Uplift Capacity of Axially Loaded Piles in Clays

Vishwas N. Khatri1 and Jyant Kumar2

Abstract: The present study aims at determining the uplift capacity of axially loaded piles in clays whose undrained cohesion increases
linearly with depth. With the application of the axisymmetric static limit analysis approach, proposed recently by the writers, the variation
of the nondimensional uplift factors with respect to changes in the embedment ratio 共H / B兲 has been obtained for several rates of increase
of soil cohesion with depth defined in terms of a nondimensional factor m for given values of soil cohesion 共c0兲 along the ground surface
and diameter 共B兲 of pile. The uplift resistance has been evaluated in the form of the uplift factors, Fcb and Fct due to the components of
the shaft resistance and the total resistance, respectively. For the given values of c0 and B, the magnitude of the uplift resistance increases
continuously with an increase in the value of m. Furthermore, the Fcb values are found to remain unaffected with the variation in the shaft
adhesion, whereas, the Fct values increase continuously with an increase in the adhesion factor of the pile shaft.
DOI: 10.1061/共ASCE兲GM.1943-5622.0000064
CE Database subject headings: Clays; Limit analysis; Piles; Plasticity; Uplifting; Axial loads.
Author keywords: Clays; Limit analysis; Piles; Plasticity; Uplift resistance.

Introduction 1985; Das 1989; Shin et al. 1993兲, for determining the uplift
resistance of piles, are based on the assumption that the soil co-
Piles are frequently used to resist the uplift forces for the different hesion remains constant with depth. However, in reality, labora-
structures such as transmission towers and dry docks. The piles tory experiments performed on saturated normally consolidated
generate the required uplift resistance from the contribution of the and lightly overconsolidated clays indicate approximately a linear
frictional resistance along the pile shaft and the base resistance at increase of undrained cohesion with depth 共Bishop 1966兲. The
the pile tip 共Meyerhof 1973; Das and Seeley 1982; Shelke and effect of increase of soil cohesion with depth on the uplift resis-
Patra 2008兲. Quite often, the piles are embedded in clays rather tance of piles has not been addressed so far. In the present study,
than in granular soils. A number of experimental studies have it has been attempted to study the effect of linearly increasing
been conducted by different researchers to determine the uplift cohesion with depth on the uplift resistance of piles. The axisym-
capacity of piles in clays 共Ali 1968; Bhatnagar 1969; Das and metric static limit analysis formulation in combination with finite
Seeley 1982; Azim 1985; Das 1989; Shin et al. 1993兲. It has been elements, proposed recently by the writers 共Khatri and Kumar
2009a,b兲 has been used in the present study. The results are pre-
generally observed from these studies that the contribution of the
sented in the form of the nondimensional uplift factors to account
tip resistance, especially at greater depths, remains generally very
for the contribution of the tip and the total resistance. The failure
small as compared to the total frictional resistance. The frictional
patterns are also examined.
resistance of the pile shaft is determined by simply multiplying
the mobilized unit shaft adhesion with the surface area of the pile
shaft. The tip resistance of the pile under tensile load for ␾ = 0 is
assumed to remain the same as under the compressive load; Problem Definition
where ␾ = friction angle of soil mass. Meyerhof 共1973兲 presented
a semiempirical expression for estimating the tip resistance of A pile is embedded in clay whose cohesion, under undrained con-
dition 共␾ = 0兲, increases linearly with depth 关refer Fig. 1共b兲兴; the
piles in clays under uplift load and it was indicated that the uplift
cohesion at a given depth is defined by the following expression:
factor, due to the tip resistance, increases with the embedment
ratio 共H / B兲 up to a limiting value of nine; where the parameters
H and B refer to the total embedment depth and the diameter of c = c0 + mc0h/B 共1兲
pile, respectively. All the available studies 共Ali 1968; Bhatnagar where c and co = values of cohesion at a depth h and at ground
1969; Sowa 1970; Meyerhof 1973; Das and Seeley 1982; Azim surface, respectively, and m = nondimensional factor, which on
multiplication with c0 / B, defines the rate at which the cohesion
1
Postdoctoral Fellow, Civil Engineering Dept., Indian Institute of Sci- increases with depth; in this expression, the cohesion is related
ence, Bangalore 560012, India. with h as the value of c increases linearly with depth, and the term
2
Associate Professor, Civil Engineering Dept., Indian Institute of Sci- B in the denominator has been simply used for the purpose of
ence, Bangalore 560012, India 共corresponding author兲. E-mail: making the factor associated with c0 to be a nondimensional. The
[email protected]
pile is subjected to vertical upward load with its point of applica-
Note. This manuscript was submitted on October 21, 2009; approved
on February 24, 2010; published online on March 12, 2010. Discussion tion coinciding with the axis of the footing. The soil mass is
period open until July 1, 2011; separate discussions must be submitted for assumed to be weightless, perfectly plastic, and it obeys an asso-
individual papers. This paper is part of the International Journal of ciated flow rule; Mohr Coulomb failure criterion is assumed to be
Geomechanics, Vol. 11, No. 1, February 1, 2011. ©ASCE, ISSN 1532- applicable with ␾ = 0. For various values of m and embedment
3641/2011/1-23–28/$25.00. ratio 共H / B兲, it is required to determine the magnitude of the pull-

INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JANUARY/FEBRUARY 2011 / 23

Downloaded 13 May 2012 to 180.211.192.67. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
 Qt

τrz= 0,σz = 0 mco H

Ground level co cb = co +
B/2 B
c
Pile S T

Depth below ground level h

-αs c ≤ τrz≤ αs c H
Line of symmetry

r
O R
-αb cb ≤ τrz ≤ αb
Lg
mco h
σz c = co +
τrz= 0

D B
τrz
σr

V U B – Diameter of pile

-z cb – Cohesion value at pile

base

(a) (b)

Fig. 1. 共a兲 Problem definition and the associated stress boundary

conditions; 共b兲 cohesion variation along the depth

out resistance of the pile. In the present analysis, the soil mass is
assumed to be weightless 共unit weight, ␥ = 0兲. For ␾ = 0, the fric-
tional resistance offered by the pile shaft remains unaffected in
the presence of ␥, on the other hand, the uplift resistance from the
pile base simply reduces by ␥H 共Meyerhof 1973兲.

Mesh and Boundary Conditions

By considering the symmetry about a vertical line passing

through the center of the footing, only one-half of a plane domain
in a r-z domain is employed, where the parameters r and z rep-
resent dimensions along the radial and vertical axes, respectively. Fig. 2. Mesh used in the analysis for: 共a兲 H / B = 1.0; 共b兲 H / B = 5.0
The domain and the associated applicable stress boundary condi-
tions are shown in Fig. 1共a兲. As shown in Fig. 1共a兲, along the
interface of soil-pile shaft and soil-pile base, the shear stress Analysis
boundary condition is specified in a bound format; that is, 共1兲
−␣sc ⱕ ␶rz ⱕ ␣sc along the pile shaft and 共2兲 −␣bcb ⱕ ␶rz ⱕ ␣bcb The axisymmetric static limit analysis formulation, proposed ear-
along the pile base, where ␣s and ␣b refer to the adhesion factor lier by Khatri and Kumar 共2009a,b兲, is used for the present prob-
along the soil-pile shaft and the soil-pile base, respectively, and lem. The nodal stresses 共␴r, ␴z, ␶rz, and ␴␪兲 are treated as basic
␶rz represents the shear stress. The correct sign of the shear stress, unknown variables. For any element, the stresses are assumed to
which maximizes the objective function, is automatically chosen vary linearly; this approximation ensures the linear nature of the
during the process of optimization. The depth 共D兲 of the domain equations 共needed for the optimization problems to be solved with
below the pile base, and the horizontal extent 共Lg兲 of the domain linear programming兲 which arrives on account of the satisfaction
from the axis of the pile, were kept in a range of 4.3B–10.3B and of 共1兲 the equilibrium conditions for an element and 共2兲 statically
6.9B–27.4B, respectively. The values of 共Lg兲 and 共D兲 were se- admissible stress discontinuities permitted at the interfaces of dif-
lected by using a number of trials such that for any value of H / B ferent adjacent elements. For a discontinuity to remain statically
and m, 共1兲 the plastic zones generated from the analysis are con- admissible, the normal and shear stresses across this path should
tained well within the chosen boundaries of the domain and 共2兲 remain to be always continuous. By the integration of the stresses
the magnitude of the collapse load does not change even if the along the interface of soil-pile base and soil-pile shaft, the expres-
size of the domain is increased further. The domain was dis- sion for the objective function, which is the value of the total
cretized into a number of three-noded triangular elements. Typical collapse load, is generated. As per the lower bound theorem of the
meshes for H / B = 1 and 5 are shown in Fig. 2. The mesh was limit analysis, in order to achieve the statically admissible stress
created in a manner such that the sizes of all the elements de- field, associated with a stress state lesser than or equal to at yield
crease continuously when approaching toward the singular point everywhere in the domain, the value of the total collapse load
共point R in Fig. 1兲. For the different cases, the total number of 共objective function兲 needs to be maximized subjected to various
elements, the total number of nodes, and the total number of equality and inequality constraints 共Sloan 1988兲.
stress discontinuities were varied in a range of 3,216–7,284; The various equality constraints are generated from the satis-
9,648–21,852; and 4,770–5,166, respectively. Greater numbers of faction of 共1兲 element equilibrium; 共2兲 conditions for the discon-
nodes, elements, and discontinuities were chosen for higher val- tinuities to be statically admissible; and 共3兲 the associated stress
ues of H / B. boundary conditions. Whereas, the enforcement of the linearized

