RE-ASSESSMENT OF FOUNDATION SETTLEMENTS FOR THE BURJ KHALIFA, DUBAI

Gianpiero Russo1

Harry G. Poulos2

John C. Small3

ABSTRACT: This paper deals with the re-assessment of foundation settlements for the Burj Khalifa Tower in Dubai. The foundation system for the tower is a piled raft, founded on deep deposits of calcareous rocks. Two computer programs, GARP and NAPRA have been used for the settlement analyses, and the paper outlines the procedure adopted to re-assess the foundation settlements, based on a careful interpretation of load tests on trial piles in which the interaction effects of the pile test setup are allowed for. It then examines the influence of a series of factors on the computed settlements. In order to obtain reasonable estimates of differential settlements within the system, it is found desirable to incorporate the effects of the superstructure stiffness which act to increase the stiffness of the overall foundation system. Values of average and differential settlements for the piled raft calculated with GARP and NAPRA were found to be in reasonable agreement with measured data on settlements taken near the end of construction of the tower.

Key words: case history, footings and foundations, full-scale tests, piles, rafts, settlement.s 730INTRODUCTION

The Burj Khalifa in Dubai was officially opened in January 2010, and at a height of 828m, is currently the world’s tallest building. The foundation system is a piled raft, a form of foundation that is being used increasingly to support tall structures where the loads are expected to be excessively large for a raft alone and where the raft and the piles are able to transfer load to the soil. The foundation design process for this building has been described by Poulos and Bunce (2008). An important component of the design of a piled raft foundation is the detailed assessment of the settlement and differential settlement of the foundation system, and their control by optimizing the size, location and arrangement of the piles, and the raft thickness. Many different methods of analysis have been devised in order to predict the behaviour of raft and piled raft foundations (Selvaduri, 1979; Clancy and Randolph, 1993; Poulos, 1994; Ta and Small, 1996; Russo and Viggiani, 1998; Viggiani, 1998; Hemsley, 1998; Hemsley, 2000), and these range from simple hand based methods to complex three-dimensional numerical analyses.

In this paper, attention is focussed on two methods that model the raft as an elastic plate and the piles as interacting non-linear springs. The computer codes implementing these methods are described very briefly and are then applied to the Burj Khalifa, currently the world’s tallest building, which is founded on a piled raft. The development of the ground modulus values is described using a combination of field test and laboratory data and the results of pile load tests. The method of interpreting the pile load test data is discussed, and the importance of allowing for interaction between the test pile 1 University of Naples, Italy2 Coffey Geotechnics, Sydney, Australia3 Coffey Geotechnics and University of Sydney

1 11 July 2011

and the surrounding reaction piles in emphasised. The two programs are then used to compare the computed settlements with available measurements of foundation settlements, and with the “Class A” predictions made by the foundation designers and the peer reviewers.

An important objective of the paper is to explore how pile load test data should be used when predicting the settlement performance of piled and piled raft foundation systems, and to examine some factors that may have an important influence on predicted foundation settlements. COMPUTER ANALYSES

The settlement analyses used in this paper for the Burj Khalifa have employed two computer programs, GARP and NAPRA, which idealize the piled raft foundation as a plate supported by non-linear interacting springs. A very brief description of these programs is given below.

Program GARP

The computer program GARP (Geotechnical Analysis of Raft with Piles, Small and Poulos, 2007) uses a simplified boundary element analysis to compute the behaviour of a piled raft when subjected to applied vertical loading, moment loading, and free-field vertical soil movements.The raft is represented by a thin elastic plate and is discretized via the finite element method, using 8- noded elements. The soil is modelled as a layered elastic continuum, and the piles are represented by elastic–plastic or hyperbolic springs, which can interact with each other and with the raft. Pile–pile interactions are incorporated via interaction factors (Poulos and Davis, 1980). Simplifying approximations are utilized for the raft-pile and pile-raft interactions. Beneath the raft, limiting values of contact pressure in compression and tension can be specified so that some allowance can be made for nonlinear raft behaviour. The output of GARP includes the settlement at all nodes of the raft; the transverse, longitudinal, and torsional bending moments within each element of the raft; the contact pressures below the raft; and the vertical loads on each pile. In its present form, GARP can consider vertical and moment loadings, but not lateral loadings or torsion.

Program NAPRAThe computer program NAPRA (Non linear Analysis of Piled Rafts, Russo 1998; Russo & Viggiani, 1998) computes the behaviour of a raft subjected to any combination of vertical distributed or concentrated loading and moment loading. The raft is modelled as a two-dimensional elastic body using the thin plate theory, and utilizing the finite element method, adopting a four or nine noded rectangular element. The piles and the soil are modelled by means of interacting linear or non-linear springs. It is assumed that the interaction between the raft and the soil (the piles) is purely vertical; accordingly, only the axial stiffness of the springs is required.

The soil is assumed to be a layered elastic continuum. The Boussinesq solution for a point load and the closed form solution for a rectangular uniformly loaded area at the surface of an elastic half-

2 11 July 2011

space are used to calculate the soil displacements produced by the contact pressure developed at the interface between the raft and the soil. The layered continuum is solved by means of the Steinbrenner approximation (Russo, 1998; De Sanctis and Russo, 2002), and as such, invokes the simple assumption that the stress distribution within an elastic layer is identical with the Boussinesq distribution for a homogeneous half-space (Russo, 1998).

