Proceedings of Indian Geotechnical Conference

December 15-17, 2011, Kochi (Paper No. N-322)

PILED RAFT BEHAVIOUR – MODEL STUDIES AND FIELD PERFORMANCE

V.Balakumar, Senior Consultant, Simplex Infrastructures Limited, Chennai. vb_kumar2002@yahoo.com

I.V.Anirudhan, Director, Geotechnical solutions, Chennai. anirudhen@eth.net

ABSTRACT: In order to establish the general applicability of piled raft foundation, the data obtained from hydro testing of
a typical ammonia tank was compared with the results of independent small scale model studies. The pile supported raft of
the tank and the model piled raft had a pile - raft area ratio of 9% and 9.25%. It was found that, the load settlement and the
load sharing response of the tank foundation and the model were identical. The study established that even when the system
was not designed as a piled raft, under favourable conditions, the ground supported raft together with the pile group would
have a tendency to behave as piled raft.

INTRODUCTION basements, wherein very thick rafts are placed at

The settlement requirement for the storage structures is
often very stringent and depends upon the nature of the
storage material and the permissible deflection of the pipe
line. The design of foundation system for such storage
tanks and structures has always been a challenge for the
geotechnical designers. Traditionally designed deep piles,
had been the automatic choice of the foundation system in
such cases. By convention, deep piles were designed to take
the entire applied load to eliminate the settlement and the
presence of raft and its capacity to share the load was being
completely ignored. The urge for economising the
foundation design has led to the conviction among the
designers to believe that reducing the settlement to a
permissible level rather than completely eliminating it can
produce considerable economy. This has led to the
recognition of the load sharing behaviour of the raft and the
pile. The load sharing behaviour of the raft and the piles
was brought out three decades before by researchers [1, 2],
and others. The foundation system with this load sharing
concept is addressed as piled raft wherein the piles are
added as settlement reducers. The piled raft system has a lot
of advantages [3] under the favourable circumstances
particularly when the raft rests on strata with adequate
bearing capacity.
Recognising the advantages and the economy that can be
achieved by reducing the settlement to a permissible level
continued research work has been carried out adopting
analytical and numerical modelling [4,5,6,7,8,9], small
scale model studies with centrifuge models [10], 1g models
[11,12] and by monitoring the proto type piled raft
[6,11,3].These extensive studies has made the piled raft
foundation a viable alternative to deep piles and many
structures have now been supported on piled raft [3,13]
successfully. However the development of piled raft
appears to have been directed more towards specific
conditions as given below.
1. In majority of the cases the supporting soil has been
deep deposits of overconsolidated clay.
2. They have been supporting predominantly tall
buildings (with very high column loadings) with deep
deeper depths. In such cases the raft gets the advantage
of relief in the overburden pressure from settlement
point of view.
However there are various other conditions where the
concept of piled rafts can be considered. For example, when
storage tanks are to be placed on loose or loose to medium
dense sand, deep piles are resorted to, as the permissible
settlement for foundation placed on sand is less than that
for the foundation placed on clay. Also the tank pad namely
the combined raft is placed very close to the ground level or
at ground level itself. One more important aspect is that the
driving of piles in sand improves the state of compaction.
Although this has not been quantitatively established, but
qualitatively accepted. If the above conditions apply, then
the pile supported raft of the storage structures will behave
as piled raft even if they have not been designed as piled
raft. In order to understand the above behaviour, the load
settlement response obtained from the results from hydro
tests conducted on a tank farm was examined in more detail
based on published results of [14]. Typically the result of
hydro test conducted on a 33m diameter ammonia tank is
discussed.
STUDY OF THE FIELD DATA
A tank farm comprising of 33m diameter ammonia tanks,
28m diameter phosphoric acid tanks and some small
storage structures were constructed as a part of a huge
manufacturing facility. The soil profile at the site had
alternate seams of soft clay, medium dense to dense sand
and stiff clay as presented through the results of cone
penetration tests in Figure 1. The tanks were supported on
closely spaced (3D, D being the diameter of the pile)
450mm dia driven cast in situ piles terminated at 10m from
ground level of which 6m length of the pile was passing
through sand. The capacity of the pile was 65 tonnes and
this was established by individual pile load tests. The
ammonia tank had 437 piles and the pile-raft area ratio was
close to 9%.The raft was 400 mm thick, being a flexible
tank. On completion, the tank was subjected to hydro test
with 17.90 m of standing water. The water was pumped in