24 / INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JANUARY/FEBRUARY 2011

Downloaded 13 May 2012 to 180.211.192.67. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
Mohr-Coulomb yield criterion together with the additional con- The inequality constraints imposed on four stresses at node i,
straints on account of the axisymmetric nature of the problem can be expressed in the following form:
forms the different inequality constraints. For the present prob-
关Aiyield兴 p⫻4兵␴i其4⫻1 ⱕ 兵biyield其 p⫻1 共7兲
lem, as per Harr-Von Karman hypothesis, the value of ␴␪ is kept
closer to ␴3. In the following paragraphs, the procedure for the The previous references of the writers 共Khatri and Kumar
generation of the various constraints and objective function ex- 2009a,b兲 can be referred for the description of the various terms
pression is explained in brief. in the different constraints equations 关Eqs. 共3兲–共5兲 and 共7兲兴.

Constraints for Element Equilibrium Objective Function

The equilibrium of each element can be ensured by means of the The objective function 共the value of the collapse load兲 for the
following two equations: axisymmetric problem can be defined by the following expres-
sion:
⳵ ␴r ⳵ ␶rz ␴r − ␴␪

冕
+ + =0 共2a兲
⳵r ⳵z r
Q = 2␲ ␴nrsds 共8兲
Ls
⳵ ␶rz ⳵ ␴z ␶rz
+ + =␥ 共2b兲 where Q = value of the collapse load; ␴n = normal stress acting
⳵r ⳵z r
over the boundary surface s; Ls = length of the boundary with two
Tensile normal stresses are considered positive, and the directions boundary nodes 1 and 2; and rs = mean radius of boundary edge
of the positive shear stress are indicated in Fig. 1共a兲. The values Ls. The above equation can be written in the matrix form as
of the terms 共␴r-␴␪兲 / r and ␶rz / r are specified at the centroid of the
Q = 兵gs其1⫻8
T
兵␴s其8⫻1
element, implying that the equilibrium equations will be satisfied
precisely at the centroid of the element not everywhere. It will After assembling the various equality and inequality constraints,
generate two equality constraints the linear programming problem can now be stated as
关Aeequil兴2⫻12兵␴e其12⫻1 = 兵beequil其2⫻1 共3兲 Maximize 兵g其T兵␴其 共9a兲

