The Design of Foundations for high rise buildings Harry Poulos.pdf

8/9/2019 The Design of Foundations for high rise buildings Harry Poulos.pdf

1/627

C I V I L E N G I N E E R I N G

doi: 10.1680/cien.2010.163.6.27

Keywordsbuildings, structures & design;

foundations; piles & piling

High-rise buildings are usually founded on some form of piled

foundation subject to a combination of vertical, lateral andoverturning forces. However, conventional methods for assessing

stability may not be adequate when designing such foundations

because they tend to focus on resistance under vertical loading. This

paper sets out an ultimate-limit-state approach for computer-based

design of pile foundation systems for high-rise buildings and provides

an example application on a 151-storey tower in South Korea.

The design offoundations forhigh-rise buildings

Proceedings of ICE

Civil Engineering 163 November 2010Pages 2732 Paper 10-00005

Harry PoulosBE, PhD, DSc Eng, AM,

FAA, FTSE

is a senior principal at CoffeyGeotechnics Pty. Ltd., Sydney, Australia

Conventional methods of assessing foun-dation stability tend to focus primarilyon foundation resistance under verticalloading but, for tall buildings, the resis-tance to combined vertical, lateral andmoment loadings must be considered.This paper sets out a limit-state designapproach for tall-building foundationsystems, with attention being focused onpiled and piled-raft foundation systems the predominant types currently used.

The key characteristics of piled rafts

are outlined and then the principlesof the design approach are set out.An example of the application of thisapproach is described for a 151-storeytower on reclaimed land in Incheon,South Korea (Figure 1).

Piled-raft foundation systems

In a piled-raft foundation system, thepiles provide most of the stiffness forcontrolling settlements at serviceabil-

Figure 1. Artists impression of the 601 m tall Incheon 151 Tower in Songdo, Korea, which is due forcompletion in 2015


2/628 PROCEEDINGS OF THE INSTITUTION OF CIVIL ENGINEERS CIVIL ENGINEERING, 2010, 163, No. CE6 ISSN 0965 089 X

POULOS

ity loads and the raft element providesadditional capacity at ultimate loading.A geotechnical assessment for design-ing such a foundation system thereforeneeds to consider not only the capacityof the pile and raft elements, but theircombined capacity and interaction underserviceability loading.

The most effective application of piledrafts occurs when the raft can provideadequate load capacity, but the settle-

ment and/or differential settlementsof the raft alone exceed the allowablevalues. Poulos (2001) has examined anumber of idealised soil profiles andfound that the following situations maybe favourable

n soil profiles consisting of relativelystiff clays

n soil profiles consisting of relativelydense sands.

It has been found that the perfor-mance of a piled-raft foundation can be

optimised by selecting suitable locationsfor the piles below the raft. In general,the piles should be concentrated in themost heavily loaded areas, while thenumber of piles can be reduced, or eveneliminated, in less heavily loaded areas(Horikoshi and Randolph, 1998).

Design requirements

Design issuesThe following issues usually need to be

addressed in the design of foundations

for high-rise buildings (Poulos, 2009).

n Ultimate capacity of the foundationunder vertical, lateral and momentloading combinations.

n Influence of the cyclic nature ofwind, earthquakes and wave loadings(if appropriate) on foundation capac-ity and movements.

n Overall settlements.n Differential settlements, both within

the high-rise footprint and betweenhigh-rise and low-rise areas.

n Structural design of the foundation

system, including the load-sharingamong the various components of thesystem (for example, the piles and thesupporting raft) and the distributionof loads within the piles. For this, and

most other components of design, itis essential to have close co-operationand interaction between geotechnicaland structural designers.

n Possible effects of externally-imposedground movements on the foundationsystem, for example movements aris-ing from excavations for pile caps oradjacent facilities.

n Earthquake effects, including theresponse of the structurefoundation

system to earthquake excitation andthe possibility of liquefaction in thesoil surrounding and/or supportingthe foundation.

n Dynamic response of the structure foundation system to wind-induced(and, if appropriate, wave) forces.

In this paper, attention will be concen-trated on the first five design issues.

