Approximate Design Method For Piled Raft Foundations
Approximate Design Method For Piled Raft Foundations
INTRODUCTION
Piled raft foundations have been used to support a variety of structures, and they are now widely recognized as
one of the most economical methods of foundation systems since Burland et al. (1977) presented the concept of
`settlement reducers'. This type of foundation has been
used in Japan since 1980s (for example, Kakurai, 2003).
As a preliminary routine design tool of piled raft foundations subjected to vertical, horizontal and moment
loading as well as free-eld ground movements, a computer program PRAB (Piled Raft Analysis with Batter
piles) has been developed by Kitiyodom and Matsumoto
(2002, 2003) and Kitiyodom et al. (2005).
In PRAB a hybrid model, in which the exible raft is
modelled as a thin plate, the piles as elastic beams and the
soil is treated as interactive springs, is employed. Both the
vertical and horizontal resistances of the piles as well as
the raft base are incorporated into the model ( see Fig.
1(a)). Pile-soil-pile, pile-soil-raft and raft-soil-raft interactions are taken into account based on Mindlin's solutions (Mindlin, 1936) for both vertical and horizontal
forces.
In this work, the approach described previously by the
authors is modied, in order to make it possible to solve
problems of large non-uniformly arranged piled raft
foundations in a time-saving way using a PC. Instead of
i)
ii)
iii)
Fig. 1.
Senior Engineer, Geotechnical & Foundation Engineering Co., Ltd., Bangkok, Thailand (pastsakorn_kgfe.co.th).
Professor, Graduate School of Natural Science and Technology, Kanazawa University, Ishikawa, Japan.
Doctoral Student, ditto.
The manuscript for this paper was received for review on November 16, 2009; approved on August 30, 2010.
Written discussions on this paper should be submitted before September 1, 2011 to the Japanese Geotechnical Society, 4-38-2, Sengoku,
Bunkyo-ku, Tokyo 112-0011, Japan. Upon request the closing date may be extended one month.
1
KITIYODOM ET AL.
Numerical Method
In PRABS, the raft is modelled as a thin elastic plate,
while the piles and the soil are treated as interactive springs attached to the raft as shown in Fig. 1(b).
The vertical soil springs, kR, at the raft nodes are estimated by Eq. (1).
k R
4G sa
1
1n s s
1exp (h/2a)t
(1)
E s*
2(1n s)
(2)
n sS ns(i) DIi/DItotal
(3)
i 1
(4)
E *s(i)Es(i)/(1n 2s(i))
(5)
where Es(i) and ns(i) are the Young's modulus and the Poisson's ratio for soil layer number i in the n-layered system.
DIi and DItotal in Eqs. (3) and (4) are the dierences between the vertical settlement inuence factors at dierent
soil depths which can be determined by Eqs. (6) and (7).
(6)
(7)
where z itop and z ibottom are the depths below the surface of
the top and bottom of layer number i. The vertical settlement inuence factor I has been given by Harr (1966).
The pile spring stiness can either be directly input into
the program after obtaining it from another analysis (for
example, the program PRAB), or else calculated from
Eq. (8) following Randolph and Wroth (1978).
4
2pr tanh ( mL)
h(1 ns)
mL
z
kpGsr0
tanh ( mL)
4
1
plh(1ns)
mL
L
r0
L
r0
(8)
(9)
w0
Fig. 2.
Fig. 4.
Fig. 5.
Fig. 3.
(10)
KITIYODOM ET AL.
EeqEs(EpEs)Atp/Ag
K R
(11)
R ns/Lp
the raft-soil stiness ratio, KR, for four dierent pile slenderness ratios, where
(12)
2ERt 3BR(1n2s)
3pEsL4R
(13)
Raft Alone
For a square raft having a length of LR subjected to an
uniform vertical load, q, resting on a deep homogeneous
layer, Fig. 6 compares the distributions of normalised
settlement, S, contact pressure, p, and the bending moment, Mx, from PRABS, with those from the piled strip
model (GARP) by Poulos (1994), and the nite element
analysis by Hain and Lee (1978). It can be seen that there
are good agreements among the solutions in all cases.
