125m to 1.20m.
For the 2layered geometry with fixed pile head, fortysix (46) analyses have been performed
with the same range of parameters. For the 3layer geometry fifty (50) numerical analyses have
been performed. These analyses cover the above mentioned range of parameters in combination
with soil crust thickness crust = 1 to 4m.
Figure 1. Static models for the (a) 2layered, (b) 3layered and (c) fixed pile head cases
4 DESIGN CHARTS
Pile design against lateral spreading must assure that, following the seismic excitation:
a) no structural failure of the pile has occurred (no development of plastic hinges at any
depth), and
b) no performance failure of the superstructure should be encountered due to excessive
superstructure displacements.
To check against these criteria, both the maximum developing moment and the maximum dis
placement of the pile head are needed.
The depth of the maximum bending moment is in general variable. For the cases considered
in this article (Figure 1), it is known beforehand that maximum moments develop at the inter
face between the liquefied soil layer and the nonliquefied base layer. Similarly, it is known be
fore hand that the maximum pile displacement develop at the pile head, for the 2 and the 3
layered cases, and near the middepth of the liquefiable soil layer for the fixed pile head case.
For these reasons, the statistical analysis of the parametric analyses results has been focused
upon the magnitude of those two design parameters and not upon the respective location along
the pile.
It should also be mentioned that the statistical processing was not blind, e.g. based only on
some algorithm that minimizes the error of the empirical predictions. On the contrary, a general
form of the prediction relations was initially obtained based on analytical solutions of the static
q
q
()
( )
P
()
(a) (b) (c)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
106
models presented in Figures 1a, 1b and 1c and subsequently the statistical processing was used
to calibrate the general relations against the results of the parametric analyses.
Figures 2a and 2b present the proposed design charts for the maximum pile displacement and
the associated bending moment for the 2layered soil profile. Alternatively, the pile head dis
placement D
pile
(m) may be computed from Figure 2a and the respective maximum bending
moment
max
(kN/m) may be subsequently estimated as:
2
max
2 . 2
liq
pile
H
EID
M = (1)
where Hliq(m) is the liquefied soil layer thickness and EI(kN/m2) is the pile stiffness.
Observe that the relation in Figure 2a is strongly nonlinear. This is due to the fact that the
Winkler springs representing the soil are elastoplastic and thus, after a certain soil displace
ment, the loads due to the lateral movement of the soil remain constant. This elastoplastic re
sponse of the soil springs is the main reason why the derivation of a simple analytical expres
sion for the pile displacement was not possible. Moreover, note that the correlations of Figure
2a are not dimensionless and thus they should always be used in conjunction with the interna
tional system unit SI (kN, m).
0.01 0.1 1 10 100
(D
h*
EI)/(H
l i q
6
D
2
)
0.01
0.1
1
10
100
(
D
p
i
l
e
*
E
I
)
/
(
H
l
i
q
6
D
2
)
b=0,4 +D=0,3m
b=0,4 +D=0,15m
b=0,2 +D=0,3m
b=0,2 +D=0,15m
b=0,1 +D=0,6m
b=0,1 +D=0,3m
b=0,05 +D=0,6m
b=0,05 +D=0,3m
(a)
200 400 600 800 1000
(D
pi l e*
EI)/H
l i q
2
0
500
1000
1500
2000
M
m
a
x
b=0.05 & D=0.3m
b=0.1 & D=0.3m
b=0.2 & D=0.3m
b=0.,4 & D=0.3m
b=0.4 & D=0.15m
b=0.2 & D=0.15m
b=0.1 & D=0.6m
b=0.05 & D=0.6m
(b)
Dr
= 50%
Dr
= 65%
D
r
= 85%
Dr
= 40%
Figure 2. Design charts (a) for the maximum pile displacement and (b) for the maximum developing
bending moment of the pile, for the 2layered soil profile
The design charts for the 3layered soil profiles are shown in Figures 3a and 3b. Observe that
the pile head displacement follows systematically the nonliquefied soil crust displacement. This
observation has been also confirmed from centrifuge experiments (Abdoun, 1999) which show
that pile head displacements are only slightly larger than soil surface displacements. In this case,
it was possible to develop simplified analytical relations, both for the pile head displacement
and the developing bending moments, namely:
h pile
D D = 22 . 1 (2)
65 . 0
2
max
18


.

\

=
liq
pile
H
EID
M (3)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
107
0.01 0.1 1 10
D
h
0.01
0.1
1
10
D
p
i
l
e
b=0,4 +D=0,3m
b=0,4 +D=0,15m
b=0,2 +D=0,3m
b=0,2 +D=0,15m
b=0,1 +D=0,6m
b=0,1 +D=0,3m
b=0,05 +D=0,6m
b=0,05 +D=0,3m
(a)
1000 2000 3000 4000
(D
pi l e*
EI)/H
l i q
2
0
1000
2000
3000
4000
M
m
a
x
b=0,05 +D=0,3m
b=0,1 +D=0,3m
b=0,2 +D=0,3m
b=0,4 +D=0,3m
b=0,4 +D=0,15m
b=0,2 +D=0,15m
b=0,1 +D=0,6m
b=0,05 +D=0,6m
(b)
Figure 3. Design charts (a) for the maximum pile displacement and (b) for the maximum developing
bending moment on the pile, for the 3layered soil profile
Finally, for the fixed pile head case, the design chart for the maximum pile displacement and
bending moment, are being presented in Figures 4a and 4b respectively.
In this case also, it was possible to phrase analytical relations for the estimation of the above
mentioned design parameters:
( )
2
2 2
3 . 0 12
EI
D
D H D
h liq pile

= (4)
2
max
18
liq
pile
H
EID
M = (5)
where is the degradation factor for the soil strength due to the liquefaction which can be taken
from Table 1.
0.0001 0.001 0.01 0.1 1 10
(D
h
)
0.3
H
liq
12
(D/EI)
2
0.001
0.01
0.1
1
D
p
i
l
e
(
m
)
b=0,4 +D=0,15m
b=0,3 +D=0,3m
b=0,2 +D=0,3m
b=0,2 +D=0,15m
b=0,1 +D=0,6m
b=0,1 +D=0,3m
b=0,05 +D=0,6m
b=0,05 +D=0,3m
(a)
20 40 60 80 100
EID
p
/H
liq
2
400
800
1200
1600
2000
M
(
k
N
m
)
b=0,4 +D=0,15m
b=0,3 +D=0,3m
b=0,2 +D=0,3m
b=0,2 +D=0,15m
b=0,1 +D=0,6m
b=0,1 +D=0,3m
b=0,05 +D=0,6m
b=0,05 +D=0,3m
(b)
Figure 4. Design charts for the (a) maximum developed pile displacement and (b) the maximum develop
ing bending moment on the pile, for the fixed pile head case
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
108
5 COMPARISON WITH DYNAMIC NUMERICAL ANALYSES RESULTS
The proposed design charts have been evaluated with the help of 2Dimension and 3Dimension
dynamic numerical analyses. The numerical analyses were performed with the aid of the non
linear Finite Difference method combined with an effective stress constitutive model which can
simulate the static and the dynamic response of cohesionless soils, including liquefaction. In
brief, this is a bounding surface model, with a vanished elastic region, which was built based
upon the Critical State Theory framework (Papadimitriou et al., 2001, Andrianopoulos et al.
2006a, 2006b and 2007). One of its main characteristics is that the monotonic and cyclic re
sponse of soils is described using a single set parameters which is soilspecific, but does not de
pend on the initial stress and density conditions.
The 2Dimension simulation of a clearly 3Dimensioned problem with the help of a plane
strain algorithm, has been achieved with the help of the special pile elements, incorporated in
the finite difference program FLAC2D (Itasca, 2005). Those elements, are in fact beam ele
ments which have the ability to connect to the FLAC grid that simulates the surrounding soil
with the help of special Winkler springs, whose maximum strength depends on the mean effec
tive stress of the surrounding soil. Those elements can simulate in a simple way the liquefac
tioninduced degradation of the soil characteristic, since their strength depend on the effective
strength and thus can be used for the simulation of piles inside laterally spreading soil (Valsamis
2008). Figure 5a shows a typical 2Dimension grid used for the simulation of the problem. Sixty
seven (67) such numerical analyses have been performed with the following range of parame
ters:
o Relative Density of liquefiable soil layer Dr = 35 to 85%
o Thickness of liquefiable soil layer liq = 2 to 8m
o Thickness of nonliquefiable soil crust Hcrust = 0 to 6 m
o Pile stiffness EI = 500 to 80000 kPam4
o Pile diameter = 0.3 to 1.0m
o Ground surface displacement Dh = 0.2 to 1.4m.
o 2layered and 3layered soil profiles have been simulated, both with free and fixed
pile head.
Moreover, thirty (30) 3dimension numerical analyses have been performed, in which the
ground has been simulated with the same constitutive model. The pile in this case was simulated
with a simple elastic model. Between the soil grid and the pile, contact elements have been used
with friction angle equal to 2/3 x (Itasca 1997), which allows the relative movement of the
liquefied ground around the pile. Figure 5b shows a typical grid used for the 3D simulation of
the problem. 3D numerical analysis has the advantage of simulating directly the problem and
thus not having any predefined parameters such as the coupling springs parameters as with the
case of the 2D numerical analysis. This increases our confidence that the 3D numerical simula
tion will be able to simulate accurately a large range of parametric lateral spreading problems.
The 3D analyses cover the following range of parameters:
o Relative Density of liquefiable soil layer Dr = 40 to 85%
o Thickness of liquefiable soil layer liq = 6 to 8m
o Pile stiffness EI = 4 to 16500 GPam4
o Pile diameter = 0.3 to 2.0m
o Ground surface displacement Dh = 0.1 to 0.3m.
Those analyses concern exclusively 2layered soil profiles, which is the geometry that has the
biggest difficulties in the simulation, since the developed moments and displacements are exclu
sively due to the forces of the liquefied soil.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
109
P
ult
= f (p')
P
y
K
Constitutive
model based
on CS Theory
Pile simulated
as a beam
following an
elastic law
(a)
Constitutive
model based
on CS Theory
Pile simulated
with a simple
elastic model
Figure 5. Typical grids (a) 2dimension and (b) 3dimension numerical simulation of single pile under
lateral spreading
(b)
The results of both 2D and 3D dynamic numerical analyses, agree in general with the design
charts which were developed by the Py methodology and have already been presented (Figures
2, 3 and 4). Differences were only encountered concerning the design chart for the evaluation of
the pile displacement in the case of the 2layered soil geometry with free pile head (Figure 2a).
For this case, Figure 6a presents the results of the 2D and 3D numerical analyses together with
the previously presented design chart. As it can be seen, the numerical analyses gave systemati
cally smaller displacements and thus the design chart presented in Figure 6b should be pre
ferred.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
110
0.01 0.1 1 10 100
(D
h*
EI)/(H
l i q
6
D
2
)
0.01
0.1
1
10
100
(
D
p
i
l
e
*
E
I
)
/
(
H
l
i
q
6
D
2
)
FLAC 2D Dr=45%
FLAC 3D Dr=45%
FLAC 3D Dr=65%
FLAC 3D Dr=85%
0.01 0.1 1 10 100
(D
h*
EI)/(H
l i q
6
D
2
)
0.01
0.1
1
10
100
(
D
p
i
l
e
*
E
I
)
/
(
H
l
i
q
6
D
2
)
FLAC 2D Dr=45%
FLAC 3D Dr=45%
FLAC 3D Dr=65%
FLAC 3D Dr=85%
D
r
= 50%
Dr
= 65%
D
r
= 85%
D
r
= 40%
D
r
=
4
0
%
D
r
=
6
5
%
D
r
=
8
5
%
(a) (b)
Figure 6. (a) Comparison between the 2D and 3D predictions for the 2layered soil profile and the design
chart from the Py analyses (b) Final design chart for the ground surface displacement in 2layered soil pro
files
6 CONCLUSIONS
In the previous paragraphs, diagrams and relations were presented for the approximate evalua
tion of the maximum displacement and bending moment of the pile due to liquefactioninduced
lateral spreading. The charts concern three different combinations of pile and ground conditions,
often encountered in practice. The proposed design charts and relations should be used with the
following limitations:
(a) They were derived pseudostatically, taking only into account the final displacement of
the ground, at the end of shaking. Any effects of the superstructure inertia are ignored.
(b) The expected freefield maximum ground surface displacement should be computed in
dependently, based on the (many) available empirical relations which are published in the litera
ture (e.g. Hamada, 1999, Youd et al, 2002, Valsamis, 2009).
(c) All the above mentioned charts and relations, and more specifically those concerning
the 2layered soil profile case, should be applied only when the soil has the capability to flow
freely around the pile under investigation. In all other cases (e.g. small distance between the
piles, sheetpile wall, etc) they may lead to nonconservative predictions of the pile displace
ment and bending moment.
(d) It has been assumed that the pile has been adequately embedded to the nonliquefiable
base soil layer so as to guarantee fixed bottom conditions during lateral ground spreading. When
the pile has not been driven adequately to the bottom (nonliquefiable) soil layer, there is the
possibility of pile extortion or significant pile base rotation which results to larger displacements
for the pile head and smaller developing moments.
7 REFERENCES
Abdoun T. H. (1999) Modeling of seismically induced lateral spreading of multilayered soil and its ef
fect on pile foundations, PHD Thesis, Rensselaer Polytechnic Institute, Troy, New York
Andrianopoulos, K.I. (2006), Numerical modeling of static and dynamic behavior of elastoplastic soils,
Doctorate Thesis, Department of Geotechnical Engineering, School of Civil Engineering, National
Technical University of Athens (in Greek).
Andrianopoulos, K.I., Papadimitriou, A.G. and Bouckovalas, G.D. (2006), Implementation of a bound
ing surface model for seismic response of sands, Proceedings of the 4th International FLAC Sympo
sium on Numerical Modeling in Geomechanics, Madrid, Spain
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
111
Andrianopoulos, K.I., Papadimitriou, A.G. and Bouckovalas, G.D. (2007), Use of a new bounding sur
face model for the analysis of earthquakeinduced liquefaction phenomena, paper no 1443, Proceed
ings of 4th International Conference on Earthquake Geotechnical Engineering.
API (1995), Recommended practice for planning, designing and constructing fixed offshore platform,
Washington, DC: American Petroleum Institute.
API (2002), Recommended practice for planning, designing and constructing fixed offshore platform,
Washington, DC: American Petroleum Institute.
Ashford S. A. & Juirnarongrit T. (2004), Evaluation of force based and displacement based analyses for
responses of single piles to lateral spreading, 11
th
International conference on Soil dynamics &
earthquake engineering, 3
rd
International conference on earthquake geotechnical engineering, 79
January 2004, Berkeley
Bhattacharya S. (2003), Pile instability during earthquake liquefaction, PHD Thesis, University of
Cambridge, UK.
Boulanger R.W., Kutter B.L., Brandenberg S.J., Singh P. and Chang D. (2003), Pile foundations in liq
uefied and lateral spreading ground during earthquakes: Centrifuge experiments and analyses Report
No. UCD/CGM03/01, Univ. of California at Davis.
Boulanger R.W., Wilson D.W., Kutter B.L. and Abghari, A. (1997), "Soilpilesuperstructure interaction
in liquefiable sand", Transportation Research Record No. 1569, TRB, NRC, National Academy Press,
5564
Brandenberg S.J. (2002), Behavior of Pile Foundations in Liquefied and Laterally Spreading Ground,
PHD Thesis, University of California, Davis
Cubrinovski M, Kokusho T. & Ishihara K. (2004), Interpretation from LargeScale Shake Table Tests
on Piles subjected to Spreading of Liquefied Soils, 11
th
International conference on Soil dynamics &
earthquake engineering, 3
rd
International conference on earthquake geotechnical engineering, 79
January 2004, Berkeley
Cubrinovsky M. , T. Kokusho and K. Ishihara (2006), Interpretation from large scale shake table tests
on piles undergoing lateral spreading in liquefied soils Soil Dynamics and Earthquake engineering,
vol.26
Dobry, R., Abdoun, T., O Rourke T.D., Goh S.H. (2003), Single piles in lateral spreads: Field Bending
Moment Evaluation, ASCE Journal of Geotechnical and Geoenvironmental Engineering, Vol. 129,
No. 10, October, pp. 879889
Hamada M. (1999), Similitude law for liquefiedground flow, Proceedings of the 7th U.S.Japan
Workshop on Earthquake Resistant design of lifeline facilities and countermeasures against soil lique
faction, pp. 191205.
High Pressure Gas Safety Institute of Japan (2000), Design method of foundation for Level 2 earthquake
motion, (In Japanese)
Ishihara K. & Cubrinovski M. (1998), Soilpile interaction in liquefied deposits undergoing lateral
spreading, XI DanubeEuropean Conference, Croatia, May 1998
Itasca (2005), FLAC version 5.0: Fast Langrangian Analysis of Continua, Itasca Consulting Group,
Minneapolis, Minnesota.
Itasca (1997), FLAC3D version 2.0: Fast Langrangian Analysis of Continua in 3 Dimentions, Itasca
Consulting Group, Minneapolis, Minnesota.
Japan Road Association (1996), "Specifications for highway bridges", Part V Seismic Design
Karamitros, D.K. (2009), Development of a numerical algorithm for the dynamic elastoplastic analysis
of geotechnical structures in two (2) and three (3) dimensions, PHD thesis , N.T.U.A.
The MacNealSchwendler Corporation (1994), MSC/NASTRAN for Windows: Reference Manual
Matlock, H. (1970). Correlations of design of laterally loaded piles in soft clay. Proc. Offshore Tech
nology Conference, Houston, TX, Vol 1, No.1204, pp. 577594.
Papadimitriou A., Bouckovalas G. and Dafalias Y. (2001), A plasticity model for sand under small and
large cyclic strains, Journal of Geotechnical and Geoenviromental Engineering, Vol.127, No. 11
Railway Technical Research Institute (1999), Earthquake resistant design code for railway structures,
Maruzen Co. (in Japanese)
Reese L.C. and Van Impe W. F. (2001), "Single piles and pile groups under lateral loading", A.A.
Balkema/Rotterdam/Brookfield, Book p.p.463.
Rollins K.M, Gerber T.M., Lane J.D. and Ashford S.A. (2005), "Lateral resistance of a fullscale pile
group in liquefied sand", ASCE Journal of Geotechnical and Geoenviromental Engineering, Vol 131,
No. 1, January, pp. 115125.
Rollins K.M., Bowles S., Brown D. & Ashford S. (2007), Lateral load testing of large drilled shafts af
ter blastinduced liquefaction, 4
th
International Conference on Earthquake Geotechnical Engineering,
Paper no 1141, June 2528, Thessaloniki, Greece
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
112
Valsamis A., Bouckovalas G. & Dimitriadi V., (2007), Numerical evaluation of lateral spreading dis
placements in layered soils, 4
th
International Conference on Earthquake Geotechnical Engineering,
Thessaloniki, Greece, June 2528
Valsamis (2008), Numerical simulation of single pile response under liquefactioninduced lateral
spreading, Doctorate Thesis, Department of Geotechnical Engineering, School of Civil Engineering,
National Technical University of Athens.
Youd L. T., Hansen M. C. and Bartlett F. S. (2002), "Revised multilinear regression equations for predic
tion of lateral spread displacement", Journal of Geotechnical and Geoenvironmental Engineering, Vol.
128, No. 12, December 1, pp 10071017.
Youd L. T., Idriss I. M., Andrus R.D., Arango I., Castro G., Christian J.T., Dobry R., Finn W.D.L.,
Harder L. F. jr, Hynes M. E., Ishihara K., Koester J. P., Liao S.S.C., Marcuson W.F. III, Martin G.R.,
Mitchell J. K., Moriwaki Y., Power M.S., Robertson P.K., Seed R. B. and Stokoe K.H. II (2001),
Liquefaction resistance of soils: summary report from the 1996 NCEER and 1998 NCEER/NSF
Workshops on evaluation of liquefaction resistance of soils, Journal of Geotechnical and Geoenvi
ronmental Engineering, Vol. 127, No. 10, October, pp 817833.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
113
1 INTRODUCTION
Pile foundations may sustain significant damage due to soil liquefaction, especially when soil
flow occurs. The problem became intensely apparent during the 1995 Great Kobe Earthquake.
Extensive soil flow was triggered along river and sea sides, causing seaward displacements to
sheet pile quaywalls in the Kobe port area accompanied by considerable translation of the
neighboring pile foundations (Ishihara, 1997, Tokimatsu et al., 1997, Yasuda et al., 1996), illus
trated in Figure 1.
soilflow soilflow
Figure 1. Soil liquefaction along sea sides may cause significant seaward displacement to quay walls af
fecting the neighboring pile foundations.
Numerous numerical and experimental researches has been performed during the last decade
in order to figure out the mechanism of soilpile interaction under soil flow conditions. The
moving soil mass provides the driving force to the pile and displaces the pile a certain amount
Piles in LiquefactionInduced Soil Flow behind QuayWall: A
Simple Physical Method versus Centrifuge Experiments
P. Tasiopoulou
Research Ast., University of California, Davis, USA
N. Gerolymos
Lecturer, National Technical University, Athens, Greece
T. Tazoh
Director, Institute of Technology, Shimizu Corporation, Tokyo, Japan
G. Gazetas
Professor, National Technical University, Athens, Greece
ABSTRACT: The paper presents a new physically simplified methodology for computing dis
placements and internal forces on piles under conditions of lateral spreading. The results com
pare well with results from centrifuge tests. To this end, 2D effective stress dynamic analysis of
a crosssection of the wallsoil system without the presence of the piles is combined with an also
2D quasistatic analysis of a horizontal slice of the system with the group of piles.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
114
depending on the relative stiffnesses between the pile and the liquefied soil (Boulanger et al.,
2003), as depicted in Figure 2. Thus, the magnitude of the soil movement, the lateral load of the
surficial nonliquefiable soil layer and the stiffness degradation in the liquefied zone are the key
parameters that need to be taken into account when evaluating the pile response due to soil flow
(Cubrinovsky et al., 2004).
Beforeliquefaction Afterliquefaction
flow
Beforeliquefaction Afterliquefaction
flow
K
Figure 2. The moving soil mass provides the driving force to the pile and displaces the pile a certain
amount depending on the relative stiffnesses between the pile, K, and the liquefied soil.
In engineering practice, several methods have been formulated based on these understand
ings, either for designing purposes or prediction of field performance. In general, the methods
can be classified into three categories: (a) the force methods, including the Japanese Road Asso
ciate Method (JRA, 1996), the limit equilibrium method (Dobry and Abdoun, 2000) and the vis
cous fluid method (Hamada 2002, Yasuda 2002), (b) the displacement methods, or else known
as pseudostatic beam on nonlinear Winkler foundation method according to which the free
field soil displacement is imposed to the pile through py springs (Boulanger et al., 2003), (c)
the hybrid forcedisplacement methods, which are a combination of the first two (Cubrinovsky
and Ishihara, 2004).
The aforementioned methods are mainly single pile analyses dependent on the soil profile and
the geometry of the problem. Moreover, lots of assumptions are required regarding the stiffness
degradation in the liquefied layer and soilpile interaction issues, such as the direction of the
load exerted on the pile by the upper nonliquefiable layer. Inevitably, considerable uncertainty
is hidden behind all methods of post liquefaction analysis (Finn and Thavaraj, 2001).
In this study, we present a new physically simplified methodology, appropriate for every soil
profile and type of pile configuration (single piles, pile groups). This methodology, described
below, falls into the displacementmethod category, but avoids the associated empirical selec
tion of stiffnessreduction factors and does not involve the use of py curves. Continuously, the
method is applied to two different centrifuge experiments (Tazoh et al., 2005, Sato et al., 2001),
reproducing the test results with satisfying accuracy for engineering purposes.
2 CENTRIFUGE EXPERIMENTS
Several series of dynamic centrifuge experiments were conducted at the Institute of Technol
ogy, Shimizu Corporation, in Japan, in order to evaluate the damage of pilefoundation systems
triggered by liquefactioninduced soil flow after quaywall collapse. One of these, Test Case
CD, presented by Tazoh et al. (2005), explores the effect of the superstructure on the response
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
115
of a 2x2 pile group under soil flow conditions, portrayed in Figure 3. A partition was placed at
the center of the laminar box behind the sheet pile quaywall in order to separate the two pile
foundation models: the one with a superstructure (C side) and the other without a superstructure
(D side), as depicted in Figure 3.
0 m
1.8 m
6.0 m
8.1 m
9 m
0.6 m
Dry SilicaSandNo8
Dr=50%
Liquefiable SilicaSandNo8
Dr=50%
Toyoura Sand
Dr=90%
SilicasandNo3
Sheetpile
Quaywall
24m
3.6 m
6m 3m Superstructure
Side D
with superstructure
Side C
no superstructure
Partition
Floating
Quaywall
0 m
1.8 m
6.0 m
8.1 m
9 m
0.6 m
Dry SilicaSandNo8
Dr=50%
Liquefiable SilicaSandNo8
Dr=50%
Toyoura Sand
Dr=90%
SilicasandNo3
Sheetpile
Quaywall
24m
3.6 m
6m 3m Superstructure
Side D
with superstructure
Side C
no superstructure
Partition
Floating
Quaywall
Figure 3. Geometry and soil properties of the centrifuge model of test Case CD, side D into the laminar
box, in prototype scale (Tazoh et al., 2005). A partition behind the quaywall separates the two pile
foundation models: the one with a superstructure (C side) and the other without a superstructure (D side).
The input motion at the base of the laminar box is shown in Figure 4 along with the recorded
time histories of excess pore water pressure in the liquefied layer at the almost free field. Liq
uefaction starts at around 4 sec and 5 sec at the depths of z = 3 m and z = 5 m, respectively. Fig
ure 5 depicts the horizontal displacements of the footings and the quaywall of the Case CD.
Evidently, the horizontal displacement of the Cside footing (with no superstructure) is larger
than that of the Dside footing, (with a superstructure), but the difference is not significant. The
effect of the inertial force of the superstructure can be identified during the excitation period;
however it can not be recognized during the liquefactioninduced soil flow some time after the
end of excitation. Figure 6 illustrates the distribution of the maximum bending strains along the
pile at the end of the shaking with and without the influence of the superstructure. It is evident
that the effect of the inertial force of the superstructure on the final pile strains is not significant.
Thus, the main load on the pile is the kinematic one coming from the soilflow.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
116
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
A:g
A:g
time(sec)
Tazohetal.(2005)
Satoetal.(2001)
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
A:g
A:g
time(sec)
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
A:g
A:g
time(sec)
Tazohetal.(2005)
Satoetal.(2001)
40
20
0
40
z=3m
t(sec)
z=5m
E
x
c
e
s
s
P
o
r
e
P
r
e
s
s
u
r
e
(
k
P
a
)
(a) (b)
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
A:g
A:g
time(sec)
Tazohetal.(2005)
Satoetal.(2001)
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
A:g
A:g
time(sec)
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
A:g
A:g
time(sec)
Tazohetal.(2005)
Satoetal.(2001)
40
20
0
40
z=3m
t(sec)
z=5m
E
x
c
e
s
s
P
o
r
e
P
r
e
s
s
u
r
e
(
k
P
a
)
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
A:g
A:g
time(sec)
Tazohetal.(2005)
Satoetal.(2001)
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
A:g
A:g
time(sec)
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
0.6
0.4
0.2
0
0.2
0.4
0.6
0 5 10 15
A:g
A:g
time(sec)
Tazohetal.(2005)
Satoetal.(2001)
40
20
0
40
z=3m
t(sec)
z=5m
E
x
c
e
s
s
P
o
r
e
P
r
e
s
s
u
r
e
(
k
P
a
)
(a) (b)
Figure 4. (a) Input accelerograms recorded at the base of the centrifuge models of the two experiments
(Tazoh et al., 2005 and Sato et al., 2001) in prototype scale, and (b) Indicative time histories of excess
pore water pressures recorded in the free field (away from the pile foundation) at the depths 3 m and 5
m below the ground surface
0 5 10 15 400 800
0.8m
0.1 m
0.15 m
Quaywall
With superstructure
No supersructure
t(sec)
D
i
s
p
l
a
c
e
m
e
n
t
Figure 5. Time histories of displacements of the quay wall and the footing with and without the super
structure (Tazoh et al., 2005).
Similar centrifuge results with the same soil profile were described by Sato et al. (2001) indi
cating the seismic performance of a 2x8 pile group, situated 3 m behind the floating sheet pile
quaywall. Two different centrifuge models were designed: one with (Case 2) and one without
the pile group (Case 1). Figure 4 shows the input wave recorded at the base of the model of
Case 1, which caused excessive pore water pressure generation in the loose saturated sand layer
and maximum seaward displacement of the quaywall of 0.8 m during the shaking. In Case 2,
the existence of the pile group limited the quaywall movement to 0.45 m approximately. The
footing of the pile foundation sustained even smaller displacement, about 6 cm. Long after the
end of shaking, when considerable dissipation of pore water pressure has occurred the quay
wall reached a displacement of 1.15 m, while the footing displacement increased by a mere 2
cm. The trend of the timedependent results is practically identical with that of the results of
Test Case CD.
3 A NEW PHYSICALLY SIMPLIFIED METHODOLOGY
To begin with, the soil response without the piles needs to be identified. On these grounds, a
2D effective stress numerical analysis of the soil profile including the quaywall hereafter
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
117
called free field is performed using the code FLAC, as depicted in Figure 7. Thus, the free
field soil response behind the quaywall is obtained in terms of:
the distribution of the horizontal soil movement with depth,
the depth and the thickness of the liquefied zone,
the strength degradation in the liquefied layer and
the shear strain distribution with depth.
0
1
2
3
4
5
6
7
8
9
0.7 0.5 0.3 0.1 0.1 0.3 0.5
bendingstrains
(10
3
)
d
e
p
t
h
(
m
)
Cside(NOsuperstructure)
Dside(WITHsuperstructure)
Figure 6. Distributions of the maximum bending strains along the piles at the end of shaking with and
without a superstructure. The influence of the superstructure in terms of bending distress of the piles is
not as significant.
Although the dynamic numerical analysis provides the required results as a function of time,
we are primarily interested in the final values after the end of the shaking, which are also the
maximum ones due to the accumulative nature of the liquefactioninduced soil flow. Obviously
our methodology is not restricted to this code.
In the next step, a horizontal slice in the middle of the liquefied zone is isolated, including the
piles and the quaywall, as demonstrated in Figure 8. Our purpose is to perform an elastic plane
strain analysis of this horizontal slice by imposing pseudostatically a unit uniform displacement
at the quaywall boundary, illustrated in Figure 9, so as to estimate the ratio of the pile dis
placement to the soil displacement in the freefield (away enough of the piles), named ratio
and depicted in the same figure. This ratio of displacements represents the soilpile interaction
due to soil flow, in quantitative terms.
The numerical model of the horizontal slice consists of the liquefied soil with uniform prop
erties (shear modulus, G
l
) surrounding the pile sections, as shown in Figure 9. So far, the pile
sections are simulated as rigid bodies into the liquefied soil, G
l
. In order to provide the required
horizontal resistance to the pile sections against the moving soil mass, an out of plane horizontal
spring, K, is connected to each pile section.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
118
Sheetpile
Quaywall
A: g
Sheetpile
Quaywall
Sheetpile
Quaywall
A: g
Sheetpile
Quaywall
Figure 7. 2D Finitedifference mesh of a numerical model of the free field without the piles, in FLAC,
before and after shaking.
G
l
G
l
HorizontalSlice
Largedisplacement
androtation
ofquaywall
G
l
G
l
HorizontalSlice
Largedisplacement
androtation
ofquaywall
Figure 8. A horizontal slice in the middle of the liquefied zone, G
l
, is isolated, including the pile and the
quaywall sections.
In retrospect, a realistic numerical simulation of the horizontal slice at the middle of the liq
uefied zone requires the appropriate calibration of the horizontal stiffness of each pile section,
K, and of the shear modulus of the liquefied soil, G
l
.
Horizontal stiffness of each pile section, K
Every single pile of the pile group is simulated as a vertical beam element with suitable
boundary conditions. The rotation at the top depends on the pile cap (Mokwa and Duncan,
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
119
2003), the number of piles of the foundation and the stiffness of the surrounding surficial non
liquefiable soil. All these features tend to restrain the rotation at the pilehead.
G
l
Unit
displacement
Pilecapconstraint
pilesection
Q
u
a
y
w
a
l
l
liquefiedsoil
G
l
Unit
displacement
Pilecapconstraint
pilesection
Q
u
a
y
w
a
l
l
liquefiedsoil
Freefielddisplacement
Disp.1
Pilegroupdisplacement
Disp.2
liquefiedsoil
Ratio =
Disp.2
Disp.1
G
l
Unit
displacement
Pilecapconstraint
pilesection
Q
u
a
y
w
a
l
l
liquefiedsoil
G
l
Unit
displacement
Pilecapconstraint
pilesection
Q
u
a
y
w
a
l
l
liquefiedsoil
Freefielddisplacement
Disp.1
Pilegroupdisplacement
Disp.2
liquefiedsoil
Ratio =
Disp.2
Disp.1
Ratio =
Disp.2
Disp.1
Ratio =
Disp.2
Disp.1
G
l
Unit
displacement
Pilecapconstraint
pilesection
Q
u
a
y
w
a
l
l
liquefiedsoil
G
l
Unit
displacement
Pilecapconstraint
pilesection
Q
u
a
y
w
a
l
l
liquefiedsoil
Freefielddisplacement
Disp.1
Pilegroupdisplacement
Disp.2
liquefiedsoil
Ratio =
Disp.2
Disp.1
G
l
Unit
displacement
Pilecapconstraint
pilesection
Q
u
a
y
w
a
l
l
liquefiedsoil
G
l
Unit
displacement
Pilecapconstraint
pilesection
Q
u
a
y
w
a
l
l
liquefiedsoil
Freefielddisplacement
Disp.1
Pilegroupdisplacement
Disp.2
liquefiedsoil
Ratio =
Disp.2
Disp.1
Ratio =
Disp.2
Disp.1
Ratio =
Disp.2
Disp.1
Figure 9. (Top) Numerical model of the horizontal slice. The pile sections sustain the same horizontal
displacement due to the pilecap constraint. (Bottom) Definition of the ratio .
The active length of the beam element depends on the depth to fixity, below the liquefied
layer. Therefore, every single pile is simulated as a vertical beam fixed at the bottom with a cer
tain degree of rotational freedom at the top depending on the aforementioned kinematic con
straints.
Based on this simulation, the pile section of the numerical model is just a section of the beam
element at a characteristic depth at the middle of the liquefied zone. Thus, the horizontal stiff
ness, K, of each pile section, is defined as the point load exerted on the beam, in order to cause a
unit displacement of the beam at the characteristic depth, as depicted in Figure 10.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
120
0.0
0.2
0.4
0.6
0.8
1.0
0.001 0.01 0.1 1 10 100 1000
K/EI
z
/
L
0.0
0.2
0.4
0.6
0.8
1.0
0.001 0.01 0.1 1 10 100 1000
K/EI
z
/
L
0.0
0.2
0.4
0.6
0.8
1.0
0.001 0.01 0.1 1 10 100 1000
K/EI
z
/
L
0.0
0.2
0.4
0.6
0.8
1.0
0.001 0.01 0.1 1 10 100 1000
K/EI
z
/
L
Figure 10. Dimensionless horizontal stiffness along the pile for free to fixed boundary conditions at the
pile head.
0.0
0.2
0.4
0.6
0.8
1.0
0.1 1 10 100
K/ G
l
r
a
t
i
o
"
"
1
0.8
0.6
0.4
0.2
0
0.1 1 10 100
singlepile
2x2pilegroup
2x8pilegroup
0.0
0.2
0.4
0.6
0.8
1.0
0.1 1 10 100
K/ G
l
r
a
t
i
o
"
"
1
0.8
0.6
0.4
0.2
0
0.1 1 10 100
singlepile
2x2pilegroup
2x8pilegroup
Figure 11. The ratio as a function of the relative stiffnesses between the pile and the liquefied soil for
three different pile configurations: single pile, 2x2 and 2x8 pile group, obtained from several parametric
numerical analyses of the horizontal slice.
Shear modulus of the liquefied soil, G
l
In the framework of an elastic analysis of the horizontal slice, an equivalent linear shear
modulus of the liquefied soil can be determined as:
0
1
liq
H
res
l liq
res
G dz
H
(1)
H
liq
is the thickness of the liquefied zone, is the residual shear stress and is the maximum
shear strain after the end of shaking, obtained from the numerical analysis of the free field. That
is how the stiffness degradation of the liquefied soil is taking into account.
The elastic pseudostatic analysis of the horizontal is conducted for two different pile con
figurations: (a) a 2x2 pile group (Tazoh et al., 2005) and (b) a 2x8 pile group (Sato et al., 2003).
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
121
0
0.2
0.4
0.6
0.8
1
0 0.5 1
w/ ( pL EI w
max
)
z
/
L
uniformload
invertedtriangleload
invertedtrapezoidload
0
0.2
0.4
0.6
0.8
1
0 0.5 1
w/ ( pL EI w
max
)
z
/
L
uniformload
invertedtriangleload
invertedtrapezoidload
Figure 12. Deformation shape of a pile with no rotation at the top and the bottom, determined for three
different load distributions: uniform, inverted triangle shaped and inverted trapezoid one.