Design criteriaIn limit state format (such as in the

Australian Piling Code AS2159-1995

(Standards Australia, 1995)), the ulti-mate limit state (ULS) criteria may beexpressed as follows

Rs* S*1

Rg* S*2

whereRs*is the design structuralstrength which is equal to s.Rus; Rg*isthe design geotechnical strength which

is equal to g.Rug; Rusis the ultimatestructural strength; Rugis the ultimategeotechnical strength (capacity); sisthe structural reduction factor; gis thegeotechnical reduction factor and S*=isthe design action effect (factored loadcombination).

The above criteria are applied to theentire foundation system, while thestructural strength criterion (Equation 1)is also applied to each individual pile.However, it is not good practice to applythe geotechnical criterion (Equation 2) toeach individual pile within the group, as

this can lead to considerable over-design(Poulos, 1999).

Rs*and Rg*can be obtained fromthe estimated ultimate structural andgeotechnical capacities, multiplied by

appropriate reduction factors. Values ofthe structural and geotechnical reductionfactors are often specified in nationalcodes or standards. The selection of suit-able values of grequires the designer toexercise judgement and take into accounta number of factors that may influencethe foundation performance.

Load combinationsThe required load combinations for

which the structure and foundationsystem have to be designed will usuallybe dictated by a structural loading codeand a large number of combinations mayneed to be considered.

For example, for the Emirates Towersproject in Dubai (Poulos and Davids,2005), 18 load combinations wereanalysed for each tower, these being oneloading set for the ultimate dead and liveloading only, four groups of four loadingsets for various combinations of ultimatedead, live and wind loads and one set forthe long-term serviceability limit state

(dead plus live loading).

Design for cyclic loadingIn addition to the normal design cri-

teria, Poulos and Davids (2005) havesuggested an additional criterion for thewhole foundation of a tall building, tocater for the effects of repetitive loadingfrom wind action, as follows

.Rgs* Sc*3

whereRgs*is the design geotechnicalshaft capacity; Sc*is the half amplitudeof cyclic axial wind-induced load and isa factor assessed from geotechnical labo-ratory testing.

This criterion attempts to avoid thefull mobilisation of shaft friction alongthe piles, thus reducing the risk of cyclicloading leading to degradation of shaftcapacity. For the Emirates Towers proj-ect, was selected as 0.5 based on labo-ratory tests. Sc*can be obtained fromcomputer analyses which give the cycliccomponent of load on each pile for vari-

ous wind loading cases.

Serviceability limit stateThe design criteria for the serviceabil-

ity limit state (SLS) are as follows


3/629ISSN 0965 089 X PROCEEDINGS OF THE INSTITUTION OF CIVIL ENGINEERS CIVIL ENGINEERING, 2010, 163, No. CE6

THE DESIGN OF FOUNDATIONSFOR HIGH-RISE BUILDINGS

maxall4

maxall5

wheremaxis the maximum computed set-tlement of foundation; allis the allowablefoundation settlement; maxis the maxi-mum computed local angular distortionand allis the allowable angular distortion.

The values of alland all depend onthe nature of the structure and the sup-porting soil. Some suggested criteriaare reported by Zhang and Ng (2006).Criteria specifically for very tall buildingsdo not appear to have been set but it maybe unrealistic to impose very stringentcriteria on tall buildings on clay deposits,as these may not be achievable.

Experience with tall buildings inFrankfurt, Germany suggests that total set-tlements well in excess of 100 mm can betolerated without any apparent impairmentof function (Katzenbach et al., 2000).

Analysis methods

Overall stabilityFor consideration of the overall stabil-

ity of the foundation system, the ULSload combinations are applied and theanalysis uses geotechnical and structuralresistances of the foundation compo-nents, which are reduced by a geotech-nical reduction factor and a structuralreduction factor respectively.

The design requirements in Equations

1 and 2 will be satisfied if the foundationsystem does not collapse under any ofthe sets of ULS loadings. In addition, acheck can be made of the cyclic actionsgenerated in the foundation elements toassess whether the cyclic loading require-ment (Equation 3) is satisfied.

If any of the above requirements are notsatisfied, then the design will need to bemodified to increase the strength of theoverall system or of those components ofthe system that do not satisfy the criteria.

Serviceability

For the serviceability analysis, the SLSloads are applied and the best-estimate(unfactored) values of foundation resis-tances and stiffnesses are employed.