The interaction factor method is used to model pile to pile interaction and a preliminary boundary element (BEM) analysis allows calculation of the interaction factors between two piles at various spacings. Interaction between axially loaded piles beneath the raft and the raft elements is accounted for via pile-soil interaction factors computed with a preliminary BEM procedure. The reciprocal theorem is used to maintain that the soil-pile interaction factor is equal to the pile-soil interaction factor.

A stepwise incremental procedure is used to simulate the non-linear load-settlement relationship of a single pile, the total load to be applied is subdivided into a number of increments, and the diagonal terms of the pile-soil flexibility matrix are updated at each step. A computation of the nodal reactions vector is made at each step, to check for tensile forces between raft and soil and an iterative procedure is used to make them equal to zero. Basically, this procedure releases the compatibility of displacements between the raft and the pile-soil system in the node where tensile forces were detected, although the overall equilibrium is maintained by a re-distribution of forces. An iterative procedure is needed since after the first run some additional tensile forces may arise in different nodes. The output of the code is represented by the distribution of the nodal displacements of the raft and the pile-soil system, the load sharing among the piles and the raft, the bending moments and the shear in the raft, for each load increment.

Abagnara et al (2011) have compared GARP and NAPRA analyses for a simple case, and have concluded that both programs give comparable results, but that some of the simplifying assumptions employed in each program give rise to differences in detail. For example, the difference in raft settlements may be due to the differences in the details of calculation of the soil layer stiffness using the Boussinesq/Steinbrenner approach. The difference in plate element types may also contribute to the differences. For the piled raft, the differences may arise because of differences in the methods used to compute the single pile stiffness values, the interaction factors and the pile-raft and raft-pile interactions.

In this paper, attention will be focussed on analyses carried out with NAPRA, although a comparison will also be presented between the GARP and NAPRA analyses.

SETTLEMENT ASSESSMENT FOR THE BURJ KHALIFA TOWER, DUBAI

Foundation layout

The Burj Khalifa project in Dubai, United Arab Emirates (UAE), comprises a 160 storey high rise tower, with a podium development around the base of the tower, including a 4-6 storey garage. The Burj Khalifa is located on a 42000 m2 site. The tower is founded on a 3.7m thick raft supported on 194 bored piles, 1.5 m in diameter, extending 47.45m below the base of the raft; podium structures

3 11 July 2011

are founded on a 0.65 m thick raft (increased to 1m at column locations) supported on 750 bored piles, 0.9 m in diameter, extending 30-35 m below the base of the raft. A plan view of foundation is shown in Fig. 1.

Figure 1. plan view of the Khalifa Tower foundation system

Ground investigation and site characterization

The investigations involved the drilling of 32 boreholes to a maximum depth of about 90 m below ground level and 1 borehole to a depth of 140 m under the tower footprint. Standpipe piezometers were installed to measure the ground water level which was found to be relatively close to the ground surface, typically at a level of 2.5m DMD. The ranges of measured SPT N values are summarised in Table 2. There was a tendency for N values to increase with depth, beyond an elevation of about -8m DMD.

Table 2 Summary of Measured SPT Values

Elevation m Range of SPT Values2.5 to -1 0-40-1 to -8 50-400-8 to -14 0-100-14 to -30 40-200-30 to -40 100-200-40 to -80 100-400

The ground conditions at the site comprise a horizontally stratified subsurface profile which is complex and highly variable in terms of the strata thickness due to the nature of deposition and the prevalent hot arid climatic conditions. The main strata identified were as follows

1. Very loose to medium dense silty sand (Marine Deposits). 2. Weak to moderately weak calcarenite, generally unweathered with fractures close to medium

spaced interbedded with cemented sand. This material is generally underlain by very weak to weak sandstone which is generally unweathered with fractures close to medium spaced interbedded with cemented sand.

3. Very weak to weak calcarenite, calcareous sandstone and sandstone; this formation is slightly to highly weathered with fractures extremely close to closely spaced and interbedded with

4 11 July 2011

cemented sand. Bands of 1 to 5 m thickness are also present of medium dense to very dense, cemented, sand with sandstone bands and locally with bands of silt.

4. Very weak to weak gypsiferous sandstone, gypsiferous calcareous sandstone occasionally gypsiferous siltstone. This material is generally unweathered to slightly weathered with fractures extremely close to closely spaced and interbedded with cemented sand. The formation is interbedded with dense to very dense, cemented, silty sand and occasionally silt with sandstone bands.

5. Very weak to weak calcisiltite, conglomeritic calcisiltite and calcareous calcisiltite. This material is generally moderately to highly weathered, occasionally slightly and completely weathered with fractures extremely close to medium spaced. Calcareous siltstone was encountered in the majority of the deeper boreholes comprising very weak to weak occasionally moderately weak calcareous siltstone in bands with a thickness of 0.5 to 14.4 m generally slightly to moderately weathered occasionally highly to extremely weathered.

6. Very weak to weak and occasionally moderately weak calcareous siltstone, calcareous conglomerate, conglomeritic sandstone and limestone. This material is generally slightly weathered and occasionally unweathered and moderately weathered to highly weathered. Occasionally encountered as calcisiltite interbedded with bands of siltstone and conglomerate.

7. Very weak to moderately weak claystone interbedded with siltstone. This material is generally slightly weathered with close to medium spaced fractures. Between -112.2 m and -128.2 m occasional bands of up to 500 mm thick gypsum were encountered. Below -128.2 m the stratum was encountered as weak to moderately weak siltstone with medium to widely spaced fractures.

