See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/260178562

PILE-SUPPORTED RAFT FOUNDATION SYSTEM

Conference Paper · October 2012

CITATIONS READS
0 4,098

1 author:

Mohab Sabry
Bechtel Corporation
7 PUBLICATIONS 50 CITATIONS

SEE PROFILE

All content following this page was uploaded by Mohab Sabry on 04 August 2016.

The user has requested enhancement of the downloaded file.

PILE-SUPPORTED RAFT FOUNDATION SYSTEM
Emre Biringen, Bechtel Power Corporation, Frederick, Maryland, USA
Mohab Sabry, Bechtel Power Corporation, Frederick, Maryland, USA

Over the past decades, there has been an increasing recognition of the strategic
use of pile-supported rafts in design of heavily-loaded structures to reduce total
and differential settlements. However, such a hybrid foundation construction
method has not been widely utilized in many countries, including the U.S., due to
code limitations. In current practice, the foundation design is based on either (i)
the bearing capacity of the raft supported by the subgrade soils, or (ii) solely the
pile capacity. This paper addresses the potential use of such hybrid piled-raft
systems, where the structural load is shared by both by the piles and the
subgrade soils directly beneath the raft, in foundation design of power plant
structures. The study assessed some of the characteristics of piled-raft behavior
by undertaking three-dimensional finite element analyses of two raft sizes, and
various pile layout patterns, including the rafts with piles distributed evenly and
only in the central area of the raft. The computer software PLAXIS 3D with the
Mohr-Coulomb constitutive soil model was used to facilitate the modeling of such
cases. The results presented in this paper indicate how strategically locating the
piles could reduce the differential and total settlements.

INTRODUCTION foundation is reduced, and the spacing between

them is increased. This saves pile costs and
Under the high applied loads coming from the reduces the installation schedule. Also, by
superstructure, the addition of piles is primarily increasing the spacing between the piles, it
required to improve the factor of safety against provides more accessibility to install conduits
failure when a raft does not provide adequate and piping below the raft, in addition to more
bearing capacity. If the bearing capacity of the flexibility for construction to add more piles in
raft is sufficient to carry the total load with a case of pile replacement. In addition to the
reasonable safety margin, then the addition of primary goal of improved bearing capacity and
piles is usually intended to reduce the settlement performance, other positive effects
settlements to an acceptable amount. In current from using pile-supported rafts can be listed as
practice, the design process for pile-supported follows: (i) reduction of soil heave inside and
raft foundations conservatively concentrates on outside the excavation, because the piles
providing solely the pile capacity to carry the improve the overall soil conditions by preventing
total structural load without taking the stress relief in the ground; (ii) minimizing
contribution of the raft into account. This paper construction measures for the control of
addresses the potential use of hybrid piled-rafts deformations of structure, facades, and
in controlling foundation settlement of power equipment; (iii) better and more economical
plant structures by considering the support of control of large load differences between
both the piles and the subgrade soils beneath heavily-loaded structures adjacent to more
the raft to carry the load. lightly-loaded structures, as well as to adjacent
properties, thus minimizing risk; (iv) ensuring
Several studies (Prakoso and Kulhawy, 2001; stability for the entire foundation, when the
Cunha et al., 2001; De Sanctis and Mandolini, foundation slab by itself does not provide
2006) have undertaken to investigate pile- sufficient stability for the large foundation loads;
supported rafts, in which the loads coming from and (v) creation of an eccentric foundation block
the structure are shared by the pile and the raft, for eccentric loading to prevent anticipated tilting
have shown that such a hybrid foundation (with subsequent centering of the resultant
system provides an efficient way of supporting structural load) by an asymmetrical arrangement
highly-loaded rafts. By sharing the load with the of the piles.
raft, the number of piles needed under the raft

1
The objectives of this paper are (i) establishing Coulomb elasto-plastic soil constitutive model
an understanding of the pile-supported raft load were utilized. The raft and piles are considered
sharing mechanism and the behavior of the to behave linear elastically. Consideration has
system in different soil and pile configuration been given to design applications where the raft
conditions, and (ii) evaluating the settlement thickness and width easily can be as much as 2
performance of pile-supported raft systems for m and 38 m, respectively. In this parametric
foundation design applications. study, the effect of variations in the pile
configurations, length of piles, pile spacing,
NUMERICAL MODELING number of piles, raft width, and raft thickness,
under uniform loading were investigated. For
In order to assess some of the characteristics of simplicity, square shaped rafts, including piled
piled raft behavior, three-dimensional finite and unpiled (for comparison purposes), were
element analyses of various piled raft modeled.
configurations have been undertaken. The
computer software PLAXIS 3D with the Mohr-

