etati

n H

alays

es thKench st

ness with standard penetration test (SPT) N value is found to be 1.5 N (MPa) with unload-ingreloading stiffness three times of tri-axial stiffness. The Hardening Soil model and thecorrelation obtained may be applied to similar soil conditions as the Kenny Hill Formation.

2013 Elsevier Ltd. All rights reserved.

ce suunne

logical settings is given by Tan [2] where the engineeringgeologic problems in these two cities were discussed. Sev-eral studies in the past have been conducted to examinethe soil parameters like stiffness. Tests for the stiffness ofsoil at very small strain were conducted in a hydraulic

complex calcula-od for estiid foundaposed by N

ura [4]. In this method, waves propagating in thetraced using the conception of the cone model, aimpulse response function can be calculated direceasily in the time domain with a good degree of accuracy.Lipinski and Wdowska [5] predicted the soil stiffness witha focus on Quaternary heavy over-consolidated stiff sandyclay. Series of tri-axial tests on reconstituted and naturalmaterial were carried out which provided data for setting

0263-2241/$ - see front matter 2013 Elsevier Ltd. All rights reserved.

Corresponding author. Tel.: +60 379675284; fax: +60 379675318.E-mail address: zubaidah_jka@yahoo.com (Z. Ismail).

Measurement 47 (2014) 645650

Contents lists available at ScienceDirect

Measure

eviehttp://dx.doi.org/10.1016/j.measurement.2013.09.030has been assuming an increasingly important role in theprediction of ground and wall deformations as highlightedby Lee et al. [1].

A general overview of the Kuala Lumpur and Ipoh geo-

a practical method that does not requiretions. Another simple and practical meththe horizontal dynamic stiffness of a rigthe surface of multi-layered soil was promatingtion onakam-soil arend thetly andsubstantially in highly urbanized areas due to scarcity ofthe land. A major concern in these developments is theability of the geotechnical engineers to predict accuratelythe wall and ground movements associated with the con-struction activities during the design stage. Nowadays,numerical analysis such as nite element method (FEM)

studied by Viggiani and Atkinson [3]. Simple expressionswere obtained which described the variation of strain interms of the current stress and over consolidation ratio.The parameters in these expressions were found to dependon plasticity index. The inuence of layered soil in soilstructure interactionwas also estimated. Themethod offersKeywords:Deep excavationKenny Hill FormationHardening Soil modelOedometer stiffnessStandard penetration testTri-axial stiffness

1. Introduction

The demand for underground spament car-park, road and railway tch as deep base-ls has increased

tri-axial cell tted with bender elements and with localaxial gauges for characterizing the non-linear stressstrainbehavior of soil formonotonic loading required for analysesof the dynamic and small strain cyclic loading of soils wasAccepted 16 September 2013Available online 2 October 2013

relation between stiffness parameters with eld standard penetration tests (SPTs) N valuefor Hardening Soil model. The correlation between tri-axial stiffness and oedometer stiff-Determination of soil stiffness paramconstruction site in Kenny Hill Form

Law Kim Hing a, Siti Zulaikha Othman b, RoslaaKH Geotechnical Services, 50603 Kuala Lumpur, MalaysiabCivil Engineering Department, University of Malaya, 50603 Kuala Lumpur, M

a r t i c l e i n f o

Article history:Received 2 December 2012Received in revised form 4 September 2013

a b s t r a c t

This paper determinment car park in thewall deection at ea

journal homepage: www.elsers at a deep excavationon

ashimb, Zubaidah Ismail b,

ia

e residual soil stiffness parameters at a deep excavation for a base-ny Hill Formation. Parametric studies revealed that the horizontalage of excavation could be reasonably predicted with a simple cor-

ment

r .com/ locate/measurement

phosed into quartzite and phyllite. The weathering process

646 K.H. Law et al. /Measurement 47 (2014) 645650up formula for calculation of Youngs modulus in a widerange of strain 102/1.0%. A series of tri-axial tests was alsoconducted by Powrie et al. [6] on samples of speswhite kao-lin, to investigate the stressstrain relations appropriate todiaphragm walls in clay. The results of the tests highlightthe inuence of the recent stress history on the behaviorof the soil. In particular, the recent stress history imposedduring wall installation was found to have signicant effecton the stiffness of the soil during the subsequent excavationstage. Although the pre-excavation stress state of the soilmay be closer to the passive than the active condition, thereversal in the direction of the stress path at the start ofthe excavation stage means that the response of the soilbehind the wall will probably be very stiff. Shaee et al. [7]conducted an experimental study investigating the pre-failure and failure characteristics of compacted sand-claymixtures under monotonic compression and extensionloading paths. Results revealed that pore pressure, secantmodulus, undrained shear strength and angle of shearingresistance increase when sand content was raised in bothcompression and extension. It was also found that thetested materials were over-consolidated by the fact thatnormalized shear strength depends on initial conningstress.

