Sunteți pe pagina 1din 43

16.

Practical design procedures for


piled raft foundations
H. G. Poulos

16.1. Introduction
It is common practice in foundation design to consider first the use of a shallow or raft
foundation to support a structure, and then if this is not adequate, to design a fully piled
foundation in which the entire design loads are resisted by the piles. Despite this design
approach, it is usual for a raft to be part of the foundation system (e.g. because of the
need to provide a basement below the structure). In the past few years, there has been an
increasing recognition that the strategic use of piles can reduce raft settlements and
differential settlements, and can lead to considerable economy without compromising
the safety and performance of the foundation. Such a foundation makes use of both the
raft and the piles, and is referred to here as a pile-enhanced raft or a piled raft. The
concept of piled raft foundations is by no means new, and has been described by several
authors, including Zeevaert [16.1], Davis and Poulos [16.2], Hooper [16.3], [16.4],
Burland et al. [16.5], Sommer et al. [16.6], Price and Wardle [16.7] and Franke [16.8],
among many others.
This chapter describes a philosophy of design for piled rafts and outlines circum-
stances that are favourable for such a foundation. A two-stage design process is
proposed, the first being an approximate preliminary stage to assess feasibility, and the
second, a more complete analysis procedure, to obtain detailed design information.
Methods of analysis are described and compared, and an assessment of the required
geotechnical parameters is outlined. Finally, some applications of piled raft foundations
are described briefly.

16.2. Design concepts


16.2.1. Design considerations
Figure 16.1 illustrates a general piled raft foundation and the overall forces and moments
which it must be designed to resist. The issues that must be considered in the design of
the piled raft foundation include:
(a) ultimate geotechnical capacity under vertical, lateral and moment loading;
(b) overall settlement and stiffness;
(c) differential settlements and angular rotations;

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
426 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

Hy, My

Hx, Mx x

V
Mx
Hx

Figure 16.1. Piled raft foundation system

(d) lateral movements and stiffness;


(e) structural design of both raft and piles.
The criteria for design depend to some extent on whether a traditional overall factor of
safety approach is adopted, or whether a limit state design philosophy is pursued. In the
former case, the design criterion for ultimate geotechnical capacity takes the form
Ru = FS Fd (16.1)
where Ru = ultimate geotechnical resistance of the foundation system
FS = overall factor of safety (typically in the range 2–3 for piled rafts)
Fd = design loading (usually the overall working load).
For limit state design, the design criterion for the ultimate limit state is
Rud ≥ Fud (16.2)
where Rud = ultimate design geotechnical resistance of the foundation system
Fud = overall loading for the ultimate limit state (usually a combination of
factored loadings such as dead, live, wind and earthquake loads).

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 427

The value of Rud is generally obtained by factoring down the ultimate geotechnical resist-
ance Ru by a geotechnical reduction factor Φg, given by
Rud = Φg Ru (16.3)
The value of Φg depends on a number of factors which contribute to uncertainty,
including analysis methods, available geotechnical information, experience with similar
situations, and the consequences of failure. Values are typically in the range 0.4–0.8,
with the higher values being appropriate only for projects in which the level of uncer-
tainty has been reduced significantly by extensive investigation and load testing, in
conjunction with appropriate experience and sound methods of design analysis.
For total settlements, differential settlements and lateral deformations, the design
criterion (irrespective of whether an overall safety factor method or a limit state method
of design is adopted) is that the maximum movement (or differential movement) must be
equal to or less than the specified allowable value, this latter value being dependent on
the type of structure.
The structural design of the foundation system requires an estimation of the following:
(a) bending moments and shear forces in the raft;
(b) axial loads, lateral loads and bending moments in the piles.
Ideally, a complete design method would be able to address all of the above issues in a
single coherent analysis. While such analyses are available using sophisticated
three-dimensional numerical analyses, it is essential that relatively simple methods be
available both for preliminary design purposes, and as a check on computer-based design
techniques. Some of these simplified methods are described in section 16.3.
In much of the available literature, emphasis has been placed on the bearing capacity
and settlement under vertical loads. While this is a critical aspect, and is considered in
detail herein, other issues must also be addressed. In some cases, for example, the pile
requirements may be governed by the overturning moments applied by wind loading,
rather than the vertical dead and live loads.

16.2.2. Alternative design philosophies


Randolph [16.9] has defined three different design philosophies with respect to piled rafts.
1. The ‘conventional approach’, in which the piles are designed as a group to carry
the major part of the load, while making some allowance for the contribution of
the raft, primarily to ultimate load capacity.
2. ‘Creep piling’, in which the piles are designed to operate at a working load at
which significant creep starts to occur at the pile–soil interface, typically at
70–80% of the ultimate load capacity. Sufficient piles are included to reduce the
net contact pressure between the raft and the soil to below the preconsolidation
pressure of the soil.
3. Differential settlement control, in which the piles are located strategically in order
to reduce the differential settlements, rather than to substantially reduce the
overall average settlement.

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
428 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

1
Curve 0:
raft only (settlement excessive)

2 Curve 1:
Piles and raft raft with piles designed for
yielding conventional safety factor
Piles yielding
Curve 2:
Load

3 raft with piles designed for


lower safety factor
No yield
0 Curve 3:
raft wth piles designed for
full utilization of capacity
Design
load
Allowable
settlement

Settlement

Figure 16.2. Load–settlement curves for piled rafts according to various design
philosophies

In addition, there is a more extreme version of creep piling, in which the full load
capacity of the piles is utilised i.e. some or all of the piles operate at 100% of their ulti-
mate load capacity. This leads to the concept of using piles primarily as settlement
reducers, while recognising that they also contribute to increasing the ultimate load
capacity of the entire foundation system.
Clearly, the latter approaches are most conducive to economical foundation design,
and will be given special attention herein. However, it should be emphasised that the
design methods to be discussed allow any of the above design philosophies to be
implemented.
Figure 16.2 illustrates, conceptually, the load–settlement behaviour of piled rafts
designed according to the first two strategies. Curve 0 shows the behaviour of the raft
alone, which in this case settles excessively at the design load. Curve 1 represents the
conventional design philosophy, for which the behaviour of the pile–raft system is
governed primarily by the pile group behaviour, and which may be largely linear at the
design load. In this case, the piles carry the great majority of the load. Curve 2 represents
the case of creep piling where the piles operate at a lower factor of safety, but because
there are fewer piles, the raft carries more load than for Curve 1. Curve 3 illustrates the
strategy of using the piles as settlement reducers, and utilising the full capacity of the
piles at the design load. Consequently, the load–settlement relation may be non-linear at
the design load, but nevertheless, the overall foundation system has an adequate margin
of safety, and the settlement criterion is satisfied. Therefore, the design depicted by

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 429

Curve 3 is likely to be considerably more economical than the designs depicted by


Curves 1 and 2.

16.2.3. Favourable and unfavourable circumstances for piled rafts


The most effective application of piled rafts occurs when the raft can provide adequate
load capacity, but the total or differential settlements of the raft alone exceed the allow-
able values. Poulos [16.10] has examined a number of idealised soil profiles, and found
that the following situations may be favourable:
(a) soil profiles consisting of relatively stiff clays;
(b) soil profiles consisting of relatively dense sands.
In both circumstances, the raft can provide a significant proportion of the required load
capacity and stiffness, with the piles acting to ‘boost’ the performance of the foundation,
rather than providing the major means of support.
Conversely, there are some situations which may be unfavourable, including:
(a) soil profiles containing soft clays near the surface;
(b) soil profiles containing loose sands near the surface;
(c) soil profiles which contain soft compressible layers at relatively shallow depths;
(d) soil profiles which are likely to undergo consolidation settlements due to
external causes;
(e) soil profiles which are likely to undergo swelling movements due to external
causes.
In the first two cases, the raft may not be able to provide significant load capacity and
stiffness, while in the third case, long-term settlement of the compressible underlying
layers may reduce the contribution of the raft to the long-term stiffness of the foundation.
The last two cases should be treated with considerable caution. Consolidation settle-
ments (such as those due to dewatering or shrinking of an active clay soil) may result in
a loss of contact between the raft and the soil, thus increasing the load on the piles, and
leading to increased settlement of the foundation. In the case of swelling soils, substan-
tial additional tensile forces may be induced in the piles because of the action of the
swelling soil on the raft. Theoretical studies of these latter situations have been described
by Poulos [16.11] and Sinha [16.12]. However, it should be noted that there may be
circumstances in which piled rafts are designed to act in tension; for example, in base-
ments to reduce excavation heave and prevent foundation uplift.

16.2.4. Design process


A rational design process for piled rafts involves two main stages:
(a) a preliminary stage to assess the feasibility of using a piled raft, and the required
number of piles to satisfy design requirements;
(b) a detailed design stage to obtain the optimum number, location and configura-
tion of the piles, and to compute the detailed distributions of settlement, bending
moment and shear force in the raft, and the pile loads and moments.

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
430 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

The preliminary stage involves relatively simple calculations which can often be
performed without a computer. The detailed stage will generally demand the use of a
suitable computer program which accounts in a rational manner for the interaction
between the soil, raft and piles. The stiffening effect of the superstructure also may need
to be considered.

16.3. Preliminary design


16.3.1. Estimation of ultimate geotechnical capacity
16.3.1.1. Vertical loading
The ultimate geotechnical capacity of a piled raft foundation can be estimated as the
lesser of the following two values:
(a) the sum of the ultimate capacities of the raft plus all the piles in the system;
(b) the ultimate capacity of a block containing the piles and raft, plus that of the
portion of the raft outside the periphery of the pile group.
Thus, the relationship between the ultimate geotechnical capacity and the number of
piles will generally have an upper limit once the ‘block’ mode of failure develops.
Conventional design approaches can be used to estimate the various capacities.