Table 1. Soil parameters

Soil Loose to Medium (M) Medium to Dense (D)

Medium (LM) Dense (MD)
Classification SP to SW SP to SW SP to SW SP to SW
Friction angle, φ (deg) 30 33 36 37
Relative density, RD (%) 35 50 75 88
Elastic modulus, Es (MPa) 21 30 45 53
3
Total Unit Weight (kN/m ) (10% moist) 19 19 19 19
Void ratio, e 0.6 0.55 0.53 0.50
Cohesion , c (kPa) 0.2 0.2 0.2 0.2
Poisson’s ratio, νs 0.3 0.3 0.3 0.3
Interface strength, Rinterface 0.67 0.67 0.67 0.67
Dilatancy angle, ψ (deg) 0 3 6 8

Soil Parameters embedded piles and rafts, and applying the load
The main parameters used in the Mohr-Coulomb to the system, the model equilibrium under
constitutive soil model are internal friction angle coefficient of earth pressure at-rest (Ko = 1-sinφ)
(φ), cohesion (c), elastic (Young’s) modulus (Es), was generated. An interface strength coefficient
Poisson’s ratio (νs), and dilatancy angle (ψ). The (Rinterface) of 0.67, as recommended by PLAXIS
stress state at failure is described with effective 3D Manual (2007), was implemented to model
friction angle and cohesion of soil (PLAXIS 3D the contact area between soil and foundation
Manual, 2007). For this parametric study, four including the raft and piles.
granular soil types (i.e., loose to medium,
medium, medium to dense, and dense sand) Pile and Raft Parameters
were selected. The soil parameters adopted are
presented in Table 1. In PLAXIS 3D, the raft and piles are considered
to behave linear-elastically. For the parametric
In order to avoid any complications during the study, the pile and raft parameters listed in
analyses, a cohesion value of 0.2 kPa was Table 2 were adopted.
adopted, as recommended by PLAXIS 3D Table 2. Pile and raft parameters
Manual (2007). Prior to introducing the

2
Parameter Raft Pile B
Elastic modulus (MPa) 34,000 30,000
t
Poisson’s ratio 0.2 0.2
3
Unit weight (kN/m ) 25 25

Note that Reul and Randolph (1994) refer to a Lp

study in which concrete samples taken from
bored piles as well as in situ integrity testing
show that the elastic (Young’s) modulus of the
piles is generally smaller than the design value
obtained from samples under less complex
production conditions. Therefore, a smaller
value was taken for the elastic modulus of the s
dp
piles than for the raft.
Configuration 1: Configuration 2:
Pile and Raft Configurations

In this parametric study, as shown in Figure 2.

six pile configurations and two raft widths were
investigated. In configurations 1 through 4,
square rafts with a width (B or Br) of 38 m were
used, whereas B was reduced to 20 m in
configurations 5 and 6. Configuration 1 had 169 n = 169 (13x13), s = 3dp n = 49 (7x7), s = 3dp
piles evenly distributed under the whole raft
area, with a spacing (s) of 3 pile diameter (dp). In Configuration 3: Configuration 4:
configuration 2, the spacing was increased to
6dp, and the total number of piles used (n) was
reduced to 49. In configuration 3, the piles were
placed only in the central area of the raft (n = 49,
s = 3dp). In configuration 4, the spacing was
increased to 6dp, and n = 16. In configurations 5
and 6, the piles (n = 49 and 16, respectively)
were evenly distributed under the whole raft
n = 49 (7x7), s = 3dp n = 16 (4x4), s = 6dp
area with s = 3dp and 6dp, respectively. Note that
in all configurations, the pile diameter (dp) was
Configuration 5: Configuration 6:
held constant at 1.0 m.