Table 3 summarizes the stratigraphy adopted for the foundation settlement analyses.

In situ and laboratory test results

A comprehensive series of in situ tests was carried out, including pressuremeter tests, down-hole seismic, cross-hole seismic, and cross-hole tomography to determine compression (P) and shear (S) wave velocities through the ground profile. The vertical profile of P-wave velocity with depth gave a useful indication of variations in the nature of the strata between the borelogs.

Conventional laboratory classification tests (moisture content of soil and rock, Atterberg limits, particle size distribution and hydrometer) and laboratory tests for determining physical properties (porosity tests, intact dry density, specific gravity, particle density) and chemical properties were carried out. In addition, unconfined compression tests, point load index tests, and drained direct shear tests were carried out. A considerable amount of more advanced laboratory testing was undertaken, including stress path triaxial tests, resonant column testing for small-strain shear modulus, undrained cyclic triaxial tests, cyclic simple shear, and constant normal stiffness (CNS) direct shear tests.

Table 3. Stratigraphic model adopted for settlement assessment.

Stratum

Description Level at the top

of the stratum

[m DMD]

Thickness

[m]

AdoptedLevel at top of layer

[m DMD]

UCSqu

[MPa]

1 Marine deposits

1.15 to 2.96 1.85 to 4.3 2.5

5 11 July 2011

2Calcarenite/Calcareous sandstone

-0.27 to -1.95 2.87 to 10.75 -1.2 2

3a Calcareoussandstone/Sandstone

-4.13 to -12.06 10.5 to 21.43-7.3 -

3b -13.5 1

4 Gypsiferous sandstone

-21.54 to -26.69

1.7 to 7.75 -24 2

5a

Calcisiltite/Conglomeritic calcisiltite -27.64 to -

31.1539.2 to 46.75

-28.5 1.3

5b Calcareous siltstone

-50 1.7

6Calcareous/

Conglomeritic Strata

-67.19 to -76.04

31 (from 140m deep BH only) -68.5 2.5

7

Claystone/ Siltstone

interbedded with gypsum

layers

-98.19 Proved to 39.6 m thickness

-90 -

Some of the relevant findings from the in situ and laboratory testing are as follows:

i. The cemented materials were generally very weak to weak; unconfined compressive strength (UCS) values ranged mostly between about 0.1–6 MPa the average values for each layer being the ones reported in the table 3.

ii. Values of the Young’s Modulus from pressuremeter tests (first and second reload cycle) were found to be in good agreement with values from correlation with shear waves velocities. From calcarenite (0 m to -7.5 m) to sandstone (-7.5 m to -24 m) Young’s Modulus is approximately constant with depth; at greater depths the average values decrease in the gypsiferous sandstone (-24 to -28.5 m) then they slightly increase in the calcisiltite (from -28.5 to -68.5 m) and finally decrease in the siltstone (from -68.5 to -91 m).

iii. Triaxial Stress Path Testing (at strain levels of 0.01% and 0.1%) was found to give results for Young’s modulus that were in good agreement with pressuremeter and geophysics testing results.

iv. Resonant Column Testing was found to give more conservative values for the Young’s Modulus when compared with values from pressuremeter tests, geophysics tests and triaxial stress tests.

v. Constant normal stiffness (CNS) tests were carried out on three samples taken from stratum 5a to assess the ultimate skin friction values and the potential for cyclic degradation at the pile-soil interface. These tests indicated values of peak monotonic shear stress ranging from 360 to 558 kPa, with only a little difference between the peak monotonic and the residual cyclic shear stress values.

Geotechnical Model

The key parameters for the assessment of the settlement behaviour of the Khalifa Tower piled raft foundation system are the values of the Young’s modulus of the strata for both raft and pile behaviour under static loading. In a non-linear analysis, the values of ultimate skin friction of piles, the ultimate end-bearing resistance of the piles, and the ultimate bearing capacity of the raft would also be required, but in this paper, only linear elastic analyses have been undertaken using NAPRA

6 11 July 2011

and GARP analyses, having explored the little influence of non linearity up to the maximum observed load level. Attention has thus been focussed on evaluating relevant values of Young’s modulus for each stratum.As a first step in obtaining these values, the relative stiffness of the various soil layers was assessed considering values of the Young’s Modulus from the following data:

1. Pressuremeter tests (initial loading, first reload, second reload cycles);2. Geophysics tests (correlation with shear wave velocities);3. Resonant column tests (Initial, 0.0001%, 0.001%, 0.01% strain levels);4. Triaxial Stress Path Tests (0.01% and 0.1% strain levels);

Values of the various Young’s Modulus values are plotted in Fig. 2, and although inevitable scatter exists among the different values, there is a reasonably consistent general pattern of variation with depth.

Layer 3b (see Table 3) has arbitrarily been chosen as the reference layer, and for each type of test, values of the Young’s Modulus for a layer i, Ei, have been related to the value for layer 3b, E3b. The values of Ei/E3b have then been averaged, using the following data: reload cycles from pressuremeter testing; seismic data; resonant column data at a strain level of 0.01%, and the triaxial stress path tests. Fig. 3 shows the different assessed relative stiffness profiles so obtained, and Table 4 summarises the average values of relative Young’s modulus that were adopted for the analyses and the interpretation of the pile load test data. The absolute values of Young’s modulus for each of the different layers have been then obtained by fitting the load settlement curves of the single piles obtained from the load tests, and the process of fitting the load-settlement curves to obtain the Young’s modulus values is described below.