16.3.1.2. Lateral loading


The same approach as used for vertical loading will apply for lateral loading i.e. the ulti-
mate capacity is the lesser of the sum of the ultimate lateral capacity of the raft plus that
of all the piles, or the ultimate lateral capacity of a block containing the piles, raft and the
soil, plus the contribution due to that portion of the raft or cap outside the periphery of
the pile group. The following points need to be noted:
(a) the response in both orthogonal lateral directions needs to be considered;
(b) the ultimate lateral capacity of the raft may include both shear resistance at the
underside of the raft and passive resistance of the embedded portion of the raft;
(c) for the ultimate lateral capacity of the piles, both ‘short-pile’ (lateral failure of
the supporting soil) and ‘long-pile’ (yield or failure of the pile itself) modes of
failure need to be considered;
(d) for the ultimate lateral capacity of the pile–soil–raft block, it will generally be
adequate to consider only the ‘short-pile’ failure of the block.
The general form of the relationship between ultimate lateral capacity and the number of
piles will be similar to that for vertical loading, with an upper limit being the ‘block’
capacity of the group. As with vertical loading, the various ultimate capacities can be
assessed from conventional foundation design procedures.

16.3.1.3. Moment loading


The ultimate moment capacity of the piled raft can be estimated approximately as the
lesser of:

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 431

(a) the ultimate moment capacity of the raft (Mur) and the individual piles (Mup);
(b) the ultimate moment capacity of a block containing the piles, raft and soil (Mub).
The ultimate moment capacity of the raft can be estimated using the approach described
by Meyerhof [16.13], from which Lee [16.14] obtains the expression

M ur 27 V   V 1 / 2 
= 1 −    (16.4)
  u  
Mm 4 Vu V

where Mm = maximum possible moment that soil can support


V = applied vertical load
Vu = ultimate centric load on raft when no moment is applied.
Considering loading in the x-direction only, for a rectangular raft, the maximum moment
Mm in the x-direction can be expressed as
pur BL2 (16.5)
Mm =
8
where pur = ultimate bearing capacity below raft
B = width of raft (in y-direction)
L = length of raft (in x-direction).
The ultimate moment contributed by the piles can be estimated from
np
Mup ≈
∑P
i =1
uui xi (16.6)

where Puui = ultimate uplift capacity of typical pile i


| xi | = absolute distance of pile i from centre of gravity of group
np = number of piles.
The ultimate moment capacity of the block can be estimated (conservatively) from the
theory for ‘short-pile’ failure of a rigid pile subjected to moment loading. Poulos and
Davis [16.15] give the solution for ultimate moment capacity MuB (if no horizontal force
is acting) as

MuB = a B pu BB DB2 (16.7)

where BB = width of block perpendicular to direction of loading


DB = depth of block
p–u = average ultimate lateral resistance of soil along block
aB = factor depending on distribution of ultimate lateral pressure with depth
= 0.25 for constant pu with depth
= 0.20 for linearly increasing pu with depth from zero at the surface.

16.3.2. Estimation of load–settlement behaviour of piled rafts


A number of simplified analyses have been developed to estimate the load–settlement
behaviour of a piled raft. The well-known equivalent raft method (e.g. Tomlinson

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
432 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

[16.16]) is one approach that can be adopted, whereby the loading is assumed to be
applied at some distance below the raft and usually over a larger area, to reflect the load
transfer along the piles.
Poulos and Davis [16.15] developed a simplified hand calculation method for
constructing the overall load–settlement curve to failure. Elastic solutions were used for
the initial stiffness of the piled raft and of the raft alone. A tri-linear load–settlement
curve was obtained, reflecting the three main portions of the relationship shown in
Figure 16.2. Only perfectly rigid or perfectly flexible rafts could be considered.
Randolph [16.9], [16.17] developed convenient approximate equations for the stiff-
ness of a piled raft system and the load sharing between the piles and the raft. The
method is restricted to linear behaviour of the piled raft system i.e. the initial portion of
the load–settlement curve. Other approaches with similar concepts have been presented
by Franke et al. [16.18] and van Impe and de Clerq [16.19]. In the latter case, the piled
raft was represented by a series of pile–raft segments having a circular cap. The various
interactions were modelled using elastic theory, and the pile behaviour was given by a
modification of the analysis of Randolph and Wroth [16.20]. While the resulting equa-
tions needed to be solved using a computer, the computation process was simple and did
not require specialised software. A limitation of the approach was that it could consider
only perfectly flexible or perfectly rigid rafts.
A method which combines and extends the approaches of Poulos and Davis [16.15]
and Randolph [16.9], [16.17] is described below. The following aspects are included:

(a) estimation of the load sharing between the raft and the piles, using the approxi-
mate solution of Randolph [16.9];
(b) hyperbolic load–deflection relationships for the piles and for the raft, thus
providing a more realistic overall load–settlement response for the piled raft
system than the original tri-linear approach of Poulos and Davis [16.15].

Figure 16.3 shows diagrammatically the load–settlement relationship for the piled raft.
The point A represents the point at which the pile capacity is fully mobilised, when the
total vertical applied load is VA. Up to that point, both the piles and the raft share the
load, and the settlement (S) can be expressed as

S= V (16.8)
Kpr

where V = vertical applied load


Kpr = axial stiffness of piled raft system.

Beyond point A, additional load must be carried by the raft, and the settlement is given by

VA V − VA (16.9)
S= +
K pr Kr

where VA = applied load at which pile capacity is mobilised


Kr = axial stiffness of raft.

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 433

Vu B

A
VA
Load V

Piles
Vpu

Vru
Raft

SA
Settlement S

Figure 16.3. Construction of load–settlement curve for piled raft

The load VA can be estimated from


Vpu
VA = (16.10)
bp
where Vpu = ultimate capacity of piles (single pile or block failure mode, whichever is
less)
bp = proportion of load carried by piles.
Use can be made of the approximate expressions described by Randolph [16.9] for Kpr in
equation (16.8) and bp in equation (16.10), namely
Kpr = X Kp (16.11)
where Kp denotes the stiffness of pile group alone and, for fairly large numbers of piles,
1 − 0.6 ( K r / K p )
X ≈ (16.12)
1 − 0.64 ( K r / K p )

bp = 1/(1 + a) (16.13)
 Kr 
a≈ 0 .2   (16.14)
1 − 0 .8 ( K r / K p )  K p 
 
If it is assumed that the pile and raft load–settlement relationships are hyperbolic, then
the secant stiffnesses of the piles (Kp) and the raft (Kr) can be expressed as
Kp = Kpi (1 – Rfp Vp /Vpu) (16.15)
Kr = Kri (1 – Rfr Vr /Vru) (16.16)

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
434 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

where Kpi = initial tangent stiffness of pile group


Rfp = hyperbolic factor for pile group
Vp = load carried by piles
Vpu = ultimate capacity of piles
Kri = initial tangent stiffness of raft
Rfr = hyperbolic factor for raft
Vr = load carried by raft
Vru = ultimate capacity of raft.
The load carried by the piles is given by
Vp = bp V ≤ Vpu (16.17)
and the load carried by the raft is
Vr = V – Vp (16.18)
where V denotes the total vertical applied load.
Substituting equations (16.10)–(16.18) in equations (16.8) and (16.9), the following
expressions are obtained for the load–settlement relationship of the piled raft system.
For V ≤ VA:
V
S=
 Rfp b p V 
XK pi 1 −  (16.19)
 Vpu 
 
For V > VA:
V − VA
S = SA +
 (V − Vpu )  (16.20)
K ri 1 − Rfr 
 Vru 
where
VA
SA = (16.21)
XK pi (1 − Rfp )

with VA given by equation (16.10).


Equations (16.19)–(16.21) provide the means for estimating the average load–settle-
ment relationship for the piled raft. Because Kr and Kp will vary with the applied load
level, the parameters X and bp will also generally vary. Thus, it may be necessary to carry
out an iterative or incremental analysis, commencing with the initial stiffnesses Kri and
Kpi. This approach is very amendable to calculation using either spreadsheets or a
proprietary mathematical program such as MATHCAD [16.21].

16.3.2.1. Immediate and final settlements


The above procedure can, in principle, be used to estimate both the immediate and final
settlements of piled rafts in clay. For immediate settlements, the pile and raft stiffnesses
are those relevant to the undrained case, and, if using elastic-based theory, are estimated

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 435

by using the undrained values of modulus and Poisson’s ratio of the soil. For long-term
settlements (immediate plus consolidation settlements, but excluding creep), the pile and
raft stiffnesses are computed using drained values of modulus and Poisson’s ratio.
Long-term ultimate capacities of the raft and the pile group are also relevant. The consol-
idation settlement is then computed as the difference between the total and immediate
settlements.
However, because of the possible non-linearity of behaviour during undrained loading
conditions, the application of the above procedure may not always be accurate. As
suggested by Poulos and Davis [16.15], it may be preferable to calculate the consolida-
tion settlement as the difference between the elastic total final and consolidation
settlements, and add this to the immediate settlement computed from a non-linear anal-
ysis. Thus, the overall total final settlement STF is then

V  1 1 
S TF = +V  − (16.22)
 K ue 
Ku  K e′
where V = applied vertical load on foundation
Ku = undrained foundation stiffness (from non-linear analysis)
Kue = undrained foundation stiffness (from elastic undrained analysis)
Ke′ = drained foundation stiffness (from elastic drained analysis).

16.3.2.2. Differential settlements


The procedure outlined above will generally only give an estimate of the average
load–settlement behaviour of the piled raft foundation. The differential settlements
within the foundation will depend largely on the distribution of applied loads, the
arrangement of the piles, and the relative rigidity of the raft. Randolph [16.9] has
proposed an approximate procedure to estimate the maximum differential settlement of a
uniformly loaded raft foundation, by relating the ratio of differential settlement to overall
settlement, to the relative rigidity of the raft. As a first approximation, this ratio may also
be used for a piled raft foundation. Convenient charts for the estimation of differential
settlement are provided by Horikoshi and Randolph [16.22].