947
 V.Balakumar & I.V.Anirudhan
at a slow rate in stages. The rate of loading for the hydro Table 1 Comparison of Settlement of pile supported raft
test was designed considering the mobilization & and unpiled raft for Ammonia Tank
dissipation of pore pressure based on the principles of Load in kN Settlement, in mm
critical state soil mechanics. The settlement was recorded in Unpiled Raft Pile Supported raft
the centre and eight symmetrically placed points along the 5129 104 6.0
periphery. 8549 163 19.0
Number of Blows for 100mm penetration, empty driving 11968 209 47.0
0
I
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
I I I I I I I I I I I I I I I 15302 252 65.0
Cone resistance qc from SCPT, Mpa
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
4 -I I I I I I I I I I I I I I I I

3- It is seen that at any load level the settlement of the pile

2- Cone resistance qc from SCPT
Pile penetration rate supported raft was much lower than the unpiled raft. The
1-
0-
load settlement response of the piled raft was characterised
Elevation, m

-1 - by plotting the load on the log-scale (y-axis) and settlement

-2 -
in normal scale (x-axis). This helps in identifying the
-3 -
-4 - various stages in load settlement response. This is presented
-5 - in Fig. 4.
-6 -
-7 -
-8 -
-9 -
-10 -
-11 -
-12 -
-13 -
-14 -
-15 -

Fig. 1 Cone resistance profile – Ammonia Tank

Fig. 4 Typical characterized Load Settlement Response

Fig.2 Load Settlement Response - Ammonia Tank
Typical load settlement response of an ammonia tank and
Phosphoric acid tanks are presented in Figure 2 and 3.
Adopting Davisson’s method the capacity of the pile group
as a pier was computed. Based on this the capacity of the
pier was found to be of the order of11000kN and 8000kN
respectively. In order to understand the effect of the pile
group on load settlement behaviour of raft, the load
settlement response of unpiled raft was computed
numerically [14]. Table 1 presents the comparison of
settlement of unpiled raft and pile supported raft.
Fig.3 Load Settlement Response - Phosphoric Acid
– Ammonia Tank
The figure 4 clearly shows a three phased behaviour of the
pile supported raft foundation (piled raft) which is identical
to that of model piled raft as shown in Fig 5.This behaviour
confirms that although the pile group was designed to carry
the entire structural load, in reality there appears to be a
load sharing process between the raft and the pile group. In
the first phase maximum load was taken by the pile group
as seen by the high stiffness. Thereafter the stiffness
reduces gradually upto an extent and then rapidly. Also the
pile group has functioned as settlement reducer. This is
evident from the fact that the settlement of the pile
supported raft is far less than the computed unpiled raft
settlement To represent the settlement reduction and the
load shared by the pile group quantitatively, the settlement
reduction ratio SR which is the ratio of difference in the
settlement between the unpiled raft and piled raft at any
load level divided by the unpiled raft settlement for a given
loading and the load sharing ratio Dpr (ratio of the load taken
by the pile group to that of piled raft for a given settlement
were computed from the settlement of the unpiled raft and
the load settlement results of hydro tests. The values are
presented below in table 2. From the table 2 it is seen that
the load sharing ratio reduces with the settlement and at the
final stages the load sharing ratio remains more or less
constant. However it is seen that the pile group has taken
72% of the load and the raft has shared 28% of the load.

