Sunteți pe pagina 1din 167

PlaxisUsersMeeting

Singapore
2012

PUMSINGAPORE
Time
9:00

9:45

Title:ModelingSoftToeinBoredPilessocketedinWeatheredRocks
Prof. TanSiewAnn NationalUniversityofSingapore

9:45

10:30

Title:Plaxis2Dand3DFEMAssessmentofTrouserLegDiaphragmWallPanelsina
DeepExcavationinSingaporeSoftClay
MrChuahSzeSeng

10:30

11:00

Break

11:00

11:45

Title:AStudyofTunnelPileInteractionusingPlaxis3D
DrShenRuiFu NationalUniversityofSingapore

11:45

12:30

Title:NumericalModellingofTunnellingandTunnelSoilPileInteraction
MrHartono WuNationalUniversity ofSingapore

12:30

1:30

Lunch

1:30

2:15

Title:PlaxisTechnical Support
MrLinXinChengPlaxisAsiaPac

2:15

3:00

Title:PlottingGroundReactionCurve UsingPlaxis NATMTunnelModelling


MrSajjad AnwarAecomSingapore

3:00

3:30

Break

3:30

4:15

Title:PerformanceofaCantileveredDiaphragmWallIntegratedwithSoilImprovement
Block
MrEdward Lim GeoEngConsultants

4:15

5:00

Title:PlaxisExpertServices
MrJosephWongPlaxisAsiaPac

13/04/2012

ModelingSoftToeinBoredPiles
SocketedinWeatheredRocks
PlaxisUsersForum2012

Date:12April2012
By:ProfHarryTan

Outline
HypotheticalCase FEMmodelofsinglepile
withfixedDeadLoadsubjectedto15m
Drawdownsustainedoveraperiodof200
days
Whatisthedifferenceofpilewithorwithout
softtoe?
3DFEMmodelofboredpilewithorwithout
softtoe
Conclusions
2

13/04/2012

FEMAxissymmetricSinglePileModel
DL=6MN
Fill

GWT

SoftMarineClay
S4ResidualSoil

HighlyWeathered
S2Jurongrocks

Model1.4mDBOREDPILEOFLENGTH=28M
FROMGROUNDLEVEL
PileisconcretemodelaselasticsolidwithE=30
GPa (withdummybeamelementtogetaxialforce
plot)
15mGWTdrawdown
Modelofpossiblesoft
toe100mmthickbelow
pilebase
Interfaceelements
appliedontheinner
surfacecontactswithpile
baseandS3rock

Thenumberand
sizeofelements
arechosentogive
averyhighquality
meshforaccurate
numerical
computation:
Numberof
elements=1247,
averagesize<1m

CheckQualityofFEMMeshforAccurateAnalysis
4

13/04/2012

PlaxisInputParameters
TABLE1PLAXISINPUTFORPIER28PILE
Identification
2KallangMClay(MC)SU3S4a/b(MC)SU=150Rin 4S3(MC)SU=900Rint=05SoftToe0.1mSU=20 1FILL03mPHI'=30RINT=0.9
Identificationnumber
1
2
3
5
6
Materialmodel
MohrCoulomb
MohrCoulomb
MohrCoulomb
MohrCoulomb
MohrCoulomb
Drainagetype
Undrained(B)
Undrained(B)
Undrained(B)
Undrained(B)
Undrained(A)
Colour
RGB25,199,215
RGB135,234,163
RGB229,114,56
RGB54,22,218
RGB198,211,13
Comments
_unsat
kN/m^3
16
20
25
20
18
_sat
kN/m^3
16
20
25
20
20
E
kN/m^2
4000
7.50E+04
3.00E+05
1000
1.00E+04
(nu)
0.3
0.3
0.3
0.3
0.3
c_ref
kN/m^2
20
150
600
20
0.1
(phi)

0
0
0
0
30
C_v,ref
m^2/day
0.2692
10.1
538.5
0.1
134.6
R_inter
0.9
0.5
0.5
1
0.9
K_0,x
0.7601
0.7
0.7
0.658
0.5
k_x
m/day
5.00E04
1.00E03
0.01
1.00E03
0.1
k_y
m/day
5.00E04
1.00E03
0.01
1.00E03
0.1

S4strengthSu=150kPa,fs=0.5Su=75kPa,E=500Su=75,000kPa
S3(weakrock)strengthSu=600kPa,fs=0.5Su=300kPa,E=500Su=300,000kPa
R_inter isthealphavalueofskinfrictionloadtransferinUndrained LoadTest

RInter

RInter=0.5

ForS4soil,Su=150kPa,
Su/Pa=1.5,RInter=0.5
ForS2,Su=900kPa,Su/Pa=9,
RInter=0.33

AlphavalueforboredpileshaftfrictionasfunctionofSu/Pa(wherePa=100kPa)
AlphavalueistheR_inter valuetoapplyinFEMpilemodelinUndrained LoadTest
6

13/04/2012

SimulatedPileLoadTestComparePile
with/withoutSoftToe
6MN
8.7MN

12.3MN

Shaftfriction
FULLYmobilized
inS3/S4

Shaftfriction
NOTfully
mobilizedin
S3/S4

SettlementHistory

At10days,initialpileloadto
6MN;pilessettle67mmw/wo
softtoe
At20days,afterdrawdown
of15m,Pilewithoutsofttoe
settlesadditional67mm
BUTPilewithsofttoesettles
additional1011mm
After200daysofsoil
consolidation,groundsettleby
30mm.

13/04/2012

ConditionofPileWithoutSoftToewith15mdrawdown

NP=20m
NSF=78006000
=1800kN

Pilewithoutsofttoe,havesignificantendbearingresistance(3800kN)

SettlementofPileWithout SoftToewith15mdrawdown
27mm

NP=20m

SoilSettlement
PileSettlement

6mmatNP
Shaftfriction
NOTFULLY
mobilizedin
S3/S4

AtNPof20mdepth,pileandsoilsettletogether=6mm
Piletoepenetration=5.74.3=1.4mmintoS2layer
10

13/04/2012

ConditionofPileWithSoftToewith15mdrawdown

NP=17m
NSF=68306000
=830kN

Shaftfriction
FULLYmobilized
inS3/S4

Pilewithsofttoe,smallendbearingresistance(600kN)

11

SettlementofPileWithSoftToewith15mdrawdown
27mm

SoilSettlement

NP=17m

11mmatNP

PileSettlement

AtNPof17mdepth,pileandsoilsettletogether=11mm
Piletoepenetration=10.93.6=7.3mmintoSoftToelayer
12

13/04/2012

Summary
WithestimateactualDL=6MN,pilewith
adequateshaftfrictionresistancewillsettleby
67mmevenwithlarge15mGWdrawdown.
However,withinadequateshaftfriction
resistanceandsofttoecondition,pile
settlementwillincreaseto1011mmupon
15mdrawdown.

13

3DFEMSimulationofPilewithSoftToe
3DFEManalysisisconductedinthisstudytoexploretheeffect
ofwaterdrawdownonthesettlementbehaviorofthe4pilegroup
usingPlaxis 3DVersion2011.
The site has ground surface at about 104mRL, a pile cap located
at about 1m below ground surface with a pile cap thickness of 2m
supported by 4 bored piles with diameter of 1.4m each with pile
toe elevation located at about 76.2mRL. The centertocenter (c/c)
spacing of the piles is 4.2m.
A working load of 44MN (or 44,000kN) was assumed on top of
the pile cap.

14

13/04/2012

Corresponding3DFEMmeshforthe4pilegroup Topview

Pilecap

4solidcircular
boredpileswith
dia.of1.4mand
shroudedwith
interface
elements

15

ExaggeratedblowoutviewofInterfaceelementssurroundingtheboredpiles

Exaggeratedblowout
viewofInterface
elementssurrounding
theboredpiles.