16.3.3. Estimation of pile loads


For the structural design of the piles, an estimate is required of the largest compressive
and tensile forces in the piles. If a limit state design approach is adopted, these forces
will generally be developed under one of the ultimate limit state load combinations. A
first estimate of the axial forces in the piles can be made using an adaptation of the ‘rivet
group’ approach. If the piles carry a proportion bp of the total vertical load, then the axial
force Pi in any pile i in the foundation system can be estimated from
Pi = Vb p / np + M x* xi / I y + M y* yi /I x (16.23)
with
M x − M y I xy / I x M y − M x I xy / I y
M x* = , M y* = (16.24)
1− 2
I xy /( I x I y ) 1 − I xy
2
/(I x I y )

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
436 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

where V = total vertical load acting at centroid of foundation


np = number of piles in group
Mx, My = moments about centroid of pile group in direction of x- and y-axes,
respectively
bp = proportion of load carried by piles
Ix, Iy = moment of inertia of pile group with respect to x- and y-axes,
respectively
Ixy = product of inertia of pile group about centroid
xi, yi = distance of pile i from y- and x-axes, respectively
Mx*, My* = effective moments in x- and y-directions, respectively, taking
symmetry of pile layout into account.
For a symmetrical pile group layout, Ixy = 0 and Mx*= Mx, My* = My. Equation (16.23) then
reduces to
M x xi M y yi
Pi = Vb p / np ± ±
np np
(16.25)
∑2
xi
i =1
yi2
∑i =1

The above approach inherently makes the following assumptions:


(a) the raft is rigid;
(b) the pile heads are pinned to the raft and no moment is transferred from the raft to
the piles i.e. the applied moments are carried by ‘push-pull’ action of the piles;
(c) the piles are vertical.

16.3.4. Estimation of raft moments and shears


It must be recognised that estimates of raft bending moment and shear force are liable to
be very sensitive to the assumptions made and the approximations adopted in any simpli-
fied approach. The precise nature and distribution of the loads also has a very significant
influence on the moments and shears. Structural designers often adopt a very simplistic
approach when designing rafts, and assume that the raft is rigid and that the contact pres-
sures are related linearly to the raft deflections.
A similar approach can be adopted for piled rafts, except that the contact pressures
below the raft balances only the load carried by the raft, with the piles carrying the
remaining load. Alternatively, the piled raft can be considered as a series of piled strip
foundations, with the behaviour of each piled strip being obtained either on the assump-
tion of the strip being rigid, or preferably, using solutions for a strip on an elastic
foundation, with the piles being treated as supports (or negative loads). With such a
simplification it is not possible to estimate the torsional moments in the raft; this requires
a more refined analysis in which the raft is treated as a plate.

16.4. Detailed design


Once the preliminary stage has indicated that a piled raft foundation is feasible, it is
necessary to carry out a more detailed assessment of settlement and decide upon the

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 437

(a)

(b) Pile represented by spring


of equivalent stiffness

At element representing pile:


(i) stiffness is increased;
(c) (ii) pile force is 'smeared' over element;
(iii) limiting compressive and tensile pile–soil
stresses are computed from compressive
and tensile capacities of pile, respectively.

Figure 16.4. Modelling of piled strip foundation: (a) actual pile; (b) pile representation;
(c) assumed contact pressures

optimum locations and arrangement of the piles. The raft bending moments and shears,
and the pile loads, should also be obtained for the structural design of the foundation.
Two broad classes of detailed analysis methods will be considered below:
(a) approximate computer-based methods;
(b) more rigorous computer-based methods.

16.4.1. Approximate computer-based analyses


16.4.1.1. Methods employing a strip-on-springs approach
A typical method in this category is that presented by Poulos [16.10] and illustrated in
Figure 16.4. A section of the raft is represented by a strip, and the supporting piles by
springs. Approximate allowance is made for all four components of interaction
(pile–raft, raft–pile, raft–raft, pile–pile), and the effects of the parts of the raft outside the
strip section being analysed are taken into account by computing the free-field soil
movements due to these parts and interacting these with the strip section. The method is
versatile and has been shown to give reasonable agreement with more complete anal-
yses. However, it does have significant limitations, especially as it cannot consider

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
438 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

torsional moments within the raft, and also because it may not give completely
consistent settlements at a given point if strips in two directions through that point are
analysed.
Brown and Wiesner [16.23] and Wiesner and Brown [16.24] have developed
boundary element analyses for a piled strip, and then suggested how such solutions may
be applied to a piled raft.

16.4.1.2. Methods employing a plate-on-springs approach


In this type of analysis, the raft is represented by an elastic plate, while the piles are
modelled as springs supporting the plate. Some early approaches in this category e.g.
Hongladaromp et al. [16.25], neglected some interaction effects and hence gave stiff-
nesses that were too large, as revealed by studies made by Brown et al. [16.26] who
compared such methods with the results of more complete analyses. Poulos [16.27]
employed a finite difference method for the plate and took account of the various interac-
tions in approximate elastic solutions. Allowance was also made for the effects of piles
reaching their ultimate capacity, the development of bearing capacity failure below the
raft, and the presence of free-field vertical soil movements acting on the foundation
system.
Clancy and Randolph [16.28] have adopted a more refined approach in which each
pile is modelled as a series of rod finite elements, while the raft is analysed using
two-dimensional thin-plate finite elements. The four components of interaction are taken
into account, although the method is restricted to analysing the elastic response of the
foundation. A similar type of analysis has been outlined by Franke et al. [16.18], but
where non-linear pile behaviour is considered via hyperbolic shaft and base response
characteristics. Yamashita et al. [16.29] mention an approach in which finite elements
are used to analyse the raft, while the soil and piles are represented by appropriate
springs. The means by which the various interactions are considered are not detailed.

16.4.2. More rigorous numerical methods of analysis


This category includes methods in which the various components of the piled raft system
are modelled in more detail than in the above categories. All methods are clearly
dependent on computer analysis, and many involve the use of special-purpose software.
Two main numerical techniques have been employed; the boundary element method and
the finite element method, sometimes in combination.

16.4.2.1. Boundary element methods


In this type of approach, both the raft and each pile within the foundation system are
discretised. An early example of such analysis was that of Butterfield and Banerjee
[16.30], who studied groups of piles in an elastic soil mass with a rigid cap resting on the
surface. Kuwabara [16.31] described an analysis for a piled raft in a homogeneous elastic
soil mass. The raft was assumed to be rigid and the compressibility of the piles was taken
into account. It was found that under elastic conditions, the raft carried only a small
proportion of the load at normal pile spacings. Poulos [16.11] extended Kuwabara’s
analysis to allow for the effects of free-field soil movements and for limiting contact

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 439

pressures between raft and soil, as well as for the development of ultimate compression
or tension loads in the piles. However, the limitation of a rigid raft remained, as is the
case for almost all methods in this category.

16.4.2.2. Methods combining boundary element and finite element analyses


Hain and Lee [16.32] published a seminal paper on piled raft analysis in which they
represented the raft as a series of thin-plate finite elements, while the characteristics of
the piles were computed from boundary element analyses. Use was made of the interac-
tion factor concept to reduce the computational effort; thus, not every element of every
pile was represented. However, the solutions obtained by Hain and Lee have been used
by many subsequent researchers as reference solutions, despite the fact that they involve
approximations. The analysis was essentially elastic, but was capable of including the
development of ultimate loads in piles via a load ‘cut-off’ limit in their computer
program. A limitation of this analysis is that the soil was represented by a homogeneous
semi-infinite elastic mass. However, their solutions provided a ready appreciation of the
importance of several key factors on piled raft response in the early stages of loading.
Franke et al. [16.18] describe a ‘mixed’ technique involving the use of boundary
element analysis for the piles, and finite element analysis of the raft. A non-linear
response of the piles is allowed for, as are residual stresses in the piles after installation.
Sinha [16.12] has developed a complete boundary element analysis in which all piles
are discretised and analysed using the boundary element method, while the raft is repre-
sented by thin-plate finite elements. The soil is assumed to be a homogeneous elastic soil
mass, but allowance for non-linear behaviour is made by including limiting raft–soil
contact pressures in both compression and tension, and also by specifying limiting
stresses between the pile shaft and soil, and beneath the pile tip. The effects of free-field
soil movements can also be taken into account, thus allowing the effects of soil swelling
or consolidation to be analysed.

16.4.2.3. Simplified finite element analyses


Simplified finite element analyses usually involve the representation of the pile group or
piled raft as either a plane strain problem (e.g. Desai [16.33]) or as an axisymmetric
problem (e.g. Hooper [16.3], Naylor and Hooper [16.34]). In each case, finite elements
are used to discretise both the raft and the soil, and it is therefore a relatively simple
matter to take account of non-linear soil and raft behaviour. Two-phase behaviour of the
soil can also be incorporated, so that time-dependency of settlement and pile load distri-
bution due to consolidation of the soil can be computed. Hooper [16.3] has shown that,
for a building in London, an axisymmetric representation predicts foundation behaviour
which is similar to that observed. The main problems in such a simplified approach are
that only regular loading patterns can be analysed, and that it is not possible to obtain
torsional moments in the raft.
Although not strictly a finite element analysis, Hewitt and Gue [16.35] have carried
out an analysis of a piled raft founded in ground containing karstic limestone. The anal-
ysis was performed with the commercially available program FLAC [16.36], which uses
an explicit finite difference code. The foundation was modelled as a plane strain

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
440 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

problem, and this analysis, although approximate, enabled the effects of cavities in the
limestone on the settlement of the foundation to be investigated.

16.4.2.4. Three-dimensional finite element analysis


In terms of the ability to model a real problem, three-dimensional finite element methods
are usually considered to be the ‘ultimate weapon’, at least as far as the analysis is
concerned (the problem of assigning appropriate parameters still remains, of course).
Ottaviani [16.37] appears to have been the first to apply such an analysis to pile founda-
tions. Zhuang et al. [16.38] and Lee [16.39] have used a linear three-dimensional
analysis to derive parametric solutions for the settlement and load distribution within
piled rafts. Among the parameters varied were the relative raft stiffness, the pile length,
and the number of piles.
Ta and Small [16.40] have developed a method involving the use of thin-plate finite
elements for the raft and a finite layer method for the soil. This method is limited to
linear soil behaviour but can handle a layered soil system very efficiently, and can also
handle piles located anywhere beneath the raft.
One of the most complete analyses which appears to have been undertaken to date is
that by Wang [16.41], who carried out a non-linear analysis of vertically loaded piled
rafts. Clearly, the computational effort is substantial in such an analysis, although the
results may provide benchmark solutions against which simpler analyses can be
checked. Indeed, for a rigid raft, the analysis revealed that simple elastic methods of
analysis using interaction factors [16.15] gave settlement and pile load distributions in
good agreement with those from the three-dimensional analysis. In addition, some of the
characteristics of foundation behaviour, which may not be revealed using simpler
methods, can be discerned e.g. the lateral response of the piles, even though the loading
is vertical, and the non-symmetric distribution of stresses along each pile. There is,
however, a price to pay when using this approach, in terms of the effort required to
prepare the data and to run the analysis. Some idea of the time involved may be obtained
from the non-linear three-dimensional finite element analysis of a laterally loaded 9-pile
group by Kimura and Adachi [16.42]. For a mesh containing 4200 nodes and 3432
elements, they reported a total run time of 85 hours on a SPARC II work-station for a
total of 100 load increments.

