P 1515- Design and Contstruction of Anchored and Strutted Sheet Pile Walls iin Soft Clay.pdf

8/14/2019 P 1515- Design and Contstruction of Anchored and Strutted Sheet Pile Walls iin Soft Clay.pdf

http://slidepdf.com/reader/full/p-1515-design-and-contstruction-of-anchored-and-strutted-sheet-pile-walls 1/36

Proceedings: Second International Conference on Case Histories in Geotechnical Engineering June 1-5 1988 St. Louis Mo. Invited Paper

Design and Construction of Anchored and Strutted Sheet Pile Walls in·Soft Clay

Bengt B. BromsProfessor Nanyang Technological lnstHute Singapore

SYNOPSIS: The design and construction of anchored and strutted sheet pi le walls in soft clay are reviewed in thepaper based on experience gained mainly in Singapore during the last 10 years where mainly strutted sheet pile wallsand contiguous bored piles are used. t is important to consider in the design also the high 1ateral earth pressureson the sheet piles below the bottom of the excavation when the depth of the excavation is large compared with theshear strength of the clay. The strut loads and the maximum bending moment in the sheet piles can e considerablehigher than indicated by a conventional analysis. Different methods to increase the stabil i ty have also beeninvestigated. With jet grouting. embankment piles and excavation under water i t is possible to reduce significantlythe maxinrum bending moment, the strut loads, the settlements outside the excavated area and the heave within theexcavation.

INTRODUCI ION

The design of anchored and strutted sheet pi le walls insoft clay had to satisfy the following criteria.

o that the sheet pile wall should be stable and thefactor of safety be adequate with respect to completecollapse both during and after the construction ofthe wall (ultimate limit state)

o that the displacements and deformations of the sheetpile wall and of the support system a t working loadsshould be small so that the sheet pi le wall willfunction as intended in the design (serviceabilitylimit state) .

o that the settlements or lateral displacements caused

by the installation of the sheet piles or of the

support system (e.g. the· driving of the sheet pilesor the installation of the anchors) should be smallso that adjacent buildings or other nearby structuresare not damaged. The settlements from anunintentional lowering of the ground water level insoft clay due to e.g. pumping can be large.

The main factors affecting the behaviour of anchored orbraced excavations in soft cla;y can be classified asfollows :

o Geometry of the excavation

(depth, width, shape and excavation sequence)

o Soil and ground water conditions

(strength and deformation properties of the soil andthe ground water level)

o Properties of the sheet piles(stiffness and depth of sheet piles and the chosenconstruction method}

o Properties of the support system(type, spacing and preloading of ground anchors or ofstruts)

o Loading conditions{surcharge and traffic loads}

o Worlananship

Thus a large number of factors can affect the behaviourof both anchored and strutted sheet pile walls. Inthis paper experience with strutted and anchored sheetpi le walls primarily in Singapore has been reviewed.Limfi:ationsOf i f f e r e ~ wall and support systems areanalyzed. Methods that can be used to calculatelateral earth pressure and the stability of deepexcavations with respect to bottom heave and excessivesettlements have·been evaluated as well as methods toincrease the stability. The following review is mainly

based on experience gained in Singapore during the las t10 years where numerous deep excavations in soft clayhave been required for high r i se building, 'subwaystations and tunnels.

SOILS <X>NDITIONS IN SINGAPORE

There are extensive deposits of very soft marine clay.and organic soil with a thickness of up to 35m or morealong the coast and in the buried river valleys inSingapore. t is mainly these soils that have causeddifficult ies during the construction of both anchoredand braced sheet pile walls, e.g. large lateraldisplacements and settlements. The organic content ofthe marine clay is normally 3X to 5 . The watercontent varies usually between 65 and 100 . Theundrained shear ~ t t r e n g t h (cu) which is usually low

close to the ground surface increases approximatelylinearly with depth. Tan {1983) has reported a c/p -ratio { c u u ~ of 0.315 based on the results from f ield

vane tests. Tan {1970) and Ahmad and Peaker (1977)

have indicated somewhat lower values, 0.27 and 0.25,respectively. The effective friction angle 4> asdetermined by consolidated undrained or drainedtr iaxial tests (CD or aJ-tests) has been veryconsistant, 21 to 22 degrees. Settlement observationsand oedometer tests indicate that the clay is slightlyoverconsolidated. The overconsolidation ratio (OCR) is

1.1 to 1.5. The coefficient of consolidation when theclay is normally consolidated is typically to

2 m2/year.

1515



W LL SYSTEMS

Different wall systems can be used as i l lus t rated inFig 1 depending on the soil conditions. In Fig l a i s

shown a conventional anchored sheet p i le wall. The

lateral earth pressure on the wall is t ransferred tothe ground anchors through wale beams, normally U-, H

or I - beams.

Soldier p i le and lagging construction is shown in

Fig lb. This support method, also cal led Berliner wall

construction, is commonly used in the United States andin Europe mainly in sand, s i l t or· gravel above theground water level. The method i s not sui table in soft

clay. The soldier pi les or beams, usually H-piles orchannels, are driven or placed in predr i l led holes and

grouted. The spacing of the pi les i s normally 1.0 to2.0 m Lagging wooden boards) i s placed during theexcavation between the flanges of the soldier pi les .

t.Jale beam

' 1 - _ _ . , ~ - - ~ - e e l :Sheet f'<le

Anchorn:xl

a :51: eel ~ h f ? des.

~ ~ - ~ ~ 1 - _ . Sold.<er pile

La'Jg<ng

AncJ orrod

Precast Concretefa ne b

C :Soldier pdes and ftrecasl

eoncrel.e p.£nel:s.

Fig la

It·. i s important that the lagging is carefully wedged.

against the soil behind the boards in order to reduce

the settlements around the excavation. Also precast orcast-in-place concrete panels can be used as shown inFig lc. The deep excavations required for some of thesubway s ta t ions in Hong Kong have been s tabi l ized by

this method.

In s i l t or in f ine s nd there i s a r isk of erosion of

the soi l below the ground water level and the resultingsettlements can be large. Soldier pi les and lagging

construction should therefore be avoided in these soi lswhen the ground water level is high. The ground water

level can, however, be lowered temporarily with well

points or f i l te r wells to prevent erosion and fai lureof the excavation by bottom heave. In s t i f f to hardclays i t may be advantageous to use pairs of channels

as soldier pi les instead of steel H-piles (Fig lc .

The wale beams can then be eliminated in order to

d.. R.ad;:;

2ails

Anchor rod

Wale b e a

Pail

ShotcreleAn char reel

Re,n.forr-erneni

Abou f t o .

Shotcre-l:ei so"'"VJle<r.krcernent

Wall sys erns

1516



reduce the costs since the anchors can be placedbetween the two channels.

Rails are used as la teral support in Fig ld. Thespacing is usually 0.2 to 0.3 m This support methodis mainly used in stony or blocky soi ls above theground water level. The ra i ls are often placed inpredri l led holes when the content of stones or bouldersis high since the ra i ls cannot be driven. The rails

are often bri t t le due to the low duct i l i ty of the steel{high strength steel ) . They are diff icul t to spl ice by

welding. Therefore, bolted joints are often used. In

dry sand above the ground water level plywood boardsare sometimes placed between the ra i ls to contain thesand. In s t i f f medium to s t i f f clays or in s i l ty

soils, the soil is normally protected by shotcrete asi l lustrated in Fig le. The reinforced shotcrete archestransfer the lateral pressure from the soi l to thera i l s The thiclmess of the shotcrete is normally

about 50 mm

Also bored piles can be used as lateral support in deepexcavations as illustrated in Fig l f In soft clay thepiles should overlap while in medium to s t i f f clayoverlapping is not required. The distance between thepiles can be relatively large. The unprotected areabetween the piles is often covered by shotcrete.Overlapping bored piles, so-called contiguous bored

piles, are common in Singapore also in soft clay asfoundation for high r ise buildings and as lateral

support.

ANCl ORS AND STRUTS

Different support systems can be used for a deep

excavation in soft clay or s i l t as illustrated in Fig 2depending on the soil and ground water conditions and

on the size {width, length and depth) of theexcavation.

The choice of support ·system depends mainly on thecosts, on restr ictions a t the worksite, on availableequipment in the area and on the experience of · theconsultant or of the contractor. For example adjacentbuildings may be damaged by excessive settlements i f acantilever sheet pile wall is used to support arelat ively deep excavation. Also water mains, sewer

lines and heating ducts can be damaged by the resul t inglarge settlements ~ lateral displacements. Excessive

settlements can also be caused by the instal la t ion ofthe anchors as well as by the driving of piles insidethe excavation. Struts may therefore, be chosen

instead of ground anchors to reduce the risks. Thesettlements can be reduced further by preloading thestruts or the ground anchors. If the anchors are le f t

permanently in the ground they may interfere withfuture construction such as the driving of sheet piles.However different anchor systems have been developed

during the last few years which can be removed after

use and where the settlements caused by the

installation of the ground anchors will be small.

The lateral earth pressure behind a. cantilever sheetpile wall {Fig 2a) is resisted by the passive earthpressure below the bottom of the excavation while for

an anchored or strutted sheet pile wall the lateralearth pressure i s resisted by ground anchors or by

st ruts as shown in Fig 2b and 2c, respectively. Groundanchors or st ruts are normally required in soft claywhen the depth of the excavation exceeds 2 to 3 m

In a large and wide excavation the length of the strutswill be large i f the s tru ts are horizontal. They hadto be braced to prevent buck ing as can be seen inFig 3. The st ruts will , however interfere with the

1517

Anchor

;...,.___.Dd' el: con

I D ~ l l . e c t - o l ?\_ /I

:· :

······

IF ; == Ji l ·.· · · ·

c) ;5/:ru.H:-ed s h e ~ : l : ?-de. walt

Fig 2 Support systems

Ground anchor

work in the excavation and reduce the efficiency.Horizontal bracing is common in Singapore.

The anchors or the s tru ts can either be horizontal orinclined. In narrow deep cuts horizontal s tru ts areused while in large and wide excavations the ·struts areoften inclined. The inclined s t ru ts are generallysupported a t the bottom of the excavation by a concreteslab or by separate individual concrete footings. I tshould be observed that the inclined s tru ts or anchors

will cause an axial force in the sheet pi les which

affects the s tabi l i ty of the wall.

A number of different ground anchor systems using bars,wires or strands have been developed during the las t 20

years as described by e.g. Hanna (1982). A relativelyhigh pressure is often used in sand or s i l t for thegrouting of the tendons in order to enlarge the hole so

that a bulb is formed around the tendons within theg r o u t ~ section, the fixed anchor length. Thetube-a-manchette method can be used especially in sand,

gravel and rock to control the grouting. The bore holeCl3Jl be enlarged mechanically in s t i f f clay, using aspecial cutt ing device in order to increase the tensiler«:sistance of the ground anchors. Also, H-beams have

been us.ed as ground anchors in Sweden in very softclay. The pull-out resistance is high due to the largesurface area.



Fig 3 Braced sheet pile wall

Rods bars) are normally used when the load in theanchon i s relatively low, less than about 400 kN,while cables wires or strands) are uti l ised as tendonswhen the load exceeds about 400 kN. The anchor rod orwires are often prestressed in order to reduce thehorizontal d i a p l n t a and the deforme.tions of thewall and thus the settlements during the excavation.Ground anchors are mainly used for temporary structuresbecause of th risk of corrosion of the tendons or ofthe anchor rods. The corrosion can be reduced forper.enent anchors by enclosing the tendons and by

introducing a fluid between the covering and thetendons. Also cathodic protection can be used.

A recent e v e l o ~ n t is expander bodies. This new typeof anchor consists in principle of a folded thin steel

sheet, which can inflated in-situ thrQU h the inject ionof c . e n t arout as shown in Fig 4 Broms, 1987). Theexpander ·bodies can either be driven into the soil orplaced in predrU led cued holes depending on the sotl

::

)mny a Pfactm rrlil onchor

··

Expender bodies

conditions. The volume of the grout required for theexpansion and the pressure should be measured in orderto check the ul ttmate resistance. The me.ximum groutpressure in grenular sotl ta 3 to 4 IIPa. The mainadventage with this new type of ground anchor ta thatthe size and the shape of the anchors are controlled.

In Sweden. the L1n48 and the JB methods where thecasing i s provided with a sacrif ic ia l dr i l l ing bi t areused for the dri l l ing of the boreholes. Also differenteccentric dri l l tng methods have been developed e.g.Odex, Exler and Alvik to facili tate the ins tal lat ion ofthe casing and to reduce the costs. An addi ttonalmethod ts the In-Situ Anchoring Method where the anchorrods are used as dr i l l rods during the dri l l ing of the

boreholes. castng is not required. However. the

allowable load 1s relatively low for this type ofanchor and the method is therefore relativelyexpensive.

The chosen method of ins tal lat ion of the struts and ofthe anchors affects both the total lateral earthpressure as well as the earth pressure distribution.When relatively st i f f struts are used, the lateralearth pressure can be considerably higher than theactive Rankine earth pressure particularly close to theground surface while a t the toe the lateral earthpressure can be lower than the active Rankine earthpressure. The reason for this difference is therelatively small lateral deflection of the sheet pilewall close to the ground surface during theconstruction since the struts are normally wedged andpre loaded.

1518

A certain small la teral deflection i s required to110biltze the shear strength of the soil behind the walland to reduce the lateral earth pressure. In densesand a la teral displacements of O.OSX of the depth ofthe excavation s normally sufficient to reduce thelateral earth pressure to the active Rankine earthpressure. When the sand is loose the required lateraldeflection i s approximately 0.2X of the depth. A ~ c hlarger deforaation i s required in soft clay.