The ber Bragg grating (FBG) sensor has been widelyused in the measurement of temperatures and moisture[810]. Jackson et al. [11] examined the feasibility of usinginexpensive wireless nanotechnology based devices forthe eld measurement of soil temperature and moisture.In their study, the design, validation, and application of anewexible ber Bragg grating (FBG) sensing beamare pre-sented for effectively measuring dynamic lateral displace-ments inside soil mass in a shaking table test. The dynamiclateral displacements at different depths of the soil massin the shaking table box throughout time history are calcu-lated by differential and integral methods Xu et al. [12].

In the past, the performance of deep excavations in Ken-ny Hill Formation was mainly evaluated using 2D nite-element back-analyses as exemplied by studies by Liewand Gan [13], Soana and Hooi [14] and Tan et al. [15].Approximation is commonly needed in 2D numerical mod-el to represent the real situations and this could lead touncertainty in the interpretation and validity of the resultsas shown by Simpson et al. [16]. The correlation of soilstiffness parameters with standard penetration test (SPT)N value, which was calibrated based on 2D back-analysesresults may not be representative of actual condition atthe site. Field data clearly indicated that the stiffeningeffect of corners lead to much smaller wall and groundmovements at the corners as compared to that measurednear the middle of the excavation wall as shown by Leeet al. [17], Ou and Shiau [18] and Ou et al. [19]. In this case,when back analyses were performed to calibrate the 2Dmodel, the soil stiffness would have to be increased inorder to match the observed wall deection. Therefore,3D geometrical or corner effect needs to be consideredwhen back-analyses were performed in order to get ameaningful empirical correlation to be adopted in thefuture in same soil conditions.

A common problem in the analysis of deep excavationin residual soils is the soil tests data often limited or lowof the rock material which is rather complex have beendescribed by Raj [22].

The soil prole at this project site consists of an upper6 m of recent alluvium underlined by Grade IV to VI resid-ual soils of Kenny Hill Formation up to depths of about30 m. Highly fractured and weathered Siltstone with RockQuality Designation (RQD) of 0% is encountered beyond30 m depth. Grading analysis revealed that the residual soilmainly consists of sandy silt and clayey silt material. Stan-dard penetration test (SPT) blow counts were low in thealluvium layer but increases beyond 50 blows/300 mmfrom depth exceeding 10.5 m. The high SPT-N valuesexceeding 150 blows/300 mm were probably due to thepresence of quartz veins or phyllite fragments encounteredin the boreholes. The bulk density of residual soil layers aregenerally ranged from 19 kN/m3 to 22 kN/m3 with depth.The moisture content of residual soil layers are close toplastic limit with plasticity index generally lying inquality due to the difculty in obtaining undisturbedin situ soil samples. Very often, acceptable data on strengthproperties of soil could be obtained through laboratorytests but not on its modulus value. Therefore, informationfrom back-analyses of the Youngs modulus based on localcase histories, if available, are often very useful for engi-neering judgment in the estimation.

This study examined the soil stiffness parameters for adeep excavation supported by diaphragm wall in weath-ered residual soils of Kenny Hill Formation. An elasto-plastic isotropic Hardening Soil (HS) model followingSchanz et al. [20], as implemented in commercial nite ele-ment program PLAXIS, was employed in this study. Theobjective is to provide data for the determination of hori-zontal displacements which can also be generally appliedto other excavation works in soil conditions similar tothe Kenny Hill Formation.

2. Material and methods

The study project is located at Lebuh Ampang, KualaLumpur city center. It is a 24-storey ofce building with5 levels of basement car-park. The construction of base-ment involved 18.5 m deep of excavation, approximately30 m wide and 35 m long, in weathered residual soils ofKenny Hill Formation. The excavations were performedusing the bottom-up method. The diaphragm wall of23 m deep and 0.8 m thick was supported by three levelsof H-section steel struts with 3.5 m horizontal spacing onaverage. A double steel section was used for 2nd and 3rdlayers strut to provide sufcient resistance against highhorizontal earth pressures at these levels. At the contactpoint between the strut and diaphragm wall, I-sectionwalers supported by angle brackets were installed to pro-vide better load transfer between the retaining wall andstruts.