The design will be satisfactory if the

computed deflections and rotations arewithin the specified allowable limits(Equations 4 and 5).

Structural design requirementsFor structural design of the raft and

piles, the results of ULS analysis are notconsidered relevant because the loadsthat can be sustained by the piles areartificially reduced by the geotechnicalreduction factor and the worst response

may not occur when the pile and raftcapacities are factored downwards.Consequently, the most rational

approach appears to be one in whicha separate ULS analysis is carried outusing the various ULS load combina-tions but in which the unfactored resis-tances of the foundation components areemployed. The consequent computedfoundation actions (i.e. pile forces, raftmoments and shears) are then multipliedby a structural action factor (for example1.5) to obtain values for structuraldesign.

Analysis program requirementsTo undertake the above analyses, a

computer program should ideally havethe following abilities. For overall stabil-ity, it should have the ability to consider

n non-homogeneous and layered soilprofiles

n non-linearity of pile and, if appropri-ate, raft behaviour

n geotechnical and structural failure ofthe piles (and the raft)

n

vertical, lateral and moment loading(in both lateral directions), includingtorsion

n piles having different characteristicswithin the same group.

For serviceability analysis, the abovecharacteristics are also desirable, togetherwith the ability to consider

n pilepile interaction, and if appropri-ate, raftpile and pileraft interaction

n flexibility of the raft or pile capn some means by which the stiffness of

the supported structure can be takeninto account.

There do not appear to be any com-mercially available software packages

that have all of the above desirable char-acteristics, other than three-dimensionalfinite-element packages such as Plaxis3D (Plaxis, 2009), or the finite-differ-ence program Flac3D (Itasca, 2009). Thecommercially available programs Repute(Geocentrix, 2009), Piglet (Randolph,2004) and Defpig (Poulos, 1990) havesome of the requirements, but fall shortof a number of critical aspects, particu-larly in their inability to include raftsoil

contact and raft flexibility.The author has developed pile-groupanalysis packages that, between them,provide most of the features listed above.The programs include Pigs, which analy-ses the settlement and load distributionwithin a group of piles subjected to axialand moment loading; Clap, which com-putes the distributions of axial and lateraldeflections, rotations, loads and momentsat the top of a group of piles subjectedto a combination of vertical load, lateralloads and moments; and Garp, whichanalyses the behaviour of a piled raft

subjected to vertical and moment loading(Small and Poulos, 2007).

Application to Incheon Tower, Korea

A 151-storey super-high-rise buildingproject is currently under constructionon reclaimed land on soft marine clay inSongdo, Incheon, Korea. Due for comple-tion in 2015, the 601 m tall 151 IncheonTower is illustrated in Figure 1 and thegeotechnical aspects are described indetail by Badelow et al. (2009).

Ground conditions and geotechnical modelThe site lies entirely within an area of

reclamation, which typically comprisesapproximately 8 m of loose sand andsandy silt, constructed over approxi-mately 20 m of soft to firm marine siltyclay, referred to as upper marine depos-its. These are underlain by approximately2 m of medium dense to dense silty sand,referred to as lower marine deposits,which overlie residual soil and a profileof weathered rock.

It is intended that the fill will be

improved to reduce the risk of liquefac-tion during earthquakes. However, asthe tower will be founded on the uppermarine deposits liquefaction will not be adirect issue.



The lithological rock units withinabout 50 m of the surface have beenaffected by weathering, which hasreduced their strength to a very weakrock or a soil-like material. This depthincreases where the bedrock is intersect-ed by closely spaced joints and shearedand crushed zones that are often relatedto the existence of roof-pendant sedimen-tary and metamorphic rocks. The geolog-ical structures at the site are complex and

comprise geological boundaries, shearedand crushed seams possibly related tofaulting movements and jointing.

From the available borehole data forthe site, inferred contours were devel-oped for the surface of the soft rockfounding stratum within the tower foun-dation footprint. These are reproducedin Figure 2. It can be seen that there is apotential variation in level of the top ofthe soft rock (the pile founding stratum)of up to 40 m across the foundation.