Table 4. Relative Values of Young’s Modulus Used in Pile Load Test Interpretation

Stratum Young’s Modulus, Relative to Value for Layer 3b

23a3b3c45a5b6

2.30.61.01.00.80.70.80.7

7 11 July 2011

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Young's Modulus [MPa]

De

pth

[m

]

Press Init AVE

Press First AVE

Press Second AVE

Seismic (x 0.2)

TRXL 0.01%

TRXL 0.1%

RC Initial

RC 0.0001%

RC 0.001%

RC 0.01%

adopted profile for drained modulus E'

Fig. 2. Summary of Young’s modulus values.

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

0 1 2 3 4

Relative Stiffness E/E3b [-]

Dep

th [

m]

Press Init AVE

Press First AVE

Press Second AVE

Seismic (x 0.2)

TRXL 0.01%

TRXL 0.1%

RC Initial

RC 0.0001%

RC 0.001%

RC 0.01%

ALL

NO INITIAL PRESS

NO INITIAL PRESS & NO TRXL0.01%

Poulos & Davids 2005

Fig. 3. Assessed soil relative stiffness.

Pile Load TestsA program of pile load testing was undertaken which involved the installation of seven test piles in the podium area near the location of the Khalifa Tower. All the test piles and reaction piles were bored cast in-situ and constructed under polymer fluid. A permanent casing, 6m long, was installed from the top of each pile to just above the highest strain gauge level for all the trial piles tested in compression and tension. Five piles, designated as P1, P2, P3, P4 and P5, were tested in compression; two, P3 and P5, were shaft grouted. Test pile P6 was tested in tension and test pile P7 was laterally loaded.

Only the compression load tests on trial piles P1, P2 and P4 have been considered for the present paper. Table 5 summarizes the main features of these piles. Figure 4a shows the load test arrangements for piles P1 and P2, which consisted of the test pile and six reaction piles, while

8 11 July 2011

Figure 4b shows the set-up for pile P4, which consisted of the test pile and four reaction piles. Steel load distribution plates were grouted to the top of the test piles and hydraulic jacks were placed between the steel plates and the reaction beams. Steel reaction beams were used to transfer the load from the hydraulic jacks to the installed reaction piles. Macalloy bars were used as reaction anchors to transfer the load from the beams to the reaction piles. Six cycles of loading were applied to trial piles P1 and P2 while nine cycles of loading were applied to trial pile P4, which was the pile designated to be tested cyclically.

Table 5. Summary of pile load tests.

Trial Pile

Diam.

[m]

Cut-off level

[m DMD]

Toelevel

[m DMD]

Length

[m]

Load Test layout

DWL*

[t]

DML**

[t]

No. of cycles

1 1.5 -4.85 -50 45.156 RP

circle with a 4.5 m radius

3000 6000 6(50%-150% DWL)

2 1.5 -4.85 -60 55.156 RP

circle with a 4.5m radius

3000 6000 6(50%-150% DWL)

4 0.9 -2.90 -50 47.14 RP

square with a 9 m side

1000 3500 9(100%-150%

DWL)

*designated working load;** designated maximum test load.

Four main types of instrumentation were used in the compression test piles:

1. Concrete embedment vibrating wire strain gauges, to allow measurement of axial strains at six levels along the pile shafts and hence estimation of the axial load distribution;

2. Extensometers, to measure change in length of the piles, and installed at the same levels as the vibrating wire strain gauges to provide back-up information on axial load distribution with depth;

3. Displacement transducers at the top of piles, to measure the vertical movement at the pile heads.

4. Load cells, to monitor the load applied to the pile via the jacks.

Back-Analysis interpretation of load tests to obtain Young’s Modulus Values

The computer program NAPRA was used to carry out the back-analyses of compression tests on the three test piles considered. Since a detailed soil profile at each trial pile location was not available, the same geotechnical model was adopted for all three piles.

For comparison purposes the three load tests were back-analysed both taking and not taking into account interaction between test piles and reaction piles. It is now well-recognised that ignoring interaction between the test pile and the reaction piles can lead to an overestimation of the pile head stiffness (Poulos & Davis, 1980; Poulos, 2000; Kitiyodom et al., 2004).

Both linear elastic (LE) and non-linear analyses (NL) were carried out. In all analyses, Young’s modulus for the piles, Ep, was assumed to be 31.8 GPa. For the linear

9 11 July 2011

analyses, the theoretical behaviour was fitted to the observed load-settlement behaviour at pile head displacements of about 0.08% of diameter and 0.2 % of diameter.

In the non-linear analyses, in order to assess the sensitivity of the back-calculated values of soil stiffness to the value of ultimate capacity, Q lim, three different values were adopted in the analyses:

1) Qlim was estimated as the asymptote to a hyperbola fitted to the whole measured load-settlement curve (HYP);

2) Qlim was based on the load transfer deduced by strain gauges readings (SG); 3) Qlim was based on the load transfer deduced by extensometer readings (EX).