16

13/04/2012

SoillayersandMohrCoulombParameters:
0~3mbelowgroundsurface:Topfillmaterial(Drained)withtypical
soilparametersofc=0.1kPa,=30 andE=10,000kPa.
3~15mbelowgroundsurface:SoftMarineclay(UnDrained)with
lowerparametersofc=0.1kPa,=22 andE=2000kPa.
15~24mbelowgroundsurface:S4material(Drained)withhigher
soilparametersofc=1kPa,=30.InviewthattheSPTNvaluesis
about40~50(averageof45)forS4layerasrevealedbythepost
boreholesafterstopworkanE=2N=90,000kPawasadopted.
24~32mbelowgroundsurface:S3material(Drained)withhigher
soilparametersofc=45kPa,=28 andE=400,000kPa.
From32mdownward:S2bmaterial(Drained)withhighersoil
stiffnessparametersofc=45kPa,=28 andE=900,000kPa(to
accountforsmallstrainhigherstiffnesseffects).
17

Soillayeringin3DFEMmesh

25kPasurchargeontheroads
besidethepier

Columnloadsappliedtopile
cap(insidetheground)

Topfillmaterial
Marineclaylayer
S4layer
S3layer

S2blayer

VERYMUCHFINER
elementsaroundand
alongthepileshafts

18

13/04/2012

TheFEMsimulationisstraightforward:
(1)Establishtheinsitustressregime;
(2)Wishinplaceofboredpiles/pilecapandapply44MN
(ignoredundrained behavior);
(3)Resetthedisplacementthusfartozero;
(4)ApplyinstantWaterdrawdownof15m.

Theinducedpile/piersettlementduetotheabovesimulated15m
waterdrawdownhasaMaxvalueofabout4.1mmasshownin
thenextslideforthiscaseofgoodnormalpileswithoutsofttoe
problem(perfectbasecontactwithS3weakrock).
19

Theinducedpile/piersettlementduetothesimulated15mwaterdrawdown
hasaMaxvalueofabout4.1mmforgoodpileswithoutsofttoeproblem.

20

10

13/04/2012

Likewise,pilecaptopsettlementcontourof
4.1mm

21

Theaboveanalysisshowsthatforgoodnormalboredpiles
supportingthePier,theinducedMaxsettlementwillbeabout
4.1mmsubjectedtothesustained15mwaterdrawdown, which
ismuchlowerthantheactualmeasurementofabout10mm.
Thequestionis:Whatelsecouldcontributetotherestofabout
5~6mmofadditionalpiersettlementwhichcannotbeaccounted
forinexpectednormalboredpiles?
WhatwouldbeeffectofaSoftToe?

22

11

13/04/2012

CorrectSimulationofSoftToeunderaBoredPile: Forthesimulationofsofttoe,
anannulusthincylindercanbecreatedbelowtheboredpilewithathinthicknessof
say100mmwithweakstiffnesslikeE=1000kPaandweakstrengthparametersof
c=0.1kPaand=22 likethoseforsoftclay.
However, foracorrectsimulationofsofttoeunderaboredpilein3DFEM
simulation,itiscriticalthediameterofthethinannulussimulatingthesofttoemust
bemadeslightlylargerthanthediameteroftheboredpile(say100mmlarger)so
thattheboredpileisallowedtosettleintothesofttoematerial.
Otherwise,ifthethinannulusissetasthesamediameteroftheboredpile,the
edgesofthebaseoftheboredpileandtherockmaterial(S3)attheedgeofbaseof
theboredpilewillsharethesameelementnodesandelementsidessuchthatthe
rockmaterialwillpreventtheboredpilefromsqueezingintothesofttoelayerto
settledownintothesofttoeannulus.
Inreality,theboredpilecanreadilysettleintotheunderlyingsofttoeevenifthe
softtoehasthesamediameteroftheboredpile.However,in3DFEMsimulation,
theabovecarefulandjudicioussimulationtechniqueisneededinordertoreplicate
theactualbehaviorofsofttoecorrectly.
23

Propersimulationofsofttoebeneaththeboredpilessimulationwithslightly
largerdiameterofthethincircularlayersimulatingthesofttoesoastoallowthe
boredpiletosettleintothesofttoematerial.

Smallsofttoeelements
aroundthepiletoeso
thatthepileconcrete
elementsandtheS3
elementsDONOTshare
commonelementnodes
andsidestoproviderigid
supporttothepile
elements

24

12

13/04/2012

Impropersimulationofsofttoebeneaththeboredpileswiththeedgesofthebaseof
theboredpileandtherockmaterial(S3)attheedgeofbaseoftheboredpileshare
thesameelementnodesandelementsidessuchthattherockmaterialwillprovide
rigidsupporttotheboredpiles,leadingtoincorrectsimulationin3DFEManalysis
(Thenextslidewillrevealthesofttoebeneaththepilebase)
Pileconcreteelements
andtheS3elementsshare
commonelementnodes
andsidestoproviderigid
supporttothepile
elements;theeffectof
thesofttoebeneaththe
pilebasewillthenbe
artificiallysuppressed
withtheWRONG3DFEM
simulationofsofttoe

25

Revealingofthesofttoebeneath1boredpiles(continuingwiththepreviousslide)

Softtoesamediameteras
pile:showedlittleimpact
tothepilesettlement
behaviorinsuchWRONG
3DFEMsimulation,as
explainedintheprevious
slide.

26

13

13/04/2012

Withthepropersimulationofsofttoebeneaththeboredpiles,whilewithoutchangingany
othersoilparametersandanalysismethodology,thepierisobservedtosettleforabout10.7mm
intothesofttoe;thatisthecorrectsimulationandrightexplanationoftheunusuallylargePier
settlementobservedatsite.

27

Likewise,pilecapsettlementof10.7mm

28

14

13/04/2012

Ontheotherhand,withtheincorrectsimulationofsofttoeasillustratedabove,thepiletoewill
berigidlysupportedbytheedgesoftheS3elementsaroundtheedgeoftheboredpilebase;
andthesofttoeelementsbeneaththeboredpileisnotabletoplayanysubstantialroleinthe
pilesettlementbehavior,whichisaWRONG 3DFEMsimulationofthesofttoe,leadingto
wrongconclusions.

Maxpilesettlementof
5.2mmonlyduetowrong
3DFEMsimulationofsoft
toe,whichisonly1.1mm
largerthanthecase
withoutsofttoeof
4.1mm,leadingtowrong
conclusionthatsofttoe
doesnothavesubstantial
impactonthepile
settlementbehavior

Softtoebeneaththepile
base

29

Conclusions(1)
Fromthe2DFEMstudyofasinglepile,itsobserved
thatforasofttoetoleadtolargerpilesettlements,
nearlyalltheshaftfrictionmustfirstbefullymobilized
forthepiletoplungeintothesofttoelayer.
PilesproperlydesignedtoFOSonshaftfriction>1.3
willnotsufferlargesettlementsevenwiththe
presenceofthinsofttoelayer(withverylowstiffness
of<1MPa).
Correctsimulationofsofttoein2DFEMcanbedone
withproperapplicationsofinterfaceelementsonthe
innersideofthesofttoelayer(interfaceelementsare
alwaysappliedonthesideoftheweakermaterial).
30

15

13/04/2012

Conclusions(2)
Fromthe3DFEMstudyofa4pilegroup,softtoe
mustbemodeledcorrectlytoallowforthepossibility
ofpilestoetoplungeintothethinsofttoelayer.
Theuseofathinlayerofsofttoematerialwithexact
diameterasthepile,withoutinterfaceontheinner
sideofthelayerwillpreventthepilefromplunging
intothesofttoelayer,evenifallshaftfrictionhad
beenmobilized.
Correctsimulationofsofttoein3DFEMcanbe
convenientlydonebyusingathincircularlayerofsoft
soilwithdiameterslightlylargerthanpileradius(say
0.1m).
31

16

16/04/2012

Sharing on Plaxis 2D and 3D FEM


Assessment of Trouser Leg
Diaphragm Wall Panels in a
Deep Excavation in
Singapore Soft Clay

Prof Harry Tan


Dr. N. Mace, C.K. Toh,
Dr. Haibo Yang,
Er. S.S. Chuah

Objectives
a)Sharing on how the Plaxis 2D was used to model
long and short D/Wall panels in DTL1 C908
b)Sharing on how Plaxis 3D was used to model 3D
problem of long and short D/Wall panels and
compared with Plaxis 2D results
c)Sharing on how the water table is considered during
excavation stage in Plaxis 2D and 3D Foundation to
ensure appropriate results were obtained
d)Sharing of parametric study of long and short panel
of D/Walls

16/04/2012

Outline of presentation
a)Project Information and description
b)Plaxis 2D and 3D modeling approach
c)Numerical Results
d)Parametric study on long and short panel
spacing
e)Conclusions

DTL1 C908 Project Information: Telok Ayer Station and Tunnels

16/04/2012

Long and short D/Wall


panels in DTL1 C908

A unique first-of-itskind Trouser Leg


diaphragm wall
panels in
Singapore
underground tunnel
project.