The footprint of the tower was dividedinto eight zones which were consideredto be representative of the variation ofground conditions and geotechnicalmodels were developed for each zone.Appropriate geotechnical parameterswere selected for the various strata basedon the available field and laboratory testdata, together with experience of similarsoils on adjacent sites. One of the criticaldesign issues for the tower foundation

was the performance of the soft uppermarine deposits under lateral and verticalloading. Typical parameters adopted forthe foundation design are presented inTable 1.

Foundation layoutThe foundation system considered

comprises 172 no. 2.5 m diameter boredpiles, socketed into the soft rock layerand connected to a 5.5 m thick raftsupporting columns and core walls. The

number and layout of piles and the pilesize were obtained from trial analysesthrough collaboration between thegeotechnical and structural designers.

The pile depth was evaluated by thegeotechnical engineer, considering theperformance and capacity of piles. Thepile layout was selected from the variousoptions considered, and is presented inFigure 3.

LoadingsTypical loads acting on the tower wereas follows

n Vertical dead plus live load: Pz(DL+LL)= 6622 MN

n Horizontal wind loads: Px(WL)=146 MN, Py(WL)= 112 MN

n Horizontal earthquake loads: Px(E)=105 MN, Py(E)= 105 MN

n Wind load moments: Mx(WL)=12 578 MN.m, My(WL)= 21 173 MN.m

n Wind load torsional load: Mz(WL)=1957 MN.m

The vertical load Pz(DL+LL)and over-turning moments Mx(WL)andMy(WL)wererepresented as vertical load componentsat column and core locations. The loadcombinations, as provided by the struc-tural designer, were adopted throughoutthe geotechnical analysis and 24 wind-load combinations were considered.

Assessment of pile capacitiesThe required pile length was assessed

from the estimated shaft friction and

end bearing capacities. For a largepile group founding in weak rock, theoverall settlement behaviour of the

POULOS

m0 20

Figure 3. Layout of the 172 no. 2.5 m diameterbored piles and 5.5 m thick raft

50m

50m

40m

50m

50m

40m

60m

60m

70m

70m

70m80m

80m

Boreholes

Figure 2. Borehole-derived contours of the soft rock surface under the IncheonTower foundation the depth varies from 40 to 80 m

Table 1. Summary of geotechnical parameters

Strata Ev: MPa Eh : MPa f s: kPa f b: MPa

Upper marinedeposits

715 511 2948

Lower marinedeposits

30 21 50

Weathered soil 60 42 75

Weathered rock 200 140 500

Soft rock (aboveelevation 50m)

300 210 750 12

Soft rock (belowelevation 50m)

1700 1190 750 12

Ev= vertical modulus, fs = ultimate shaft f riction, Eh= horizontal modulus, fb= ultimate end bearing


5/631ISSN 0965 089 X PROCEEDINGS OF THE INSTITUTION OF CIVIL ENGINEERS CIVIL ENGINEERING, 2010, 163, No. CE6

THE DESIGN OF FOUNDATIONSFOR HIGH-RISE BUILDINGS

pile group controlled the required pilelengths rather than the overall geotech-nical capacity.

In this case, the soft rock layer wasconsidered to be a more appropriatefounding stratum than the overlyingweathered rock. In particular the softrock below elevation 50 m providesa more uniform stiffness and is likelyto result in more consistent settlementbehaviour of the foundation.

The basic guidelines to establish thepile founding depth were

n minimum socket length in soft rock =two diameters

n minimum toe level = elevation 50m.

The pile design parameters for theweathered/soft rock layer are shown inTable 2 and were estimated on the basisof the pile test results in the adjacent siteand the ground investigation data, suchas pressuremeter tests and rock corestrength tests.

Pile load tests prior to commencementof the main piling works are planned.Based on the interpreted findings ofthese tests, the pile capacities will beverified and the pile design confirmed.

Overall stabilityThe ULS combinations of load were

input into a series of Clap computerprogram analyses, with the pile axial andlateral capacities reduced by geotechni-cal reduction factors (0.65 for axialload, 0.40 for lateral load). The smaller

factors for lateral load reflected thegreater degree of uncertainty for lateralresponse.

In all cases analysed, the foundationsystem was found to be stable, that isthe computed foundation movementswere finite and generally the maximumcomputed settlement under ULS loadingswas less than 100 mm.