The ground condition at the site generally consists ofresidual soils and weathered rocks of the Kenny HillFormation. This formation is also referred by Komoo [21]as meta-sedimentary, considering that the sedimentaryrocks (e.g. sandstone, siltstone) have been partly metamor-

between 15% and 30%. The groundwater table is located atdepth of 4.5 m below ground surface.

The movements of the diaphragm wall, the ground andthe adjacent buildings were monitored during excavationusing standard monitoring devices. Fig. 1 shows the exca-vation site along with the instruments for monitoring loca-tions. Eight inclinometer casings (I-1 to I-8) were installedinside 800 mm thick diaphragm wall to monitor the lateraldisplacements of diaphragm wall. All the inclinometer cas-ings were installed to a depth of 3 m below diaphragmwalltoe level, so that the toe movement of the diaphragm wallcan be measured.

Six water standpipes (P-1 to P-6) were also installedoutside the excavation area to monitor the uctuationof groundwater table during the entire excavationprocess.

2.1. Mesh and boundary conditions

The excavation geometry of the case history was carriedout for a plan area of approximately 30 m by 35 m. The

are listed in Table 1.Due to friable nature of the material, the recovery of

suitable undisturbed samples for laboratory testing is oftenlimited or low-quality. The effective stress strength (c0 and/

0) parameters as shown in Table 1 have been selected as

representative effective strength parameters, as reportedby Nithiaraj et al. [25] and Wong and Muhinder [26]. Anunloadingreloading Poissons ratio, vur of 0.2 and stress-dependent Yongs modulus parameter, m of 0.5 wereadopted for the residual soil layers. In the numericalback-analyses, only the triaxial secant modulus was opti-mized while other parameters remain unchanged. Theoedometer stiffness and unloadingreloading stiffnessparameters were set as suggested by Tan et al. [15].

The diaphragm wall was modeled with 6-noded isotro-pic linear elastic plate element. The Youngs modulus ofthe diaphragm wall was determined using the correlationE = 4.7 106(fcu)0.5 kN/m2, as recommended by Ou [27],where fcu is the 28 days uniaxial compressive strength of

K.H. Law et al. /Measurement 47 (2014) 645650 647ratio of excavation length to width is about 1.2 suggestingthat a plane strain 2D model may not be appropriate due tocorner effect of the excavation [68]. The numerical backanalyses of this case history have therefore been conductedby 3D nite element analyses using the program PLAXIS3D FOUNDATION Version 2.2.

Fig. 2 shows the nite element mesh adopted in thenumerical back analyses. The side boundaries of the meshare prevented from movement in the horizontal plane butare free to move vertically and the bottom boundary of themesh is fully xed. The 0.8 m thick diaphragm wall wasmodeled with isotropic linear elastic plate elements. Soilelements are 15-node wedge elements which are createdby projection of 2D, 6-node triangular elements. Support-ing steel struts were modeled with isotropic linear elasticbeam elements. The diaphragm wall was assumed to bewished-in-place. The installation effect of diaphragmwall was not considered.

Fig. 1. Instrumentation monitoring layout.2.2. Constitutive models and selection of parameters

The Hardening Soil (HS) model as implemented in niteelement program PLAXIS 3D FOUNDATION Version 2.2 wasused to study the Youngs modulus of the residual soil ofKenny Hill Formation. It has been successfully used forthe modeling and analysis of retaining wall structures inweathered residual soil of Kenny Hill Formation as demon-strated by Liew and Gan [13] and Tan et al. [15]. Thestressstrain curve of HS model is represented by non-linear hyperbolic curve as proposed by Duncan and Chang[23]. In HS model, three Youngs modulus, namely triaxialsecant, oedometer and unloadingreloading Youngs mod-ulus at the reference pressure are required to be input intothe numerical model. In contrast to the MohrCoulombmodel, the three Youngs moduli of the HS model representthose at the reference pressure rather than at the in situstate.

According to Tan [24] the residual soils of Kenny HillFormation may be assumed as drained material. Hence,an effective drained analysis was adopted in the 3Dnumerical modeling. The effective stress strength and stiff-ness parameters adopted in the numerical back-analyses

100 m

50 m

100 m

S1S3

S2

S4S5

S6

Fig. 2. 3D model of excavation.