The ratio , is obtained as function of the relative stiffness between the pile and the lique
fied soil, K/G
l
, shown in Figure 11.
When the relative stiffness tends to zero (K/G
l
0), the ratio tends to unity, which
means that the pile sections move just like the soil, as a rigid body. On the contrary, when the
relative stiffness tends to infinity (K/G
l
), the ratio tends to zero. This is due to the fact
that the soil has practically zero shear strength (G
l
0) and flows around the pile without exert
ing any significant load on them. Moreover, Figure 11 indicates that increasing the number of
piles, the resistance of the foundation to the moving soil mass becomes stronger. It is worth
mentioned at this point that the numerical modeling of the pilegroup section into the liquefied
soil is based on the assumption that the piles sustain the same horizontal displacement due to the
pilecap constraint.
In conclusion, as long as the relative stiffness, K/G
l
, is determined, the ratio can be esti
mated. Eventually, multiplying the ratio with the freefield soil displacement in the middle
of the liquefied zone, at the position where the piles would be present behind the quaywall, we
can calculate the pile displacement at the same characteristic depth.
In the last step of the methodology, the whole pile deformation and mainly the pile displace
ment at the top, remains to be evaluated. The deformation shape of each pile, simulated as a
beam element, is defined primarily by its boundary conditions and secondarily by the load dis
tribution along it, as depicted in Figure 12. A potential load distribution, is imposed on the pile,
so as to obtain its deformation shape as a function of the unknown load value, p.
Until now, we have just estimated the pile displacement in the middle of the liquefied zone.
Using this known displacement, we can calibrate the shape function of the pile with depth and
finally, estimate the unknown load value, p. Continuously, the distribution of the pile dis
placements with depth, as long as the bending moments along the pile, can be determined by
imposing the already fully known load distribution on the beampile.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
122
0
1
2
3
4
5
6
7
8
9
15 10 5 0 5 10 15 20
shearstress(kPa)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
0 2 4 6 8 10 12 14 16
shearstrain(10
2
)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
15 10 5 0 5 10 15 20
shearstress(kPa)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
0 2 4 6 8 10 12 14 16
shearstrain(10
2
)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
freefielddisplacement(m)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
15 10 5 0 5 10 15 20
shearstress(kPa)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
0 2 4 6 8 10 12 14 16
shearstrain(10
2
)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
15 10 5 0 5 10 15 20
shearstress(kPa)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
0 2 4 6 8 10 12 14 16
shearstrain(10
2
)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
freefielddisplacement(m)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
(i)
(ii)
0
1
2
3
4
5
6
7
8
9
15 10 5 0 5 10 15 20
shearstress(kPa)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
0 2 4 6 8 10 12 14 16
shearstrain(10
2
)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
15 10 5 0 5 10 15 20
shearstress(kPa)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
0 2 4 6 8 10 12 14 16
shearstrain(10
2
)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
freefielddisplacement(m)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
15 10 5 0 5 10 15 20
shearstress(kPa)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
0 2 4 6 8 10 12 14 16
shearstrain(10
2
)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
15 10 5 0 5 10 15 20
shearstress(kPa)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
0 2 4 6 8 10 12 14 16
shearstrain(10
2
)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
0
1
2
3
4
5
6
7
8
9
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
freefielddisplacement(m)
d
e
p
t
h
(
m
)
Tazohetal.(2005)
Satoetal.(2001)
(i)
(ii)
Figure 13. Distributions of horizontal soil displacements, shear strains and stresses, obtained from the
numerical analysis of the two models: (i) Tazoh et al., 2005 (Case CD) and (ii) Sato et al. (2001) without
the piles.
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0 2 4 6 8 10 12
time(sec)
Q
u
a
y
w
a
l
l
d
i
p
l
a
c
e
m
e
n
t
a
t
t
o
p
(
m
)
Tazohetal.(2005)
Satoetal.(2005)
(i)
(ii)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0 2 4 6 8 10 12
time(sec)
Q
u
a
y
w
a
l
l
d
i
p
l
a
c
e
m
e
n
t
a
t
t
o
p
(
m
)
Tazohetal.(2005)
Satoetal.(2005)
(i)
(ii)
Figure 14. Horizontal quay wall displacements, obtained from the numerical analysis of the two models:
Tazoh et al., 2005 (Case CD) and Sato et al. (2001) without the piles.
4 COMPARISON WITH RESULTS FROM CENTRIFUGE TESTS
The results obtained from the numerical analysis of the free field including the quaywall in
the first step for the two centrifuge experiments (Tazoh et al., 2005 and Sato et al., 2001) are il
lustrated in Figures 13 to 16. The only differences between the two models, both the centrifuge
and the numerical ones are the input wave motions, shown in Figure 4 and the distance of the
quay wall from the boundaries parallel to it, depicted in Figure 3. According to the excess pore
pressure time histories, demonstrated in Figures 15 and 16, the liquefaction seems more exten
sive in case of Tazoh et al. model. This is one of the reasons why the quaywall displacement of
this model is larger than the one of the Sato et al. model, as portrayed in Figure 14.
In the second step, every single pile of the pile groups is simulated as a beam fixed both at
the top and the bottom, following the assumption that the pile cap does not allow any rotation at
the pile heads. The horizontal stiffnesses, K, are calculated at the characteristic depths depicted
in Figure 17 which coincide with the middle of the liquefied layer. The values required to esti
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
123
mate the pile displacement at the characteristic depth are given in table 1. Once the relative
stiffness is estimated for each model, the ratio is determined, choosing the 2x2 pilegroup
curve in case of the Tazoh et al. model and the 2x8 pilegroup curve for the Sato et al. model.
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12
time(sec)
E
x
c
e
s
s
p
o
r
e
p
r
e
s
s
u
r
e
r
a
t
i
o
z=4m
z=3.5m
z=5.5m
z=6m
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12
time(sec)
E
x
c
e
s
s
p
o
r
e
p
r
e
s
s
u
r
e
r
a
t
i
o
z=4m
z=3.5m
z=5.5m
z=6m
Figure 15. Time histories of excess pore pressure ratio at depths of 4, 3.5, 5.5, 6 m below the ground sur
face, obtained from the numerical analysis of the Sato et al. model (2001) without the piles.
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12 14
time(sec)
E
x
c
e
s
s
p
o
r
e
p
r
e
s
s
u
r
e
r
a
t
i
o
z=4m
z=3.5m
z=5.5m
z=6m
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12 14
time(sec)
E
x
c
e
s
s
p
o
r
e
p
r
e
s
s
u
r
e
r
a
t
i
o
z=4m
z=3.5m
z=5.5m
z=6m
Figure 16. Time histories of excess pore pressure ratio at depths of 4, 3.5, 5.5, 6 m below the ground sur
face, obtained from the numerical analysis of the Tazoh et al. model (2005) without the piles.
In the last step, five different potential load distributions are imposed on the pilebeam. A
range of estimated values regarding the displacements and the bending moments along each pile
of the pile group are illustrated in Figure 18 for the two models. The range of pile displacements
is not very sensitive to the shape of the load distribution whereas the bending moments vary
significantly, especially on the top. However, the maximum bending moment for both of the
two models was recorded at the pile tip, during the centrifuge experiments.
5 CONCLUDING REMARKS
In this paper, a new simple physicallymotivated methodology is proposed for the evaluation
of pile response due to liquefactioninduced soil flow. The main characteristics of this method
ology verified above are:
It avoids the associated empirical selection of stiffnessreduction factors and does not in
volve the use of py curves.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
124
1.3 m
L=7.5m 9 m
Liquefiedlayer
2 m
6 m
3 m
P
1.3 m
L=7.5m L=7.5m 9 m
Liquefiedlayer
2 m
6 m
3 m
P
EI=17000kNm
2
0.6 m
L=8.0m 9 m
Liquefiedlayer
2 m
6 m
3.5 m
P
0.6 m
L=8.0m L=8.0m 9 m
Liquefiedlayer
2 m
6 m
3.5 m
P
EI=6480kNm
2
1.3 m
L=7.5m 9 m
Liquefiedlayer
2 m
6 m
3 m
P
1.3 m
L=7.5m L=7.5m 9 m
Liquefiedlayer
2 m
6 m
3 m
P
EI=17000kNm
2
0.6 m
L=8.0m 9 m
Liquefiedlayer
2 m
6 m
3.5 m
P
0.6 m
L=8.0m L=8.0m 9 m
Liquefiedlayer
2 m
6 m
3.5 m
P
EI=6480kNm
2
1.3 m
L=7.5m 9 m
Liquefiedlayer
2 m
6 m
3 m
P
1.3 m
L=7.5m L=7.5m 9 m
Liquefiedlayer
2 m
6 m
3 m
P
EI=17000kNm
2
0.6 m
L=8.0m 9 m
Liquefiedlayer
2 m
6 m
3.5 m
P
0.6 m
L=8.0m L=8.0m 9 m
Liquefiedlayer
2 m
6 m
3.5 m
P
EI=6480kNm
2
1.3 m
L=7.5m 9 m
Liquefiedlayer
2 m
6 m
3 m
P
1.3 m
L=7.5m L=7.5m 9 m
Liquefiedlayer
2 m
6 m
3 m
P
EI=17000kNm
2
0.6 m
L=8.0m 9 m
Liquefiedlayer
2 m
6 m
3.5 m
P
0.6 m
L=8.0m L=8.0m 9 m
Liquefiedlayer
2 m
6 m
3.5 m
P
EI=6480kNm
2
Figure 17. Each pile of the 2x8 pile group of the Sato et al. (2001)model (Top) and of the 2x2 pile group
of the Tazoh et al. (2005) model (Case CD  Bottom), is simulated as a vertical beam element with no ro
tation at the top and the bottom. The horizontal stiffness, of the pile is estimated at a characteristic depth
of 3 m (3.5 m for the Tazoh et al model) from its top, which coincides with the middle of the liquefied
layer.
Table 1. Table of the values needed to determine the ratio and the pile displacements at the middle of
the liquefied layer for the two models:
depth
(m)
G
l
(kPa)
K
(kN/m)
K /G
l
ffdisp
(m)
ratio
""
piledisp.
(m)
3.50 90 370 4.1 0.5 0.152 0.076
depth
(m)
G
l
(kPa)
K
(kN/m)
K /G
l
ffdisp
(m)
ratio
""
piledisp.
(m)
3.00 200 1020 5.1 0.42 0.105 0.044
Tazohetal.
(2005)
Satoetal.
(2001)
It introduces an elastic pseudo static numerical analysis of a horizontal slice into the lique
fied layer including the piles, in order to estimate the soilpile interaction under soil flow
conditions in quantitative terms. This interaction is determined as a function of the relative
stiffness between the pile and the liquefied soil. The pile stiffness is assumed to remain lin
earelastic throughout the analysis.
It can be applied to any type of soil profile and pile configuration (single pile or pile
group).
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
125
The effectiveness of the new methodology in combination with suitable engineering judg
ment and reasonable assumptions can provide sufficient accuracy for designing and evalu
ating purposes.
0
1
2
3
4
5
6
7
0 0.02 0.04 0.06 0.08
pilediplacement(m)
d
e
p
t
h
(
m
)
Centrifugeexperiment
Rangeofpredictedvalues
0
1
2
3
4
5
6
7
0 0.02 0.04 0.06 0.08
pilediplacement(m)
d
e
p
t
h
(
m
)
0
1
2
3
4
5
6
7
0 0.02 0.04 0.06 0.08
pilediplacement(m)
d
e
p
t
h
(
m
)
Centrifugeexperiment
Rangeofpredictedvalues
Centrifugeexperiment
Rangeofpredictedvalues
0
1
2
3
4
5
6
7
600 400 200 0 200 400 600
bendingstrains()
d
e
p
t
h
(
m
)
Centrifugeexperiment
Rangeofpredictedvalues
0
1
2
3
4
5
6
7
600 400 200 0 200 400 600
bendingstrains()
d
e
p
t
h
(
m
)
Centrifugeexperiment
Rangeofpredictedvalues
Centrifugeexperiment
Rangeofpredictedvalues
0
1
2
3
4
5
6
7
8
0 0.05 0.1 0.15
pilediplacement(m)
d
e
p
t
h
(
m
)
Centrifugeexperiment
Rangeofpredictedvalues
0
1
2
3
4
5
6
7
8
0 0.05 0.1 0.15
pilediplacement(m)
d
e
p
t
h
(
m
)
0
1
2
3
4
5
6
7
8
0 0.05 0.1 0.15
pilediplacement(m)
d
e
p
t
h
(
m
)
Centrifugeexperiment
Rangeofpredictedvalues
Centrifugeexperiment
Rangeofpredictedvalues
0
1
2
3
4
5
6
7
8
1500 1000 500 0 500 1000 1500
bendingstrains()
d
e
p
t
h
(
m
)
Centrifugeexperiment
Rangeofpredictedvalues
Centrifugeexperiment
Rangeofpredictedvalues
0
1
2
3
4
5
6
7
0 0.02 0.04 0.06 0.08
pilediplacement(m)
d
e
p
t
h
(
m
)
Centrifugeexperiment
Rangeofpredictedvalues
0
1
2
3
4
5
6
7
0 0.02 0.04 0.06 0.08
pilediplacement(m)
d
e
p
t
h
(
m
)
0
1
2
3
4
5
6
7
0 0.02 0.04 0.06 0.08
pilediplacement(m)
d
e
p
t
h
(
m
)
Centrifugeexperiment
Rangeofpredictedvalues
Centrifugeexperiment
Rangeofpredictedvalues
0
1
2
3
4
5
6
7
600 400 200 0 200 400 600
bendingstrains()
d
e
p
t
h
(
m
)
Centrifugeexperiment
Rangeofpredictedvalues
0
1
2
3
4
5
6
7
600 400 200 0 200 400 600
bendingstrains()
d
e
p
t
h
(
m
)
Centrifugeexperiment
Rangeofpredictedvalues
Centrifugeexperiment
Rangeofpredictedvalues
0
1
2
3
4
5
6
7
8
0 0.05 0.1 0.15
pilediplacement(m)
d
e
p
t
h
(
m
)
Centrifugeexperiment
Rangeofpredictedvalues
0
1
2
3
4
5
6
7
8
0 0.05 0.1 0.15
pilediplacement(m)
d
e
p
t
h
(
m
)
0
1
2
3
4
5
6
7
8
0 0.05 0.1 0.15
pilediplacement(m)
d
e
p
t
h
(
m
)
Centrifugeexperiment
Rangeofpredictedvalues
Centrifugeexperiment
Rangeofpredictedvalues
0
1
2
3
4
5
6
7
8
1500 1000 500 0 500 1000 1500
bendingstrains()
d
e
p
t
h
(
m
)
Centrifugeexperiment
Rangeofpredictedvalues
Centrifugeexperiment
Rangeofpredictedvalues
Figure 18. Calculated range of distributions of (a) the horizontal displacement, and (b) the bending strain,
in comparison with the centrifuge test results by Sato et al., 2001 (Top), and Tazoh et al., 2005 (Bottom).
ACKNOWLEDGMENTS
This work forms part of an EU 7
th
Framework research project funded through the European
Research Council (ERC) Programme Ideas, Support for Frontier Research Advanced Grant,
under Contract number ERC2008AdG 228254DARE.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
126
REFERENCES
Boulanger R. W., Kutter B. L., Brandenberg S. J., Singh P., and Chang P., (2003), Pile Foundations in
Liquefied and Laterally Spreading Ground during Earthquakes: Centrifuge Experiments and Analyses,
Report No. UCD/CGM03/01, Department of Civil and Environmental Engineering, University of
California at Davis.
Cubrinovsky M., and Ishihara K., (2004), Simplified Method for Analysis of Piles undergoing Lateral
Spreading of Liquefied Soils, Soils and Foundations, Vol. 44, No. 5, pp. 119133.
Cubrinovsky M., Kokusho T., and Ishihara K., (2004), Interpretation from largescale shake table tests
on piles subjected to spreading of liquefied soils, Proceedings of 11
th
Int. Conf. Soil Dynamics and
Earthq. Engrg. / 3
rd
Int. Conf. Earthq. Geotech. Engrg., Berkeley, USA, Vol. 2, pp. 463470.
Dobry R., and Abdoun T. H., (1998), PostTriggering Response of Liquefied Soil in The Free Field and
Near Foundations, Stateoftheart paper, Proc. ASCE 1998 Specialty Conference on Geotechnical
Earthquake Engineering and Soil Dynamics (P. Dakoulas, M. Yegian and R. D. Holtz, eds.), Univer
sity of Washington, Seattle, Washington, August 36, Vol. 1, pp. 270300.
Dobry R., and Abdoun T. H., (2000), Recent studies on seismic centrifuge modeling of liquefaction and
its effect on deep foundations, Proceedings of the fourth International Conference on Recent Ad
vances in Geotechnical Earthquake Engineering and Soil Dynamics , San Diego.
Finn Liam W. D., and Thavaraj T., (2001), Deep Foundations in liquefiable soils: Case Histories , cen
trifuge tests and methods of analysis , Proceedings of the fourth International Conference on Recent
Advances in Geotechnical Earthquake Engineering and Soil Dynamics, San Diego.
Ishihara K., (1997), Geotechnical aspects of the 1995 Kobe Earthquake, Proceedings of 14
th
Int. Conf.
SMFE, Terzaghi Oration, Hamburg.
Mokwa R. L., Duncan J. M (2003), Rotational restraint of pile caps during lateral loading, Journal of
Geotechnical and Geoenvironmental Engineering, ASCE, 129(9), pp. 829837.
Ramos R., Abdoun T. H., and Dobry, R., (2000), Effect of Lateral Stiffness of Superstructure on Bend
ing Moments of Pile Foundation Due to Liquefactioninduced Lateral Spreading, Proc. 12th World
Conf. on Earthquake Engineering, Auckland, New Zealand, Jan. 30  Feb. 4, 8 pages.
Sato M., Tazoh T. and Ogasawara M. (2001), Reproduction of Lateral Ground Displacement and Lateral
Flow Earth Pressure Acting on Pile Foundations using Centrifuge Modeling, Proceedings of the
Fourth International Conference on Recent Advances in Geotechnical Earthquake Engineering and
Soil Dynamics, San Diego.
Tazoh T., Gazetas G. (1996), Pile Foundations Subjected to Large Ground Deformations: Lessons from
Kobe and Research Needs, Proceedings of the Eleventh World Conference on Earthquake Engineer
ing, Acapulco, Mexico, paper 2081.
Tazoh T., Sato M., and Gazetas G., (2005), Centrifuge Tests on PileFoundationStructure Systems Af
fected by LiquefactionInduced Soil Flow after Quay Wall Failure, Proceedings of the 1
st
Greece
Japan Workshop: Seismic Design, Observation, and Retrofit of Foundations, Athens, Oct. 1112 pp.
79106.
Tokimatsu K., Ohoka H., Samoto Y., Nakazawa A., and Asaka Y., (1997), Failure and Deformation
Modes of Piles Caused by LiquefactionInduced Lateral Spreading in 1995 Hyogoken Nambu Earth
quake, Proceedings of the Third Kansai International Forum on Comparative Geotechnical Engi
neering, (KIGForum 97), pp. 239248.
Yasuda S., Ishihara K., Harada K., and Namura H., (1996), Area of Ground Flow Occurred behind
Quaywalls due to Liquefaction, Proceedings of the Third Kansai International Forum on Compara
tive Geotechnical Engineering, (KIGForum 97), pp. 8593.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
127
FINITE ELEMENT ANALYSIS OF PILESOIL INTERACTION
SYSTEM BY OVERLAYING MESH METHOD
A. Ohta & F. Miura
The graduate school of science and engineering, Yamaguchi University
Y. Ono & J. Kiyono
The graduate school of engineering, Kyoto University
1 INTRODUCTION
The Overlaying mesh method (OMM) is an analytical approach that overlaps two or more
independent differentsizemesh models. In the OMM, detailed mesh model is used in elected area
under consideration, with coarser mesh model else where, in order to optimize calculation effort. In
the previous study, differentsizemodels are used to express a complex area with different material
constants, but same type elements, such as plane strain two dimensional elements are used. In this
research, we propose a new application method of the overlaying mesh method using different
elements such as beam elements and solid elements. We analyzed two types of pile foundation
models using OMM, and proved that the proposed method is valid.
2 THEORY OF THE PROPOSED METHOD
2.1 Derivation of the fundamental equations for the OMM
In the OMM, two or more differentsizedmesh models are used, one is for modeling the wide area,
which we call Global area, the other/others is/are used to model detailed area(s), which we call Local
area, where we want to know the detailed behavior. In the soilstructure interaction problem, for
example, the former is used to model the ground which widely extends, and the latter is used to
model the structure of which shape is complex.
Let designate the Global area as
G
, the Local area as
L
and the boundary between these areas as
GL
. The image of the relationship of them is illustrated in Figure1.
Figure 1. Superimposition of global and local areas
Displacement fields are independently defined in each
G
and
L
, i.e., u
i
G
and u
i
L
, respectively.
The actual displacement u
i
in
L
is defined as the sum of u
i
G
and u
i
L
, while u
i
is equal to u
i
G
outside the
L
. Namely, the displacement u
i
is defined as the following equations.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
128
L L
i
G
i i
u u u + = in
(1)
L G G
i i
u u = in
(2)
To satisfy the continuity of the displacement at the boundary
GL
, the following condition is
needed.
on 0
GL
=
L
i
u
(3)
Displacements u
i
G
and u
i
L
in
G
and
L
are expressed by using shape function matrices N
G
and N
L
and nodal displacement vectors u
i
G
and u
i
L
as follows.
G
j
G
ij
G
i
u N u =
(4)
L
j
L
ij
L
i
u N u =
(5)
By partially differentiating Eq.(1) and using above equations, we obtain strain
ij
as,
L
ij
G
ij ij
+ = (6)
In which
G
k
G
ijk
G
ij
u B =
(7)
L
k
L
ijk
L
ij
u B =
(8)
By using the principle of virtual work, we can obtain the next equation.
+ = d t u d b u d D
i i i i kl ijkl ij
(9)
Where,
ij
,
ij
,
i
u , b
i
, t
i
, D
ijkl
are virtual strain, strain, virtual displacement, body force, surface
traction and constitutive tensor, respectively. The left side of the equation stands for the virtual
work due to the internal strains and the right side represents the virtual work done by the external
forces. By substituting Eq.s (1), (6), (7) and (8) into Eq. (9), we can obtain the following equations.
+ + + =
+ +
d t u u d b u u
d D
i
L
i
G
i i
L
i
G
i
L
ij
G
ij ijkl
L
ij
G
ij
) ( ) (
) ( ) (
(10)
+ + + =
+ +
d t u N u N d b u N u N
d u B u B D u B u B
i
L
i
L
ij
G
i
G
ij i
L
i
L
ij
G
i
G
ij
L
m
L
ijm
G
m
G
ijm ijkl
L
m
L
ijm
G
m
G
ijm
) ( ) (
) ( ) (
(11)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
129
By rewriting the above equations in the matrix form, we obtain the following equation.
L
G
L
G
L LG
GL G
f
f
u
u
K K
K K
(12)
Where :
+ =
+ =
=
=
=
=
d t N d b N f
d t N d b N f
d B D B K
d B D B K
d B D B K
d B D B K
i
L
i i
L
i
L
i
G
i i
G
i
G
L L
kl ijkl
L
ij
L
L G
kl ijkl
L
ij
LG
L L
kl ijkl
G
ij
GL
G G
kl ijkl
G
ij
G
L
L
L
G
(13)
In which K
G
and f
G
are stiffness matrix and external force vector for the global area
G
, and K
L
and f
L
are stiffness matrix and external force vector for the local area
L
, respectively.
2.2 Linking the beam element and the plane strain solid element
According to the previous work
1)
, linkage matrices between global and local plane strain elements,
K
GL
and K
LG
, are obtained from Eq.(13). Linkage matrices between plane strain elements and beam
elements, however, cannot be obtained in the same manner, because the strains are different between
the beam element and the solid element. It is, therefore, necessary to develop a new method to link
them.
The global nodal displacement at the same position as that of the local node, u
G
can be obtained
by using the global shape function N
G
and global nodal displacements
G
k u as Eq. (14).
G
k
G
kl
G
l
u N u = (14)
Global strain at arbitrary point,
G
, can be obtained from Eq.(7), and also obtained using other
element if the point is included inside the element and the coordinate of the nodal points of the
element. Therefore, global strain can be obtained by using u
G
and local shape function B
L
.
G
l
L
ijl
G
k
G
ijk
G
ij
u B
u B
=
=
(15)
Using Eq. (14), we can obtain the following relationships.
G
kl
L
ijl
G
ijk
G
k
G
ijk
G
l
L
ijl
G
k
G
kl
L
ijl
N B B
u B
u B u N B
=
=
=
(16)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
130
Therefore, K
LG
can be obtained as follows.
[ ]
[ ][ ]
G L
G
mn
L L
klm ijkl
L
ij
L G
mn
L
klm ijkl
L
ij
L G
kl ijkl
L
ij
LG
N K
N d B D B
d N B D B
d B D B K
L
L
L
=
=
=
=
(17)
In the same manner, K
GL
is expressed in the following way.
[ ] [ ] [ ]
L
T
G LG
K N K =
(18)
2.3 Constitution of the local mesh
Figure 2 shows the total system which includes global model and local model. The local model
contains beam elements of which area is designated by
C
. The local area modeled by solid elements
is expressed by
B
and the global area by
A
. It is assumed that the areas
A
and
C
are not in
contact. The constants of elasticity in the areas
A
and
B
are the same and expressed as D
1
ijkl
and in
the area
C
, D
1
ijkl
in the global model and D
L
ijkl
in the local model.
As for the boundaries, the boundary between
A
and
B
is designated by
AB
, in the same
manner, the boundary between
B
and
C
is designated by
BC
. The boundary is divided into
A
,
B
and
C
according to the division of the areas
A
,
B
and
C
, respectively.
With the definitions above, K
G
, K
L
and K
GL
are obtained as follows.
[ ] [ ] [ ][ ]
+ +
=
C B A
d B D B K
G
kl ijkl
T
G
ij
G 1
(19)
[ ] [ ] [ ][ ] [ ] [ ][ ]
+ =
C B
d B D B d B D B K
L
kl
L
ijkl
T
L
ij
L
kl ijkl
T
L
ij
L 1
(20)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
131
[ ] [ ] [ ][ ] [ ] [ ][ ]
+ =
C B
d B D B d B D B K
L
kl
L
ijkl
T
G
ij
L
kl ijkl
T
G
ij
GL 1
(21)
Eq. (11) can be written in the tensor form as;
+ + + =
+ +
+
d t u d t u d b u d b u
d D d D
d D d D
i
L
i i
G
i i
L
i i
G
i
L L
kl ijkl
L
ij
L G
kl ijkl
L
ij
L L
kl ijkl
G
ij
G
kl ijkl
G
ij
L
L L
L
(22)
The displacements can be written in the following equation, in which symbols G, L, A, B and C stand
for Global, Local and areas A, B and C.
+ = +
+ = +
=
=
in
in
in
C LC
i
GC
i
L
i
G
i
B LB
i
GB
i
L
i
G
i
A GA
i
G
i
i
u u u u
u u u u
u u
u
(23)
As for the global displacement concerning the virtual displacement
G
i
u , and strain
G
ij
, we can
obtain the following equation.
+ + +
+ + =
+ +
+ +
d t u d t u d t u
d b u d b u d b u
d D d D
d D d D d D
i
GC
i i
GB
i i
GA
i
i
GC
i i
GB
i i
GA
i
LC
kl
L GC
ij
LB
kl
GB
ij
GC
kl
GC
ij
GB
kl
GB
ij
GA
kl
GA
ij
C B A
C
ijkl
B
ijkl
A
ijkl
B
ijkl
A
ijkl
1
1 1 1
(24)
By partially integrating the left part of Eq.(24) using the Greens formula, the following equation is
obtained.
{ } { }
{ }
{ } { }
{ }
{ } 0 ) ( ) (
) (
) ( ) (
) (
) (
1 1
1 1
1 1
1
, ,
1
, ,
1
,
1
= + + +
+ +
+ + + +
+ + +
+ + +
d u n D D D
d u n D D
d u t n D D d u t n D
d u t n D d u b D D
d u b D d u b D
GBC
i
C
j
LC
kl
L
ijkl
GC
kl ijkl
LB
kl
GB
kl ijkl
GAB
i
B
j
LB
kl
GB
kl ijkl
GA
kl ijkl
GC
i i
C
j
LC
kl
L
ijkl
GC
kl ijkl
GB
i i
B
j
LB
kl
GB
kl ijkl
GA
i i
A
j
GA
kl ijkl
GC
i i
LC
l kl
L
ijkl
GC
l kl ijkl
GB
i i
LB
l kl
GB
l kl ijkl
GA
i i
GA
l kl ijkl
BC
AB
C B
A C
B A
(25)
As the global displacements u
i
G
is continuous in area , the following relations can exist.
AB GAB
i
GB
i
GA
i
u u u = = on
(26)
BC GBC
i
GC
i
GB
i
u u u = = on
(27)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
132
On the other hand, as for the local displacement concerning the virtual displacement
L
i
u , and
strain
L
ij
, we can obtain the following equation.
+ + + =
+ +
+
C B C B
C B
C B
d t u d t u d b u d b u
d D d D
d D d D
i
LC
i i
LB
i i
LC
i i
LB
i
LC
kl
L
ijkl
LC
ij
LB
kl ijkl
LB
ij
GC
kl
L
ijkl
LC
ij
GB
kl ijkl
LB
ij
1
1
(28)
In the same manner as in the global area, Eq. (28) can be written as,
{ }
{ }
{ } { }
{ }
{ } 0 ) ( ) (
) (
) ( ) (
) (
1 1
1 1
1 1
, ,
1
, ,
1
= + + +
+ +
+ + + +
+ +
+ +
d u n D D D
d u n D D
d u t n D D d u t n D
d u b D D
d u b D
GBC
i
C
j
LC
kl
L
ijkl
GC
kl ijkl
LB
kl
GB
kl ijkl
GAB
i
B
j
LB
kl
GB
kl ijkl
GA
kl ijkl
GC
i i
C
j
LC
kl
L
ijkl
GC
kl ijkl
GB
i i
B
j
LB
kl
GB
kl ijkl
GC
i i
LC
l kl
L
ijkl
GC
l kl ijkl
GB
i i
LB
l kl
GB
l kl ijkl
BC
AB
C B
C
B
(29)
And,
AB
on 0 =
LB
i
u
(30)
In the Eq.s (25) and (29), as the virtual displacements are arbitrary, we obtain the following
equations.
A
i
GA
l kl ijkl
b D = + in 0
,
1
(31)
B
i
LB
l kl
GB
l kl ijkl
b D = + + in 0 ) (
, ,
1
(32)
C
i
LC
l kl
L
ijkl
GC
l kl ijkl
b D D = + + in 0
, ,
1
(33)
C
i
LC
l kl
L
ijkl
GC
l kl
L
ijkl
b D D = + + in 0
, ,
(34)
A
i
A
j
GA
kl ijkl
t n D = on 0
1
(35)
B
i
B
j
LB
kl
GB
kl ijkl
t n D = + on 0 ) (
1
(36)
C
i
C
j
LC
kl
L
ijkl
GC
kl ijkl
t n D D = + on 0 ) (
1
(37)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
133
C
i
C
j
LC
kl
L
ijkl
GC
kl
L
ijkl
t n D D = + on 0 ) (
(38)
{ }
AB B
j
LB
kl
GB
kl ijkl
GA
kl ijkl
n D D = + on 0 ) (
1 1
(39)
{ }
BC C
j
LC
kl
L
ijkl
GC
kl ijkl
LB
kl
GB
kl ijkl
n D D D = + + on 0 ) ( ) (
1 1
(40)
{ }
BC C
j
LC
kl
L
ijkl
GC
kl
L
ijkl
LB
kl
GB
kl ijkl
n D D D = + + on 0 ) ( ) (
1
(41)
By subtracting Eq. (34) from Eq. (33), Eq. (38) from Eq. (37), Eq. (41) from Eq. (40), we obtain
Eq.s (42), (43), (44), respectively.
C GC
l kl
L
ijkl ijkl
D D = in 0 ) (
,
1
(42)
C C
j
GC
kl
L
ijkl ijkl
n D D = on 0 ) (
1
(43)
BC C
j
GC
kl
L
ijkl ijkl
n D D = on 0 ) (
1
(44)
From Eq.s (43) and (44), equilibrium of stress is independently satisfied within the global model
on the boundaries
BC
and
C
, and normal stress outward direction is 0. From Eq.s (42), (43), (44) we
can get the next relationship.
C GC
kl
L
ijkl ijkl
D D = in 0 ) (
1
(45)
This means that the stress of beam elements due to global model is 0 on the boundary of area
C
.
And from Eq. (45)
C GC
kl
= in 0
(46)
Eq.s (33) and (34) become
C
i
LC
l kl
L
ijkl
b D = + in 0
,
(47)
In the same way, Eq. (40) and (41) become
{ }
BC B
j
LC
kl
L
ijkl
LB
kl
GB
kl ijkl
n D D = + on 0 ) (
1
(48)
This means that the stresses due to displacements in local model within area
C
on the
boundary
BC
, equilibrium to those within area
B
. Stresses in the beam elements, therefore, can be
expressed only by the local model and obtained only by the stiffness of the beam elements.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
134
Figure 5 Finite element mesh using the OMM
Figure 3 Vertical pile model
Figure 4 Ordinal finite element mesh of vertical
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
135
3 ANALITICAL EXAMPLES
3.1 Vertical pile model
The vertical pilefootingground model used in this analysis is illustrated in Figure 3. Ground and
piles and footing are assumed to be elastic materials and material constants of ground and footing are
the same for the simplicity. Youngs modulus, sectional area, and moment inertia of the section of
the pile are 206GPa, 0.009628m
2
, and 7.86x10
5
m
4
, respectively. Shear wave velocity, unit weight
and Poissons ratio are 300m/s, 16.67kN/m
3
, 0.3, respectively. Finite element models of the model
are shown in Figures 4 and 5. Figure 4 is the ordinal finite element model and Figure 5 is overlaying
mesh model.
Numerical analysis results are compared in Figures 6, 7 and 8. Horizontal placements of the
beam elements are illustrated in Figure 6, vertical displacements in Figure 7 and rotational angles in
Figure 8. In these figures, normal means the results from the ordinal element model, and OMM
from the overlaying mesh model.
The difference of the horizontal response displacements is about 0.1mm, and this is very small
compared with the maximum response of the system of 2.9mm, in the vertical direction, the
difference is about 0.03mm, while the maximum response is about 13.5mm. The distributions of the
response displacements in the total system are shown in Figures 9, 10, 11 and 12. From these figures
the results are almost same in two models. This means the validity of the proposed method.
Figure 6. Comparison of horizontal displacements Figure 7. Comparison of vertical displacements
Figure 8. Comparison of rotational angle pile
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
136
Figure 9. Distribution of horizontal displacement
from normal finite element mesh.
Figure 10. Distribution of horizontal displacement
from finite element mesh with OMM.
Figure 12. Distribution of vertical displacement
from finite element mesh with OMM.
Figure 11. Distribution of vertical displacement
from normal finite element mesh
Figure 13. Battered pile model.
Figure 14. Normal finite element mesh of battered
piles.
Figure 15. Finite element mesh of battered piles with OMM
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
137
3.2 Battered pile model
Figure 13 shows the battered pilegroundfooting model. The material constants are the same as
those of the vertical pile models. Figures 14 and 15 are ordinal mesh model and OMM, respectively.
The mesh of the former model is very complicated to express the battered piles, on the other hand, the
mesh is very simple for the latter model as shown in Figure 15. Same global mesh as in the vertical
pile is used for the OMM.
The comparisons of horizontal and vertical displacements and rotational angle are made in Figures
16, 17 and 18. The differences between these two models are little larger than those from the vertical
piles model, especially for rotational angle. One of the possible reasons is the irregularity of the
ordinal mesh model. To generate the mesh shown in Figure 14, we employed automatic mesh
generation software, thats why the mesh is not symmetry about the centerline. We need to seek and
examine other cause of the difference of the results.
The comparisons of the distributions of horizontal and vertical displacements in the global system
are shown in Figures 19 and 20. As can be seen in Figures 16 and 17, the difference of the responses
can be observed at the top and the tip of the piles.
The CPU time to analyze models shown in Figures 14 and 15 are almost same in both cases. To
generate the OMM is very easy, because we just put the battered piles models on the global model
(ground model). This is a typical advantage of using OMM.
Figure 16. Comparison of horizontal displacements Figure 17. Comparison of vertical displacements
Figure 18. Comparison of rotational angle of pile
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
138
4 CONCLISIONS
We derived the OMM in application of the soilstructure interaction system. Then we examine the
validity of the method. For vertical pile model, we could get good agreement between the ordinal
model and the OMM, but in the analysis of battered pile model, the difference is little larger than
those for vertical pile models. We need to examine the reason and establish the analysis method for
the soilstructure interaction problem, and more we need to extend the method to three dimensional
problem in which the advantage of the method will be remarkable.