Fig. 7.
Fig. 8.
Piled Raft
Figure 7 shows the solutions for maximum settlement,
Smax, of a uniformly loaded square raft supported by 64
piles, in a deep homogeneous elastic soil layer. The normalised maximum settlement is plotted as a function of
Fig. 6.
load carried by the piles. Despite the approximations involved, PRABS can provide solutions of adequate accuracy for the settlement and pile load distribution within
a piled raft with L/D less than 25 and s/D greater than 3
which are common for piled raft foundation employed in
practice.
CASE STUDIES
Sonoda et al. (2009) have described the case of a large
piled raft foundation for a commercial building called
Amuplaza that was constructed in Kagoshima City,
Kyushu, in 2003 to 2004. The building is 7-storied with a
basement oor having a building area of 9000 m2, a oor
area of 50000 m2, and a maximum height of 45 m ( see
Fig. 10). A piled raft foundation was employed for the
building in a sandy ground to reduce the average settlement as well as the dierential settlement. The building
was constructed using a reverse construction method, in
which construction of the superstructure (building) and
the substructure (foundation) were constructed simultaneously, in order to reduce the construction period.
Therefore the foundation was regarded as a free standing
pile group without contribution of the raft resistance in
earlier stages of construction, while the foundation
behaved as a piled raft after the construction of the mat
slab (raft) was completed. A static vertical pile load test
was carried out at the construction site. Moreover, during
the construction stage, settlements of the foundation and
the water pressure beneath the raft were monitored.
A test pile was constructed additionally at a location
indicated by `star' symbol in Fig. 10. The test pile was a
cast-in-situ concrete pile having a length of 32.0 m and a
diameter of 1.0 m.
The programs PRAB and PRABS were employed to
analyse the behaviour of the whole piled raft system. The
analyses were carried out in two stages. The rst stage
was the deformation analysis in the stage of pile group
where the raft resistance was not expected. The analysis
in the stage of piled raft was carried out after the end of
the rst stage, considering the existence of the raft
resistance. The stress conditions at the end of the rst
stage were used for the initial conditions in the second
stage. Shear moduli of the soils at small strain, G0, were
derived by Sonoda et al. (2009) on the basis of SPT and
PS-logging tests as shown in Fig. 11.
In order to determine the soil parameters appropriately, back-analysis of the vertical load test of the test pile
was carried out using PRAB, prior to the analysis of the
whole foundation. The test single pile and the ground
were modelled as Fig. 12. Young's modulus of the pile Ep
2.27107 kPa was employed. The maximum shaft friction, fmax, of each section obtained from the static vertical
pile load test results was adopted in the back-analysis.
Figure 13 shows comparison of the analysed and measured load-settlement curves of the pile head and the pile
base. Good matching was obtained if the shear modulus
of the soil obtained from PS-logging was reduced by a
factor of 2 for the soils surrounding the pile shaft and by
Fig. 10.
Fig. 11.
KITIYODOM ET AL.
Fig. 14. Relationship between shaft resistance and local pile displacement
Table 1.
Soil layer
Fig. 12. Seating of test pile, and soil prole and SPT-N values obtained at borehole EB-2
1
2
3
4
to
to
to
to
15.5
21.5
26.5
31.5
G (kPa)
tf (kPa)
q (kPa)
51,840
77,440
77,440
77,440
99.5
44.6
54.3
54.3
900
900
Fig. 13.
of the ground to a depth of 63 m is needed when analysing such a large piled raft foundation because the inuence of the wide length of the raft of 156 m reaches to
deeper depths. Note here that SPT N-values for depths
greater than 63 m were very large and the depth of 63 m
was assumed to be a bed stratum.
For the analysis using PRAB with the equivalent pier
concept, Fig. 16 shows the arrangement of the piles and
the equivalent piers. Properties of the equivalent piers are
summarised in Table 2. It can be seen that the values of R
(dened in Eq. (5)) in all types of equivalent piers are less
than 2.