16/04/2012

Construction Sequence of C908 Tunnel Typical Section

Deepest Tunnel Section for Diaphragm Wall panels where


geotechnical analysis and monitoring results were compared
Shaded Panels are Long
Panels

16/04/2012

16/04/2012

16/04/2012

LSS
Backfill

Fill
F1

JGP
UMC

F1

LMC
F2

FCBB

Model I For Short D-wall panels : Out of plane presence of long panels is
modeled using anchor elements that do not interfere with the potential base
heave mechanism

LSS
Backfill

Fill
F1

JGP
UMC

E
LMC
F2
FCBB

Model II For Long D-wall panels


Long panel length (below base of short panels) is modeled as a separate beam
element, with a stiffness divided by the long panel spacing

16/04/2012

Plaxis 3D
Foundation Model

Full thickness of 0.8m D/Wall panel is


modeled
Reduced E value of 19.56E6 kN/m2 to
achieve 70%EI to be similar to Plaxis 2D
70%EI where E=28E6 kN/m2
Geological profile for 2D model applies
to full 3D model

16/04/2012

Plaxis2D:ClusterPhreaticlevelis
intheshadedarea(inJGPzone)

Plaxis2D:GeneralPhreaticlevelissetfor
thehighlightedzonebelowJGPlayer

16/04/2012

JGP is treated as
Undrained material
Work Plane is at 1m interval

below formation level at 89.4.


At RL 89.4, water below is
set at 89.4.
At RL 88.4, the water below
is set as Interpolate from
adjacent clusters or lines.
Similarly, at RL87.4 and
86.4, water level is set at
Interpolate from adjacent
clusters or lines.
At RL 85.4, soffit of JGP
layer, water below is set at
General Phreatic Level.
Enlarged Print

By c-phi reduction calculation,


FoS = 1.52

10

16/04/2012

11

16/04/2012

12

16/04/2012

13

16/04/2012

A Parametric Study to vary Spacing to 18m and


24m

14

16/04/2012

15

16/04/2012

16

16/04/2012

Observation: Apparently, there is no great difference between the magnitude


of results in wall movement, bending moments and shear forces for short and
long panels. This affirms that the complex long and short diaphragm wall
panels are close to a 2D problem with plane strain conditions.

Conclusions
1. Complex short and long diaphragm wall panels
for a tunnel box can be modelled in 2D analyses
with two different models, one with anchor
elements to hold down the short panel walls, and
another one with plate elements to simulate long
panel walls.
2. Spacing of long panel walls does not significantly
affect the results of the wall movements, bending
moments and shear forces. This affirms that the
complex short and long diaphragm wall panels
are close to a 2D problem with plane strain
conditions.

17

16/04/2012

Thank You
!

18

SINGAPORE PLAXIS USER MEETING AT


The National University of Singapore Society
(NUSS)

Performance of Cantilevered Diaphragm Wall


Integrated with Soil Improvement Block
by Edward Lim KH
12 April 2012

Content

Introduction
Design Basis
Design Construction Sequence
Actual Construction Sequence
Action Taken
Conclusion

Introduction

Introduction

Deep Soil Mixing


(DSM) or Jet Grouting
Piles (JGP) are
commonly used to
improve the soft soil
layer beneath the base
of excavation to
enhance the basal
heave and toe stability
of a temporary earth
retaining system.

Introduction

Jet Grouted mass has also been used as self standing


retaining system for canal reconstruction work.

Introduction

In the typical
application of
DSM and JGP,
the entire
improved soil
mass is
grouted by
interlocking
grout piles

Introduction

In this presentation the design and site issues faced in a


construction of a cantilevered diaphragm wall (D-wall)
supported by soil improvement blocks, in rectangular cells
formed by DSM and JGP, will be discussed.

Introduction

A 1.0m thick cantilevered diaphragm wall (D-wall) was


designed to support a terraced platform with a maximum
retained height of 15.8m.
Soil improvement blocks, in rectangular cells formed by
Deep Soil Mixing (DSM) method, was carried out on the
active and passive sides of the D-wall, forming an integrated
earth retaining system together with the D-wall.
Jet Grouting Piles (JGP) were used to fill the gaps between
the DSM and D-wall.

Cross Section of D-wall in Permanent Condition


RL112.0

RL104.5
RL102.0

JGP
FEL96.2

JGP
D-wall

Types of DSM Grid Patterns in Case Study


RL112.0

RL104.5
RL102.0

FEL96.2

Types of DSM Grid Patterns in Case Study


RL112.0

RL104.5
RL102.0

FEL96.2

Types of DSM Grid Patterns in Case Study


RL112.0

RL104.5
RL102.0

FEL96.2

Types of DSM Grid Patterns in Case Study


RL112.0

RL104.5
RL102.0

FEL96.2

Design Basis

Design Basis of In Situ Soil Mixing

In situ soil mixing Topolnicki, M. (2004)

Design Basis of In Situ Soil Mixing

INDIVIDUALCOLUMNS

IMPROVEDMASS
(DSM+SOIL)

IMPROVEDMASS
(DSMONLY)

Design Basis of In Situ Soil Mixing

Example in this project:

Respective total area, A =


3.0m x 10.1m = 30.3m2
Net Area of Soil Mixing
inside resepective total
area, At = 16.47m2
Ratio of area
improvement, ap = At/A =
16.47/30.3 = 0.54
c = 0.54 * 500 = 270kPa;
= 0.54*200,000=108,000kPa

Design Construction Sequence

Design Construction Sequence

RL112.0

RL112.0

Design Construction Sequence

RL112.0

RL112.0
RL109

Design Construction Sequence

RL112.0

RL112.0
RL109

Design Construction Sequence

RL112.0

RL102.5

Design Construction Sequence

RL112.0

FEL
RL96.2

Actual Construction Sequence

Actual Construction Sequence

RL112.0

RL112.0

Actual Construction Sequence

RL112.0

RL112.0
RL109

Actual Construction Sequence

RL112.0

RL112.0
RL109

JGP installation works


behind D-wall.

Progress of JGP Works

Excavation Side
(RL109)

8th March 2010

Inclinometer IW-1

1000mm THK D-wall

Inclinometer IW-2

12th March 2010


18th March 2010

Retained Soil Side


(RL112)

April 16, 2012

28

Inclinometer IW-1

17/3/2010
10/3/2010

22/3/2010

After JGP
installation
Move > 40mm

12/3/2010
JGP installed
just behind
IW-1

24/3/2010
Close-up
Diagram of
Inclinometer
Reading

Inclinometer IW-2
17/3/2010

22/3/2010

After JGP
installation
Move > 46mm

18/3/2010
JGP installed
just behind
IW-2

24/3/2010
Close-up
Diagram of
Inclinometer
Reading

Time History of IW-2 and Actions Taken

Original WSL = 95mm


Original AL = 67mm

End of June
2010

End of March
2010

STOP WORK. ASSESS


WALL INTEGRITY

Assessed Wall
integrity at this
stage. Still OK.
CURRENT
STAGE OK,
WITHIN
CAPACITY.

SUBSEQUENT
STAGE HOW TO
ASSESS WALL
INTEGRITY?

Stage of excavation during breaching of WSL


1/3 of Final Exc Level
completed. Remaining
2/3 yet to commence.

End of June
2010

Action Taken

Action Taken - Back-Analyses

To properly assess the BM in the D-wall, the shape of the


deflected D-wall must first be reproduced in the FEM
analysis.
To reproduce the measured wall deflection, the cause of the
wall movement has to be identified.
This will gives greater confidence to subsequent predictions
of wall response on resume of remaining 2/3 excavation.
First step to adjust soil parameters to match measured
deflection from stage 1 excavation before JGP work

Back-Analyses Original Soil Parameters

Soil Description

Bulk
Density
(kN/m3)

Cu
(kPa)

C'
(kPa)

Phi'
(deg)

E'50 (kPa)

*E'ur
(kPa)

Ko

k
(m/s)

Loose clayey SAND

18

--

0.5

30

8,000

N.A

0.5

1x10-6

Backfill material

19

--

30

10,000

N.A

1x10-6

Marine CLAY

16

15

20

0.87 x400 x cu

N.A

1.0

1x10-9

Peaty CLAY

16

15

20

0.87 x 400 x cu

N.A

1.0

1x10-9

OA (N<10)

19

5N

30

1,500 N

2.5* E'50

0.8

1x10-8

OA (10<N<30)

19

5N

33

1,500 N

2.5* E'50

0.8

1x10-8

OA (30<N<50)

20

5N

33

1,500 N

2.5* E'50

0.8

1x10-8

OA (50<N<100)