Ultimate skin friction values inferred from the axial load distribution and from the extensometer readings were employed to assess pile shaft capacity up to depths above -30m, -38m and -30m for piles P1, P2 and P4 respectively. From pile–soil interface load-strain curves at various depths along the shafts, these values were found to be representative of the ultimate values in the upper (cased) part of the shaft. Below these depths, ultimate values of shaft friction were estimated from correlations with the unconfined compressive strength (UCS) of the rock.

Table 6 summarises the values of Qlim obtained from these three procedures. As might be expected, the hyperbolic extrapolation procedure gives the largest values, and probably over-estimates the capacity. There is some difference between the values assessed on the basis of the strain gauge and extensometer readings, but from the point of view of settlement prediction, such differences are not very significant.

Figures 5 and 6 show typical fits (for Pile P2) of the computed non-linear load-settlement behaviour and the observed load-settlement behaviour. Figure 5 is for the interpretation taking account of interaction, while Figure 6 shows the corresponding fit with interaction between the test pile and reaction piles being ignored. In both cases, very reasonable fits are obtained with the measured data.

Table 6. Assessed pile capacity with different methods.

Pile

Qlim

[kN] Hyperbolic Extrapolation

(HYP)Strain Gauge

Readings(SG)

Extensometer Readings(EX)

TP1 108,800 93,800 73,200

TP2 115,900 97,300 100,200

TP4 82,600 50,500 59,900

10 11 July 2011

(a) Piles P1 and P2. (b) Pile P4

Fig. 4 Set-up for pile load tests

0

5

10

15

20

0 10000 20000 30000 40000 50000 60000 70000Q [kN]

w [m

m]

measured

E3b=1000 Mpa (NL)HYP

E3b=1200 Mpa (NL)SG

E3b=1200 Mpa (NL)EX

0

5

10

15

20

0 10000 20000 30000 40000 50000 60000 70000Q [kN]

w [m

m]

measured

E3b=1500 Mpa (NL)HYP

E3b=1700 Mpa (NL)SG

E3b=1700 Mpa (NL)EX

Fig. 5 Predicted and measured load–settlement Fig. 6 Predicted and measured load–settlement for pile P2(interaction considered). for pile P2 (interaction ignored). Back-calculated values of the Young’s Modulus for stratum 3b, E3b, are reported in Table 7. In the linear elastic analyses the first point on the measured load-settlement curve has been considered. In this way back-calculated values of soil stiffness in linear analyses are affected by the loading procedure adopted in the load tests. In the cases of piles P2 and P4, values of back-calculated soil stiffness are in close agreement with values back-calculated in the non linear analysis (values of w/D are 0.0008 - 0.0009) while in the case of pile P1, the first point is at a higher displacement (0.21%), and so the back-calculated value is lower. It should be noted that had the interaction between the test pile and the reaction piles not been taken into account, the back-calculated values of pile-soil relative stiffness would have been considerably larger.

Table 7. Young’s Modulus values derived from load tests.

TEST PILE

E3b [MPa] with interaction accounted for E3b [MPa] with interaction not accounted for

Linear Analysis(w/D=0.0008)

Linear Analysis(w/D=0.002)

Non-linear Analysis Linear Analysis(w/D=0.0008)

Non-linear Analysis

P1- 350 650 (HYP)-850 (SG) - 900 (HYP)-1100 (SG)

P2 700 6501000 (HYP)-1200 (SG-

EX)1200

1500 (HYP)-1700 (SG-EX)

11 11 July 2011

P4850 550 650(EX)-850(SG) 1100 850(EX)-1100(SG)

Notes: HYP denotes values derived from hyperbolic extrapolationEX denotes values derived from extensometer readingsSG denotes values derived via strain gauges

From Table 7, the following points can be noted:1. The consideration of interaction between the test pile and the reaction piles results in

backfigured modulus values which are considerably less than those for which interaction has been ignored. Thus, there would be a tendency to under-estimate foundation settlements if interaction effects are ignored.

2. The back-calculated values from the three tests are scattered.

In order to partially overcome the described limitation in the back-analysis of load tests, and to show its effects on the average settlement assessment, sensitivity analyses have been carried out with NAPRA by adopting two different values of soil stiffness assessed to be representative of lower and upper bound values. In the GARP and NAPRA analyses described below, for the assessment of the average and differential settlements, the values of E3b shown in Table 8 were adopted, on the basis of the non-linear analysis of the load test results.

Table 8. Values of Young’s Modulus (E3b) of Layer 3b Adopted for Foundation Analyses

CaseYoung’s Modulus of Layer 3b (E3b) MPa

Best Estimate Upper Bound Value Lower Bound ValueReaction pile interaction

considered900 1000 650

Reaction pile interaction ignored

1200 1500 900

PROCEDURE FOR FOUNDATION SETTLEMENT RE-ASSESSMENT

The majority of the foundation settlement re-assessment was carried out using linear elastic analyses with the computer program NAPRA. The mesh adopted for the NAPRA analyses is shown in Figure 7, and in this mesh, the columns were spaced 1.7 m apart. Preliminary analyses indicated that using a finer mesh than this produced no change in the results. The actual shape of the raft was modelled by adopting a piecewise approximation.

Only long-term conditions have been considered, and for the majority of the early analyses, an average load per pile of 23.21 MN has been used (this is representative of the design dead plus live loading) and has been applied as a point load on each of the 194 piles. This load corresponds to a uniformly distributed load on the tower raft of about 1250 kPa.