The pile length and raft thickness assigned for

each configuration are presented in Table 3. The
pile length (Lp) was selected as 19 and 38 m for
configurations 1 through 4, and 20 and 40 m for n = 49 (7x7), s = 3dp n = 16 (4x4), s = 6dp
configurations 5 and 6. Thus, Lp/B was set equal
to either 0.5, 1 or 2. Considering the commonly Figure 2. Pile configurations
used raft dimensions is 1- and 2-m thick rafts (t
or tr) were selected for the analysis. For
comparison purposes, the settlement behavior System Configuration
of the unpiled rafts is taken as the reference for Due to two-fold symmetry of the problem, only
the settlement behavior assessment of piled one quarter of the piled raft was modeled, as
rafts. shown in Figure 3. The foundation level was set
at the ground surface. Only vertical movement
was set along the symmetry plane boundaries.
In order to avoid any boundary effects on
stresses and displacements, the distance to the

3
vertical boundaries in the horizontal direction the maximum pile length modeled (i.e., 2Br +
was set to ten times the width of the raft 2/3Lp). For comparison purposes, H for a
modeled (i.e., 10Br/2) (Reul-Randolph, 2004), specific piled-raft configuration was kept the
and the total depth (H) to the lower rigid same for an unpiled raft. The values of H
boundary in the vertical direction was set to two assigned for each configuration are tabulated in
times the full width of the raft plus two-thirds of Table 6.

Table 3. Loading cases and conditions

For Configurations 1 through 4: For Configurations 5 and 6:

Case 1 2 3 4 Unpiled 1 Unpiled 2 5 6 7 8 Unpiled 3 Unpiled 4
Raft width, B (m) 38 38 38 38 38 38 20 20 20 20 20 20
Raft length, L (m) B B B B B B B B B B B B
Raft thickness, t (m) 2 1 2 1 2 1 2 1 2 1 2 1
Pile diameter, dp (m) 1 1 1 1 none none 1 1 1 1 none none
Pile length, Lp (m) 19 19 38 38 - - 20 20 40 40 - -
Ratio of Lp/B 0.5 0.5 1 1 - - 1 1 2 2 - -
Pile spacing, s (m) 3dp and 6dp - - 3dp and 6dp - -

Construction Process and Load Type the rafts are considered. The maximum uniform
The objective of this parametric study was pressure, including the weight of the raft, was
mainly to investigate the global response of the 250 kPa.
various piled rafts under uniform loading with
increasing intensity up to 250 kPa (5 ksf). The
step-by-step construction process in the finite
element analyses was as follows:
1. Generate in-situ stress state using Ko.
2. Install embedded piles.
3. Re-set all displacements to zero. Apply a
vertical pressure of 25 or 50 kPa, equivalent
to the weight of the raft concrete (unit weight
times thickness).
4. Install raft with actual stiffness. Remove the
vertical pressure from Step 3.
5. Apply vertical uniform pressure (q) with
increments of 25 kPa until the total applied
pressure, including the raft weight, reaches
250 kPa.

In all models, following the pile installation, the

weight of the raft was applied to the soil.
Simulating the real construction process, once
the raft concrete was set, the stiffness of the raft
was included in the model. By altering the raft
thickness from 1 to 2 m, the corresponding
variation in raft-soil stiffness ratio was
investigated. All results presented in this study
are related to the situation after the installation Figure 3. Finite element mesh as defined in
of the piles, so deformations due to the weight of PLAXIS 3D

4
RESULTS AND DISCUSSION rafts with a centralized pile group (configurations
3 and 4) the location of the maximum settlement
For a total vertical pressure of 250 kPa, the shifts from the raft center towards the edge.
settlement values were measured at the center,
at the mid-point between center and edge, at the Figure 6 shows the settlement performance of
mid-point of the edge, and at the corner of the piled rafts with varying pile length to raft width
rafts (points A, B, C, and D in Figure 4, ratios (Lp/Br) from 0.5 to 2. The values presented
respectively). The average settlement (δavg) is are the normalized central settlements beneath
expressed as a function of settlement at points the rafts supported on evenly distributed piles
A, B, C, and D, as follows. (configurations 1, 2, 5 and 6). Thus, the ratio of
pile group width to raft width (Bg/Br) is equal to
1.0. Note that for configurations 1 and 2 the raft
δ A+ 4δ B + 4δ C + 4δ D width (Br) is 38 m, whereas Br = 20 m for
δ avg = Eq. 1
13 configurations 5 and 6. The settlement
performance is observed for the raft thicknesses
of 1 and 2 m, while the pile spacings of 3dp and
6dp are used under a uniform pressure of 250
kPa. The values are normalized by the
A B C settlements of unpiled rafts with the respective
raft thickness.