DFSIGN PRINCIPLFS

The following four steps are normally followed in the

design of a sheet pi le wall :

o Evaluation of the magnitude and the dis t r ibut ionof the lateral ear th pressure behind the sheet

pi le wall

o Calculation of the required penetration depth

o Determination of the moment distr ibution in thesheet piles

o Estimation of the axial force in the ground

anchors or in the st ruts

Extensive investigations are normally required in the

f ie ld and in the laboratory to determine the depth and

the thickness of the diflerent soil s t r a t a and of theunderlying rock as well as their strength and

deformation propert ies as indicated, for example, inthe Brit ish Code of Practice CP2001). Penetrationtests are mainly used in cohesionless soils sand and

gravel) in order to estimate the re la t ive density, theangle of internal fr ic t ion and unit weight. Conepenetration tests (CPT) and weight soundings WST) are

preferred belore the standard penetration test (SPJ')because of the uncertainties connected with this

testing method. However, representative samples areobtained with SPJ' so that the soi l can be classified.The size and the shape of the soil part icle areimportant as well as the gradation since theseparameters affect the frict ion angle of the soi l .

The driving of the sheet piles are affected by stonesand boulders in the soi l . The stone and boulder

content of the dif ferent st ra ta and the dif f icul t ies

that ~ y be encountered during the driving of the sheetpiles can normally be evalauted from weight (WSf) orram soundings (DP) or from cone penetrat ion tests

(CPT) . Driving tests with ful l s ize sheet piles may be

required lor large jobs. Stress wave measurements can

be helpful to determine the driving resistance and the

elliciency of the driving. I t is also important todetermine the location and possible variat ions o£ theground water level.

For anchored or strut ted walls the depth of an y so l t

clay or ai l t layers below the bottom of the excavation

and the variation of the thickness of these layers areparticularly important since the stabi l i ty of the walldepends to a large extent on the passive earth pressurethat can develop a t the toe of the sheet pi le wall.

The depth to a l irm layer below the bottom of theexcavation can usually be determined by penetrat iontests . Also seismic methods can be used.

Penetration tests especial ly cone penetrat ion tests

(CPT) and wei ht soundings WST) are useful in cohesive

soi l s in order to determine the sequence and the

thickness or the different layers. The undrained shearstrength of the clay i s normally evaluated by f ie ld

vane tests. Undisturbed samples obtained preferably by

a thin-walled piston sampler are usually required when

the shear strength of the soil i s evaluated in thelaboratory by, for example; unconfined compression,

fall-cone or laboratory vane tests. Undrained t r iaxial

teats are often used to determine the undrained shear

strength of s t i f f f issured clay. The water content,the l iquid and plast ic l imits or the clay should alaobe - . u r e d . Drained t r iaxial or direct shear tests

are required for heavily overconsolidated clays in

order to evaluate +d o r + . The difference between the

1519

two angles is usually only a few degrees. An estimateof the long term ground water level and the changes

that may occur with time is also necessary . Percussion

dr i l l ing and coring are normally required in rock. The

quality of the rock can often be estimated from the

dr i l l ing rate . The compressive and tensi le strengthscan be determined by unconfined compression and or

point load tests.

The condi t iona of the adjacent structures should alsobe investigated dilapidation survey). The type of

foundation spread footings, raf t . or pi les) i s

important since i t can affec t the choice or support

system .

C a u ~ t ol .fa/lure IOilurt m t h o m ~ m

/Odurt o riruloron 'I

.

\. .a.

;:otlure ofmiddlt-slrul o/ anchor

:;j

b.

Oilurt of /()l.«r rt n r l or anchor

.

.

c.

11oment Capt;lCI:fy d ~n s u l k c i ~ n l ol lhetop.

d .

o m ~ n l C O j X l ~ td rrisullictirrl -the

.

c ~ r r f r e.

e.

Penefrofton depth rnd moment co-

.

p::zdf L _o : t n ~ u l f i -IP Ctenf

• ,Piosftc h,n1e

> - lOtlure r slrut oranC. IJo,n

Fig 5 Failure mechanisms



In the design o£ anchored or braced sheet p i le walls t

is preferable to use character is t ic strengths and

characteris tic loads which takes into account the

uncerta int ies connected with the determination o£ theshear strength o£ the soi l or o£ the rock and the

loading conditions. A design s t rength rd = fii Tm i s

used in the calculat ion of the la teral ear th pressureswhere fk i s the character is t ic strength of the so i l or

the rock nd ''m i s a par t ia l · £actor of safety la rge r

than 1.0 . External loads are treated in a similar way.

A design load Fd = Fk T£ where Tf i s a par t ia l

coeffic ient and Fk is the character is t ic load, i s then

used in the calculat ions of the la tera l ear th

pressures. The probabi l i ty that the character is t ic

load will be exceeded in the f ie ld should not be

greater than 5%. The fa:ilure mode or fai lure mechanism

and the deformation required to mobilize the peak

resistance of the so i l should a lso be considered whenthe required par t ia l £actor o£ safety is evaluated as

well as cracks and f issures . A s ta t i s t i ca l analysis o£

tqe tes t resul ts may in some cases be helpful .

A global factor of safety F s is often used in the

design of both anchored and s t ru t ted sheet p i le walls .A value o£ 1.5 on Fs i s often chosen for cleys wi th

respect to the required penetrat ion depth in order to

prevent fai lure by ro tat ion of the sheet p i le wallabout the anchor level . For cohesionless soi l s aglobal factor of sa fe ty of 2.0 i s normally required.

LATERAL EARTII PRESSURE

Possible failure mechanisms of anchored or s t ru t t ed

sheet pi le walls supported a t several levels are shownin Fig 5 . Failure may occur when the anchors or st ru ts

rupture or buckle Figs Sa, 5b or 5c} or when the

moment capaci ty of the wall bas been exceeded 5d, 5eor 5f . The deformations of the sheet pi les during the

excavation affect both the magnitude and thedistr ibution or the la tera l earth pressure behind thewall. The la teral ear th pressure can be considerably

lower than the active Rankine earth pressure between

the support levels due to arching when the la tera ldeflect ions of the wall are large. At the s tru t or

· anchor levels the la tera l earth pressure can beconsiderably higher that the active Rankine ear th

pressure. as pointed out by e.g. Rowe 1957}.

The earth pressure dis t r ibut ion for temporary

structures in clay i s shown in Fig 6. This

distr ibution is in principle the same as that proposed

by Terzaghi and Peck 1967}. A trapezoidal ear th

pressure dis t r ibut ion can be used in the calculation of

the force in the anchors and in the st ru ts as well as

of the required penetrat ion depth. The la tera l ear th

pressure i s assumed to be [pH - 4c ] above the bottom

of the excavation when the deptn uof the excavation

exceeds 4cu/p nd 0.35pH when the depth i s less than

4cu/p.

Below. the bottom of the excavation the net pressure ,the di fference in the la teral ear th pressure on bo th

sides of the wall i s (pH - Ncbcu } where Ncb i s the

bearing capaci ty factor o£ the soi l with respect tobottom heave. This factor depends on the dimensions ofthe excavation depth, width and length). The netpressure wil l be negative and contribute to the

stabi l i ty when pH < Ncbcu and positive when pH >Ncbcu.

1520

Fig 6

b

Design of anchored and braced sheet p i l e

walls in sof t clay

I t i s proposed to use the net pressure below the bottom

of the excavation a t the design instead of thecoeffic ient m as proposed by Terzaghi and Peck ( 1967}

to take into accoWlt the increase o£ the s tru t oranchor loads when the shear strength o£ the clay i s low

below the bottom of the excavation compared with· the

to tal overburden pressure . A similar calculat ionmethod has been proposed by Aas 1984) and by Karlsrud

1986).

t bas been assumed in the calculat ion of the net earthpressure that the adhesion ca) along the sheet pi les

corresponds to the Wldrained shear strength of the clay

cu). The bearing capacity factor Ncb wil l be reduced

when ca < cu. For an inf in i te ly .long excavation Ncb =

4cu when ca = 0, a reduct ion by about 30%.

A relat ively large la tera l deflect ion i s required todevelop the passive la tera l earth pressure in front ofthe wall and thus the net pressure when the shear

strength of the clay i s low. Adjacent bui ldings can be

damaged by the resul t ing large settlements. I t maytherefore be advisable for sof t clay to use a lower

la tera l earth pressure than the net pressure in thecalculat ion of the required penetrat ion depth.

The to tal la tera l ear th pressure when the depth of the

excavation i s less than the c r i t i ca l depth 4cu/p

corresponds approximately to the la tera l ear th pressurea t rest (K

0.7 to 0.8}. This earth pressure may be

used in the design of permanent structures in sof t

clay. The preload in the anchors and in the s t ru t s

should preferably be adjusted periodically especiallyin sof t clay to compensate for creep

andconsolidat ion

of the so i l behind the wall .

In a heavily overconsolidated clay i t i s important that

the la tera l ear th pressure is s u f f i i ~ t l y high closeto the ground surface to el iminate any tens i le s t resses

in the soi l nd to prevent cracking of the clay.Vert ical tens i le cracks may reduce the shear s t rength

of the clay and increase the la tera l pressure when the

cracks are f i l l ed with water af ter a heavy rainstorm.



BO'ITOM HEAVE

In the design of a strutted or anchored sheet pi lewalls in soft clay 1 failure by bottom heave had to beconsidered as illustrated in Fig 7. The part of thesheet piles that extends below the bottom of theexcavation in Fig 7a must resist a lateral earthpressure that depends on the depth of thll excavationand on the undrained shear strength of the clay.

t is proposed to use the net earth pressure as shown

in Fig 6 for the part of the sheet pi le wall thatextends below the lowest strut level. This part of thewall functions as a cantilever which carries the loadcaused by the lateral earth pressure behind the sheet

piles. This load is partly resisted by the passiveearth pressure between the two sheet pile walls.

The passive earth pressure is affected by the distanceB) between the two walls. f this distance is less

than approximately the penetration depth D) then thepassive earth pressure a t the bottom of the sheet pilescan be evaluated from the relationship

CT = 2 c Dp 2 c DIBp u u

{1)

When the distance B between the sheet piles exceeds thepenetration depth D B>D) i t is proposed to evaluatethe passive earth pressure from the followingrelationship Janbu, 1972)

2)

where = ca cu. t should be noticed that the passive

undrained shear strength as determined from tr iaxialextension tests should be used in the calculations.This shear strength may be lower than that determinedby e.g. field vane tests.

A load factor equal to 1.0 has been used with respectto the. unit weight of the soil and the water. In thesoft clay below the bottom of the excavation the net

lateral pressure is [ ~ f q pH1 - pw w - N c b c u / ~ m ] where

Ncb is the stabil i ty factor with respect to bottom

heave Fig 8). In the intermediate sand layer the netpressure will be positive and contribute to thestabil i ty of the wall. The lateral earth pressure willto a large extent depend on the pore water pressure inthis layer.

A 2.0 m thick unreinforced concrete slab will be castbelow water a t the bottom after excavation down to therequired depth to prevent heaving when the water levelin the excavation is lowered.

f the adhesion ca) along the sheet piles corresponds

to the undrained shear strength cu) of the clay, then

CT =2.83 c + Dpp u 3)

When the penetration depth is large compared with thewidth B, the passive pressure between the two rows willnormally be larger than the outside earth pressure andthe sheet piles will be supported at least partly by

the passive earth pressure between the two walls.

The uplif t pressure a t the bottom of the sheet pi lewall depends on the depth of the excavation H, thepenetration depth D. the undrained shear strength ofthe clay as well as on the shape of the excavation

1521

....:JL.. _,

17

1 J.

~ ~ £ + ~8~ t t t H i-

I

U J I I ~ c uc{( "" 5 0 tlb

a Eorlh -"':f5ut' didn"bulion h. Boll-om heare

Fig 7 Design of braced sheet J?ile walls in softclay

B/L). The uplif t pressure at the bottom of the sheetpiles Fig 7b) can be evaluated from the equation

{4)

where Ncb is a stabi l i ty fe,ctor {Fig 8) which can be

determined from the following relat.ionships Bjerrumand Eide, 1956).

Ncb =5 {1 + 0.2 HIB) {1 0.2 B/L)

when HIB 2.5 and from

Ncb = 7.5 {1 + 0.2 B/L)

when HIB > 2.5.

5)

{6)

This upl i f t pressure had to be resisted by the weightof the soil below·the bottom of the excavation and bythe adhesion ca of the clay along the sheet piles.

7)

In the calculation of the required penetration depth isis advantageous to use load factors ( ~ f ) and partial

safety factors_ c ~ m > as mentioned previously.

The proposed design method is i l lustrated in Fig 9a for·

a braced sheet pi le wall. The sheet piles have been.driven through sof t marine clay upper Marine Clay, M)into an underlying intermediate layer with sand {F1).Below this intermediate layer is a second layer withsoft marine clay Lower Marine Clay·, M). The shearstrength of the clay is low.

t is anticipated that the excavation of the f i l l andthe soft clay will be carried out below water in orderto prevent failure of the excavation by bottom heave



1tab/illy fador { 6/0 r 1 >/ v,

8 . . ..