20

5N

35

1,500 N

2.5* E'50

1.0

1x10-9

Back-Analyses Revised Soil Parameters

Soil Description

Bulk
Density
(kN/m3)

Cu
(kPa)

C'
(kPa)

Phi'
(deg)

E'50 (kPa)

*E'ur
(kPa)

Ko

k
(m/s)

Loose clayey SAND

18

--

0.5

30

15,000

N.A

0.5

1x10-6

Backfill material

19

--

30

10,000

N.A

1x10-6

Marine CLAY

16

15

20

0.87 x400 x cu

N.A

1.0

1x10-9

Peaty CLAY

16

15

20

0.87 x 400 x cu

N.A

1.0

1x10-9

OA (N<10)

19

5N

30

1,750 N

2.5* E'50

0.8

1x10-8

OA (10<N<30)

19

5N

33

1,750 N

2.5* E'50

0.8

1x10-8

OA (30<N<50)

20

5N

33

1,750 N

2.5* E'50

0.8

1x10-8

OA (50<N<100)

20

5N

35

1,750 N

2.5* E'50

1.0

1x10-9

Back-Analyses Revised Soil Parameters

Soil parameters calibrated to


match site deflection as close
as possible using reasonable
soil parameters.
Measured Site Deflection at this
stage, site = 7.9mm
Black Line Using Original Soil
Parameters, Max = 28mm, far
from measured deflection.
Blue Line After revised
parameters on clayey SAND
and OA soil, Max = 22mm.
Still far from measured
deflection, but better.
These parameters are still
conservative to be adopted.

Clayey
SAND

Revised
Soil
Parameters

Peaty
CLAY
OA,
SPT-N=6
OA,
SPT-N=17
OA,
SPT-N=37
OA,
SPT-N=50

Original
Soil
Parameters

Measured
Deflection

Back-Analyses Revised Soil Parameters


Predicted
Line

JGP effects not considered.


The magnitude of the FE
prediction is less than the
measured value.
The deflection profile also
does not resemble the trend
measured by IW-2.
So, JGP installation effects
has to be taken into
consideration.

Measured
deflection
during 1/3
exc to FEL

Back Analyses: JGP installation effects

How to simulate the JGP installation effects?


Constant Volume Strain?
Varying Volume Strain?

Back Analyses: JGP installation effects

Apply Volume Strain in Plaxis How?


Plaxis Reference
Manual

Back Analyses: JGP installation effects


3% v

Apply CONSTANT VOLUME


STRAIN to see the effects.
Constant Volume Strain were
applied throughout the JGP
layer, using a number of
1%,3% and 5% strain.
None of the predicted
displacements at this stage
shows similar trend to what is
measured at site.

1% v
JGP
thickness
5% v

Measured
Site
Deflection

Back Analyses: JGP installation effects

VARYING VOLUMETRIC
STRAIN was imposed & the
result is Line A.
It is not very close to the
measured deflection, but the
trend is similar.

Predicted
deflection at
Stage Exc to
RL109

Measured Site
Deflection after JGP
works (24/3/2010)

Line A

Back Analyses: JGP installation effects

Note that this is not a prediction,


but a back-analysis by imposing
volumetric strain to fit the
measured wall deflection.
Also note that the imposed
volume strain at a particular soil
cluster is not always fully applied.
The actual wall movement
depends on the stiffness of the
wall and surrounding soils. That is
why Line A is a bit far-off from the
measured deflection on the first
trial.
Therefore, some iterations have
to be carried out to match the
measured deflection.

Predicted
deflection at
Stage Exc to
RL109

Measured Site
Deflection after JGP
works (24/3/2010)

Line A

Back Analyses: JGP installation effects

After a few iterations, a final


profile of Line B, which is
closer to measured deflection,
was obtained.
Once this was done, the
simulation of the remaining
excavation is carried out.
And the deflection and forces
in the D-wall at the end of the
excavation was assessed and
check against its structural
capacity.

Measured
Site
Deflection

Line A

Line B

Install JGP behind Wall (apply volume strain)

April 16, 2012

45

Install Soil Improvement in front of Wall

April 16, 2012

46

Exc to RL104.5

April 16, 2012

47

Excavate to RL96.2 (1/3 of area only)

April 16, 2012

48

Excavate to RL96.2 (1/3 of FEL only)

April 16, 2012

49

Final Profile

April 16, 2012

50

Wall Deflection at the End of Excavation

Inclino
Reading
after install
JGP

Line B

Inclino
Reading at
End of
Construction
Predicted
deflection at
End of
Construction
Considering
JGP installation

Wall Deflection at the End of Excavation

Line B

Inclino
Reading at
End of
Construction

Predicted
deflection at
End of
Construction
Considering
JGP installation

Inclino
Reading
after install
JGP

Conclusion

Conclusion

Design Issue:

The use of DSM in grid/cell form is effective.


The design concept of Soil Improvement Block using the approach of
area replacement ratio in Plaxis works well in this project, but further
research in understanding the actual mechanism is necessary.
The function of Volumetric Strain in Plaxis could be used to simulate
the effect of ground/wall movement due to grouting in back-analysis.
More research to determine the volumetric strain caused by various
types of grouting in different soil condition has to be done before one
could predict the ground movement due grouting work using the
volumetric strain method.

Site Issue:

Be aware of the sequence for JGP installation during construction


period. Do not underestimate the effects of JGP installation process.

Thank you

Back Analyses: After revising soil parameters,


then?

After revised soil parameters, the subsequent steps are:


Simulate JGP installation
effects behind D-wall after
Stage excavate to RL109

B.M. Capacity of Wall


were found insufficient,
Remedial Works required.

Proceed with subsequent


construction sequence until
permanent stage to include
permanent condition of Wall

Remedial
Works
thickening
of
considered.

Assess
Wall.

B.M.

Capacity

of

Will only focus


on this part in
this presentation

by
Wall

Excavate to 1/3 of FEL area.

April 16, 2012

57

Comparison with Inclinometer Reading at End of


Construction

If JGP installation effects is


not considered, this behavior
may not be captured.

Line B

Inclino
Inclino at
Reading
Reading
at
End
of
End
of
Construction
Construction

Predicted
Predicted
deflection
deflection
at at
End ofEnd
Construction
of
without
Construction
Considering JGP
installation

16/04/2012

AStudyofTunnelPileInteraction
usingPlaxis3D
RF Shen

Scenario1:
Impactoftunnelling
onexistingpiles

16/04/2012

Scenario2:
Impactofpiling
loadingonexisting
tunnels

(1)Aproposeddevelopmentwaslocatedadjacenttothefuture
developmentMRTtwintunnels;
(2)ThepilingsystemwithinMRTProtectionZoneadoptsboredpiles
soastominimizethedynamicimpactduringconstruction.
RCpilesoutsideMRT
ProtectionZone
Boredpiles
withinMRT
ProtectionZone

MRTProtection
Zone

Future MRTtwin
tunnels

16/04/2012

HOWtosimulatetheproblemusingPlaxis3D?

Mostcritical
sectionadoptedfor
thepresent3DFEM
analysis

Typicalcrosssection
Roadsurface

Boredpiledia.
1000mmwith40m
lengthwith28minto
underlyingOAsoils

16/04/2012

Boreholesatthislocalareaareadoptedfortheinterpretation
ofsubsurfacesoilprofile

GIBR soilparametersareadoptedfortheanalysis.Effectivedrained
parametersareadoptedduetothelongtermnatureoftheproject

Boredpiledia.
1000mmwith40m
lengthtorestonthe
underlyinghardOA
withSPT N>100

16/04/2012

IllustrationofeffectivedrainedsoilparametersfollowingGIBR
adoptedin3DFEManalysis

IllustrationofeffectivedrainedsoilparametersfollowingGIBR
adoptedin3DFEManalysis

10

16/04/2012

Constructingthe3DFEMmesh

11

3DFEMmeshwithsubsurfacesoilprofiles,pilegroups,tunnels
Pilegroups

25kPasurcharge

Workingload
onpilecap

Topfill

OA(E)
OA(D)
OA(C)
OA(B)

F1
F2

Upperand
closertunnel

Lowerand
farthertunnel

Underlyinghard
OA(N>100)
12

16/04/2012

Hidingofsomesoilelementstorevealthetunnelsandpiles
25kPasurcharge

Pilegroups
Workingload
onpilecap

Boredpiles
dia.1mwith
40mlength

tunnels

13

Scenario1:
Pilegroupsassumedtobeconstructedfirst;
Effectof2tunnelling(with2%volumelosseach)
ontheadjacentpilegroups