948
 Piled Raft Behaviour – Model Studies And Field Performance
Table 2 SR and Dpr at different of load from hydro test for The characteristic load settlement response obtained from
typical Ammonia Tank. the 1g model tests for 3 different area ratio is presented in
% of Loading SR Dpr Fig.6. Comparing Fig.6 with Fig.4 it is seen that the
25 95 90 behaviour is identical in both the cases. In the three phased
50 90 84 response stage OA represents the elastic stage, AB
75 82 77 represents the elasto plastic stage and BC represents plastic
100 73 72 stage wherein even for small increase in the load the piled
raft loses its stiffness rapidly. Fig.7 presents the load
1g MODEL STUDIES sharing behaviour wherein the load sharing ratio Dpr is
In order to validate the behaviour of the pile supported raft, plotted against the settlement. It can be seen that up to a
the results of the hydro test were compared with the results settlement level of 2mm the load sharing ratio is very high
of the 1g model tests conducted on circular piled raft. These and then gradually decreases and beyond 10mm settlement
tests were part of the elaborate and well planned research which corresponds to nearly 50% of the final load applied.
programme that was carried out to study the behaviour of The SR and Dpr value at different % of loading is given in
piled raft on sand keeping in mind that the piles would be Dpr table 3.
relatively smaller diameter for representing driven piles
with moderate capacity and the raft placed close to the Table 3 Settlement reduction and load sharing at different
ground level. Rafts typical of storage tanks (circular) and settlement levels.
structures like buildings (non circular) were taken up for the % of Loading SR (%) Dpr (%)
study. Among the various methods available it was decided 25 72 40
to study the behaviour of piled raft on sand adopting 1g 50 60 30
model studies. Primarily 1g model study was preferred 75 54 29
mainly because it easier to conduct parametric studies and 100 50 27
represents more or less field conditions. Although the
results cannot be used directly for field problems, the
This confirms that the pile group at higher load adds the
results do give good idea of the performance of the system.
required capacity for the raft to take a higher load compared
The details of the test setup and the model piled raft are to the unpiled raft. Comparing the progression of SR and Dpr
presented in earlier publications. The bed was prepared by values estimated in table 2 and 3 it is seen that the
sand raining process and recalibrated compaction procedure behaviour of model piled rafts and the pile supported raft
in layers. The variation of angle of internal friction with bed are more or less identical.
density is presented in Figure 6.
ANGLE OF INTERNAL
10

FRICTION, deg

B
L O A D , kN

A
1

D = 200mm
Area ratio UNIT WEIGHT, kN/m3
t = 8mm
9.25%
L = 160mm
6.25% d = 10mm Fig.6 Variation of angle of internal friction with
O 4.25% Medium dense unit weight D
0 .1

PR

0.0 5.0 10.0 15.0 20.0 25.0 0.0 0.2 0.4 0.6 0.8 1.0
0
SETTLMENT, mm
2
Fig. 5 Characterization curves – 1g model circular raft -
4
various area ratios t = 8mm
6
d= 8mm
SETTLEMENT,mm

The model piles were installed with a specially made 8 N = 21

template and a guide. The sequence of installation by 10
R.A. = 36º
D = 200mm
driving was done in similar way as in the field. The raft was 12 Bed = Medium dense
connected to the piles with countersunk stainless steel 14
screws. The loading was applied in small increments and Area
Pile Ratio
length
16
9.25%
for each increment the time versus settlement was also 18 5.25%
recorded. The tests were carried out by varying the
20 4.25%
parameters relating to the pile namely pile-raft area ratio,
22
pile diameter and pile length.
Fig. 7 Settlement v/s LS ratio Įpr for different area