REFERENCES
T. Belytchko, J. Fish and A. Bayliss (1990), The spectral overlay on finite elements for problems with high
gradients, Computer methods in applied mechanics and engineering, Vol.81 pp.7189.
Figure 19. Distribution of horizontal displacement from ordinal finite element mesh.
Figure 20. Distribution of horizontal displacement from finite element mesh with OMM.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
139
1 INTRODUCTION
Many researches have been done on the liquefaction of sand experimentally, empirically, and
mathematically. Cyclic mobility due to soil skeleton dilation at large shear strain excursions is
one of typical mechanical behaviors of sand during liquefaction. A large number of computa
tional models have been, and continue to be developed for simulation of nonlinear soil re
sponse, and the dilatancy effects. Research related to testing methods and modeling of the cy
clic mobility of sand can be found in many publications, e.g., the work by Oka et al. (1992);
LIQCA program by Yashima et al. (1991). In recent years, research on constitutive model for
soils has been developing very quickly. For instance, the concept of subloading proposed by
Hashiguchi and Ueno (1977), and the concept of superloading proposed by Asaoka et al.
(1998), make it possible not only to describe remolded soils (Roscoe et al., 1963 and Schofield
and Wroth 1968), but also naturally deposited soils in which overconsolidation, structure and
anisotropy of soils play a very important rule in determining the mechanical behaviors of soil.
Asaoka et al. (2002) proposed an elastoplastic constitutive model based on the evolution rules
which describe the collapse of the soil skeleton structure (the concept of superloading), the loss
in overconsolidation (the concept of subloading), and the development of anisotropy during
shearing. In their paper, the importance of stressinduced anisotropy is introduced.
In liquefaction analysis, a constitutive model that can properly describe stressstrain relations of
sand during liquefaction is most important. Liquefaction analysis is a solution of boundary
value problem (BVP). If the stress and strain path of sand cannot be expressed rightly in ele
mental level, the solution of a BVP is impossible or has no meaning. However, almost constitu
tive models can describe the stressstrain relations obtained from different values of material
parameters for sands with variant densities. This is unreasonable in a liquefaction process be
cause the density of ground is not constant and will increase due to flow out of the pore water
from ground.
Zhang et al. (2007) proposed a new constitutive model for soils that can describe liquefaction
behaviors of sands with different densities. The model provides an approach for describing the
Unique description of liquefaction behavior of Toyoura sands
with different densities
F. Zhang & Y. Jin
Nagoya Institute of Technology, Nagoya, Japan
ABSTRACT: In this paper, based on the model proposed by Zhang et al.(2007), a minor modi
fication for the evolution equation of overconsolidation is carried out at first and then a unique
description of the overall mechanical behaviors of Toyoura Standard Sand (TSS), a typical
clean sand, is conducted, in which the eight parameters for describing TSS is kept constant no
matter what kind of loadings or drainage conditions may be. In the theoretical simulation, based
on conventional drained triaxial compression tests and undrained triaxial cyclic loading tests,
the material parameters of TSS are determined. The capability of the model to describe
uniquely the overall behaviors of the sand under different drainage conditions and different
loadings with one set of fixed parameters is verified.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
140
stressinduced anisotropy with a new evolution rule for changes in overconsolidation, by which
the mechanical behavior of soils subjected to cyclic loading under undrained conditions, includ
ing the cyclic mobility of medium dense sand, can be uniquely described. Moreover, different
soils can be explained efficiently with the concepts of overconsolidation, structure and stress
induced anisotropy. In his paper, the proposed constitutive model is used to describe the overall
behaviors of Toyoura Standard Sand with different densities subjected to various kinds of load
ing under different drainage conditions with a unique set of parameters. The material parame
ters of the sand are determined based on conventional triaxial tests. By comparing the test re
sults with predicted ones, it is possible to confirm the performance of the constitutive model.
In this study, based on conventional drained triaxial compression tests and undrained triaxial
cyclic loading tests, material parameters of Toyoura Standard sand is determined at first and
then other tests under different loadings and drainage conditions are simulated by the constitu
tive model. The purpose of the research is to verify if the model can describe uniquely the over
all behaviors of the sand under different drainage conditions and different loadings with one set
of fixed parameters.
2 MODIFICATION OF ORIGINAL CONSTITUTIVE MODEL
In the original model (Zhang et al. 2007), apart from the concepts of subloading (Hashiguchi
and Ueno, 1977) and superloading (Asaoka et al., 1998), a new evolution rule for the develop
ment of overconsolidation was proposed. Here we use the word of development instead of
loss just want to emphasize that during plastic loading the degree of overconsolidation some
time may even increase, not always the case in which overconsolidation only develops in elastic
unloading process. In the original model, the changing rate of overconsolidation is assumed to
be controlled by two factors, namely, the plastic component of stretching that was employed as
the only factor in the work by Asaoka et al. (2002), and the increment in anisotropy, in other
words,
MD M
p
R f
R JU
= +
o
&
D
(1)
in which,
o
3
,
1
and
3
are the maximum and minimum principal effective stresses). In
the case of small amplitude of cyclic loading, it is seen that it needs more cyclic loading num
bers before cyclic mobility take places than those of large amplitude of cyclic loading. To de
termine the evolution parameter, element simulations are conducted under the same condition
with experiment.
Figure 2 shows theoretical results with different amplitude of cyclic loading. The tendency of
predicted effective stress paths and stressstrain relations are well coincident with test ones,
while the cyclic number necessary for causing cyclic mobility is less than test results.
A unique set of material parameters of Toyoura sand are listed in Table 2. Eight parameters are
involved in the constitutive model, among which five parameters, M, N,
%
, % , and are the
same as in the Camclay model obtained from laboratory tests. The values of the parameters ex
cept a and m are the same as the those in the works by Ye et al. (2007).
From the parametric study shown in Figure 2, the evolution parameter a which controls the deg
radation of structure is decided to be 0.5. Degradation parameter of overconsolidation state m
has already been determined by the works of Sago et al. (2008). The value of parameter b
r
is
determined by Ye et al. (2007) that is based on the performance of the soil influenced by the
development of the stressinduced anisotropy when the soil is subjected to shearing or com
pression. On the other hand, it is difficult to determine the initial conditions of the state parame
ters for soils because the values are not only dependent on present stress state but also on its
history. The initial conditions of the state parameters for the sand are listed in Table 3. Initial
anisotropy
0
is determined to be `zero` , which means the soil is assumed as isotropic at the
beginning if the specimen is not disturbed. Structure R
*
which influenced by the process of the
soils in its deposition, is quickly collapse during shearing and never recover. Therefore initial
degree of structure R
0
*
is assumed to be a relatively large value R
0
*
=0.75 for the medium dense
sand. While the initial degree of overconsolidation 1/R
0
is determined to be 70 after considering
the experiment results.
Table 1. Experiment conditions of sand with different amplitude of cyclic loading
Amplitude of shear stress ratio (q/2
0
) 0.15 0.2 0.25
Initial void ratio e
0
0.748 (Dr=0.62) 0.753 (Dr=0.61) 0.77 (Dr=0.56)
Initial mean effective stress p (kPa) 98 98 98
Cyclic loading frequency (Hz) 0.01 0.01 0.01
Note: e
max
=0.97; e
min
=0.61
60
30
0
30
60
5 2.5 0 2.5 5
Axial strain a()
D
e
v
i
a
to
r
s
t
r
e
s
s
q
(
k
P
a
)
60
30
0
30
60
0 50 100 150
Mean effective stress p`(kPa)
D
e
v
i
a
t
o
r
s
tr
e
s
s
q
(
k
P
a
)
60
30
0
30
60
5 2.5 0 2.5 5
Axial strain a()
D
e
v
i
a
to
r
s
t
r
e
s
s
q
(
k
P
a
)
60
30
0
30
60
0 50 100 150
Mean effective stress p`(kPa)
D
e
v
ia
t
o
r
s
t
r
e
s
s
q
(
k
P
a
)
60
30
0
30
60
5 2.5 0 2.5 5
Axial strain a()
D
e
v
i
a
t
o
r
s
t
r
e
s
s
q
(
k
P
a
)
60
30
0
30
60
0 50 100 150
Mean effective stress p`(kPa)
D
e
v
ia
t
o
r
s
t
r
e
s
s
q
(
k
P
a
)
(a) q/2
0
=0.15 (b) q/2
0
=0.20 (c) q/2
0
=0.25
Figure 1. Experiment results of cyclic triaxial test under undrained condition with different
amplitude of cyclic loading
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
142
Table 2. Material parameters of Toyoura Sand
Compression index 0.05
Swelling index 0.0064
Critical state parameter 1.30
Reference void ratio N (p=98 kPa on N.C.L.) 0.87
Poissons ratio 0.30
Degradation parameter of overconsolidation state m 0.01
Degradation parameter of structure a 0.5
Evolution parameter of anisotropy b
r
1.5
60
30
0
30
60
0 50 100 150
Mean effective stress p`(kPa)
D
e
v
ia
to
r
s
tr
e
s
s
q
(
k
P
a
)
60
30
0
30
60
5 2.5 0 2.5 5
Axial strain a()
D
e
v
ia
to
r
s
tr
e
s
s
q
(
k
P
a
)
60
30
0
30
60
0 50 100 150
Mean effective stress p`(kPa)
D
e
v
ia
to
r
s
tr
e
s
s
q
(
k
P
a
)
60
30
0
30
60
5 2.5 0 2.5 5
Axial strain a()
D
e
v
ia
to
r
s
tr
e
s
s
q
(
k
P
a
)
60
30
0
30
60
0 50 100 150
Mean effective stress p`(kPa)
D
e
v
ia
to
r
s
tr
e
s
s
q
(
k
P
a
)
60
30
0
30
60
5 2.5 0 2.5 5
Axial strain a()
D
e
v
ia
to
r
s
tr
e
s
s
q
(
k
P
a
)
(a) q/2
0
=0.15 (b) q/2
0
=0.20 (c) q/2
0
=0.25
Figure 2. Analysis results with different amplitude of cyclic loading
Table 3. Initial conditions of sand with different amplitude of cyclic loading
Amplitude of shear stress ratio(q/2
0
) 0.15 0.2 0.25
Initial void ratio e
0
0.81 0.81 0.81
Initial mean effective stress p (kPa) 98 98 98
Initial degree of structure R
0
*
0.75 0.75 0.75
Initial degree of overconsolidation 1/R
0
70 70 70
Initial anisotropy
0
0.0 0.0 0.0
3.2 Drained triaxial compression test and its simulation
Triaxial compression test and element simulation under constant mean principal stress are con
ducted to confirm the performance of the constitutive model. The test results on Toyoura sand
are reported from the works by Nakai et al. (2004). Loose specimen was prepared by depositing
the saturated sand slowly in deaired water using a funnel with an opening of 3mm. Dense
specimen was prepared by pouring the saturated sand into the mold in several layers and com
pacting each layer with a rod whose diameter was 6mm. The initial void ratio of the loose sand
is e
0
=0.851, and the dense sand is e
0
=0.666.
It is, however, very different to identify the reference void ratio N in a small confining stress
condition. Therefore, by extension of eln p relation to the small stress range, the reference void
ratio N of Toyoura sand is determined to be 0.87. Material parameters used in the analysis are
the same as the Table 2, and the initial values of the state parameters determined from the ele
ment simulation are listed in Table 4. Initial degree of structure is determined from the fact that
well remolded sand has extremely low structure. Figure 3 shows comparison between experi
ment and analysis results obtained from triaxial compression test sands with different densities.
The dots and the lines in the figures represent the relations between principal stress ratio q/p,
deviator strain
d
and volumetric strain
v
. In the Figure 3(a), theoretical result of loose sand
represents well the test result. The theoretical result for dense sand as shown in Figure 3(b) is
reproduced relatively well before peak strength while the residual stress and strain relations are
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
143
discrepancy slight from test results. On the whole, however, the constitutive model can describe
the test results qualitatively and quantitatively.
Table 4. Initial conditions of sand with different densities
Loose sand Dense sand
Initial void ratio e
0
0.78 0.69
Initial mean effective stress p (kPa) 196 196
Initial degree of structure R
0
*
0.99 0.99
Initial degree of overconsolidation 1/R
0
4.0 30.0
Initial anisotropy
0
0.0 0.0
1
0.5
0
0.5
1
1.5
2
0 5 10 15
Axial strain
a
(%)
P
r
i
n
c
i
p
a
l
s
t
r
e
s
s
r
a
t
i
o
q
/
p
0
10
5
0
5
V
o
l
u
m
e
t
r
i
c
s
t
r
a
i
n
v
(
%
)
0
Experimental result
Analysis result
1
0.5
0
0.5
1
1.5
2
0 5 10 15
Axial strain a (%)
P
r
i
n
c
i
p
a
l
s
t
r
e
s
s
r
a
t
i
o
q
/
p
0
10
5
0
5
V
o
l
u
m
e
t
r
i
c
s
t
r
a
i
n
v
(
%
)
0
Experimental result
Analysis result
(a) Loose sand (b) Dense sand
Figure 3. Experiment and analysis results obtained from triaxial compression test
3.3 Numerical prediction of sands with different densities
The behavior of sand is known as being dependent on its density. In order to verify the influ
ence of the initial density on the behavior of sand, a series of calculations with different initial
densities are conducted at the same confining pressure.
It is known that loose sand will undergo large volume change and get denser under repeated
perturbation. Firstly, loose sand subjected to repeated triaxial compaction under constant lateral
pressure is simulated under drained condition to explain the phenomenon of densification of
loose sand. Similar to the work by Nakai K. (2005), in the simulation, the density of specimen
is adjusted with a small vibrating compact with amplitude of 2.3kPa on the sand under a small
confining pressure 10kPa. After the compaction, these sands with different densities are isot
ropically consolidated to the prescribed confining pressure of 294kPa. The set of sands with
eight different densities are prepared by different numbers of vibration compaction, as shown in
Figure 4. All these sands have the same five parameters which are listed in Tables 2 and 5. It
means that these eight sands are made from the same sand material. The initial sand is supposed
to be extremely loose condition with a very large void ratio which is normally consolidated and
highly structured without stressinduced anisotropy. Table 6 presents initial conditions of sands
obtained from compaction procedure. By using these material parameters, the sands with dif
ferent densities in the laboratory tests under both drained and undrained conditions subjected to
monotones and cyclic loading, are calculated systematically.
Table 5. Reference conditions of sand before vibrating compaction
Initial void ratio e
0
1.19
Initial mean effective stress p (kPa) 10.0
Initial degree of structure R
0
*
0.1
Initial degree of overconsolidation 1/R
0
1.0
Initial anisotropy
0
0.0
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
144
0.5
0.6
0.7
0.8
0.9
1
1.1
1 10 100 1000
Mean effective stress (kPa)
V
o
i
d
r
a
t
i
o
e
294
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
Figure 4. Set of sands with different densities prepared from loose sand by vibration compaction and iso
tropic compression
Table 6. Initial conditions of sands obtained from compression procedure
No.
Vibration
number
n
Initial
void ratio
e
0
Initial degree of
overconsolidation
OCR
Initial degree of
structure
R
0
*
Initial
anisotropy
0
[1] 2 0.903 1.30 0.103 6.44E06
[2] 15 0.842 4.77 0.113 5.28E06
[3] 30 0.785 15.99 0.124 7.30E06
[4] 35 0.771 21.70 0.128 1.01E05
[5] 40 0.758 28.14 0.132 1.70E05
[6] 50 0.737 42.96 0.139 8.13E05
[7] 70 0.707 78.88 0.152 1.75E03
[8] 120 0.660 208.47 0.168 2.64E02
3.3.1 Mechanical behaviors of sand subjected to undrained/drained cyclic loading
In the simulation, confining stress of the specimen is 294kPa and amplitude of the cyclic load
ing in shear stress ratio (q/
m
) is 0.12.
Figure 5 shows the stress paths, stressstrain relations with different densities subjected to cy
clic triaxial test under undrained condition. It is clear from the figures that very loose sands pro
ducing a large failure strain along the way towards the zero effective stress state before cyclic
mobility has a chance to occur, as shown in Figures 5 (a) to (d). For medium dense sand, how
ever, cyclic mobility occurs and the strain which results in limited deformation gradually gets
larger, as shown in Figures 5 (e) to (g). On the other hand, dense sand having a small strain
never shows cyclic mobility, as depicted in Figure 5 (h). The above results mean that the me
chanical behavior of sand subjected to cyclic loading without drainage, can be uniquely and
properly described by the constitutive model no matter what density it may have.
Figure 6 displays the result of representative loose sand (No.2) subjected to cyclic loading un
der drained condition. Other simulate conditions are the same as aforementioned undrained test.
As the cyclic loading goes on, the loose sand is compacted and undergoes a large volumetric
stain, which is the same as the results reported by Nakai K.et al. (2004), in which the phenom
ena of consolidation and liquefaction of sand were just dependent on drainage condition. That
is, if the compaction conducted under drained condition, then the consolidation of sand will oc
cur; while under undrained condition, the liquefaction will occur.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
145
100
50
0
50
100
0 50 100 150 200 250 300
Mean effective stress p`(kPa)
D
e
v
ia
to
r
s
tr
e
s
s
q
(
k
P
a
)
[1]
100
50
0
50
100
8 4 0 4 8
Shear strain a()
D
e
v
ia
to
r s
tre
s
s
q
(
k
P
a
)
[1]
100
50
0
50
100
0 50 100 150 200 250 300
Mean effective stress p`(kPa)
D
e
v
ia
to
r
s
tr
e
s
s
q
(
k
P
a
)
[2]
100
50
0
50
100
8 4 0 4 8
Shear strain a()
D
e
v
ia
to
r s
tre
s
s
q
(
k
P
a
)
[2]
100
50
0
50
100
0 50 100 150 200 250 300
Mean effective stress p`(kPa)
D
e
v
ia
to
r
s
tr
e
s
s
q
(
k
P
a
)
[3]
100
50
0
50
100
8 4 0 4 8
Shear strain a()
D
e
v
ia
to
r s
tre
s
s
q
(
k
P
a
)
[3]
100
50
0
50
100
0 50 100 150 200 250 300
Mean effective stress p`(kPa)
D
e
v
ia
to
r
s
tre
s
s
q
(
k
P
a
)
[4]
100
50
0
50
100
8 4 0 4 8
Shear strain a()
D
e
v
ia
to
r s
tre
s
s
q
(
k
P
a
)
[4]
(a) No.1 (b)No.2 (c)No.3 (d) No.4
100
50
0
50
100
0 50 100 150 200 250 300
Mean effective stress p`(kPa)
D
e
v
ia
to
r
s
tr
e
s
s
q
(
k
P
a
)
[5]
100
50
0
50
100
8 4 0 4 8
Shear strain a()
D
e
v
ia
to
r s
tre
s
s
q
(
k
P
a
)
[5]
100
50
0
50
100
0 50 100 150 200 250 300
Mean effective stress p`(kPa)
D
e
v
ia
to
r
s
tr
e
s
s
q
(
k
P
a
)
[6]
100
50
0
50
100
8 4 0 4 8
Shear strain a()
D
e
v
ia
to
r s
tre
s
s
q
(
k
P
a
)
[6]
100
50
0
50
100
0 50 100 150 200 250 300
Mean effective stress p`(kPa)
D
e
v
ia
to
r
s
tr
e
s
s
q
(
k
P
a
)
[7]
100
50
0
50
100
8 4 0 4 8
Shear strain a()
D
e
v
ia
to
r s
tre
s
s
q
(
k
P
a
)
[7]
100
50
0
50
100
0 50 100 150 200 250 300
Mean effective stress p`(kPa)
D
e
v
ia
to
r
s
tr
e
s
s
q
(
k
P
a
)
[8]
100
50
0
50
100
8 4 0 4 8
Shear strain a()
D
e
v
ia
to
r
s
tr
e
s
s
q
(
k
P
a
)
[8]
(e) No.5 (f) No.6 (g) No.7 (h) No.8
Figure 5. Stress paths, stressstrain relations of the sand specimens with different densities subjected to
cyclic triaxial test under undrained condition
100
50
0
50
100
0 100 200 300 400
Mean effective stress p`(kPa)
D
e
v
i
a
t
o
r
s
t
r
e
s
s
q
(
k
P
a
)
0.6
0.7
0.8
0.9
0 100 200 300 400
Mean effective stress p`(kPa)
V
o
i
d
r
a
t
i
o
e
e=0.189
(a) (b)
Figure 6. Stress paths, stressstrain relations of the loose sand specimen
(No.2) subjected to cyclic triaxial test under drained condition
3.3.2 Mechanical behaviors of sand under undrained/drained compression
The relationships between stress, strain and void ratio under undrained/drained triaxial com
pression tests on different densities of sand are discussed in this section. The parameters of the
sand used in the calculation are the same as in Table 2 and Table 6.
Under undrained conditions, three different types of typical stressstrain relations can be ob
served in the simulation, as shown in Figure 7. For loose sands marked with [1] to [2], the sands
exhibit peak strength at first in small strain level, then collapse and flow rapidly to the original,
showing a typical strain hardening and strain softening behavior. For medium dense sands
marked with [3] to [7], stiffness of the sands decrease abruptly in small strain but strain harden
ing continues within an intermediate range of shear strains and finally followed by a limited
strain softening, in which a relative higher strength can be approached in spite of a limited
strain softening. Dense sand marked with [8], however, only shows strain hardening and large
shear strength can be reached. Figure 7(c) depicts the traces of the states of all sands during
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
146
shearing in eln p curves, showing that all sands change towards the C.S.L. horizontally at criti
cal states.
Figure 8 displays the simulated results of stress, strain and dilation of the sands under drained
condition. It is observed that all sands approached to the same values of ultimate shear stress
and the same void ratio, irrespective of different initial densities at the beginning of shearing. It
is also known that initial dense sands show typical strain hardening  strain softening and dila
tion while initial loose soils only show strain hardening along with monotonic compression.
The simulated facts that the behaviors of sands subjected to shearing under drained/undrained
condition are largely dependent on their initial densities, are already wellknown in laboratory
tests of sands and it is not necessary to give any comparison between test and simulation.
0
150
300
450
600
0 150 300 450 600
Mean effective stress p`(kPa)
D
e
v
i
a
to
r
s
t
r
e
s
s
q
(
k
P
a
)
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
0
150
300
450
600
0 2 4 6 8 10
Shear strain a()
D
e
v
i
a
to
r
s
t
r
e
s
s
q
(
k
P
a
)
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
0.6
0.7
0.8
0.9
1
1 10 100 1000 10000
Mean effective stress p`(kPa)
V
o
id
r
a
ti
o
e
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
294
start point
end point
(a) (b) (c)
Figure 7. Stress paths, stressstrain relations of the sand specimens with different densities subjected to
triaxial compression test under undrained condition
0
300
600
900
1200
1500
0 2 4 6 8 10
Shear strain a()
D
e
v
i
a
to
r
s
t
r
e
s
s
q
(
k
P
a
)
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
6
4
2
0
2
4
6
8
10
0 2 4 6 8 10
Shear strain a()
V
o
lu
m
e
tr
i
c
s
t
r
a
i
n
v
(
)
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
0.5
0.6
0.7
0.8
0.9
1
1.1
0 300 600 900 1200 1500
Mean effective stress p`(kPa)
V
o
i
d
r
a
t
io
e
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
start point
end point
(a) (b) (c)
Figure 8. Stressstraindilatancy relations of the sand specimens with different densities subjected to triax
ial compression test under drained condition
3.4 Confiningstress dependency of sand in undrained monotonic loading tests
Verdugo and Ishihara (1996) reported their experimental results of TSS, in which undrained
triaxial compression tests on sands with the same void ratio but different confining pressures
were conducted under very high confining pressures (up to 3MPa). The test results given in
Figure 9 show that under the same void ratio, if a confining stress is large, the sand behaves like
a loose sand, while if the confining stress is small, the sand behaves like a dense sand. Such a
phenomenon is called as confiningstress dependency of sand, originally defined in the re
search by Ishihara (1993). Nakai (2005) also reported the same phenomenon in his tests on sil
ica sands.
The parameters of the sand used in the simulation are the same as Table 2. The initial values of
the void ratios are set to be equal to 0.78, 0.70 and 0.65 respectively and are listed in Table 7
with other initial conditions of the sands. Figure 10 shows that the simulated results on the
whole, coincide well with the test results quantitatively and qualitatively. It is also known that
the mechanical behavior of sands with the same density but different confining stresses can also
be reproduced uniquely with one set of the same material parameters in all different conditions.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
147
0
1
2
3
4
0 5 10 15 20 25
Axial strain a (%)
D
e
v
i
a
t
o
r
s
t
r
e
s
s
q
(
M
P
a
)
'0=2.0 MPa, e=0.908
'0=1.0 MPa, e=0.906
'0=0.1 MPa, e=0.908
Toyoura sand
( e=0.907, Dr=18.5% )
0
1
2
3
4
0 1 2 3 4
Effective mean stress p' (MPa)
D
e
v
i
a
t
o
r
s
t
r
e
s
s
q
(
M
P
a
)
Toyoura sand
( e=0.907, Dr=18.5% )
0
1
2
3
4
0 5 10 15 20 25
Axial strain
a
()
D
e
v
i
a
t
o
r
s
t
r
e
s
s
q
(
M
P
a
)
0
1
2
3
4
0 1 2 3 4
Mean effective stress p'(MPa)
D
e
v
i
a
t
o
r
s
t
r
e
s
s
q
(
M
P
a
)
[3]
(a) Loose sand (e=0.907, Dr=18.5%) (a) Loose sand (e=0.78)
0
1
2
3
4
0 1 2 3 4
Effective mean stress p' (MPa)
D
e
v
i
a
t
o
r
s
t
r
e
s
s
q
(
M
P
a
)
Toyoura sand
( e=0.833, Dr=37.9%)
0
1
2
3
4
0 5 10 15 20 25
Axial strain a (%)
D
e
v
i
a
t
o
r
s
t
r
e
s
s
q
(
M
P
a
)
Toyoura sand
( e=0.833, Dr=37.9% )
'0=3.0 MPa
'0=2.0 MPa
'0=1.0 MPa
'0=0.1 MPa
0
1
2
3
4
0 5 10 15 20 25
Axial strain a (%)
D
e
v
i
a
t
o
r
s
t
r
e
s
s
q
(
M
P
a
)
0
1
2
3
4
0 1 2 3 4
Effective mean stress p' (MPa)
D
e
v
i
a
t
o
r
s
t
r
e
s
s
q
(
M
P
a
)
(b) Medium dense sand (e=0.833, Dr=37.9%) (b) Medium dense sand (e=0.70)
0
1
2
3
4
0 1 2 3 4
Effective mean stress p' (MPa)
D
e
v
i
a
t
o
r
s
t
r
e
s
s
q
(
M
P
a
)
Toyoura sand
( e=0.735, Dr=63.7%)
0
1
2
3
4
0 5 10 15 20 25
Axial strain a (%)
D
e
v
i
a
t
o
r
s
t
r
e
s
s
q
(
M
P
a
)
Toyoura sand
( e=0.735, Dr=63.7% )
'0=3.0 MPa
'0=2.0 MPa
'0=1.0 MPa
'0=0.1 MPa
0
1
2
3
4
0 5 10 15 20 25
Axial strain a (%)
D
e
v
i
a
t
o
r
s
t
r
e
s
s
q
(
M
P
a
)
0
1
2
3
4
0 1 2 3 4
Effective mean stress p' (MPa)
D
e
v
i
a
t
o
r
s
t
r
e
s
s
q
(
M
P
a
)
(c) Dense sand (e=0.735, Dr=63.7%) (c) Dense sand (e=0.65)
Figure 9. Test results of TSS Figure 10. Simulation of the test results in Fig.9
(Verdugo and Ishihara, 1996)
Table 7 Initial conditions of TSS samples before undrained triaxial compression
No.
Confining
stress
p' (MPa)
Initial
void ratio
e
0
Initial degree of
overconsolidation
OCR
Initial degree of
structure
R
0
*
Initial
anisotropy
0
0.1 0.776 40.0 0.21 0.0
1.0 0.781 3.0 0.20 0.0
2.0 0.781 1.6 0.15 0.0
0.1 0.698 111.0 0.45 0.0
1.0 0.698 12.0 0.30 0.0
2.0 0.703 9.0 0.16 0.0
3.0 0.709 5.3 0.15 0.0
0.1 0.651 270.0 0.55 0.0
1.0 0.651 35.0 0.30 0.0
2.0 0.651 16.0 0.30 0.0
3.0 0.651 12.0 0.25 0.0
3.5 Dense sand subjected to drained cyclic loading
Finally, behaviors of dense sand subjected to drained cyclic loading under constantmean
effectivestress constant mean effective stress are simulated. The confining pressure of the sand
is 196 kPa and cyclic loading condition is that the mean effective stress is kept constant and a
maximum principal stress ratio (
1
/
3
) is loaded to 4. Figure 11 shows the test results by Hi
nokio (2000), in which the stressstrain curves are plotted in terms of effective stress ratio
1
/
3
, a dimensionless normalized stress. The volumetric strain shows dilatancy at the very be
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
148
ginning under cyclic loading and then turns to compression until it reaches a steady state at
which the compression almost stops, as shown in Figure 11 (b). For deviatory stressstrain rela
tion, at the beginning, it shows a relatively large loop, as the cyclic loading number increases,
however, the stiffness of the sand grows up and the stressstrain relation comes into an almost
fixed loop as shown Figure 11(c). In the simulation, the parameters of the sand are the same as
those listed in Table 2 and initial conditions of dense sand are shown in Table 8. In determining
the initial conditions of the sand, it is assumed that the sand is wellremolded one with ex
tremely low structure and relatively high OCR. As can be seen in Figure 12, the overall charac
teristics of the sand predicted by the theory, for instance, the changes in dilatancy and stress
strain relations, agree qualitatively well with the test results, but showing a slight over
estimation of volume strain. It should be emphasized here that in the simulation, volumetric
compression also stopped automatically after certain cycles of loadings, which is completely
coincides with the reality. The reason why the model can describe this behavior is quite simple.
Taking a look at Figures 12(d) and (e), in which the changes of overconsolidation and stress
induced anisotropy are plotted, it is easy to find out that during plastic loading the degree of
overconsolidation sometime may even increase, not always the case in which overconsolidation
only develops in elastic unloading process. One of the most important features of the model is
that the changing rate of overconsolidation is assumed to be controlled by two factors, namely,
the plastic component of stretching and the increment in anisotropy. The physical meaning of
the assumption is clear because stressinduced anisotropy is usually dependent on the roundness
of soil particles and their orientation of deposition. Therefore a sand with strong anisotropy will
have a stronger resistance against volumetric change than those of weak one in which soil parti
ciples deposit rather randomly. It is clear from the figures that during cyclic shearing, overcon
solidation gets higher and higher, in other words, the density is getting higher, resulting in the
difficulty to further compression.
(a) (b) (c)
Figure 11. Test results of dense sand in drained cyclic loading (after Hinokio, 2000)
Table 8. Initial conditions of sand before drained cyclic loading
Initial void ratio e
0
0.661
Confining stress p (kPa) 196.0
Initial degree of structure R
0
*
0.99
Initial degree of overconsolidation 1/R
0
55.0
Initial anisotropy
0
0.0
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
149
4
1.5
1
3.5
6
0 150 300
Mean effective stress p`(kPa)
S
t
r
e
s
s
r
a
t
i
o
1
/
3
comp.
ext.
196
a/r=5.4
r/a=5.4
3.5
6
2
0
2
4
6
8
4 1.5 1 3.5 6
Stress ratio
1
/
3
V
o
l
u
m
e
t
r
i
c
s
t
r
a
i
n
v
(
%
)
0
comp. ext.
6 3.5
4
1.5
1
3.5
6
0.5 0 0.5 1 1.5 2
Deviator strain
d
(%)
S
t
r
e
s
s
r
a
t
i
o
1
/
3
comp.
ext.
3.5
6
(a) (b) (c)
0
300
600
900
1200
4 1.5 1 3.5 6
Stress ratio
1
/
3
O
C
R
comp. ext.
6 3.5
0
0.25
0.5
0.75
1
4 1.5 1 3.5 6
Stress ratio
1
/
3
D
e
g
r
e
e
o
f
a
n
i
s
o
t
r
o
p
y
comp. ext.
6 3.5
(d) (e)
Figure 12. Simulated stress paths, stress ratiostrain relations of dense sand in drained cyclic loading under
constant mean effective stress
4 SUMMARY
In this study, based on conventional drained triaxial compression tests and undrained triaxial
cyclic loading tests, material parameters of Toyoura Standard sand is determined and other tests
under different loadings and drainage conditions are simulated by the constitutive model. The
capability of the model to describe uniquely the overall behaviors of the sand under different
drainage conditions and different loadings with one set of fixed parameters, is verified and the
following conclusions can be given.
1) The evolution parameter a, the parameter controlled the degradation of structure, is deter
mined to be 0.5 from the parametric study based on the undrained cyclic triaxial test. Mean
while, from the parametric study based on the drained triaxial compression test, reference
void ratio N is determined to be 0.87.
From the comparison between experiment and theoretical results obtained from undrained
triaxial cyclic loading tests and drained triaxial compression tests, it is known that the con
stitutive model can describe the test results relatively well.
2) The behavior of sand is known to be largely dependent on its density. For this reason, a se
ries of simulations on sands with different densities originally compacted from the same
sand are conducted with one set of material parameters. The results reveal the fact that the
mechanical behaviors of sand subjected to cyclic loading under drained/undrained condition
can be uniquely and properly described by the constitutive model no matter what density it
may have. Furthermore, it is concluded theoretically that if the compaction conducted under
drained condition, then the consolidation of sand will occur; while under undrained condi
tion, the liquefaction will occur.
3) Under undrained triaxial compression test, loose sands exhibit peak strength at first in
small strain level, then collapse and flow rapidly to the original, showing a typical strain
hardening and strain softening behavior. For medium dense sands, stiffness of the sands de
crease abruptly in small strain but strain hardening continues within an intermediate range of
shear strains and finally followed by a limited strain softening, in which a relative higher
strength can be approached in spite of a limited strain softening. Dense sand, however, only
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
150
shows strain hardening and large shear strength can be reached. In the drained condition, it is
observed that all sands approached to the same values of ultimate shear stress and the same
void ratio, irrespective of different initial densities at the beginning of shearing.
4) Confiningstress dependency of sand, a typical behavior of sands with the same density but
different confining stresses, can also be reproduced with the set of fixed material parameters.
5) Dense sand subjected to drained cyclic loading under constant mean effective stress is also
simulated. The overall characteristics of the sand is predicted well by the theory, for in
stance, the changes in dilatancy and stressstrain relations are qualitatively the same as the
test results with a slight overestimation of volume strain.
It cannot say that the model can perfectly describe the colorful behaviors of TSS, but that the
model can give a unified description of TSS to a rather satisfactory level if considering the fact
that only eight material parameters are employed in the model. Meanwhile, if the influence of
intermediate principal stress can be properly taken into consideration like the t
ij
concept (Nakai
and Mihara, 1984), the model would be much better.
REFERENCES
Asaoka, A., Nakano, M. and Noda. T. (1998): Super loading yield surface concept for the saturated struc
tured soils, Proc. of the Fourth European Conference on Numerical Methods in Geotechnical Engi
neeringNUMGE98, 232242.
Asaoka, A., Noda, T., Yamada, E., Kaneda, K. and Nakano, M. (2002): An elastoplastic description of
two distinct volume change mechanisms of soils, Soils and Foundations, Vol.42, No.5, 4757.
Hashiguchi, K. and Ueno, M. (1977): Elastoplastic constitutive laws of granular material, Constitutive
Equations of Soils, Pro. 9th Int. Conf. Soil Mech. Found. Engrg., Spec. Ses. 9, Murayama, S. and
Schofield, A. N. (eds.), Tokyo, JSSMFE, 7382.
Hinokio, M. (2000): Deformation characteristic of sand subjected to monotonic and cyclic loadings and
its application to bearing capacity problem, Doctoral Dissertation, Nagoya Institute of Technology,
3338 (in Japanese).
Ishihara, K. (1993): Liquefaction and flow failure during earthquake, The 33
rd
Rankine Lecture, Geotech
nique.
Nakai, K. (2005): An elastoplastic constitutive modeling of soils based on the evolution laws describing
decay of soil skeleton structure, loss of overconsolidation and development of anisotropy , Doctoral
Dissertation, Nagoya University, 115118 (in Japanese).
Nakai, K., Nakano, M., Noda, T. and Asaoka, A. (2004): Description of Compaction and Liquefac
tion behavior of sand based on evolution of soil skeleton structure, Proc. 2
nd
International Workshop
on New Frontiers in Computational Geotechnics, Fortaleza, Brazil, Zhang et al. (eds), 135144.
Nakai, T. and Mihara, Y. (1984): A new mechanical quantity for soils and its application to elastoplastic
constitutive models, Soils and Foundations, Vol.24, No.2, 8294.
Nakai T. and Hinokio M. (2004): A simple elastoplastic model for normally and over consolidated soils
with unified material parameters, Soils and Foundations, Vol.44, No.2, 5370.
Oka, F., Yashima, A., Kato, M. and Sekiguchi, K. (1992): A constitutive model for sand based on the non
linear kinematic hardening rule and its application, Proc. 10th World Conf. Earthquake Engineering,
Madrid, Vol.5, Balkema, 25292534.