14

16/04/2012

Simulationsequence:
25kPasurcharge

Pilegroupswith
loadingsappliedfirst

TunnelsNOT
constructedyet
15

Lowertunnelactivatedwith2%volumeloss

Lowertunnel
activatedwith
2%volumeloss
16

16/04/2012

Theinvertofthetunnelwasrestrainedfromheavingup,soas
toinducemaximumtunnelshrinkinginwardwithmaximum
impacttosurroundingground

Crosssectionofmodeltunnel

3Dview

Asurprise:tunnelhasanoverallshrinkingin,therestraintat
theinverthasNOTeffect

18

16/04/2012

Asurprise:tunnelhasanoverallshrinkingin,therestraintat
theinverthasNOTeffect

hexagon
hexagon
tunnel
tunnel
composedof30
composedof24
sides,each12
sides,each15
19

Arelieftoremovetheunpleasantsurprise

Correctrestraintof
invertoftunnel

20

10

16/04/2012

Followedbytheuppertunnelactivatedwith2%volumeloss

Followedbythe
uppertunnel
activatedwith
2%volumeloss

21

Finaltunnelvolumelossshapes(scaledupby25times)

22

11

16/04/2012

Finalgroundmovementcontourplot
Maxgroundmovementaroundtunnelcrown,anddissipatesawayfromthetunnels
Immediatelyabovethetunnel,theinducedgroundsurfacesettlementisabout25mm;
whilethegroundmovementattheadjacentsiteisabout10mm
10mm
25mm

23

Theinducedmaxpiledeflectionisonlyabout6mmduetothe
2tunnelling with2%volumelosseach

24

12

16/04/2012

Theinducedmaxpilesettlementislessthan5mm

25

Maxpileaxialforceof5386kNbeforetunnelling;and5766kNafter2
tunnelling,anincrementof380kN,orabout7%incrementonly.

AxialforceBEFORE2tunnelling

AxialforceAFTER2tunnelling 26

13

16/04/2012

MaxpileBMtowardstunnels(M22)of90kNmbeforetunnelling;and104kNmafter2
tunnelling,anincrementof14kNmwhichisnegligibleforaboredpileof1mdiameter.

Bendingmomenttowardstunnels
M22BEFOREtwotunnelling

Bendingmomenttowardstunnels
M22AFTERtwotunnelling 27

MaxpileBMparalleltotunnels(M33)of60kNmbeforetunnelling;and63kNmafter2
tunnelling,indicatingnegligibleincrementofBMparalleltothetwotunnelling.

Bendingmomenttowardstunnels
M33BEFOREtwotunnelling

Bendingmomenttowardstunnels
M33AFTERtwotunnelling 28

14

16/04/2012

Finalpilemaxloading
condition:
FinalMaxworkingaxialforce=
5766kN;
FOS=1.4;
Factoredaxialforce=5766*1.4=
8072kN
MaxworkingBM:
M22=104kNm;M33=63kNm;
SoCompositeBM=122kNm;
FOS=1.4;
FactoredBM=170*1.4=170kNm
Thefinalloadingstateislocated
wellwithintheMNplot
envelope
29

Scenario2:
Tunnelsassumedtobeconstructedfirst;
Effectofpilegrouploadingsontheadjacent
tunnels;

30

15

16/04/2012

Simulationsequence

Twotunnelsin
placefirst

31

Pilegroupssubsequentlyinstalledandloadedsubsequently
Pilegroups
Workingload
onpilecap

Somesoil
elements
hiddentoshow
thepilegroups
32

16

16/04/2012

Additionaltunnelmovement(predominatelyverticalsettlement,scaledupby250times):

4.8mmforthetop/neartunnel
1.3mmforthelower/fartunnel
Botharequitesmallandtolerable
Bluecolortunnelis
original position
beforepileloadings
Redcolortunnelis
shifted position
afterpileloadings

33

Theadditionaltunnelmovementsaremainlysettlementwithminorlateralmovement:
Verticalsettlement:4.7mmforthetop/neartunnel
1.0mmforthelower/fartunnel
Horizontalmovement:1.2mmforthetop/neartunnel
0.7mmforthelower/fartunnel

Verticalsettlement

Horizontalmovement

34

17

16/04/2012

Totalnormalstressonthetunnelwallis957kPabeforepile
loadings;andincreasesto964kPaafterpileloadings,an
incrementofabout7kPa,farlessthanthespecified15kPa.
Interfaceelements
aroundthetunnels

LTA required
BEFOREpilegroup Incrementalpressure
ontunnel<15kPa
loadings

AFTERpilegroup
loadings

35

(1)ItisnottheintentionofthepresentanalysistoevaluatetheBMonthetunnel,but
justtoevaluatetheorderofincrementofBMonthetunnelduetopilegrouploadings.
(2)Thesagging/hoggingBMvaluesbeforepileloadingsis77/66kNm/m,respectively;
whileBMvaluesincreaseto80/67kNm/mafterpilegrouploadings,withanincrement
of1~3kNm/m onlywhichisquitenegligible.

BEFOREpilegroup
loadings

AFTERpilegroup
loadings

36

18

16/04/2012

Conclusions(1)
1) The1st3DFEManalysiswasconductedfortheScenario#1withpile
groupsassumedtobeconstructedfirst,followedbythe2tunnelling
with2%volumelosseachtoevalute theimpactonthepilegroups.
2) Inthe3DFEMvolumelosssimulation,theinvertofthetunnelwas
restrainedfromheavingup(becarefulofthesurprise!),soasto
inducemaximumtunnelshrinkinginwardwithmaximumimpactto
surroundingground.
3) TheMaxgroundmovementoccursaroundtunnelcrowns,and
dissipatesawayfromthetunnels.Immediatelyabovethetunnel,the
inducedgroundsurfacesettlementisabout25mm;whiletheground
movementattheadjacentsiteisabout10mm.
4) Theinducedmaxpiledeflectionisonlyabout6mmduetothe2
tunnelling with2%volumelosseach,andtheinducedmaxpile
settlementislessthan5mm,botharetolerabletothepilefoundation.
5) ThefinalforcesactingonboredpilesarecheckedtobewithintheMN
envelopeplot,indicatingthestructuralsoundnessoftheboredpiles
37
aftertunnelling impact.

Conclusions(2)
1) The2nd 3DFEManalysiswasconductedfortheScenario#2with
thetunnelsassumedtobeconstructedfirst,followedbythe
applicationoftheadjacentpilegrouploadingsandevaluateits
impacttothetunnels.
2) Theadditionaltotaltunnelmovementduetothepilegroup
loadingsisonlyabout4.8mmforthetoptunnelandisonlyabout
1.3mmforthelowertunnel,botharequitesmallandtolerable(<
15mm).
3) Theinducedincrementofpressureonthetunnelwallisonly
about7kPa,whichisfarlessthanthespecified15kPa.
4) Accordingly,theincrementofsagging/hoggingBMvaluesofthe
tunnelisonlyabout1~3kNm/mwhichisquitenegligible.
5) The3DFEManalysisillustratesthatdebondingoftheupper
portionofthepileslocatedwithintheMRTProtectionZoneis

notnecessary.

38

19

16/04/2012

Thankyouforyourpatience!

39

20

16/04/2012

Plotting Ground Reaction


Curve Using Plaxis
NATM Tunnel Modeling

Presented by
SAJJAD ANWAR
AECOM SINGAPORE
Slide 1

Overview

Slide 2

16/04/2012

GROUNDREACTION

MStage =0

MStage >0

MStage =1

Slide 3

PATTERNOFDEFORMATION
(withoutSupport)

Tunnel
Face

Slide 4

16/04/2012

GROUNDREACTIONCURVE

Slide 5

CONVERGENCECONFINEMENTCURVE
CONVERGENCECURVE
(GROUNDREACTIONCURVE)

CONFINEMENTCURVE
(SUPPORTREACTIONCURVE)

Deformation
before
installationof
support

Support
Elastic
Deformation

Slide 6

16/04/2012

LONGITUDINALDEFORMATIONPROFILE

Slide 7

SUPPORTDESIGN

AdequateSupportDesign/Capacity

Slide 8

16/04/2012

SUPPORTDESIGN

InadequateSupportDesign/Capacity

Slide 9

Empirical Method

Slide 10

16/04/2012

EMPIRICALEQUATIONS

Slide 11

REFERENCES

Slide 12

16/04/2012

SUPPORTCAPACITYESTIMATION

Slide 13

CONVERGENCECONFINEMENTCURVE

Slide 14

16/04/2012

LONGITUDINALDEFORMATIONPROFILE

Slide 15

Slide 16

16/04/2012

Numerical Modeling
Plaxis 2D
Plaxis 3D Tunnel

Slide 17

Plaxis2D
Plaxis3D

WhyNumerical
Modeling?

