Professional Documents
Culture Documents
net/publication/269196621
CITATIONS READS
13 517
2 authors:
Some of the authors of this publication are also working on these related projects:
Geotechnical mitigation strategies for structures impacted by earthquake induced fault ruptures View project
Laboratory and Numerical Analysis of Piled Rafts and Offshore Foundations Considering Interaction and Bearing Mechanism View project
All content following this page was uploaded by Abbasali Taghavi Ghalesari on 24 April 2016.
l
Graduate Student, Department of Civil Engineering, Iran University of Science and
y
Technology, Tehran, Iran; Rasouli.Habib@gmail.com
on
ABSTRACT: Piled raft foundations are often used when a raft foundation alone does
se
not satisfy the design requirements or the raft settlements exceed allowable values.
The addition of piles may improve both the ultimate load capacity and the settlement
u
performance of the raft. In this paper, the behavior of piled raft foundation is
investigated using three-dimensional finite element analysis and 100g centrifuge tests.
a l
The effect of a gravel layer on the maximum and differential settlement behavior of
piled raft is considered. Based on the results, the behavior of piled raft foundation is
INTRODUCTION
on
dependent on the thickness of the gravel layer and the particle size.
r s
Pile enhanced raft has been experienced as one of the most effective ways for
p e
increasing the bearing capacity of a foundation system. The system is known as a
hybrid foundation or a piled raft foundation (PRF). The piled raft foundation is a
o r
recent design concept as one of the effective methods to reduce the settlements of
superstructures. In soft soils where a raft alone may not provide the bearing
capacity, connected piles to enhance the performance of the raft or non-connected
F
piles to reinforce the supporting soil are usually added. Traditionally, one common
perception of engineers is that more piles can lead to a greater reduction in raft
settlements. However, as pointed out by Poulos and Davis (1980), the number of piles
required to reduce settlements under the working load to a tolerable limit is usually
small, and any further addition of piles may result in only marginal further reductions
in settlements. Conventionally, there are two main design methodologies for piled raft
foundation. The first design method is based on reducing raft foundation settlement by
adding a certain number of piles. Another design method assumes that all structure
loads to be transferred to piles without considering contribution of load taken by the
raft.
Piled raft foundations has received much attention recently. In order to address the
design of piled rafts, many numerical and experimental methods has been developed.
Prakoso and kulhawy (2001) proposed a simplified linear elastic and nonlinear plane
strain finite element model to evaluate the effects of raft and pile group system
geometries on the average and differential displacements of the piled rafts. Reul and
Randolph (2003) presented a three-dimensional finite element method to investigate
the behavior of piled rafts with associated and non-associated flow rules for modeling
the plastic behavior of soil. Zheng et al. (2008) performed finite element analyses to
investigate the behavior of composite CFG (Cement-Flyash-Gravel)–lime pile
Downloaded from ascelibrary.org by Ondokuz Mayis Universitesi on 05/23/14. Copyright ASCE. For personal use only; all rights reserved.
on
resting on pile-reinforced sand. El Sawwaf (2010) studied the effectiveness of using
short piles in connected or non-connected piled rafts under eccentrically load instead
of long piles.
se
Some studies have been devoted to the effect of sand or gravel cushion on the non-
connected piled raft performance (Liang et al. 2003), while no studies were aimed to
l u
the connected ones. To consider the effects of a surface gravel layer on the settlement
and bearing behavior of unpiled and piled raft foundation, 100g centrifuge tests and
a
three-dimensional finite element analyses have performed in this study.
EXPERIMENTAL PROCEDURE
on
Model Box and Materials
r s
Full-scale field tests to consider the behavior of piled raft foundation is costly and
p e
time consuming. Furthermore, it is difficult to control the permanent changes in
groundwater level. Therefore, simulation of in-situ condition using model tests such as
r
centrifuge tests can be useful. In centrifuge modelling, a model geometrically scaled
down N times and prepared from the prototype material is accelerated at N times
F o
Earth’s gravity: the centrifuge acceleration reproduces the same stress and strain in the
model as in the prototype. The centrifuge used for the present study is installed at Iran
University of Science and Technology. A model scale of 1/100 was used with a
nominal centrifugal acceleration of 100g. The observations from the model can be
converted to the prototype scale using the scaling factors reported in Table 1.
The laboratory model tests were conducted in a test box, having inside dimensions
of 0.62 m 0.20 m in plan and 0.18 m in depth. The box was made from Plexiglas with
steel frame. The dimensions of square raft (steel plate) was 55 55 mm (5.5 m at
prototype scale) with the thickness of 10 mm (1 m). Model piles with 5.0 and 5.6 mm
diameters made of aluminum bars (E=7 104 MPa) were used. The length of the piles
was 84 mm (8.4 m).