16.4.3. Comparison of capabilities of methods


As a means of summarising the capabilities of some of the methods mentioned above
(including the simplified methods), Table 16.1 lists their main features and their ability
to predict the response of the foundation system as outlined in section 16.2.1.
Poulos et al. [16.43] have compared some of these methods when applied to the ideal-
ised hypothetical problem shown in Figure 16.5. The following six methods were used:
(a) simplified non-linear method of Poulos and Davis [16.15];
(b) simplified linear method of Randolph [16.9];
(c) strip-on-springs analysis, using the program GASP (Geotechnical Analysis of
Strip with Piles), after Poulos [16.10];

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
Table 16.1. Capabilities of various methods of piled raft analysis

Response characteristics Problem modelling

Raft Raft Non-linear Non-linear Non-uniform


Total Differential Pile bending torsional soil pile soil Raft
Method settlement settlement load moment moment behaviour behaviour profile flexibility

Poulos & Davis [16.15] x x


Randolph [16.17] x x
van Impe & de Clerq [16.19] x x
Equivalent raft x x
Poulos [16.10] x x x x x x x x
Brown & Wiesner [16.23] x x x x x
Clancy & Randolph [16.28] x x x x x x x
Poulos [16.52] x x x x x x x x x
Kuwabara [16.31] x x

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
Hain & Lee [16.32] x x x x x x x
Sinha [16.12]; Franke et al. [16.18] x x x x x x x x
Hooper [16.3] x x x x x x x x
Hewitt & Gue [16.35] x x x x x x
Lee [16.39] x x x x x x
Ta & Small [16.40] x x x x x x x
Wang [16.41] x x x x x x x x x

Note: x indicates capability


PR AC TIC AL DE SIG N PRO CE DU RES FO R PI LED R AFT FO UN DA TI ON S
441
442 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

A A 1m
V1 V2 V1
2m
A A
V1 V2 V1 x
2m Bearing capacity of raft = 0·3 MPa
A A Load capacity of each pile
V1 V2 V1 1m = 0·873 MN (Compression)
= 0·786 MN (Tension)

1m 2m 2m 2m 2m1m

Ep = Er = 30 000 MPa V2
νp = νr = 0·2 V V1
1
t = 0·5 m

E = 20 MPa
l = 10 m
ν = 0·3

H = 20 m

d = 0·5 m
s=2m 2m 2m 2m

Figure 16.5. Hypothetical example used to compare results of various methods of piled
raft analysis

(d) plate-on-springs approach, using the program GARP (Geotechnical Analysis of


Raft with Piles), after Poulos [16.27];
(e) finite element and boundary element method of Ta and Small [16.40];
(f) finite element and boundary element method of Sinha [16.12].
Figure 16.6 compares the computed behaviour of a raft supported by 9 piles, one under
each column, when the overall factor of safety at the design load is 2.15. The total
applied load is 12 MN, with V1 = 1 MN and V2 = 2 MN, and piles labelled A are not
present. These loads exceed the ultimate capacity of the piles alone, and hence there is
some non-linear behaviour. Despite differences among the various methods, most of
those which incorporate non-linear behaviour give somewhat similar results for settle-
ment and load sharing, although there are significant differences among the computed
raft bending moments. However, it would appear that, provided the analysis method is
soundly based and takes into account the limited load capacity of the piles, similar
results may be expected for similar input parameters.

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 443

50 10

Differential settlement: mm
Average settlement: mm

40 8

30 6

F.E. + B.E. Sinha

F.E. + B.E. Sinha


Poulos & Davis

F.E. Ta & Small

F.E. Ta & Small


Plate (GARP)

Plate (GARP)
Strip (GASP)

Strip (GASP)
20 4
Randolph

10 2

0 0
(a) (b)

1·2 100
Maximum moment: MNm/m

1·0 80
Load on piles: %

0·8
60

F.E. + B.E. Sinha


F.E. + B.E. Sinha

0·6

F.E. Ta & Small


F.E. Ta & Small

Plate (GARP)
Plate (GARP)

Strip (GASP)
Strip (GASP)

40
0·4
Randolph
20
0·2

0 0
Method Method
(c) (d)

Figure 16.6. Comparative results for hypothetical example (raft with 9 piles, total
applied load 12 MN): (a) average settlement; (b) differential settlement (centre to
mid-point of shorter side); (c) maximum bending moment Mx; (d) proportion of load
carried by piles

16.5. Further characteristics of foundation behaviour


In order to examine some of the characteristics of piled raft behaviour, a more detailed
study has been made of the preceding hypothetical case. The ‘standard’ parameters
shown in Figure 16.5 have been adopted, but consideration has been given to the effects
on foundation behaviour of variations in the following parameters:
(a) the number of piles;
(b) the nature of the loading (concentrated or uniformly distributed);
(c) raft thickness;
(d) applied load level.
The analyses have been carried out using the computer program GARP [16.27]. This
program has the capability of considering the following factors:
(a) non-homogeneous or layered soil profile;
(b) limiting pressures below the raft, in both compression and uplift;
(c) non-linear pile load–settlement behaviour, including limiting pile capacity in
compression and tension;

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
444 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

(d) piles of different stiffness and load capacity within the foundation system, with
easy alteration of the location and numbers of piles;
(e) applied loading consisting of concentrated loads, moments, and areas of uniform
loading;
(f) effects of free-field vertical soil movements, such as those arising from consoli-
dation or soil swelling.
For the case analysed, the raft has been divided into 273 elements, and for simplicity,
both the raft and the piles have been assumed to exhibit elastoplastic load–settlement
behaviour. The stiffness and interaction characteristics of the piles have been computed
from a separate computer analysis using the program DEFPIG [16.44]. For the purposes
of this example, the length and diameter of the piles have been kept constant.

16.5.1. Effect of number of piles and type of loading


Figure 16.7 shows the effect of the number of piles on maximum settlement, differential
settlement, maximum bending moment, and the proportion of load carried by the piles.

60 10
Concentrated
50 loading
Maximum settlement: mm

Differential settlement: mm

8
Uniform loading
40 Concentrated
loading 6
30
4
20 Uniform loading

2
10

0 0
(a) (b)

1 100
Concentrated loading
Maximum moment Mx: MNm/m

Uniform loading
0·75 75
Load on piles: %

Concentrated and
uniform loading
0·5 50

0·25 25

0 0
0 10 20 30 40 50 0 10 20 30 40 50
Number of piles Number of piles
(c) (d)

Figure 16.7. Effect of number of piles on piled raft behaviour for hypothetical example
(total applied load 12 MN): (a) maximum settlement; (b) differential settlement (centre to
corner piles); (c) maximum bending moment Mx; (d) proportion of load carried by piles

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 445

The raft thickness in this case is 0.5 m and the total applied load is 12 MN. The first
three piles are located below the central loads (on the y-axis), the next six are below the
outer loads, and then any additional piles are located evenly between the loads. Both
concentrated loading (V1 = 1 MN and V2 = 2 MN in Figure 16.5) and a uniformly
distributed load (0.2 MPa) have been analysed. The following characteristics are
observed.
1. The maximum settlement decreases with increasing number of piles, but becomes
almost constant for 20 or more piles.
2. For small numbers of piles, the maximum settlement for concentrated loading is
larger than for uniform loading, but the difference becomes very small for 10 or
more piles.
3. The differential settlement between the centre and corner columns does not
change in a regular fashion with the number of piles. For the cases considered, the
smallest differential settlements occur when only 3 piles are present, located
below the central portion of the raft. The largest differential settlement occurs for
9 piles, because the piles below the outer part of the raft ‘hold up’ the edges which
do not settle as much as the centre.
4. The maximum bending moments for concentrated loading are substantially
greater than for uniform loading. Again, the smallest moment occurs when only 3
piles, located under the centre, are present.
5. The percentage of load carried by the piles increases with increasing pile
numbers, but for more than about 15 piles, the rate of increase is very small. The
type of loading has almost no effect on the total load carried by the piles, although
it does of course influence the distribution of load among the piles.

16.5.2. Effect of raft thickness


Figure 16.8 shows the effect of raft thickness on raft behaviour, for the case of concen-
trated loading shown in Figure 16.5 (V1 = 1 MN, V2 = 2 MN). Neither the maximum
settlement nor the percentage of load carried by the piles is very sensitive to raft thick-
ness. However, as would be expected, increasing the raft thickness reduces the
differential settlement between the centre and corner columns, but generally increases
the maximum bending moment. It should be noted that for zero piles (i.e. raft only), the
raft behaviour is quite non-linear for small raft thicknesses, and the development of
plastic zones below the raft tends to reduce the differential settlement. Once again, the
raft with only 3 piles performs very well, and clearly demonstrates the importance of
locating the piles below those parts of the foundation which most require support. This is
in accordance with the philosophy of designing piled rafts for differential settlement
control.

16.5.3. Effect of load level on settlement


Figure 16.9 shows computed load–settlement curves for the 0.5 m thick piled raft with
various numbers of piles, for the case of concentrated loading (Figure 16.5). Clearly, the

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
446 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

Number
of piles
0

15

45

100 20

Differential settlement: mm
Maximum settlement: mm

75 15

50 10

25 5

0 0
(a) (b)

1 100
Maximum moment Mx: MNm/m

0·75 75
Load on piles: %

0·5 50

0·25 25

0 0
0 0·25 0·5 0·75 1 0 0·25 0·5 0·75 1
Raft thickness: m Raft thickness: m
(c) (d)

Figure 16.8. Effect of raft thickness on piled raft behaviour for hypothetical example
(total applied load 12 MN): (a) maximum settlement; (b) differential settlement (centre
to corner columns); (c) maximum bending moment Mx; (d) proportion of load carried by
piles

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 447

60

Number of piles
below raft
48
45

25
36
Total load: MN

15
9
24
3

0
12

0
0 50 100
Central settlement: mm

Figure 16.9. Load–settlement curves for various piled raft foundation systems (concen-
trated loading; see Figure 16.5)

settlement increases with increasing load level, and the beneficial effect of adding piles
as the design load level increases are obvious. Provided that there is an adequate safety
margin, the addition of even a relatively small number of piles can lead to a considerable
reduction in the maximum settlement of the foundation.