Roscoe, K. H., Schofield, A. N. and Thurairajah, A. (1963): Yielding of clays in states wetter than critical,
Geotechnique, Vol.13, No.3, 250255.
Sago, H. (2008). The evaluation of earthquake resistant capability on selfstanding steel sheet pile, Master
Dissertation, Nagoya Institute University (in Japanese)
Schofield, A. N. and Wroth, C. P. (1968): Critical State Soil Mechanics, McGrawHill.
Verdugo, R. and Ishihara, K. (1996): The steady state of sandy soils, Soils and Foundations, Vol.36, No.2,
8191.
Yashima, A., Oka, F., Shibata, T., and Uzuoka, R. (1991): Liquefaction analysis by LIQCA, Proceedings
of JGS Conference on Liquefaction of Ground and its Counter measure, 165174 (in Japanese).
Ye, B., Ye, G., Zhang, F. and Yashima, A. (2007): Experiment and numerical simulation of repeated liq
uefactionconsolidation of sand, Soils and Foundations, Vol.47, No.3, 547558.
Zhang, F., Ye, B., Noda, T., Nakano, M. and Nakai, K. (2007): Explanation of Cyclic Mobility of Soils:
Approach by StressInduced Anisotropy, Soils and Foundations, Vol.47, No.4, 635648.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
151
APPENDIX: BRIEF DESCRIPTION OF CONSTITUTIVE MODEL
The model (Zhang et al., 2007) is proposed based on the concepts of subloading (Hashiguchi and
Ueno, 1977) and superloading (Asaoka et al., 1998). Here a brief description of the model is given.
The similarity ratio of the superloading surface to normal yield surface R
*
and the similarity ratio of the
superloading surface to subloading surface R are given as,
'
* *
' ' '
, 0 1
p q q q
R R and
q p p p
= = < =
% % %
%
A(1)
'
' ' ' '
, 0 1,
p q q q q
R R and
q p p p p
= = < = =
%
%
A(2)
where,
' ' '
( , ), ( , ) and ( , ) p q p q p q % % represent the present stress state, the corresponding normally con
solidated stress state and the structured stress state at pq plane, respectively. The normally yield surface is
given in the following form as:
' 2 2 2
'
' 2 2
0 0
0
M
( , , ) tr MDln MDln tr 0
M
t t
p p
p
f p J d J d
p
+
+ = + + =
% %
% %
%
D D A(3)
where,
* *
= % , and the other variables involved in Equations, (1), (2) and (3) are defined as:
' ' '
3 / 2, , tr / 3
' '
p p , p
= = + = S/ S T I T = , = = , = = , = = , = A(4)
2/ 3 = , 2/ 3 = A(5)
where, S SS S is the deviatory stress tensor; is the anisotropic stress tensor, and
'
T is the Cauchy effec
tive stress tensor and is assumed to be positive in tension. In Equation A(3), J is the Jacobean determina
tion of deformation gradient tensor F and can be expressed as:
0 0
1
det
1
v e
J
v e
+
= = =
+
F A(6)
where v and v
0
are the specific volume at the current time (t) and the specific value at the reference
time (t=0). D is the dilatancy parameter which can be expressed by ,
%
% the compression and the swell
ing index, respectively, as follows:
0 0
D
M(1 ) Mv e
= =
+
% %
% %
A(7)
D
p
denotes the plastic component of stretching D which is assumed to be positive in tension, and is re
lated to the plastic volumetric strain rate in the following form under the condition that the compressive of
the volumetric strain is supposed to be positive:
0
tr
t
p p
v
J d =
D
A(8)
By substituting Equations A(1) and A(2) into Equations A(3), subloading yield surface can also be
written as:
' *
0
' 2 2 2
' 2 2
0
0
( , , ) MDln MDln tr
M
MDln MDln MDln MDln tr 0
M
t
p
t
p
f p R R J d
p
R R J d
p
+ +
+
= + + + =
D
D
A(9)
An associated flow rule is employed in the present model, namely,
'
/
p
f = D T A(10)
The consistency equation for the subloading yield surface can then be given as:
'
MD MD tr 0
p
f f R R
J
R R
+ + + =
o
o & &
T D
T
'
A(11)
where,
o
T
'
and
o
are the GreenNaghdi rates of stress tensor T
'
and anisotropic stress tensor , re
spectively. is material spin tensor. It is easy to obtain the following relation:
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
152
2 2 2
2 2 2
' ' ' 2 2 2 '
1 ( ) M
MD( /(M )) MD
(M )
f
p p p p
= + + =
+
A(12)
From Equation A(12), it is clear that the C.S.L., defined by the condition in which
'
/ 0 f p = , always
satisfies the relation =M, implying that the C.S.L., as the threshold between plastic compression and
plastic expansion, does not move with the changes in the anisotropy.
Evolution rule for the anisotropic stress tensor is defined as:
(M ) 2 / 3
D
p
r s
J
b =
o
D
A(13)
where, <M provides a natural physical limitation on the development of anisotropy automatically., it is
known from the evolution Equation A(13) that development of anisotropy will stop at the state
when = .
Evolution rule for degree of structure R*, which is the same as in the work by Asaoka et al. (2002), is
adopted:
2 / 3
p
s
R JU
=
&
D , (1 ) / D U aR R
= , ( 0 1 R
< ) A(14)
The changing rate of overconsolidation is assumed to be controlled by two factors, namely, the plastic
component of stretching and the incremental anisotropy as,
MD M
p
R f
R JU
= +
o
&
D
A(15)
In which,
o
is proportional to the norm of the plastic component of stretching
p
s
D and U is given
by the following relation as:
' ' 2
0
' ' 2
0
( / )
( ) ln
D ( / ) 1
p p m
U R
p p
=
+
A(16)
where
'
0
p =98.0 kPa is reference stress.
If the stretching is divided into elastic and plastic components, and the elastic components follow
' '
'
, ,
e e p
f
= = + =
o o
T ED D D D T ED E
T
A(17)
The positive valuable ( =) can be rewritten as :
'
2 2
s ' ' 2 2 *2 '
MD
(M )
(M )
f
f f
J
p
=
+
+
ED
E
T T
T A(18)
where
' ' 2
2 2 *2 2 2 2 0
s ' ' 2
0
*2 2
* *
2 2 *2 2 2
( / ) Mln 1
M M { } 6 (M )
( / ) 1 3
6M (M ) (2M 3 )
2 M(1 ) (1 )
M (M )(M )
r
p p m R
R p p
b
a R
= +
+
+
+
A(19)
The loading criteria are given below:
0
0
0
loading
neutral
unloading
>
<
A(20)
In most cases, the denominator is positive, therefore, 0 > is equivalent to the following relation:
'
0
f
>
ED
T
A(21)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
153
2.
EarthquakeObservations,GroundMotions
i
becomes 1 only
when a sample ( th sample of ith group) fall into a subcategory of each item (k category of j
item).
=
0
1
) ( jk
i
(1)
To discriminate individuals into several groups, sample scores are computed by the following
linear equation.
= =
=
R
j
c
k
i jk i
j
jk a Y
1 1
) (
(2)
where
jk
is unknown coefficients. As already known in the analysis of variance, total sum of
squared differences between scores on
i
and the grand mean is partitioned into sum of
squared differences between group means and the grand mean (between groups variability), and
sum of squared difference between individual scores and their respective groups means (within
group variability).
a
Y
= = = = =
+ =
K
i
n
i i
K
i
i i
K
i
n
i
i i
Y Y Y Y n Y Y
1 1
2
1
2
1 1
2
) ( ) ( ) (
(3)
where Y and
i
Y are mean values of total sample scores (grand mean) and samples scores
within subgroups. Correlation ratio is calculated using the following equation as the ratio the
sum of squared differences between group means and the grand mean to total sum of squared
differences between groups means and the grand mean. Unknown coefficients are deter
mined so that the correlation ratio becomes the maximum.
jk
a
max
) (
) (
1 1
2
1
2
2
= =
=
K
i
n
i
K
i
i i
i
Y Y
Y Y n
(4)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
174
Table 3 Item categories
Item Description Category
before 1971 1
19721981 2 Construction years
after 1982 3
1 5 1 Stories above
ground 6 or more 2
0 1
B1 2
Basement
levels
B2 or more 3
less than 80 1
PGV(cm/s)
90 or more 2
Level 1 3 1
Damage level
Level 4 6 2
Implementing some mathematical manipulation, the problem attributes to generalized
eigenvalue proble
Table 4 Results
Item Category
Number
of items
Category score Range
Partial Cor
relation Co
efficient
before 1971 29 0.6630
19721981 23 0.0725
Construction
years
after 1982 20 0.8779
1.5409 0.3426
1 5 28 0.0293 Stories above
ground 6 44 0.0186
0.0479 0.0130
0 16 0.5180
B1 32 0.4176
Basement
levels
B2 or more 24 0.9021
1.4201 0.3359
less than 80 31 0.3918
PGV(cm/s)
90 or more 41 0.2963
0.6881 0.1911
Level 13 39 0.4629
Damage level
Level 46 33 0.5470
Partial correlation coefficient = 0.2568
m.
3.2 Categories for the Quantification II
The original datasets were recategorized for the analysis by quantification II as shown in Table
3. Dummy variables were, then, introduced based on the new category to quantify the qualita
tive data.
3.3 Results
Results calculated by the quantification II are shown in Table 4. Looking at the category scores
of the outside variable (damage level), it is recognized that higher damage level has a lower
category score. Hence, category scores of each item have an inverse relationship (i.e. larger
score contributes to reduce damage). From the values of the range and the partial correlation co
efficient, it is apparent that items of construction year and presence of adequate basements have
about the same value as also shown in Figure 2. This indicates that these items have an influ
ence on damage to the same degree. Paying attention to the category scores of each item, we can
understand that buildings that have two or more basement levels show less damage degree in
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
175
0.0 0.5 1.0 1.5 2.0
PGV
basement levels
stories above ground
construction years
Item Range
Figure 2 Item range
addition to buildings constructed after 1982. Thus, these categories contribute to lessen damage
to building structures.
4 CONCLUDING REMARKS
This study statistically investigated the effect of embedment due to presence of basement on
damage mitigation of reinforced concrete building structures during the 1995 HyogokenNanbu
Earthquake. The results indicate that the effect of embedment has a remarkable influence on re
ducing damage to buildings when there exists sufficient embedment (two or more basement lev
els). The degree of reducing damage by embedment is similar to the effect of construction year
due to the major revision of the Japanese building standard law in 1981. Although this result
may support more positive adoption of embedment effect in design practice in Japan, the fol
lowing two points should be noted: (1) The number of data used in the analysis was only 72
which may not be adequate. (2) The result reflects only a single earthquake event (1995
HyogokenNanbu Earthquake). Therefore, verification of our findings using some other earth
quakes and their damage data is needed in the future.
ACKNOWLEDGMENTS
The authors would like to express their gratitude to Kansai Geoinformatics Network (KGNET)
for providing soil information in Kobe area. The authors would also like to thank Mr. Tsutomu
Seo, Ms. Rie Yoshioka and Ms. Sakiko Tsujino for their support to establish the database.
REFERENCES
Architectural Institute of Japan. 2004. Recommendations for Loads on Buildings: Architectural Institute
of Japan (in Japanese).
Architectural Institute of Japan. 1997. Report of damage to reinforcedconcrete building structures dur
ing the 1995 HyogokenNanbu Earthquake: Architectural Institute of Japan (in Japanese).
Day, S.M. 1977. Finite element analysis of seismic scattering problems, Ph.D. Dissertation, University of
California, San Diego.
Elsabee, F. & Morray, J.P. 1977. Dynamic behavior of embedded foundations, MIT report, R773.
Hayashi, Y., Miyakoshi, J. & Tamura, K. 1997. Study on the distribution of peak ground velocity based
on building damage during the 1995 HyogokenNanbu Earthquake, J. Struct. Constr. Eng., 502: 61
68. (in Japanese)
Hayashi, C. 1952. On the prediction of phenomena from qualitative data on the quantification of qualita
tive data from the mathematicostatistical point of view, Ann. Inst. Statist. Math. 3, 6998.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
176
Global Increase of Natural Disasters and
International Cooperation for Disaster Mitigation
Masanori Hamada
Professor, Dept. of Civil and Environment Engineering, Waseda University, Tokyo, Japan
Xu Wu
Dr. Eng., Technical Center, OYO Corporation, Saitama, Japan
1 INTRODUCTION
ABSTRACT: This paper describes the current situation of global increase of natural disasters
such as earthquake, tsunami, flood and storm. The change of natural environments is increasing
disasters, and the vulnerability of the society proves serious menace to safe and secure life of
mankind. This paper touches upon the basic concept of the policy against natural disasters under
change of natural and social environments. This document also introduces the international co
operation by Japan Society of Civil Engineering and a nonprofit organization, Engineers with
out Borders, Japan for reduction of natural disasters in the world.
Natural disasters such as earthquake, tsunami, storm and flood have been increasing in the re
cent years in the world. There are two main factors underlying the increase of natural disasters.
Those are the natural environmental change and the social vulnerability.
The natural environmental change such as global warming, heat island phenomena in mega
cities, the decrease of the forest, desertification and erosion of rivers, are resulting in extremely
heavy rains and snows, huge typhoons and hurricanes, abnormally high temperature, and high
tidal waves. The social environment is also changing and it is becoming fragile against natural
disasters. Those are highly congested urban areas, depopulation of rural areas, human habitation
on disasterprone lands, lack of cooperation and communication among recent urban societies,
and insufficient infrastructures for the disaster mitigation.
This paper briefly reviews the recent natural disasters in the world and the associated sub
jects, and describes the basic concept of the policy for the reduction of future natural disasters in
future. Furthermore, the authors introduce internationally cooperative activities for the disaster
mitigation and the restoration of affected societies.
2 NATURAL DESASTERS AND CLIMATE CHANGE IN THE WORLD
2.1 Recent Natural Disasters in the World
During recent few years, the disastrous earthquakes and tsunamis have attacked the Asian coun
tries. The 2008 Wenchuan earthquake caused an unprecedented damage to the vast and
mountainous areas of the Sichuan Province of China. The 2004 Sumatra earthquake and conse
quent tsunami killed more than 200,000 people in the areas around the Indian Ocean (JSCE 2005).
In 2005, about 70,000 people were killed in Pakistan (JSCE and AIJ 2005), and in 2006, a
disastrous earthquake attacked the Java Island, Indonesia.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
177
N
U
M
B
E
R
O
F
E
V
E
N
T
S
Asia America
Europe Others
7
6
5
4
3
2
1
0
8
7
6
5
4
3
2
1
0
8
1946 51 56 61 66 71 76 81 86 91 96 2001
50 55 60 65 70 75 80 85 90 95 2000 05
Figure 1 Damaging Earthquakes and Tsunamis in
the World in the Last 60 Years (Events
with more than 1000 casualties)
Figure 1 shows the number of earthquakes and tsunamis with more than 1,000 deaths in each
five years period during the last 60 years (1946~2005) in the world. The number of events has
drastically increased in the last two decades.
The number of earthquakes with magnitudes more than 7.0 and 6.0 in the world during the
last 60 years is shown in Figure 2. On the contrary of the increase of the earthquake and tsunami
disasters, the number of occurrences of earthquakes with magnitudes more than 7.0 has been
decreasing during the last 60 years. The number of the earthquakes with more than 6.0 slightly
increased during the last decade, but was not consistent with the rapid increase of the number of
the earthquake and tsunami disasters. This suggests that the reason of the increase of the earth
quake and tsunami disasters is the increase of the vulnerability of our human societies.
0
20
40
60
80
100
120
140
160
180
0
100
200
300
400
500
600
700
800
900
1946
50
51
55
56
60
61
65
66
70
71
75
76
80
81
85
86
90
91
95
96
2000
01
05
N
u
m
b
e
r
o
f
E
v
e
n
t
s
M
o
r
e
t
h
a
n
M
7
.
0
N
u
m
b
e
r
o
f
E
v
e
n
t
s
M
o
r
e
t
h
a
n
M
6
.
0
51
55
56
60
61
65
66
70
71
75
76
80
81
85
86
90
91
95
96
2000
01
05
1946
50
1946
50
51
55
51
55
56
60
56
60
61
65
61
65
66
70
66
70
71
75
71
75
76
80
76
80
81
85
81
85
86
90
86
90
91
95
91
95
96
2000
96
2000
01
05
01
05
N
u
m
b
e
r
o
f
E
v
e
n
t
s
M
o
r
e
t
h
a
n
M
7
.
0
N
u
m
b
e
r
o
f
E
v
e
n
t
s
M
o
r
e
t
h
a
n
M
6
.
0
b
e
f
o
r
e
D
i
s
a
s
t
e
r
s
E
m
e
r
g
e
n
c
y
R
e
s
p
o
n
s
e
RestorationandReconstruction
Figure 11 Cycle of Disaster Management (Cabinet Of
fice, Government of Japan 2006).
4 INTERNATIONAL COOPERATION FOR NATURAL DISASTER MITIGATION BY
JAPAN SOCIETY OF CIVIL ENGINEERS AND ENGINEERS WITHOUT BORDERS,
JAPAN
As mentioned previously, most of the natural disasters concentrated in the Asian region in the
recent years. Therefore, reduction of natural disasters in this area is one of our primary subjects.
The Japan Society of Civil Engineers (JSCE), which was established in 1914 with a mission that
it should contribute to the advancement of scientific culture and the creation of a safe and secure
society by developing technology and knowledge in civil engineering field. Therefore, the
members of JSCE have been continuously promoting the transfer of technologies and experi
ences for natural disaster mitigation, technical assistance for restoration works for the damaged
infrastructures and societies, and education and training of people for future disaster mitigation
in the world, particularly in the Asian countries.
The Engineering without Borders, Japan (EWBJ), a nonprofit organization was organized by
members of JSCE and the Architectural Institute of Japan (AIJ) with purposes to promote inter
national activities for natural disaster mitigation, with supports governmental institutions and
construction industries of Japan. JSCE and EWBJ have promoted various kind activities for
natural disaster mitigation in the Asian region, which are introduced below.
4.1 Technical Assistances for Restoration and Reconstruction
After the 2004 and 2005 Sumatra earthquakes, JSCE and EWBJ jointly dispatched their teams
to the counties around the Indian Ocean for technical assistances for restoration of the affected
areas and reconstruction of damaged infrastructures. In northern Sumatra, they proposed a strat
egy for reconstruction of the road along the west coast of Sumatra, which was washed away and
many bridges were destroyed by the tsunami.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
183
Steeltrussbridgeswerewashedaway 2004 Sumatra Tsunami 2005 Sumatra Earthquake
Seminar and Workshop after the 2008 Wenchuan Earthquake
Figure 12 Technical Assistances for Restoration and
Reconstruction by JSCE and EWBJ.
In Nias island, where located in the off coast of the west of Sumatra was attacked by the sec
ond earthquake in 2005. The dispatched members by JSCE and EWBJ proposed a plan of re
construction of bridges and buildings damaged by soil liquefaction.
After the 2008 Wenchuan earthquake, JSCE and its associated other engineering and science
societies of Japan dispatched their members for the discussions on the restoration methods of
damaged buildings, bridges, tunnels and slopes with the Chinese counterparts. They introduced
some examples of restoration of structures damaged by Japanese earthquakes including the 1995
Kobe earthquake.
4.2 Educational Activities
Student members of JSCE and EWBJ have continued educational program for pupils of primary
and middle schools in Indonesia after the tsunami tragedy. They are teaching children the
mechanism of the occurrence of earthquake and tsunami, and how to save lives from future nat
ural disasters.
Members of JSCE constructed one hundred monuments in Banda Ache, the tops of which
show the height of the tsunami at each location. The epitaph of the monuments describes the
tragedy and the mourning for the dead. These monuments have been used for the education of
the children.
In Chengdu, Sichuan of China, JSCE and its associated other engineering and science socie
ties started a series of lectures on seismology and earthquake engineering for the students and
young engineers. They are lecturing the system of the earthquake prediction, estimation of
ground motion, earthquake resistant design of structures, and strategies of midand long term
restoration of the affected areas
Series of Lecture on Seismology and Earthquake Engineering
Education Program in Indonesia Monuments of Tsunami
Figure 13 Educational Actives by JSCE and EWBJ.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
184
4.3 Transfer of Technology
Figure 14 shows the tsunami warning system which was proposed by JSCE and EWBJ members
to respond the request from the North Sumatra provincial government.
EWBJ members are also installing seismographs in Padang area of West Sumatra to develop
the earthquake and tsunami warning system, and to survey the characteristics of the earthquake
ground motion in the areas, under cooperation with the researchers in Indonesia.
In Bangladesh EWBJ made a manual for the construction of shelters against future cyclones,
and has been training regional engineers by using this manual under financial aid of Ministry of
Foreign Affairs, Japan. It is planned that 10 thousand shelters will be constructed in the coastal
areas of Bangladesh.
Satellit e
Seismic st at ion
Cent ral disaster
miti gati on center
Broadcast
center
Televi sion Radio
Sat ell ite
Mosque
Figure 14 Proposal of Regional Tsunami Warning Sys
tem for North Sumatra Provincial Government.
5 CONCLUSIONS
In recent years, global societies especially in the Asian region have experienced many nature
disasters such as earthquake, tsunami, flood and storm. These nature disasters produced great
human suffering and economic loss.
Due to the natural environmental change and the social vulnerability, the increasing risk asso
ciated with such disasters seems continue to grow, unless effective measures are developed and
implemented to mitigate the disasters.
Over the last few decades Japan has gathered many invaluable experiences and knowledge in
both the technological and societal fields which can be shared in the natural disasters mitigation.
REFERENCES
[1] Japan Society of Civil Engineer (2005), The Damage Induced by Sumatra Earthquake and Associ
ated Tsunami of December 26, 2004.
[2] Japan Society of Civil Engineers and Architectural Institute of Japan (2005), A quick Report on
Kashmir Earthquake.
[3] Japan Meteorological Agency (2005), A Report on Abnormal Weather 2005.
[4] Japan Society of Civil Engineers (2004), General Investigation about the Damage of Social Infra
structure Systems due to the 2004 Niigata Earthquake / Result of Investigation and Urgent Proposals.
[5] Cabinet Office, Government of Japan (2006), White Paper on Disaster Prevention 2006.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
185
Tsunami Damage Studies and Construction of the Memorial
Poles in Banda Aceh
Hirokazu Iemura
Kinki Politchnic College, Kishiwada, Osaka,Japan
Mulyo Harris Pradono
Agency for the Assessment and Application of Technology, Jakarta, Indonesia
Megumi Sugimoto
Graduate School of Global Environmental Studies, Kyoto University, Kyoto, Japan
ABSTRACT: During a survey led by the first author in Banda Aceh in Indonesia and surround
ing area after the great Sumatra Earthquake, questionnaires have been distributed to the people
affected by the earthquake and tsunami. One important result of the questionnaires shows that
even if people had started running away just after the big earthquake, the percentage of ex
pected survivors would have been less s than 100%. The practical implication from this result is
that education, socialization, escape structures, warning systems, and wave resisting structures
are among important factors for people to be safer against future tsunami attacks. Moreover, for
learning and educational purposes, the memorial poles showing the height of the tsunami attack
were constructed.
1 INTRODUCTION
Huge earthquake and tsunami on December 26th, 2004 has caused a great number of casualties
around Indian Ocean rim countries especially in Banda Aceh city and Aceh Besar Subprovince,
in Sumatra Island, Indonesia. Lessons from this huge disaster shall be learnt by locals and peo
ple all around the world.
In a quick response to the disaster, a Japanese group of researchers led by the first author de
parted to Banda Aceh and surrounding areas in attempt to study the lessons by the huge earth
quake and tsunami. The authors also conducted tsunami questionnaires during their first visit to
Banda Aceh and surroundings area (Iemura, et al, 2006).
2 IMPORTANT RESULTS FROM THE QUESTIONNAIRES
The important results from the questionnaires are that the local peoples knowledge on tsunami
was very low. The question was Did you know that tsunami would come after a big earth
quake? Figure 1 shows the result.
Figure 1. Respondents Knowledge on Tsunami
Kne w if tsunami would c ome ?
No
97%
Yes
3%
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
186
Other important result is that, according to the respondents, even if people had started running
away just after the big earthquake, the percentage of expected survivors would have been less
than 100% (Figure 2). The practical implication is that education, socialization, escape routes,
escape structures, warning system, wave resisting structures are among important factors for
people to be safer against future earthquake and tsunami attacks.
Figure 2. Percentage of Survivors in Banda Aceh City (Number in Parenthesis Shows Percentage of Sur
vivors if People Would Have Escaped Immediately after the Earthquake)
3 TSUNAMI HEIGHT MEMORIAL POLES AND DISASTER EDUCATION
Since tsunami is a rare event, the future generations may forget the disaster easily. People need
encouraging and reminding words that would be written on lasting memorial structures. One of
the ideas prompted by the first author to local people is to build poles with the height of tsunami
run up throughout the affected areas in the city.
The poles have many important purposes: (1) to encourage people to be prepared for the next
one, (2) to keep the memory of tsunami attack, (3) to educate next generation the important les
sons from the tsunami, (4) to mourn the passed away people and to restore and reconstruct
Banda Aceh from the disaster, (5) to keep accurate data of tsunami height for future planning,
(6) to be escaping sign with the tsunami height, (7) to encourage local people to live with hope
and ease under tsunami risk, and (8) to be a symbol of Banda Aceh as the tsunamiattacked city.
The idea was supported by local people and the Embassy of Japan in the Republic of Indone
sia. The Embassy made available a grant up to US $ 91,411 (US $ Ninety one thousand, four
hundredeleven) by March 31, 2006. It contributes to the execution of the construction by the
Yayasan Umi Abasiah. The project entitled The Project for Supporting Education of Tsunami
Disaster Prevention in Nanggroe Aceh Darussalam.
The first author also provides technical assistance. The pole structure should be strong and
yet reasonably economical and feasible in the material availability point of view. Hybrid struc
ture consisting of masonry and reinforced concrete was finally adopted as the material. The first
drawing for designing the poles is shown in Figure 3.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
187
Figure 3 First Drawing of the Pole for Design Purpose
Figure 3. First Drawing of the Pole for Design Purpose
The poles would be built along the evacuation routes. The purpose is for providing evacuees
with escaping sign with height, so that the evacuees escape toward the lower poles. After con
sidering many aspects, including limitation of the land for construction, the realized location of
the poles is as shown in Figure 4. The height of the poles is the inundation of tsunami water,
when the tsunami wave was relatively flat and around ten minutes later, the tsunami water re
turned back to the sea.
Figure 4. Locations of 85 Tsunami Height Memorial Poles in Banda Aceh and Aceh Besar
3
1
17
7
50
51
52
63 61
59
46
13
14
12
11
15
5
4
2
10
68
8
9
18
19
21
44
23
25
1.5 km
North
T TS SU UN NA AM MI I H HE EI IG GH HT T M ME EM MO OR RI IA AL L P PO OL LE ES S
B Ba an nd da a A Ac ce eh h a an nd d A Ac ce eh h B Be es sa ar r, , I In nd do on ne es si i a a
Malaka Strait
Yayasan Umi Abasiah
Banda Aceh,
Indonesia
in collaboration with
Embassy of Japan
in Indonesia
and
Kyoto University
Japan
B BA AN ND DA A A AC CE EH H C CI IT TY Y
41
48
47
49
54
56
57
58
60
45
40
62
16 28
29
6
31
30
37
65
27
26
20
66
22
24
39
Indian
Ocean Jakarta
Singapore
Malaysia
Sumatera
Indonesia
Epicenter, 26 Dec. 2004
Simelue Is.
Nias Is.
Padang
Medan
Banda Aceh
A AC CE EH H B BE ES SA AR R
D DI IS ST TR RI IC CT T
B BA AN ND DA A
A AC CE EH H C CI IT TY Y
8 85 5
8 84 4
8 83 3
8 81 1
8 82 2
8 80 0
7 79 9
7 78 8
7 77 7
7 76 6
7 75 5
7 73 3
7 72 2
7 71 1
6 69 9
7 70 0
6 68 8
7 74 4
32
36
35
33
3
4
67 38
42
43
53 55
64
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
188
Figure 5. The Firstly
Built Tsunami Pole
(February 2006)
Pole No.: 01,
Height: 3.15 m, Dis
tance from Shore:
2.80 km, Location:
Gampong Lamdin
gin, Kecamatan
Kuta Alam, Banda
Aceh.
The placard in front of
the pole shows the in
undation height, dis
tance from shore, time
of tsunami arrival,
time of earthquake, lo
cation name, words of
wisdom, developer,
and donor
Figure 6. The Lowest Tsu
nami Pole
Pole No.: 28, Height: 0.90 m,
Distance from Shore: 4.30 km,
Location: Dinas Pertanian dan
Tanaman Pangan, Banda Aceh.
Figure 7. Tsunami Pole at a School
Pole No.: 15, Height: 1.52 m, Distance from Shore:
3.90 km, Location: SD Negeri 28, Kp. Keuramat,
Banda Aceh
Figure 8. The Tallest Tsunami Pole
Pole No.: 69, Height: 9.00 m (w),
Distance from Shore: 0.50 km, Lo
cation: Masjid Lam Tengoh, Peukan
Bada, Banda Aceh.
The wave height was justified from
witnesses, because there was no in
undation mark left by the tsunami.
Figure 9. Tsunami Pole at a School
Pole No.: 49, Height: 3.80 m, Distance from Shore: 3.10 km,
Location: SDN 2, Punge Jurong, Banda Aceh
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
189
Table 1. List of 85 Tsunami Height Memorial Poles in Banda Aceh and Aceh Besar
1. 3.15 m 2.80 km GPG.
LAMDINGIN KEC.
KUTA ALAM
23. 1.75 m 3.50 km Asrama
Mahasiswa Unsyiah,
Darussalam
45. 2.30 m 3.20 km
Kantor Pengadilan
Negeri Banda Aceh
67. 1.30 m 3.30 km RUMAH
BP. ALAMSYAH UMAR
JALAN SYIAH KUALA
2. 2.90 m 3.00 km GPG.
LAMBARO SKEP
KEC. KUTA ALAM
24. 1.45 m 3.60 km Kantor
Rektorat Iain ArRaniry,
Darussalam
46. 2.00 m 3.50 km
Taman Kanak2
YKA Taman Sari
68. 5.70 m 1.50 km MASJID
BAITUL MAGHFIRAH
PEUKAN BADA
3. 1.40 m 3.75 km MAN 1
25. 2.00 m 3.40 km MAN 3,
RUKOH
47. 3.40 m 3.20 km
SMA NEGERI 1
BLANG PADANG
69. 9.00 m (w) 0.50 km
MASJID LAM TENGOH
PEUKAN BADA
4. 1.95 m 3.40 km SMP
NEGERI 2
26. 1.00 m 4.00 km SMP
NEGERI 18
48. 1.80 m 3.40 km
SMPN 17 BLANG
PADANG
70. 7.00 m 0.40 km MASJID
INDRA PURWA
LAMGURON
5. 1.39 m 3.55 km SD
NEGERI 25
27. 1.80 m 3.85 km SMP
NEGERI 6
49. 3.80 m 3.10 km SDN
2 PUNGE JURONG
71. 2.50 m 2.50 km
MEUNASAH TANJONG
KEC.LHOKNGA
6. 1.84 m 3.70 km Masjid
Agung Al Makmur
Bandar Baru
28. 0.90 m 4.30 km DINAS
PERTANIAN
TANAMAN PANGAN
50. 4.50 m 2.00 km
Rumah Ny. Zulkifli
Narukaya Blang Oi
72. 5.50 m 2.50 km SDN
KAJHU
KEC.BAITUSSALAM
7. 2.60 m 3.60 km SMK
NEGERI 2/STM
NEGERI
29. 1.60 m 3.80 km SMK
NEGERI 3
51. 5.80 m 1.90 km
Masjid Syech Abdul
Rauf Blang Oi
73. 7.00 m 2.20 km Tanah
Widari Dawam Dawood
Sp, Cot Paya
8. 2.60 m 3.40 km SD
NEGERI 80, PRADA
30. 2.45 m 3.50 km Kantor
Dekranas Taman Ratu
Safiatuddin
52. 7.00 m 0.50 km
Rumah Bp. Bachtiar
Zakaria Deah Baro
74. 3.50 m 2.70 km SMAN1
BAITUSSALAM
9. 2.60 m 3.50 km Kanwil
Kehakiman Dan
HAM
31. 2.65 m 3.40 km KAN
TOR BAWASDA
53. 3.90 m 1.80 km
SMPN 11
LAMJABAT
75. 5.10 m 1.50 km MASJID
LAMBADA LHOK
BAITUSSALAM
10. 1.55 m 4.10 km
DINAS KOPERASI
DAN UKM
32. 1.00 m 3.85 km
Direktorat Politeknik
Kesehatan
54. 3.70 m 2.10 km SDN
95 GAMPONG
BARO
76. 4.60 m 2.00 km
GAMPONG LABUI
BAITUSSALAM
11. 3.20 m 3.50 km SMA
NEGERI 2
33. 1.80 m 3.75 km
Politekkes NAD Jurusan
Keperawatan
55. 2.20 m 2.90 km Mas
jid Baitul Muqarra
bin Punge Blang Cut
77. 4.00 m 1.50 km
GAMPONG LAMNGA
JALAN KRUENG RAYA
12. 2.70 m 3.70 km SD
NEGERI 20, POCUT
BAREN
34. 2.00 m 3.40 km SD 35
LAMPRIET
56. 2.20 m 2.90 km SDN
18 PUNGE BLANG
CUT
78. 3.40 m 1.30 km
MEUNASAH NEUHEUN
MASJID RAYA
13. 3.50 m 3.60 km SMP
NEGERI 9,
PEUNAYONG
35. 1.80 m 3.35 km MASJID
JAMIK SILANG
RUKOH
57. 3.40 m 2.70 km
Universitas Iskandar
Muda, Surien
79. 2.20 m 0.40 km
MEUNASAH DURONG
MASJID RAYA
14. 2.57 m 3.90 km
KANTOR BKPMD
36. 1.20 m 3.80 km SMP
NEGERI 8
DARUSSALAM
58. 2.30 m 3.00 km SDN
97 LAMTEUMEN
TIMUR
80. 3.30 m 1.00 km Meunasah
Paya Kameng Masjid Raya
15. 1.52 m 3.90 km SD
NEGERI 28, KP.
KEURAMAT
37. 3.40 m 3.40 km SDN 27
GAMPONG MULIA
59. 2.00 m 3.10 km SDN
93 LAMTEUMEN
TIMUR
81. 3.40 m 0.50 km MASJID
KRUENG RAYA
16. 0.89 m 3.90 km
DARUL ULUM,
YPUI
38. 3.50 m 3.35 km MIN
Merduati Jalan
Malahayati, GP.MULIA
60. 1.40 m 3.30 km
PGSD FKIP
Unsyiah Goheng
82. 3.20 m 0.80 km
GAMPONG MEUNASAH
KULAM
17. 1.52 m 3.70 km SD
KARTIKA XIXI,
LAMPRIET
39. 4.60 m 3.00 km MASJID
AL MUKARRAMAH
GP. MULIA
61. 2.00 m 3.30 km MIN
TELADAN
LAMTEUMEN
83. 3.20 m 0.80 km
GAMPONG MEUNASAH
MON
18. 1.91 m 3.40 km Badan
Perpustakaan
Wilayah
40. 4.50 m 2.50 km Masjid
Tgk. Dianjong
PEULANGGAHAN
62. 1.00 m 3.50 km
Rumah Zakaria
Ismail Lamteumen
84. 2.50 m 0.30 km MASJID
LAMREH KRUENG
RAYA
19. 1.80 m 3.80 km 41. 7.00 m 1.80 km SMPN 63. 2.35 m 3.30 km Biro 85. 3.10 m 0.50 km
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
190
MASJID
LAMGUGOB
12 GAMPONG JAWA Logistik Polda NAD
Lamteumen
Pasantren/Dayah Al
Mahfuzhah Krueng Raya
20. 1.40 m 3.80 km MIS
LAMGUGOB
42. 6.00 m 2.00 km SDN 6
KEUDAH
64. 3.80 m 2.30 km
SMPN 15 LAMPOH
DAYA
21. 1.00 m 3.90 km
STIES/AMBA
43. 8.00 m (w) 1.80 km SDN
8/38 MERDUATI
65. 3.70 m 2.70 km SDN
61 JEULINGKE
22. 1.30 m 3.70 km SD
NEGERI 69,
DARUSSALAM
44. 2.70 m 2.70 km SD
MUHAMMADIYAH
LAMPASEH
66. 3.20 m 2.70 km SDN
106 RUKOH
Notes:
 Data in [m] is height and
data in [km] is distance from
shore
 Poles 43 & 69 show wave
height. Others show inundation
height
 red in Banda Aceh, blue in
Aceh Besar
Figure 10. Tsunami Height Memorial Poles as part of Education Tools on Hazard Mitigation
In May 2007, all of the 85 poles were constructed throughout Banda Aceh city and surround
ing areas. Some photos are shown in Figures 6 to 10 and the list of the poles is shown in Table
1.