6 ::\: ---c B;i -o

4 ty; ·/ o

(Circa/or or:yuorJ

I-/

0 :-_..___...___.......__ _J

0 I Z 3 -¥j

Pofto lf ./J

;o /ure b. boffom ht CIYe

a l f . e ~ B;errum I i d ~ 95f}

Fig 8 Stabil i ty factor Ncb

due to the very low shear strength of the clay. Thewater level in the excavations will be kept a t or abovethe ground level in order to increase the stabili ty ofthe excavation. Bored piles are used to support thebottom slab. The piles will be installed before thes tar t of the excavation and provided with a permanent.casing to prevent necking of the concrete during thecasting because of the low shear strength of the ~ l a y . ·

The earth pressure-distribution when the excavation has

reached the maximum depth is shown in Fig 9a. Thelateral earth pressure above the bottom of the

excavation c o r r e s p o n d ~ to [ ~ f q + pH1 - 4 c u / ~ m J where

is a load factor and is a partial factor of ·safety.

The upl i f t pressure on the concrete slab will vary. Ahigher upl i f t pressure q1) is expected on the slab

next to the two sheet pi le walls compared with thatq3) at the center of the slab as shown in Fig 9b and

Fig 9c, respectively.

The upl i f t pressure q1 in Fig 9b depends on the total

overburden pressure ~ f q + pH1) outside the sheet pi le

wall at the level of the concrete slab, on the lateralresistance of the sheet piles on the shear strength

~ ~ r .·H,·H tt J

Fig 9a Proposed design method for a strutted sheetpi le wall i·n soft clay

1522



r

Fig 9b Bottom heave (upper cla:y layer)

of the cla:y cul and on the s tabi l i ty factor 'Ncb This

upl i f t pressure will act on a s tr ip with a width thatcorresponds to the depth of the clay layer below theconcrete slab.

The s tabi l i ty niDDber for the excavation (DIL =0.58) i s

5.9 when the excavation i s long compared with the width

{BIL 0 as can be seen from Fig 7. However, a

relat ively large deformation will be required tomobilize the average· shear strength of the cla:y. A

partial factor of safety of about 1.4 is required tol imit the maximum wall movement to 1 of the excavationdepth {Mana and Clough, 1981). .

The uplif t pressure within the center par t . of theexcavation can be estimated as shown in Fig 96. Thisupl i f t pressure q3 will be lower than that next to the

two sheet pile walls (q1) because of the relatively

high shear strength or the lower marine clay (cu2).

The overburden pressure a t the bottom of the fluvialmaterial Fl depends on the average uni t weight of thesoi l above th is layer.

The confining pressure q4 below the bottom of the

intermediate layer {Fl) a t the centre of the excavationcan be estimated from the equation.

{8)

where lQs i s the total skin fr iction resis tance per

unit length along the sheet piles and the piles in the

marine clay and in the F1 material {fs l and

1523

r

etay l1)

Fig 9c Bottom heave (lower clay layer)

respectively) and B is the total width of the

excavation. The adhesion {ca) along the sheet pi les

and the pi les in the sof t clay is estimated to O.Scu'.

where cu is the undrained shear strength as determined

by e.g. f ie ld vane tes ts . I t i s suggested thai: the

unit skin f r ic t ion resis tance in the sand {Fl) can be

taken as 1 .of qc' where qc is the cone resis tance as

determined by cone penetration tests (CPT). I t hasthus been assumed that the total skin f r ic t ion

resis tance along the pi les and the sheet pi les can bedistributed uniformly over the total width of theexcavation.

SHEET PilE WALLS SUPPORTED BY INO..INED ANQ.IORS

An anchored sheet p i le walls may fa i l when the ver t icalbearing capacity of the sheet pi les is exceeded asi l lust ra ted in Fig 10 in the case the anchors areinclined. The inclined anchors produce a ver t ical

force _in the sheet piles which may cause the sheetpiles to se t t le i the embedment depth is notsufficient . A settlement {6v) will also cause the wall

to move outwards («\) a distance 6v tan a where a is

the incl inat ion of the anchor rods or of the cables a tthe level of the anchor (Fig 10). The incl inat ion ofsoi l anchors in so i l is often 20 degrees while for rockanchors the incl inat ion is normally 45 degrees. Theinclination can be increased in order to reduce thelength of the anchor rods or of the cables and thus thecost. The ver t ical component of the anchor force alongthe sheet pi les i s therefore, .often higher when thesheet pi les have been d;riven into :rock compared with



a. Ru ure m e h o m ~ m

;:5y _

v :5 :5t.i?

h. fOrce j22ly f-or;.

Fig 10 Vert ical stabi l i ty of sheet pi le wall withincl ined anchors

o Anchorecf ~ h lf- lt wall

b /}raced :5hel "t

;Q It? wall

so i l anchors because of th e difference in inc l ina t ionof the tendons. The sheet pi les can general ly be

driven to a higher resistance when competent rock i s

located close to the bottom of the excavation and rockanchors are used. I t i s then re la t ive ly easy to re s i s t

the high vert ical force in the sheet pi les

When the depth to rock or to a layer with a high

bearing capaci ty is relat ively -large and soi l anchors

had to be used then t i s dif f icul t to re s i s t the

vert ical component of the anchor force by adhesion or

by f r i c t ion along the sheet piles I t may then be moreeconomical to reduce the inclination of the anchors and

to increase the length of the anchor rods or of the

cables. Then the length of the sheet pi les c n be

reduced because of the reduced axia l force.

Figs l l a and l lb i l lu s t ra te the forces act ing. on abraced and anchored sheet p i l e walls in clay.respect ively. The normal force Nand the shear force TT is proport ional to the active undrained shear

strength of the so i l cu) act along the assumed fa i lure

plane. The weight W) of the s l id ing soi l wedge is

approximately the same for the two cases . The force

Ca) along the sheet pi les depends on the adhesion {ca)

between the sheet pi les and the clay below the bottom

of the excavation. The inclination and the magnitude

of the force R) in the anchors or in the s t rut s wi l l

however, be di ffe rent

I t c n be seen from the two force diagrams in Fig 11

that both the normal force N on the fai lure plane and

the passive ear th pressure force P which are requiredp

for equilibrium wil l be · larger for an anchored sheet

pi le wall when the anchors are incl ined than for abraced or a st rut ted wall when the st ruts are

horizontal . Thus a larger penetrat ion depth and ahigher passive earth pressure will be required for an

anchored wall where the tendons are inclined comparedwith a braced wall .

The s t ab i l i t y of an anchored sheet p i l e wall can be

expressed by the s t ab i l i t y factor Ncb defined by the

Fig Stabi l i ty of anchored and braced sheet p i l ewalls

1524



equation (pHcr ~ f q =Ncb c u / ~ m where (pHcr ~ f q i s

the total overburden pressure a t the bottom of the

excavation Her is the cr i t ical depth and cu is the

undrained character is t ic shear strength of the clay.The total overburden pressure depends on the c r i t i ca l

depth of the excavation Her ( the maximum depth when the

excavation i s s t i l l stable) . the unit weight of thesoil p and on the surcharge load q.

The stabi l i ty factor Ncb as shown in Fig 12 is a

function of the inclination of the anchors (a), thepenetration depth D) of the sheet pi les below the

bottom of the excavation and the adhesion (ca) between

the sheet pi les and the clay. At = 1.0 the adhesionp

corresponds to the undrained shear strength of the soi l

cu. At =0 the adhesion is equal to zero. t can

be seen from Fig 12 that the stabi l i ty factor Ncb

5tabtk- ylaclor ~ h6 o

5.o

4 o

3 0

0 I 2 3

Fig 12 Stabi l i ty Factor Ncb

increases with increasing value on and with

increasing force R in the anchors unt i l a c r i t i ca l

value has been reached. f this c r i t i ca l value i s

exceeded then Ncb will decrease.

In order to simplify the calculat ions Sahlstrom and

St i l le (1979) have proposed for sof t normally

consolidated clay that the stabi l i ty factor Ncb should

be taken as 5.1 when the sheet pi les are driven to ahard stratum so that the end bearing capacity of thesheet pi les will be sufficient to resis t the axialforce caused by the inclined anchors. In the case the

525

. :. · : ·jI I

I /iJ ? /

/

a /0;/ure mtchan/517

N

Fig 13 Total s tab i l i ty of an anchored sheet p i l e

wall

sheet pi les have not been driven to refusal in a hard

layer and the ver t ical s tabi l i ty of .the wall i s low

then a value on Ncb of 4.1 should be used in the

calculat ions.

The s t ab i l i t y may be reduced especial ly in s i l ty clays

when pi les have been driven close to an exis ting sheetpi le wall due to the remoulding of the soi l and the

resulting increase of the pore water pressures thattake place during the driving. In th i s case a value

equal to 3.6 on Ncb c n be used.

In most cases fai lure takes place in the undis turbed

soi l between the flanges = 1.0) of the sheet pi lesp

since the perimeter area i s large. Usually a layer ofclay will cling to the surface and come up together

with the sheet pi les when they are pulled.

The length of the anchors should be sufficient so tha tthe s tabi l i ty of the sheet pi le wall wil l be adequate

with respect to a deep-seated failure. In Fig 13 is

shown the forces act ing on an anchored sheet p i l e wallin a cohesionless soi l and the corresponding force

diagram. The rear face of the indicated sliding wedge

had to res is t the la teral earth pressure Pa. The

required passive earth pressure Pp req a t equi l ibrium

can be determined as shown in Fig 13 (Broms, 1968)

which is a modification of the Kranz method which is



widely used in Germany and Austria. I t has beenassumed in the analysis that the critical failuresurface is located a/2) from the end of the anchors,where a is the spacing of the anchors. I t has thusbeen assumed that the inclination of the failuresurface behind the anchors is 45° + 1/2 ~ · . The mainadvantage with the proposed calculation method is i ts

simplicity.

I t is also necesary to check the .stability of the wedgelocated above the fixed anchor length as i l lustrated ,in

Fig 14. The failure surface has been assumed to extenda distance a/2) from the end of the anchor block asshown. The passive resistance of the soil in front ofthe sliding soil wedge should be sufficient to resis tthe lateral displacement of the wedge. I t is proposedto use partial safety factors and load f,ru:tors in the

calculations.

a fcll1tlrt mechom m

w

,1 I

n /~ II=< ' I

I I· ; i I

Fig 14 Stability of anchor block

STRENGTH OF NCHORS

The design of ground anchors has been reviewed by

Littlejohn 1970). The method that can be used tocalculate the tensile resistance of soi l anchors is

illustrated in Fig 15. The ultimate tensile resistance

~ t depends on the fr iction resistance Qskin along the

grouted part of the anchor and on the end resistanceQend as expressed by the relationship

9)

. : •,

• .

Fig 15 Tensile resistance of ground anchors

The displacement required to develop the maximum skinfriction is small, a few mm, compared with the relative

large displacement which is required to mobilize theend resistance.

In cohesionless soils sand and gravel) the pull-outresistance sa) depends on the effective overburden

Pressure a and on the friction angle .P betweenvo a

grouted part of the anchors and the soil as expressed

by the equation

10)

The fr iction angle .P is normally assumed to correspond.a

to the angle of internal friction of the soil .P or .pd.

The coefficient K depends mainly on the . relativedensity of the soi l . This coefficient can for dense,

coarse and wellgraded sand or gravel be as high as 2 to3 due to the dilatancy of the soil . In loose fine sandand s i l t the coefficient K can be as low as 0.5. Theassumed value on K should be verif ied by load tests.

The tensile resistance can also be estimated from thegrout pressure used during the installation of theanchors, from the grout pressure required for theexpansion of the expander bodies or from thepenetration resistance as determined by e.g. conepenetration tests (CPT), standard penetration tests(SPT) or weight soundings {Wsr .

I t is proposed to use the equations suggested by

Baquelin e t al (1978) for bored piles to estimate thepull-out resistance from the maximum grout pressure

p . The tensile resistance of the anchorsgrout

increases generally with increasing grout pressure

especially in hard rock and in dense s nd and gravel.The capacity of the anchors wi 11 also increase withincreasing length of the grouted zone, the fixed anchorlength. In s nd and gravel there is , however, amaximum effective length. If this effective length is

exceeded then there is no further increase of theanchor force. The cri t ical length is about 6 m forsand and gravel. Cyclic loading will , however, reducethis length. The fixed anchor length is usually 3 to6 m.

1526



According to Baguelin e t al {1978) the net baseresistance of a bored pi le qend can be evaluated from

the limit pressure p2 determined from pressuremeter

tests

(11)

where p0 i s the ini t ial total horizontal pressure inthe grounct a t the base of the pile and k is acoefficient that depends on the embedment length and on

the magnitude of the limit pressure.

I t is expected, however, that the limit pressure willcorrespond to the maximum grout pressure.

Pe = Pgrout z Pgrout <12)

where p t is the grout pressure a t the groundgrou

surface, Pgrout is the unit weight of the grout and z

is the depth.

For the case the tensile resistance corresponds to 70of the ultimate bearing capacity of a bored pi le thenthe end resistance of the anchors can be calculatedfrom the equation

Qend = 0 ·7 k Pgrout Aend (13)

where k is a coefficient that depends on the embedmentlength and on the magnitude of the limit pressure and

Aend is the cross-sectional area.

The unit skin friction resistance fs of a pile in sand

or gravel will normally be 0.5% to 2 of the pointresistance {Meyerhof, 1956). The skin friction willgenerally increase with decreasing particle size andincreasing cone resistance. It is suggested for sandand gravel that the skin friction resistance should betaken as 1 of the unit end resistance. For s i l t 2 is

proposed.