16.5.4. Summary
The foregoing simple example demonstrates the following important points for practical
design.
1. Increasing the number of piles, while generally of benefit, does not always
produce the best foundation performance, and there is an upper limit to the useful
number of piles, beyond which very little additional benefit is obtained.
2. The raft thickness affects differential settlement and bending moments, but has
little effect on load sharing or maximum settlement.
3. For control of differential settlement, optimum performance is likely to be
achieved by the strategic location of a relatively small number of piles, rather than
using a large number of piles evenly distributed over the raft area, or increasing
the raft thickness.
4. The nature of the applied loading is important for differential settlement and
bending moment, but generally is much less important for maximum settlement or
load sharing between the raft and the piles.

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
448 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

16.5.5. Guidelines for economical design


Horikoshi and Randolph [16.45] have studied the performance of piled rafts under
vertical load and examined several of the factors considered above. They have suggested
the following guidelines for the ‘optimum design’ of uniformly loaded rafts.
1. Piles should be distributed over the central 16–25% of the raft area.
2. The pile group (or equivalent pier) stiffness should be approximately equal to the
axial stiffness of the raft alone.
3. The total pile capacity should be designed for between 40 and 70% of the design
load, depending on the ratio of the area occupied by the pile group to that of the
raft, and the Poisson’s ratio of the soil. The degree of mobilisation of pile capacity
should not exceed about 80%, to avoid significant increases in differential
settlement.
For concentrated loading, some of the above guidelines may not be appropriate, espe-
cially in relation to the concentration of the piles near the centre of the raft, but in
general, they provide a useful starting point for design.

16.6. Geotechnical parameter assessment


The design of a piled raft foundation requires an assessment of a number of geotechnical
and performance parameters, including:
(a) raft bearing capacity;
(b) pile capacity;
(c) soil modulus for raft stiffness;
(d) soil modulus for pile stiffness.
While there are a number of laboratory and in-situ procedures available for the assess-
ment of these parameters, it is common for at least initial assessments to be based on the
results of simple field tests such as the Standard Penetration Test (SPT) and the Static
Cone Penetration Test (CPT). Typical of the correlations which the author has employed
are those based on the work of Décourt [16.46], [16.47] using the SPT, namely:
raft ultimate bearing capacity; pur = K1 Nr (kPa) (16.26)
pile ultimate shaft resistance; fs = a (2.8 Ns + 10) (kPa) (16.27)
pile ultimate base resistance; fb = K2 Nb (kPa) (16.28)
soil Young’s modulus below raft; Esr = 2Nr (MPa) (16.29)
soil Young’s modulus along pile; Es = 3Ns (MPa) (16.30)
soil Young’s modulus below pile tip; Es = 3Nb (MPa) (16.31)
where Nr = average SPT-value (corrected to 60% energy ratio) within a depth of
one-half of raft width
Ns = average SPT-value along pile shaft
Nb = average SPT-value close to pile tip
K1, K2 = factors shown in Table 16.2
a = 1 for displacement piles in all soils and non-displacement piles in clay
= 0.5–0.6 for non-displacement piles in granular soil.

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 449

Table 16.2. Correlation factors K1 and K2 for ultimate bearing capacity

K2 : Displacement K2: Non-displacement


Soil type K1 : Raft piles piles

Sand 90 325 165


Sandy silt 80 205 115
Clayey silt 80 165 100
Clay 65 100 80

16.7. Typical applications


16.7.1. Westend Tower, Frankfurt, Germany
The Westend 1 Tower is a 51-storey, 208 m high building in Frankfurt, and has been
described by Franke [16.8] and Franke et al. [16.18]. A cross-section and foundation
plan of the building are shown in Figure 16.10. The foundation for the tower consists of
a piled raft with 40 piles, each about 30 m long and 1.3 m in diameter. The central part of
the raft is 4.5 m thick, decreasing to 3 m at the edges. While full details of the geotech-
nical profile are not available in the published literature, it appears that the building is
located on a thick deposit of relatively stiff Frankfurt clay. On the basis of pressuremeter
tests, an average reloading soil modulus of 62.4 MPa has been reported by Franke et al.
[16.18].

Main Side building raft


tower

Side
208 m building

Main tower raft


(40 piles)

(b)
15 m

30 m

(a)

Figure 16.10. Westend 1 Building, Frankfurt, Germany (after Franke et al. [16.18]):
(a) sectional elevation; (b) foundation plan

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
450 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

200 80

150 60
Settlement: mm

Pile load: %
100 40

Poulos & Davis

Franke et al.
Franke et al.

Ta & Small
Ta & Small

Measured
Measured
Randolph

Randolph
50 20

GARP
GARP

GASP
GASP

Sinha
Sinha
0 0
(a) (b)

20 20
Maximum pile load: MN

Minimum pile load: MN


15 15

10 10

Franke et al.
Franke et al.

Ta & Small
Ta & Small

Measured
Measured

5 5
GARP

GASP
GARP

GASP

Sinha
Sinha

0 0
Method Method
(c) (d)

Figure 16.11. Comparison of analysis methods for piled raft foundation, Westend 1
Tower, Frankfurt, Germany: (a) central settlement; (b) proportion of load carried by
piles; (c) maximum pile load; (d) minimum pile load

Calculations have been reported by Poulos et al. [16.43] to predict the behaviour of the
building, using a number of different analysis methods:

(a) a finite element analysis (Ta and Small [16.40]);


(b) the GARP analysis described earlier in this chapter;
(c) a piled strip analysis (Poulos [16.10]);
(d) the simple hand calculation method described by Poulos and Davis [16.15];
(e) the approximate linear method developed by Randolph [16.9], [16.17];
(f) the combined finite element and boundary element method developed by Sinha
[16.12];
(g) the combined finite element and boundary element method described by Franke
et al. [16.18].

Figure 16.11 compares the predicted performance using the above methods with the
measured values. The calculations have been carried out for a total vertical load of
968 MN, which is equivalent to an average applied pressure of 323 kPa.
The following points are noted.

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 451

1. The measured maximum settlement is about 105 mm, and most methods tend to
over-predict this settlement. However, most of the methods provide an acceptable
design prediction.
2. The piles carry about 50% of the total load. Most methods tended to over-predict
this proportion, but from a design viewpoint, most methods give acceptable
estimates.
3. All methods capable of predicting the individual pile loads suggest that the load
capacity of the most heavily loaded piles is almost fully utilised; this is in agree-
ment with the measurements.
4. There is considerable variability in the predictions of minimum pile loads. Some
of the methods predicted larger minimum pile loads than were actually
measured.

This case history clearly demonstrates that the design philosophy of fully utilising pile
capacity can work successfully and produce an economical foundation which performs
satisfactorily. The available methods of performance prediction appear to provide a
reasonable, if conservative, basis for design in this case.

16.7.2. Messe Turm Tower, Frankfurt, Germany


This building is one of the pioneering structures designed to be supported on a piled raft
foundation. It has been described extensively in the literature e.g. Sommer et al. [16.48];
Sommer [16.49], El-Mossallamy and Franke [16.50], Tamaro [16.51]. The Messe Turm
tower block is 256 m high and, at the time of its construction, was the tallest building in
Europe. It is supported by a raft 6 m thick in the central portion, decreasing to 3 m at the
edges. There are 64 piles arranged in three concentric circles below the raft. The piles are
1.3 m diameter, and vary in length from 26.9 to 34.9 m. The distance between the piles
varies from 3.5 to 6 pile diameters. Figure 16.12 shows details of the foundation. The
piles were designed to develop their full geotechnical capacity and to carry about
one-half of the design load.
Extensive instrumentation was installed to monitor foundation performance, with
measurements including foundation settlement and rotation, sub-surface settlement,
pile-head loads, and the distribution of load along the length of the piles.
The foundation behaviour was complicated by drawdown of the groundwater table
arising from a nearby subway excavation. Figure 16.13 shows the measured time–settle-
ment behaviour of the tower [16.51], and indicates that the total settlement of the
building was about 115 mm at the end of 1995, approximately 7 years after the
commencement of construction. Also shown in Figure 16.13 is the predicted time–settle-
ment behaviour, which agrees reasonably well with the measurements.
This project again demonstrates the feasibility of designing piled raft foundations with
the piles developing their full capacity. Tamaro [16.51] gives an interesting comparison
between the Messe Turm Tower and the Commerz Bank building, which is now the
tallest building in Europe (300 m). The foundations of the latter building have been
designed as conventional piled foundations, with some account being taken of the

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
452 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

Core Exterior corner


column

14 m
6m

34·9 m
Total vertical load
1880 MN

64 piles 1·3 m diameter

58·8 m 28 piles 26·9 m long


20 piles 30·9 m long
16 piles 34·9 m long

Figure 16.12. Piled raft foundation for Messe Turm Tower, Frankfurt, Germany

contribution of the 2.5 m thick raft. Table 16.3 shows this comparison, and highlights the
economy which can be achieved in the piling by designing the piles to act primarily as
settlement reducers, and to operate at or close to their full capacity.

Table 16.3. Comparison between foundation characteristics of two building towers in


Frankfurt, Germany

Quantity Messe Turm Tower Commerz Bank

Total load (MN) 1880 1300


Effective pressure (kPa) 470 550
Base area (m2) 3457 2150
Bottom of excavation (m) –14 –7
Number of piles 64 111
Pile diameter (m) 1.3 1.8 (upper 25 m)
1.5 (below 25 m)
Pile length (m) 26.9–34.9 45
Predicted settlement (mm) 150–200 60–70
Measured settlement (mm) 115 19
Raft thickness (m) 6 (to 3 at edge) 4.45 beneath cores
2.5 between cores

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 453

1600

Assumed for calculations


Building load: MN

1200

800
Actual
400

1988 1989 1990 1991 1992 1993 1994 1995 1996


0

1st subway 2nd subway


40 dewatering dewatering
Settlement: mm

80
Measured settlement
at raft centre
120

160
Calculated range
of settlement
200

Figure 16.13. Calculated and measured settlements for Messe Turm Tower, Frankfurt,
Germany (after Tamaro [16.51])

16.7.3. Akasaka building, Sao Paulo, Brazil


Poulos [16.52] has described the application of the piled raft design concept to the
Akasaka commercial building in Sao Paulo. The building consists of a multi-storey
block, occupying a total rectangular footprint of 44.5 m by 26.8 m. The foundations
consist of individual footings below each column, with piles below the more heavily
loaded columns to reduce differential settlements. Figure 16.14 shows the foundation
plan and the geotechnical profile.
Analyses were carried out for footing SP11, in order to assess the number of piles
required to satisfy the design requirement of a maximum settlement of 30 mm. Precast
reinforced concrete piles 520 mm in diameter, and extending about 12 m below the base-
ment raft level, were assumed in design. The estimated ultimate load capacity of each
pile was about 2500 kN.
Preliminary design calculations were carried out to give the required number of piles
for various values of the factor of safety of the piles. A conventional design approach,
assuming a safety factor of 2.5 for the piles and ignoring any load on the base of the
footing, would require 23 piles. In a design based on the concept of full utilisation of pile
capacity, only 8 piles are required, and the overall factor of safety for the piles and the
footing is 2.5.