Cyclic loading assessmentTable 3 summarises the results of the

cyclic loading assessment. Figure 4 showsthe assessed factor for each pile within

the foundation system.The assessment indicates that the cri-

terion in Equation 3 is satisfied and thatdegradation of shaft capacity due to cyclicloading in unlikely to occur.

Assessment of vertical pile behaviourThe vertical pile head stiffness values

for each of the 172 foundation pilesunder serviceability loading (dead andlive load) were assessed using the Clapprogram, which provided the geotechni-cal capacities, interaction factors andstiffness values for each pile underserviceability loading for input into the

group assessment.Table 4 presents a summary of theassessment, including the average verti-cal pile stiffness values. The pile stiffnessvalues and interaction factors were thenused in the Garp program.

SettlementThe maximum settlements predicted

were about 50 mm (using the upperbound modulus) with differential settle-ments between the centre and edge ofthe slab of about 30 mm.

For the tower alone, an elastic analysis

using an equivalent block approach viaa two-dimensional axi-symmetric finite-element analysis gave maximum andminimum settlement values of 43 mmand 18 mm, respectively, which were ingood agreement with those predictedby one of the project reviewers usingan equivalent block hand calculationapproach (43 mm and 19 mm respec-tively for the tower alone).

The predicted settlements using thevarious analysis methods were reasonablyconsistent, giving a level of confidence to

Pile number

1 917

25

33

41

49

57

65

73

81

89

97

105

113

121

129

137

145

153

161

169

0.50

0.45

0.40

0.35

0.30

0.25

0.20

0.15

0.10

0.05

0

R

ation

Figure 4. Assesed values of for each pile rom cyclic loading analysis using a load case 0 .75 (dead load +live load + wind load)

Table 2. Ultimate capacities for pile analysis

Material Ultimate friction, f s: kPa Ultimate end bearing, f b: MPa

Weathered rock 500 5

Soft rock 750 12

Table 4. Summary of vertical pile stiffness values

Foundation

layout

Minimum vertical pile

stiffness: MN/m

Maximum vertical pile

stiffness: MN/m

Average vertical

pile stiffness: MN/m

172 piles 600 1300 820

Table 3. Summary of cyclic loading assessment

Quantity Value

Maximum cyclic axial load Sc*:MN 74.00Maximum ratio, = Sc*/Rgs* 0

.43

Cyclic loading criterion satisfied? Yes

Figure 5. Typical set of computed settlementcontours for the IncheonTower

10

20

30

40

50

60

70

80

90

100

Settlement: mm

Scale: 20 m



POULOS

the settlement values predicted. Figure 5shows a typical set of settlement contoursderived from the Garp analyses, whileFigure 6 show the computed distributionof pile loads.

Assessment of lateral pile behaviourA critical design issue for the tower

foundation was the performance of thepile group under lateral loading. Thenumerical modelling packages used in

the analyses comprised the following

n three-dimensional finite-element com-puter program Plaxis 3D Foundation

n computer program Clap.

Plaxis 3D Foundation provided an assess-ment of the overall lateral stiffness of thefoundation. Clap was used to assess thelateral stiffness provided by the pile groupassuming that the raft is not in contact withthe underlying soil and then a separate cal-culation was carried out to assess the lateralstiffness of the raft and basement.

Table 5 presents typical values of com-puted lateral stiffness for the piled-matfoundation obtained from the analyses.

Conclusion

This paper has set out an approachfor the design of pile foundation systemsfor high-rise buildings using a limit-statedesign approach. This approach involvesthree sets of analyses.

n An overall stability analysis in whichthe resistances of the foundation com-ponents are reduced by the appropriategeotechnical reduction factor and theULS load combinations are applied.The design requirements are satisfied ifthe foundation system does not collapseunder any of the sets of ULS loadings.

n A serviceability analysis, in which thebest-estimate (unfactored) values offoundation resistances and stiffnesses

are employed and the SLS loads areapplied. The design is satisfactory if thecomputed deflections and rotations arewithin the specified allowable limits.

n For structural design of the raft andthe piles, the results of the above ULSanalysis are not considered to be relevantbecause the loads that can be sustainedby the piles are artificially reduced bythe geotechnical and structural reduc-tion factors. The most rational approachappears to be to carry out a separateULS analysis in which the ULS loadcombinations are applied but in which

the unfactored resistances of the foun-dation components are employed. Theconsequent computed foundation actions(i.e. pile forces and, if appropriate, raftmoments and shears) are then multipliedby a structural action factor to obtain thevalues for structural design.

