PILED RAFT DESIGN PROCESS FOR A HIGH-RISE BUILDING ON THE GOLD COAST, AUSTRALIA PARAN MOYES Senior Geotechnical Engineer, Coffey Geosciences Pty Ltd 8/12 Mars Road, Lane Cove West NSW 2066 HARRY G POULOS Senior Principal, Coffey Geosciences Pty Ltd Emeritus Professor. Professor. Department Department of Civil Engineering, Engineering, University of of Sydney, Australia Australia JOHN C. SMALL Professor. Department Department of Civil Engineering, Engineering, University University of Sydney, Sydney, Australia FRANCES BADELOW Principal, Coffey Geosciences Geosciences Pty Ltd Ltd
This paper describes the process of design of a piled raft foundation for a high rise residential building on the Gold Coast in Queensland, Australia. The design process comprised an initial stage of geotechnical site characterization using the results of a series of investigation investigation boreholes boreholes to prepare prepare a subsurface subsurface model and and derive geotechnical geotechnical parameters parameters for raft raft and pile design design from empirical correlations. Following this a preliminary analysis was undertaken using a combination of elastic theory and allowances for non linear behaviour of the piled raft system to assess the viability of such a foundation system and any potential advantages of a piled raft over conventional fully piled foundation systems. Finally, a detailed analysis was undertaken using the GARP8 computer program. These detailed analyses were used to design a more efficient piled raft system and to provide design actions for structural design of the foundation system for a variety of load combinations.
1.
Introduction
Paradise, Queensland Australia. This assessment required the evaluation of a piled raft system subjected to large ultimate loadings with consideration also being given to the satisfaction of stringent differential settlement criteria.
The development of tall buildings on land previously regarded as unsuitable for large structures presents several geotechnical problems with regard to the design and assessment assessment of foundations. Design of the buildings must take into account both the short and long term deformations of the foundations (serviceability limit state) and the strength of the foundations at ultimate loading (ultimate (ultimate limit limit state). state). Piled raft raft foundations foundations utilise piled support for control of settlements with piles providing most of the stiffness at serviceability loads, and the raft element providing additional capacity at ultimate loading. A geotechnical assessment for design of such a foundation system therefore needs to consider not only the capacity of the pile elements and the raft elements, but their combined capacity and interaction under serviceability loading. This paper presents details of the geotechnical assessment and design carried out for a piled raft foundation system for a residential tower development located in Surfers
2.
Residential Tower Project
The project comprises a 30 storey, 176 unit residential tower located in Surfers Paradise, Queensland. The construction of the development commenced in late 2004 with anticipated completion in late 2005. 2.1. Geological and Geotechnical Conditions
The Surfers Paradise area is underlain by alluvial sediments comprising sands and clays. Beneath the alluvial deposits is a residual soil strata of Silty Clay overlying the meta siltstone bedrock.
1
2 2.1.1 Site geotechnical conditions The geotechnical investigations at the site comprised nine boreholes drilled in two phases of fieldwork. Figure 1 illustrates the locations of the geotechnical investigation boreholes. The first phase of investigation comprised the drilling of six boreholes (BH1 to BH6) that were terminated within about 1m penetration into a sandy gravel layer, some 5m above the weathered rock layer. A second phase of investigation boreholes followed (“X” series) which were extended through the sandy gravel layer and encountered extremely weathered rock, typically described as sandy clay/clayey sand and sandy gravel.
3.
Geotechnical Model
On the basis of the information in Figures 2 and 3, a geotechnical model was developed in order to analyse the piled raft foundation option. As the conditions at BH5 were the least favourable for foundation design, this borehole was used as the main basis for the geotechnical model. Figure 3 shows the stratigraphy adopted for the model and the relevant geotechnical parameters selected for the various strata appropriate for either the raft or the pile design, where: Es(raft) = soil modulus for assessment of raft behaviour
BH 2
BH 1
level is generally about RL+5.5m, the ground water table is at about RL+0.7m, and the upper 6m or so of soil was to be excavated beneath the tower to allow for construction of the basement.