Slide 18

16/04/2012

Because!
EmpiricalMethodbasedonCloseForm
Solution
SuitableforCircularTunnelProfile/Geometry
HomogeneousGroundCondition
IsotropicStressRegime
Slide 19

WhyNumerical
Modeling?
NumericalMethodnotbasedonCloseForm
Solution
SuitableforAnyTunnelProfile/Geometry
InhomogeneousGroundCondition(Multilayers)
AnisotropicStressRegime
Slide 20

10

16/04/2012

Plotting Ground
Reaction Curve by
using Plaxis 2D

Slide 21

GROUNDPROFILE

63m
Slide 22

11

16/04/2012

CURVECONSTRUCTIONSEQUENCE

Slide 23

RELAXATIONFACTOR( FACTOR)

Slide 24

12

16/04/2012

TUNNELDEFORMATIONPROFILE

0,41.02

Slide 25

SELECTIONOFPOINTTOPLOTCURVE

0,41.02

Slide 26

13

16/04/2012

PLOTTINGCURVE

Slide 27

GROUNDREACTIONCURVE

Slide 28

14

16/04/2012

GROUNDREACTIONCURVE

Slide 29

Plotting Longitudinal
Deformation Curve
by using Empirical
Method

Slide 30

15

16/04/2012

LONGITUDINALDEFORMATIONPROFILE(LDP)

Thisprofilecanbeusedto
Establishadistanceconvergencerelationshipfor2D
modelingorforanalyticalsolutions
ThefollowingEquationsareproposedby
Vlachopoulos andDiederichs (2009a)toestimate
LDP
Slide 31

LONGITUDINALDEFORMATIONPROFILE(LDP)
InputParameters

RadiusofPlasticZone(Rp)

TunnelRadius(Rt)

TotalTunnelDeformation(Umax)

DistanceIntervalAheadandBehindof
Tunnelface(X)

Slide 32

16

16/04/2012

TUNNELANDPLASTICZONERADIUS

(FinalStage)
Slide 33

TOTALTUNNELDEFORMATION

Slide 34

17

16/04/2012

LONGITUDINALDEFORMATIONPROFILE(LDP)

Slide 35

Plotting Support
Reaction Curve

Slide 36

18

16/04/2012

SUPPORTREACTIONCURVE

Slide 37

SUPPORTREACTIONCURVE

SFPsm/peq>1.0

psm
equilibrium

peq

Slide 38

19

16/04/2012

DESIGNSPREADSHEET

Slide 39

RELAXATIONFACTORDETERMINATION

MStage =0.7

15.6mm
Slide 40

20

16/04/2012

COMPARISON

Slide 41

Verification by
Plaxis 3D

Slide 42

21

16/04/2012

3DTUNNELMODELMESH

Slide 43

TUNNELDEFORMATION

Slide 44

22

16/04/2012

TUNNELDEFORMATION

Slide 45

TUNNELDEFORMATION
Comparison

3D

2D

Slide 46

23

16/04/2012

TUNNELRADIALDEFORMATION
DEFORMEDMESH

Slide 47

TUNNELRADIALDEFORMATION

Slide 48

24

16/04/2012

TUNNELRADIALDEFORMATION

Slide 49

TUNNELRADIALDEFORMATION

Slide 50

25

16/04/2012

TUNNELDEFORMATION

RADIAL

FACE

Slide 51

RADIALDEFORMATION

Slide 52

26

16/04/2012

FACEDEFORMATION

Slide 53

DEFORMEDPROFILE

Slide 54

27

16/04/2012

DEFORMEDPROFILE

Slide 55

DEFORMATIONPROFILE

Slide 56

28

16/04/2012

DEFORMATIONPROFILE

Slide 57

PLANVIEW

TunnelFace

TunnelDrive

SECTIONALVIEW
Slide 58

29

16/04/2012

With Support

Slide 59

DEFORMEDPROFILE
WithSupport

Slide 60

30

16/04/2012

DEFORMEDPROFILE
WithSupportandExcavation

Slide 61

TUNNELDEFORMATIONPROFILE
WithSupportandExcavation

PLANVIEW

Slide 62

31

16/04/2012

TUNNELDEFORMATIONPROFILE
WithSupportandExcavation

SIDEVIEW

Slide 63

COMPARISON

Slide 64

32

16/04/2012

THANKYOU

Slide 65

33

Numerical modelling of tunnelling and tunnel-soil-pile interaction

Hartono
Research Scholar
The National University of Singapore
Courtesy of Keller

Plaxis Users Meeting, Singapore, 12 April 2012

Content

Ground movements due to tunnelling

Current predictive methods

Numerical modelling of tunnelling-induced ground movements

Case histories: Second Heinenoord Tunnel and C704 NEL Tunnel

Behaviour of piled foundation due to tunnelling

Back analysis of C704 NEL tunnel-soil-pile interaction

Concluding remarks

Typical TBM EPB machine


11
1

9
5

1.

Tunnel face

2.

Cutter head

3.

Excavation chamber

4.

Bulkhead

5.

Airlock

6.

Thrust cylinders

7.

Screw conveyor

8.

Lining erector

9.

Lining segments

10.

Belt conveyor

11.

TBM shield

10
4

2
7
3
6
Courtesy of Herrenknecht

Ground movements due to tunnelling

Tunnelling inevitably causes ground movements


Surface displacements
- Settlement trough
(transverse/longitudinal)
- Horizontal displacement
- Horizontal strain

Subsurface displacements

Greenfield or
tunnel-structure
interaction

- Subsurface settlement
- Subsurface horizontal
displacement

These movements must be quantified

Predictive methods:
- Empirical (Gaussian)
- Analytical
- Numerical (FEM)

Surface ground displacements can be estimated accurately using empirical


and analytical methods, but not so for the subsurface displacements

Results of FEA are very sensitive to many variables, but not so for
empirical and analytical methods

Ground movements due to tunnelling

Five (5) primary sources of ground movements

The ground movements, mainly due to over excavation, are usually


quantified as a lump parameter of volume loss.

Volume Loss - Definition

Gaussian settlement trough

y/i
-3

-2

-1

3
0

0.2
0.4

0.6

S/Smax

0.8

The volume of settlement trough is expressed as a proportion of the theoretical volume of


the tunnel excavation. The percentage is termed as a useful index of loss of ground or
volume loss (Peck, 1969)

The volume loss, VL, is the amount of ground lost in the region close to the tunnel.

Volume of settlement = Volume of over-excavation


Volume of settlement < Volume of over-excavation
Volume of settlement > Volume of over-excavation

no volume change
dilation
compression

However, for convenience, it is expressed as: (Mair and Taylor, 1997)


Volume Loss

Volume of settlement trough Area of settlement trough


=
Volume of excavated tunnel
Area of excavated tunnel

Ground movements due to tunnelling Gaussian

The most widely used method in practice due to its simplicity as only two parameters are
required to establish the curve, namely volume loss (VL) and trough width parameter (K).

The trough width parameter (K) has been established on the basis of the worldwide field
database. K is typically 0.4-0.6 for clays and 0.25-0.45 for sands and gravels (Mair and
Taylor, 1997).

OReilly and New (1982)

Loganathan, Poulos & Xu (2001) Analytical Solution

u z 0 o R 2

4 H 1

1.38 x 2
exp

2
H x
H cot R
2

zH
zH
2 z x 2 z H 2
u z 0 R 2
3 4

2
x 2 z H 2
x 2 z H 2
x 2 z H 2

1
3 4 4 z z H
u x 0 R 2 x

x 2 H z 2 x 2 H z 2
x 2 H z 2

exp

1.38 x 2

H cot R 2

4R 2

0.69 z 2

H 2


1.38 x 2
0.69 z 2
exp

2
H 2
H cot R

4 gR g 2

100%

Hypothetical problem
D = 6.5m; Zo = 30 m; VL = 1.5%
Offset from tunnel center-line (m)
-40

-20

20

40

-60

60

10
15
20

K=0.30 (Sand)
K=0.40
K=0.50 (Clay)
K=0.60

25
30

Surface settlement (mm)

Surface settlement (mm)

-60

-20

20

40

60

10
15
v=0
v=0.25
v=0.50

20
25
30

Gaussian

Offset from tunnel center-line (m)

-40

Loganathan, Poulos & Xu (2001)

Gaussian and analytical solution are not sensitive to a wide range of


variables, but not so in FEA

Back-analysis of settlement trough using FE

In literatures, there is no clear conclusion which facets of modelling affect the predicted
results
Soil model

Simulation technique of tunnelling (2D or 3D)

Simple soil model versus advanced soil model (Gunn, 1993; Dasari, 1996; Addenbrooke
et al., 1997)
Stiffness nonlinearity (Mair, 1993; Atkinson, 2000)
Anisotropic (Lee and Rowe, 1989; Simpson et al., 1996)
2D analysis: Lining contraction method (Vermeer and Brinkgreve, 1993), stress
reduction Beta method (Panet and Guenot, 1982), GAP method (Lee et al., 1991),
tunnel-volume loss method (Addenbrooke et al., 1997), grout pressure method (Moller
and vermeer, 2008), etc.
3D analysis: step-by-step procedure that simulating face pressure, shield advancement,
over-excavation due to pitching or yawing, tail-void grouting, lining installation, etc.