In addition, a check can be carried outto assess the ratio of cyclic load amplitudeto factored-down pile shaft resistance. It issuggested that if this ratio for a pile is lessthan about 0.5, there should be a low risk

of cyclic degradation of shaft resistanceoccurring.For the 151 Incheon Tower in Korea, the

process was used to assess overall stability,foundation settlements and pile head stiff-ness values as part of the design process.

Acknowledgements

The author gratefully acknowledges thecontributions of Frances Badelow, TristanMcWilliam, Helen Chow and Patrick Wongin relation to the analyses for the tower

described in the paper. S.H. Kim, AhmadAbdelrazaq and their teams in Korea hada major involvement in the Incheon Towerfoundation design. Professor John Small hasbeen pivotal in developing the Garp programand implementing it in a user-friendly form.

What do you think?If you would like to comment on this paper,

please email up to 200 words to the editor at

[email protected] you would like to write a paper of 2000 to 3500

words about your own experience in this or any

related area of civil engineering, the editor will be

happy to provide any help or advice you need.

ReferencesBadelow F, Kim SH, Poulos HG and Abdelrazaq

A (2009) Foundation design for a tall tower ina reclamation area. In 7th Int. Conf. Tall Buildings,Hong Kong, (Au FTK (ed.), Research Publishing,Hong Kong, pp. 815823.

Geocentrix (2009) Repute 2 user manual.Geocentrix, Surrey, UK.

Horikoshi K and Randolph MF (1998) Acontribution to the optimum design of piledrafts. Gotechnique48(2): 301317.

Itasca (2009)Users manual, FLAC3D 4.00.49. ItascaConsulting Group Inc., Minneapolis, USA.

Katzenbach R, Arslan U and Moorman C (2000)Piled raft foundation projects in Germany.In Design Applications of Raft Foundations,(Hemsley JA (ed.), Thomas Telford, London,pp. 323391.

Plaxis (2009) Users manual, PLAXIS 3D. Plaxis bv,Delft, the Netherlands.

Poulos HG (1990) DEFPIG users manual. Centre

for Geotechnical Research, University ofSydney, Australia.Poulos HG (1999) The design of piles with

particular reference to the Australian PilingCode.Australian Geomechanics34(4): 2539.

Poulos HG (2001) Piled rafts design andapplications. Gotechnique51(2): 95113.

Poulos HG (2009) Tall buildings and deepfoundations Middle East challenges. In Proc.17thInt. Conf. Soil Mechs. Geot. Eng., Alexandria,(Hamza M, Shahien M and El-Mossalamy Y(eds), IOS Press, Amsterdam, 4: 31733205.

Poulos HG and Davids AJ (2005) Foundationdesign for the Emirates twin towers, Dubai.Canadian Geotechnical Journal42(3): 716730.

Randolph MF (2004) PIGLET 5.1 users manual.University of Western Australia, Perth,Australia.

Small JC and Poulos HG (2007) A method of

analysis of piled rafts. Proc. 10thAustralia NewZealand Conf. on Geomechanics, Brisbane, 1:550555.

Standards Australia (1995)AS2159 AustralianStandard Piling Design and Installation. StandardsAustralia, Sydney.

Zhang L and Ng AMY(2006) Limiting tolerablesettlement and angular distortion for buildingfoundations. Geotechnical Special PublicationNo. 170, Probabilistic Applications in GeotechnicalEngineering, ASCE (on CD Rom).

Table 5. Summary of lateral stiffness of pile group and raft

Horizontalload: MN

Direction Pile groupdisplacement: mm

Lateral pilestiffness: MN/m

Lateral raftstiffness: MN/m

Total lateralstiffness: MN/m

149 X 17 8760 198 8958

115 Y 14 8210 225 8435

Figure 6. Computed distribution of axial pile loadsfor the Incheon Tower

Documents

The Design of Foundations for high rise buildings Harry Poulos.pdf