The total skin friction resistance Qskin of the

expander bodies will be 12 of the total end resistancefor sand and gravel and 24 for s i l t . Then for sand

and gravel

o 1t = 0.78 k p t A dU grou en

(14)

where k iS a bearing capaicty factor which depends onthe embedment depth. For s i l t

= 1. 24 Qend = 0 ·86 k Pgrout Aend (15)

The ultimate pull-out resistance of the expander bodiesas determined by Equs {14) and {15} has been plotted inFig 16 as a function of the maximum grout pressure. I tcan be seen that the tensile resistance increasesrapidly with increasing grout pressure. I t should ·beobserved that the depth of the exp Ulder bodies shouldbe a t

leasteight times the diameter. Otherwise the

resistance will be reduced.

The tensile resistance can also be calculated from thepenetration resistance of different penetration testssuch as cone penetration tests (CPT) standardpenetration tests {SPI') and weight soundings (Wsr). Acomparison between the different penetration tests is

shown in Table I for cohesionless soils {si l t sand and

gravel). For example, a standard penetrationresistance (N30 ) of 30 blows/0.30 m in a medium sand

Ten:>de rest stonce, QuU ; MN

6.or------------------------------------5 <J

4 o

3 o

:Z.o

i oO.e

O G

o s

:·o 4

0 3

O.i-

0 1

8

6

O os

. ·

· · . : · ......•... :· ... :

>· ·· ): Q ~ u

_:. - ~ - - - .//

//

/

/

0.01.;. ------- t - - - - . .1 - . . - :L--- - - - ---L-- - - . . . .._L_ J _ . J . . . . __ L__j

0 1 0.2. 04 06 as lo 2.o 4 o

. 0.3 0.5 0. 7 0.9 3 o 5 oMa umu rn ctrou t Dressu re o , MPct

v ' / rgr-out:)

Fig 16 Pull-out resistance of Expander Bodies

corresponds a cone penetration resistance of about

10 MPa. I t should be noted that the results areaffected. for example, by the particle size. the depthbelow the ground surface and the location of the groundwater level. For s i l t sand ,md gravel the cone

penetration resistance in MPa is approximately 0.2 N30 •

1527

0.4 N30 and 0.6 N30 , respectively.

However, the result from the weight soundings are a tlarge depths (> 10 m} influenced b y the fr iction alongthe sounding rod since a casing .is not used, while a tSPI' the results are affected by the method used to l i f tand to release the hammer. The energy delivered by afree

fallinghammer i s considerably higher than that.

when the rope and pulley method is used.

Load tests indicate that the end bearing· capacitycorresponds closely to the cone penetration resistance(CPT} within a zone that extends one pile diameterbelow and 3.75 pile diameters above the pile point (van

der Veen and Boersma, 1952}. In cohesionless soils thetensile resistance will be lower than the end bearingcapacity because of the reduction of the over-burdenpressure as mentioned above. I t is therefore,



TABLE I

a>MPARISON BETWEEN DIFFERENT PENETRATION TESTS

after Broms and Bergdahl, 1982)

Cone Penetrat ion Tests (CPT),

Relative Point Resistancee n s ~ t y qs MPa

Very loose 2.5

Loose 2.5 - 5

Medium 5 - 10

Dense 10 - 20

Very dense > 20

suggested that the tensi le resis tance of soi l anchors

should be taken s 70 of the bearing capacity of anequivalent pi le .

Test data indicate also that the tensi le resis tance ofthe expander bodies wil l decrease with increasingdiameter. I t is , therefore , suggested that the uni ttensi le resis tance of 0.5 m and 0.8 m diameter expander

bodies should be taken as 80 and 50 , respectively of

the resistanye of expander bodies with 0.3 m diameter.

The net end resis tance in clay can be estimated from

(16)

when the anchor is located a t least four diameters

below the ground surface.

Also the skin resis tance (ca) wil l depend on the

undrained shear strength cu of the clay

s a ca u

(17)

where i s a reduction coefficient which decreases withincreasing shear strength. I t is suggested thatshould be taken s 0.8 for sof t clays (cu 50 kPa) and

s 0.5 for medium to s t i f f clays when cu >50 kPa.

I t should be noted that the tensi le resis tance wil lgradually increase with time af ter the instal lat ion due

to the reconsolidation of the clay. Particulary theskin fr ict ion resis tance is affected. About 1 to 3

months wil l be required in sof t clay to reach theXimum resis tance while in medium to s t i f f clay the

calculated tensi le resis tance usually wil l be obtainedwithin a few weeks. In weathered rock and residualsoils a value 0 .45 C i s CODIIIOnly USed. The tensile

·

resis tance can be increased further by enlarging the

boreholes by underreaming.

The pull-out resis tance of ground anchors in rock hasbeen correlated with the unconfined compressive

strength. The allowable shear resis tance is oftentaken as 0.1 where is the unconfined compressive

Standard Penetrat ion Tests (SPT),

Penetration

Resistance N20 ,

Weigth Sounding

Tests, Penetra

t ion Re.sistanceblows/30 em Nw' ht /0 .2 m

4

4 - 10

10 - 30

3 0 - 5 0

>50

4

10 - 30

3 0 - 6 0

6 0 - 100

> 100

strength of small diameter rock cores. The maximum

shear resis tance is normally limited to 4 MPa.HQwever, the spacing and the orientation of the jo in t

in the rock can have a large influence on the pull-outresis tance. The reduction of the shear resis tance hasbeen related to the RQD-value of the rock. Failure of

rock anchors located close to the ground surface (D <1.5 m) often occurs when a cone of rock i s pulled outtogether with the anchor rod or the cable . The tensi leresis tance will in that case correspond to the weight

of the rock cone and thus to the uni t weight of the

rock mass.

SETil EMENTS AND LA'FERAL DISPLACEMENTS

Deep excavations in sof t clay can cause settlementsaround the excavation. As a resul t surrounding

buildings can be damaged. The damage can be related toei ther the angular distort ion, the relat ive deflection(sagging and hogging) or the la tera l deformation of thebuilding. Buildings are in general more af fec ted by

large re la t ive deflections or by large la tera l

deformations than by an angular distort ion. Structuresare also more sensi t ive to hogging than to sagging.

Buildings located close to an excavation are oftenloaded in compression while buildings located furtheraway are subjected to lateral tension (elongation) and

. may therefore crack. The locat ion of the buildingwithin the settlement trough around an open excavationis thus important.

1528

The la tera l displacement of the soi l around deep

excavations nd i t s effect on nearby buildings has

attracted so far re la t ive ly l i t t l e at tent ion. The

resulting la tera l movement can damage buildings closeto the excavation and other structures . A tensi lest ra in of only 0.1X to 0.2X is often suff icient tocause extensive cracking of ma.sona.ry s t ructures . E.g.O Rourke (1981) has observed large la tera l s t ra insbehind an 18 m deep excavation. The resulting la tera l

displacements were high enouih to cause extensivecracking of ma.sonary structures located up to 9 mbehind the excavation.

Some settlements wil l always occur even when the bestavailable construct ion teclmique i s been used and the



soil conditions are favourable. The installation ofthe top strut is particularly important. When thecosts of different methods to reduce settlements areestimated, i t is important to consider also theindirect costs e.g. loss of time and business caused bycongestion around the si te due to the construction.Grouting and freezing require, for example, space fordrilling rigs, mixing and refrigeration units, pipesand pumps as well as for storage of various chemicalsand aggregates. For a particular job i t is importantthat the total costs including indirect costs should be

as low as possible.

The f in i te element method (FEM) provides an alternativeapproach to analyze deep excavations with respect tosettlements and lateral displacements. This method canhandle complicated soil and boundary conditions. Thenonlinear behaviour of the soil and of the supportsystem can be considered as well as the constructionsequence. Many case records have been reported in theliterature where FEM has been used to analyze theresults (D'Appolonia, 1971; Clough and Davison , 1977;Burland e t al 1979; Karlsrud et al , 1980; Mana andClough, 1981 and O'Rourke, 1981). Both braced and

anchored excavations have been investigated (Egger,1972; Clough and Tsui, 1974; Stroh and Breth, 1976,Clough and Mana, 1977 and Clough and Hansen, 1981).These studies show that the settlements and the lateraldisplacements of sheet pile or diaphragm walls in soft

clay are to a large extent affected by the factor ofsafety with respect to base heave, by the stiffness ofthe wall. by the support system (ground anchors andstruts), by the geometry of the excavation and by thechosen construction method.

Settlements should be measured frequently during the

excavation by level surveying and the results should beplotted and evaluated so that remedial measures, inecessary, can be taken in time. Inclinometers can bee.g. used to determine the lateral displacements ofsheet pi le or of diaphragm walls. There areinclinometers available with a high resolution(1:10,000) so that lateral displacements as small as 1to 2 can be detected. FEM can be helpful to locatethe source of the settlements or of the lateraldisplacements. Lee et al (1986) have recentlydescribed the monitoring of a deep excavation in soft

clay in Singapore.

The lateral displacements of braced and anchored sheetpile or of diaphragm walls depend to a large extent onthe stiffness of the walls. The displacement is often

expressed in terms of a stiffness factor E l 41E Is w w

where e is the vertical spacing of the struts or the

anchors, Es and Ew are the moduli of elasticity of the

soil and of the wall material respectively, and Iw is

the second area of moment of the sheet piles .

Field observations as well as FEA indicate that i t is

important to place the struts as soon as possible afterthe excavation has reached the strut level. Frequently

the struts are not installed until the excavation had

advanced an addi tiona tWo to three meters. In that

case the settlements and the lateral displacements caneasily increase 50 to 100 . I t is also important thatthe wale beams are tightly wedged against the sheetpiles in order to reduce the settlements. Gaps should

be fi l led with concrete or be shimmed.

The lateral displacements can be reduced by increasingthe stiffness of the wall or by decreasing the verticalspacing of the struts or of the anchors E.g. adiaphragm wall can be used instead of sheet piles .Anchors are very effective since they. can be placed

1529

close to the bottom of the excavation and be preloaded.Raked struts nd temporary berms can then be avoided.

One difficulty with FEM is the choice of parameterssince they should reflect both the in-situ behaviour ofthe soil or of the rock as well as the effect of e.g.workmanship and time. I t is important to check thecalculations at an early stage with field measurements.The design should be reanalyzed using back soilproperties i f the discrepancy is large.

J]iSTALLATION OF SHEET PILES

I t is often diff icul t to drive the sheet pilessufficiently deep into the underlying rock in order toprovide sufficient lateral resistance so that the highlateral earth pressure behind the wall can be resistedespecially when the depth of the excavation is large.This is frequently the case in Sweden where the softclay often is underlain by unweathered hard granitewith a compressive strength of 150 to 200 MPa or more.Steel dowels are often used which are driven into therock or placed in predrilled holes nd grouted in orderto increase the lateral resistance of the sheet pi lesas i l lustra ted in Fig 17a. The drill ing is normallydone through steel pipes which have been attached tothe sheet piles before the driving. The lateralresistance of the steel dowels depends on the strengthof the rock nd on the dimensions of the dowels. I t is

also possible to install addi tiona ground anchorsclose to the bottom of the excavation as shown in Fig17b to increase the lateral resistance of the sheetpiles in order.

'Another common case is i l lustra ted in Fig 18a where i th s not been possible to drive the sheet pilessufficiently deep because of stones or boulders in thesoil which interfere with the driving. Addi tionaanchors may be required a t the toe of the sheet pi lesin order to increase the lateral resistance. However,an additional row of anchors will increase the vert icalforce in the sheet pi le which bas to be considered.

a SteeL b Aciddt. onaL

Fig 17

d.owe .:s groun onchor.s

Prevention of toe failure for anchored sheetpi le walls



a. J h ~ z d f i d e n tj2.UlefMion do/ f_h

Pion

.··

r / . 1 /1 <. . . . .

..... :...z: <

b. 5fab/lr2nlon wtl 6

~ l e e 1 / - f ? l e ' ~ .

Fig 18 Vertical s tabi l i ty of anchored sheet pile

walls

Erosion may even occur below the boulders or the stonesi the surface of the cut is not protected by, forexample, shotcrete. Drain holes will be required toreduce the high water pressure that otherwise maydevelop behind the shotcrete layer.

Fig 18b i l lus tra tes the case when the ver t icalstability of the sheet pile wall i s not sufficient andthe vertical force in the sheet piles from the inclinedanchors will cause the sheet piles to set t le Thevert ical s tabi l i ty of the wall can be increased by

driving steel H-piles in front of the wall as shown.The H-piles should be welded to the sheet piles so thatthe vertical force from the anchors can be transferredto the pi les . The bearing capacity of the H-pilesshould be sufficient ly high so that they will be ableto carry the vertical force.

IMPROVEMENT OF 1HE SI'ABILITY IN SOFT a..AY

Different methods can e used to increase the s tabi l i ty

of braced or anchored sheet pi le wall in soft clay as

i l lustrated in Figs 19 through 22. Lime or cementcolumns have been installed in Fig 19 in front of orbetween the two rows of sheet pi les in order toincrease the average shear strenght of the clay and

thus the passive resis tance of the soil .

The lime or cement columns can also be installed insuch a way that they form a series of continuous wallsbetween the two sheet pi le walls to keep them apart .The lateral earth pressure acting on the sheet pi lesbelow the bottom of the excavation will then be

1530

transferred through the walls. In th is case, thecolumns will function as an additional level of s tru ts

below the bottom of the excavation. The requiredspacing of the 1 me or cement columns depends on theincrease of the shear strength that can be obtainedwith lime quick lime) or with cement. This can beinvestigated in the laboratory by mixing the clay withdifferent amounts of lime and cement. The optimum lime

content i s usually 6% to 10% with respect to the dryunit weight. About 15% to 25% cement is usuallyrequired in order to reach the required shear strength

of the stabilized soi l . Gypsum in combination withquicklime can be beneficial in organic soils.