Subjected to 关A1兴兵␴其 = 兵b1其 共9b兲

Constraints for Statically Admissible Discontinuities
关A2兴兵␴其 ⱕ 兵b2其 共9c兲
The statical admissibility of discontinuity generates four equality
constraints on 16 nodal stresses associated with four nodes defin- where 兵␴其T = 关␴r,1 ␴z,1 ␶rz,1 ␴␪,1 ␴r,2 ␴z,2 ␶rz,2 ␴␪,2 , . . . , ␴r,N
ing the line of discontinuity ␴z,N ␶rz,N ␴␪,N兴. Like the previous studies 共Khatri and Kumar
2009a,b; Kouzer and Kumar 2009兲, the optimization in this study
关Astat
dc
兴4⫻16兵␴dc其16⫻1 = 兵bstat
dc
其4⫻1 共4兲
was also performed with the usage of LINPROG, a library pro-
gram in MATLAB.
Boundary Conditions
The stress boundary conditions, along any boundary edge l, can Definition of Uplift Factors Fct and Fcb
be defined in the following manner: ␴ln,1 = q1, ␴ln,2 = q2, ␶lnt,1 = t1,
and ␶lnt,2 = t2; where q1, q2, t1, and t2 = values of respective nodal After obtaining the value of the collapse load 共Q兲, the nondimen-
normal and shear stresses prescribed along the boundary. These sional uplift factors Fcb and Fct, with regard to the base and total
constraints can be easily expressed as capacity, for a weightless 共␥ = 0兲 soil medium are defined by using
the following expression:
关Albound兴4⫻8兵␴l其8⫻1 = 兵blbound其4⫻1 共5兲
Qi
= cbFci 共10兲
␲共0.5B兲2
Yield Condition
In the Eq. 共10兲, the subscript i in Qi and Fci should be replaced
For the present problem, as per Harr-Von Karman hypothesis, the with the letters b and t to define the uplift factors Fcb and Fct
value of ␴␪ is kept closer to ␴3. To ensure that the value of ␴␪ corresponding to the base resistance Qb and total resistance Qt
remains close to ␴3, the following three additional inequality con- 共comprising of both the shaft and base capacity兲, respectively;
straints are specified: 共1兲 ␴␪,i ⱖ ␴r,i; 共2兲 ␴␪,i ⱖ ␴z,i; and 共3兲 ␴␪,i where cb = value of the soil cohesion at the pile base. It should be
ⱕ ␴3f,i; where ␴␪,i, ␴r,i, and ␴z,i = stresses associated with node i, mentioned that in the present analysis, the total uplift resistance
and ␴3f,i is the minor principal stress 共least compressive normal 共Qt兲 has been optimized, and from the obtained solution, the
stress兲 at failure. value of Qb has then been obtained simply from the integration of
To ensure that the finite-element formulation leads to a linear the normal stresses along the pile base.
programming problem, following Bottero et al. 共1980兲, the origi-
nal yield surface is approximated by a regular polygon of p sides
inscribed to the parent yield surface. It will lead to Results and Comparisons
Ak␴r + Bk␴z + Ck␶rz ⱕ ␧; k = 1,2, . . . ,p 共6兲
The computations are performed for different values of 共1兲 H / B
where Ak = cos共2␲k / p兲; Bk = −cos共2␲k / p兲; Ck = 2 sin共2␲k / p兲; and ranging from 0 to 15 and 共2兲 m varying from 0 to 5. In all the
␧ = 2c cos共␲ / p兲 cases, the adhesion factor corresponding to the pile base 共␣b兲 has

INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JANUARY/FEBRUARY 2011 / 25

Downloaded 13 May 2012 to 180.211.192.67. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
 100 15 Table 1. Percenatge of the Load Shared by Pile Base and Shaft with
Fcb
Fct
αb =1, αs = 1.00 ␣b = 1 and ␣s = 1
80 0.50 12
0.25
m=0 m=5
60 9
Load shared Load shared Load shared Load shared

Fcb
Fct

by base by shaft by base by shaft

40 6
H/B 共%兲 共%兲 共%兲 共%兲
20 3 0.25 86.42 13.58 92.43 7.57
m=0
0.5 77.00 23.00 86.15 13.85
0 0
1 64.28 35.72 77.37 22.63
0 2 4 6 8 10 12 14 16
2.5 46.43 53.57 61.39 38.61
H/B (a)
50 12 5 33.31 66.69 47.43 52.57
Fcbαb =1, αs = 1.00
0.50 7.5 26.52 73.48 39.36 60.64
40 Fct 0.25
9 10 22.24 77.76 33.91 66.09
15 17.05 82.95 26.90 73.10
30

Fcb
6
Fct

20
the present analysis is much more rigorous and it does not make
3 any explicit assumption about the development of local plastic
10
m=1 zone for greater values of H / B.
0 0 It needs to be noted that for smaller values of H / B, the values
0 2 4 6 8 10 12 14 16 of both Fcb and Fct increase continuously with an increase in m,
H/B (b) on the other hand, for greater values of H / B, a decrease is noted
50 12 in the values of Fcb and Fct with an increase in m. From this
αb =1, αs = 1.00
0.50
Fcb
0.25 observation, one should not draw a wrong conclusion that the
40 Fct
9 uplift resistance decreases with an increase in the value of m. It
30
needs to be noted that both Fct and Fcb are defined with respect to
6 soil cohesion 共cb兲 at the pile base, however, the terms cbFcb and
Fcb
Fct

20 cbFct become always continuously higher for greater values of m

共not shown兲. It can also be seen that values of Fcb values remain
3
10 almost unaffected with the variations in the shaft adhesion 共␣s兲,
m=5 whereas, the values of Fct, as expected, increase continuously
0 0
with an increase in ␣s.
0 2 4 6 8 10 12 14 16