949
V.Balakumar & I.V.Anirudhan
COMPARISON
Comparing the behaviour, it is seen that the pile group
acted as settlement reducer and also exhibited ductile
behaviour. In the initial stage of both the cases, the pile
group takes a higher proportion of the load and at the final
stages the load shared by the pile group reduces, to about
70% in the case of storage tank under study. Very similar to
the model behaviour, the load shared by the pile group
reduces with the settlement and tends to remain constant
suggesting that, eventhough the tank foundation was not
designed as a piled raft system, it behaved like piled raft.
But in the case of storage tank foundation under study, the
pile group shared much higher proportion of the applied
load under large settlements also, perhaps due to the fact
that the soft clay layers immediately below the raft and the
thin fill has very limited stiffness. It is also possible that the
installation procedure of very closely spaced driven piles
enhanced the density of the supporting sand layers below
the soft clay, thereby increasing the individual pile
capacities. However, the soft clay does not achieve higher
stiffness by the pile driving process.
CONCLUSION
The study has clearly established that the principles of piled
raft can be applied in general under favourable conditions,
when the raft is seated on a layer with reasonable bearing
capacity. It is evident from the fact that even when the
system was not designed on principles of load sharing by
interaction the load sharing has taken place leading to the
foundation system behaving as a piled raft.
In case of tank foundation the pile group appear to have
shared much higher load mainly because driving of closely
spaced piles has densified the sand layer enhancing the
confining stress of the soil around the pile and this has
resulted in the pile group taking much higher proportion of
the applied load compared to that of the model piled raft.
The study has established the general applicability of the
principles of piled raft when the settlement alone governs
the design.
REFERENCES
1. Burland J.B., Broms B.B. and de Mello V.F.B. (1977),
‘Behaviour of Foundations and Structures’, Proc. 9
ICSMFE Tokyo 2, pp 495 – 546.
950
2. Hooper J.A. (1974), ‘Observations on the Behavior of
a Piled Raft Foundation on London Clay’, Proc.
Institution of Civil Engineers, Part 2, Vol. 55, pp. 855-
877.
3. Poulos H.G. (2008), ‘The Piled Foundation for the Burj
Dubai – Design and Performance’, IGS – Ferroco
Terzaghi Oration.
4. Clancy P. (1993), Numerical Analysis of Piled Raft
Foundations, University of Western Australia, PhD
Thesis.
5. Gandhi S.R and Maharaj D.K. (1996), ‘Analysis of
Piled Raft Foundations’, 6th International Conference
on Piling and Deep Foundations, Bombay, pp. 1.11.1-
1.11.7.
6. Katzenbach R. and Reul O. (1997), ‘Design and
Performance of Piled Rafts’, Proc. XIV Inst.
Conference on Soil Mechanics and Foundation
Engineering, pp. 2253-2256.
7. Poulos H.G. (2001), ‘Piled Raft Foundation: Design
and Application’, Geotechnique, Vol. 51, No. 2, pp
111-113
8. Prokoso W.A. and Kulhawy F.H. (2001), ‘Contribution
of Piled Raft Foundation’, Journal of Geotechnical and
Geoenvironmental Engineering, ASCE, pp 17-24.
9. Small J and Poulos, H.G. (2007) “A Method of
Analysis of Piled Raft”, 10th Australia Newzeland
Conference on Geo Mechanics, PP. 555.
10. Horikoshi K. (1995), ‘Optimum Design of Piled Raft
Foundations’, Dissertation submitted for the degree of
Doctor of Philosophy, University of Western Australia.
11. Balakumar V. and Ilamparuthi K. (2006),”Performance
of model piled raft on sand”, Proceedings Indian
Geotechnical Conference, Chennai, India, pp 463-298.
12. Turek.J. and Katzenbach. R. 2003. Small scale model
tests with combined pile raft foundations. Proceedings
of the 4th International Geotechnical Seminar on Deep
Foundation on Bored and Augured piles. Ghent,
Belgium, 409-413
13. Yamashita (2011), ‘Field Measurements on Piled Rafts
with Grid –Form Deep Mixing Walls on Soft Ground.
Geotechnical Engineering Journal of the SEAGS &
AGSSEA Vol – 42; No.2; June 2011, pp 2 -10.
14. Anirudhan & Balakumar (2010), ‘Pile Foundation as
Settlement Reducer for Large MS Storage Tanks’,
Indian Geotechnical Conference 2010, pp