The tsunami height memorial poles are also education tools for hazard preparedness. One ex
ample is by showing the students the record of the data mentioned in the poles and why they
were built. The on going survey now is to map the three dimension coordinate of the poles. This
data is invaluable for future tsunami disaster mitigation.
The tsunami height memorial poles will not only provide hazard data dissemination but also
education for the present and next generations. The poles are expected to last long and convey
the messages to every generation.
4 RECOMMENDATION FOR THE RECONSTRUCTION
Other than the tsunami poles and disaster education above (Not forget but understand) above,
institute or center for earthquake and tsunami research, tsunami and earthquake museum
(monuments, facts, data, education materials, and so on), international collaborations among re
search institution, and tsunami and earthquake safe structural design (technologies and codes),
shall be realized.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
191
5 CONCLUSION
1) From the questionnaires about the tsunami damage in Banda Aceh, it was found that local
people did not know the tsunami behavior, even by words.
2) Even if people had started to escape just after the earthquake ground motion, some people
could not have survived because of long evacuation distance.
3) The tsunami memorial poles were constructed with many purposes for future safety of
Banda Aceh city, with financial aid from Japanese people and understanding and support of
local people.
ACKNOWLEDGEMENT
The authors would like to express their sincere gratitude to the Embassy of Japan in the Re
public of Indonesia, to the Japan Society for the Promotion of Science, to the Ministry of Edu
cation, Culture, Sports, Science, and Technology, and to the people of Nanggroe Aceh Darussa
lam, Indonesia, for their strong interest and support.
REFERENCES
Iemura, H., Takahashi, Y., Pradono, M. H., Sukamdo, P., and Kurniawan, R. (2006) Earthquake and
Tsunami Questionnaires in Banda Aceh and Surrounding Areas, Disaster Prevention and Manage
ment, Emerald, Vol. 15, No. 1, pp. 2130.
Agusssalim bin Husen, Thantawi Jauhari, Hirokazu Iemura, Muryo Harris Pradono, Megumi Sugimoto
and Takashi Furuwatari (2007) Tsunami Height Memorial Poles in Banda Aceh for disaster Mitiga
tion and Education Proc. APRU/AEARU Research Symposium, Jakarta, Indonesia, June, 2007.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
192
The Algerian Experience of Covering Earthquake Damages and
Applied Techniques of Reinforcement
Moulay Ali
President General Manager CTC OUEST
Civil Engineer, DEA INSA LYON, PGD IISEEBRI TOKYO
ABSTRACT: Algeria is located on the northern edge of the African plate, which is converging
with the European plate, with a shortening rate of about 48 mm/yr. Northern Algeria is a
highly seismic area, as evidenced by the historical (13651992) seismicity. During the last two
decades, northern Algeria experienced several destructive moderatetostrong earthquakes.
Since 1980, El Asnam earthquake (M
s
7.3), which claimed over 2700 lives and destroyed about
60,000 housings, many moderate, but destructive, earthquakes occurred, such as the Constantine
October 27, 1985 (M
s
5.7), Chenoua October 29, 1989 (M
s
6.0), Mascara August18, 1994
(M
s
5.6), Algiers September 4, 1996 (M
s
5.6), Ain Temouchent December 22, 1999 (M
s
5.6), and
Beni Ourtilane November 10, 2000 (M
s
5.5) earthquakes. On May 21, 2003, the Algiers
Boumerdes region was struck by a magnitude 6.8 (M
w
) earthquake, which caused considerable
damages and claimed over 2300 lives.
The paradox is that this region was trough all ages a convergence point of migration of
populations and a Carrefour of cultural exchange and commercial established currents
especially along the Mediterranean coast.
The historical heritage of many types of ancient constructions and significant cultural
monuments and archaeological sites is remaining as a witness of the melting common history of
generations of populations.
This heritage is in danger. Its exposure to strong earthquake events beside the weight of time
and the carelessness of users are worsening the situation, the assessment is traduced in
catastrophic results. The reserved potential for restoration of existing patrimony is remaining
weak and insignificant. Many cases of collapse occurred and many valorous monuments and
sites did extinguish or disappeared embedded under inappropriate extension works and
anarchical use.
Because of a rather high growth of the population, important developing construction
programs are engaged mainly in this northern part of the country where the climate is mild and
the whole part of populations is concentrated. Huge construction programs are planned to face
the continuously increasing needs for all purposes of use (housing, education, sanitary,
administrations...)
The challenge is to protect both of existing historical sites and representative classified
monuments and buildings and new construction programs from potential natural risks and to
promote a policy of maintenance and restoration of old buildings and specially to keep alive the
historical symbols of the past civilisation and the stamps of historical occupancies (ROMAN
EMPIRE ,OTOMAN EMPIRE, SPANISH, etc)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
193
1 CHARACTERISTICS OF THE ALGERIAN SEISMICITY
The characteristics of the Algerian seismicity are based upon the information recorded in the
historical seismicity, the seismotectonic studies, the seismology, the paleoseismology, seismic
hazard. Today we possess important indications on the seismic activity in Northern Algeria. The
Earthquake monitoring is conduced trough a seismological network implemented with a
countrywide accelerometer reseal.
This activity is essentially focused in the Northern area of the country, although with much
sporadic way, some micro quakes are recorded in the subSaharan side in the Saharan party. In
the Northern area, the Tellian region is the most active one.
The established zoning shows that the regions of El Chelif, Oran and Algiers are the most
threatened areas since in those zones the most important earthquakes occurred.
This seismicity is linked to the complex tectonic movements of the convergence of the
African plate in the south towards the Eurasian plate in the north.
The seismic zoning of the Algerian territory shows that the Tellian strip, notably in its coastal
fringe, is subject to the highest degree of seismic hazard.
The lack of strong ground motion data was significantly experienced when elaborating the
first Algerian aseismic building code in 1976. Revisions were gradually introduced after mean
earthquakes in 1981, 1983, 1988, 1990, 1999 and 2003.
Actually it becomes urgent to set a new policy of urban planning and occupancy of this
important area considering its exposure to the earthquake risks.
The developing construction programs should concern more and more the middle region of
the high plateaux and the large Sahara in the south to fix more and more population
considering the total inexistence of earthquake risk there and the rare presence of consistant
monuments. ( projects of the two main new cities of Hassi Messaoud and Bouzeghoul)
However, the existing constructions of the most important cities and urban areas mainly
concentrated in the northern part of Algeria requires a continuous effort of maintenance and
protection of the established historical patrimony, beside the cumulated economic
infrastructures.
Figure 1. Algeria Erthquaque Hazard Map
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
194
Figure 2. Algiers, Casbah in XVII Century
2 STATUS OF THE EXISTING FRAME IN ALGERIA
The evolution of the housing sector in Algeria remains obviously marked by the history of
the country. The actual stock of houses is consisting of more than 6.5 million units. More than
25 % of existing constructions are threatened by an advanced decomposition process.
The identified construction typologies are closely related to the construction periods.
In such a way the main cities of the northern part are still keeping particular styles of
construction as a heritage of all the historical development of construction technologies.
Thus, the frame heritage is presented as a stratification of construction techniques.
TYPOLOGY OF EXISTANT CONSTRUCTIONS
Figure 3. Typical street and houses arrangement in Medina
2.1 Category A:
Encloses the remaining rural and urban structures of houses and districts built before 1920
(old medina), including chiefly constructions made of adobe, clay, random rocks, with wooden
stiffener: (Casbah of Algiers, Oran, Constantine, Dellys, Cherchel etc)
On the urban scale they are organized in:
BLOCKS: varying number of houses, individualized by communication streets which
surround them. This makes each block reacts independently from the others during the shakings.
THE HOUSES: are semidetached, overlapping and leaning against each other forming a
compact homogeneous unit.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
195
Figure 4. Covered built Circulation Figure 5. Masonry wood reinforced walls
THE SABATS: Number of streets are covered by galleries on top of which the houses extend
and thus creating roofed passageways.
They can be flat with wooden logs incorporated or vaults built out of stones or bricks.
Bearing stone walls are linked one to another by alternate crossing of wood logs
(consolidating the angles) every 50 cm in height, wood logs of about 2 m long.
Partition brick walls are connected to main walls by wood logs.
On a structural scale, floors are built by superposition of two layers of wood logs inserted in
all width of bearing walls (rigid diaphragm).
Corbellings for frontal balcony are set overhanging and supported by wood logs forming an
angle with bearingwalls.
Wooden framing is found around the openings.
2.2 Category B:
Buildings with loadbearing walls: carved rocks, stone and vaulted metallic flours.
This type of construction represents 90% of old urban construction frames (colonial texture)
which constitute the main central heart of the towns in major important cities like Algiers, Oran,
Constantine, Annaba. Generally, the predominant type of building is raised on 01 to 05 floors.
This class is in a rather altered state, with an average of age reaching the century.
2.3 Category C:
Early in the sixties, new constructions were set on the existing urban occupancy mainly made
of columnbeam reinforced concrete frames which represent the major part of individual private
houses and public buildings in the new urbanized areas around old centers of the cities. A
massive exude of population from rural areas to the urban sites chiefly for economical trend and
employment The main reason
A great part of this class of constructions suffered hardly under destructive earthquakes
occurred in main old towns ( El Asnam OCTOBER 10,1980 (M
s
7.3), Constantine October 27,
1985 (M
s
5.7), Mascara August18, 1994 (M
s
5.6), Ain Temouchent December 22, 1999 (M
s
5.6),
ALGIERSBoumerdes May 21, 2003(M
s
6.8).
Although many of them seem moderate, the poor quality of construction and bad soil
conditions were the main reasons of massive destructions and wide collapse.
The cases of Ain Temouchent city extended in a river bed or the surrounding districts of
Algiers built in an anarchical way with a very poor quality of materials remain typically
representative of the fragility and the extreme vulnerability of this urban texture category.
The behavior of this class of construction under earthquake has been traducing the lack of
the seismic code and the inexistence of quality control procedures.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
196
2.4 Category D:
This category consists of rigid panels concrete construction systems. This technology has
been applied intensively after El Asnam earthquake in 1980 and the standard revision
recommendations consequently introduced.
2.5 Category E:
New designed important structures are nowadays more and more engineered with regards to
advanced earthquake resisting methods and various construction processes.
From these categories, those of A & B are usually called old constructions
This heritage is in danger. Both of the weight of time and the carelessness of users are
worsening the situation, the assessment gave catastrophic results.
The requalification of this old heritage should inevitably goes through a multi dimensional
scientific step (Urban, technical, social and economical). The urban fabric includes a certain
number of buildings that represents a culture and the marks of countrys History.
The indepth study of the old constructions as an architectural & urban creation holds the
marks of a particular civilization and a specific transformation; the old districts are
sentimentally more attractive than the new or modern ones, and more pleasant in appearance.
This leads back to sociological and cultural referents.
The consciousness of the critical status of the old constructions showed up the importance of
saving it. This requires a global strategy of intervention and specific means given the extreme
fragility of the traditional urban fabrics.
The status of these constructions and the high density of population give a particular
vulnerability to the Algerian great metropolis.
Indeed, 30% of the population lived in urban areas in 1966, 62% in 1998 and according to
the latest RGPH results; more than 80% in 2008. The towns of more than 100 000 inhabitants
jump from 3 in 1962 to 32 in 1987 and more than 60 nowadays.
10 have more than 200 000 inhabitants. The urban areas which are the result of conurbations
and the densification of the urban networks become larger and larger, while the general
population of Algeria had tripled during that period. The urban population from its side was
multiplied by 10.
The public authorities are interested by the old constructions; this interest is materialized
today by lots of rehabilitations conducted over buildings and notably over historical sites and
monuments, we cite some instances regarding that in what follows:
3 EXPERTS RENOVATION EXPERIENCES CASES
3.1 Rais Palace (Bastion 23) ALGIERS
Figure 6. Bastion 23 General plan distribution
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
197
Figure 7. General view before intervention
Summary sheet of building revamping:
Country : Algeria Written by: BERDJANI Mustapha
Monument : Ras Palace (Bastion 23)
Identification:Description of the original building
Revamping goals: To restore Ras Palace (Bastion23) and to create the centre of arts and
culture in order to promote arts & culture for a large public.
Address: 23, AMARA Rachid Boulevard  Bab El Oued Algiers.
Location of the Building: Included in the saved perimeter of Algiers Casbah (World
classification in 1992)
Built environment: Weak density
Protection for Building: Yes, since 1909 (carried on by the
Law n67281 of 20/12/1967)
Owner: National agency of archeology (Ministry of Culture)
Contractor: SCIMBM Contractors (Italy)
Main companies: SCIMBM Contractors (Italy), Company of creation of cultural heritage
(ERPC)/ Algerian architects.
Date of renovation: Novembre1987 End 1994
Overall cost of the operation (Taxes included): Twelve million of dollars (12.000.000 $)
Financing: Equipment budget financed by the government.
Documentation source: Study performed by (YapiMerkezi) Turkish engineering office.
Archeological research (Algerian team) written data, photographic and iconographic.
Bastion 23 is an architectural set composed of palaces and houses of Moorish type. It was
part of the urban fabric of the ancient town, which was surrounding a set of districts that
constituted ElDjazair town at the Ottoman time between the XVI XIX century.
Walls: Mixed bearing walls, made of bricks and stone masonry.
Floor: Traditional floor made of thuyas log with marble and baked clay covering.
Waterproof quality: Sand covering, lime and baked clay mortar.
Outer covering: Coating used: sand, traditional lime.
Openings and front elements: Faades with discrete elements toward outside.
Water supply: Rain water collecting system by canalization made of baked clay as well as the
presence of three wells.
Purification system: Peripheral gutter system.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
198
Figure 8. Site works view
Figure 10. Architectural improvement of the main side Figure 9. Internal decoration of
Diagnosis conducted on constructive elements:
Walls and columns: Analysis of the structure and materials.
Floors: Analysis of the structure and materials.
Roofing: Analysis of the impairment status of the traditional waterproof quality.
Coverings: Analysis and localization of the altered areas.
Renovation works
Walls:Reinforcement of the existent walls
in order to consolidate and stabilize the structure and the rebuilt of the collapsed walls.
Floor:
Reconstruction of floors at 100%, with the reuse of bearing elements that was in good
conditions.
Waterproof quality:
New layers have been applied Pose (sand, cement, baked clay)
Coverings:
General cleaning of the impaired coatings. A Treatment of all deep and superficial cracks has
been conducted. Marble, baked clay and schist have been used.
Paving:
Lying of paving on outside areas.
Equipments:
Renovation of different networks (electricity, gas, AEP, phone network, purification system)
Outer arrangement:
The outer area was arranged as a parking with a covering made of cobblestone Based on the
overall analysis of all main characteristics of the various components, a multidisciplinary
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
199
Figure 11. Protection retaining wall
Figure 12. General view after completion
Turkish engineering team (Yapi Merkezi), has elaborated a synthesis and an assessment of the
building.
Restoration, renovation and arrangement of Monuments surrounding areas. Excavation
works had been conducted in order to determine the exact plan of the Monument, the
reinforcement of walls, reconstruction of collapsed parts (faade walls etc), the treatment of
faades and architectural corrections after restoration and consolidation of the entire original
structure, renovation of different networks (electricity, gas, telephone network, purification
system), mending the pavement of the soil surface.
Evaluation of the results
Integration of the building: after nearly seven years of restoration, Bastion 23 became the
urban or scenery environment: it was the first experience of retrieving a historical and presented
to the public.
Compliance with the initial program: The program that has been applied in this project is
regarded as a public one. Its conversion to a museum and to an Arts & culture centre contributes
to the promotion of culture and to keeps alive the memory of Algiers.
Compliance of usage: Air conditioning, fire security and all the necessary comfort.
Social balance: Employment source, a place for meetings and cultural exchange.
Bastion 23 plays a sociocultural & economic role.
Maintenance of the building: The Arts & Culture center is permanently kept under a
continuous and global maintenance program.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
200
Figures 13, 14. Walls cracks repairing and reinforcement
Figure 15. Reinforcement of poof dome support and stone walls
3.2 Abdellah Ibn Salem Mosque restoration
The construction of Abdellah Ibn Salem mosque goes back to the colonial period and its
inauguration dates from the year 1908.
Abdellah Ibn Salem Mosque is one of the historical sites in Oran city thanks to its arab
moorish architectural style.
The mosque was subject to time vagary and the lack of maintenance, it showed important
degradations and this lead the public authority to decide its rehabilitation.
The bearing structure is composed of bearing stone walls with vaulted veined bricks by IPN
steel profiles. The prayer room which occupies the central part is protected by a false ceiling in
half circle made of lathing reeds with plaster covering surmounted by a double side tiles roofing
which lays over mixed steelwood framework.
The two minarets are placed over the two angles of the main faade are made by rock blocks
of large seize linked by a hydraulic lime mortar.
The restoration works consist of:
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
201
 The preparation of cracked and broken walls.
 The rehabilitation of damaged vaulted roof.
 The rehabilitation of impaired wooden framework elements.
Figure 16. Internal view of the buildings Figure 17. Waterproofing coating
protection
Figure 18. Original view of Arenas Figure 19. Before renovation
3.3 Orans Arenas rehabilitation
The arenas have been built between 1910 and 1911.In the 40s, Oran was still more Spanish
than Arab. The city was founded by the Andalousians in the year 903; the second
Spanish occupation does back to 1774 and after the French conquest, the city kept without
doubt the Spanish urban stamp.
In the 50s, a big majority of the inhabitants of ORAN were from Spanish origin. Some are
political refugees, but most of them were staying there since many generations
The best world toreadors came to perform in these large arenas which could contains
thousands of persons. They confronted the Toros of the most famous Spanish Cowboy boots
or herd of bulls.
This is not surprising at all, especially with the presence of the Eckmhl site arenas found in
the Southern working class area in ORAN, the tradition of bullfighting is still felt there
This imposing architecture which was abandoned and wrongly used had to wait for the
heritage to retrieve its place in the cultural priorities, and then Orans arenas were
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
202
programmed for the rehabilitation of the heritage sites of the local council program, while
waiting to be indexed in the list of classified sites.
Since then, local concerned authority started to look over this site and planed reinforcement
and renovating operation
STRUCTURAL DESCRIPTION OF THE ARENAS
The arenas have a circular form, composed of a hybrid structure made of stones, which are in
staggered rows, and reinforced concrete for the extension added part.
It is founded on a channeled linear feet. The link between the different stones is ensured by a
lime link mortar.
The brace of the external and internal surrounding walls of the arenas in the radial
direction is ensured by thick masonry shear walls in form of a multi span stone arc that lies over
stone poles.
Discharging arches built out of stones or bricks are set in a circular arrangement of a thick
framework, to allow transfer of horizontal loads to the ground.
Construction is erected as a compact dynamic block.
IPN steel profile tie beams are arranged at different highs to ensure the rigidity of stone poles.
The hollow block roofing lies directly over a rack which at its turn lies directly over the brace
rock walls. Circular openings are laid out at the level of the surrounding walls of the arenas in
order to ventilate the place and avoid the condensation of the air saturated by water on the walls
(humidity). The access to the terraces for the public is ensured by stairs of stone masonry.
The hollow block roofing lies directly over a rack which at its turn lies directly over the brace
rock walls.
Figure 20. Supporting walls and arch distribution
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
203
Figure 21.22.23 Repairing works (concrete Jacketing)
4 CONCLUSIONS
Two main reasons has been developed as an explanation of omitting to maintain the heritage
patrimony and monuments in acceptable environmental conditions and stable configuration.
The first is the great demand of housing and construction programs concentrated in the
existing metropolis in the north putting all the potential on this aim.
The second cause of deterioration is due to age effect and the weight of time added to
destructive sporadic earthquakes increasing the secular collapse process.
A new appearing reason is an economical tendency, thus the price of land inside urban areas
has grown to incredible values in main cities due to the high density of occupation of the soil.
This is actually encouraging the squatting of the latest free area of land after total collapse of
existing constructions and ancient monuments.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
204
REFERENCES
Architectural project of renovation of BASTION23 from Ministry of Culture by Turkish engineering
office : YapiMerkezi
Project undertaken with the help of The European Union, Euromed HERITAGE and The International
Cooperation Agency of Spain (AECI).
Architectural design of Ibn Salem Mosque by IBDAA Architects Oran 2007
Design architectural group of BARCO Architects Oran for Arenas Project 2008
CTC OUEST diagnostic and expertise works on the three Exposed cases 2003 2006 2008
CTC reinforcement methods and applicable renovation techniques 2006
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
205
1
1 INTRODUCTION
Nonlinear dynamic response is one of the distinguished characteristics of the ground under
strong shaking. The nonlinearity has been reported and investigated by many researchers for
more than 40 years. The investigation has been mainly concerned with the following two sub
jects;
(1) Shift of predominant frequency of acceleration towards lower frequency [e.g., Idriss and
Seed (1968), Idriss (1990)].
(2) High spikes on horizontal acceleration time histories due to cyclic mobility in relatively
dense sand deposit [e.g., 1987 Superstition Hills earthquake (Holzer, et al. 1989), 1993 Ku
shiroki, Japan, earthquake (Iai, et al. 1995), 1994 Northridge earthquake (Bardet and Davis
1996), 1995 Hyogoken Nanbu, Japan, earthquake (Iwasaki and Tai 1996), 2001 Nisqually
earthquake (Frankel, et al. 2002), 2004 Niigataken Chuetsu, Japan, earthquake, 2007 Niiga
taken Chutsuoki, Japan, earthquake].
The phaseshift of predominant frequency (1) is caused by inelastic response of the ground
material against large ground strains. This nonlinearity can be indirectly observed by comparing
the frequency transfer functions between small and large earthquakes as a shift of the peak fre
quency of large earthquakes towards lower frequency. In recent studies, to derive such a transfer
functions, time histories obtained from seismometers forming the vertical array have been inten
sively employed, with which simultaneous recordings of the base at certain depth and surface
motions are achieved. Mechanism of the nonlinearity has been investigated through laboratory
experiments for soils under cyclic loadings, and it has been understood that large cyclic shear
strain amplitude causes degradation of soil stiffness. Because of this type of stiffness degrada
tion predominant frequency of the ground is shifted toward lower ones.
Highspikes on horizontal acceleration records (2) are typically found in locations where the
ground material constitutes dense saturated sand. Results of stress controlled undrained cyclic
Numerical Analysis of NearField Asymmetric Vertical Motion
T. Tobita, S. Iai & T. Iwata
Disaster Prevention Research Institute, Kyoto University, Kyoto, Japan
ABSTRACT: An unprecedented vertical surface acceleration of nearly four times of gravity was
recorded during the 2008 IwateMiyagi Inland, Japan, earthquake (Mw 6.9). The motion was
recorded at the KiKnet, IWTH25 station, located 3 km southwest of the epicenter. The station
is equipped with threecomponent accelerometers, installed at both the freesurface and the bot
tom of a 260m borehole. Wave form of the vertical acceleration shows clear asymmetric form
with large amplitude in upward direction. To study this recently discovered nonlinear behavior
of the surface ground motion, numerical analysis with the finite element method has been con
ducted with parameters derived from the borehole data at the station. The analysis successfully
simulates the asymmetric vertical motion. The analysis indicates that the asymmetric motion
may be characterized by the existence of lower bound of negative acceleration and high positive
pulses caused by the compression stress of the disturbed surface ground material.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
206
2
triaxial tests for dense sands show increase of volumetric strain due to stress dilatancy at large
strains. This increase of volumetric strain causes rapid reduction of the excess pore water pres
sure, and results in the rapid recovery of shear strength. When this phenomenon occurs in real,
the rapid increase of shear stress may be observed as spiky response on acceleration records
(Ishihara 1993). When the spikes appear, other frequency components are attenuated due to
build up of excess pore water pressure reaching to liquefaction state. These spikes may give
overestimation of the maximum acceleration amplitude and duration of shaking (Bonilla, et al.
2005).
In addition to the above mentioned two types of nonlinear response of the ground, the 3rd
nonlinearity under strong ground motion has been recently discovered as,
(3) Asymmetric from of the surface vertical acceleration amplitude (Aoi, et al. 2008).
In the 2008 IwateMiyagi Inland, Japan, earthquake, very large acceleration amplitude which
exceeded 40 m/s
2
(3 component combined) was recorded at the KiKnet, IWTH25 station (Fig.
1 and 2). The site is located about 3 km southwest of the epicenter on the hanging wall of the
seismic fault. The earthquake was caused by the inland reverse fault with 30 km strike and 20
km in depth. At IWTH25 station, 3 components seismometers are installed at the surface and
G.L. 260 m. The site is located in volcanic region and locally it is on the river sediments under
lain by igneous rocks, such as tuff. The recorded wave form of the surface vertical acceleration
has large amplitude only in positive direction. The maximum amplitude of vertical acceleration
is 4 times larger than acceleration of gravity and 2 times larger than its horizontal components.
Aoi, et al. (2008) reported and qualitatively explained the mechanism of this phenomenon by
the analogy of bouncing a matter on a trampoline and called the Trampoline effect.
The asymmetric form may be attributed to physical characteristics of granular media, which
shows asymmetric response against normal compression and extension force. That is, granular
media, such as dry sands, has less resistance against tension force. Objective of the present
study is to simulate this asymmetric response of the surface vertical acceleration and give physi
cal background to explain this phenomenon.
Figure 1 Location of the epicenter of the 2008 IwateMiyagi Inland, Japan, earthquake and IWTH25,
KiKnet station.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
207
3
Figure 2 Measured time histories and Fourier amplitudes of surface acceleration at IWTH25 after the
2008 IwateMiyagi Inland, Japan, earthquake. Asymmetric amplitude appears on the UD record.
2 ANALYTICAL MODEL
Numerous analytical models which simulates the nonlinear dynamic ground motion have
been proposed. Equivalentlinear method [SHAKE (Schnabel, et al. 1972)EERA (Bardet, et
al. 2000)] has been widely used for its simple configuration and precise results for low to inter
mediate levels of ground shaking. To directly simulate nonlinear stressstrain response of soils,
the HD model (Hardin and Drnevich 1972) and the RO model (Ramberg and Osgood 1943)
have been widely used. By combining these nonlinear constitutive models with Masings law
(Masing 1926), hysteresis response of soils can be simulated [e.g., NERA (Bardet and Tobita
2001)]. Iwan (1967) and Mroz (1967) proposed a model for soil nonlinearlity and hysteretic be
havior by using multiple springs and sliders.
By taking advantage of plasticity theory started in the field of metal engineering, constitutive
models based on the theory of plasticity have been formulated and widely used in the frame
work of finite element analysis (Roscoe, et al. 1963, Roscoe and Burland 1968, Matsuoka 1974,
Dafalias and Popov 1975, Sekiguchi and Ohta 1977, Mroz, et al. 1978, Hashiguchi 1980,
Towhata and Ishihara 1985, Nakai 1986). These models generally require many parameters,
which are more or less derived from experimental results, to simulate more realistic and compli
cated behavior of soils.
To simulate soil liquefaction, dynamic solid and fluid coupling behavior has to be incorpo
rated into constitutive modeling. Constitutive models based on the effective stress concept have
been proposed for such a material (Martin, et al. 1975, Prevost 1985, Iai, et al. 1992, Oka, et al.
1999). In these models, by applying dependency on the effective mean stress to shear behavior
of soils, liquefaction as well as dilatant behavior can be successfully simulated.
In this study, the multiple simple shear mechanism proposed by Towhata and Ishihara (1985)
and Iai, et al. (1992) is implemented.
2.1 Multiple simple shear mechanism
The finite element code called FLIP (Iai, et al. 1992) is implemented for nonlinear site re
sponse analysis. Total stress analysis is conducted, i.e., no excess pore water pressure buildup
during shaking is assumed. The code utilizes the multiple simple shear mechanism as nonlinear
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
208
4
constitutive relation (Towhata and Ishihara 1985). In this model, contact forces between sand
particles are idealized by evenly distributed multiple springs whose property is characterized by
the hyperbolic type (Fig. 3). The model automatically accommodates the principal stress rota
tion which plays an important roll in the cyclic behavior of anisotropically consolidated sands.
In what follows, core of the modeling is briefly introduced. Details are found in Ozutsumi
(2003). Let us consider the following vectors of stresses and strains,
{ } { } ' ' ' '
T
x y xy o o o t = (1)
{ } { }
T
x y xy c c c = (2)
where ' x o and ' y o are the normal stress, x c and y c are normal strain, 'xy t and xy
are shear stress and shear strain, respectively. Incremental form of constitutive relationship can
be written as follows,
{ } { }{ } { }
/ /
'
T
L U L U
d R n n d o c = (3)
where the vector { }
/ L U
n specifies direction of stress increment, the scalar
/ L U
R defines mag
nitude of stress increment per unit strain increment along the direction { } n , vector { } n gives
the direction of strains. The subscripts L/U indicate the components are different in the direction
of loading L and unloading U. Iai et al. (1992) postulated that the incremental constitutive
relation, Eq. (3), is given by I+1 separate mechanisms for i=0 to I in associated form, i.e.,
{ } { }
/ L U
n n = , as
{ } { }{ } { }
( ) ( ) ( )
/
0
'
I
T
i i i
L U
i
d R n n d o c
=
=
(4)
To specify the volumetric behavior, the first term of Eq. (4), i=0, is independently defined so
that Eq. (4) becomes,
{ } { }{ } { } { }{ } { }
(0) (0) ( ) ( ) ( )
/
1
'
I
T T
i i i
p L U
i
d K n n d d R n n d o c c c
=
= +
(5)
where K is rebound modulus and
{ }
p
dc is an additional volumetric strain increment due to di
latancy given by
{ } { }
/ 2 / 2 0
p p p
d d d c c c = (6)
Direction vectors are given as
{ } { }
(0)
1 1 0
T
n = (7)
{ } { }
( )
cos cos sin
T
i
i i i
n u u u = (for i = 1 to I) (8)
where i u is the angle of the ith spring from the horizontal axis, and given as,
( 1) i i u u = A (for i = 1 to I) in which / I u t A = .
Rewriting Eq. (5) yields,
{ } { }{ } { }{ } { } { }{ } { }
{ } { }{ } { }
(0) (0) ( ) ( ) ( ) (0) (0)
/
1
(0) (0)
'
=[D]
I
T T T
i i i
L U p
i
T
p
d K n n R n n d K n n d
d K n n d
o c c
c c
=
(
= +
(
(9)
in which matrix [D] is,
{ }{ } { }{ }
(0) (0) ( ) ( ) ( )
/
1
[ ]
I
T T
i i i
L U
i
D K n n R n n
=
= +
(for i = 1 to I) (10)
Rebound modulus K is expressed as
0
0
'
'
m
K K
o
o
 
=

\ .
(11)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
209
5
where K
0
is the rebound modulus at a reference confining stress
0
' o , ' o is an effective stress,
and m is a parameter which correlates K and stress ratio
0
'/ ' o o . In the frame work of the mod
eling, tangent shear modulus is given as,
( )
/
( )
i i
L U i i
dF
R d
d
u
= A (12)
in which the function ( ) i F is a hyperbolic function which is equivalent to the spring force
due to the displacement of a multispring in one radian. By assuming Masings law for cyclic
response and defining damping of spring, the shape of function ( ) i F is determined (see
Ozutsumi (2003) for detail).
In each shear mechanism, loading and unloading conditions can be defined by the sign of,
{ } { } ( )
( )
cos sin
T
i
i x y i xy i
d n d d d d c c c u u = = + (13)
where
{ } { } ( )
( )
cos sin
T
i
i x y i xy i
n c c c u u = = + is the virtual shear strain of the mechanism
i= 1 to I. Each tangent modulus depends on the current state and history of each virtual simple
shear strain
i
.
In matrix form, Eq. (10) can be written as,
  1 2 3
1 1 0 1 1 0 0 0 1 0 0 0
1 1 0 1 1 0 0 0 1 0 0 0
0 0 0 0 0 0 1 1 0 0 0 1
D K G G G
       
   
= + + +
   
   
\ . \ . \ . \ .
(14)
where
2
1
1
( )
cos
I
i
i
i
dF
G
d
u u
=
= A
(15)
2
1
( )
cos sin
I
i
i i
i
dF
G
d
u u u
=
= A
(16)
2
3
1
( )
sin
I
i
i
dF
G
d
u u
=
= A
(17)
In the finite element code, FLIP, if the volumetric strain is positive, judged by the following
relation,
( ) 0
x y p
c c c + > (Tension) (18)
then, all the stress components are set to be zero, i.e., ' ' ' 0
x y xy
o o t = = = . As it will be shown
in the next section, this is the property that simulates asymmetric behavior of granular media
with no cohesion.
Figure 3 Schematic view of the multiple simple shear mechanism
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
210
6
2.2 Model behavior under normal compression and tension stress
Behavior of the multiple simple shear model is investigated by applying vertical compression
and tension forces to a single element (Fig. 4). Model parameters for this particular study are,
shear modulus=3.210
5
(kPa), Poissons ratio=0.33, density=1.8 (t/m
3
), and friction an
gle=35. Computation was carried out under drained condition. Firstly the element is isotropi
cally consolidated with stress o
0
= 49 (kPa) with nodal constraints shown in Fig. 4(a). Then, it is
stretched and compressed vertically by enforcing cyclic normal force to nodes 1 and 4 (Fig. 4b).
Here, the horizontal displacement of nodes 1 and 4 are constrained (Fig. 4b). Analytical time
step of At=10
4
was employed. Figure 5(a) depicts time histories of mean stress,
o
m
=(o
xx
+o
yy
)/2, and volumetric strain, c
v
=c
xx
+c
yy
, on the left and right vertical axes, respec
tively. For both strains and stresses, tension is taken as positive. As shown in Fig. 5(a), when the
mean stress is zero, tensile strain abruptly increases, that is, the element is vertically stretched.
Stress strain curve in Fig. 5(b) gives another view of this property. Namely, the curve starts at
the point indicated as 1 in Fig. 5 where mean stress is 49 kPa, then gradually stretched and
once the mean stress reached to zero at 2, volumetric strain abruptly increases up to about
0.3%. In the compression side, in the given range of enforced vertical stress, the stressstrain re
lationship is within an elastic range following the line whose slope is defined by the rebound
modulus of K=605 MPa.
Figure 4 Illustration for application of isotropic confining pressure (a) and cyclic axial stress (b) to a sin
gle element.
Figure 5 Time histories of (a) mean stress and volumetric strain, and (b) mean stress versus volumetric
strain curve.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
211
7
2.3 Model behavior against vertical input acceleration
To simulate the asymmetric vertical ground motion with simple model, 1D column of 10 m
deep with 20 elements are implemented. Each mesh is a square with 0.5 m at the side. Dis
placement degree of freedom of nodes at the bottom mesh is fixed in both horizontal and verti
cal directions. A pair of nodes at the same height was given the same displacement degree of
freedom to simulate halfspace ground motion. Analysis is carried out by assuming dry condi
tion. Sinusoidal input acceleration with 10 Hz and the amplitude of 19.6 m/s
2
(=2 g) is given at
the base of the model. In this study, only the vertical motion is given. Computational time step
of At=0.001 (s) was employed.
Figure 6(a) is a section of the time histories from 0.5 to 0.8 (s) of volumetric strain (c
v
=c
x
+c
y
),
input and surface accelerations. In the same manner, Fig. 6(b) shows mean stress
[o
m
=(o
x
+o
y
)/2] of the surface element, in which vertical scale of input and surface accelerations
is magnified 8 times for clarity. Here, tension is taken as positive. In Fig. 6(a), tensile strain can
be observed while tensile stress is not generated [Fig. 6(b)] as it was shown in the previous sec
tion. As shown in Fig. 6(b), surface acceleration start to abruptly increase when the input verti
cal motion becomes lower than 9.8 m/s
2
. This is because, at this moment, the surface element is
free falling and compression force is generated in an element, and reaction to this compression
force may create acceleration pulse in the vertical upward direction. Again from Fig. 6(b), lower
bound of surface acceleration is limited by the minimum amplitude of vertical input accelera
tion. Thus, the asymmetric motion of the vertical acceleration can be characterized by the exis
tence of lower bound of negative acceleration and positive pulse in the vertical acceleration due
to compression of soils.
Figure 7 compares mean stress and volumetric relationship of the surface element for various
amplitudes of sinusoidal vertical input accelerations. As shown in Fig. 7(a), when the vertical
input acceleration is small, response is in the compression side and follows the line whose slope
is defined by the initial elastic rebound modulus. As amplitude of input acceleration increases,
the curve starts to show nonlinearity. If the amplitude of input vertical acceleration exceeded 9.8
m/s
2
, tensile strain appears while mean stress is kept to be zero.
Figure 6 Time histories of (a) volumetric strain, vertical input and vertical surface acceleration, (b) mean
stress, vertical input and vertical surface acceleration of the surface element. Scale of acceleration in (b)
is magnified 8 times of that of (a).
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
212
8
Figure 7 Mean stress versus volumetric strain relationship of the surface element for sinusoidal input ac
celeration with various amplitudes: (a) 4.9 m/s
2
, (b) 9.8 m/s
2
and (c) 19.6 m/s
2
.