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
454 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

S4
S2
Rua das Caneleiras

S5

SP11

S1 S3

Scale:
0 10 m Indicates borehole location

(a)

SPT N -value
0 10 20 30 40 50 60 70 80 90 100

0 'Porous' silty clay S1


S2
Medium clay S3
S4
–5 Medium dense clay S5

Stiff clay Foundation level


Datum level: m

–10 Clayey sand

Stiff–hard clay
–15

Fine–medium
dense sand
–20
Stiff clay
Dense sand
–25
(b)

Figure 16.14. Akasaka building, Sao Paulo, Brazil: (a) foundation plan; (b) typical
geotechnical profile

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 455

Overall safety factor


1·57 2·0 2·5 3·0
60

Raft thickness
50
t=1m
t = 0·75 m
t = 0·5 m
Maximum settlement: mm

40

30

20

10

0
0 5 10 15 20
Number of piles

Figure 16.15. Computed relationship between maximum settlement and number of piles,
Akasaka building, Sao Paulo, Brazil

For a detailed analysis of the various design options, the program GARP was used,
with the geotechnical parameters being estimated on the basis of correlations with SPT
data [16.46]. Figure 16.15 shows the computed variation of maximum settlement with
the number of piles, for raft thicknesses of 0.5, 0.75 and 1 m. In this case, the settlement
ranged from over 50 mm for an unpiled footing, to about 20 mm for some 10 or more
piles. The characteristics of behaviour are very similar to those in Figures 16.7 and 16.8;
that is, there is little benefit in adding piles beyond a certain number (in this case, about
10), and there is little effect of raft thickness on the maximum settlement.
For a maximum settlement of 30 mm, Figure 16.15 indicates that only about 6 piles
would be required; such a foundation system would have an overall factor of safety of
about 2.25, and was in fact recommended by the consulting engineer on the project as the
appropriate design for that foundation.

16.7.4. Piled strip foundation for printing facility, Sydney, Australia


A printing press facility was constructed for a large international publishing organisation
in the Sydney suburb of Chullora during 1993–94. The printing press for this facility was
a highly sophisticated and expensive piece of equipment with a low tolerance to differen-
tial settlements. Extensive geotechnical and geophysical investigations were carried out
to characterise the site and to facilitate the development of geotechnical models for
various areas.

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
456 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

Beneath the press line, the geotechnical profile consisted primarily of a thin layer of
stiff residual clay underlain by shale which varied from extremely weathered to fresh,
and was in turn underlain at depth by fresh sandstone. In order to obtain data on the stiff-
ness of these strata, use was made of both in-situ seismic shear wave measurements and
pressuremeter testing. The inferred values of Young’s modulus from these tests are
shown in Figure 16.16(a). The values from the pressuremeter tests were obtained from
unload–reload loops.
Soil stiffness values were required for two purposes:
(a) to assess the dynamic stiffness and damping of the foundation under dynamic
forces imposed by the machinery;
(b) to assess the long-term settlement of the printing press foundation.
For the dynamic response, values of Young’s modulus were assessed primarily from the
results of the geophysical shear-wave measurements. For the long-term settlements, it
was recognised that both strain-level effects and time effects from consolidation and
creep would reduce the Young’s modulus considerably. Consequently, long-term
modulus values were assessed to be one-quarter of the values used for the dynamic
response analyses when considering the settlement of the strip footings, and one-third of
the dynamic values when considering the long-term settlement and stiffness of the piles.
Figure 16.16(b) summarises the key features of the geotechnical model developed for
design.
Two foundation alternatives were investigated for the press line.
1. A strip foundation about 117 m long without piles; widths of 5 m and 9 m were
considered, with a thickness of 0.8 m.
2. A strip foundation, with piles extending either 7.5 m or 12 m below the base of a
5 m wide strip, 0.8 m thick. The piles were 0.9 m in diameter and arranged in
pairs beneath each pair of columns, which were spaced about 4.3 m apart.
The computer program GASP [16.10] was used to carry out the analyses of the various
foundation options. For the long-term loading case, the computed settlement profiles
along the press line foundation are shown in Figure 16.17 for the piled raft options, and
for the corresponding 5 m wide raft without piles. These results show that both total and
differential settlements for the piled raft are significantly less than for the raft alone. It is
also evident that the benefits of using the longer 12 m piles are only marginal.
Additional analyses were carried out to assess whether there would be benefits in
replacing the upper layer of residual clay by a compacted and stabilised layer of much
greater stiffness (Young’s modulus 200 MPa). These analyses indicated that:
(a) without piles, the total and differential settlements are reduced but they are still
considerably greater than for the piled strip;
(b) with piles, there is little benefit in replacing the clay layer.
As a consequence of these analyses, it was decided to use a strip foundation 5 m wide
and 0.8 m thick, with pairs of bored piles of 0.9 m diameter, typically extending about
7.5 m into moderately weathered shale. It was also decided to remove the upper clay

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 457

Young's modulus: MPa


0 1000 2000 3000 4000 5000 6000 7000
40
From shear-wave measurements
Symbol BH
PF1
35 PF3
Level of base of press line PF4
PF5
PF7
Datum level: m

30
From pressuremeter tests

25
Note: Values plotted for mean
datum level of shear-wave
measurements
20

15
(a)

Long-term Long-term
Dynamic
Young's modulus Young's modulus
Description Young's modulus:
for raft: for piles:
MPa
MPa MPa
33
Residual clay 150 38 50

Shale, very
weak to weak,
400 100 130
extremely
weathered
28
Datum level: m

Shale, various
classes, weak 1400 350 470
to medium

1800 450 600


23

5000 1250 1660

Shale, medium to
strong, relatively 6000 1500 2000
fresh
18
(b)

Figure 16.16. Ground stiffness data, Printing Facility, Sydney, Australia: (a) values
of Young’s modulus from in-situ tests; (b) geotechnical model adopted for analysis

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
458 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

B A A C C C A A A A A C C C A A A A A C C C A A A A C B

0
12 m long, 0·9 m diameter piles

2
Computed settlement: mm

7·5 m long, 0·9 m diameter piles

4
Long-term loads A 2400 kN
on column pairs: B 1200 kN
C 1800 kN
6

No piles

10

Figure 16.17. Computed settlement profiles for strip foundation (117 m long, 5 m wide,
0.8 m thick) for Printing Facility, Sydney, Australia

layer beneath the foundation to avoid the possibility of undesirable shrink–swell move-
ments. However, rather than stabilising the replacement fill, the specification was that it
need only be properly compacted and non-reactive.
Although detailed settlement measurements have not been undertaken on the press
line, the continuing high quality of colour printing emanating from the printing press
suggests that the foundation is performing its function of limiting both dynamic and
long-term movements to tolerable values.

16.8. Worked example


To illustrate the application of the design procedure outlined in this chapter, the example
shown in Figure 16.18 is considered. The problem has deliberately been kept simple to
avoid unnecessary complication, although it is still not possible to present the complete
working because of the limitations of space.
It is required to assess the adequacy of the piled raft foundation shown in Figure 16.18,
with respect to the following design criteria:

(a) a minimum overall factor of safety of 2.5 against bearing capacity, overturning
and lateral failure for the ultimate load case;
(b) a maximum long-term average settlement of 50 mm and a maximum differential
settlement not exceeding 10 mm.

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 459

V
Mx

Hx
t = 0·5 m

15 m

25 m
Eu = 30 MPa
E ′ = 15 MPa
ν′ = 0·3 Ultimate loading
d = 0·6 m
V = 20 MN
Mx = 25 MNm
Hx = 2 MN

Long-term loading
y V = 15 MN
1m 4m 4m 1m Mx = 0
Hx = 0

1m
2m
6m
2m x

1m

10 m

Figure 16.18. Piled raft foundation used in worked example

All 9 precast concrete piles are of 0.6 m diameter, each driven to a penetration depth of
15 m. The average ultimate shaft friction is assumed to be 60 kPa in compression and
42 kPa in tension, while the ultimate end-bearing capacity is 900 kPa. For each pile, the
ultimate axial capacities are then calculated to be 1.95 MN in compression and 1.20 MN
in tension.

16.8.1. Assessment of vertical load capacity


For the raft, it is assumed that the ultimate bearing capacity is 6cu = 0.6 MPa, where cu
denotes the undrained shear strength of the soil. The total bearing capacity of the raft is
therefore 10 × 6 × 0.6 = 36 MN.
If the raft and pile capacities are added, the total capacity of the foundation (in
compression) is 36 + (9 × 1.95) = 53.55 MN.
The bearing capacity of a block containing the raft and piles must now be considered,
assuming that the ‘shaft’ failure of the block occurs around the outer perimeter of the pile
group. The bearing capacity of the cap outside the perimeter is also added. The ‘block’
capacity is then

2 × (8.6 + 4.6) × 0.100 × 15 + 8.6 × 4.6 × 0.900 + (10 × 6 – 8.6 × 4.6) × 0.6
= 39.60 + 35.60 + 12.60 = 87.46 MN.

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
460 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

This exceeds the sum of the raft and pile capacities, and thus the design value of ulti-
mate capacity of the foundation is 53.55 MN. The corresponding factor of safety is
53.55/20 = 2.67, which satisfies the design criterion.