As would be expected, the normalized central

settlement decreases with increasing pile length,
D as the proportion of load carried by the piles
increases. Comparing the settlement behaviors
Figure 4. Settlement observation points in of the evenly piled rafts in four soil types (LM, M,
PLAXIS 3D MD and D) indicates that the configurations with
Lp/Br = 1 are effective in reducing the central
Figure 5 shows the performance of piled rafts settlement by 30 to 50%.
with varying number of piles from 16 to 169
under a uniform pressure of 250 kPa. For the six Figure 7 shows the settlement performance of
pile configurations considered, the color piled rafts with varying pile lengths of 19 and 38
contours show the vertical settlement under six m. The values presented are the normalized
square rafts with a thickness (t or tr) of 1 m in central settlement and the normalized differential
soil type LM. In Figure 5, the pile length (Lp) is settlement between the center and corner points
19 m for configurations 1 through 4 and 20 m for of the piled rafts supported on evenly distributed
configurations 5 and 6. Due to two-fold piles (configurations 1 and 2) and on a central
symmetry of the problem, only one quarter of the pile group (configurations 3 and 4). Thus, the
piled raft is presented. The first set of settlement ratios of pile group width to raft width (Bg/Br) are
contours consisting of configurations 1 through equal to 1.0 and 0.5, respectively. Note that a
4, with Br = 38 m, is presented using a scale constant raft width (Br) of 38 m is used in the
from 100 to 320 mm. Due to less impact analysis. The settlement performance is
because of the smaller raft size, the second set observed for Lp/Br ratios (pile length to raft width)
of settlement contours consisting of of 0.5 and 1.0, while the raft thicknesses (tr) of 1
configurations 5 and 6, with Br = 20 m, is and 2 m and the pile spacings of 3dp and 6dp are
presented using a scale from 60 to 120 mm. used, under a uniform pressure of 250 kPa. The
values are normalized by the settlements of
The results show that the layouts with evenly unpiled rafts with the respective raft thickness.
distributed piles (configurations 1, 2, 5 and 6)
show less intensity of settlement as the pile As would be expected, both the normalized
spacing (s) is reduced from 6dp to 3dp, which central settlement and the differential settlement
indicates reduction in settlement regardless of decrease with increasing pile length, as the
the raft size selected. It is also noticeable that proportion of load carried by the piles increases.
for rafts supported on evenly distributed piles Comparing the settlement behaviors of the
(configurations 1 and 2) the maximum evenly piled rafts with Bg/Br = 1 (configurations 1
settlement occurs at the center, whereas in the and 2) and the centrally piled rafts with Bg/Br =

5
0.5 (configurations 3 and 4) indicates that the The addition of piles to the raft is effective in
configurations with Bg/Br = 1 are effective in reducing the central settlement. However, there
reducing the central settlement. On the other is an upper limit to the useful number of piles,
hand, for a constant raft thickness, the centrally beyond which little additional benefit is obtained.
piled rafts with Bg/Br = 0.5 are more effective in Between configurations 1 and 2, the
reducing the differential settlement, and the improvement achieved in central settlement
layouts with s = 3dp provide the best solution for control by increasing the number of piles from
differential settlement control. 49 to 169 is only 10%.