3 NUMERICAL ANALAYSIS
3.1 Model parameters for site response analysis of KiKnet, IWTH25 station
Physical model parameters, such as, shear modulus, poissons ratios, and densities, are esti
mated from the PS velocity profile obtained at the KiKnet, IWTH25 station. Table 1 summa
rizes model parameter implemented in the following analysis. Based on the velocity profile,
model ground is divided into 7 layers from the surface to G.L. 242 m (Fig. 8). From P and S
wave velocities, shear modulus and poissons ratios of each layer can be estimated. Borehole
profile deeper than 120 m shows that the tuff whose S wave velocity exceeds 1,300 m/s is pre
dominant. Above it, mud stones and terrace deposit constitute the layers. Densities of material
are given as follows; 2.2 t/m
3
for surface cover soil, 2.3 t/m
3
for sediments, and tuff and mud
stones are set uniformly 2.6 t/m
3
(Toyota 2009).
In the analysis, the ground is modeled with single column elements (133 elements) whose
width is 1 m, and height varies from 0.2 m (layer 1) to 5.43 m (layer 7). With this model
ground, frequency component of the maximum 50 Hz can be properly simulated unless being
attenuated.
Table 1 Model parameter of the ground profile at the KiKnet, IWTH25 station
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
213
9
0
80
160
240
0 4500
Velocity (m/s)
D
e
p
t
h
(
m
)
P
S
Layer 1
Layer 2
Layer 3
Layer 4
Layer 5
Layer 6
Layer 7
Figure 8 P and S wave velocity profile at IWTH25 station (NIED 2009) and assumed layering system for
numerical analysis.
3.2 Site response analysis with small input motion
Site response analysis with small input motion is conducted to examine the model parameter
defined above (Table 1). The small input motion used in the analysis (M
w
=6.8, Depth 108 km
on 24/7/2008) was recorded at the same site, IWTH25, where large acceleration was measured
during the 2008 IwateMiyagi Inland, Japan, earthquake. In the analysis, both the horizontal
(NS) and vertical motion (UD) were simultaneously input at the base of the model. The meas
ured maximum accelerations of NS and UD components at the base are, respectively, 1.2 m/s
2
and 0.83 m/s
2
and at the surface, 1.1 m/s
2
and 0.71 m/s
2
, respectively. Time step of the analysis
was At=0.005 (s).
Computed time history of the horizontal motion (NS) in Fig. 9 is in good agreement with the
recorded motion in terms of both maximum amplitude and the shape of envelope, while the
maximum amplitude of vertical motion is over estimated about 40%. As shown in Fig. 9, in the
case of small input motion, amplitude of both measured and computed vertical motions are
symmetric.
Transfer functions of both horizontal and vertical motions in Fig. 10 show significant agree
ment below 10 Hz. However, for response above 10 Hz is under estimated. This may be because
high frequency components in the input acceleration were too small to be transmitted to the sur
face without attenuation. In reality, there are many sources of high frequencies in the ground,
such as inclined boundaries between soil stratum and/or heterogeneity, which may be properly
included in twodimensional analysis.
Figure 9 Measured and computed surface acceleration (NS and UD components) for small input motion.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
214
10
Figure 10 Transfer functions of measured and computed acceleration for small input motion.
3.3 Site response analysis for the 2008 IwateMiyagi Inland, Japan, earthquake
Site response analysis is conducted to simulate asymmetric vertical motion observed in the
2008 IwateMiyagi Inland, Japan, earthquake. Recorded input motion of both horizontal (NS)
and vertical (UD) components at the base of the KiKnet array at IWTH25 was input at the base
of the model ground with time step of 0.01 s which was the same time step as measured accel
eration. As shown in Fig. 11, computed surface acceleration has similar asymmetric shape in the
vertical component as measured one, i.e., positive spikes and bounded negative amplitude. The
maximum amplitude of measured vertical acceleration was 33 m/s
2
, while the one computed
overestimates about double to be 62 m/s
2
. Negative amplitudes of surface vertical acceleration
show, on average, significant agreements between observation and analysis, and it is close to the
acceleration due to gravity. This might indicate the temporal freefall of the ground.
Transfer function of computed vertical acceleration [Fig. 12(b)] shows a peak at 5 Hz which
is not seen in the observation. Cause of this is under investigation. Except the peak, transfer
function is over all in good agreement. Higher frequency components are properly simulated
because input acceleration might be large enough for higher frequency components to be trans
mitted to the surface. Transfer functions of horizontal motions also show in good agreement
[Fig. 12(a)]. However, there are no clear peaks in the transfer function of computed horizontal
motion.
Figure 11 Measured and computed surface acceleration (NS and UD components) for the 2008 Iwate
Miyagi Inland earthquake.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
215
11
Figure 12 Transfer functions of measured and computed acceleration for the 2008 IwateMiyagi Inland
earthquake. (a) NS and (b) UD components.
Figure 13 compares measured and computed horizontal and vertical acceleration. Computed ho
rizontal acceleration [Fig. 13(a)] shows no clear peaks when vertical acceleration is negative.
When the vertical acceleration is negative, elements near the surface is freefalling, and at this
state confining pressure may be very small, and shear waves cannot be transmitted in such a soft
material. While as shown in Fig. 13(b), this is not clear in the measured acceleration. One possi
ble reason may be that in reality there are sources of higher frequencies as mentioned earlier.
Computed time histories from 4.5 to 5 s of volumetric strain [Fig. 14(a)], mean stress [Fig.
14(b)] with computed vertical surface acceleration, and mean stress and volumetric strain rela
tionship [Fig. 14(c)] are plotted to see the model behavior of the surface element. When the sur
face acceleration is negative, tensile volumetric strain indicated by number 3 to 5 in Fig. 14(a)
appears at zero mean stress. When the surface element is under compression indicated by num
ber 6 and 7 in Figure 14, positive spikes appear on the surface vertical acceleration. In the same
manner as Chapter 2, mean stress and volumetric strain relation [Fig. 14(c)] follows a quadratic
curve which is a volumetric stress and strain relationship indicated by Eq. (11).
Figure 13 Computed (a) and measured (b) time histories of NS and UD components of surface accelera
tion with magnified time scale for 3 to 8 seconds of Fig. 9.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
216
12
Figure 14 Computed time histories from 4.5 to 5.0 (s); (a) volumetric strain, (b) mean stress. In (a) and
(b), computed UD components of surface acceleration is plotted as a reference. (c) Mean stress versus
volumetric strain relationship.
4 CONCLUSIONS
During the 2008 IwateMiyagi Inland, Japan, earthquake, very large acceleration record
which exceeded 40 m/s
2
(3 components combined) was observed at the vertical array site, KiK
net, IWTH25 station. The surface vertical acceleration record had asymmetric amplitude. In the
present study, analytical studies were carried out for this newly discovered nonlinear site re
sponse associated with the vertical strong ground motion.
Firstly, model behavior against normal compression and tension was examined by using sin
gle element. Then response against vertical motion was investigated by using a simple 1D col
umn mesh with 10 m depth and sinusoidal input motion. The analysis indicated that the mecha
nism of the asymmetry on vertical surface acceleration might be characterized by the existence
of lower bound of negative acceleration and positive pulse due to compression of soils.
Then, observed ground motion was simulated with model ground of 1D column whose model
parameters were derived from the borehole data at the vertical array, KiKnet, IWTH25 site.
The ground model constituted with 7 layers from the surface to G.L. 246 m which was assumed
to be the base of IWTH25 station. With the base records of NS and UD components as input
motions, 1D site response analysis was conducted. Vertical acceleration at the surface clearly
showed asymmetric form as can be seen in the observed acceleration.
ACKNOWLEDGEMENT
Strong motion data used in this study was provided by the digital strongmotion seismograph
network, KiKnet, National Research Institute for Earth Science and Disaster Prevention
(NIED), Japan.
REFERENCES
Aoi, S., Kunugi, T. and Fujiwara, H. 2008. Trampoline effect in extreme ground motion. Science 322: 727730.
Bardet, J.P. and Davis, C.A. 1996. Performance of San Fernando dams during the 1994 Northridge Earthquake. Journal of Geotech
nical Engineering, ASCE 122 (7): 554564.
Bardet, J.P., Ichii, K. and Lin, C.H. 2000. EERA: A computer program for Equivalentlinear Earthquake site Response Analyses of
layered soil deposits. A report to Department of Civil Engineering, University of Southern California,
http://geoinfo.usc.edu/gees.
Bardet, J.P. and Tobita, T. 2001. NERA: A computer program for Nonlinear Earthquake site Response Analysis of Layered Soil De
posits. Department of Civil Engineering, University of Southern California, http://geoinfo.usc.edu/gees.
Bonilla, L.F., Archuleta, R.J. and Lavallee, D. 2005. Hysteretic and dilatant behavior of cohesionless soils and their effects on
nonlinear site response: Field data observations and modeling. Bulletin of the Seismological Society of America 95 (6): 2373
2395.
Dafalias, Y.F. and Popov, E. 1975. A model of nonlinearity hardening materials for complex loading. Acta Mechanica 21 (3): 173
192.
Frankel, A.D., Carver, D.L. and Williams, R.A. 2002. Nonlinear and linear site response and basin effects in Seattle for the M 6.8
Nisqually, Washington earthquake. Bulletin of the Seismological Society of America 92 20902109.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
217
13
Hardin, B.O. and Drnevich, V.P. 1972. Shear modulus and damping in soils: design equations and curves. J. Soil Mech. Found. Div.
1: 289.
Hashiguchi, K. 1980. Constitutive equation of elastoplastic materials with elastoplastic transition. Journal of Applied Mechanics,
ASME 102(2): 266272.
Holzer, T. L., Youd, T. L. and Hanks, T. C. 1989. Dynamic liquefaction during the 1987 Superstition Hills, California, earthquake.
Science 244: 5659.
Iai, S., Matsunaga, Y. and Kameoka, T. 1992. Strain space plasticity model for cyclic mobility. Soils and Foundations 32(2): 1 15.
Iai, S., Morita, T., Kameoka, T., Matsunaga, Y. and Abiko, K. 1995. Response of a dense sand deposit during 1993 Kushirooki
earthquake. Soils and Foundations 35 (1): 115131.
Idriss, I.M. and Seed, H.B. 1968. Analysis of ground motions during the 1957 San Francisco earthquake. Bulletin of the Seismologi
cal Society of America 58: 20132032.
Idriss, I.M. 1990. Response of soft soil sites during earthquakes. Proceedings of the H. Bolton Seed Memorial Symposium 2 J. M.
Duncan (Editor), BiTech Publisher, Vancouver: 273289.
Ishihara, K. 1993. Liquefaction and flow failure during earthquakes. The 33rd Rankine lecture, Geotechnique 43 (3): 351  415.
Iwan, W.D. 1967. On a class of models for the yielding behavior of continuous and composite systems. Journal of Applied Mechan
ics, ASME 34: 612617.
Iwasaki, Y. and Tai, M. 1996. Strong motion records at Kobe Port Island. Special Issue of Soils and Foundations, Japanese Geo
technical Society 2940.
Martin, G.R., Finn, W.D.L. and Seed, H.B. 1975. Fundamentals of liquefaction under cyclic loading. Journal of Geotechnical Engi
neering, ASCE 101 (5): 423438.
Masing, G. 1926. Eiganspannungen und verfestigung beim messing. Proceedings of Second International Conference of Applied
Mechanics: 322335.
Matsuoka, H. 1974. Stressstrain relationship of sands based on the mobilized plane. Soils and Foundations 14 (2): 4761.
Mroz, Z. 1967. On the description of anisotropic workhardening. Journal of Mechanics and Physics of solids 15: 163175.
Mroz, Z., Norris, V.A. and Zienkiewicz, O.C. 1978. An anisotropic hardening model for soils and its application to cyclic loading.
International Journal for Numerical and Analytical Methods in Geomechanics 2: 203221.
Nakai, T. 1986. A generalized elastoplastic constitutive model for clay in a threedimensional stresses. Soils and Foundations 26 (3):
8189.
NIED 2009. Digital StrongMotion Seismograph Network, Kiknet. http://www.knet.bosai.go.jp/knet/.
Oka, F., Yashima, A., Tateishi, Y., Taguchi, Y. and Yamashita, S. 1999. A cyclic elastoplastic constitutive model for sand consider
ing a plasticstrain dependence of the shear modulus. Geotechnique 49 (5): 661680.
Ozutsumi, O. 2003. Development of a numerical method for earthquake damage estimation of soil and structure systems on liquefi
able ground. Doctor's dissertation submitted to the Department of Civil and Earth Resources Engineering, Kyoto University (in
Japanese)
Prevost, J.H. 1985. A simple plasticity theory for frictional cohesionless soils. Soil Dynamics and Earthquake Engineering 4 (1): 9
17.
Ramberg, W. and Osgood, W.R. 1943. Description of stressstrain curve by three parameters. Tech. Note 902, National Advisory
committee for Aeronautics, Washington, D.C.
Roscoe, K.H., Schofield, A.N. and Thurairajah, A. 1963. Yielding of clay in states wetter than critical. Geotechnique 13 (3): 211
240.
Roscoe, K. H. and Burland, J. B. 1968. On the generalised stressstrain behaviour of 'wet' clay. Engineering Plasticity, Cambridge
Press 535609.
Schnabel, P.B., Lysmer, J. and Seed, H.B. 1972. SHAKEA computer program for earthquake response analysis of horizontally lay
ered sites. Earthquake Engineering Research Center, College of Engineering, University of California, Berkeley Report No.
EERC 7212.
Sekiguchi, H. and Ohta, H. 1977. Induced anisotropy and time dependency in clays. Proc. 9th ICSMFE, Speciality Session 9 229
238.
Towhata, I. and Ishihara, K. 1985. Modelling soil behaviour under principal stress axes rotation. Proceedings of 5th International
Conference on Numerical Methods in Geomechanics, Nagoya 1: 523530.
Toyota, H. 2009. Personal communication.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
218
Effect of PreYielding Elasticity on Sliding Triggered by
NearFault Motions Modeled as Idealized Wavelets
E. Garini, G. Gazetas, N. Gerolymos
Soil Mechanics Laboratory, National Technical University of Athens, Greece
ABSTRACT: The influence of elastic preyielding on the response of a mass resting on an
inclined plane is investigated in this paper. The ultimate shearing capacity of the interface
obeys Coulombs friction law. The slope is subjected to nearfault triggering by two types of
idealized wavelets: (i) a Ricker wavelet, representative of forward directivity affected motions,
containing strong longperiod acceleration pulses, and (ii) an onecycle sinusoidal wavelet,
representative of flingaffected motions, containing an onesided velocity pulse with an ensuing
permanent displacement. The asymmetric sliding response is analyzed and the effect of a
number of parameters is explored. They include: the critical acceleration ratio, a
C
/a
H
, the
excitation frequency, f
o
, the changing polarity of excitation, and the magnitude of elastic
preyielding displacement, dy.
1 INTRODUCTION
Several applications in geotechnical earthquake engineering require an understanding of the
dynamic sliding response of a block of mass m supported on seismically vibrating base through
an asymmetric frictional contact. n his 1965 seminal Rankine Lecture, Newmark proposed that
the seismic performance of earth dams and embankments be evaluated in terms of permanent
deformations which occur whenever the inertia forces on a potential slide mass are large enough
to overcome the frictional resistance at the failure surface. He proposed the analog of a rigid
block on inclined plane for a simple way of analytically obtaining approximate estimates of
these deformations. Newmarks analog has seen numerous applications and extensions: seismic
deformation analysis of earth dams and embankments, displacements associated with landslides,
seismic deformation of landfills with geosynthetic liners, seismic settlement of surface
foundations, movements of wedges in rock slopes, and even potential sliding of concrete gravity
dams. The extension of the analog by Richards & Elms (1979)
to gravity retaining walls has
met worldwide acceptance, and has found its way into seismic codes of practice. Several other
generalized applications have also appeared.
A numerical study has been recently presented by the authors (Garini et al 2007, Gazetas et
al 2009) for a rigid block supported through a rigidplastic frictional contact surface on an
inclined plane, and subjected to slopeparallel excitation. The latter was described with near
fault seismic records strongly influenced by forwarddirectivity or flingstep effects. Our
study had consistently and repeatedly revealed a profound sensitivity of both maximum and
residual slippage: (i) on the sequence and even the details of the pulses contained in the
excitation, and (ii) on the polarity of shaking. A few of the findings contradicted some of the
prevailing beliefs that have emanated from statistical correlation studies in literature.
However, all these finding were based on the extreme assumption of a perfectly rigidplastic
interface. Since in most realistic systems some presliding elasticity is unavoidable, this paper
investigates an elasticplastic support interface. Fig.1 illustrates the problem studied herein.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
219
(a)
A
(t)
D
(t)
m
=
f(D
)
T(t)
D(t)
mg(cos sin)
mg(cos+sin)
T(t)
D(t)
mg(cos sin)
mg(cos+sin)
dy
b) (c)
(a)
=
f(D
)
=
f(D
)
D
(t)
D
(t)
Figure 1. (a) The problem studied in the paper sliding on inclined plane undergoing excitation parallel
to the slope, (b) ideally rigidplastic behavior of the interface as studied by Garini et al (2007), and (c)
elasticperfectly plastic sliding response studied here.
2 DIRECTIVITY AND FLING IN NEARFAULT MOTIONS
Earlier studies of Newmarksliding were based on records available in the late 1970s and
1980s. Very few of those motions were nearfault records from largemagnitude (M > 6.5)
events. Today, however, such nearfault records are known to often contain either longperiod
highamplitude acceleration pulses, or large residual displacements the outcome,
respectively, either of the coherent arrival of seismic waves when the fault rupture propagates
towards the site, or of tectonic permanent displacement (offset) of the earth in the proximity of
the seismogenic fault rupture. The terms forwardrupture directivity and fling step have
been given to the two phenomena (Singh, 1988; Somerville et al, 1996; Abrahamson, 2000; and
Bolt, 2004).
Fig. 2 illustrates in idealized form some fundamental characteristics of these two types of
nearfault motions. For strikeslip earthquakes, the signature of forward rupture directivity
appears in direction normal to the fault; whereas, the fling step is significant in the parallel
component of motion in close proximity to the fault, especially if the latter emerges on the
surface with a large static offset. The two phenomena (and directivity in particular) have been
the subject of seismological (theoretical and instrumental) as well as earthquake engineering
research.
Two idealized motions (wavelets) are used as excitation in this paper to represent in a
simple way typical directivity and fling affected ground motions. They are the onecycle
sinusoidal and the Ricker wavelets: the former modeling a typical fling affected motion, and the
latter a directivity affected motion. Four characteristic frequencies are utilized for each of these
motions as follows:
Richer: f
o
= 0.35 Hz, 0.75 Hz, 1.5 Hz, and 3,5 Hz
Sinusoidal: f
o
= 0.57 Hz, 1.25 Hz, 2.5 Hz, 6.67 Hz
(
A
(t)
m
A
(t)
m
T(t)
D(t)
mg(cos sin)
mg(cos+sin)
T(t)
D(t)
mg(cos sin)
mg(cos+sin)
T(t)
D(t)
mg(cos sin)
mg(cos+sin)
dy
T(t)
D(t)
mg(cos sin)
mg(cos+sin)
dy
b) (c) (
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
220
The response spectra of corresponding pairs of these motions are approximately matched to give
almost the same maximum spectral acceleration at nearly the same dominant period. For
instance, Fig.3 illustrates the spectral matching of two of the pairs.
0
2
4
6
3 5 7 9 11
0
2
4
6
3 5 7 9 11
0.7
0
0.7
1.4
3 5 7 9 11
3
0
3
3 5 7 9 11
10
5
0
5
10
3 5 7 9 11
D
residual
= 0
Fault
strike
Directivity pulse (Normal component
of Displacement)
Fling step (Parallel component
of Displacement)
Site
(t) :
m/s
2
V(t) :
Fault
strike
Directivity pulse (Normal component
of Displacement)
Site
Figure 2. Explanatory sketch of the forwarddirectivity and flingstep phenomena as reflected in the
displacement records; and examples of simple wavelets bearing the signature of the two effects.
3 RESULTS: THE EFFECT OF PREYIELDING ELASTICITY ON SLIDING
Fig.4 depicts the asymmetric sliding response in terms of acceleration, velocity, and
displacement timehistories for a mass of m = 1 Mgr on a slope of inclination = 25
o
. The
system is subjected to a Ricker wavelet of peak acceleration 1 g and central frequency 1.5 Hz.
The right handside portrays the timehistories corresponding to a rigidplastic interface with
critical ratio of a
C
/a
H
= 0.1. Notice that the mass is moving in unison with the base as long as
the critical acceleration a
C
is not exceeded. Whenever base acceleration exceeds the critical
m/s
D(t) :
Fling Step
General Example:
(Onecycle Sinus pulse )
t : s
m
Forward Directivity
General Example:
(Ricker wavelet)
t : s
V 5.6
10
5
0
5
10
3 5 7 9 11
V 5.3 m/s
D
maxl
= 1.31 m
D
residual
= 4.9 m
0
2
4
6
3 5 7 9 11
0
2
4
6
3 5 7 9 11
0.7
0
0.7
1.4
3 5 7 9 11
3
0
3
3 5 7 9 11
10
5
0
5
10
3 5 7 9 11
D
residual
= 0
Fling step (Parallel component
of Displacement)
Fault
strike
Directivity pulse (Normal component
of Displacement)
Site
Fling step (Parallel component
of Displacement)
(t) :
m/s
2
V(t) :
s m/
D(t) :
Fling Step
General Example:
(Onecycle Sinus pulse )
Fling Step
General Example:
(Onecycle Sinus pulse )
t : s
m
Forward Directivity
General Example:
(Ricker wavelet)
Forward Directivity
General Example:
(Ricker wavelet)
t : s
V 5.6
10
5
0
5
10
3 5 7 9 11
V 5.3 m/s
D
maxl
= 1.31 m
D
residual
= 4.9 m
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
221
acceleration, the block slides either downward (usually) or upward (rarely). As a result, a
residual yielding displacement of 0.36 m occurs.
The left handside of Fig.4 illustrates the elasticplastic response of the same massbase
system. Here, the block can displace elastically up to 0.1 m before yielding. As a consequence,
a phase difference between input excitation and induced mass response occurs. Notice this
phase shift in Fig.4 in the acceleration and velocity timehistories (see arrows). It is emphasized
that preyielding displacement exists in both the upward and downward direction. Thus, even
though no uphill sliding happens after 1 sec in Fig.4, an elastic upward displacement does take
place reversal of accumulated slippage. In addition, the block does not rest after the last
sliding period but continues uphill and downhill elastic oscillations after the end of triggering.
The forcedisplacement hysteresis loop in Fig.5 shows the two yielding events and the
accompanying elastic branches.
0
1
2
3
0 1 2 3 4
Ricker, f
o
= 0.35 Hz, 1.50 Hz
Onecycle Sinus, f
o
= 0.57 Hz, 1.25 Hz
T : s
S
A
: g
Onecycle Sinus
Ricker
1 g
1 g
1 g
0.6 g
0
1
2
3
0 1 2 3 4
Ricker, f
o
= 0.35 Hz, 1.50 Hz
Onecycle Sinus, f
o
= 0.57 Hz, 1.25 Hz
T : s
S
A
: g
Onecycle Sinus
Ricker
Onecycle Sinus
Ricker
1 g
1 g
1 g
0.6 g
1 g
0.6 g
Figure 3. The two idealized time histories used as base excitation, with their response acceleration spectra
S
A
:g T. (Peak ground acceleration: 1 g).
4 COMPILATION OF RESULTS
All our numerical results with Ricker wavelets as base excitations are compiled in Figs 6 and 7.
Fig.6 depicts the permanent slippage with respect to the critical acceleration ratio, a
C
/a
H
, for
four values of preyielding deformation, dy, and four excitation frequencies, f
o
. As expected,
when the acceleration ratio, a
C
/a
H
, increases the induced slippage decreases. This general trend
is independent of the existence or not of the preyielding displacement. Notice that the
existence of preyielding elasticity may lead to larger or smaller permanent displacements,
depending on f
o
and a
C
/a
H
.
Furthermore, in Fig.7 observe the influence of frequency on sliding displacement. For
frequency, f
o
= 0.35 Hz, as the elastic region dy increases the slippage D also increases. The
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
222
same is valid for f
o
= 0.75 Hz. However, when the frequency, f
o
, increases up to 1.50 Hz the
response changes. Observe that while slippage becomes greater for preyielding region between
0 and 0.01, for larger values of dy the sliding response decreases, in the case of 1.50 Hz. The
behavior is more complicated when the frequency takes the value of 3.50 Hz.
Figure 4. Acceleration, velocity, and displacement time histories for a maximum elastic deformation of
dy = 0.1 m are presented at the left, and for dy = 0 are presented at the right (excitation: Ricker wavelet of
frequency f
o
= 1.5 Hz).
5 EFFECT OF CHANGING POLARITY
The next two Figures (Figs 8 and 9) address a most astonishing effect: that of the reversal in
polarity (i.e., change from + to direction in which the excitation is applied). [This is the same
as having two identical slopes, one opposite to the other (across the street so to speak),
subjected to the same excitation, as sketched at the top of Figs. 8 and 9.]
A few researchers and only in recent years (Kramer & Lindwall, 2004; Fardis et al, 2003)
appear to have published on the importance of the polarity of shaking. This has much to do
0
0.1
0.2
0.3
0.4
0 1 2 3
10
5
0
5
10
0 1 2 3
0.14 m
2
nd
slide 0.29 m
1
st
slide
t : s
:
s
2
s
m/
V :
m/
D :
m
Ground
Sliding block
25
o
C
/
= 0.1
dy = 0.1 m
t : s
dy = 0 m
10
5
0
5
10
0 1 2 3
0.8
0.4
0
0.4
0.8
0 1 2 3
0.36 m
0.8
0.4
0
0.4
0.8
0 1 2 3
0
0.1
0.2
0.3
0.4
0 1 2 3
0
0.1
0.2
0.3
0.4
0 1 2 3
10
5
0
5
10
0 1 2 3
0.14 m 0.14 m
2
nd
slide 0.29 m
1
st
slide
t : s
:
s
2
s
m/
V :
m/
D :
m
Ground
Sliding block
25
o
25
o
C
/
= 0.1
C
/
= 0.1
dy = 0.1 m dy = 0.1 m
t : s
dy = 0 m dy = 0 m
10
5
0
5
10
0 1 2 3
0.8
0.4
0
0.4
0.8
0 1 2 3
0.8
0.4
0
0.4
0.8
0 1 2 3
0
0.1
0.2
0.3
0.4
0 1 2 3
0.36 m
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
223
T : kN
4
3
2
1
0
1
2
0 0.1 0.2 0.3 0.4 0.5
1
st
slide 2
nd
slide
0.08
Initial elastic deformation induced
by the m g sin component
D : m
T : kN
4
3
2
1
0
1
2
0 0.1 0.2 0.3 0.4 0.5
1
st
slide 2
nd
slide
0.08
Initial elastic deformation induced
by the m g sin component
D : m
Figure 5. Forcedisplacement response for the case of an elastoplastic sliding system with dy = 0.1 m,
= 25
o
, a
C
/a
H
= 0.1 and a Ricker excitation of 1.5 Hz frequency. Notice that yielding occurs only in one
direction, as it was expected.
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
dy = 0
dy = 0.01 m
dy = 0.05 m
dy = 0.1 m
0
3
6
9
12
15
18
21
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
C
/
H
S
l
i
p
p
a
g
e
:
m
1 g
0.6 g
S
l
i
p
p
a
g
e
:
m
C
/
H
25
o
dy = 0
dy = 0.01 m
dy = 0.05 m
dy = 0.1 m
dy = 0
dy = 0.01 m
dy = 0.05 m
dy = 0.1 m
dy = 0
dy = 0.01 m
dy = 0.05 m
dy = 0.1 m
Ricker 0.35 Hz Ricker 0.75 Hz
Ricker 1.50 Hz Ricker 3.50 Hz
m
=
1
M
g
r
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
dy = 0
dy = 0.01 m
dy = 0.05 m
dy = 0.1 m
0
3
6
9
12
15
18
21
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
C
/
H
S
l
i
p
p
a
g
e
:
m
1 g
0.6 g
1 g
0.6 g
S
l
i
p
p
a
g
e
:
m
C
/
H
25
o
25
o
25
o
dy = 0
dy = 0.01 m
dy = 0.05 m
dy = 0.1 m
dy = 0
dy = 0.01 m
dy = 0.05 m
dy = 0.1 m
dy = 0
dy = 0.01 m
dy = 0.05 m
dy = 0.1 m
Ricker 0.35 Hz Ricker 0.35 Hz Ricker 0.75 Hz Ricker 0.75 Hz
Ricker 1.50 Hz Ricker 1.50 Hz Ricker 3.50 Hz Ricker 3.50 Hz
m
=
1
M
g
r
Figure 6. Influence of maximum elastic deformation, dy, on asymmetric sliding response triggered by
Ricker wavelets of maximum acceleration 1 g and of different frequencies.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
224
S
l
i
p
p
a
g
e
:
m
S
l
i
p
p
a
g
e
:
m
dy : m dy : m
0
3
6
9
12
15
18
21
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.05
0
1
2
3
4
5
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.05
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.05
0
0.03
0.06
0.09
0.12
0.15
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.05
C
/
H
0.05
0.1
0.02
1 g
0.6 g
25
o
Ricker 0.35 Hz Ricker 0.75 Hz
Ricker 1.50 Hz Ricker 3.50 Hz
S
l
i
p
p
a
g
e
:
m
S
l
i
p
p
a
g
e
:
m
dy : m dy : m
0
3
6
9
12
15
18
21
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.05 0.05
0
1
2
3
4
5
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.05
0
1
2
3
4
5
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.05
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.05
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.05
0
0.03
0.06
0.09
0.12
0.15
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.05
0
0.03
0.06
0.09
0.12
0.15
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.05
C
/
H
0.05
0.1
0.02
C
/
H
0.05
0.1
0.02
1 g
0.6 g
1 g
0.6 g
25
o
25
o
25
o
Ricker 0.35 Hz Ricker 0.35 Hz Ricker 0.75 Hz Ricker 0.75 Hz
Ricker 1.50 Hz Ricker 1.50 Hz Ricker 3.50 Hz Ricker 3.50 Hz
Figure 7. Effect of characteristic frequency, f
o
, and of
C
/
) and a
yield acceleration ratio a
C
/a
H
= 0.2, we notice the following : When the first sinusoidal
acceleration halfpulse is downward [as in Fig. 8(a) on the right] the block remains almost
attached to the base. Only a mere 0.18 m uphill displacement takes place (see the upward
slippage region enclosed by the dotted lines, starting at 1 sec and ending at 1.8 sec). Even in
this small 0.18 m deformation, the yielding part is particularly smaller than the elastic. The
subsequent, second (and last), upward halfpulse acceleration of the base initiates an uninhibited
downslope slippage of the block, which lasts for a long time after the excitation has terminated
t 2.4 sec on t 5.2 sec. The result is a huge 7.70 m.
In stark contrast, when the first sinusoidal acceleration halfpulse of the base is upward [as in
Fig. 8(b) at the left] the block starts sliding downslope almost immediately. But it soon comes
to a stop after about 1.4 seconds, as the upward base motion decelerates and then reverses. The
resulting residual slip is only 3.14 m, almost 2.5 times smaller than the 7.70 m produced with
the reverse motion!
This effect of reversing the polarity of shaking is of profound importance, especially with
fling type motions (as the sinus pulse studied above). It may not however be as dramatic with
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
225
directivity affected motions if they contain several competing cycles of pulses, as seen in
Fig. 9 with the Ricker wavelet.
12
10
8
6
4
2
0
2
4
1 0 1 2 3 4 5 6 7 8
0
1
2
3
4
5
6
7
8
0 1 2 3 4 5 6
12
10
8
6
4
2
0
2
4
1 0 1 2 3 4 5 6 7 8
A
(
t
)
:
m
/
s
2
t : s
15
10
5
0
5
10
15
0 1 2 3 4 5 6
D
(
t
)
:
m
T
:
k
N
D : m
25
o
25
o
+
15
10
5
0
5
10
15
0 1 2 3 4 5 6
1
0
1
2
3
4
5
6
7
8
0 1 2 3 4 5 6
t : s
D : m
Ground
Sliding block
C
/
= 0.2
3.14 m
7.70 m
(b) (a)
12
10
8
6
4
2
0
2
4
1 0 1 2 3 4 5 6 7 8
0
1
2
3
4
5
6
7
8
0 1 2 3 4 5 6
12
10
8
6
4
2
0
2
4
1 0 1 2 3 4 5 6 7 8
A
(
t
)
:
m
/
s
2
t : s
15
10
5
0
5
10
15
0 1 2 3 4 5 6
D
(
t
)
:
m
T
:
k
N
D : m
25
o
25
o
+
25
o
25
o
+
15
10
5
0
5
10
15
0 1 2 3 4 5 6
1
0
1
2
3
4
5
6
7
8
0 1 2 3 4 5 6
t : s
D : m
Ground
Sliding block
C
/
= 0.2
C
/
= 0.2
3.14 m
7.70 m
(b) (a)
Figure 8. Acceleration and displacement time histories for a stiff elastoplastic system with yield
displacement dy = 0.05 m. The third row of figures illustrates the forcedisplacement response.
(Excitation: onecycle Sinus of 0.57 Hz frequency).
6 CONCLUSIONS
The solutions portrayed graphically in the paper (through the acceleration and velocity time
histories of the base and mass) are easy to understand, offering considerable insight into the
dynamics of asymmetric sliding when elastic preyielding is taken into account. The effects
of elastoplastic yielding on the final accumulated displacement are: (a) the time shift in the
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
226
sliding timehistories of the mass, and (b) the increase of uphill deformation for accelerations
smaller than the critical because of elasticity.
Forwarddirectivity and flingstep affected motions, containing severe acceleration pulses
and/or large velocity steps, may have an unpredictablydetrimental effect on residual slip,
especially for small values of the critical acceleration. The unpredictability of asymmetric
response arises from the sensitivity of the sliding on the sequence, duration, and details of
motion.
Changing the polarity of excitation (i.e., applying it in the + and then in the direction) has a
significant effect on the accumulated slippage.
25
o
25
o
+
0
9
18
27
36
45
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
C
/
H
+
Figure 9. The impact of alternating excitation polarity on the induced yielding for both the Ricker and
onecycle sinus pulses. Three different values of elastic deformation, dy, are presented.
7 ACKNOWLEDGEMENT
The research presented in this paper was financially supported by the Secretariat for Research
and Technology of Greece, under the auspices of PENED Programme with Contract number
03ED278. We thank Dr. Ioannis Anastasopoulos for his thoughtful help in programming
aspects of this study.
REFERENCES
Abrahamson, N. A. (2000), Effects of rupture directivity on probabilistic seismic hazard analysis, Pro
ceedings, 6
th
International Conference on Seismic Zonation, Palm Springs, California, Earthquake En
gineering Research Institute.
Bolt, B.A. (2004), Earthquakes, W. H. Freeman & Co, Fifth Edition, New York.
Boore, M. D. (2001), Effect of baseline corrections on displacements and response spectra for several Re
cordings of the 1999 ChiChi, Taiwan, Earthquake, Bulletin of the Seismological Society of America,
Vol. 91, No. 5, pp. 11991211.
Bray, J.D., and Rathje, E.M. (1998), EarthquakeInduced Displacements of SolidWaste Landfills, Jour
nal of Geotechnical and Geoenvironmental Engineering, ASCE, Vol. 124, pp. 242253.
Constantinou, M. C., and Gazetas, G. (1987), Probabilistic Seismic Sliding Deformations of Earth Dams
and Slopes, Proceedings of the Specialty Conference on Probabilistic Mechanics and Structural Reli
ability, ASCE, Berkeley, pp. 318321.
S
l
i
p
p
a
g
e
:
m
C
/
H
C
/
H
f
o
= 0.57 Hz,
f
o
= 0.57 Hz,
f
o
= 0.35 Hz,
f
o
= 0.35 Hz,
OneCycle Sinus
Ricker
}
}
+
+
dy = 0.01 m dy = 0.05 m
dy = 0.1 m
25
o
25
o
+
25
o
25
o
+
0
9
18
27
36
45
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
C
/
H
C
/
H
C
/
H
S
l
i
p
p
a
g
e
:
m
0
9
18
27
36
45
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
C
/
H
C
/
H
C
/
H
f
o
= 0.57 Hz,
f
o
= 0.57 Hz,
f
o
= 0.35 Hz,
f
o
= 0.35 Hz,
OneCycle Sinus
Ricker
}
}
+
S
l
i
p
p
a
g
e
:
m
+
f
o
= 0.57 Hz,
f
o
= 0.57 Hz,
f
o
= 0.35 Hz,
f
o
= 0.35 Hz,
OneCycle Sinus
Ricker
}
}
++
++
V
S,30 V
s
V
s
V
S ( z = H )
V
R
V
R
V
R
z
(
m
)
V
s
V
S,30
V
S,30
V
S,30 V
s
V
s
V
S ( z = H )
V
R
V
R
V
R
Aegion Sakarya
Dayhook
Superstition Mtn.
Gilroy Stone Canyon
Lucerne
0.4
0.4
0
(g)
Aegion Sakarya
Dayhook
Superstition Mtn.