The columns will increase the average undrained shearstrength of the soi l In soft clay the average shearstrength can usually be doubled i the 0.5 m diameter

lime or cement columns are spaced 1.4 to 1.5 m apart .Lime or cement columns can also be placed behind thesheet pi les in order to reduce the la teral earthpressure acting on the wall.

The soil a t the ground surface has been excavated inFig 20 in order to reduce the total overburden pressureat the bottom of the excavation. The reduction of thelateral earth pressure on the wall will be large below

the excavation especial ly when the total overburdenpressure a t the bottom of the excavation is

approximately equal to Nc cu. The excavated soil can

be replaced by light weight f i l l e.g. expanded shale,slag or flyash. In the Scandinavian countries and inFinland sawdust, bark and peat are often used. Withslag or flyash, pollut ion of the ground water might

become a problem.

Also jet grouting and quick lime columns can be used toincrease the s tabi l i ty as shown in Fig 20 as has been

the, case in Singapore. At the quicklime column method

v' ' t T rf

1---

E evcrlton

~ · r n e or

Cem ent

Co ..u.rnn5

r ement coLumn.>_j,

_ ~ m e oI-

• i·· ·

- y

}• • :-.. - •••• -

•. • •••

J• .y•• ...A A

A.Uernative I

A l ~ e r n a f : L · v e [

v

Fig 19 Stabilization with lime or cement columns



Fig 20

Fig 21

et Jrot.dt:mJor ~ w e i f me

coturnns

Qu,·cl:. [,·me.

coi..W??ns

Stabilization with l ight-weight f i l lgrouting or quicklime columns

je t

Stabi l iza t ion with Bakau pi les and embankment

pi les

1531

large diameter holes which are f i l led with quicklime

are used. At th is method, the expansion that takesplace when the unslaked lime reacts with water is

uti l ized. The method i s mainly effect ive in s i l ty

soi l s with a low plast ic i ty index where a small change

of the water content will have a large effect . on theshear strength. The effectiveness of the method is

however, reduced when the soi l is s tra t i f ied . Then theexpansion of the quicklime columns wil l occur faster

than the consolidat ion of the sof t so i l around thecolumns. As a resul t the soi l wil l be displaced and

heave rather than consolidate.

Embankment or Bakau p i l es are used in Fig 21 in orderto reduce the la teral earth pressure acting of thesheet pi le wall. The pi les will carry part of theweight of the clay due to the fr iction or adhesion

along the pi les . The efficiency of the embankmentpiles can be increased i f the pi les are provided withconcrete caps which will transfer the weight· of thesoi l above the caps to the pi les . P i le caps are

required especially when concrete or steel p i les withhigh bearing capacity are used because of the largelength required to t ransfer the load from the soi l tothe pi les though adhesion or fr iction along the pi les .

The transfer length wil l be large because of therelat ively high p i l e loads which are required in order

to make the method economical. Embankment pi les arecommon in Sweden, Finland and Norway par t icular ly in

sof t clay. Bakau pi les are extensively used as

embankment pi les in Southeast Asia. They have theadvantage that the surface area i s large, that thetransfer length i s small and that they are cheap. The

diameter i s usually SO to 100 mm The maximum lengthi s about 6 m f longer pi les are required they had tobe spliced.

The stabi l iz ing ef fect of embankment pi les i s

equivalent to that caused by an increase of the uni t

weight of the soi l below the excavation bottom as

i l lust ra ted in Fig 22. The equivalent uni t weight ..,eff

of the soi l when the embankment pi les are used to

stabilize an embankment or slope can be estimated from

the equation

where d =-diameter of the pi les

ca = adhesion of the c lay along the pi les

a = spacing of the pi les

'f = uni t weight of the soi l between the pi les

n elCBIIIple where an 7. 6 m deep excavation in soft

marine clay was successfully stabi l ized with 6 m long

Bakau pi les has been described by Broms and Wong1985).

Other methods· that have been used to increa.Se thes tabi l i ty with respect to bottom heave are shown in Fig

0 . The s tab i l i ty can be improved by driving a few

sheet pi les to a so i l layer with high bearing capacity

so that part of the weight of the soi l can be carriedby the skin f r ic t ion along the sheet piles. I t is a l so

possible to use inclined anchors in order to increasethe ver t ical s tab i l i ty of the sheet pi le wall as shown.This method can be economical if there i s a concreteslab next to the excavation. The s tab i l i ty can beincreased as well by placing the bottom level of s t ru t s

in trenches below the bottom of the excavation.Thereby the effect ive length of the sheet pi les below

the lower s tru t level will be reduced.



Fig 22

0

_. l o 0 0

J v 1i c

v v

iy y y

Elevofti:m

Increase a£ the equivalent uni t weight using

embankment pi les

FAILURE OF A SINGLE ANCHOR

The redis tr ibution a£ the load that takes place when

one or several a£ the anchors or s t ruts £ail has been

investigated by St i l l e {1976) and by St i l l e and Brems

1976). In Fig 24 i s s ~ w the load redistributionthat was observed for an anchored sheet pi le wall a t

M8lntorp. Sweden in a very soft clay with an average

shear strength o£ 18 kPa when one or two a£ the anchors

were unloaded. For thi s sheet pi le wall which was

anchored a t two levels i t was observed that the maximum

increase of the load in the adjacent anchors was 9

when one anchor was unloaded and that the lOad

increased by an addit ional 8X when the load in a second

anchor was released. I t is interesting to note thatthe total increase of the load in al l anchors was only

36 o£ the in i t ia l load in the unloaded anchor. Thus

the total lateral earth pressure on the sheet p i l e walldecreased by 64 of the in i t ia l load in the unloaded

anchor. When the second anchor was unloaded then thetotal increase of the load in the adjacent anchors was

only 16 of the in i t ia l load in that anchor. Thus thetotal lateral earth pressure on the wall decreased by

84 with respect to the in i t ia l anchor load.

The corresponding load redis tr ibution for a sheet p i l e

wall a t Bergshamra, Sweden with three anchor levels i s

shown in Fig 25. In this case Panel.Bl) the maximum

increase of load in the adjacent anchors was to 35 of

the in i t ia l anchor force before the f i r s t anchor was

unloaded. The to tal la teral earth pressure on the wall

increased by 32 with respect to the in i t ia l anchor

load. In a second panel Panel Cl) the maximum

increase of the anchor. force in the adjacent anchors

was 14 with respect to the in i t ia l load when the load

in one of the anchors was released. In thi s case the

IL___.

5/:ru l:s £n{renches

heet ptlesdrt ven c nto

/ . / : < 5 ::< :· /

a ~ r r t Y t.aye;

ond t

Fig 23 Inclined anchors and lowering of the strutlevel

to tal la teral ear th pressure on the wall increased by

4 with respect to the load in the unloaded anchor

compared with a decrease of 64 a t M8lntorp. The

behaviour of this sheet Pi.le was thus different . Thisdifference in behaviour can be explained by the

difference in mobilized shear strength of the claybehind the wall.

Fig 24

1532

·~ z e a + a r m m a a e ; ; y : ; a ~ ~ ; ; ; ; : w ; ; o : i i A , a ; r ; & E O It

8 ·

Load redistribution a t Molntorp, Sweden a t

failure o£ one or two ground anchors afterSti l le , 1976)



Fig 25

Fig 26

Load redistribution a t Bergshamra, Sweden at

failure of one or two ground anchors afterStil le, 1976

fvlobd t zcrtt on o

; heqr sirenJ- h

Load redistribution due to mobilization ofshear strength

1533

The lateral earth pressure acting on a braced or ananchored sheet pi le wall depends on the lateraldisplacement required to mobilize the shear strength ofthe soil behind the wall and on the factor of safetyused in the design. The wall will deflect laterallywhen the load in one of the anchors is released or theanchor fails. The increase of the lateral deflect ionof the wall is generally .sufficient to mobilize theshear strength of the clay along a potential failuresurfaces behind the wall as illustrated in Fig 26. Arelative small deflection is normally required to

develop the maximum shear strength of even soft clay

compared with the displacement required to develop theultimate resistance of the anchors or of the struts.In the case the factor of safety ini t ia l ly isrelatively high then only a small part of the availableshear strength will ini t ia l ly be mobilized. Areduction of the force in one of the anchors will thenmainly increase the average shear stress alongpotential failure surfaces in the clay. In this case,the increase of the load in the adjacent anchors willbe small and the total lateral earth pressure on the

wall will decrease when one of the anchors is unloaded

or fails as was the case at Molntorp.

If on the other hand the factor of safety is low andthe shear strength of the clay has been fully mobilizedbefore the release of the force in one of the anchorsthen the failure of one of the anchor will result in alarge increase of the load in the adjacent anchors.

The total load on the sheet pile wall m y even increasewhen the peak strength of the clay has been exceededand the residual shear strength is lower than the peak

strength. This was the case at Bergshamra where thetotal force acting on the sheet pile wall increasedwhen the load in one of the anchors was released.

The consequences when one of the anchors fai l will thus

depend to a large part on the chosen factor of safety.If a relatively high factor of safety has been used inthe design 1.5) and only part of the shear strengthof the soil will be mobilized a t working loads then theincrease of the load in the adjacent anchors will besmall when one of the anchors fails. If on the other

hand the factor of safety is close to 1.0 then thefailure of one of the anchors will cause a largeincrease of the load in the adjacent anchors which alsomay fa i l The total lateral earth pressure on thesheet pi le wall may ev.en increase and cause aprogressive failure of the whole wall {zipper effect).

SfABILI1Y OF THE BASE OF A SHEET PILE WAlL

Several failure of anchored walls have been occurred inSweden in soft clay. In Fig 27 is shown an anchored

wall constructed of large diameter bored piles {Bromsand Bjerke, 1973). The exposed clay between the pileswas shotcreted during the excavation. Clay started toflow into the excavation below the shotcreted part ofthe wall almost like tooth paste squeezed out of a tubewhen the depth of the excavation was 5.5 m. Within afew minutes the excavation was f i l led with softremoulded clay due to the high sensitivity of the clay.Failure took place when the total overburden pressureat the bottom of the excavation was about 6 c where c

u u

is the undrained shear strength of the clay asdetermined by f ield vane tests. The factor 6.0corresponds to the stabi l i ty factor Ncb This type of

construction using bored piles and shotcrete istherefore not suitable for soft clay when the depth ofthe excavation is large and the total overburden

pressure at the bottom of the excavation exceeds N bec_ u



5ho-l:cretc- ,...,s.srn Bored pde

.:5hotcre.f:e-

Sol : do j_,

c4 12 /;Pet

A

Fig 27 Failure of a vertical cut in soft. clay (after

Brems Bjerke, 1973)

(Brems and Bennerma.rk, 1967). Steel sheet piles orcontiguous bored pi les should have been used instead.

Several fai lures have also occurred in Sweden when the

sheet pi les have been driven to rock through a deep

layer of soft clay. Because of the high compressive

strength of the granite i t is not possible to drive thesheet piles into the rock. Soft clay was squeezed intothe excavation through the triangular openings which

were formed between the bottom of the sheet pi les andthe rock as shown in Fig 2S since the surface of therock was inclined. Large settlements were observedoutside the wall. The diameter of the depressionscorresponded approximately to the depth of theexcavation.

STABILITY OF DEEP EXCAVATIONS IN SOFT CLAY IN SINGAPORE·

T-hree deep excavations in soft marine clay in Singapore

have been analyzed using a modified version of the

computer program EXCAV In the original program which

was developed a t the University of California a tBerkeley by hang and Duncan (1977) a non-linearhyperbolic soil model (Duncan et al 1980) is uti l izedto describe the soi l behaviour. The program can modelthe excavation layer by layer, the installation and thepreloading of the struts and the application of asurcharge load.

The f irst project involves a braced sheet pile wall,where the sheet pi les have been driven into a deep

stratum of soft marine clay. In the second project theexcessive plast ic yielding of a braced sheet pile wall

has been investigated. The third project is concernedwith the prediction prior to the construction of wall

movements for a deep excavation in soft clay.

The short term conditions have been investigated with atotal stress analysis using the undrained shearstrength of the soft clay. The soft marine clay hasbeen assumed to be saturated and incompressible. APoisson's ratio of 0.495 has been used in the analysis.The elast ic modulus (Eu) that corresponds to undrained

conditions has been assumed to 100 cu to 200 cu. This

equivalent modulus corresponds to the in i t ia l tangentmodulus, Ei of the soft clay. The tangent modulus, Et'

is a function of E. and of the stress level.l

ljl-l-roct..l.

~ t .-rn·r-,-

1534

\ _v f Dc,z c <::::: /;{ / z ..c t.

C o I OTYie

{n here

:5ectt.. on A-A

Fig 2S Failure by bottom heave



Proiect A. This project i s located in the CentralBusiness Dist r ic t (CBD) of Singapore Fig 29). The

size of the 11.1 m deep excavation i s 42.6 m x 27.0 m.

The walls of the excavation were supported by 30 m long

sheet pi les (FSP IIIA) which were driven 19 m below thebottom of e x c a ~ a t i o n Six levels of st ruts supported

the wall. The ver t ica l spacing of the st rut variedbetween 1.5 m to 2.5 m. The horizontal spacing wasabout 6 m.

The excavation proceeded in stages. The s t ru t s

supporting the sheet pi les were instal led during eachexcavation stage 0.5 m above the bottom of theexcavation and they were preloaded to 15 percent of the

design load. The s i te was divided into three sect ionsduring the excavation. In the present study thebehaviour of the sheet p i l e wall in the middle sect ion

of the excavation has been analyzed.