The majority of analyses were undertaken using the best-estimate modulus value of 900 MPa derived from the proper interpretation of the load test data (see Table 8).

A series of sensitivity analyses was undertaken to examine the following issues:

12 11 July 2011

1. The influence of non-linear pile response;2. The influence of the range of back-figured values of Young’s modulus;3. The influence of using “correct” and “incorrect” back-figured values of Young’s modulus;4. The effect of not considering the raft in the foundation settlement analysis;5. The differences between analyses using NAPRA and GARP;6. The influence of the assumed loading pattern;7. The effect of incorporating the stiffness of the superstructure;8. The effect of including the podium loading.

The results of these sensitivity studies are described below.

Fig.7 Model adopted in NAPRA analyses

RESULTS OF SENSITIVITY STUDIES

Influence of Non-Linear Pile Response

For the non-linear NAPRA analyses, the ultimate axial capacity of each pile has been assumed to be 112.5MN. Table 9 compares the computed maximum (S max) and central settlements (S centre), and the maximum differential settlement (DS max), from the linear and non-linear analyses. There is very little difference between the two analyses in this case, as it could be expected being the global safety factor on each pile in the range 3 to 5. Thus the comparison indicates that the foundation response is essentially elastic under the dead plus live loadings. Accordingly, only linear analyses have been employed for the remainder of the sensitivity studies.

Table 9. Computed Settlements (mm) from Linear and Non-Linear Analyses

Linear Analysis Non-Linear AnalysisS max S centre DS max S max S centre DS max

52 51 27 53 53 27

The influence of the range of back-figured values of Young’s modulus

Table 10 summarises the computed settlements from the NAPRA analysis, using the range of values of Young’s modulus back-figured from the correct interpretation of the pile load tests (see Table 8). As would be expected, the computed settlement for the lower-bound modulus value is considerably greater than that for the upper-bound modulus value, although the ratio of the computed settlement values is less than the ratio of the modulus values. This may be explained by

13 11 July 2011

the non linearity source provided by the iterative check on the tensile forces at the raft-soil interface.

Table 10. Influence of Using Upper and Lower Bounds of Backfigured Modulus Values

Modulus Value for Layer 3b S max

S centre DS max

Lower bound (E3b= 650 MPa) 81 68 50

Upper bound (E3b=1000 MPa) 56 46 35

The influence of using “correct” and “incorrect” back-figured values of Young’s modulus

Table 11 shows the influence on the computed settlements of using the best-estimate modulus values for Layer 3b obtained from the “correct” interpretation (considering test pile-reaction pile interaction) and the “incorrect” interpretation (ignoring this interaction). The settlements computed using the “incorrect” modulus interpretation are about 21% less than those using the “correct” interpretation, and it is therefore important to properly interpret the test pile load-settlement data to avoid the tendency to under-estimate the foundation settlements and differential settlements.

Table 11. Influence of Modulus Value on Computed Settlements

Modulus Value UsedComputed Settlements mm

S max

S centre DS max

“Correct” InterpretationE3b= 900 MPa

52 51 27

“Incorrect” InterpretationE3b=1200 MPa

41 40 22

Effect of Not Considering the Raft in the Analysis

14 11 July 2011

NAPRA has been used to analyse the foundation system, both as a piled raft, and as a pile group in which there is no raft joining the piles. The “correct” best-estimate modulus of Layer 3b of 900 MPa has been used. Table 12 shows the computed settlements for both these cases. The difference between the computed central settlements is negligible, but there is a considerable difference between the computed maximum settlements and differential settlements. In this case, the conservatism introduced by ignoring the raft would lead to a 17% increase in the computed maximum settlement but a 40% increase in the maximum differential settlement. Therefore it is desirable to incorporate the effect of the raft when computing the settlement distribution within the foundation system.

Table 12. Influence of Ignoring Pile Cap on Computed Settlements

Case Analysed Computed Settlements mmS max S centre DS max

Pile Group (no raft)61 51 38

Piled Raft52 51 27

Analyses Using NAPRA and GARP

Similar meshes have been used for both the NAPRA and GARP analyses and identical analysis assumptions have been made in both cases. Figure 8 shows the computed profile of settlement along Wing C from both analyses, and reveals that they are almost identical. Thus, pleasingly, for the same input, each program is capable of giving very similar results.

0

10

20

30

40

50

60

0 10 20 30 40 50 60 70

Distance along wing cross section [m]

w [m

m]

GARP AV 23210 kN NAPRA AV 23210 kN

Figure 8 Calculated settlements along wing C from NAPRA and GARP

The Influence of the Loading Pattern

15 11 July 2011

The preceding results have all been obtained assuming that the average design load (23.21 MN) has been applied to each pile location. In reality, the loads will be applied via wall and column locations, and consequently, NAPRA has been used to examine the influence of the loading pattern on the computed settlement profile for two cases:

a. Equal loads on all the piles;b. The actual design loadings are applied at the wall and column locations.

The computed settlement profiles along Wing C in Figure 9 show a difference in the computed settlement patterns, with the equal load assumption giving smaller maximum settlement than the other case. Thus, it would appear desirable to employ the actual load pattern in design calculations.