BH 4
pu = ultimate bearing capacity of the raft BH 5
Es (pile) = soil modulus for assessment of pile behaviour
BH 3
f s = ultimate pile shaft friction BH 6 OUTLINE OF PROPOSED DEVELOPMENT
Figure 1. Borehole locations
The ground conditions typically consist of a relatively thick layer of dense to very dense sand, underlain by a layer of peat with some sand present. Below the peat there is another dense sand layer and then a layer of stiff to very stiff, silty clay with some sand. Underlying this layer is extremely weathered rock, which is described as gravel in some of the boreholes. The extremely weathered rock is underlain by slightly weathered metasiltstone rock at between 40.5m and 42.5m depth below existing ground level. Based on the subsurface information obtained from boreholes X1, X3, BH1, BH4 and BH5, a geotechnical summary of the borehole details was formulated (refer Figures 2 and 3). The quantitative information is limited and consists of SPT data. The existing ground surface
f b = ultimate pile end bearing capacity. In assessing these parameters, use was made of the correlations between SPT data and foundation stiffness and capacity parameters suggested by Decourt (1995), together with the authors’ previous experience in parameter selection for use in piled raft design.
BH1 RL +5.4 0 +5
BH5 RL +6.0
BH4 RL +5.5 SPT
0
50
FILL
+5
SPT
+5
SAND L
SAND L
SPT
0
50
FILL
50
FILL
3
SAND L 62
0
0
SAND MD
0
ASSUMED LEVEL OF BASE OF RAFT
SAND D 75
-5
64
SAND VD
-10
62
-5
-5
SAND D (+ SHELLS & GRAVELS)
62
SAND VD
SAND/PEAT PEAT
-10
PEAT
-10
) m (
) m (
SAND VD
62
) m ( L R
L R
L R
PEAT + SAND MD
SAND VD -15
81
SAND VD
-15
-15
PEAT S 64
SILTY CLAY
SAND D
SAND VD -20
-25
-20
-20
SILTY/CLAY WITH SAND St
SILTY/SANDY CLAY V-St
-25
SILTY/CLAY WITH SAND V-St
-25
SANDY CLAY
-30
128
XW ROCK
-30
-30
106
GRAVEL (XW ROCK)
100
GRAVEL (XW ROCK)
Figure 2: Summary of Geotechnical Investigation Boreholes – BH Series BH X1 RL +5.5 (ASSUMED) 0
SPT
BH X2 RL +5.5
BH X3 RL +5.5 (ASSUMED) 50
+5
0
SPT
SPT
+5
SAND MD
0
50 +5
20
40
60
80
FILL (LOOSE)
SAND MD SAND L 52
0
-5
0
SAND D-VD
0
SAND D 60
60
60
60
SAND VD -5
60
SANDY GRAVEL VD -5
SAND VD
60
SAND VD
SAND VD 60 -10
-10
CLAYEY PEAT CLAYEY SAND
60
) m-15 ( L R
60 SAND D
SAND MD PEAT SANDY CLAY
-10
60 60
ASSUMED LEVEL OF BASE OF RAFT
SAND MD
60
WOOD/PEAT L
SAND VD ) m ( -15 L R
SAND VD
) m ( -15 L R
60
CLAY F-St HIGH ORGANIC CONTENT
PP=95 PP=170-220 PP=270
60 SAND VD -20
-20
-20
SANDY CLAY V-St
SANDY CLAY V-St
PP=500 SANDY CLAY H
CLAY H -25
-25
CLAY H
SANDY GRAVEL VD
-35
CLAYEY SANDY GRAVEL VD
GRAVELLY SAND D
GRAVELLY SAND D
GRAVELLY SAND MD
SANDY GRAVEL VD
XW ROCK
CLAYEY SAND MD
SAND VD -30
-35
PP=600
-25
CLAYEY SAND MD
SAND D -30
PP=460-500
SAND MD -30 SANDY GRAVEL etc. D
SAND MD -35
SANDY GRAVEL VD METASILTSTONE (SW)
-40
Figure 3: Summary of Geotechnical Investigation Boreholes – X Series
4
Es (RAFT) MPa
pu (RAFT) MPa
Es (PILES) MPa
f s kPa
f b MPa
90
5.4
120
100
9.9
8
0.5
20
22
0.7
-
90
5.4
120
100
9.9
14
250
25
1.5
40
60
2.0
SAND WITH SOME GRAVEL MD
25
-
37.5
2.25
50
48
4.1
SANDY GRAVEL
100
-
150
9.0
200
100
10.0
METASILTSTONE SW
-
2000
-
2000
-
10.0
Su kPa
DESCRIPTION
Av. SPT
SAND D-VD
60
-
PEATY CLAY (SOME SAND) F-St
10
80
SAND D
60
SANDY CLAY (H) /CLAYEY SAND MD
0
ASSUMED BASE OF RAFT
-5
-10
-15
) m ( -20 L R
-25
-30
-35
-
-40
Figure 4. Stratigraphic and Geotechnical Parameters Adopted
In addition to the parameters shown in Figure 4, it was assumed that, below the raft, the soil modulus (Es) for reloading is 3 times the value for initial loading (shown in Figure 3). This assumption was made to evaluate the benefits of excavation of the upper 6m of soil, and the consequent partial compensation that this excavation provides.