16.8.2. Moment capacity


From equation (16.5), the maximum ultimate moment sustained by the soil is

pur BL2
Mm =
8
= 0.6 × 6 × 10 2 / 8
= 45 MNm

From equation (16.4), the ultimate moment capacity of the raft is

27 20  20 
Mur = 45 × × 1−
4 53.55  53.55 
= 44.1 MNm

Considering now the contribution of the piles; from equation (16.6),


9

Mup =
∑P
i =1
uui xi

= 1.20 × (3 × 4 + 3 × 4 + 3 × 0)
= 28.8 MNm

The total moment capacity is therefore 44.1 + 28.8 = 72.9 MNm.


A check must now be made of the moment capacity of the block containing the piles
and the soil, using equation (16.7). For the block, the length is 2.5 times the width, so
that the average ultimate lateral pressure along the block, p– u , is approximately 4.5 × 0.1
= 0.45 MPa. Thus

MuB = 0.25 × 0.45 × 6 × 152


= 151.9 MNm

This far exceeds the value of 72.9 MNm computed above.


The factor of safety for moment loading is therefore 72.9/25 = 2.92, which also satis-
fies the design criterion.

16.8.3. Lateral load capacity


The ultimate lateral load capacity of single piles is computed using the solutions of
Broms [16.53], assuming that the pile heads can be considered as fixed. For short-pile

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 461

failure, Broms’ theory gives an ultimate lateral load capacity of 7.6 MN per pile. For
long-pile failure, taking the yield moment of the pile itself to be 0.45 MNm, the ultimate
lateral load is found to be 0.61 MN, which is obviously the critical value. For the nine
piles, the total lateral load capacity is 5.49 MN. This value is found to be less than the
corresponding value for the block. Thus, the factor of safety against lateral failure is
5.49/2.0 = 2.74, which satisfies the design criterion.

16.8.4. Load–settlement behaviour


The following calculations will be carried out.

1. A non-linear analysis to estimate the relationship between load and immediate


settlement.
2. A linear analysis of both undrained and drained behaviour to obtain, by differ-
ence, the consolidation settlement. This can then be added to the computed
immediate settlement to obtain the long-term (immediate plus consolidation)
settlement.

The average axial stiffness of the raft can be estimated from the elastic solutions repro-
duced by Poulos and Davis [16.54] for a rigid circular foundation on a finite layer, using
a circle of equal area to the actual rectangular foundation. The following values of initial
(elastic) raft stiffness Kri are obtained:

(a) undrained case; Kri = 420 MN/m


(b) drained case; Kri = 169 MN/m.

On considering the piles, use of the solutions presented by Randolph and Wroth
[16.20] gives single pile stiffness values of 217 MN/m and 122 MN/m for the
undrained and drained cases, respectively. Assuming that the group factor is approxi-
mated as np (where np is the number of piles), the following initial pile group
stiffnesses are obtained:

(a) undrained case; Kpi = 651 MN/m


(b) drained case; Kpi = 366 MN/m.

From equation (16.11), the initial stiffness of the piled raft foundation is

Kpri = X Kpi

For the undrained case,

1 − 0.6 (420 / 651)


X =
1 − 0.64 (420 / 651)
= 1.044
K ue = 1.044 × 651 = 680 MN/m

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
462 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

For the drained case,

1 − 0.6 (169 / 366)


X =
1 − 0.64 (169 / 366)
= 1.026
K e′ = 1.026 × 366 = 375 MN/m

The proportion of load carried initially by the piles, bp, is given by equation (16.13).
For undrained conditions,
0.2 420
a= ×
1 − 0.8 × 420 / 651 651
= 0.267
b p = 1 / 1.267 = 0.79
For drained conditions,
0.2 169
a= ×
1 − 0.8 × 169 / 366 366
= 0.146
b p = 1 / 1.146 = 0.87

For the undrained case, the non-linear analysis is tabulated in Table 16.4, assuming that
the hyperbolic factors are Rfr = 0.75 and Rfp = 0.5. For each applied load, the values of bp
and X from the previous load are used, starting with the initial values for the first load.
The computed load–settlement curve is shown in Figure 16.19. At the long-term
design load of 15 MN, the calculated immediate settlement is 31 mm.

Table 16.4. Calculation of load–settlement curve for piled raft foundation in worked
example (undrained case)

V Vp Vr Kr Kp VA S
(MN) X bp (MN) (MN) (MN/m) (MN/m) (MN) (mm) V >VA

0 1.044 0.790 0 0 420.0 651.0 22.2 0 No


5 1.044 0.790 3.95 1.05 410.8 577.7 22.2 8.3 No
10 1.052 0.751 7.51 2.49 398.2 511.7 23.4 18.6 No
15 1.062 0.708 10.62 4.38 381.7 454.0 24.8 31.1 No
20 1.073 0.661 13.22 6.78 360.7 405.8 26.6 45.9 No
25 1.082 0.619 15.48 9.52 336.7 363.9 28.3 63.5 No
30 – – 17.55 12.45 311.1 325.5 28.3 85.8 Yes
35 – – 17.55 17.45 267.3 325.5 28.3 105.3 Yes
40 – – 17.55 22.45 223.6 325.5 28.3 132.6 Yes
45 – – 17.55 27.45 179.8 325.5 28.3 173.1 Yes
50 – – 17.55 32.45 136.1 325.5 28.3 239.6 Yes
52 – – 17.55 34.45 118.1 325.5 28.3 280.0 Yes

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 463

60

50
Vertical applied load: MN

40

30

20

10

0
0 50 100 150 200 250 300
Settlement: mm

Figure 16.19. Calculated load–settlement curve for piled raft foundation in worked
example (undrained case)

It will be assumed that the final consolidation settlement (SCF) can be computed as
the difference between the total final and immediate settlements from purely elastic
analyses, so that

V V
S CF = − (16.32)
K e′ K ue

where K e′ and Kue are defined in equation (16.22). Then using the pile–raft stiffness
values computed previously,

 1 1 
S CF = 15  − 
 375 680 
= 0.0179 m

Thus, the estimated total final settlement is 0.0311 + 0.0179 = 0.0490 m (49 mm). This
just satisfies the design criterion of 50 mm maximum long-term settlement.

16.8.5. Differential settlement


It is only possible to obtain a rough estimate of differential settlement in the approximate
approach, as the distribution of applied loading is not specified. The simplifying assump-
tion is made that the vertical load is uniformly distributed, and use is made of the results

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
464 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

of Horikoshi and Randolph [16.22] for the differential settlement of a rectangular raft.
The raft–soil stiffness is defined therein as
1/ 2 3
E (1 − 2
s ) B t
K rs = 5.57 r     (16.33)
Es (1 − 2
r )  L L
The parameters assumed here (for long-term conditions) are Er = 30 000 MPa, Es =
15 MPa, vs = 0.3, vr = 0.2. Also, B = 6 m, L = 10 m and t = 0.5 m. Thus Krs = 1.022, and
from the above reference, the ratio of the maximum differential settlement to the average
settlement is 0.22. Assuming that this ratio applies also to the piled raft, the maximum
long-term differential settlement (centre-to-corner) is 0.22 × 0.049 = 0.011 m. This
exceeds the specified value of 10 mm, and it is found that the raft thickness needs to be
increased slightly to 0.52 m.

16.8.6. Pile loads


To assist in the structural design of the piles, equation (16.24) can be used to estimate the
maximum and minimum axial pile loads. At the design ultimate load of 20 MN, the
proportion of load carried by the piles (from Table 16.4) is given by bp = 0.661. Then

20 × 0.661 25 × 4
Pmax = +
9 96
= 1.47 + 1.04
= 2.51 MN
Pmin = 1.47 − 1.04
= 0.43 MN

It will be noted that the maximum calculated pile load exceeds the ultimate geotechnical
pile load capacity of 1.95 MN, thus implying that the capacity of the outer piles is fully
utilised. It would, however, be prudent to design the piles structurally to carry the calcu-
lated maximum load, in case the geotechnical capacity of the piles has been
under-estimated.

16.8.7. Raft bending moments and shears


For simplicity, the applied loading is assumed to be uniformly distributed, and the
long-term case (purely vertical loading) is considered. For this case, the average applied
pressure is 0.25 MPa and the calculations above indicate that the piles take 87% of the
applied load. Assuming the reaction pressures also to be uniformly distributed, the
average raft contact pressure is 0.0325 MPa, and the average load in each pile is
1.45 MN.
On dividing the raft into three strips of equal width (in each direction), calculations
based on simple statics show the maximum positive (sagging) bending moments to be
0.484 MNm/m in the x-direction, and zero in the y-direction. The corresponding
maximum negative (hogging) bending moments are –0.109 MNm/m in both
directions.

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 465

16.8.8. Comparisons with computer analysis


The example problem has been analysed for long-term vertical loading, using the
computer program GARP [16.27]. For this analysis, a regular mesh with a total of 273
nodes was employed.
Table 16.5 shows some of the computed performance characteristics of the piled raft,
together with values from the simplified analysis described herein. The following obser-
vations are made:

Table 16.5. Comparison between simplified method and computer analysis for piled raft
foundation in worked example (long-term loading)

Computed value

Quantity Simplified method Computer analysis (GARP)

Max. settlement (mm) 49.0 43.4


Max. differential settlement (mm) 11.0 8.4
Max. moment in x-direction (MNm/m) 0.484 0.499
Min. moment in x-direction (MNm/m) –0.109 –0.220
Max. moment in y-direction (MNm/m) 0 0.201
Max. pile load (MN) 1.45 1.64
Min. pile load (MN) 1.45 0.95
Proportion of applied load carried by piles 0.87 0.83

(a) there is reasonable agreement for the maximum central and differential
settlements;
(b) there is reasonable agreement for the proportion of load carried by the piles;
(c) there is fair agreement for the maximum and minimum pile loads;
(d) the maximum moments from the simplified method are, with one exception, not
in agreement with the values computed by GARP.
The simplified method therefore appears to give a reasonable estimate of overall behav-
iour of the piled raft system, but should not be relied upon to provide design bending
moments for the raft.