Velocity Time
1 N N 1
(undrained conditions) (dynamic)
Mass N3 1 Frequency 1 N
Experiments were performed using dry Firouzkouh (No. 161) siliceous sand. This
kind of sand has uniform aggregation and in the present study, it has relative density
of about 55%. For modeling gravel layer, Firouzkouh D11 sand was used which
l y
n
models granular layer by scaling. Some specification of used soil is presented in Table
2. The specific gravity of the soil particles (Gs) was determined by the gas jar method.
The particle size distribution was determined using the dry sieving method and the
o
e
results are shown in Fig. (1).
n
%
Firouzkouh sand (161) 2.658 0.97 0.55 0.30 32 0.97 2.58 0.20
Firouzkouh sand (D11)
*fine content
2.650
s o
0.89 0.63 1.15 - 0.96 1.43 0.15
e r
r p
F o
l
required to produce a specific relative density, was first weighed and placed in the
y
on
tank and tamped by a steel plate until achieving the required layer height. The relative
density of sample was 55%. After sampling, the piles were installed singly one-by-one
using a special guide system which held the piles vertical during the installation. No
se
visible movement in the sand surface was observed during the installation process.
Then the sand surface was leveled again and finally the model raft was placed (Fig.
2b). Some points on the piles and raft or in the soil were instrumented by LVDTs
l u
(linear variable displacement transducers). A data logger instrument are applied to
gathering required data related with piles, pile cap and soil parameters.
n a
s o
e r
r p
F o (a) (b)
FIG. 2. (a) Soil layers (b) Structural components setup
elements. Fig. (4) shows a typical finite element mesh used in the numerical study.
Material properties of soil and piled raft are tabulated in Table 3.
Downloaded from ascelibrary.org by Ondokuz Mayis Universitesi on 05/23/14. Copyright ASCE. For personal use only; all rights reserved.
l y
on
se
l u
n a
s o
FIG. 3. Model schemes
e r
r p
F o
(a) (b)
FIG. 4. (a) Typical finite element mesh (b) Pile-soil interface modeling
Internal Dilation
Young’s Unit
Poisson’s Cohesion, friction angle,
Material modulus, weight,
ratio, c (kPa) angle, (deg)
E (MPa) (kN/m3)
(deg)
Sand 20 0.20 18 0.3 32 2
Downloaded from ascelibrary.org by Ondokuz Mayis Universitesi on 05/23/14. Copyright ASCE. For personal use only; all rights reserved.
Gravel 35 0.15 18 - 30 -
Pile 25000 0.15 24 - - -
Raft 20000 0.15 24 - - -
y
The pile head was connected to the raft rigidly. A relatively fine mesh was used
near the structural components while a coarser mesh was used further from the pile
and raft. Lateral boundaries are restrained by horizontal translations to allow
n l
o
downward movement of the soil layers and the beneath of the model is fixed in three
orthogonal directions. For the far-field boundaries, the distance of the boundary from
e
the edge of the raft was greater than the pile length. This limit prevents the effects of
s
boundaries. After initial equilibrium, the vertical loading was applied on the top of the
raft surface. Since modeling of the entire pile installation process is rather
Constitutive Modeling
n a
s o
The material behavior of the soil was modeled with a Mohr–Coulomb model, and
to simplify the analysis process, average values of material parameters (as mentioned
e r
in Table 3) were adopted for the soil layer. Since the raft and piles have great Young’s
modulus in comparison with the soil, they remain in elastic range. Due to the
p
aforementioned reason, they were modeled with a non-porous linear elastic model.
Interface elements were used at the soil-pile interface. The modeling techniques
o r
used for the pile-soil interface were generally divided into two types: thin layer
element and slip element. The former was used by Jeong et al. (2004), in which the
F
slip behavior between the adjacent surfaces could be considered. The latter was used
by Reul and Randolph (2003) and a middle layer is used to model the interface using
the behavior of the soil. In this study, slip interface elements was used at the pile-soil
interface. When no slip occurs between the soil and pile elements; the nodes at the
interface have identical coordinates and the distance between the two surfaces is zero.
When contact occurs, the relationship between shear force and normal pressure ( P ) is
governed by a modified Coulomb’s friction theory. The slip interface elements are
completely defined by their geometry, a friction coefficient ( ), where tan .
Interface friction angle, ( ) can be estimated by the following equation (Jeong et al.
2004) using soil internal friction angle ( ):
sin cos
(1)
1 sin 2
settlement relationship for different types of foundation. In this figure, Smax is the
calculated and measured settlement at the raft center, Br is raft width and q is the total
applied load. Piled raft with 4 piles has pile spacing to diameter (S/Dp) of 4.9 and this
ratio for piled raft with 9 piles is 3.7.
l y
on
se
l u
n a
(a)
s o (b)
e r
FIG. 5. Load-settlement curve from the (a) test results (b) numerical analyses for
different foundation type
r p
As expected, in a certain value of normalized settlement, bearing capacity of raft is
o
lower than that of piled raft with different number of piles from both numerical and
F
experimental results. This is more obvious for higher levels of the applied load and
this shows that the ultimate bearing capacity of the raft is reached. Comparing Fig.