4.
Preliminary Geotechnical Assessment
4.1. Introduction
Prior to the detailed geotechnical assessment, a feasibility assessment was conducted of various foundation schemes. A geotechnical assessment was carried out for the following foundation schemes: • A raft alone, without piles; • A raft with 50 piles; A raft with 70 piles; • • A raft with 140 piles (approximating of the proposed design as per the concept foundation drawing prepared by the piling contractor).
For the purposes of the preliminary assessment, the piles were assumed to be 0.7m diameter Continuous Flight Auger (CFA) piles extending to the thin sand layer directly above the stiff to very stiff clay layer, and having an average length of about 18m. The raft was taken to be a square area 50m by 24m in plan, and approximately 0.8m thick. The thickness is of little consequence for the overall load-settlement behaviour, but will of course influence the differential settlements and the bending moments and shear forces in the raft. Based on the results of the initial assessment it was concluded that while a raft foundation alone would have an overall factor of safety of more than 10 with respect to combined dead and live loading, the foundation design would be governed by settlement considerations, rather than by ultimate bearing capacity. The presence of a dense to very dense sand deposit allows the raft to develop significant vertical load capacity and stiffness.
4.2. Method of analysis
For preliminary assessment, an extension of an approximate analysis of the piled raft load-settlement behaviour described by Poulos (2002) was used. This method uses the equations developed by Randolph (1994) to compute the stiffness of a piled raft, in terms of the raft and pile group stiffness values, and also the load sharing between the raft and the piles. A tri-linear load-settlement curve is derived from this process. An extension to this procedure was developed by Poulos (2005) for compensated piled rafts. In this extended procedure, account is also taken of the increase in soil stiffness during the re-loading phase after excavation has been carried out. This increase in soil stiffness increases the stiffness of the raft during the reloading process, and consequently, the raft carries a greater proportion of the load until the soil pressure below the raft reaches the previous pressure at which virgin loading conditions prevail; this will usually occur when the raft pressure balances the pressure reduction due to the excavation. Thereafter, the raft stiffness is controlled by the virgin loading soil stiffness, which will be smaller than the reloading stiffness. Thus, at that stage, the piles will then tend to take a larger proportion of the further applied load. For the preliminary assessment, it was assumed (somewhat conservatively) that the soil modulus for
5 reloading was 3 times the value for first or virgin loading. The analysis was carried out using a MATHCAD worksheet developed by the authors.
Table 1. Summary of Computed Settlements Number of Piles Below Raft
Average Settlement at Serviceability Load (257.8 MN)
Average Settlement at 2*Serviceability Load (515.6 MN)
0
98 mm 58 mm 49 mm 37 mm
234 mm 153 mm 106 mm 79 mm
4.3. Analysis results
The results of the preliminary geotechnical assessment are illustrated in Figures 5 and 6. The computed loadsettlement relationships are shown in Figure 5, for loads up to 3 times the serviceability load. This figure shows that the settlement tends to decrease as the number of piles increases, but at a decreasing rate. This figure also indicates the transition from reloading to virgin loading conditions at relatively low load levels, and the consequent reduction in stiffness once virgin loading conditions are re-established. 900
800
Raft only Raft + 50 piles
700
Raft + 70 piles Raft + 140 piles
600
N 500 M d a o L
400
300
200
100
0 0
50
100
150
200
250
300
350
400
450
50 70 140
Average Settlement at 3*Serviceability Load (773.4 MN)
370 294 247 121
mm mm mm mm
The following observations can be made: 1. The average settlement of the raft alone is likely to be excessive; 2. The average settlement at the serviceability load decreases significantly when a relatively small number of piles are included in the foundation; 3. There is relatively little benefit, at the serviceability load, in doubling the number of piles from 70 to 140. However, the settlement at higher load levels is decreased significantly. Only the overall load-settlement behaviour was considered in this preliminary assessment. A number of other issues including differential settlements, the thickness of the raft, and the size, length and required locations of the piles were not considered in detail at this stage, but were addressed at the detailed design stage.