Effect of initial stress conditions


Effect of boundary conditions

Nevertheless, it has been well accepted that soil models with stiffness nonlinearity
better predict the ground movements compared to simple soil models (e.g., Mohr
Coulomb)

Finite element analysis of shield tunnelling 2D model


ground

ground

GAP

GAP method

Lining contraction

Vermeer and Brinkgreve, 1993

ground

Rowe et al. (1983)

ground

ground

-T/i

Pcrown
grout

Po

Stress reduction method

Tunnel volume loss method

Panet and Guenot (1982)

Addenbrooke et al. (1997)

PA

Grout pressure method

Moller and Vermeer (2008)

On Grout Pressure Method

FEA is very sensitive not only to soil models but also applied pressure. Moller
and Vermeer (2008) also arrived with a conclusion that the estimated ground
displacements are very sensitive to the grout pressure.

Comparison with the actual tunnelling process?

On Grout Pressure Method

The pressure at the tunnel crown could be taken from the pressure-volume
loss curve according to the expected volume loss

The unit weight of water is taken as a rate of increment

Volume loss is a prescribed value rather than a predicted value

Pressure (kPa)

ground
Pcrown
w

Modified grout pressure method

Case history Second Heinenoord Tunnel in the Netherland

8.3 m diameter tunnel with 15.6 m depth of the tunnel axis

Plaxis 2D v.9.0 was used. 1110 number of 6-noded triangular elements. Plane strain model.

Field record of surface settlement trough and subsurface horizontal ground movement was used as
a comparison with back-analysis results (Data from Moller, 2006)

Soil parameters were derived from Moller (2006) based on the soil report (Bakker, 2000)

Drained response is assumed for all soil layers


40m
35

Fill

19.75 m
23.25 m
27.5 m

Sand
Clay

27.5m

Sand

Max. surface settlement (mm)

0m
1.5 m
4m

30
25
20
15
10
5
0
0

1000

2000

3000 4000 5000


No. of elements

6000

7000

Hardening soil (HS)

Stress dependent stiffness


ref c cos 3 sin
E50 E50
c cos p ref sin

ref c cos 3 sin


Eur Eur
c cos p ref sin

Modulus

ref c cos 1 sin


Eoed Eoed
c cos p ref sin

m=0

0<m<1

Schanz, Vermeer and Bonnier (1999)

Hardening soil with small strain (HSSmall)

Atkinson & Sallfors (1991); Mair (1993)

Santos & Correia (2001)

q = 1 - 3

Gs

G0

1 0.385
0 .7

c cos 3 sin

G0 G0ref
c cos p ref sin

1
Benz (2006)

m=1

Case history Second Heinenoord Tunnel in the Netherland


Soil Parameters: MC and HSSmall
sat
(kN/m3)

Eoed
(MPa)

c
(kPa)

(kPa)

Ko

17.2

0.34

27

0.58

Sand

20

0.30

40

0.01

35

0.47

Sand

20

0.30

120

0.01

35

0.47

Clay, local parts of sand

20

0.32

48

31

0.55

Soil
Fill

ref
Eoed
(MPa)

ur

Soil

ref
E50
(MPa)

Eurref
(MPa)

G0ref
(MPa)

0.7

OCR

Fill

0.20

14

14

42

52

5.10-4

0.5

Sand

0.20

35

35

105

175

5.10-4

0.5

Sand

0.20

35

35

105

175

5.10-4

0.5

Clay, local parts of sand

0.20

12

35

88

5.10-4

0.9

Stiffness, E50 (MPa)


0
20
40
60

Stiffness, Eoed (MPa)


20
40
60

Stiffness, Eur (MPa)


20 40 60 80 100 120 140

1
Stiffness, Go (MPa)
40 80 120 160

200

0
5

Depth (m)

10
15
20
25
HS - HSSmall
MC

30
35

Case history Second Heinenoord Tunnel in the Netherland


250

134
Max. settlement (mm)

130
128
126
124
122
120

200
150
100
Gaussian

50

Current method

118

0
0

2
3
4
Volume Loss (%)

2
3
4
Volume Loss (%)

y/D

-5

-4

-3

-2

-1

0
0
0.2
0.4
0.6
0.8
1

Sv/Smax

Pressure (kPa)

132

Case history Second Heinenoord Tunnel in the Netherland


-40

-30

Distance from centre line (m)


-20

-10

Measurement
Gaussian (K=0.42) VL=0.8%
HS-Contraction(C=0.95%) VL=1.40%
HS-Contraction with invert fixity (C=0.8%) VL=1.43%
HS-Stress reduction (0.747) VL=1.39%
HS-Grout pressure (135 kPa) VL=1.21%
HS-Modified grout pressure (142 kPa) VL=1.23%

Settlement (cm)

Case history Second Heinenoord Tunnel in the Netherland


-30

Distance from centre line (m)


-20

-10

0
0

Measurement
Gaussian (K=0.42, GLR=0.8%)
HSSmall-Contraction(C=0.85%) VL=0.96%
HSSmall-Contraction with invert fixity (C=0.8%) VL=0.97%
HSSmall-Stress reduction (0.702) VL=0.90%
HSSmall-Grout pressure (118.5 kPa) VL=0.73%
HSSmall-Modified grout pressure (125 kPa) VL=0.80%

Settlement (cm)

-40

Case history Second Heinenoord Tunnel in the Netherland


Horizontal displacements (mm)
-5
0
5
10
15

-5

Horizontal displacements (mm)


0
5
10
15

0
5

Depth (m)

10
Tunnel axis

15
20
25

Measurement
HSSmall-Contraction (0.85%)
HSSmall-Contraction with total fixity (0.8%)
HSSmall-Stress reduction (0.702)
HSSmall-Grout pressure (118.5 kPa)
HSSmall-Modified grout pressure (125 kPa)

30
35

6m

10 m

Case history Second Heinenoord Tunnel in the Netherland

Contraction

100 times scale

GAP

Grout pressure

Stress reduction

Modified grout pressure

Case history Second Heinenoord Tunnel in the Netherland


Subsurface settlement (mm)
0m
-10
0

Depth (m)

10

20

5m
30

40

-10 0

10 20 30 40 0

10 m
4

12

16

20

HSSmall-Contraction
HSSmall-Contraction with invert fixity
HSSmall-Stress reduction
HSSmall-Grout pressure
HSSmall-Modified grout pressure

10

15

20

25

30

Case history C704 NEL Tunnel in Singapore

Case history C704 NEL Tunnel in Singapore

6.5 m diamater EPB shield machine was used. The tunnel axis was 21 m. The distance of SB and
NB tunnel was 16 m. SB tunnel was advancing first and later followed by NB tunnel

1.9 km viaduct was being constructed along the twin tunnels of Contract 704

The viaduct was supported by piers seating on bored piles

Comprehensive field measurements of ground movements and pile responses were reported by
Pang (2005)
62 m
30 m

0m
G4a
16 m
G4b

G4a

74 m

30 m
G4c

NB
tunnel

SB
tunnel

44 m

G4b
G4c

G4d
G4d

140 m

74 m
1623 of 6-noded triangular elements
(Plaxis 2D v9.2.)