648 K.H. Law et al. /Measurement 47 (2014) 645650concrete in MPa. For this case history, the adopted concretestrength of diaphragm wall was 35 MPa. A Poissons ratioof v = 0.15 was adopted for the concrete diaphragm wall.Considering the possibility of long term cracking and creepof concrete due to bending, the stiffness of the diaphragmwall was reduced by 30% as recommended Gaba et al. [28]in CIRIA Report C580.

The steel strut was modeled using 3-noded linearelastic beam element with its stiffness determined by EA,where E is the Youngs modulus of the steel, E = 205kN/mm2 and A is the cross-sectional area of the steel strut.No preloading was applied in the 3D numerical modelingon each strut right after it was installed.

The interaction between diaphragm wall and surround-ing soil is modeled using interface elements. The reason forincluding the interface elements in the modeling of thewall-soil interaction is to limit the friction, which is mobi-lized between the soil and the wall. For this study, an inter-face reduction factor, Rinter of 0.67 was adopted to modelthe interaction between the residual soils and concretediaphragm wall.

Table 1Soil parameters.

Symbol Unit S1 S2

c0

kPa 1 5/

0() 27 31

w () 0 0

Eref50MPa 6 45

ErefoedMPa 6 45

Erefur MPa 18 135

csat kN/m3 18 19cunsat kN/m3 18 19m (-) 0.5 0.5vur (-) 0.2 0.2pref kPa 100 100

KNCo (-) 0.546 0.485

Rf (-) 0.9 0.9Rinter (-) 0.7 0.67

Note: Eref50 for S1 to S5 is taken as 1.5 N (MPa).The groundwater level outside the excavation area waslocated at about 4.5 m below ground surface with ground-water pressure in hydrostatic condition.

3. Results and discussion

Fig. 3 presents a comparison of predicted versusmeasured lateral wall deections at inclinometer I-3 forconstruction stages 14. Only I-3 is used for comparison,as this section would most probably satisfy the 2D planestrain condition. Comparison of 3D back analyses resultswith eld measured data revealed that the horizontal walldeections at each stage of excavation could be reasonablyestimated with a simple correlation between stiffnessparameters with eld standard penetration tests (SPT) Nvalue for HS model. In this case history, the correlationbetween triaxial stiffness, and oedometer stiffness, withSPT N value is found to be 1.5 N (MPa) with unloadingreloading stiffness, three times of triaxial stiffness. The pre-dicted lateral wall deections agree well for most stagesbut with slightly conservative result at 2nd stage of exca-vation. For the 1st stage of excavation, the predicted canti-lever mode deection prole and magnitude of the wallmatches well with eld measured prole and values. Forthe 2nd stage of excavation, 3D analysis predicts deep-seated (bulging) movements toward the excavation sidewith wall top movement restrained by L1 strut. In contrast,the measured deection prole still in cantilever mode.However, close scrutiny of the measured deection prolehas revealed that the bulging movement of the wall has infact started to develop. The discrepancy between the mea-sured and predicted wall deection prole could be due tothe reason that fully drained condition has yet to beachieved in the residual soil layers with the wall responsecloser to undrained behavior. As for 3rd and 4th excavationstage, the predicted deection prole and magnitude of thewall below L2 strut level matches reasonably well witheld measured data. However, large discrepancy in bothdeection prole and magnitude has been observed abovethe L2 strut level. This discrepancy may be attributed tothe effect of strut pre-loading carried out in the eld,

S3 S4 S5 S6

8 15 20 10033 35 35 350 0 0 090 180 225 500

90 180 225 500

270 540 675 1500

20 20 20 2220 20 20 220.5 0.5 0.5 0.50.2 0.2 0.2 0.2100 100 100 1000.455 0.426 0.426 0.426

0.9 0.9 0.9 0.90.67 0.67 0.67 1.0which was not modeled in the back analysis.Fig. 4 compares the 2D plane strain and 3D analysis

results based on the above established correlation. Itshould be noted that pre-loading of strut has been includedin the 2D numerical simulation. The results clearly demon-strate that geometrical or corner effect has signicant im-pact on the induced wall and ground deformations. As theexcavation depth increases the discrepancy between 2Dand 3D result getting wider, implying that as the excava-tion gets deeper relative to its length more restraint is pro-vided by the sides of excavation as well as the arching ofthe soil across the corners. The above result is consistentwith the nding of Finno et al. [29]. They showed that largedifferences between 2D and 3D responses are apparentwhen L/He ratio is less than 2, where L is the length of walland He is the total excavation depth. For this case history,the L/He is approximately 1.89.

The above results may generally be applied to caseswith similar soil conditions as that of the Kenny HillFormation.