H/B
(c)
Percentage of Load Shared by the Pile Base and Shaft
Fig. 3. Variation of Fcb and Fct with H / B for: 共a兲 m = 0; 共b兲 m = 1; and The division of the total collapse load into that associated with the
共c兲 m = 5 shaft and base helps in understanding that which of the two com-
ponents provides a major contribution in finding the uplift capac-
ity. The percentage of the load shared by the pile base and the pile
shaft for different values of embedment ratio, corresponding to
been kept equal to 1. On the other hand, three different values of ␣s = 1 and ␣b = 1 and with m = 0 and 5, is presented in Table 1. It
shaft adhesion factor 共␣s兲, namely, 0.25, 0.5, and 1.00, were being can be observed that with an increase in the value of H / B, for all
used. The chosen range of ␣s practically covers various possibili- the values of m, the percentage contribution of the load due to the
ties for steel and concrete piles in clays 共Shin et al. 1993; Tian shaft resistance increases continuously. It can also be noted that
and Cassidy 2008兲. Based on the previous analysis of the writers the percentage of the resistance offered by the pile base becomes
共Khatri and Kumar 2009a兲, the number of sides of the yield poly- higher for greater value of m.
gon 共p兲 is kept equal to 24 共Bottero et al. 1980兲. The results from
the analysis are summarized herein. Failure Patterns
After determining the solution, the Mohr’s circle can be drawn at
Variation of Uplift Factors the centroid of each element to define the state of stress in a r-z
The variation of Fcb and Fct with H / B for m = 0, 1, and 5 is shown plane. Similar to the previous studies of the writers 共Khatri and
in Fig. 3. It can be noted that for a given value of m, the values of Kumar 2009a,b兲, the states of all elements, with respect to the
Fcb and Fct increase continuously with an increase in embedment shear failure, are defined in terms of a ratio, a / c; where a
ratio 共H / B兲. As compared to the Fct, after a certain embedment = radius of Mohr circle defining the actual state of stress, that is,
ratio, the increase in Fcb with H / B becomes only marginal. For a = 冑关共␴r − ␴z兲 / 2兴2 + 共␶rz兲2. For a point at yield, the radius of the
H / B = 15, the maximum value of Fcb from the present analysis for Mohr circle will become simply equal to c; the value of c is
m = 0 has been found to be 12.23 which is greater than nine as computed at the centroid of each element. If a / c = 1, the point will
recommended by Meyerhof 共1973兲. It should be mentioned that be in a state of shear failure. On the other hand, if a / c ⬍ 1 it will
the solution provided by Meyerhof 共1973兲 is based on a simple indicate that the point will remain in a nonplastic state. The fail-
limit equilibrium approach and it considers the development of ure patterns are sketched in such a manner that the color of the
only the local shear failure for the deep piles. On the other hand, element becomes continuously darker when approaching toward

26 / INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JANUARY/FEBRUARY 2011

Downloaded 13 May 2012 to 180.211.192.67. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
Fig. 4. Failure patterns for ␣s = 1, ␣b = 1, and m = 0 with: 共a兲 H / B
Fig. 5. Failure patterns for ␣s = 1, ␣b = 1, and m = 1 with: 共a兲 H / B
= 1.00; 共b兲 H / B = 5.00
= 1.00; 共b兲

the shear failure 共a / c = 1兲. For a pile with the rough shaft and
rough base 共␣s = 1 and ␣b = 1兲, Figs. 4 and 5 provide the failure
patterns for H / B = 1 and 5; these two figures are associated with equal to zero. In the presence of ␾ and nonzero m, however,
m = 0 and 1, respectively. In all the cases, a nonplastic zone adja- there still exists a scope of carrying out a further rigorous
cent to the pile shaft has been observed especially for greater computational work.
values of H / B. It can also be noticed that an increase in the value • The present study deals with the determination of the failure
of m leads to a reduction in the size of the plastic zone both in the load with the consideration of the yield failure of the soil
lateral and vertical downward direction. mass. No attempt has, however, been made to find the failure
load with respect to settlement consideration.