100 mm

60 mm
Config. 1, Case 2, s = 3dp Config. 2, Case 2, s = 6dp

Scale for Config. 1 through 4

Scale for Config.5 and 6

Config. 3, Case 2, s = 3dp Config. 4, Case 2, s = 6dp

120 mm
320 mm

Config. 5, Case 5, s = 3dp Config. 6, Case 5, s = 6dp

Figure 5. Vertical settlement contours for tr = 1m, q = 250 kPa, Lp = 19 (and 20) m, soil type LM

Matching some prerequisites at the lowest cost of the piles. The values presented are the
is another factor in determining the optimum normalized central settlement and the
layout for settlement control. The cost of the normalized differential settlement between the
foundation is broadly linked to the total length of center and corner points of the piled rafts
the piles, (nLp). Figure 8 shows the settlement supported on evenly distributed piles
performance of piled rafts versus the total length (configurations 1 and 2) and on a central pile

6
group (configurations 3 and 4). Thus, the ratios (n = 16 to 49), it is clear that increasing the
of pile group width to raft width (Bg/Br) are equal length of the piles is, for this case, a more
to 1.0 and 0.5, respectively. Note that a constant effective design strategy for improving
Br of 38 m is used in the analysis. The foundation performance than increasing the
settlement performance is observed for Lp/Br number of piles. The results show that at a given
ratios (pile length to raft width) of 0.5 and 1.0, total length of the piles (nLp), the layouts with s =
while the raft thicknesses of 1 and 2 m and the 6dp provide the better solution for differential
pile spacings of 3dp and 6dp are used, under a settlement control compared to the layouts with
uniform pressure of 250 kPa. The values are s = 3dp. If the layouts that have the same total
normalized by the settlements of unpiled rafts length of piles are considered (i.e.,
with the respective raft thickness. configurations 2 and 3), then the centrally piled
rafts with Bg/Br = 0.5 are more effective in
Comparing the settlement behaviors of the reducing the differential settlements.
uniformly piled rafts with Bg/Br = 1 (n = 49 to
169) and the centrally piled rafts with Bg/Br = 0 .5

1 1.0

0.9 0.9

0.8 0.8

0.7 0.7

0.6 0.6

0.5 0.5

0.4 0.4

0.3 0.3
Normalized Central Settlement

0.2 0.2

0.1 0.1

0 0.0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Soil LM Soil M
1.0 1.0

0.9 0.9

0.8 0.8

0.7 0.7

0.6 0.6

0.5 0.5

0.4 0.4

0.3 0.3

0.2 0.2

0.1 0.1

0.0 0.0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Soil MD Soil D
Pile Length / Raft Width

Figure 6. Ratio of pile length to raft width vs. central settlement.

7
 1
1.0

0.9
0.8

0.8
0.6

0.7
0.4

0.6
0.2

0.5
0.0

0.4
0 5 10 15 20 25 30 35 40 -0.2

Soil LM 0 5 10 15 20 25 30 35 40

Soil LM
1.0 1.0

Normalized Centre- to-Corner Differential Settlement

0.9 0.8

0.8 0.6

0.7
0.4

0.6
Normalized Central Settlement

0.2

0.5
0.0

0.4
0 5 10 15 20 25 30 35 40 -0.2
0 5 10 15 20 25 30 35 40
Soil M
Soil M
1.0 1.0

0.9
0.8

0.8
0.6

0.7
0.4
0.6

0.2
0.5

0.4 0.0
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40

Soil MD Soil MD
1.0 1.0

0.9
0.8

0.8
0.6

0.7

0.4
0.6

0.2
0.5

0.4 0.0
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
Soil D Soil D
Pile Length (m)

Conf.1 Bg/Br=1.0 tr=1m s=3m n=169 Conf.1 Bg/Br=1.0 tr=2m s=3m n=169
Conf.2 Bg/Br=1.0 tr=1m s=6m n=49 Conf.2 Bg/Br=1.0 tr=2m s=6m n=49
Conf.3 Bg/Br=0.5 tr=1m s=3m n=49 Conf.3 Bg/Br=0.5 tr=2m s=3m n=49
Conf.4 Bg/Br=0.5 tr=1m s=6m n=16 Conf.4 Bg/Br=0.5 tr=2m s=6m n=16
0

Figure 7. Pile length vs. central and differential settlements

8
 1 1.0

0.9 0.8

0.8 0.6

0.7 0.4

0.6 0.2

0.5 0.0

0.4 -0.2
0 1000 2000 3000 4000 5000 6000 7000
0 1000 2000 3000 4000 5000 6000 7000 Soil
LM Soil LM
1.0 1.0