Settlement mm
Figure 5: Computed Load Settlement Curves from Preliminary Analysis
Table 1 summarizes the computed settlement of the foundation system at 1, 2 and 3 times the serviceability loads, while Figure 6 plots these settlements versus the number of piles. 400
350
4.4. Recommendations from preliminary geotechnical analysis
Based on the results of the preliminary geotechnical assessment, it was recommended that a piled raft foundation system should provide a cost-effective solution for the tower. The average settlement at the serviceability load was estimated to be of the order of 35-60mm, depending on the number of piles.
Serviceability Load 2*Serviceability Load
300
3*Serviceability Load No piles-Serviceability
250 No piles - 2*serviceability
m m t n e 200 m e l t t e S
5.
Detailed Foundation Design
No piles - 3*Serviceability
150
100
50
0 50
75
100
125
150
Number of Piles
Figure 6: Effect of Number of Piles on Settlement from Preliminary Analysis
Following the preliminary geotechnical assessment, a piled raft foundation was adopted for the proposed building. The foundation design comprises a 0.8m thick reinforced concrete raft founded at RL -1m AHD. Beneath the raft, 136 piles were initially located to support loading from columns, the building core and a number of walls. The number of piles was modified on the basis of the analysis undertaken.
6 The pile design that was assessed to be appropriate for this foundation system comprised 18m long, 0.7m diameter Continuous Flight Auger (CFA) piles. These piles were designed to be founded on the relatively thin, dense sand layer, but allowance was made in the analyses for the possibility of the pile tip extending into the underlying sandy clay/clayey sand layer. Accordingly, the analysis carried out assumed the pile tip capacity to be that for the sandy clay/clayey sand. The ultimate geotechnical capacity of each pile was assessed to be about 4.2MN. 5.1. Loading
The load data supplied by the project structural engineer was used to carry out a number of geotechnical assessments based on different combinations of loading. The loading data provided comprised a series of vertical loadings at column locations. Table 2 shows the loading combinations that were assessed based on Section 2.1 of the Australian Loading Code AS1170.0-2002. Table 2. Loading Cases for Piled Raft Assessment Limit State Dead Load (G)
Live Load (Q)
Load Factors Positive Wind Load (permissible) (W p+)*
Serviceability
1
0.4
-
-
Ultimate
1.2
1.5
-
-
1.2
0.4
1.35
-
1.2
0.4
-
1.35
0.9
-
1.35
-
0.9
-
-
1.35
Positive Wind Load (permissible) (W p-)
* A load factor of 1.35 was adopted to factor the permissible wind load to ultimate wind load
5.2. Serviceability limit state (SLS) assessment of piled raft
The load data supplied by the structural engineer was used to carry out a geotechnical assessment based on serviceability loading for assessment of deflections of the proposed piled raft foundation. The serviceability loading combination was assessed, based on the information provided in Section 2.1 of AS1170.0-2002. The geotechnical assessment of the piled raft was undertaken using the GARP8 program initially
developed by the Centre for Geotechnical Research at Sydney University for analysis of piled raft foundations. GARP (General Analysis of Rafts with Piles) is based on a finite element analysis of the raft, and a boundary element analysis of the piles. The contact stress that acts between the raft and the soil is assumed to be made up of a series of uniform blocks of pressure that act over each element in the raft. Each of the piles is assumed to apply a reaction to the raft at a point (corresponding to a node in the raft). The boundary element analysis is used to calculate the interaction of pairs of piles, or of a pile with the raft. In doing this, it is assumed that the soil is an elastic material. If the soil is layered, a weighted average of the properties of the soil layers is used in determining the equivalent elastic properties of the overall soil mass. If the behaviour of the piles is non-linear, this is modeled by allowing the stiffness of the piles to reduce with load level according