Six slope indicator pipes were instal led behind thesheet pi le wall as shown in Fig 29. Surface monumentswere established to determine the settlements behind

the sheet pi les Stra in gages were attached to

selected st ruts in order to evaluate the st rut loads.

A typical soi l prof i l e i s shown in Fig 30. A sandy

f i l l about 1 to 2 m thick i s located a t the ground

surface . The f i l l was followed by a deep layer withsof t marine clay which belonged to the Kallang

Formation. The clay consis ts of two dist inct members,an upper layer which i s approximately 25 m th ick and an

approximately 7 m thick lower layer. The two layersare separated by a layer of loose to medium dense s i l ty

sand. A layer of s t i f f sandy s i l ~ basically decomposed

granite was found below the marine clay.

Fl

Upper morine

day

Lower mor ne

cloy

2

·

Legend

G Inclinometer

r J Heave point

Fig 29

0

Scale

10 2 m

Instrumentation - Project A

The water contents of the upper and lower members ofthe sof t marine clay were 70 and 50 , respectively.

The l iquid and p las t i c l imits of the upper marine clayvaried between SO and 105 and between 60 and 70 ,respectively. The l iquid and the plast ic l imi ts of the

lower marine clay were 70 and 50 , respectively.

Oedometer tests indicated that the marine clay wass l ight ly overconsolidated. The undrained shear

strength for the upper and lower members of the marine

clay increased almost l inear ly with depth.

Lafera di¥1acemenf1 mm50 1 6 2Q::) 25

Fintle e l ~ r n e n lo n o f r . ~ 6{Fc-A

l?anqe fromtncllnomelerMadti?j5

Fig 30 Measured and calculated wall deflections -Project A

1535



Field measurements indicated that the wall graduallymoved inwards with increasing depth of the excavation.The maximum deflection of the middle section of theexcavation was 150 to 170 nm when the excavation had

reached i t s final depth of 11 m. This is about 1.5 ofthe excavated depth A comparison between the measuredand the computed deflections is shown in Fig 30.

The observed surface settlements when the depth ofexcavation was 5. 75 m and 11.1 m are shown in Fig 31.

The lateral displacements of the wall thus caused large

settlements that spread far behind the wall. Themaximum settlement was about 1 of the f inal excavationdepth. t occurred a t a distance from the excavationequal to about half the excavation depth.

The measured settlements are plotted in Fig 32 asproposed by Peck 1 9 6 ~ ) . t can be seen that thesettlements even at a distance of 3.5 times theexcavation depth were large. This behaviour can beexplained the restraint of the lateral deformations and

of the settlements of the sheet pile wall by t h ~ sandlayer at the toe of the wall as i l lustrated in Fig 33.

This was confirmed by FEM.

Distance from sheet piles, m

0o.------5..-- ....o _

~ . . . ~~ 5 . 7 5 m excavation

100

/ l l . lm excavation

Settlement mm

Fig 31 Measured ground settlements - Project A

The maximum bending moment in the wall has been backcalculated from the curvature of the sheet piles whichwas determined from the inclinometer measurements.These measurements indicated that local yielding of thesheet pi le occurred during the final stage of theexcavation as indicated in Fig 34. The yield moment ofFSP lilA sheet piles is about 380 kN/mlm. The computedmaximum bending moment by FEM was 372 kN/mlm. The

finite element analysis also indicated that the wallwas highly stressed down to about 6 m below the bottomof excavation.

Both field measurements and FEA indicate that the strutload increased rapidly with increasing depth of theexcavation. The strut loads reached a maximum justbefore the installation of the next level of struts.Thereafter the strut loads decreased slightly withincreasing excavation depth.

A comparison between measured and computed strut loadsis shown in Fig 35 for the top three levels. Themeasured strut loads agreed closely with thosecalculated by FEM, The pressure distributiondetermined by the tributary area method is shown in Fig

1536

D / ~ f a n c e from e tcaYalton£ w::; v ftbn depfh

I 2 4

Fig 32 Normalized settlements (Peck, 1969)

36. t can be seen that the measured and the computedvalues are in close agreement. t appears that theapparent pressure diagram proposed by Terzaghi and Peck(1967) at m = 0.4 is conservative. A better match is

obtained with m = 0.7.

The penetration depth of the sheet piles {19 m belowthe bottom of the excavation was 1.73 times the depth{11.1 m). An analyses using FEM indicated that thepenetration depth could have been reduced by 13.5 mwithout any significant increase of the strut loads.This conclusion concurs with the observation by Peck{1969) that very l i t t l e is gained in soft to mediumst i ff clay by driving the sheet piles far below thebottom of the excavation provided the stabi l i ty of theexcavation with respect to bottom heave is sufficient.

a)

Fig 33

{bj

Relationship between settlements and lateraldisplacements - Project A



0

-1/ .; m 10

~ ~ / r 7 / / T o t ~ ~ L : T / ~ ; r r r l l ~ ;I

i

Fig 34

"

"

"/

Fig 36

20

Moment distribution at different stages ofconstruction - Project A

LeJ-end.

-lII

III

IIII

I

I

II1 m ~ o ~

l.a- en:;d earlhprt>:5.5Ur€J i Pa

0 f ietd m e a s u r e m e n f ~- H..r11le elerneYJf ano y.:st:S

Terzctyl j .Oecf.

Comparison between measured and calculateds tru t loads

1537

ZOO

fltostlred (a culcrfed

IJet-lh ol euova.fton1.. 2

Fig 35 Strut loads - Project A

Project B. This project , which is located just outsidethe Central Business District in Singapore i l lust rates

the influence of the construction sequence on theperformance of braced excavations. The size of the14.7 m deep excavation was 200 m x 35 m. A

cross-section of the excavation · i s shown in Fig 37.FSP IV sheet piles with a total length of 18.5 m were

driven 3. 8 m below the bottom of the excavation D =0.26H). The sheet pi les were supported a t six levels.The vertical spacing of the struts varied between 2 mto 2.5 m. The horizontal spacing was about 5.5 m.

The soil condition a t this s i t e was highly variable. Asoil profile along section A-A is shown in Fig 38. Onthe west side, the sheet pi les were driven into a st i ff

sandy s i l t or clay decomposed granite). On the eastside, the sof t marine clay extended the full depth of

the excavation. The ground water level was locatedabout 1.0 m below the ground surface.

The soil profile on the east side of the excavation i s

similar to that at Project A. The upper and lower

members of the Kallang Formation with sof t marine clay

are separated by a layer of loose to medium dense sand.Below the marine clay is a deep stratum of decomposedgranite, a s t i f f sandy s i l t or clay. The upper marine

clay is organic peaty) with an average undrained shearstrength of about 10 kPa. The average undrained shearstrength of the lower marine clay is 15 kPa. Thedecomposed g,ranite has an estimated undrained shear

strength of about 70 kPa. This material was verydiff icul t to sample and to test.



The soft clay on the east side was adopted in the FEM

analysis since i t is more crit ical than the st i ff soilon the west side. The excavation was carried out inseven stages. t should be noted that the struts Slwere placed after level El had been reached Fig 37}.The excavation proceeded down to level E2 prior to theinstallation of the struts a t this level. Thissequence was continued down to level E7. The strutswere preloaded to 70 of the design load.

The observed wall movements are shown in Fig 3Sa. The

maximum deflection was 270 mm which is 1.8 of thefinal excavation depth. This deflection is relativelylarge since the sheet piles were driven into a st i ff

soil . The computed maximum deflection was only about200 mm regardless of the strength and stiffness of thesoils when the wall was assumed to be l inearly elastic,i .e . non-yielding.

• Inc t"nom.e-1-er

:Janel·.:_1n ~ u t Sf·•.· 5Z

~:54

:55

/ t c o m ; o ~ ~

The computed deflections a t the different stages of theexcavation are shown in Fig 38b for the case when thewall yields. t can be seen that the computed lateraldeflections are in good agreement with the measuredvalues.

The maximum settlement, 100 mm occurred at a distancefrom the excavation which corresponded to aboutone-half the excavation depth. The computedsettlements fal l within Zone I of the normalizedsettlement chart proposed by Peck (1969}.

The measured strut loads were low. A compar isonbetween the measured and computed apparent lateralearth pressures is shown in Fig 40. The measured loadswere considerably smaller than those computed· by FEM

except for the two strut levels at the bottom of theexcavation. One possible explanation of this behaviour

0

~ c o l e10

(7 7

2 m

; ~ n d

yronile Hom ol eLcall c:dtonf I

,

( ~ - ; ; t e 0 :5 1 m

wall ( 'PN)

Fig 37 Plan and cross-section for Project B

1538

[ ) p c o m ; - ~ c l.Jml?rfe



D e p 1 h . ~ m

t ;

8

Z

16

.20

E.tcavatLol }

//efl e/

E4

E7

1 7

300

HORIZONTAL MOVEMENTmm

a) Measured deflection

300 200 100 0

HORIZONTAL MOVMENTmm

b) Computed deflection

El

E3

Fig 38 Measured and computed wall deflectionsProject B

is the st i f f soil a t the west side of excavation. I t

has been assumed in the analysis that the soft clayextended over the e ~ t i r e excavation.

The measured maximum wall deflection, 275 mm,

corresponds to about 1. 9% of the depth of theexcavation which is rather high for a sheet pile walldriven into st i f f soil. This large deflection couldhave caused by yielding of the sheet piles a t an earlystage of the excavation.

The analysis indicates that yielding occurred when the

excavation reached Level E3, only 7 m below the groundsurface due to overexcavation prior to the installationof the struts. Especially the f i rs t level of struts isaffected.

The installation of the struts lagged behind theexcavation of the soft clay by as much as 2.0 m whichundoubtly increased the bending moments in the sheet

pile wall. I t is thus very important to limit thedifference between the st rut level and the excavationlevel as much as possible when the struts areinstalled. This difference should not exceed 0.5 m.

The effect of the construction sequence was alsoinvestigated assuming that the depth of the excavationand the strut levels are the same when the struts areinstalled. In this case, the computed maximum

deflection is only 120 as shown in Fig 40 which is

less than half the measured values. The computedsurface settlements and the strut loads were also muchsmaller. In fact the maximum bending moment in thewall never reached the yield moment of the sheet piles ,590 kN/mlm. ·

51

52

54_.;

I.

........ I

: _

_.:::r :-h.I 'I ..........

I~ F E M iI I

L . ~ ·II

r------- -------t•1I

•'I

I

I

II

II

I

I

II

erzaghi & Peck

m=l.O) 2...j

I I~ ~ ~ J _ _ ______ • __J

.50 100 1.50 200

end:

Obse r ved

------FEM·Terzaghl & Peck 1967)

kPa2.50

Fig 39 Computed and measured lateral earth pressures- Project B

1539



Fig 40

D t ~ f o n c c • from E> ecovcrlton

E ecavofion depth

o a ; oJ..-..---- ------ ---

1I J f r u l ~ n led /mme

\ d t o l e ~ aller e tch uro

\ vafton tYe(\

,...... _ ... .

_____

Computed surface settlements - Project B

Proiect C With the lmowledge gained from Projects A

and B, an attempt was made to predict the wall movementa t Site C prior to construction. This si te is locatedabout one kilometer away from Project B. The length ofthe excavation is about 66 m. The width varies from6.0 m to about 12.0 m as shown. in Fig 42. The totaldepth is 15.0 m. The field instrumentation inCludedone inclinometer pipe, a number of s tra in gages

attached to selected st ruts 8nd several survey markers.

Steel sheet piles FSP VIL) supported a t five levelswere used as shown in Fig 42. The 26 m long sheetpiles were driven 11 m D = 0.73 H) below the bottom ofthe excavation. The vertical spacing of the strutsvaried between 2.0 and 3.5 m. The horizontal spacingwas 3.7 m.

0 10 zom

• tA

Two series of analysis were performed. The f i r s t wasdone prior to construction while the second series wascarried out af ter the excavation had been completed.Soil data from only three boreholes were availableprior to excavation. The soil conditions variedconsiderably between the three holes which were locatedrelatively far from the si te Fig 42}. Both the upper

and the lower members of the soft marine clay werepresent in Borehole A whereas only the upper membercould be found in Boreholes B and C. The depth to thebottom of the soft clay layer was 16.7 m a t Borehole A,11.5 mat Borehole Band 9.4 mat Borehole C.

The soil conditions at Borehole A was used in theanalysis Case I) since i t was the closest of the threeboreholes to the investigated section. The average

undrained strength of the upper and lower layers of thesoft marine clay was 10 and 15 kPa, respectively. Thisis about the same shear strength as that observed inProject B. Because of the close proximity and thesimilarity of soi l conditions between Projects Band C,the soil parameters· in Project B were used in theanalysis. An Eu/cu ratio of 150 was used for the upper

layer since the upper marine clay was less peaty thana t Project B. A value of 200 was used on the

E /c -ratio for the lower marine clay.u u

A comparison of the measured and computed walldeflections after the excavation had reached the f inal

depth is shown in Fig 43. The calculated maximumdeflection, 75 mm, was only about half of the observedmaximum deflection 150 mm).

A parametric study was done prior to the excavationbecause of the variable soil conditions, in order toassess the effect .of ;the thiclmess of the soft clay.

In one case,. the soft marine clay was assumed to extenddown to 20 m depth. This assumption was later verifiedby a cone penetration tes t CPT) next, to Section A-A.For this case, the computed maximum wall deflection was142 mm Fig 44) which agreed closely with the observedmaximum deflection of 150 mm. The surface f i l l as wellas the intermediate sand layer were assumed to beabsent.