(5a) and (5b), the effect of adding piles to a raft foundation is more significant from
the test results in comparison with numerical analysis. Numerical analyses
underestimates the bearing capacity of the foundation except for raft foundation.
with cushion. In this study, the gravel layer was modeled with Firouzkouh D11 sand
and it was simulated according to Table 3 in the numerical analysis. Fig. (6) shows the
effect of the gravel layer thickness on the settlement and bearing behavior of the raft
and piled raft with 9 piles.
Downloaded from ascelibrary.org by Ondokuz Mayis Universitesi on 05/23/14. Copyright ASCE. For personal use only; all rights reserved.
l y
o n
(a)
se (b)
n a
s o
At a given applied load, the settlement of piled raft initially decreased with gravel
layer thickness and then increased for piled raft with 1.5 m gravel layer (tg=1.5 m). It
e r
is due to increasing the elastic settlements of the gravel layer caused by self-weight
and decreasing the piles performance caused by load distribution over a larger area.
Thus, the required gravel layer thickness to minimize the piled raft settlements is
tg=0.5 m or less.
r p
A comparison between the results for raft (solid line) and piled raft (dotted line)
F o
with Fig. (5) shows that adding gravel layer leads to reduction in the effect of piles on
the foundation performance. It is probably due to the fact that 1.5 m gravel layer cause
a reduction in the perimeter length of piles. Another point that realized from the figure
is that numerical analyses, however, underestimate the bearing capacity of the
foundations.
Particle Size
From the test results, the effect of the soil particle-size on the settlements and
bearing capacity of the foundation was considered. Fig. (7) shows the load-settlement
relationship for piled raft (4 piles) with mean particle-size of D50=0.30 and 1.15. As
shown, particle size has a negligible effect on the foundation performance until a
certain applied load (of about 600 kPa) and after that, soil with D50=1.15 is more
efficient.
l y
FIG. 7. Load-settlement curve for different particle-size from test results
on
Center to Corner Differential Settlement
se
In general, the effect of differential settlement on the bending moment and
l u
optimum design of piled raft foundations is not negligible. For this reason, it is
considered as an important factor in this study. Fig. (8) shows the variation of unpiled
numerical analysis.
n a
and piled raft differential settlement (Sd) with or without gravel layer from the
s o
e r
r p
F o
FIG. 8. Load-settlement curve for different particle-size from test results
As shown, the gravel layer can reduce the differential settlement of unpiled and
piled raft foundation. Clearly, piled raft with larger gravel layer thickness has lower
differential settlement. Considering the differential settlement reduction, adding piles
or gravel layer have little differences.
CONCLUSIONS
applied load, grain-size also affected on the bearing and settlement behavior of the
foundation.
REFERENCES
Bolton, M.D., Gui, M.W., Garnier, J., Corte’, J.F., Bagge, G., Laue, J., and Renzi, R.
(1999). “Centrifuge Cone Penetration Tests in Sand.” Géotechnique, 49(4), 543–
552.
l y
Cao, X.D. Wong, I.H. and Chang, M.F. (2004). “Behavior of model rafts resting on
on
pile-reinforced sand.” Journal of Geotechnical and Geoenvironmental Engineering,
130(2), 129–138.
se
Corte’, J. F., Garnier, J., Cottineau, L. M. & Rault, G. (1991). “Determination of
model soil properties in the centrifuge.” Proceedings of the international
u
conference on the centrifuge, Boulder, Colorado, pp. 607–614.
El Sawwaf, M. (2010). “Experimental study of eccentrically loaded raft with
on
Horikoshi, K. and Randolph, M.F. (1996). “Centrifuge modelling of piled raft
foundations on clay.” Géotechnique, 46(4), 741–752.
Jeong, S., Lee, J. and Lee, C. (2004). “Slip effect at the pile–soil interface on
r s
dragload.” Computers and Geotechnics, 31, 115–126.
Liang, F.Y., Chen, L.Z. and Shi, X.G. (2003). “Numerical analysis of composite piled
443–453.
p e
raft with cushion subjected to vertical load.” Computers and Geotechnics 30(6),
o r
Nguyen, D.D., Jo, S.B. and Kim, D.S. (2013). “Design method of piled-raft
foundations under vertical load considering interaction effects.” Computers and
Geotechnics, 47, 16–27.
F
Poulos, H.G. and Davis, E.H. (1980). “Pile foundation analysis and design.” New
York. Wiley.
Prakoso, W.A. and Kulhawy, F.H. (2001). “Contribution to piled raft foundation
design.” Journal of Geotechnical and Geoenvironmental Engineering, 127(1), 17–
24.
Reul, O. and Randolph, M.F. (2003). “Piled rafts in overconsolidated clay: comparison
of in situ measurements and numerical analyses.” Géotechnique, 53(3), 301–315.
Zheng, J.J., Abusharar, S.W. and Wang, X.Z. (2008). “Three–dimensional nonlinear
finite element modeling of composite foundation formed by CFG–lime piles”
Computers and Geotechnics. 35, 637–643.