Fig. 16.
Table 2.
Pier type
1
2
3
Deq (m)
11.17
20.31
11.17
Lp (m)
20
20
25
Eeq (kPa)
6
2.5110
1.76106
2.51106
1.27
1.90
1.14
Fig. 18. Time histories of the total load from the building and measured water pressure at the raft base
that.
The raft base was located at 6.5 m below the original
ground surface. The original ground water table (3.0 m
below the ground level) was lowered to 7.5 m below the
ground level until the end of February 2004, by means of
deep wells. Then, the lowered ground water table was
recovered to the original water table. The measured increase in the water pressure of 35 kPa corresponded to
this recovery of the ground water table.
In the deformation analysis of the whole structure,
rigidity of the superstructure was neglected and vertical
loads from the superstructure were directly applied on the
raft nodes. Figure 19 shows the distributions of loads on
the raft. In the analysis using PRAB with the equivalent
pier concept, the loads acting on the top of the piles,
which were modelled as an equivalent pier, were summed
up and placed on the top of the equivalent pier node. In
analysis for the stage of pile group foundation, load increments shown in Fig. 19(a) were applied, while in analysis for the stage of piled raft foundation, load increments of Fig. 19(b) were applied on the raft. Note here
that the ground water level was recovered at the construction stage of the piled raft as mentioned earlier. The
buoyancy force due to the water pressure at the raft base
was also taken into account in addition to the load increments of Fig. 19(b).
Figures 20 and 21 show the distributions of calculated
and measured settlements of the raft in the x-direction at
y40.5 m ( see Figs. 10 and 16, section A-A?) and those
in the y-direction at x34.8 m ( see Figs. 10 and 16, section B-B?), respectively. Increment of settlements in stage
of pile group are shown in Figs. 20(a) and 21(a), those in
stage of piled raft are shown in Figs. 20(b) and 21(b), and
the total settlements at the nal construction stage are
shown in Figs. 20(c) and 21(c). In the gures, the calcu-
KITIYODOM ET AL.
Fig. 19.
Fig. 20.
CONCLUDING REMARKS
This paper presents an approximate method of analysis
of piled raft foundation in which the raft is modelled as a
thin plate and the piles and the soil are treated as interactive springs. The method makes it possible to solve problems of large non-uniformly arranged piled rafts in a
time-saving way using a PC. The method is implemented
via the computer program PRABS. Moreover, the
equivalent pier concept is presented.
From the parametric study on the interaction factors
and comparison between the existing solutions and those
from PRABS indicate that the proposed approximate
method can provide solutions of acceptable accuracy for
the foundation with L/D less than 25 and s/D more than
3 which are common for piled raft foundation employed
Fig. 22.
REFERENCES
Fig. 21.
in practice.
A case study demonstrates that the analysis using
PRABS and PRAB with the equivalent pier concept can
predict reasonably well the settlements of a full-scale
piled raft containing a large number of piles and the calculation time of the analysis are less than the calculation
time of the full model analysis.
ACKNOWLEDGEMENTS
The authors deeply thank Kyushu Railway Company
and Kagoshima Terminal Building Corporation and
Yasui Architects & Engineers, Inc. for their permission to
use the valuable eld measurement data.
10
KITIYODOM ET AL.
rafts, Proc. 13th Int. Conf. on SMFE, New Delhi, India, 5, 6182.
18) Russo, G. (1998): Numerical analysis of piled rafts, International
Journal for Numerical and Analytical Methods in Geomechanics,
22, 477493.
19) Sonoda, R., Matsumoto, T., Kitiyodoom, P., Moritaka, H. and
Ono, T. (2009): Case study of a piled raft foundation constructed
using a reverse construction method and its post-analysis, Canadian
Geotechnical Journal, 46(2), 142159.
20) Yamashita, K., Yamada, T. and Kakurai, M. (1998): Method for
analysis piled raft foundations, Proc. Deep Foundation on Bored
and Auger Piles, Ghent, Belgium, 457464.