. \..,_ / r;coftbn o/

borehole

Fig 41 Site ·plan - Project C

1540



Ji7 2SI> q l 8Sm

--'L.

-7.5 m

L...- ;o.srn. . L

- ] . ( ) ' ". L _

- ' ' ~ · 7....f._

J O O ~ .. JI.s ti B

].S()J< ()Jtl31fdl

J S ) ~ t } S O S . 1 3 7 l l l l

~ 0 . < ( ) JC 13l_CllI

I .-, /

( "

lbrehole A

J:i /1· : ·

u per 17Klri Jt

day

<:5anct~ r m o r i n eclay

/ ~ ~ e d~ n l ' f e

Fig 42 Sect ion A A - Project C

IJ

Shortly af t e r the excavation had been completed, asecond ser ies of analyses were carried out using an

updated soi l prof i le based on the cone penetrat ion test

next to Section A-A. There were no other changes. Acomparison between of the computed and measured wall

deflect ions a t dif feren t stages of the excavation i s

shown in Fig 45. I t can be seen that the computed

values were about 20 percent smaller than the measured

values. However the computed shape of the deflectedsheet pi les compares well with that which was measured.

There are a number of factors that can account for therelat ively small computed wall deflect ions. I t has

been assumed in the analyses that the excavation depthsand the st rut levels were the same when the st ruts were

ins ta l led . However the excavation levels during the

construct ion were a t least 0.5 m lower than the s tru t

levels. In fact, the f i r s t level of s t ru t s was not

instal led unt i l the excavation was 2.0 m to 2.5 m belowthe ground surface. This accounts for the largela teral deflect ions observed a t the f i r s t excavationlevel as shown in Fig 45. Furthermore, the lowest

level of s t ru t s (85) was not instal led unt i l the final

depth of the excavation had been reached. This

accounted for the large observed deflection during the

f inal stage of the excavation. Also the measured

ground settlements were much larger than thosecomputed.

The s t ru t loads were not measured a t th i s project. The

computed st rut loads are shown in Fig 45. The highst rut load a t level S4 was caused by the intermediatesand layer. A similar phenomenon was observed a t

Project B.

The analysis indicates that an Eu/cu-rat io of 100 to

200 gives reasonable resul t s for the sof t marine clayin Singapore with respect to settlements, la teral

displacements and s tru t loads and that la teral

deflect ions can be reduced s ign if ican t ly by insta l l ing

the st rut as early as possible and by preloading orprestressing the st ruts

For a f loating sheet p i le wall in a deep stratum ofsof t clay, the depth of penetrat ion has l i t t l e effec t

on the overal l behaviour. A penetrat ion depth equal toone-ha1f the excavation depth appears to be adequate

provided that the c r i t i ca l depth wil l not be exceeded.

J.otera/ de/led/on_ rnm

8

16

Fig 43

/i() 8 ) 120 /60 200

Measured and calculated wall deflect ions -Project C

1541



Fig 45

S1 -

s2-

53 -

S 4 -

.,..

0

DISPLACEMENT mm)

-20 0 20 40 60 80 100 120 140 160 180

4

8

12

]:r: 16

C..t ll0 20

24

28

32measured

- - - - computed

3 6 ~ - - - - ~ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ~

Fig 44 Measured and computed wall deflections

Project C

, / erzaghi £ Peck 1967), . . . m=l.O)

STABILIZATION OF DEEP EXCAVATIONS IN SOFf CLAY

IIII

IIII

f-FEM(Case 111):I

III

:I

IIIII------..J

100 200 300

Bottom heave is frequently a problem for deep

excavations in soft marine clay in Singapore. Failureof bottom heave can occur when the excavation depthexceeds about 5 m to 6 m due to the very low shearstrength of the clay. Different methods can be used toincrease the stabi l i ty. The effect of jet grouting,excavation under water and embankment piles (soilnailing) has been investigated for a 11 m deep and 33 mwide excavation in soft clay using a modified version

of the computer program EXCAV Chang and Duncan, 1977).I t has been assumed in the analysis that the sides ofthe excavation are stabilized by 33 m long sheet pi lesFSP IIIA which have been driven 22 m below theexcavation bottom. The sheet piles are supported by

struts a t four levels. The vertical spacing of thestruts is 2.5 m. The top level is located 1 m below

the ground surface.

The 50 m deep layer with soft marine clay has

assumed to be slightly overconsolidated down todepth. Below i t is normally consolidated.

L.aferol eorf.h J r l 6 S u ~ f.p lundrained shear strength (cu) is constant. 16 kPa.

been

11m

Thefrom

the ground surface down to a depth of 11 m. Below. cu

= 16 + 1.25Z kPa where Z is the depth in metres below

El. -11 m. The increase of cu corresponds to a c/p

ratio of 0.25 c u / a ~ = 0.25).

Calculated earth pressures - Project C

The short term conditions have been evaluated using atotal stress analysis. The soft marine clay has been

assumed to be saturated and incompressible. APoisson s ra t io of 0.495 has been used in the analysis.The Eu/cu ra t io has been assumed to be 200. A value of

1542

0.9 has been used to estimate the lateral earthpressure a t res t K0 } with respect to the total stress.



0

Homen+ f.Hm/m

t5.o ~E.ecavaftol?

20 o

1 2 ~ t o r Y 1 )':field momenr

3/?AJ J.Nmjm

I uklhoul

jt ____ unloadt't?g5-o

30 o f

:leo ) ).{) ) 1 011

Momeni, l mjm

Fig 46 Effect of unloading

These values have been found to be appropriate for deep

braced excavations in sof t marine clay in Singapore

{Broms e t a l 1986} and the predicted performance hasagreed well with that which has been observed.

Lo-1-ero/ ear-1-h jYt 'X>ure1 ~0 1 10 80 / ).Q

Fig 4_7 Effect of unloading on s tru t loads

1543

The la teral deflections of the sheet p i l e wall areshown in Fig 46 when the depth of the excavation i s

11 m The maximum la teral deflection of the wall is

about 400 DUD The ground settlements outside the

excavtion and the base heave within the excavation arelarge as shown in the figure. The calculated maximum·

settlement and the maximum base heave are about 200

and 600 respectively. The analysis indicates thatthe maxinrum bending moment in the sheet p i l e wall

increases rapidly with increasing depth of the

excavation. The maximum bending moment approached the ·yield strength of the FSP IIIA sheet pi les SO kNm/m.

The lateral def lec t ion of the sheet pi les the

settlements around the excavation and the bottom heave

are also shown in Fig 46 when a 10 m wide s t r ip of thesoi l h s been removed along the excavation. I t can beseen that the unloading had only a marginal effec t on

the settlements the la teral deflections of the sheetpi le wall and on the base heave. Also the effec t on

the s tru t loads is small as can be seen from Fig 47.

Jet grouting has also been used in Singapore to improve

the sof t marine clay {Miki 1985}. At th is method

contiguous or overlapping cylindrical cement columns

are formed in-si tu in the clay. The diameter of thecolumns can be up to 2.0 m The method has forexample been used to stabi l ize a 15 m deep excavation

for the Newton Circus Stat ion of the Mass Rapid TransitSystem {MRT} in Singapore and to stabi l ize tunnelsexcavated in the sof t marine clay and in loose sands.

The construct ion sequence followed a t the j e t groutinghas been modelled in the FEM-analysis. First thes tabi l i ty of the sheet p i l es during the insta l la t ion

has been analyzed. Thereafter the ef fect of the

jet-grouting. of a 3 m thick zone of soft clay between

the two sheet p i le walls below the bottom of the



Fig 49

1°0 J . . I I 2-+/- _ _ _ _ ~ _ / . 1 1J.I Momen f J i.Hnym

{0 :1.0 30 110 ;700 0 = 4cO

::ZS.o

ao o

II

I\

D e p l ~ ' ---1'---'-+--1---t---0

\

:<co o :u;o -'00

Mon?eniJ l ty/"?

Fig 48 Effect of jet grouting (3 m

Lo ferol eorl h pre5:5ure; tr:b

0 1-1 ) 80 120 14

Effect of jet grouting on strut loads

1544

excavation has been investigated. An undrained shearstrength of 150 kPa has been assumed for the stabilized3m thick layer. Cores of the grouted soil from actualprojects indicate that the· shear strength of thejet-grouted material can be much higher than 150 kPa

A comparison with the case where no soil improvementhas been used shows as indicated in Fig 48 that theperformance of the excavation is improved considerablyby the jet grouting and · that the maximum lateral

deflection of the sheet piles is reduced by about 50percent. Also the settlements and the strut loads arereduced significantly as shown in Fig 49 as well as themaximum bending moment in the sheet piles. Jetgrouting has been found to be a very effective methodto improve the overall stabi l i ty of excavations in softclay.

A further improvement can be obtained by increasing thethickness of the j e t grouted zone to 6 m as can be seen

in Fig 51. Mainly the deflections of the wall and thebottom heave are reduced. The strut loads are alsoreduced significantly at al l levels (Fig 52}. Thelargest reduction was observed for the bottom level ofstruts as shown in Table II as could be expected.

FEM has been used to evaluate the stabilizing effect ofembankment piles. t was assumed in the analysis that

the spacing of the 6 m long Bakau piles with 100diameter is 0.5 m. The piles are driven below thebottom of the excavation using a follower. The t iplevel i s located 17 m below the ground surface. Theresults of the analysis are presented in Fig 52 and inTable and compared with the case without soilimprovement. The analysis indicate that for a 11 mdeep and 33 m wide excavation, four to eight rows of



Fig 51

o, j fonce1 m

0 2/./ 311 41-/~ ~ ~ ~ ~ f l lomeni1 f:.flrn/m

0 10 20 30 0 ,21l ti(XJ

f5P .lllA

30.o

0

Del edoon, mrY)

= o :uo .t;ao

Momenf1 i.Nnyn-;

Fig 50 Effect of j t grouting 6 m

Lafera/ earth p r e 5 s u r ~ i.Pq_UJ/Ihoufj e f

g'foqltn.J

11cutmum w ~ / 14tl?mm

dd edton;mm

#o-Limum / t.Jo·moment Lfl%_m

J4qdmum .:wr/actp209mm

:5e#lement, mm

f.ile1dmum h o ~590mm

heort7; mm

Ncu/mum :Hrull a z d ~ L:o/m

Lei'd I t7ZLePet.Z .hs-

L e v ~ /.J 32.9J . e Y ~ I 358

7etgmr.dti? ?

:l()Bmm

/80

/20mm

.J96mm

1:<6177

ISZ.,?09

Jt>l-j'rrx.rhi?q

f ilm

/6/mm

.WZ}

/65"

(arY.%)

tOOmm

«eX)

.3.:xJmm("5"?<?:)

9 (S..Y7.)/53 (487,)

/ .3 (- </ '%)

s f s<c:>::;?

Effect of j t grouting {3 m and 6 m on s tru t

loads

Table Effect of j t grouting on the performance ofa 33 m wide and 11 m deep braced excavation

1545



tim

( ~ . .F3P .IllA

E ·:ZOO c

0

3o.o

I

I

30 40

-----------

wdhout

&l:eu-<.pdes

:zoo

H o m e n ~ f:.Hmjm

0 200 "-00

:Jidd rnome'?T.3ao.tl mjm

0 J.J()(j

De{/ectio0 mm

200 0 21X) i<OO

Momcn- 1 i flrnjm

Fig 52 Effect of Bakau piles

Bakau piles in front of each wall could reduce themaximum wall deflection by up to 29 and the maximum

bending moment in the sheet piles by 35 . The resultsalso indicate a substantial reduction of the strutloads t the two bottom levels Fig 53) and an increaseof the passive pressure in front of the wall. Theeffectiveness of the Bakau piles was found to increasewith decreasing width of the excavation.

Ba./;a.u.pt'l'ff's

<j-toomm

Fig 53

Laferol eorlh p r t 5 : : s u r e , ~ i:.Pq

0 1 10 80 /; .0 160

Wtth Bal::a.u.P-t?es

Effect of Bakau pi les on strut loads

1546

){odrnarn wallddltdfo0 mm

a.timttm b a ~heave; mm

#Qltinum sln:d( o : ; , d ~ I:Njm

/..eye/ IL ~ Y e 2

l..erd 3/ . - t Y ~ I

~ 3 1 mm J J:Z.mm ..106 mm

('tt7.)

Table III Effect of embankment pi les Bakau piles) onthe performance of a 33 m wide and 11 m deep

braced excavation



Fig 54

2. Oewal.,.rtng and 11? _

dol /a l ton of' :;.J.ruilevel z

3. DewafertnJ ancl tn=>1 a I a t"(:>n of ::rlrut{ere/ 3

Construction sequence

100

. :

Z-45

33m

. .L

Fig 55

15.o

20.o

25•

0

0

()

\

·\

The s t bi l i ty of deep excavations in soft clay can alsobe increased by excavating the soft clay under water.The ini t i l excavation can be done dry until the f irst

one or two rows of s tru ts have been installed. Next,

the excavation is flooded so that the soft clay can be

excavated down to the final depth. After the base slabhas been cast under water the excavation is dewatered

and the intermediate struts are installed.

A 15 m deep and 33 m wide excavation has been analyzed

using FEM. The sides of the excavation are supported

by sheet pi les Z-45 with a section modulus (SM} equal

to 4500 cm3/m. Two sets of analysis were conducted.In the f i r s t set a conventional excavation method with

five levels of s tru ts was investigated. The second setwas concerned with the excavation of the soft clayunder water. Three levels of s tru ts are used tosupport the sheet piles . The 2.0 m thick base slabwill be cast under water as shown in Fig 54.