16

18

20

22

24

26

28

300

K.H. Law et al. /Measurement 47 (2014) 645650 649Lateral Wall Deflection (mm)

16.5

18.5

20.5

22.5

24.5

26.5

28.5

30.5-5 0 5 10 15 20 25 30 35 40 45 54. Conclusions

The performance of a deep excavation in stiff residualsoils of Kenny Hill Formation has been described. Applica-tion of the HS model in this practical deep excavation prob-lem has shown that the model is suitable not only foranalyzing the case for the Kenny Hill Formation but mayalso be applied for similar soils having these types of prob-lems from a practical point of view. The case history pre-sented here shows that whilst it is important to denethe soil modulus parameters, it is equally important to takeinto consideration the geometrical or corner effect whenevaluating the performance of an excavation.

Lateral Wall Deflection (mm) 6.5

8.5

10.5

12.5

14.5

Measured Stage 1

3D Stage 1

6

8

10

12

14

Fig. 3. Comparison of predicted versus m

Lateral Wall Deflection (mm)

Lateral Wall Deflection (mm) 6.5

8.5

10.5

12.5

14.5

16.5

18.5

20.5

22.5

24.5

26.5

28.5

30.5

Measured Stage 1

3D Stage 1 (1500N)

2D Stage 1 (1500N)

6

8

10

12

14

16

18

20

22

24

26

28

30-5 0 5 10 15 20 25 30 35 40 45 50

Fig. 4. Comparison of the 2D plaLateral Wall Deflection (mm)

.5

.5

.5

.5

.5

.5

.5

.5-5 0 5 10 15 20 25 30 35 40 45 50Acknowledgements

The study was made possible by the support of thePostgraduate Research Fund (PPP), University of Malaya(Project No: PV057-2011B) and the research faculties ofthe Civil Engineering Department, University of Malaya.

References

[1] F.H. Lee, S.H. Hong, Q. Gu, P. Zhao, Application of large three-dimensional nite-element analyses to practical problems, Int. J.Geomech. 11 (6) (2011) 529539.

[2] B.K. Tan, Urban geology of Kuala Lumpur and Ipoh, Malaysia,IAEG2006, Paper number 24, The Geological Society of London, 2006.

Lateral Wall Deflection (mm) .5

.5

.5

.5

.5

Measured Stage 2

3D Stage 2

easured lateral wall deections.

Lateral Wall Deflection (mm)

Lateral Wall Deflection (mm) .5

.5

.5

.5

.5

.5

.5

.5

.5

.5

.5

.5

.5-5 0 5 10 15 20 25 30 35 40 45 50

Measured Stage 2

3D Stage 2 (1500N)

2D Stage 2 (1500N)

ne strain and 3D analysis.

[3] G. Viggiani, J.H. Atkinson, Stiffness of ne-grained soil at very smallstrains, Gotechnique 45 (2) (1995) 249265.

[4] N. Nakamura, A practical method for estimating dynamic soilstiffness on surface of multi-layered soil, Earthquake Engng. Struct.Dyn. 34 (11) (2005) 13911406.

[5] M. Lipinski, M. Wdowska, A stress history and strain dependentstiffness of over-consolidated cohesive soil, Faculty Civil Environ.Eng., Warsaw Univ. Life Sci. 43 (2) (2012) 207216, http://dx.doi.org/10.2478/v10060-011-0056-y.

[6] W. Powrie, H. Pantelidou, S.E. Stallebrass, Soil stiffness in stress pathsrelevant to diaphragm walls in clay, Gotechnique 48 (4) (1998)483494.

[7] A. Shaee, H.R. Tavakoli, M.K. Jafari, Undrained behavior ofcompacted sandclay mixtures under monotonic loading paths, J.Appl. Sci. 8 (2008) 31083118.

[8] L. Dong, Z. Ibrahim. Z. Ismail, Added advantages of using a taperedber Bragg grating sensor to determine the early age setting time forcement pastes, Measurement 46 (2013) 43134320.

[9] L. Dong, Z. Ibrahim, H.Z. Yang, Z. Ismail, Use of tapered optical bersensors in study of the hydration process of cement paste, Sens. J.,IEEE 13 (9) (2013) 34153420.

[10] L. Dong, Z.Ibrahim.B. Xu, Z. Ismail, Optimization of the geometries ofbiconical tapered ber sensors for early-age curing temperaturemonitoring of cement paste, Comput. Aided Civil Infrastruct. Eng.28 (7) (2013) 531541.