0 10 20 30 40 50 60 70

0

10

20

30

40

50

60

70

80

Average pile loads applied

Loads applied at wall & column locations

d [m]

w [

mm

]

Figure 9 Influence of Assumed Loading Pattern on Computed Settlement Profile

The Influence of the Podium Structure on the Tower Settlement

The podium structure was assumed to be founded at the same depth of the tower raft, but the two rafts were assumed to be unconnected. The length of the podium piles was taken to be 30m, and the columns and rows in the NAPRA mesh were spaced at 3m up to a distance of 30 diameters and 4 m at larger distances. An average load per pile of 23.21 MN was applied as a point load on each of the 194 tower piles while the loads acting on the low-rise area were modelled as point loads of between 2 and 8 MN acting on the podium piles, depending on their location.

The maximum computed settlement was 54mm which was only 2mm larger than the value computed for the tower only. The effect of the 750 piles of the podium was thus very small in this case, primarily because of the significant distance of many of the podium piles from the tower, and the relatively small loads that they carried.

The Influence of Superstructure Stiffness

In order to investigate the effect on the computed settlement and differential settlement, and to try and obtain a more accurate estimate of the pattern of settlement, the stiffening effect exerted by the superstructure on the raft was taken into account, in various ways, by increasing the bending stiffness of the raft in each wing (estimated by the structural designers to be equivalent to an

16 11 July 2011

increase of 25200 kNm2 per wing). Six alternative methods of incorporating this increased bending stiffness were adopted:

a. Increasing the thickness of the whole raft to reflect the bending stiffness of the entire tower (Model 1).

b. Increasing the raft thickness over the central part of the wings and on the core, as shown in Figure 10, to reflect the bending stiffness of the entire tower. This is denoted as Model 2.

c. Increasing the raft thickness only below the shear walls (see Figure 11), to reflect the bending stiffness of the entire tower; this case is denoted as Model 3.

d. Model 1, with only 10% of the stiffness of the tower considered (Model 1M).e. Model 2, with only 10% of the stiffness of the tower considered (Model 2M).f. Model 3, with only 10% of the stiffness of the tower considered (Model 3M).

In each case, the actual pattern of loading via the columns and walls was applied, with only the dead load component considered.

Figure 12 compares the various computed profiles of settlement across the tower, together with those in which no account is taken of superstructure stiffness. Also shown is the design profile developed by Poulos and Bunce (2008), which was for combined dead plus live load, and therefore not directly comparable. Clearly, there is a considerable difference between the extreme profiles, those taking all the superstructure stiffness into account, and that in which no account is taken of the superstructure stiffness. It would appear reasonable to assume that the profiles from Models 1M, 2M and 3M may be more reasonable approximations to reality, and this appears to be borne out by the comparisons with the measured settlements, as described below.

Fig.10. Raft model 2. Fig.11. Raft Model 3.

17 11 July 2011

0 10 20 30 40 50 60 700

10

20

30

40

50

60

70

80

90

NAPRA Model 1

NAPRA Model 2

NAPRA Model 3

NAPRA Model 1M

NAPRA Model 2M

NAPRA Model 3M

NAPRA-No Structure Stiffness Allowance

Design Values (Poulos & Bunce (2008)

Distance along wing [m]

Se

ttle

me

nt

[mm

]

Figure 12 Comparison Between Various Calculated Settlement Profiles

Comparisons Between Calculated and Measured Settlements

Detailed settlement measurements were only available up to February 2008, before the end of construction and well before the commissioning and occupation of the building in January 2010. Nevertheless, anectodal evidence indicated that the additional settlements between February 2008 and January 2010 were relatively small, of the order of 1-2mm.

Figure 13 shows comparisons between the latest available measured profile of settlement in February 2008, and the calculated settlement profiles from Models 1M, 2M and 3M. The following observations are made from an examination of Figures 12 and 13:

1. Without allowance for superstructure stiffness, the calculated maximum final differential settlement is about 35mm which is considerably larger than the measured value of about 14mm. The computed maximum settlement is also larger than the measured value, although some additional settlement would be expected after the building has been in operation for some years.

2. When allowance is made for the superstructure stiffness, the computed maximum settlement is similar in magnitude to the measured value. However, for Models 1, 2 and 3, in which the full superstructure stiffness is incorporated (albeit approximately), the computed settlement pattern differs somewhat from the measured pattern, and the computed differential settlements are significantly smaller than those measured. It seems clear that it is not appropriate to allow for the bending stiffness of the entire structure when trying to modify the foundation stiffness.

3. When the allowance for superstructure stiffness is reduced by a factor of 10 (Models 1M, 2M and 3M), there is better agreement between the computed and measured profiles, with a computed maximum differential settlement ranging between 15 and 21 mm for the three models, similar to the measured value. In this case, the stiffness of the raft is approximately 53 times its original value, and this latter value is much larger than the value of 10 times adopted by Hooper (1973) for the Hyde Park Barracks in London and by Sales et al. (2010) for the Skyper Building in Frankfurt. Interestingly, and almost certainly coincidentally, the

18 11 July 2011

profile for this case is rather similar to that obtained for the case when the average load is imposed on each pile.

4. There remain some differences between the measured and computed settlement profiles in the vicinity of the edge of the wing. There may well be scope to refine the method by which the superstructure is incorporated into the geotechnical foundation analysis.

5. The calculated settlements from the design phase are considerably greater than those obtained from the analyses in this paper. The main reason for these larger settlements is that the settlements were for both dead and live load acting, and in addition, conservative values of Young’s modulus were used in these analyses, with a somewhat different distribution of ground stiffness with depth being adopted in those calculations. Once again, this comparison emphasises the importance of appropriate selection of the ground stiffness values if accurate foundation settlement predictions are to be made.