16.9. References
16.1. ZEEVAERT, L. Compensated friction-pile foundation to reduce the settlement of buildings on the
highly compressible volcanic clay of Mexico City. Proc. 4th Int. Conf. Soil Mech. Foundn
Engng, London, 1957, 2, 81–86.
16.2. DAVIS, E. H. and POULOS, H. G. The analysis of piled raft systems. Australian Geomech. J.,
1972, G2, 1, 21–27.
16.3. HOOPER, J. A. Observations on the behaviour of a piled-raft foundation on London Clay. Proc.
Instn Civ. Engrs, Part 2, 1973, 55, Oct., 855–877.
16.4. HOOPER, J. A. Review of behaviour of piled raft foundations. Rep. No. 83, CIRIA, London, 1979.
16.5. BURLAND, J. B., BROMS, B. B. and DE MELLO, V. F. B. Behaviour of foundations and structures.
Proc. 9th Int. Conf. Soil Mech. Foundn Engng, Tokyo, 1977, 2, 495–546.
16.6. SOMMER, H., WITTMANN, P. and RIPPER, P. Piled raft foundation of a tall building in Frankfurt
Clay. Proc. 11th Int. Conf. Soil Mech. Foundn Engng, San Francisco, 1985, 4, 2253–2257.

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
466 D ESIG N AP PLIC A TIO NS OF RA FT FOU N DA TIO N S

16.7. PRICE, G. and WARDLE, I. F. Queen Elizabeth II Conference Centre: monitoring of load sharing
between piles and raft. Proc. Instn Civ. Engrs, 1986, 80, 1, 1505–1518.
16.8. FRANKE, E. Measurements beneath piled rafts. Keynote Lecture, ENPC Conf., Paris, 1991,
1–28.
16.9. RANDOLPH, M. F. Design methods for pile groups and piled rafts. State-of-the-Art Report, 13th
Int. Conf. Soil Mech. Foundn Engng, New Delhi, 1994, 5, 61–82.
16.10. POULOS, H. G. Analysis of piled strip foundation. Comp. Methods & Advances in Geomech.
(eds G. Beer et al.). Balkema, Rotterdam, 1991, 1, 183–191.
16.11. POULOS, H. G. Piled rafts in swelling or consolidating soils. J. Geotech. Engng, ASCE, 1993,
119, 2, 374–380.
16.12. SINHA, J. Piled raft foundations subjected to swelling and shrinking soils. Ph.D. Thesis, Univ.
Sydney, Australia, 1997.
16.13. MEYERHOF, G. G. The bearing capacity of footings under eccentric and inclined loads. Proc. 3rd
Int. Conf. Soil Mech. Foundn Engng, 1953, 1, 440–445.
16.14. LEE, I. K. Foundations subject to moment. Proc. 6th Int. Conf. Soil Mech. Foundn Engng,
Montreal, 1965, 2, 108–112.
16.15. POULOS, H. G. and DAVIS, E. H. Pile foundation analysis and design. John Wiley, New York, 1980.
16.16. TOMLINSON, M. J. Foundation design and construction. Longman Scientific and Technical,
Harlow, 1986, 5th edn.
16.17. RANDOLPH, M. F. Design of piled raft foundations. Proc. Int. Symp. on Recent Devel. in Lab.
and Field Tests and Anal. of Geotech. Problems, AIT, Bangkok, 1983, 525–537.
16.18. FRANKE, E., LUTZ, B. and EL-MOSSALLAMY, Y. Measurements and numerical modelling of
high-rise building foundations on Frankfurt Clay. Geot. Spec. Publ. No. 40 (eds A. Yeung and
G. Felio), ASCE, New York, 1994, 2, 1325–1336.
16.19. VAN IMPE, W. F. and DE CLERQ, Y. A piled raft interaction model. Geotechnica, 1995, 73, 1–23.
16.20. RANDOLPH, M. F. and WROTH, C. P. Analysis of deformation of vertically loaded piles. J.
Geotech. Engng Div., ASCE, 1978, 104, 12, 1465–1488.
16.21. MATHSOFT. Mathcad 7 user guide. Mathsoft Inc., Cambridge, Mass., USA, 1997.
16.22. HORIKOSHI, K. and RANDOLPH, M. F. On the definition of raft–soil stiffness ratio for rectangular
rafts. Géotechnique, 1997, 47, 5, 1055–1061.
16.23. BROWN, P. T. and WIESNER, T. J. The behaviour of uniformly loaded piled strip footings. Soils
and Foundations, 1975, 15, 13–21.
16.24. WIESNER, T. J. and BROWN, P. T. Behaviour of piled strip footings subject to concentrated loads.
Australian Geomech. J., 1976, G6, 1–5.
16.25. HONGLADAROMP, T., CHEN, N. J. and LEE, S. L. Load distributions in rectangular footings on
piles. Geotech. Engng, 1973, 4, 2, 77–90.
16.26. BROWN, P. T., POULOS, H. G. and WIESNER, T. J. Piled raft foundation design. Proc. Symp. on
Raft Foundations, Perth, 1975, CSIRO (Australia), 13–21.
16.27. POULOS, H. G. An approximate numerical analysis of pile–raft interaction. Int. J. Num. Anal.
Meth. Geomech., 1994, 18, 73–92.
16.28. CLANCY, P. and RANDOLPH, M. F. An approximate analysis procedure for piled raft foundations.
Int. J. Num. Anal. Meth. Geomech., 1993, 17, 12, 849–869.
16.29. YAMASHITA, K., KAKURAI, M., YAMADA, T. and KUWABARA, F. Settlement behaviour of a
five-storey building on a piled raft foundation. Proc. 2nd Int. Symp. on Deep Foundns on Bored
and Auger Piles, Ghent, 1993, A. A. Balkema, Rotterdam, 351–356.
16.30. BUTTERFIELD, R. and BANERJEE, P. K. The elastic analysis of compressible piles and pile groups.
Géotechnique, 1971, 21, 1, 43–60.
16.31. KUWABARA, F. An elastic analysis for piled raft foundations in a homogeneous soil. Soils and
Foundations, 1989, 28, 1, 82–92.
16.32. HAIN, S. J. and LEE, I. K. The analysis of flexible raft–pile systems. Géotechnique, 1978, 28, 1,
65–83.

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
PRA CT IC AL DES IG N O F P ILED R AFT S 467

16.33. DESAI, C. S. Numerical design analysis for piles in sands. J. Geotech. Engng Div., ASCE, 1974,
100, GT6, 613–635.
16.34. NAYLOR, D. J. and HOOPER, J. A. An effective stress finite element analysis to predict the short-
and long-term behaviour of a piled-raft foundation on London Clay. Proc. Conf. on Settlement
of Structures, Cambridge, 1974, 394–402.
16.35. HEWITT, P. B. and GUE, S. S. Piled raft foundation in a weathered sedimentary formation, Kuala
Lumpur, Malaysia. Proc. Geotropika, Malacca, Malaysia, 1994, 1–11.
16.36. ITASCA. FLAC user manual. Itasca Corp., Minneapolis, Minn., USA, 1991.
16.37. OTTAVIANI, M. Three-dimensional finite element analysis of vertically loaded pile groups.
Géotechnique, 1975, 25, 2, 159–174.
16.38. ZHUANG, G. M., LEE, I. K. and ZHAO, X. H. Interactive analysis of behaviour of raft–pile foun-
dations. Proc. Geo-Coast ’91, Yokohama, 1991, 2, 759–764.
16.39. LEE, I. K. Analysis and performance of raft and raft–pile systems. Keynote Lecture, 3rd Int.
Conf. Case Hist. in Geotech. Engng, St. Louis, 1993 (also Res. Rep. R133, ADFA, Univ. NSW,
Australia).
16.40. TA, L. D. and SMALL, J. C. Analysis of piled raft systems in layered soils. Int. J. Num. Anal.
Meth. Geomech., 1996, 2, 57–72.
16.41. WANG, A. Private communication, 1995.
16.42. KIMURA, M. and ADACHI, T. Analyses on laterally loaded cast-in-place concrete piles. Proc. 6th
Int. Conf. Piling and Deep Foundns, DFI, Bombay, India, 1996, 3.9.1–3.9.6.
16.43. POULOS, H. G., SMALL, J. C., TA, L. D., SINHA, J. and CHEN, L. Comparison of some methods for
analysis of piled rafts. Proc. 14th Int. Conf. Soil Mech. Foundn Engng, Hamburg, 1997, 2,
1119–1124.
16.44. POULOS, H. G. DEFPIG user manual. Centre for Geotech. Res., Univ. Sydney, Australia, 1990.
16.45. HORIKOSHI, K. and RANDOLPH, M. F. A contribution to optimum design of piled rafts. Géotech-
nique, 1998, 48, 3, 301–317.
16.46. DÉCOURT, L. The Standard Penetration Test: State-of-the-Art Report. Proc. 12th Int. Conf. Soil
Mech. Foundn Engng, Rio de Janeiro, 1989, 4, 2405–2416.
16.47. DÉCOURT, L. Predictions of load–settlement relationships for foundations on the basis of SPT-T.
Ciclo de Conf. Int. ‘Leonardo Zeevaert’, UNAM, Mexico, 1995, 85–104.
16.48. SOMMER, H., TAMARO, G. and DEBENEDITTIS, D. Messe Turm, foundations for the tallest
building in Europe. Proc. 4th DFI Conf., Stresa, Italy, 1991, 139–145.
16.49. SOMMER, H. Development of locked stresses and negative shaft resistance at the piled raft foun-
dation – Messeturm Frankfurt/Main. Deep Foundns on Bored and Auger Piles, 1993 (ed. W. F.
Van Impe). Balkema, Rotterdam, 347–349.
16.50. EL-MOSSALLAMY, Y. and FRANKE, E. Piled rafts: numerical modelling to simulate the behaviour
of piled raft foundations. The Authors, Darmstadt, 1997.
16.51. TAMARO, G. J. Foundation engineers: why do we need them? 1996 Martin S. Kapp Lecture,
ASCE, New York, 1996.
16.52. POULOS, H. G. Alternative design strategies for piled raft foundation. Proc. 3rd Int. Conf. Deep
Foundns, Singapore, 1994, 239–244.
16.53. BROMS, B. B. Lateral resistance of piles in cohesive soils. J. Soil Mech. Foundn Div., ASCE,
1964, 90, 2, 27–63.
16.54. POULOS, H. G. and DAVIS, E. H. Elastic solutions for soil and rock mechanics. John Wiley, New
York, 1974.

Downloaded by [ University of Liverpool] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.

S-ar putea să vă placă și