Gilroy Stone Canyon
Lucerne
0.4
0.4
0
(g)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
233
Fig. 5: Compilation of response spectra of ground surface motions from all the equivalentlinear
analyses. (a) Conventionally normalized spectra ; (b) BiNormalized spectra. The thick curves are
the mean response spectra.
7. CONCLUSION
One unique BiNormalized Spectrum (BNS), for all soil categories and most likely seismic exci
tations, emerged from the comprehensive set of wavepropagation analyses reported in this arti
cle. This unique spectrum is sketched in Fig. 7 and is approximated with the following algebraic
expressions :
S
a
/A = exp (1.35 [T / T
p
]
) for T/T
p
< 1
S
a
/A = 3.75 ( T/T
p
)
1.2
for T/T
p
1
The potential benefits from adopting this simple spectrum have been highlighted in the article.
However, the imprecise definition of T
p
and the difficulty in predicting it in reality remain the
main obstacles in adopting it at present.
T (sec)
H=60m
BiNormalizedSpectra
1
S
a
/A
T / T
P
BiNormalized Spectra
S
a
/A
3.75
T (sec)
H=60m
BiNormalizedSpectra
1
S
a
/A
T / T
P
BiNormalized Spectra
S
a
/A
3.75
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
234
Fig. 6: Comparison of the mean spectra of Figure 5 (Soil: V
S,30
= 180 360 m/s) , with the mean
spectra for a much softer soil (V
S,30
= 100 m/s). The differences of the conventionally normalized
spectrum (top) almost disappear in the normalized spectra (bottom)
T ( sec )
C
C
C
S
a
/ A
S
a
/ A
T / T
p
T ( sec )
C
C
C
S
a
/ A
S
a
/ A
T / T
p
S
a
/ A
T /
S
a
/ A
T /
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
235
Fig. 7: Mean BiNormalized Spectra (BNS) from the equivalent linear wave propagation (SHAKE)
and from the inelastic wave propagation (NLDYAS) studies, and the idealized smooth spectrum
proposed for design. The algebraic expressions for two branches of this spectrum are given in the text.
ACKNOWLEDGMENTS
This work forms part of an EU 7
th
Framework research project funded through the European
Research Councils (ERC) Programme Ideas, Support for Frontier Research Advanced
Grant, under Contract number ERC2008AdG 228254DARE.
REFERENCES and RELATED BIBLIOGRAPHY
Afra H. and Pecker A. (2001), Calculation of free field response spectrum of a nonhomogeneous soil
deposit from bed rock response spectrum, Soil Dynamics and Earthquake Engineering, 22, 157165.
Bachman R and Bonneville D (2000), The seismic provisions of the 1997 Uniform Building Code, Earth
quake Spectra, 16(1), pp. 85100.
Bouckovalas G. D., Papadimitriou A.G., Karamitros D. (2006), Compatibility of EC8 ground types and
site effects with 1D wave propagation theory, Geotechnical Evaluation and Application of EC8, 53
67.
Borcherdt RD (1994), Estimates of sitedependent response spectra for design (methodology and justifi
cation), Earthquake Spectra 10 (4): 61753.
Building Seismic Safety Council (BSSC), The 2003 NEHRP recommended provisions for new buildings
and other structures. Part 1: Provisions, Federal Emergency Agency (FEMA 450), Washington D.C.
Dobry R, Borcherdt RD, Crouse CB, Idriss IM, Joyner WB, Martin GR, Power MS, Rinne EE and Seed
RB (2000), New site coefficients and site classification system used in recent building seismic code
provisions, Earthquake Spectra , 16(1), 4167.
Drosos V, Gerolymos N and Gazetas G (2007), Calibration and verification of nonlinear wave propaga
tion method, Proceedings of the 4
th
Int. Conference on Earthquake Geotechnical Engineering, K. Piti
lakis, editor, Paper No. 1594.
EC8 (2001), Part 1 General Rules, seismic actions and rules for buildings, Eurocode 8, Draft 4 (prEN
1998), European Committee for Standardization.
Gazetas G (1982), Vibrational characteristics of soil deposits with variable wave velocity, Numerical and
Analytical Methods in Geomechanics, 1982, 6 (1), 120.
Gazetas G (2006), Seismic design of foundations and soilstructure interaction, 1
st
European Conference
on Earthquake Engineering and Seismology, Geneva, Switzerland, 38 September.
Gazetas G and Bianchini G (1979), Field evaluation of body and surfacewave soilamplification theo
ries, Proceedings of the 2
nd
U.S. National Conference on Earthquake Engineering, EERI, 603612.
Gerolymos N and Gazetas G (2005), Constitutive model for 1D cyclic soil behavior applied to seismic
analysis of layered deposits, Soils and Foundations, JGS, 147159..
Hartzell S, Bonilla LF, and Williams RA (2004), Prediction of nonlinear soil effects, Bull. Seism. Soc.
Am. 94, 16091629.
Holmes WT (2000), The 1997 NEHRP Provisions for seismic regulations for new buildings and other
structures, Earthquake Spectra, 16 (1), pp. 101114.
Housner GW (1959), Behavior of structures during Earthquakes, Proceedings. ASCE, 85, October.
Idriss IM (1990), Response of soft soils during earthquakes, Proceedings of H. Bolton Seed memorial
symposium, Vol. 2, J. Michael Duncan (ed.), Bitech, Vancouver.
Mylonakis G and Gazetas G (2000), Seismic soilstructure interaction: beneficial or detrimental?, Journal
of Earthquake Engineering, 4 (3), 277301.
NEHRP (2000), Recommended provisions for seismic regulations for new buildings and other structures:
Part 1, provisions, FEMA 368, Building Seismic Safety Council (BSSC), Washington D.C.
Pitilakis K. (2004), Site effects, recent advances in earthquake geotechnical engineering and microzona
tion, A. Ansal editor, Kluwer Acacemic Publishers, 139193.
Pitilakis K., Gazepis C., Anastasiadis A. (2006), Design response spectra and soil classification for seis
mic code provisions. Geotechnical Evaluation and Application of EC8, G. Bouckovalas editor, 3752.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
236
Rathje EM, Abrahamson NA and Bray JD (1998), Simplified frequency content estimates of earthquake
ground motions, Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 124(2) 150
159..
RodriguezMarek A, Bray JD and Abrahamson NA (1999), Task 3: Characterization of site response
general site categories, Pacific Earthquake Engineering Research Center (PEER) Report 1999/03.
Schnabel PB, Lysmer J and Seed HB (1972), SHAKE: A computer program for earthquake response
analysis of horizontally layered sites, Rep. EERC 7212, University of California, Berkeley.
Schnabel PB, Seed HB and Lysmer J (1972), Modification of seismological records for effects of local
soil conditions, Bulletin of the Seismological Society of America, 62, 16491664.
Seed HB, Ugas C and Lysmer J (1976), Sitedependent spectra for earthquakeresistant design, Bulletin
of the Seismological Society of America, 66, 221243.
Somerville P (1998), Emerging Art: Earthquake ground motion, geotechnical special publication, 75
Geotechnical Eng. & Soil Dynamics III, ASCE, Dakoulas P, Yegian M and Holtz RD editors.
Tazoh T., Sato M., Shimizu K. and Koyama K. (1988), Nonlinear seismic response analysis of horizon
tally layered soil deposits, Proceedings of Ninth World Conference on Earthquake Engineering, To
kyoKyoto, II, 495501.
Tazoh T., Shimizu K., Sato M., Hirosse T. and Kouama K., (1988), Identification of timevarying pre
dominant period and damping constant in nonlinear seismic response of soft soil deposit, Proceedings
of Ninth World Conference on Earthquake Engineering, TokyoKyoto, II, 483488.
Towhata, I. (1996), Seismic wave propagation in elastic soil with continuous variation of shear modulus
in the vertical direction, Soils and Foundations, JGS, 36(1), 6172.
Travasarou Th. and Gazetas G., (2004), On the linear seismic response of soils with modulus varying as a
power of depth the Maliakos marine clay, Soils and Foundations, JGS, 44(5), 8593.
Kramer SL, Geotechnical Earthquake Engineering, Prentice Hall, 1996.
Li S. and Xie L (2007), Effect of hanging wall and forward directivity in 1999 ChiChi earthquake on
inelastic displacement response of structures., Earthquake Engineering and Engineering Vibration,
6(1) , 7784.
Stewart JP, Liu AH, Choi Y and Baturai MB (2001), Amplification factors for spectral acceleration in ac
tive regions, Pacific Earthquake Engineering Research Center (PEER) Report 2001/10.
Vucetic M and Dobry R (1991), Effect of soil plasticity on cyclic response, J. Geotech. Eng. ASCE 117,
89107.
Xu L and Xie L (2004), BiNormalized response spectral characteristics of the 1999 ChiChi Earthquake,
Earthquake Engineering and Engineering Vibration , 3 (2), 47155
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
237
3.
SoilStructureInteraction
1 INTRODUCTION
The Building Standard Law in Japan and its related enforcement and notices were revised for
the direction to the performancebased design in 1998. The calculation method of response and
limit strength was provided for checking structural serviceability and safety of buildings (Mido
rikawa et al. 2000, Kuramoto et al. 2000, Kuramoto et al. 2002). The soil structure interaction
(SSI) effects should be considered when the effect will be not negligible. The appropriate me
thod for the structural safety of buildings during severe earthquake needs to be proposed.
A method for incorporating SSI and calculating the equivalent period and damping effect is
presented. To obtain relationships between force and displacement of the building with a pile
foundation and surrounding soil, a pushover analysis is conducted (Watanabe et al. 2004). To
get the damping effect of the building with SSI, a radiation damping through piles is considered
(AIJ 2006). Based on these results, the building, pile foundation and surrounding soil are re
placed with an equivalent single degree of freedom (ESDOF) model. The spectrumbased
method is applied to earthquake responses of a residence building with a span in short direction.
The results by the spectrumbased method are compared with those by the time history analy
sis of the SSI model. Through comparison, the applicability of the method is discussed.
Calculation of seismic response of building based on pushover
analysis of SSI model
M. Iiba
Building Research Institute, Tsukuba, Japan
Y. Umemura
Ando Corporation, Minatoku, Tokyo, Japan
O. Kurimoto
Obayashi Corporation, Kiyose, Tokyo, Japan
T. Akita
Chiba University, Inageku, Chiba, Japan
M. Teshigawara
Nagoya University, Chikusaku, Nagoya, Japan
K. Watanabe
Urban Renaissance Agency, Shinjyukuku, Tokyo, Japan
ABSTRACT: The soil structure interaction (SSI) effect should be considered when the SSI ef
fect would be not negligible. The appropriate method for incorporating the SSI effect to res
ponses of structure to evaluate structural safety needs to be proposed. To obtain relationships
between force and displacement of a building with a pile foundation and surrounding soil, a pu
shover analysis is conducted. To evaluate a damping effect of the building with SSI, a radiation
damping through piles is considered. The building, pile foundation and soil system are replaced
with an equivalent single degree of freedom system and responses of the system are calculated
based on the acceleration spectrumbased method. The responses of the building are compared
with those by the time history analysis. The responses almost agree with those by time history
analysis and the proposed method is demonstrated to have enough accuracy for a practical use.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
238
2 EVALUATION PROCEDURE
The evaluation procedure involves the application of the equivalent linearization technique us
ing the ESDOF system and the response spectrum analysis (Kuramoto et al. 2000), as shown in
Figure 1. The building is replaced with the ESDOF system with equivalent period and damping
factor. This is based on the result of the nonlinear pushover analysis with the distribution in
proportion to the first mode of the building vibration (Kuramoto et al. 2002).
Transition Curve
Demand Spectra
Capacity Spectrum
for Inelastic (h
eq
=7%, 10%,15%)
for Elastic (h
eq
=5%)
T
5%
T
15%
S
d
S
a
h
eq
15%
Performance Point (Response)
h
eq
at Performance Point
5%
T
7%
T
10%
7%
10%
Limit Value
Figure 1. Illustration of seismic response evaluation procedure based on response spectrum.
The flow for response evaluation of buildings during earthquake is shown in Figure 2. The
procedure consists of several steps as follows;
a) The ARS at ground surface is set. The acceleration (S
a
) and displacement response spectra
(S
d
) are drawn up (Demand Spectrum in Fig. 1).
b) The nonlinear pushover analysis for the model with a building, a foundation and a surround
ing soil as shown in Figure 3. (Watanabe et al. 2004)
c) The forcedisplacement relationship of the ESDOF system (Capacity Spectrum in Fig. 1). is
established based on the results by the nonlinear pushover analysis.
d) Equivalent damping coefficient based on dissipation energy is calculated.
e) Based on the equivalent period and equivalent damping factor, the response of the ESDOF
system is obtained (Performance Point in Fig. 1).
The earthquake motion is given as the ARS at the outcropped engineering bedrock whose
shear wave velocity is about 400 m/s. The time histories of earthquake are fitted to the ARS
with random phases. By using equivalent linear analysis (Program SHAKE), the time history
responses on the ground surface whose average ARSs are calculated. A distribution of external
horizontal force to the superstructure is obtained through the time history analysis of the SSI
model. In the transverse direction, the distribution of horizontal force is the combination of a
uniform and a triangular distribution through height (Umemura et al. 2004).
The building is replaced with an ESDOF system as shown in Figure 4 (Kuramoto et al. 2002).
The forcedisplacement relationship of the ESDOF system is given by equations (1) and (2),
when the force corresponds to the base shear (
B 1
Q ), and its displacement (
1
) corresponds to
the displacement at an equivalent height (
e
H ) where the modal participation function is equal to
1.0 ( { } 0 . 1
1 1
= u )
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
239
a 1
1
B 1
S M Q = (1)
d 1 1
S = (2)
where, M 1 = equivalent mass corresponding to the 1st mode, and
a 1
S and
d 1
S = accelera
tion response and displacement response for the 1st mode, respectively.
START
Building, ground and earthquake motion, etc. conditions
Response acceleration spectrum
(considering amplification
properties of surface ground)
Response of each story
Pushover analysis of building,
foundation and soil model
ESDOF system (force
vs. displacement,
equivalent period
Damping coefficient and equivalent
damping factor of ESDOF
Response of ESDOF (shear force and displacement)
END
Figure 2. Flow for response evaluation.
Column
Girder
Footing
girder
Shear wall
Horizontal
Soil spring
Seismic
horizontal
force
Pile end bearing spring
Pile
Pile
peripheral
friction
spring
Figure 3. Overall model for pushover analysis.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
240
K
inel
1
SDOF System
Q
B
Structure Model
1
Q
B
H
e
Force Vector and
Displaced Mode
h
e
: equivalent height, Q
B
: base shear force
1
: horizontal Displacement at equivalent height
Figure 4. Multistory model and ESDOF system of building
3 RELATIONSHIP BETWEEN LOAD AND DEFLECTION FOR PUSHOVER ANALYSIS
3.1 Modeling of members of superstructure
Column members are modeled as a rigid beam element with a rotational spring at the column
capital and column base. Beam members are also modeled as the rigid beam element with the
rotational spring in both ends, as shown in Figure 5. In these cases, the bending moment (M)
and rotating angle () make the model with a trilinear type curve.
As far as the shear walls are concerned, as shown in Figure 6, beams and columns are able to
bear only axial tension only (beam ends, column capitals and bases are supported by hinges).
Wall member has axial and shear springs with the rotational springs at top and base. In other
words, wall member is modeled as virtual columns where the wall members can bear bending
and shearing forces.
McCracking moment
MyYielding moment
K0Elastic stiffness
KcStiffness after cracking
KyStiffness after yielding
Mc
M
My
Kc
K0
Ky
Figure 5. Model of beams.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
241
Nonlinear shear
spring
Nonlinear rotary spring
Rigid beam
Nonlinear axial spring
Figure 6. Model of shear walls.
3.2 Model of piles
Pile members are modeled, as shown in Figure 7. The relationship between the bending moment
(M) and curvature () (M  curve) is set to a model as a trilinear type skeleton curve. In the
modeling, the pile head axis force of superstructure required horizontal yield strength (at equiv
alent to required Ds) was used. Then based on the M curve calculated from the crosssection
equilibrium, the member characteristics of body structure was set up so as to obtain the expres
sion: (Area 1) = (Area 2). Moreover, the relationship of the concrete stressstrain is specified
by using the bilinear type skeleton curve, where the material strength of reinforcement is speci
fied 1.1 times larger than the design standard strength, according to e function method. More
over, the dissecting length of pile members is standardized as 1/2 of the pile diameter. In addi
tion, the curvature distribution in the dissecting zone is specified constant. It was assumed that
the pile and pile cap is jointed rigidly.
M
Mc
MyMu
c y u
McCracking moment
cCurvature of cracking
MyYielding moment
yCurvature of yielding
MuUltimate moment
uCurvature of ultimate
Muu
E1I1
E2I2
Mcc
Crosssection Analysis
Area2
Area1
Figure 7. Model of piles (M relation).
3.3 Model of sugrade reactions
The superstructure is modeled to a threedimensional frame, and columns and girders are re
placed with beam elements and bearing walls are replaced with the column beam. Spring mod
els for horizontal subgrade reaction, skin friction and end bearing are set based on the recom
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
242
mendation by Architectural Institute of Japan (AIJ 2001, Watanabe et al. 2004). The models of
subgrade reactions are drawn in Figure 8. The pile group effect is not included in the analysis.
a) Horizontal subgrade reaction
b) End bearing reaction
c) Skin friction
R
F
10mm(c)20(s)
0.8R
F
3mm(c)5(s)
cClayersSandy
R
u
1/2R
u
0.1D
0.6mm 10mm
P
max
A (=2)BC (=1/3)k
ho
y
P
A (=2)BC (=1)k
ho
A (=2)BC (=4)k
ho
k
h
= y
1/2
k
ho
Figure 8. Springs of soil reactions to piles.
Each spring is expressed as follows;
a) Horizontal subgrade reaction
Horizontal subgrade reaction is expressed as follows;
k
h
=ABCk
ho
(3)
where, k
h
= horizontal subgrade reaction per unit area; A, B and C = factors for dependency of
creep at static loading test, of strain at soil response and of soil nonlinearity, respectively; k
ho
=
horizontal subgrade reaction at displacement of 10mm (= 80E
o
d
3/4
); where E
o
= modulus of
deformation of ground; and d = diameter of piles.
The value of P
max
is a plastic subgrade reaction and is expressed as follows;
P
max
= 3.0 K
p
z d for sandy soil
= 9c (z 2.5d) for clayer soil (4)
=2(1+1.4z/d) cz < 2.5d
where K
p
= coefficient of passive earth pressure; = unit weight; z = depth; and c = value of co
hesion.
b) End bearing reaction
The relationship between force and displacement at pile tip is expressed as follows;
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
243
n
p pu
p p
p pu
p p p
A R
A R
A R
A R d S
+ =
/
/
) 1 (
/
/
1 . 0
/
(5)
where, = 0.23 (sandy soil) and 0.12 (gravel); and n = 2.70 (sandy soil) and 3.31 (gravel). And
S
p
= displacement at pile tip; Rp = vertical load; Ap = cross sectional area of pile; and R
pu
= ulti
mate end bearing capacity. The value of R
pu
is expressed as follows;
R
pu
100N
ave
A
p
(6)
where N
ave
= average N value obtained from the standard penetration test.
c) Skin friction along pile
The ultimate skin friction of pile (R
F
) is as follows;
R
F
= 10/3NsLs for sandy soil (7a)
R
F
= 0.5q
u
Lc for clayer soil (7b)
where Ns = N value for sandy soil layers ( 30); Ls = thickness of sandy soil; q
u
= unconfined
compressive strength ( 200kN/m
2
); Lc = thickness of clayer soil; and = perimeter of pile.
4 EVALUATION OF EQUIVALENT DAMPING FACTOR
4.1 Evaluation of equivalent damping coefficient
As the contribution of damping effects by SSI, following items are mentioned;
i) Hysteretic damping due to soil nonlinearity at contact points between piles or embedment and
soil
ii) Material damping of soil due to nonlinearity with shear strain dependency during Swave
propagation in the ground
iii) Radiation damping through pile or embedment to ground
In the study, only radiation damping is considered (AIJ 2006). A horizontal dashpot (viscous
damper) is set in parallel to the springs, as shown in Figure 9.
BWidth of Pile
D
u0Horizontal Displacement at Pile
Pile Tip
Ground Surface
C i
K i
uiHorizontal Displacement at Pile
Figure 9. Soil spring and dashpot in horizontal motion.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
244
The damping force in the perpendicular direction to pile axis is evaluated as follows (Iiba et
al. 2007);
i Hij Hij
u C F & = (8)
ij j Li i Hij
D d V C = (9)
where C
Hij
= viscous damping coefficient; u
i
= response displacement of all piles at ith layer in
horizontal direction;
i
= density of soils at ith layer; V
Li
= Lysmers velocity of soils at ith
layer; d
j
= diameter of jth pile; and D
ij
= length of division along jth pile. When the dis
placement at all pile head is set to be u
0
and the equivalent viscous damping coefficient at pile
head for jth pile is to be C
H0j
, the damping force at jth pile head is obtained as follows;
( )
= =
j i
i Hij j H j H
u C u C F & &
0 0 0
(10)
The total equivalent viscous damping coefficient of all piles (C
sw
) is shown.
) (
0 0 0
u C u C
j H sw
& & = (11)
The C
sw
is expressed by the damping coefficient of each dashpots (C
H0i
) as follows;
0
/ ) ( u u C C
i Hij
j i
sw
=
(12)
In case of rotational motion, the radiation damping model is illustrated in Figure 10. The ver
tical displacement at ith layer and jth pile is set to be v
ij
. The equivalent viscous damping coef
ficient at pile head is expressed as follows;
=
j i
j ij Vij ro
L v C C
0
/ ) (
(13)
ij j i S i Vij
D d V C = (14)
where
0
= average rotational angle at pile head; L
j
= distance from center of rotation; C
Vij
=
equivalent viscous damping coefficient at pile; and Vs
i
= Swave velocity at ith layer.
LjDistance to Pile
Center of Rotation
0Average Rotational Angle at Pile Head
iAverage Rotational Angle at Pile
vijVertical Displacement at Pile
Figure 10. Soil spring and dashpot in rocking motion.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
245
4.2 Equivalent damping factor of buildings with SSI system
The equivalent viscous damping factor (he) of the total system is expressed as follows;
3 3 3
=
eq
b
b
eq
ro
ro
eq
sw
sw e
T
T
h
T
T
h
T
T
h h (15)
where T
b
, T
sw
, T
ro
and T
eq
= predominant periods for superstructure, sway, rocking and whole
system in the ESDOF; and the h
b
, h
sw
and h
ro
= equivalent viscous damping factors of super
structure, sway and rocking system corresponding to predominant period of each mode.
The damping factors are obtained using following equations;
03 . 0
1
1
1
+
b
h
(16)
M
C
h
sw
sw
sw
1
2
= ,
2
1
2
e ro
ro
ro
MH
C
h
= (17)
where
1
= value dependent on structural system of superstructure (0.25 for general RC struc
ture); = the plastic ratio (ratio of story drift to yield drift) of superstructure; and
sw
and
ro
=
predominant circular frequencies for sway and rocking motions.
4.3 Reduction factor for Acceleration response spectrum
To get the acceleration and displacement responses according to the damping factor of system, a
following reduction factor (F
h
) for the ARS is used;
e
h
h
F
10 1
5 . 1
+
=
(18)
5 BUILDINGPILESOIL MODEL FOR RESPONSE CALCULATION
5.1 Outline of Models
Three buildings with 5, 8 and 14story and one soil ground are selected. Elevations of 8 stories
building are drawn in Figure 11. The short span (transverse) direction of the residential building
is analyzed, which consists of continuous bearing walls through the height in the direction. The
response of building in transverse direction is much influenced by sway and rocking motions of
the foundation. The height, mass distribution and other properties of the buildings are presented
in Tables 1, 2. The soil properties of ground are summarized in Table 3 and Figure 12. The soil
deposits have 0.73 s in predominant period at small shear strain region.
a) Longitudinal direction b) Transverse direction
Figure 11. Elevations of 8 stories building.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
246
Table 1. Dimensions and properties of buildings.
_______________________________________________________________________
Structure Floor 5story Floor 8story 14story
Height Mass Height Mass Height Mass
________________________________________________
m t m t m t
_______________________________________________________________________
Superstructure R 42.7 764
14 39.9 674
13 37.2 682
12 34.4 687
11 31.7 703
10 28.9 711
9 23.7 765 26.1 728
8 20.9 667 23.3 737
7 18.2 680 20.5 748
6 15.3 900 15.4 683 17.7 755
5 12.6 686 12.7 696 14.9 775
4 9.80 681 9.90 706 12.1 787
3 7.05 680 7.15 708 9.20 795
2 4.30 700 4.40 715 6.35 799
Foundation 1 1.10 805 1.00 1038 2.00 1405
Total 4452 6658 11749
Effective Value 9.32 4314 14.1 6094 27.5 9690
at first mode*
Embedment 1.30 1.40 3.20
______________________________________________________________________
* To include the depth of embedment
Table 2. Dimension of piles.
________________________
Story Diameter Tip depth
m m
_______________________
5 1.31.5 32
8 1.61.8 32
14 2.2 32
________________________
Table 3. Soil properties of ground model.
_________________________________________
Depth Soil type Density Vp Vs
_________________________________________
m t/m
3
m/s m/s
_________________________________________
14 clay 1.6 500 130
20 Fine sand with silt 1.8 450 150
28 Fine sand 1.8 575 200
>28 Gravel 2.0 1000 400
_________________________________________
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.00
0.20
0.40
0.60
0.80
1.00
1.0E05 1.0E04 1.0E03 1.0E02 1.0E01
h
G
/
G
0
Shear Strain
Clayey Soil
Sandy Soil
Figure 12. Nonlinear characteristic of soil.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
247
5.2 Earthquake motion
The earthquake motion is given as the ARS at the outcropped engineering bedrock as drawn in
Figure 13. The time history responses on the ground surface whose average ARS is drawn in the
same Figure, are calculated. Table 4 summarized the equivalent Swave velocities and damping
factors which are obtained by the equivalent linear analysis.
0
5
10
0.1 1.0 10.0
A
c
c
e
l
e
r
a
t
i
o
n
(
m
/
s
/
s
)
Period (s)
Ground Surface
Outcropped Engineering Bedrock
Figure 13. ARS of engineering bedrock and ground surface.
Table 4. Equivalent Swave velocity and damping factor.
___________________________________________________
Depth Initial Values Values during earthquake
Vs h Vs h
___________________________________________________
m m/s m/s
___________________________________________________
2.33 130 0.02 123 0.023
4.66 130 0.02 106 0.071
6.99 130 0.02 92.2 0.104
9.32 130 0.02 83.9 0.122
11.7 130 0.02 77.7 0.134
14.0 130 0.02 73.0 0.143
17.0 150 0.02 58.8 0.193
20.0 150 0.02 55.6 0.196
24.0 200 0.02 97.8 0.173
28.0 200 0.02 89.3 0.182
>28.0 400 0.02 400 0.02
___________________________________________________
5.3 Load and deflection relationship of buildings
Figure 14 presents the relationships between shear force and drift at each story of 8story build
ing through the nonlinear pushover analysis. The accelerationdisplacement relationship of the
ESDOF system is obtained as shown in Figure 15. In the figure, the displacements of super
structure, sway motion, rocking motion and total at the equivalent height of building are drawn.
The equivalent period of the system at any horizontal displacement, as shown in Figure 16, is
calculated using a following equation.
) / (
2
1 1
1
B
eq
Q
M
T = (19)
The equivalent period gradually increases with displacement due to the nonlinear characteris
tics of structural members and soil reaction.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
248
0
10000
20000
30000
40000
50000
0 5 10 15
S
t
o
r
y
S
h
e
a
r
F
o
r
c
e
(
k
N
)
Story Drift (cm)
1F
2F
3F
4F
5F
6F
7F
8F
Pile Head
Figure 14. Shear force vs. drift at each story (8F building).
0
200
400
600
800
1000
0 5 10 15 20 25
S
p
e
c
t
r
a
l
A
c
c
e
l
e
r
a
t
i
o
n
(
c
m
/
s
e
c
2
)
Spectral Displacement (cm)
Building Rocking
Sway
1
S
a

1
S
d
Figure 15. Relationship between
1
S
a
and
1
S
d
(8F building).
0.0
0.5
1.0
1.5
0 5 10 15 20
E
q
u
i
v
a
l
e
n
t
P
e
r
i
o
d
(
s
)
Spectral Displacement (cm)
Figure 16. Equivalent period with displacement (8F building).
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
249
5.4 Equivalent damping coefficient
Based on horizontal and vertical displacements of piles based on pushover analysis and the
equivalent swave velocities of soil deposits (Table 4), the equivalent viscous damping coeffi
cients at pile head are obtained as three lines drawn in Figure 17. The dependency of displace
ment or rotational angle at pile head on the damping coefficients is not so remarkable. The
damping coefficients tend to increase with horizontal displacement or rotational angle.
a) Sway motion
0.0E+00
1.0E+04
2.0E+04
3.0E+04
4.0E+04
5.0E+04
6.0E+04
0 1 2 3 4 5 6 7 8 9 10
V
i
s
c
o
u
s
C
o
e
f
f
i
c
i
e
n
t
(
k
N
s
/
m
)
Displacement (cm)
5F2c
8F2c
14F2c
5F2c(TLE,h=2%)
8F2c(TLE,h=2%)
14F2c(TLE,h=2%)
b) Rocking motion
0.0E+00
2.0E+06
4.0E+06
6.0E+06
8.0E+06
1.0E+07
1.2E+07
1.4E+07
1.6E+07
0.000 0.002 0.004 0.006 0.008 0.010
V
i
s
c
o
u
s
C
o
e
f
f
i
c
i
e
n
t
(
K
N
m
s
/
r
a
d
)
Rotation Angle (rad)
5F2c
8F2c
14F2c
5F2c(TLE, h=2%)
8F2c(TLE, h=2%)
14F2c(TLE, h=2%)
Figure17. Equivalent viscous damping coefficients.
5.5 Bending moment and curvature relationship of piles
The pile arrangement of 8story building is shown in Figure 18. Three different diameters of
piles are used in the building. The relationships between bending moment and curvature for two
piles are drawn in Figure 19. The characteristics of pile at position (X4, Y1) are calculated on
condition that the pile has an additional pulling force due to an overturning moment during
earthquake. On the other hand, the characteristics of pile at position (X4, Y2) are calculated un
der the additional compressive force due to the overturning moment.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
250
Y2
Y1
X1 X2 X3 X4 X5 X6 X7
Diameter of pile
: 1600 : 1700 : 1800
Figure18. Pile arrangement of the foundation (8F building).
0
2000
4000
6000
8000
0 0.001 0.002 0.003 0.004 0.005 0.006
B
e
n
d
i
n
g
M
o
m
e
n
t
(
k
N
m
)
Curvature (1/m)
PilepositionX4,Y1
0
2000
4000
6000
8000
10000
12000
14000
0 0.001 0.002 0.003 0.004 0.005 0.006
B
e
n
d
i
n
g
M
o
m
e
n
t
(
k
N
m
)
Curvature (1/m)
PilepositionX4,Y2
Figure19. Pile arrangement of the foundation (8F building).
6 RESPONSE OF BUILDINGS
Based on the equivalent period and equivalent damping factor, the response of the ESDOF sys
tem is obtained through Figure 1. The cross points between demand and capacity spectra give
the maximum acceleration and displacement of the system. Figure 20 shows the responses of 8
and 14story buildings under the acceleration responses at ground surface and the force
displacement relationship of the system with SSI.
Table 5 summarizes the maximum acceleration at equivalent height, the maximum shear
force at both 1
st
story and pile head and the maximum displacement of sway, rocking and build
ing at building top. The maximum response accelerations of buildings are about 4 m/s. The
shear forces at pile head is 1020% larger than those at 1
st
story. For 5story building, the ratio
of sway to total is 59% and the effect of sway on the total displacement is remarkable. For 14
story building, the effect of rocking is remarkable. In case of 8story building, the effects of
sway and rocking are similar.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
251
a) 8story building
0
200
400
600
800
1000
0 5 10 15 20 25 30
Spectral Displacement (cm)
S
p
e
c
t
r
a
l
A
c
c
e
l
a
r
a
t
i
o
n
(
c
m
/
s
e
c 2
)
SaSd Spectrum (h=5%)
SaSd Spectrum (h=14.7%)
Capacity Specturm
Te=0.675s
Response
b) 14 story building
0
200
400
600
800
1000
0 5 10 15 20 25 30 35 40
S
p
e
c
t
r
a
l
A
c
c
e
l
a
r
a
t
i
o
n
(
c
m
/
s
e
c
2
)
Spectral Displacement (cm)
SaSd Spectrum (h=5%)
SaSd Spectrum (h=12.3%)
Capacity Specturm
Te=1.49s
Response
Figure 20. Response of ESDOF system.
Table 5. Maximum response of buildings.
a) Acceleration and shear force
_______________________________________________
Story Acceleration shear force
Pile head 1
st
story
_______________________________________________
m/s
2
kN kN
_______________________________________________
5 3.94 1.70E+04 1.37E+04
8 3.87 2.36E+04 2.07E+04
14 4.77 4.65E+04 4.22E+04
_______________________________________________
b) Displacement (sway, rocking and building)
_____________________________________________
Story Displacement
Total Sway Rocking Building
_____________________________________________
cm cm cm cm
_____________________________________________
5 2.62 1.54 0.87 0.21
8 5.85 2.18 2.90 0.77
14 36.0 5.94 25.6 14.45
_____________________________________________
The equivalent periods and damping factors of buildings are summarized in Table 6. The
equivalent period is calculated as follows;
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
252
2 2 2 2
b ro sw e
T T T T + + = (20)
a total e
S u T
1
/ 2 = (21)
where u
total
= the displacement at equivalent height. The mode whose period is the largest
among the periods is predominant in the SSI system. For 5story building, the sway mode is
predominant. For 14story building, the mode of rocking is remarkable.
Table 6. Equivalent periods and damping factors of buildings.
a) Equivalent periods
_________________________________________
Story T
b
T
sw
T
ro
SSI(T
e
)
_________________________________________
s s s s
__________________________________________
5 0.12 0.39 0.23 0.47
8 0.24 0.47 0.42 0.68
14 0.61 0.70 1.17 1.49
__________________________________________
b) Equivalent damping factors
_______________________________________________________
Story Building(h
b
) Sway(h
sw
) Rocking(h
ro
) SSI(h
e
)
_______________________________________________________
% % % %
_______________________________________________________
5 3.46 23.5 30.6 17.3
8 4.32 25.6 24.0 14.7
14 4.07 31.0 18.5 12.3
_______________________________________________________
The maximum responses of bending moment of piles at positions (X4, Y1) and (X4, Y2) are
plotted in Figure 19. The distributions of bending moment at these piles are drawn in Figure 21.
0
5
10
15
20
25
30
4000 2000 0 2000 4000
D
e
p
t
h
(
m
)
Bending moment (kNm)
(X4,Y1)
0
5
10
15
20
25
30
6000 3000 0 3000 6000
D
e
p
t
h
(
m
)
Bending moment (kNm)
(X4,Y2)
Figure 21. responses of bending moment of piles.
7 DISCUSSION ON PILE STIFFNESS, DAMPING COEFFICIENT AND EARTHQUAKE
RESPONSE
7.1 Equivalent stiffness of pile head and viscous damping coefficient
The equivalent stiffness of pile and viscous damping coefficients during earthquake by pushov
er analysis are compared with those by the thin layer element (TLE) analysis. In the TLE analy
sis, the damping factor of soil deposits is set to be 2%. Because that the material damping of soil
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
253
deposit due to nonlinearity is not considered in this study. The dynamic impedance of the pile
foundation for horizontal and rocking motions are obtained, based on the linear approximation.
The dynamic impedances of pile foundation of 8story building are drawn in Figure 22. As the
impedance functions have a frequency dependency. Because the spring and dashpot values are
assumed to be constant in the following time history analysis, the real parts of impedances at
frequency of 0.1Hz are adopted as the spring constant, and the minimum values of the damping
coefficients are selected.
a) Sway motion
0.0E+00
2.0E+06
4.0E+06
6.0E+06
8.0E+06
1.0E+07
1.2E+07
0 1 2 3 4 5
I
m
p
e
d
a
n
c
e
(
k
N
/
m
)
Frequency (Hz)
Real Part
Imaginary Part
Spring Constant at f=0.1 Hz
Damping Coefficient(3.11*E+04 kNs/m)
b) Rocking motion
0.0E+00
2.0E+08
4.0E+08
6.0E+08
8.0E+08
1.0E+09
1.2E+09
0 1 2 3 4 5
I
m
p
e
d
a
n
c
e
(
k
N
m
/
r
a
d
)
Frequency (Hz)
Real Part
Imaginary Part
Spring Constant at f=0.1 Hz
Damping Coefficient(3.09*E+06 kNms/rad)
Figure 22. Impedance functions of foundation of 8F building.
The equivalent spring constants and viscous damping coefficients are compared with the re
sults by the TLE analysis in Table 7. And the equivalent viscous damping coefficients by the
TLE analysis are plotted in Figure 17. The sway spring values by the pushover analysis are a
good agreement to those by the TLE analysis. The difference is increased with height of build
ing. The rocking spring values by pushover analysis are lower than the TLE results. The reason
is the difference of the characteristics of spring models for end bearing and skin friction of piles.