Normalized Centre- to-Corner Differential Settlement

0.9 0.8

0.8 0.6

0.7 0.4
Normalized Central Settlement

0.6 0.2

0.5 0.0

0.4 -0.2
0 1000 2000 3000 4000 5000 6000 7000 0 1000 2000 3000 4000 5000 6000 7000

Soil M Soil M
1.0 1.0

0.9
0.8

0.8
0.6

0.7

0.4
0.6

0.2
0.5

0.4 0.0
0 1000 2000 3000 4000 5000 6000 7000 0 1000 2000 3000 4000 5000 6000 7000

Soil MD Soil MD
1.0 1.0

0.9
0.8

0.8
0.6

0.7

0.4
0.6

0.2
0.5

0.4 0.0
0 1000 2000 3000 4000 5000 6000 7000 0 1000 2000 3000 4000 5000 6000 700
Soil D Soil D
Total Pile Length (m)

Conf.1 Bg/Br=1.0 tr=1m s=3m n=169 Conf.1 Bg/Br=1.0 tr=2m s=3m n=169
Conf.2 Bg/Br=1.0 tr=1m s=6m n=49 Conf.2 Bg/Br=1.0 tr=2m s=6m n=49
Conf.3 Bg/Br=0.5 tr=1m s=3m n=49 Conf.3 Bg/Br=0.5 tr=2m s=3m n=49
Conf.4 Bg/Br=0.5 tr=1m s=6m n=16 Conf.4 Bg/Br=0.5 tr=2m s=6m n=16
0
Figure 8. Total pile length vs. central and differential settlements

9
 CONCLUSION Geoenvironmental Engineering, pp. 707-
708.
A close assessment of the results leads to the
4. Poulos, H.G. (2000). “Chapter 16: Practical
following conclusions for practical design:
Design Procedures for Piled Raft
Foundations”, Design Applications of Raft
1. The addition of piles to the raft is effective in Foundations, Hemsley, J.A. (Editor),
reducing the central settlement. However, Thomas Telford Ltd., London.
there is an upper limit to the useful number
5. Poulos, H.G. (2001). “Pile Raft Foundations:
of piles, beyond which little additional benefit
Design and Applications”, Geotechnique,
is obtained. This limit is usually less than the
Vol. 51, No. 2, pp. 95-113.
number of piles conventionally used to
support the full load from the structure. 6. PLAXIS 3D Foundation, Version 2, Finite
Element Code for Soil and Rock Analyses.
2. The longer the piles, the more effective they
are in reducing the central and the 7. PLAXIS 3D Foundation Material Models
differential settlements. Manual (2007). Version 2, EDs, Brinkgreve,
R.B.J. and Swolfs, W.M.
3. For control of differential settlement, if
loading is uniformly distributed, the optimum 8. Reul, O., and Randolph, M.F. (2004).
performance is likely to be achieved by “Design Strategies for Piled Rafts Subjected
concentrating the piles near the centre to Nonuniform Vertical Loading”, J.
area,rather than using a large number of Geotechnical and Geoenvironmental
evenly distributed piles beneath the raft Engineering, Vol. 130, No. 1, pp. 1-13.
area, or increasing the raft thickness.
4. For each value of pile length, an optimum
value of the quantity nLp exists,
corresponding to the maximum reduction of
the differential settlement and to values of
Bg /Br in the 0.5 range.
5. From the presented analysis and method of
design, reduction in number of piles could
be achieved by taking into consideration the
load sharing capacity of the soil. By applying
this method, considerable reduction in the
foundation cost could be achieved.

REFERENCES

1. Prakoso, W.A., and Kulhawy, F.H. (2001).

“Contribution to Piled Raft Foundation
Design”, Journal of Geotechnical and
Geoenvironmental Engineering, Vol. 127,
No. 1, pp. 17-24.
2. Cunha, R.P., Poulos, H.G., and Small, J.C.
“Investigation of Design Alternatives for a
Piled Raft Case History”, Journal of
Geotechnical and Geoenvironmental
Engineering, Vol. 127, No. 8, pp. 623-641.
3. De Sanctis, L. and Russo, G. (2002).
“Discussion of ‘Contributing to piled raft
foundation design’ by Prakoso, W.A. and
Kulhawy, F.H., Vol. 127, No. 1, pp. 17-24.”,
Journal of Geotechnical and

10

View publication stats