to a hyperbolic law. However, the interaction between piles and between the piles and raft is assumed to be constant (i.e. to have the values for the original pile stiffness). Loading on the raft can include point loads, uniformly distributed loads, and moment loadings. As well, the raft can have different thicknesses assigned to the elements that make up the mesh. The deflections, shear forces and moments in the raft and the vertical loads on the piles due to the loading were assessed. The GARP8 analysis modelled the piled raft as a 0.8m thick raft with 18m long, 0.7m diameter piles located as per the initial foundation concept drawing. An initial assessment was carried out using a piled raft foundation with 136 piles. The results of the initial assessment are provided in Table 3. Table 3. Results of Preliminary GARP Analysis for Settlement Under Serviceability Loads Criteria
Value
Maximum settlement
36 mm
Minimum settlement
0.4 mm
Rotation (θ)x max
0.0014 radians (1 in 714)
θx
min
-0.0020 radians (1 in 500)
θy
max
0.0025 radians (1 in 400)
θY
min
-0.0024 radians (1 in 416)
7 Following the preliminary assessment, analyses were undertaken to refine the piled raft design. This comprised the removal of some of the piles and subsequent assessment of the performance of the revised design under the serviceability loading. The revised foundation design reduced the number of piles from 136 piles to 123 piles. Table 4. Results of GARP Analysis for Settlement Under Serviceability Loads Criteria
Value
Maximum settlement
44 mm
Minimum settlement
1 mm
Maximum differential settlement between adjacent columns
10mm (1/444)
It should be noted from Table 4 that the maximum computed settlement of 44mm was less than the maximum specified allowable value of 50mm, while the maximum differential settlement between adjacent columns was in the order of 1/400, which occurred between the a relatively lightly loaded exterior column and a highly loaded interior column. Following discussion with the piling contractor, further assessment was carried out with an additional pile located at the centre of the pile group beneath one of the most heavily loaded columns to reduce the differential settlement. The results of the further assessment indicated a differential settlement of 1/444 or 0.00225. The calculated settlement contours for this case are shown in Figure 7.
Figure 7. Calculated Settlement Contours (in m) – Serviceability Loading
This revised pile layout was then used to in the assessment of the ultimate limit state design. 5.3. Ultimate limit state (ULS) assessment of piled raft
Using the Ultimate Limit State (ULS) loading combinations summarized in Table 2, assessment of the piled raft performance was made. The resultant computed maximum and minimum values for the various structural actions are summarized in Table 5. Table 5. Results of GARP Analysis for ULS Load Cases Item Mx max Mx min My max My min Mxy max Mxy min Vx max Vx min Vy max Vy min Pile Load (compression) Pile Load (tension)
Value
Loading
2.66 MNm/m
1.2G + 1.35W p + ΨQ
-0.89 MNm/m
1.2G + 1.35W p + ΨQ
3.27 MNm/m
1.2G + 1.35W p + ΨQ
-1.11 MNm/m
1.2G + 1.35W p + ΨQ
0.82 MNm/m
1.2G + 1.35W p + ΨQ
-0.78 MNm/m
1.2G + 1.35W p + ΨQ
7.59 MN/m*
1.2G + 1.35W p + ΨQ
-7.11MN/m*
1.2G + 1.35W p + ΨQ
7.02 MN/m*
1.2G + 1.35W p + ΨQ
-6.75 MN/m*
1.2G + 1.35W p + ΨQ
6.37 MN
1.2G + 1.5Q
-0.24 MN
0.9G + 1.35W p
The values of moment and shear force are for the raft slab. The GARP8 program also provided contour plots of the moments and shears calculated for the various ultimate load cases, and distributions of b ending moment along selected sections of the piled raft for use in the structural design of the raft slab. Some of the pile loads calculated by GARP8 exceeded the maximum design capacity of 4.2MN. These results were based on the GARP8 analysis being elastic only, whereas the actual behaviour of the piles in the piled raft system would restrict the pile from carrying more load than its ultimate capacity. Subsequent upgrades have been made to the GARP program, so that when a pile reaches its design capacity, the program redistributes any excess reactions in the piled raft system.