[11] T. Jackson, K. Manseld, M. Saa, T. Colman, P. Romine, Measuringsoil temperature and moisture using wireless MEMS sensors,Measurement 41 (4) (2008) 381390.

[12] D.-S. Xu, J.-H. Yin, Z.-Z. Cao, Y.-L. Wang, H.-H. Zhu, H.-F. Pei, A newexible FBG sensing beam for measuring dynamic lateraldisplacements of soil in a shaking table test, Measurement 46 (1)(2013) 200209.

[13] S.S. Liew, S.J. Gan, Back analysis and performance of semi top-downbasement excavation in sandy alluvial deposits, in: Proc. 16thSoutheast Asian Geotech. Conf., vol. 1, Kuala Lumpur, 2007, pp.833837.

[14] B.T. Soana, K.Y. Hooi, Prediction versus observedmovement of a very

[16] B. Simpson, H. Yeow, A.K. Pillai, B. Grose, Benets derived from useof 3D nite element analysis in the design of deep excavations andtunnels, in: Proc. Int. Conf. Deep Excavations, vol. 1, Singapore, 2006,pp. 15.

[17] F.H. Lee, K.Y. Yong, C.N. Quan, K.T. Chee, Effect of corners in struttedexcavations: led monitoring and case histories, J. Geotech.Geoenviron. Eng. 124 (1998) 339348.

[18] C.Y. Ou, B.Y. Shiau, Analysis of the corner effect on excavationbehaviors, Can. Geotech. J. 35 (1998) 532540.

[19] C.Y. Ou, B.Y. Shiau, I.W. Wang, Three-dimensional deformationbehavior of the Taipei National Enterprise Center (TNEC) excavationcase history, Can. Geotech. J. 37 (2000) 438448.

[20] T. Schanz, P.A. Vermeer, P.G. Bonnier, The Hardening Soil model:formulation and verication, Beyond 2000 in ComputationalGeotechnics, vol. 1, Roterdam:Balkema, 2000. pp. 281296.

[21] I. Komoo, Engineering properties of weathered rock properties inPeninsular Malaysia, in: Proc. 8th Southeast Asian Geotech. Conf.,vol. 1, Kuala Lumpur, 1985, pp. 3-813-86.

[22] J.K. Raj, A study of residual soils and the stability of their cut slopesin Peninsular Malaysia, PhD Dissertation, University of Malaya,Malaysia, 1983.

[23] J.M. Duncan, C.Y. Chang, Nonlinear analysis of stress and strain insoils, Soil Mech. Found Div. ASCE 96 (SM5) (1970) 16291653.

[24] S.A. Tan, One north station excavation in 30 m of jurong residualsoils in Singapore, Earth Retention Conf, Florida 1 (2010) 732739.

[25] R. Nithiaraj, W.H. Ting, A.S. Balasubramaniam, Strength parametersof residual soils and application to stability analysis of anchoredslopes, Geotech. Eng. 27 (1996) 5581.

[26] J. Wong, S. Muhinder, Some engineering properties of weatheredKenny Hill Formation in Kuala Lumpur, in: Proc. 12th SoutheastAsian Geotechnical Conf., vol. 1, Kuala Lumpur, 1996, pp. 179187.

[27] C.Y. Ou C, Deep excavation: theory and practice, 1st ed., Taylor &Francis, 2006.

[28] A.R. Gaba, B. Simpson, W. Powrie, D.R. Beadman, Embeddedretaining walls guidance for economic design, CERIA C580, 2003.

[29] R.J. Finno, J.T. Blackburn, J.F. Roboski, Three-dimensional effects forsupported excavations in clay, J. Geotech. Geoenviron. Eng. 133(2007) 3036.

650 K.H. Law et al. /Measurement 47 (2014) 645650deep diaphragm wall basement in Kenny Hill residual soils, in: Proc.Malaysian Geotech. Conf., vol. 1, Kuala Lumpur, 2004, pp. 359366.

[15] Y.C.Tan, S.S. Liew, S.S.Gue, Anumerical analysisof anchoreddiaphragmwalls for a deep basement in Kuala Lumpur, Malaysia, in: Proc. 14thSoutheast Asian Geotech. Conf., vol. 1, Hong Kong, 2001, pp. 16.

Determination of soil stiffness parameters at a deep excavation construction site in Kenny Hill Formation1 Introduction2 Material and methods2.1 Mesh and boundary conditions2.2 Constitutive models and selection of parameters

3 Results and discussion4 ConclusionsAcknowledgementsReferences