Remarks
Conclusions
• The computational results presented in this study can be used
for any given 共1兲 nonzero value of c0, 共2兲 the value of m The uplift capacity of an axially loaded pile resting on clayey
between 0 and 5, and 共3兲 the value of H / B lesser than 15. It strata, whose undrained cohesion increases linearly with depth,
has been assumed in the analysis that the soil mass remains in has been computed by using the static limit analysis approach in
contact with the perimeter of the pile surface including the combination with finite elements. The values of the uplift factors
base. In case if the separation occurs between the pile tip and Fcb and Fct are found to increase continuously with an increase in
adjoining soil mass, the pile base will not offer any resistance H / B. The rate of increase of Fcb with depth, however, reduces
to the uplift load. continuously with an increase in H / B. The values of Fcb at larger
• It should be mentioned the solution provided in this study embedment ratio are found to be greater than those reported in
cannot be proved to be a true lower bound since an assumption literature based on the simple limit equilibrium approach. For a
has been made with regard to the value of ␴␪. This is the given cohesion along the ground surface, the value of the uplift
reason that the methodology is referred to as the static limit resistance increases continuously with an increase in the value of
analysis rather than the lower bound limit analysis. m. The values of Fct increase further with an increase in the
• The analysis presented in this study is only for perfectly un- adhesion factor 共␣s兲 of the shaft-soil interface. The value of Fcb,
drained case. In reality, however, the value of ␾ may not be on the other hand, remains almost unaffected with the changes in

INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JANUARY/FEBRUARY 2011 / 27

Downloaded 13 May 2012 to 180.211.192.67. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
the values of ␣s. The computational results provided in this study Khatri, V. N., and Kumar, J. 共2009a兲. “Bearing capacity factor Nc under
will be useful for making an estimate of the vertical uplift resis- ␾ = 0 for piles in clays.” Int. J. Numer. Analyt. Meth. Geomech.,
tance of piles for various values of H / B and m with ␾ = 0. 33共9兲, 1203–1225.
Khatri, V. N., and Kumar, J. 共2009b兲. “Vertical uplift resistance of circu-
lar plate anchors in clay under undrained condition.” Comput. Geo-
tech., 36共8兲, 1352–1359.
References Kouzer, K. M., and Kumar, J. 共2009兲. “Vertical uplift capacity of equally
spaced horizontal strip anchors in sand.” Int. J. Geomech., 9共5兲, 230–
Ali, J. I. 共1968兲. “Pull-out resistance of anchor plates and anchored piles 236.
in soft bentonite clay.” MSc thesis, Duke Univ., Durham, N.C. Meyerhof, G. G. 共1973兲. “Uplift resistance of inclined anchors and piles.”
Azim, M. F. 共1985兲. “Ultimate uplift capacity of group piles in clay.” MS Proc. 8th Int. Conf. on Soil Mechanics and Foundation Engineering,
thesis, Univ. of Texas, E1 Paso, Tex. Moscow, 167–172.
Bhatnagar, R. S. 共1969兲. “Pull-out resistance of anchors in silty clay.” MS Shelke, A., and Patra, N. R. 共2008兲. “Effect of arching on uplift capacity
of pile groups in sand.” Int. J. Geomech., 8共6兲, 347–354.
thesis, Duke Univ., Durham, N.C.
Shin, E. C., Das, B. M., Puri, V. K., Yen, S. C., and Cook, E. E. 共1993兲.
Bishop, A. W. 共1966兲. “The strength of soils as engineering materials.”
“Ultimate uplift capacity of model rigid metal piles in clay.” Geotech.
Géotechnique, 16, 91–128.
Geologic. Eng., 11, 203–215.
Bottero, A., Negre, R., Pastor, J., and Turgeman, S. 共1980兲. “Finite ele-
Sloan, S. W. 共1988兲. “Lower bound limit analysis using finite elements
ment method and limit analysis theory for soil mechanics problem.” and linear programming.” Int. J. Numer. Analyt. Meth. Geomech., 12,
Comput. Methods Appl. Mech. Eng., 22共1兲, 131–149. 61–77.
Das, B. M. 共1989兲. “Uplift capacity of metal piles in clay.” Proc., 8th Int. Sowa, V. A. 共1970兲. “Pulling capacity of concrete cast in situ bored
Conf. on Offshore Mechanics and Arctic Engineering, Vol. 1, ASME, piles.” Can. Geotech. J., 7, 482–493.
New York, 519–524. Tian, Y., and Cassidy, M. J. 共2008兲. “Modelling of pile-soil interaction
Das, B. M., and Seeley, G. R. 共1982兲. “Uplift capacity of pipe piles in and its application in numerical simulation.” Int. J. Geomech., 8共4兲,
saturated clay.” Soils Found., 22共1兲, 91–94. 213–229.

28 / INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JANUARY/FEBRUARY 2011

Downloaded 13 May 2012 to 180.211.192.67. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org