8 5.4. Sensitivity Assessment
In addition to the serviceability and ultimate limit state assessments detailed previously, additional assessment was undertaken to evaluate the sensitivity of the proposed piled raft design to variability in the foundation system. The sensitivity assessment was carried out by considering the effect of a pile of reduced or increased stiffness located below the most heavily loaded column. The pile was modelled with a ±40% change in pile stiffness and was applied for both the serviceability limit case and also the ultimate case where the pile had previously been assessed as having the greatest load (Load Case ART 3b -1.2G + 1.35W p + ΨQ). Table 6 provides a summary of the effect of the ±40% change in pile stiffness. The deflections were evaluated using the serviceability load while the moments were assessed using the ultimate load case. Table 6. Results of GARP Analysis for Sensitivity Assessment – Serviceability Case Item
Deflection beneath load
Standard Pile Increase Pile Stiffness Stiffness (+40%)
Decreased Pile Stiffness (-40%)
39mm
38mm
38mm
8mm
8mm
Differential between column 9mm and edge of raft
The results of the assessment illustrate that the redundancy inherent in a piled raft foundation system redistributes the loading and settlement across the piled raft with relatively little influence on the overall behaviour. Table 7 provides a summary of the resultant maximum and minimum values for the various structural actions with the changed pile stiffnesses.
Table 7. Results of GARP Analysis for Sensitivity Assessment – Ultimate Cases Standard Pile Stiffness Item Mx max Mx min My max My min Mxy max Mxy min Vx max* Vx min* Vy max*
Value
Increased Pile Decreased Pile Stiffness Stiffness (-40% (+40%) Value Value
2.66 MNm/m
2.67 MNm/m 2.67 MNm/m
-0.86 MNm/m
-0.86 MNm/m -0.86 MNm/m
3.27 MNm/m
3.26 MNm/m 3.26 MNm/m
-1.11 MNm/m
-1.11 MNm/m -1.11 MNm/m
0.82 MNm/m
0.84 MNm/m 0.84 MNm/m
-0.79 MNm/m
-0.75 MNm/m -0.75 MNm/m
7.59 MN/m
7.59 MN/m
7.59 MN/m
-7.10 MN/m
-7.10 MN/m
-7.10 MN/m
7.02 MN/m
7.02 MN/m
7.02 MN/m
-6.73 MN/m -6.73 MN/m -6.73 MN/m Vy min* Pile Load* 6.23 MN 6.22 MN 6.22 MN (compression) Pile Load -0.10 MN -0.10 MN -0.10 MN (tension) * Pile load calculated based on elastic theory only. Maximum pile loads equal to maximum design capacity of 4.2MN.
The redistribution of the structural actions is reflected in the minor differences in the assessed values of moment and shear forces in the ultimate loading cases. The minor differences between the standard pile stiffness case and the reduced and increased pile stiffness cases can again be attributed to the effect of the 0.8m thick raft redistributing stresses and forces. 6.
Conclusions
This paper has illustrated the process of design of a piled raft foundation for a large residential development using a three stage procedure, consisting of an initial assessment of the feasibility of the design, a middle stage of refining pile locations and depths, and a detailed design stage of assessing the behaviour of the foundation under various loading cases. The GARP program provided an efficient computational method for the analysis of a complex geotechnical and structural problem and delivered design actions from the analysis which could be readily
9 used by structural engineers for structural design of the piled raft foundation. The utilization of a piled raft foundation versus a conventional piled only foundation delivered the required serviceability performance with regard to total and differential settlements while providing cost savings estimated to be of the order of 30% versus the original pile-only solution. References
1.
Decourt, L. (1995). “Prediction of load settlement relationships for foundations on the basis of the SPT-T”. Ciclo de Conferencias Inter. “Leonardo Zeevaert”, UNAM, Mexico, 85-104. 2. Poulos, H.G. (2005). “Piled raft and compensated piled raft foundations for soft soil sites” Geotechnical Spec Publications No 129, ASCE, 214 – 234. 3. Poulos, H.G. (2002). “Simplified design procedure for piled raft foundations”. Deep Foundations 2002, Ed. M.W. O’Neill & F.C. Townsend, ASCE Spec. Geot. Pub. 116, 1:441-458. 4. Randolph, M.F. (1994). “Design methods for pile groups and piled rafts”. State of the Art Report, 13 ICSMFE, New Delhi, 5: 61-82.