329,872 of 10-noded tetrahedron elements


(Plaxis 3D General 2011)

Case history C704 NEL Tunnel in Singapore


HSSmall was used as a soil model. Basic soil parameters were obtained from Pang (2005).

sat
(kN/m3)

E50
(MPa)

G4a (0 < N < 15)

18

0.30

8.7

20

28

1.0

G4b (15 < N < 30)

19

0.30

40

30

30

1.0

G4c (30 < N < 60)

20

0.30

65

30

30

1.0

G4d (60 < N <100)

20

0.30

86.7

30

30

1.0

Soil

ref
Eoed
(MPa)

E50ref
(MPa)

Eurref
(MPa)

G0ref
(MPa)

G4a (0 < N < 15)

0.20

8.7

8.7

26.1

G4b (15 < N < 30)

0.20

40

40

120

G4c (30 < N < 60)

0.20

65

65

G4d (60 < N <100)

0.20

86.7

86.7

Stiffness, E50 (MPa)


0
40
80
120
10
20
Depth (m)

30
40
50
60
70
80

(kPa)

ur

Soil

c
(kPa)

Stiffness, Eoed (MPa)


40
80
120

Ko

0.7

Pref
(kPa)

179

2.10-5

0.5

100

534

2.10-5

0.5

250

195

907

2.10-5

0.5

350

260

1523

2.10-5

0.5

Stiffness, Eur (MPa)


75
150 225 300
HS - HSSmall
MC

500
Stiffness, Go (MPa)
400 800 1200 1600 2000

Case history C704 NEL Tunnel in Singapore


3.0

HSSmall-2D
HSSmall-3D General

FE Vol. Loss (%)

2.5
2.0
1.5
1.0
0.5
0.0
140.0

160.0

180.0
200.0
Pressure (kPa)

220.0

240.0

Case history C704 NEL Tunnel in Singapore


2D FE analysis Single Tunnel
0

Distance from SB centre line (m)


10
20
30
40

50

-50

Distance from NB centre line (m)


-40
-30
-20
-10

0
5
10

10

15

20

Measurement - Face loss

Measurement - Face loss

Measurement - Shield loss


Measurement
Measurement--Face
Taill loss
void closure
- Face 1.07,
loss 1.38%)
Measurement
- Shield
loss
GaussianMeasurement
(K=0.50,
VL=0.68,
Measurement
Measurement
- Taill void- Shield
closureloss
HSSmall-Modified grout pressure (161 kPa)
- Taill 1.07,
void 1.38%)
closure
Gaussian Measurement
(K=0.50, VL=0.68,
HSSmall-Modified grout pressure (154.5 kPa)

Measurement - Shield loss

SB Tunnel

Measurement - Taill void closure


Measurement
Face
loss 1.00, 1.67%)
Measurement
- -Face
loss
Gaussian
(K=0.50,
VL=0.45,
Measurement - Shield loss
HSSmall-Modified
groutloss
pressure (162.5 kPa)
Measurement - Shield
Measurement - Taill void closure
HSSmall-Modified
grout
(150.5 kPa)
Measurement
- Taill
voidpressure
closure
Gaussian (K=0.50,
VL=0.45,
1.00, 1.67%)

NB Tunnel

15
20
25
30

Settlement (mm)

Settlement (mm)

Case history C704 NEL Tunnel in Singapore


3D General
-50

-40

Distance from SB tunnel centre line (m)


-30 -20 -10
0
10
20
30

40

50

-20

60

10

10

20

-10

Lateral displ. (mm)


0
10
20

20

30

Depth (m)

Settlement (mm)

-60

40
Measurement - SB+NB (at pier P20)
3D-FE-SB+NB tunnels
3D-FE-Tunnel-pile (SB + NB)

50
60

30

40

50

60

70

Measurement (SB)
Loganathan & Poulos (1998)
FE-SB Tunnel
FE-SB+NB Tunnels

1 m away from SB tunnel

Case history C704 NEL Tunnel in Singapore


Subsurface ground settlement due to SB only
-20

-10

Subsurface settlement (mm)


0
10
20
30
40

50

-20

60

Offset
10, 8, 6, 4, 0 m

Offset
6, 4, 0 m

10
5

Depth (m)

Depth (m)

20

30

40

60

70

Subsurface settlement (mm)


0
10
20
30
40

50

-10

10

15
Mair et al. (1993) - y=0 m
Mair et al. (1993) - y=4 m
Mair et al. (1993) - y=6 m
Mair et al. (1993) - y=8 m
Mair et al. (1993) - y=10 m
Loganathan
& Poulos
(1998) - y=0 m
Mair
et al. (1993)
- y=0 m
Loganathan
& Poulos
(1998) - y=4 m
Mair
et al. (1993)
- y=4 m
Loganathan
& Poulos
(1998) - y=6 m
Mair
et al. (1993)
- y=6 m
Loganathan
& Poulos
(1998) - y=8 m
Mair
et al. (1993)
- y=8 m
Loganathan
&
Poulos
(1998)
- y=10 m
Mair
et
al.
(1993)
y=10
m
FE Modified grout pressure
FE
FEModified
Modifiedgrout
groutpressure
pressure

20

25

50

60

30

Content

Ground movements due to tunnelling

Current predictive methods

Numerical modelling of tunnelling-induced ground movements

Case histories: Second Heinenoord Tunnel and C704 NEL Tunnel

Behaviour of piled foundation due to tunnelling

Back analysis of C704 NEL tunnel-soil-pile interaction

Concluding remarks

C704 Tunnel Tunnel-soil-pile interaction

6.5 m

329,872 of 10-noded tetrahedron elements

C704 Tunnel Tunnel-soil-pile interaction


3D analysis Pile group P1
-10

Settlement (mm)
10 20 30 40

50

Axial Load (kN)


2000
4000
6000

60 0

8000

10

Depth (m)

20

30

40

50

60

70

Measurement - P1 (SB)
Measurement - P1 (SB+NB)
FE-P1 (SB)
FE-P1 (SB+NB)

FE-Soil settlement (SB)


FE-Soil Settlement (SB+NB)
FE-P1 (SB)
FE-P1 (SB+NB)

C704 Tunnel Tunnel-soil-pile interaction


3D analysis Pile group P1
-1000
0

Bending Moment (kN.m)


-500
0
500

1000

-15

-10

Lateral displc. (mm)


-5
0
5
10

10

Depth (m)

20

30

40

50

60

70

Measurement - P1 (SB)
Measurement - P1 (SB+NB)
FE-P1 (SB)
FE-P1 (SB+NB)

FE-Soil lateral displc. (SB)


FE-Soil lateral displc. (SB+NB)
FE-P1 (SB)
FE-P1 (SB+NB)

15

20

C704 Tunnel Tunnel-soil-pile interaction


3D analysis Pile group P2
-10

Settlement (mm)
10 20 30 40

50

60

Axial Load (kN)


2000
4000
6000

8000

10

Depth (m)

20

30

40

50

60

70

FE-Soil settlement (SB)


FE-Soil Settlement (SB+NB)
FE-P2 (SB)
FE-P2 (SB+NB)

Measurement - P2 (SB)
Measurement - P2 (SB+NB)
FE-P2 (SB)
FE-P2 (SB+NB)

C704 Tunnel Tunnel-soil-pile interaction


3D analysis Pile group P2
-400
0

Bending Moment (kN.m)


-200
0
200
400

600 -15

-10

Lateral displc. (mm)


-5
0
5
10

10

Depth (m)

20

30

40

50

60

70

Measurement - P2 (SB)

FE-Soil lateral displc. (SB)

Measurement - P2 (SB+NB)

FE-Soil lateral displc. (SB+NB)

FE-P2 (SB)

FE-P2 (SB)

FE-P2 (SB+NB)

FE-P2 (SB+NB)

15

Content

Ground movements due to tunnelling

Current predictive methods

Numerical modelling of tunnelling-induced ground movements

Case histories: Second Heinenoord Tunnel and C704 NEL Tunnel

Behaviour of piled foundation due to tunnelling

Back analysis of C704 NEL tunnel-soil-pile interaction

Concluding remarks

Concluding Remarks

Two case histories have been back-analysed using HSSmall

Stiffness nonlinearity (e.g., HSSmall) is a necessary soil model


feature for better estimation of ground movements induced by
tunnelling.

Soil constitutive models play a more influential role, compared


to the numerical tunnelling techniques, on the estimated
surface settlement trough.

Tunnelling methods influence the curvature of subsurface


settlement profile particularly that closes to the tunnel

Concluding Remarks

Pile behaviour due to tunnelling of C704 NEL has been


successfully back-analyzed using HSSmall coupled with the
modified grout pressure method.

More works probably need to be done on the 3D simulation


technique before it is being used routinely in practice

Acknowledgement

The National University of Singapore

Prof. C.F. Leung

Prof. Y.K. Chow

Prof. Harry Tan

Thank you for you kind attention

S-ar putea să vă placă și