The spring values by the pushover analysis decrease with the relative displacement between pile
and soil. On the other hand, the properties of soil are treated to be equivalent linear in the TLE
method. While the horizontal damping coefficients by the radiation effect have a good agree
ment to those by TLE results, the rotational damping coefficients are overestimated and about
three times those by TLE results.
In order to investigate the effect of spring values for end bearing, the springs which are al
most the same as those in the TLE analysis are set at the pile tip. The equivalent viscous damp
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
254
ing coefficients at pile head are drawn in Figure 23. By using the same amount of spring at the
pile tip, the rotational damping coefficients are about two times those by TLE results.
Table 7. Comparison of spring values and viscous damping coefficients.
a) Sway motion
_____________________________________________________
Story Pushover analysis Thin layer element
Spring Damping Spring Damping
_____________________________________________________
kN/m kNs/m kN/m kNs/m
_____________________________________________________
5 1.12E+06 3.24E+04 1.51E+06 2.81E+04
8 1.11E+06 4.16E+04 1.65E+06 3.11E+04
14 7.83E+05 5.41E+04 1.91E+06 3.68E+04
_____________________________________________________
b) Rocking motion
__________________________________________________________
Story Pushover analysis Thin layer element
Spring Damping Spring Damping
__________________________________________________________
kNm/rad kNms/rad kNm/rad kNms/rad
__________________________________________________________
5 2.82E+08 7.78E+06 5.37E+08 2.31E+06
8 2.80E+08 1.04E+07 6.97E+08 3.09E+06
14 2.11E+08 1.45E+07 1.04E+09 4.70E+06
__________________________________________________________
0.0E+00
2.0E+06
4.0E+06
6.0E+06
8.0E+06
1.0E+07
1.2E+07
0.000 0.001 0.002 0.003 0.004
V
i
s
c
o
u
s
C
o
e
f
f
i
c
i
e
n
t
(
K
N
m
s
/
r
a
d
)
Rotation Angle (rad)
5F2c
8F2c
14F2c
5F2c(TLE, h=2%)
8F2c(TLE, h=2%)
14F2c(TLE, h=2%)
Figure 23. Equivalent viscous damping coefficients for rocking motion as to another model.
7.2 Earthquake response of building
In order to investigate the amount of the responses by spectrumbased method, the time his
tory analysis on multi degree of freedom model of 5, 8 and 14story buildings is conducted. The
relationship between shear force and story drift of superstructures is a trilinear skeleton curve
based on the results by pushover analysis with base fixed condition. A standard rule under
unloading and reloading processes is used as the hysteretic characteristics. The damping factor
of superstructure is 3% in proportional to the initial stiffness of the story. In order to analyze the
SSI model, the spring and dashpot are added to the superstructure (SR model). The spring and
dashpot values are assumed to be constant in the time history analysis, as before mentioned. As
the spring constant, the real parts of impedances at frequency of 0.1Hz are adopted and the min
imum values of the damping coefficients are selected (Fig. 22).
The distributions of maximum shear force and maximum displacement for 8story building
by the spectrumbased method are compared with those by the time history analysis, as drawn in
Figure 24. While the maximum shear force by the spectrumbased method is a little smaller than
those by the time history analysis, the displacement is larger. The reason why the displacement
is larger is that the rocking (vertical) spring values by pushover analysis are lower than the TLE
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
255
results (Table 7). Since the sway and rocking spring value of 8story building by pushover anal
ysis is two third and one forth times those by TLE method respectively, the displacement of
sway and rocking motions is remarkable in whole displacement.
The reason why the story shear force is smaller is that the response of building does not in
clude the effect of higher vibration modes than 1
st
mode. From Table 1, the effective mass of
building for 1
st
mode is 0.92 times total mass. The effect of higher modes on the response needs
to be considered.
0
1
2
3
4
5
6
7
8
9
0 10000 20000 30000
S
t
o
r
y
Story Shear Force (kN)
Proposed
Method
Time
History
0
1
2
3
4
5
6
7
8
9
0 5 10
S
t
o
r
y
Displacement (cm)
Proposed
Method
Time
History
Figure 24. Story shear force and displacement distribution with height.
8 CONCLUDING REMARKS
Concluding remarks in the responses of buildings with SSI through the spectrumbased method
are summarized as follows.
1) The equivalent period of SSI system is estimated based on the pushover analysis of the build
ing, foundation and surrounding soil model.
2) The equivalent viscous damping factor is estimated based on the radiation damping effect
through piles.
3) The sway spring values by the pushover analysis are a good agreement to those by the TLE
analysis. The difference is a little increased with height of building. The rocking spring values
by pushover analysis are lower than those by the TLE analysis.
4) The horizontal damping coefficients by radiation effect based on the pushover analysis have
a good agreement to those by TLE analysis. The rotational damping coefficients are about three
times those by TLE analysis. In case where the springs of pile tip are set to be almost the same
as those used in the TLE analysis, the difference of equivalent viscous damping coefficients be
comes small.
5) The maximum shear forces by the spectrumbased method are a little smaller than those by
the time history analysis, while the displacements are larger. Because of low spring values,
sway and rocking displacements are more remarkable in the results by pushover analysis. As to
the smaller story shear force, the effect of higher vibration modes on the response of building
needs to be considered.
ACKNOWLEDGEMENTS
The authors express their sincere thanks to members of the committee Response evaluation
method and its example using spectrumbased method based on pushover analysis of building,
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
256
foundation and soil model, for their suggestion related to the model of pushover analysis and
the method of spectrumbased method.
REFERENCESS
Architectural Institute of Japan, 2001. Recommendation for the design of building Foundations (in Japa
nese)
Architectural Institute of Japan, 2006. Seismic response analysis and design of buildings considering dy
namic soilstructure interaction (in Japanese)
Iiba, M., Umemura, Y., Kurimoto, O. & Inoue, Y. 2007. A study on seismic response of building based
on response spectrum using pushover analysis of combined super and substructures, Structural field,
Summaries of technical papers of annual meeting, Architectural Institute of Japan, 99106 (in Japa
nese)
Kuramoto, H., Teshigawara, M., Okuzono, T., Koshika, N., Takayama, M. & Hori, T. 2000. Predicting
the earthquake response of buildings using equivalent single degree of freedom system, Proceedings
of 12th World Conference on Earthquake Engineering, Paper No. 1039
Kuramoto, H., Midorikawa, M. Teshigawara, M., Ueda, T. & Yoshimura, M. 2002. An overview of per
formancebased seismic code of buildings in Japan, Advances in Mechanics of Structures and Mate
rials, 333338, A.A. Balkema Publishers
Midorikawa, M., Hiraishi, H., Ohkawa, I., Iiba, M., Teshigawara, M. & Isoda, H. 2000. Development of
seismic performance evaluation procedures in building code of Japan, Proceedings of 12th World
Conference on Earthquake Engineering, Paper No. 2215
Umemura, Y., Kurimoto, O., Iiba, M., Watanabe, K., Kuramoto, H. & Tamura, M. 2004. A study on the
seismic load distribution along vertical direction considering soilbuilding interaction, Structural field,
Summaries of technical papers of annual meeting, Architectural Institute of Japan, 217220 (in Japa
nese)
Watanabe, K., Umemura, Y., Teshigawara, M., iiba, M. & Tamura M. 2004. A study on seismic response
evaluation by monotonic analysis of building and its foundation, Proceedings of 13th World Confe
rence on Earthquake Engineering, Paper No. 1939
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
257
1 INTRODUCTION
For more available method on the seismic designs, it is very important to develop application of
the performance based design method. The strength demand spectrum based on the ductility fac
tor is one of the most useful methods, which can be treated with nonlinear effects on the struc
ture subjected to seismic forces (Iemura et al., 1998). However, on the performance evaluation
due to the strength demand spectrum, it is not always clear how to evaluate the damage on the
structures with respect to the hysteretic energy and input seismic motion energy. The strength
demand spectrum based on the damage evaluation of the structure has also been examined (Mi
kami et al., 1998; Kimura et al., 2007). It has been indicated that one of the important factors on
the damage evaluations could be given by the total energy of the seismic motion on the structure
(Park et al., 1995). However, these researches have been carried out in the point of view on the
influence of the main shock.
It is presumed that the earthquake with severe intensity is so considerable effect due to after
shocks that it is important to examine effects on the damage evaluation. Since the damage of
structure may be closely related to the earthquake input energy, it is essential to evaluate accu
mulated energy on the structure due to aftershocks. For the evaluation of the damage of struc
ture by the aftershock, it is important to clarify the accumulated damage on the performance
based design due to earthquake. The appropriate estimation of the deterioration due to after
shock plays the important roles on evaluation of accumulated damages on the nonlinear re
sponse situation due to aftershocks.
In this present study, effects on the aftershocks are examined with the damage index for the
soil structure interaction system (SSI), which is modeled with the sway rocking model. It is in
dicated that the damage index is closely related with accumulated damage estimation of the
structure due to aftershocks. It is suggested that an increase of seismic intensity ratio of after
shock to main shock leads to an increase of accumulated damage of the structure. Therefore, it
Aftershock Effects on Damage Evaluations for Soil Structure
Interaction System
K. Kawano & Y. Kimura
Kagohisma University, Japan
ABSTRACT: An available evaluation on aftershocks plays important roles on the performance
based design method to seismic motions. In this present study, it is examined about the uncer
tainty effects on the damage index, which is evaluated with combining the ductility factor with
the hysteretic energy for the severe damage level of the structure for the soilstructure interac
tion system. While there are a little bit influences of the damage index due to after shocks, it is
shown that the uncertainty on the maximum seismic motion plays important roles on the dam
age index evaluation. For the performance based design to the seismic motion, the evaluation of
the uncertainty to the damage situation becomes so important that it can be carried out with the
fragility on the assigned damage level.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
258
is shown that the fragility on the assigned damage index leads to the available estimation on
damages due to aftershocks.
2 FORMULATION
Many researchers have examined for the dynamic SSI and supposed that the SSI plays impor
tant roles on the seismic response evaluations. Taking into account the dynamic response repre
sented with a few dominate vibration mode, the superstructure can be expressed with a single
degreeoffreedom (SDOF) system. The nonlinear characteristics on the structure can be ex
pressed with the trilinear model. The response evaluation of the total system may be represented
with a simplified model as shown in Figure 1. The governing equation of motion for the nonlin
ear SSI system is expressed with
[ ]{ } { } [ ( )]{ } { } M x C x K t x F + + = (1)
in which [M] and [C] denote the mass and damping matrix, and [K(t)] denotes the stiffness ma
trix on each time step including the soil foundation system with linear properties. {F} denotes
vector for the input seismic motion. { } x , { } x and { } x denote the acceleration responses, the ve
locity responses and the displacement responses, respectively. If the dynamic response is carried
out within linear region, the governing equation of motion can be solved with modal analysis
and spectral analysis. On the other hand, if the response of the structure causes to be nonlinear,
it is hard to solve the equation in the frequency domain.
Figure 1. An idealized structuresoilfoundation model.
Thus, applying the incremental method for the equation (1), this equation can be solved with
the increment method because of the nonlinearity due to the structure. The equation (1) can be
expressed with the incremental method as follows :
[ ]{ } { } [ ( )]{ } { } M x C x K t x F + + = (2)
in which
[ ( )]{ } { } K t x F = (3)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
259
2
4 2
[ ( )] [ ( )] ( )[ ] [ ]
K t K t M C
t t
= + + (4)
2
4
{ } { } { }{( ){ } 2{ }} [ ](2{ })
F F M x x C x
t
= + + + (5)
Therefore, the increment of the responses can be determined by solving the equation (3).The
iterative procedure can be carried out using the Newton Raphson method. The damage evalua
tion is conveniently conducted using the drift displacement and the energy ratio of the hysteretic
energy to the total input seismic energy. The drift displacement is depended upon the seismic
motion properties, the hysteretic characteristics and structural properties. For the damage
evaluations by means of combining the ductility ratio with the seismic response energy, the
damage index to the RC structure by Park et al is applicable to assess the damage situation by
means of the maximum displacement and the hysteretic energy as follows:
1
1 1
M
u y u
x
D dE
x Q x
= +
(6)
in which x
1M
and x
1u
stand for the maximum displacement and ultimate displacement, respec
tively. The ultimate displacement can be determined by the corresponding allowable ductility
ratio. The allowable ductility ratio is 5.0 in this present study. Q
y
denotes the yield force of the
structure and denotes the coefficient depending on the characteristics of the structural member
(Fajfar, 1992). The damage index is practically related to the damage situations by means of the
damage assessment of real structures experienced to seismic forces as shown in Table 1. In this
present study, the strength demand spectra based on the damage index from 0.1 to 1.0 are exam
ined.
Table 1. Relationship between Damage Index and Damage level.
3 RESULTS AND DISCUSSIONS
The accumulated damage estimation due to aftershocks is examined using typical seismic mo
tions in Japan. In the present study, the damage evaluation is carried out with the three seismic
motions of near field type, Kobens component, Takans component and Portns component as
shown in Figure 2, which are obtained by Kobe earthquake (1995). All the seismic motions
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
260
have a dominant frequency of relatively narrow range, for example, for Kobe ns from 0.3 sec to
0.8 sec.
In order to clarify the damage evaluation of the SSI system as shown in Figure 1, the seismic
response analysis of the SSI system is carried out using typical seismic motions corresponding
to the design spectra as shown in Figure 2. The soil foundation is supported by ground condition
with two layers, and the sway and rocking spring constant of soil foundation can be determined
by the each ground condition. Figure 3 shows the relation between the natural period and
ground condition of each seismic motion site. The natural period of SSI system is supposed to
be mainly depended on the superstructure.
Figure 2. Acceleration response.
Figure 3. Natural period due to each ground condition.
4 SEISMIC RESPONSES WITH AFTERSHOCK
For the evaluation on the influence of aftershock, it is necessary to simulate earthquake ground
motions including the aftershock. In this present study, it is assumed that substantial aftershock
would be occurred after the main shock. Assumption is made by the maximum acceleration of
aftershock, which yields the intensity of seismic motion from 10% to 70 % of main shock and
the same characteristic of motion.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
261
Figure 4 shows the time history of displacement response due to simulated earthquake ground
motions of the input intensity of 50% to Kobens. The two lines denotes the displacement re
sponse due to the damage index D=0.1 and 0.6, respectively. The natural period of the structure
is 0.5 sec. If the damage index is 0.6, the response leads to severe nonlinear response and to drift
of the displacement. Figure 5 similarly shows the time history of the displacement due to the
same seismic motion including aftershock. The natural period is 1.0 sec. The two responses are
denoted to the damage index D=0.1 and 0.6, respectively. The displacement response becomes
larger responses than the natural period 0.5 sec because of the dominant frequency of the input
motion. It is understood that the dynamic response due to aftershocks is affected for the damage
index 0.6 and the natural period of the structure.
Figure 4. Time history of displacement responses due to aftershock.
Figure 5. Time history of displacement responses due to aftershock.
Figure 6 shows the relation between restoring force and displacement to Kobens. Each line
denotes responses due to the main shocks and aftershock. The abscissa denotes the ratio of the
displacement to the yield one and the ordinate does the ratio of the restoring force to yield one.
The natural period of the superstructure is 0.5 sec and the damage index 0.6. It is understood
that the maximum nonlinear response is caused by the main shock of input seismic motion and
aftershock plays important roles on increase of the hysteretic energy.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
262
Figure 6. Response of restoring force.
5 DAMAGE INDEX DUE TO AFTERSHOCK
In order to evaluate the influence of aftershock, the seismic response analysis is carried out with
the simulated earthquake ground motion. Figures 7 and 8 show the damage index with respect to
the influence of the aftershock, which is assumed to be the input intensity of 50%, to Takans,
and Kobens, respectively.
The effects on the aftershock are estimated with the difference from the corresponding target
damage index. When the damage level of the structure is slight such as 0.2, a constant damage
index is obtained in spite of the natural period of the structure. It is understood that there are a
little bit influences on the damage index due to aftershock. However, if the severe damage level
is allowed in structure, it is indicated that the influence of aftershock leads to somewhat increase
of the damage index in comparison with the damage index due to only the main shock. More
over, it is understood that the damage index is considerably influenced by the natural period of
the structure. Therefore, the aftershock plays the important roles on the damage evaluation for
the performance based evaluation of the structure.
Figure 7. Damage index with aftershock (Taka ns).
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
263
Figure 8. Damage index with aftershock (Kobe ns).
6 ENERGY RESPONSE DUE TO AFTERSHOCK
As previous mentioned, it is indicated that the hysteretic energy plays important roles on the
damage index evaluation. The damage index due to the aftershock is essentially influenced by
the hysteretic energy. The damage evaluation is so important roles that it could be demonstrated
by combining the ductility factor with the energy ratio of the hysteretic energy to the total input
energy. Figure 9 shows the energy responses to Takans with the influence of the aftershock.
The natural period of the structure is 1.0 sec and the target damage index is 0.6. Each line corre
sponds to the damping energy, the hysteretic energy, and the input seismic energy, respectively.
It is noted that the hysteretic energy of main shock comes out suddenly sharp increment be
tween 5.0 sec and 10.0 sec in accordance with increase of energy of the input seismic motion. It
is known that increase of the hysteretic energy after about 60.0 sec is brought about the influ
ence of the aftershock. Moreover, the increment of hysteretic energy aftershock indicates rela
tively smaller than the main shock. However, the hysteretic energy due to the aftershock is sup
posed to extend as input intensity of the seismic motion increases. Therefore, in order to
evaluate the influence of aftershock, it is very important to clarify the increment of hysteretic
energy due to the aftershock.
7 FRAGILITY ON ASSIGNED DAMAGE INDEX
A few of uncertainty are inherently involved in the input seismic motion and the strength
characteristics of the structure. For the available estimation of the damage index, it is so signifi
cant roles that it is essential to clarify the effects on the uncertainty for the damage index
evaluation. Moreover, it is suggested that the uncertainty effects with respect to the damage
evaluation play important contributions on the damage evaluations for the performance based
design of the structure (Kimura et al, 2007). The damage index is evaluated with the yield dis
placement, the maximum displacement response, the expected ultimate displacement and hys
teretic energy. Taking into accounts the random property of the input seismic motion, the effects
of the uncertainty are so important that it should be involved with the damage evaluation of the
structure. In the present study, it is examined about the effect on the damage index due to the
uncertainty of the input maximum acceleration with the Monte Carlo simulation. The maximum
input acceleration of seismic motion is assumed to have a lognormal distribution. The coeffi
cient of variation is 10%.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
264
Figure 9. Time history of energy response (Taka ns).
The target design damage index is assumed to be 0.6 for the simulation. The probability of
excess is evaluated for the assigned damage index such as 0.2, and 0.4. The damage index 0.2
corresponds to the slight damage situation and 0.4 to the moderate damage situation as shown in
Table 1.
Figure 10 shows the fragility on the assigned damage index of the SSI system. The input
seismic motion is Kobens and the natural period 0.5sec. The abscissa denotes the mean value of
the maximum acceleration of the input seismic motion. When the assigned damage index is 0.2,
the excess is occurred from about 3.2m/s
2
, and the damage index simulated becomes above
the 0.2 for the maximum acceleration above about 5.0m/
s2
. For the damage index assigned
0.4, the excess is similarly appeared above the maximum acceleration 5.0m/s
2
and the prob
ability of excess above 0.4 becomes almost 100% for the maximum acceleration about
7.5m/s
2
. The differences between the damage index 0.2 and 0.4 are so considerable that it is
important to figure out the probability of excess for the target design damage index.
8 CONCLUSIONS
The effects on damage evaluation due to aftershocks of SSI system are examined. The main re
sults are summarized as follows:
(1) The damage index combined the maximum displacement with the hysteretic energy is
useful to evaluate the damage situation for the SSI system. The ground condition and the dy
namic characteristics of input seismic motion play significant roles on evaluations of the dam
age index.
(2) While the performance based design of the light level of damage can be evaluated by the
maximum ductility factor of the structure, it is available for the damage evaluations of the SSI
system to carry out combining the ductility factor with the hysteretic energy to the severe dam
age level of the structure.
(3) For the performance based design to the seismic motion, it is significant to clarify the un
certainty to the damage situation as demonstrated the fragility on the damage level assigned to
SSI system.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
265
Figure 10. Fragility on assigned damage index (Kobe ns).
REFERENCES
Iemura,H., Igarashi,A. & Takahashi,Y.(1998) Ductility and Strength demand for near field earthquake
ground motion: Comparative study on the Hyogoken Nanbu and the Northridge earthquakes, Struc
tural safety and Probability, pp.17051708
Kawashima,K., Macrae,G.A., Hoshikuma,J. & Nagaya,K.(1994.), Residual displacement response spec
trum and its application, Journal of structural mechanics and earthquake engineering, JSCE, No.501/
29, pp.183192 (in Japanese)
Kimura,Y. Kawano,K. & Nakamura,Y( 2007), Damage evaluation on nonlinear SSI systems due to
forces, Applications of statistics and probability in civil engineeringKanda,Takada&Furuta (eds), pp
18
Mikami,T., Hirao,K., Sasada,S., Sawada,T. & Nariyuki,Y.( 1998), A study on design spectra of seismic
intensity for level 2 earthquake, The 10
th
Earthquake Engineering Symposium, pp.30613066 (in
Japanese)
Park,Y.J. & Ang,A.H.S.(1995), Mechanistic seismic damage model for reinforced concrete, Journal
of Structural Engineering, Vol.111, No.4, pp.722739
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
266
SoilStructure Interaction Problem of a School Building Based on
Earthquake Records and Dynamic Substructure Method
M. Nakamura, Y. Kitamura, J. Suzumura & K. Hanada
Department of civil engineering, College of science and technology, Nihon university, Tokyo, Japan
1 INTRODUCTION
ABSTRACT: The importance of dynamic interaction has been often pointed out for common
low and mediumrise buildings however effects of dynamic interaction have not been actively
discussed at the stage of structural design. One of the reasons is that observation records for dis
cussing soilbuilding interaction have not been accumulated sufficiently. College of science and
technology, Nihon university, has established the earthquake observation network for the
ground and school buildings in Funabashi Campus. The authors discuss inertial and kinematic
interaction effects for a SRC school building that has five stories above the ground and one
story under the ground for which earthquake observations are performed. Inertial interaction ef
fects on the subject building's behaviors during an earthquake, which was recorded over 100gal
on the surface ground, are small, and such behaviors almost greatly depends on conditions of
the ground around the embedment of the building.
Earthquake resistance of socially important structures such as nuclear facilities has been evalu
ated by earthquake observation, vibration tests and theoretical analyses with dynamic interac
tion considered conventionally (Hanada 1987). On the other hand, the importance of dynamic
interaction has been often pointed out for common low and mediumrise buildings however ef
fects of dynamic interaction have not been actively discussed at the stage of structural design
(Tobita et al. 2009). One of the reasons is that observation records for discussing soilbuilding
interaction have not been accumulated sufficiently (Tobita et al. 2009).
College of science and technology, Nihon university, has established the earthquake observation
network for the ground and school buildings in Funabashi Campus as major subjects for under
standing input earthquake motions to structures and their behaviors during earthquakes (Naka
mura et al. 2006). Therefore, in this paper, the authors focus on a school building, which is a
typical low and mediumrise building, and discuss inertial and kinematic interactions. The soil
and building is modeled with lumped mass system, respectively. The dynamic substructure
method is applied to connect the soil and building with interaction springs. For the interaction
springs, solutions of the circular foundation by the theory of elastic halfspace are applied.
Moreover, inputs in the case of with and without excavation forces are discussed.
2 TARGET SITE AND OBSERVATION RECORDS
2.1 Subject structure
The subject structure is Building 14 of a steel framed reinforced concrete (SRC) structure lo
cated around the center of the campus shown in Figure 1. Figures 2 and Photo 1 show outlines
and an overview of the building, respectively. The north and east wings of Building 14 are in L
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
267
type form and neighbor each other. In the foundation of both wings, castinsitu concrete piles
are placed at 52 spots (27 for the north wing and 25 for the east wing) with a bearing stratum of
GL26m. Superstructures of both buildings are connected through expansion joints. The subject
for this study is the north wing, which has five stories above the ground (steel framed structure)
and one story under the ground (reinforced concrete structure). We focus only on the records
obtained in the lengthwise direction (Xdirection) of X5Y8 in RFL. The Xdirection almost
corresponds to the EWdirection. The east wings behaviors are excluded because observation
has been just started therefore both quantities and quality of the data are not sufficient. More
over, the behavior of north wings Ydirection is also excluded because it is presumed to be
greatly affected by the east wing. Some insitu soil investigation results indicated the building
region is almost on horizontally layered soil locally (Adachi et al 1995).
Building 14
(The north wing)
PointD
N
0m 50m 100m
Blg.8
Blg.4
Blg.13
Blg.9
Blg.12
Blg.11
Blg.10
Blg.6
Blg.5
Blg.3
Blg.2
Blg.1
Blg.7
Sports
hall
Library
Figure 1. Campus layout.
unit:mm
RFL
B1FL
5FL
4FL
3FL
2FL
1FL GL
X1 X2 X3 X4 X5 X6 X7 X8 X9
2
2
,
4
0
0
6
,
4
5
0
(1) Vertical section view.
(2) Horizontal section view.
60,600
X1 X2 X3 X4 X5 X6 X7 X8 X9
Y9
Y8
Y7A
Y7
Y6
Y5
1
9
,
6
5
0
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
268
(3) Pile location.
X1 X2 X3 X4 X5 X6 X7 X8 X9
Y9
Y8
Y7A
Y7
Y6
Y5
Y5A
Figure 2. Subject building.
The north wing
The east wing
Photo 1. A view of subject building.
2.2 Free field
Figure 3 shows insitu soil investigation results conducted at PointD shown in Figure 1. Inves
tigation depth was from GL46.5m to GL0m. From GL46.5m to GL20m, the Swave velocity
(Vs) propagation varies in the range of 300450m/s and it suddenly falls from GL15m to the
ground surface. As for soil properties, the ground surface is Kanto loam, the soil down to GL
20m is clay and loose sand, and those deeper are dense sand layers containing silt at intervals.
Nvalue exceeds 50 in the region of GL26m, which is a bearing stratum of Building 14. The
vertical array observations are carried out at PointD. In this study, ground surface behaviors of
PointD are assumed free field behaviors.
0
10
20
30
40
Vs (m/s) Damping
Density (t/m
3
)
Soil investigation
Lumped mass model
Multiple reflection model
seismograph
Figure 3. The results of insitu soil investigation at PointD.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
269
2.3 Observation result
The transfer functions estimated by observed records to be referred as an index of numerical
model creation are described. Transfer functions are estimated by Hv estimation (Bendat et al.
2000) for ten records of small earthquakes. Figure 4 shows epicenter distributions of those small
earthquakes. Epicenters of those small earthquakes are seen around Kanto and their epicentral
distances are below 60km. Figure 5 shows acceleration response spectrum at GL0m of PointD.
Dominant frequencies are seen around 5.5Hz in each record. Figure 6 shows transfer functions
of the free field system (GL0/GL46.5m) and soilbuilding system (RFL/GL46.5m).
M5
M4
Figure 4. Epicenter distributions.
0.01
0.1
1
10
100
1000
1 1
0
Frequency (Hz)
0.2
Figure 5. Acceleration response spectrum.
0
10
20
30
40
50
60
0 2 4 6 8 10
Frequency (Hz)
0
5
10
15
20
25
30
35
0 2 4 6 8 1
0
Frequency (Hz)
(1)GL0/GL46.5m. (2)RFL/GL46.5m.
Figure 6. Transfer functions.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
270
3 NUMERICAL MODEL
3.1 Interaction model
Figure 7 shows numerical interaction models. The soil and building were modeled with lumped
mass system. The dynamic substructure method is applied to connect the soil and building with
interaction springs for the purpose of examining inertial interaction effects. Rocking behavior is
ignored because the number of pile foundations placed in the subject building is large and only
the lengthwise behavior is focused on.
ModelA is the case without excavation force and, ModelB is that with excavation force. Re
sponses and excavation force is calculated from the earthquake responses obtained from a one
dimensional multiple reflection model. They are inputted to multiple points at the same time as
shown in Figure 8.
(1) ModelA. (2) ModelB.
Figure 7. Interaction model.
Figure 8. Considering excavation force.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
271
3.2 Ground model
First, a lumped mass model of the soil is examined. A depth range of the modeling is from
GL0m to GL46.5m. Vs distributions were obtained by revising the insitu soil investigation re
sults. The values used for this model are shown in Figure 3. Moreover, in order to assure that
the soil behaves as a free soil without being affected by building's behaviors, a total mass of the
soil was determined 1,000 times as large as the total mass of the subject building.
Next, a onedimensional multiple reflection model is examined to evaluate excavation forces.
The soil specifications used for Figure 3 are already shown. Same as the lumped mass modeling,
we revised the Vs distribution so that the dominant frequency of observed values can corre
spond to the model.
Transfer functions of GL0/GL46.5m obtained from the lumped mass model and multiple re
flection model are shown in Figure 9 together with observed values.
Over 7Hz for the multiple reflection model, differences occur from observed value however we
focus on effects of excavation forces here therefore we presume that these differences do not af
fect the following discussions.
0
5
10
15
20
25
0 2 4 6 8 10
Observed
Lumped mass model
Multiple reflection model
Frequency(Hz)
Figure 9. Transfer functions (GL0/GL46.5m).
3.3 Interaction springs
A foundation part of the subject building is rectangle. However, the authors predicted that there
would be no significant differences between the spring value, which is a detailed explanation for
the subject building, and approximate solutions. Therefore, in this study, we used a simple equa
tion to substitute the embedded rectangular foundation for a circular foundation with a radius of
r
0
whose area is equivalent to the rectangular foundation as shown in Figure 10 (1).
For the interaction springs between the bottom face of the building's embedment and the soil,
we applied equations derived from Tajimi's vibration admittance theory (Tajimi 1968). As
shown in Figure 10 (3), the authors discuss the case that rigid distribution (RD) is assumed as
reaction distribution of the soil in the case that a circular rigid plate on an antiinfinite soil is ex
cited horizontally on the interaction spring at the bottom face, and in the case of flat distribution
(FD). Novak's interaction springs are used as the interaction springs between the side face of the
structure's embedment and the soil. For ModelB with excavation force considered, dynamic
soil spring is approximated to a static spring with an assumption that Novaks interaction spring
on the building side are assumed uniform.
(1) Substitution of foundation. (2)Side, bottom, soil. (3)Circular rigid plate horizontally excited.
Figure 10. Substitution of foundation and circular rigid plate horizontally excited.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
272
3.3.1 Effects of structures stiffness
Effects of the buildings stiffness on the dynamic characteristics are discussed by comparing
transfer functions of the numerical model with observed values. Table 1 shows values of mass
and stiffness in the case that the subject building is substituted to lumped mass system. The
mass was calculated from design drawings. For the stiffness, the values calculated from the de
sign drawings were multiplied by 1.6 times so as to correspond it to the first dominant fre
quency of transfer functions obtained from the observed values. Figure 11 shows transfer func
tions of RFL/GL46.5m obtained with the parameters shown in Table 1. It was reported that
stiffness often do not correspond with small earthquake and/or microtremor observation results
when the specifications in the design drawings were used. A similar tendency was confirmed in
this study. This is because of the effects of socalled secondary members such as window
glasses, window frames and the walls between rooms, the authors presume.
Figure 11 also shows results estimated with RD and FD for reaction distribution of the bottom
face. As mentioned later, the behaviors do not greatly vary in either case. In other words, dy
namic characteristics of the transfer function expressing roofs / foundations of the building de
pend only on the buildings stiffness. We assume that the buildings stiffness behavior was
properly evaluated hence proceed to the following discussion.
Table 1. Parameters of the subject building.
Mass (t) Stiffness (kN/m)
m
SR
950.9
m
S5
824.9 k
S5
461744
m
S4
830.9 k
S4
669680
m
S3
840.7 k
S3
949616
m
S2
1020.0 k
S2
1031408
m
s1
2266.8 k
S1
1182112
m
s0
3456.2 k
S0
61573400
0
10
20
30
40
50
60
0 2 4 6 8 10
Observed
RD
FD
Frequency (Hz)
Figure 11. Transfer functions (RFL/GL46.5m)
3.3.2 Effects of interaction spring
Comparing the transfer functions of the numerical model with the observed values, we focus on
the effects of differences in reaction distributions of the bottom face spring given to dynamic
characteristics. We have discussed the case of RD for a bottom face spring of the building and
that of FD, same as the transfer functions in Figure11. There are no significant differences be
tween those two cases. The subject soil is comparatively soft and has a rigid foundation on it
therefore we use RD for the following study.
Moreover, great variations are not seen even after changing the spring on the bottom face of the
building therefore we cannot expect dramatic changes from a discussion on pile models.
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
273
4 EARTHQUAKE RESPONSE ANALYSIS
4.1 Input earthquake motion
An earthquake response analysis is performed and interaction effects are duscussed. EW
direction record observed at GL46.5m of PointD shown in Figure 12 is used as an input earth
quake motion which was recorded over 100gal on the ground surface. The earthquake wave in
Figure 12 is inputted into the bottom of the each model and was performed an earthquake re
sponse analysis in time domain.
50
25
0
25
50
0 10 20 30 40 50 60 70 80
Time (s)
Figure 12. Input earthquake motion.
4.2 Comparison of acceleration response spectrum
Figure 13 shows absolute acceleration response spectrum obtained from the top of the models as
a result of the earthquake response analysis. The figure also shows observed values in the build
ing RFL. The peak frequencies almost agreed with the observed values in ModelA however the
response amplitudes were much lower than the observed values. On the other hand, both the
peak frequencies and the response amplitude agreed with observed values in ModelB.
(1) Observed value. (2) Without excavation force. (3) With excavation force.
(ModelA) (ModelB)
Figure 13. Absolute acceleration response spectrum at RFL.
5 CONCLUSION
In this study, the authors have discussed inertial and kinematic interaction effects for a SRC
school building that has five stories above the ground and one story under the ground for which
earthquake observations are performed. In conclusion, it has been revealed that inertial interac
tion effect on the building's behavior during an earthquake is small and only those of input
should be considered. The followings are the conclusion obtained from this study.
1) Effects of interaction springs on the subject building's behavior during small earthquake is
small and such behavior almost depends on the buildings stiffness.
2) The subject building's behavior during an earthquake which was recorded over 100gal on
the surface ground is greatly affected by conditions of the soil around the embedment of the
building and we simulated building responses almost successfully by considering excavation
force.
0
200
400
600
800
1000
1200
1 10
h=0.05
h=0.20
h=0.40
A
c
c
e
l
e
r
a
t
i
o
n
r
e
s
p
o
n
s
e
(
g
a
l
)
Frequency(Hz)
0
200
400
600
800
1000
1200
1 10
h=0.05
h=0.20
h=0.40
A
c
c
e
l
e
r
a
t
i
o
n
r
e
s
p
o
n
s
e
(
g
a
l
)
Frequency(Hz)
0
200
400
600
800
1000
1200
1 10
h=0.05
h=0.20
h=0.40
A
c
c
e
l
e
r
a
t
i
o
n
r
e
s
p
o
n
s
e
(
g
a
l
)
Frequency(Hz)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
274
REFERENCES
Hanada, Kazufumi 1987. The modal Identification and System Identification for the SoilStructure Sys
tem. The report of Central Research Institute of Electric Power Industry UO5: (in Japanese)
Tobita, J., Fukuwa, N., Kojima, H. & Hamada, E. 2009. Development of SoilStructure Earthquake Re
sponse Observation for Evaluation of Dynamic Characteristics of Buildings, Journal of JAEE 9(3):
3756. (in Japanese)
Nakamura, Masataka (The earthquake motion and earthquakeresistant structure research group of CST
Nihon university) 2006. Earthquake observation system in Funabashi area, the state of the data man
agement: Journal of the research institute of science and technology Nihon university 112: (in Japa
nese)
Adachi, H. (The earthquake motion and earthquakeresistant structure research group of CST Nihon uni
versity) 1995. Study on earthquake motion characteristics based on earthquake observation system in
Narashino campus, Special report of the research institute of science and technology Nihon university
19: (in Japanese)
Bendat, J.S. & Piersol, A.G. 2000. Random data analysis and measurement procedure. New York: Wiley
Interscience.
Tajimi, H., Soilstructure interaction. 1968. Earthquake engineering. Tokyo: Shokokusha. (in Japanese)
Proceedings, 3rd GreeceJapan Workshop: Seismic Design, Observation, and Retrofit of Foundations. Santorini 2009
275
ABSTRACT: Soil interventions applied locally at the foundation level of structures result inevi
tably in the modification of the foundation soil properties. The modified stiffness of the soil be
low the structures affects in some extent both seismic wave propagation and soilstructure inter
action mechanisms that may finally lead to a substantially different structural seismic response.
Such interventions may include either conventional soil mitigation techniques, oriented for up
grading soil strength, or innovative solut
Molto più che documenti.
Scopri tutto ciò che Scribd ha da offrire, inclusi libri e audiolibri dei maggiori editori.
Annulla in qualsiasi momento.