0 10 20 30 40 50 60 700

10

20

30

40

50

60

NAPRA Model 1M

NAPRA Model 2M

NAPRA Model 3M

Measured (February 2008)

Distance along wing [m]

Se

ttle

me

nt [

mm

]

Figure 13 Measured and Computed Settlement Profiles along Wing C

CONCLUSIONS

1. The case history of the Burj Khalifa Tower in Dubai has been re-assessed for the prediction of average and differential settlement of the piled raft foundation system. The comprehensive ground investigation and pile testing program carried out for this project has enabled the site to be characterized in some detail. The pile load tests is particular have been an important factor in enabling reasonable settlement prediction to be made.

2. The ground stiffness or modulus is a key factor in the prediction of foundation settlements. If this is to be derived from pile load test data, then the interpretation of the load-settlement should take into account

19 11 July 2011

interaction effects with the reaction system, otherwise the ground stiffness is likely to be over-estimated and the foundation settlements subsequently under-predicted.

3. Sensitivity studies have been carried out to explore the effects of a number of factors on predicted settlement behaviour of the Burj Khalifa tower. In addition to the ground stiffness or modulus, the consideration of the effects of the raft and superstructure stiffness may be important factors in influencing both the maximum settlement and the maximum differential settlement.

4. The assessment of the differential settlement of the Khalifa tower foundation has been by explored by adopting three different models to account for the stiffening effect of the superstructure, and they have been found to give reasonably similar results. However, if the foundation, or parts of it, are stiffened to represent the bending stiffness of the entire structure, the consequent foundation response is too rigid, and the differential settlements tend to be under-predicted considerably. For the Burj Khalifa tower, an additional stiffness equivalent to about 10% of the entire bending stiffness has been found to give improved, but by no means perfect, results when compared with measured settlement profiles. This result suggests that there is a limit to the stiffness that the structure can provide to the foundation system, and so the full structure should not be taken into account when calculating the effective increase of stiffness.

5. In this case at least, consideration of the low-rise podium structure leads to only a small increase in the settlement under the tower footprint.

6. The method of analysis may be a less significant factor in the prediction of piled rafts settlements, provided the method is sound. For the same input data, the computer programs GARP and NAPRA produced similar settlements of the tower.

REFERENCES

Abagnara, V., Poulos, H.G. and Small, J.C. (2012). Comparison of two piled raft analysis programs. Submitted for 12th Australia- New Zealand Conf. Geomechanics, Melbourne.

Clancy, P. & Randolph, M.F. 1993. An approximate analysis procedure for piled raft foundations. Int. Journ. For Num. and Anal. Meth. in Geomech, 17(12): 849-869.De Sanctis, L., Russo, G. and Viggiani, C. 2002. 21.Piled raft on layered soils. Proc. Ninth International Conference on Piling and Deep Foundations, ___ -___Nice 2002Hemsley, J.A. 1998. Elastic Analysis of Raft Foundations. Thomas Telford, London.Hemsley, J.A. (ed) 2000. Design Applications of Raft Foundations. Thomas Telford, London.Kitiyodom, P., Matsumoto, T. and Kanefusa, N. 2004. Influence of reaction piles on the behaviour of a test pile in static load testing. Canadian Geotechnical Journal, 41, 408 – 420Poulos, H.G. & Davis, E.H. 1980. Pile foundation analysis and design. New York, John Wiley.

20 11 July 2011

Poulos, H.G. 1994. An approximate numerical analysis of pile-raft interaction. Int. Journ. For Num. and Anal. Meth. in Geomech., 18, 73-92.Poulos, H.G. 2000. Pile testing – From the designer’s viewpoint. Statnamic Loading Test ’98, Balkema, Rotterdam, 3-21.Poulos, H.G. and Bunce, G. (2008). Foundation design for the Burj Dubai – the

world’s tallest building. Proc. 6th Int. Conf. Case Histories in Geot. Eng., Arlington, Virginia, Paper 1.47 CD volume.

Russo, G. 1998. Numerical analysis of piled rafts. Int. Journ. For Num. and Anal. Meth. In Geomech., 22, No 6, 477-493.Russo, G. And Viggiani, C. 1998. 15.Factors controlling soil-structure interaction for piled rafts. Proc. International Conference on Soil-Structure Interaction in Urban Civil Engineering, Ed. R. Katzenbach & U. Arslan, Darmstadt,___ - ___.Sales. M.M., Small, J.C., Poulos, H.G. 2010. Compensated piled rafts in clayey soils: behaviour, measurements, and predictions. Can. Geotech. J.47, 327-345.Selvaduri, A.P.S. 1979. Elastic Analysis of Soil-Foundation Interaction. Elsevier Publishing Co., New York.Small, J.C. and Poulos, H.G. (2007). Nonlinear analysis of piled raft foundations.

Geotech. Special Publication GSP158, ASCE, CD Volume, GeoDenver 2007.Ta, L.D. & Small, J.C. 1996. Analysis of piled raft systems in layered soils. Int. Journ. For Num. and Anal. Meth. inGeomech. 20: 57-72.Viggiani, C. 1998. Pile groups and piled rafts behaviour. Proc. 3rd Int. Geot. Seminar on Deep Foundations on Bored and Auger Piles, Ghent, 77-94

21 11 July 2011