The results show a significant improvement of theoverall . performance Fig 55). The maximum walldeflect ion was reduced by 53 . A 44 reduction of the

base heave and a 50 reduction of the ground settlementwere also obtained. The loads in the second and thirdlevel s tru ts are reduced significantly as well Fig 56

and Table IV). Because the base slab will be installedbefore the· second and third level struts, the axialload in the base slab will be high compared with that

in the two levels of s tru ts Table IV).

The FEM analysis indicates that excavation under waterdown to 15 m depth is feasible. The calculated maximumwall deflection, 130 mm, and a maximum ground

settlement of 150 are much less than those observedfor actual excavations using conventional methods even

when the maximum depth is less than 11 m.

) , : ; l - a n c e . ~ f7?

t./ 2/.1 3/.1 /.t./.f11omen i:Nmjtn

10 20 30 IJ() /fOO 400 I JOO

- · = = = = = = - - = = - ~>dtleme ni/ rnn,

\\

\\

\\II

N'nol -sf:Jr/Iunder 'ater-

hnal :5" ogewd-h concrele::stab

Normal con::Jt<dtbl') (\

\I

2.00 l.t-00 4oO 0 JOO ax>

[)e "ledc'on) mmt1omen.f- i.Hmjm

Effect of excavation under water

1547



Ld-era/ earlh ;;n 5.sure_; I:J1::t

0 J-1() 80 120 160o r · ~ ~ ~ ~ - - ~ - - - - ~ ~l:::;....;;;;;;;;;;;:;;;;l

9

\E e caval-ton

under wclft 'r

Fig 56 Effect on s tru t loads of excava-tion underwater

Excavation under water has the main advantage that the

stabi l i ty with respect to base heave is governed by thesubmerged uni t weight of the sof t marine clay (about

6 kN/m3 ) rather than the total unit weight (about 16

Norma/ A. ler e.tr:ofC hon

c o n 5 ~ r u c f t " " o n fo lim, h ~ o r e N n a / ~ olCor creftnf c o ~ u ion

lt/Q.timum wall . 3 ~ / m m 1.24/mm 165 mmd t l t ~ d t o n J mm

#ozimum

mome"" iNmjm 2tl? 80 110

Mo.tii7K.Jm ~ ~ Rh t a r ~ mm 1 ~ ~ mm ~ m m JotJmrJ?

J1o.ttmam flmlfood:J l :o/m

/.tl't/ I Iff/ 187' 138

l.erd z . J ~ 30 .

L/tl l .J ~ , 3 70f . t ;d II 55/ 95"3 I a ~

Table IV Effect of excavation under water on the

performance of a 33 m wide excavation

1548

kN/m3 ) . The wall movements during the dewatering and

the insta l la t ion of the st ru ts af ter the i n s t a l l ~ i o nof the base slab under water will mainly occur below

the slab. The la tera l deflect ions of the sheet pi les

above the slab wil l oe small.

SUMM RY

The design and construct ion of anchored and s t ru tedsheet p i l e walls in sof t clay have been reviewed. Most

failures have been caused by insuf f ic ien t penetrat iondepth of the sheet p i l e s when the walls ro ta te aroundthe level of the anchors or of the st ru ts . Failure can

also be caused by rupturing of the anchor rods or by

buckling of the st ru ts . The st ru t or anchor loads can

for deep cuts especially a t the bottom of theexcavation be considerably higher than those calculated

by a conventional method. Failure by bottom heave is

also a possib i l i ty which must be considered in thedesign.

When inclined anchors are used t i s a l so important totake into account the ver t ical force caused by the

inclined anchors or by the struts. This ver t ical forcecan reduce considerably the stabi l i ty of par t icular ly

anchored sheet p i l e walls . Several fa i lu res have

occurred which have been caused by insuf f ic ien tver t ical s tabi l i ty of the sheet pi les and where thever t ica l force caused by the inclined anchors was notconsidered in the design.

Failure of one of the anchors or s t r u t s may lead toprogress ive fai lure and complete collapse (zippereffect) of the wall . f a suf f ic ien t high factor of

safety i s used in the design then the increase of the

load in the adjacent anchors or st ru ts wil l be small a t

failure of one of the anchors.

REFERENCES

Aas, G. (1984}. Stabi l i ty Problem in a Deep Excavation

in Clay, Proc. Int . Conf. on Case Histories in

Geotechnical EDgg., StLouis , Vol. 1, pp 315-323 .

Ahmad, A. S. and Peaker, K. R (1977). Geotechnical

Propert ies of Soft Marine Singapore Clay, Proc. In t .Symposium on sof t Clays, Bangkok. Thailand, pp 3-14.

Baguelin, F. • Jezequel, J. F. and Shields, D. H.

( 1978}. ''The Pressuremeter and Foundation

Engineering, Trans. Tech. Publications, ClausthalGermany, 617 pp. · '

Bjerrum, L. and Eide, 0. (1956). Stabi l i ty of Stru t tedExcavations in Clay, Geotechnique, Vol. 65, No. 1,pp. 32 - 47.

Broms, B. B. (1968). Swedish Tie-Back System for

Sheet Pi le Walls, Proc. 3rd Budapest Con£. on SoilMech. a . Found. Engg., Budapest, pp. 391 - 403.

Broms, B. B. (1984). Design of Sheet Pile Walls,

Proceedings of Seminar on Precautions and Solutionsto Environmental Problems Caused by Tunnelling and

Excavations organized by ABV LCM Engineering Pte Ltd,April 1984, Singapore, 27 pp.



Broms, B. B. andBennermark, H (1967). Stabi l i ty ofClay a t Vert ical Openings, Journ. Soil Mech. a .

Found. Div., ASCE, Vol. 93, No. SMI. pp 71 - 94.

Broms, B. B. and Bergdahl U. (1982). The Weight

Sounding Test (WST), State-of-the-Art Report, SOPT

II , Amsterdam, 12 pp.

Broms, B. B. and Bjerke, H. (1973). Extrusion of SoftClay Through a Retaining Wall, Canadian

Geotechnical Journal, Vol. 10, No. 1 pp 103 - 108.

Broms, B. B. and Wong, I . H (1984). Foundation ofTall Buildings, Proceedings of Int. Conf. on TallBuildings, Singapore, pp 547 - 590.

Broms, B. B., Wong, I . H. and Wong, K. S. (1986).Experience with Finite Element Analysis of Braced

Excavations in Singapore. Proc., 2nd Int. Symp. on

Numerical Models in Geomechanics, NUMOG I I , Ghent,

Belgium, pp 309 - 324.

Burland, J . B., Simpson, B. and St. John, H. D. (1979).

Movements Around Excavations in London Clay,

Proceedings of Seventh European Con£. on SoilMechanics and Foundation Engg., Vol. 1, Brighton, pp.

13 - 29.

Chang, C. S. and Duncan (1977). EXCAV : A Computer

Program for Analysis of Stresses and Movements in

Excavations, Geotechnical Report No TE 77-4,University of California , Berkeley, 35 pp.

·Clough, G. W. and Davidson, R. R. {1977). Effects ofConstruction on Geotechnical Performance,

Proceedings, Special ty Session I I I , Ninth Int. Con£.on Soil Mech. a. Found. Engg., Tokyo, 1977, pp. 1 5 -53.

Clough, G. W. and Hansen, L. A. (1981}. Clay

Anisotropy and Braced Wall Behaviour, Journ. of the

Geotechnical Div., ASCE, Vol. 107, No. GT 7, July,pp. 893 - 914.

Clough, G. W. and Mana, a. I . 1977). Lessons Learnt

in Finite Element Analysis of Temporary Excavations

in Soft Clay, Proc. 2nd Int. Con£. on Numerical

Methods in Geomechanics, Blacksburg, Virginia, ASCE,

Vol. 7, pp. 496- 510.

Clough, G. W and Tsui, Y (1974}. Performance ofTied-Back Walls in Clay, Journ. of the Geotechncial

Div., ASCE, Vol. 100, No. GT12, pp. 1259- 1274.

D' Appolonia, D. J . (1971). Effects of Foundation

Construction on Nearby Structures, Proc. 4th Pan

American Con£. on Soil Mech. a. Found. Engg., PuertoRico, Vol. 1, pp. 189- 236.

Duncan, J . M., Byrne, P., Wong, K. S. and Mabry, P.

(1980). Strength, Stress-Strain, and Bulk Modulus

Parameters for Fin i te Element Analysis of Stressesand Movements in Soil Mass . Dept of Civil Engg.,

University of California , Berkeley, 70 pp.

Egger, P. (1972}. Influence of Wall Sti f fness and

Anchor Prestressing on Earth Pressure Distr ibution,Fif th European Con£. on Soil Mech. a. Found. Engg.,

Vol. 1. Madrid, pp. 485 - 498.

1549

Hanna, T. H. {1982). Foundations in Tens ion- Ground

Anchors, Trans Tech Publ icat ions , McGraw Hill BookCo., Series on Rock and Soil Mech., Vol. 6,Clausthal , Germany, 573 pp.

Janbu, N. (1972). Earth Pressure Computations in

Theory and Pract ice , Proc. 5th European Con£. on

Soil Mech. a . Found. Engg., Vol. 1, Madrid, Spain,pp. 4 5 - 54.

Karlsrud, K. (1986). Performance Monitoring of DeepSupported Excavations in Soft Clay, Proc. 4th In t .

Geotechnical Seminar, Field Instrumentation andIn-Situ Measurements, Nanyang TechnologicalInsti tute, Singapore, pp. 187 - 202.

Karlsrud, K. DiBiagio, E., and Aas, G. (1980).

Experience with Slurry Walls in Soft Clay. Proc.,Symp. on Slurry Walls for Underground Transportat ion

Faci l i t ies , U.S. Dept. of Transportation, Report No.FHWA-TS-80-221, March, pp. 383 - 408.

Lee, S. L . Yong, K. Y., Karunaratne, G. P. and Chua,L. H. 1986). Field Insturmentation for a Stru t tedDeep Excavation in Soft Clay, Proc. 4th In t .

Geotechn. Seminar, Field Instrumentation and In-SituMeasurements, Nanyang Technological Inst i tu te ,

Singapore. pp. 183 - 186.

Littlejohn, G. S. (1970). Soil Anchors, Proc. Conf.

on Ground Engg. The Ins t i tu t ion of Civil Engineers,London, England, pp. 33 - 44.

Mana, A. I . and Clough, G. W. {1981). Clay Anisotropy

and Braced Wall Behavior, Journ. of the GeotechnicalDiv., ASCE, Vol. 107, No. GT 6, pp. 759 - 778.

Meyerhof, G. G. {1956). Penetrat ion Tests and Bearing

Capacity of Cohesionless Soi ls , Journ. Soil Mech. a .Found. Div., ASCE, Vol. 82, SMl, pp. 1 - 19.

Miki, G. (1985). Soil Improvement by Jet Grouting,Proc. 3rd Int. Seminar, Soil Improvement Methods,

Singapore, pp. 229 - 236.

O'Rourke, T. D. {1981). Ground Movements Caused by

Braced Excavations, Journ. of the Geotechnical

Div., ASCE, Vol. 107, No. GT 9, pp. 1159- 1178.

Peck, R. B. (1969). Deep Excavations and Tunnelling

in Soft Ground, Proc of Seventh In t . Con£ on SoilMech. a . Found. Engg., State-of-the-Art-Volume, pp.225 - 290. .

Rowe, P. W. (1957). Sheet Pile Walls in Clay, Proc.of Ins t i tu t ion of Civil Enggs., Vol. 7 , Jul 1957, pp.629 - 654, London, England.

Sahlstrllm, P. 0. and St i l le , H. {1979}. Forankrade

Sponter (Anchored Sheet Pile Walls), Swedish Council

for Building Research, Report No T30 1979,

Stockholm, Sweden, 128 pp.

Sti l le , H. (1976). Behaviour of Anchored Sheet PileWalls, Thesis, Royal Inst . of Tech., Dept. of Soiland Rock Mech., 192 pp.

Sti l le , H. and Broms, B. B. {1976). Load

Redistribution Caused by Anchor Failures in Sheet

Pile Walls, Proc. 6th European Con£. on Soil Mech.a. Found. Engg .• Vienna, Austria, Vol. 1.2, pp. 197- 200.



Stroh, D and Breth, H {1976). Deformation of DeepExcavations, Proc., Conf. on Numerical Methods inGeomechanics, Vol. II . Blacksburg, June, pp. 686 -

700.

Tan, S. B. (1970). Stabi l i ty Investigations of the

Whampoa Slip, Journ. South East Asia Geotechnical

Society, Vol. 1, No. 2, pp. 95 - 101.

Tan, S. L (1983). Geotechnical Properties and

Laboratory Testing of Soft Soils in Singapore, Int .

Seminar on Construction Problems in Soft Soi ls ,Nanyang Technological Ins t i tute , Singapore, 1 - 3December 1983, 7 pp.

Terzaghi, K and Peck, R B. {1967). Soil Mechanics

in E:ngg Practice , 2nd Edition, John Wiley and Sons,

New York, 729 pp.

van der Veen, C. and Boersma, L (1957). The Bearing

Capacity Predetermined by Cone Penetrat ion Tes t ,Proc. 4th Int. Conf. Soil Mech. a . Found. Engg., Vol.

2, pp. 72 - 75.

P 1515- Design and Contstruction of Anchored and Strutted Sheet Pile Walls iin Soft Clay.pdf

Documents

Transcript of P 1515- Design and Contstruction of Anchored and Strutted Sheet Pile Walls iin Soft Clay.pdf