Strain Wedge Model Capability of Analyzing Behavior

of Laterally Loaded Isolated Piles, Drilled Shafts,

and Pile Groups
Mohamed Ashour, M.ASCE1; Gary Norris, M.ASCE2; and Patrick Pilling, M.ASCE3
Downloaded from ascelibrary.org by Universite De Sherbrooke on 08/31/21. Copyright ASCE. For personal use only; all rights reserved.

Abstract: This paper demonstrates the application of the strain wedge 共SW兲 model to assess the response of laterally loaded isolated
long piles, drilled shafts, and pile groups in layered soil 共sand and/or clay兲 and rock deposits. The basic goal of this paper is to illustrate
the capabilities of the SW model versus other procedures and approaches. The SW model has been validated and verified through several
comparison studies with model- and full-scale lateral load tests. Several factors and features related to the problem of a laterally loaded
isolated pile and pile group are covered by the SW model. For example, the nonlinear behavior of both soil and pile material, the soil-pile
interaction 共i.e., the assessment of the p-y curves rather than the adoption of empirical ones兲, the potential of soil to liquefy, the
interference among neighboring piles in a pile group, and the pile cap contribution are considered in SW model analysis. The SW model
analyzes the response of laterally loaded piles based on pile properties 共pile stiffness, cross-sectional shape, pile-head conditions, etc.兲 as
well as soil properties. The SW model has the capability of assessing the response of a laterally loaded pile group in layered soil based
on more realistic assumptions of pile interference as compared to techniques and procedures currently employed or proposed.
DOI: 10.1061/共ASCE兲1084-0702共2002兲7:4共245兲
CE Database keywords: Lateral loads; Piles; Piles groups; Soil-pile interaction; Drilled shafts; Bridge foundations.

Introduction the predicted results tremendously. Researchers have attempted to

improve the performance of the p-y method by evaluating the
The problem of a laterally loaded single pile and pile group has p-y curve based on the results of the pressuremeter 共Smith 1983;
been under investigation and research for more than three de- Briaud et al. 1984兲 or dilatometer test 共Robertson et al. 1989兲.
cades. Many approaches, such as Broms’ method 共1964a, b兲, the The strain wedge 共SW兲 model analyzes the response of later-
elastic method 共Poulos 1971a,b兲, and the p-y curve approach ally loaded piles based on a representative soil-pile interaction
共Matlock 1970; Reese 1977; Murchison and O’Neill 1984; etc.兲, that incorporates pile and soil properties 共Ashour et al. 1998a兲.
have been employed in the analysis of laterally loaded pile re- The SW model allows the designer to predict the associated p-y
sponse. These methods consider some factors while neglecting curve at any point along the deflected part of the loaded pile. The
others. Therefore, limitations exist with respect to all these ap- effect of pile properties and the surrounding soil profile on the
proaches. As a result, designers have to switch from one method nature of the p-y curve has been presented by Ashour and Norris
to another to best satisfy their needs. 共2000a兲.
The p-y method, which was developed by Matlock 共1970兲 and p multipliers for piles in different rows have been suggested by
Reese 共1977兲, represents the most commonly used and convenient Brown et al. 共1988兲 for analysis of pile interference effects. Re-
procedure for the analysis of laterally loaded piles. The confi- cent tests 共McVay et al. 1995, 1998; Rollins et al. 1998兲 indicate
dence that designers have in this method derives from the fact that that such multipliers are a function of pile stiffness, pile spacing,
the p-y curves employed have been obtained 共back calculated兲 load or deflection level, and soil type. Therefore, the p-y method
from full-scale field tests, albeit only a very few tests. However, is presently insufficient for the task of accurately evaluating pile
the p-y method does not account for some significant factors such group response. In contrast, the SW model utilizes the mobilized
as various pile properties and soil continuity which could affect geometry of the passive wedge of soil in front of the pile 共hori-
zontally and vertically兲 to assess the interference of overlapping
1
Research Assistant Professor, Dept. of Civil Engineering, Univ. of shear zones among the piles in the group. Consequently, the con-
Nevada, Reno, NV 89557. E-mail: [email protected] tinually changing effect of pile group interference on the associ-
2
Professor, Dept. of Civil Engineering, Univ. of Nevada, Reno, NV ated p-y curves and the modulus of the subgrade reaction (E s )
89557. along every pile in the group can be evaluated.
3
Vice President, Black Eagle Consulting, Inc., 1345 Capital Blvd., The SW model has the capability of representing the un-
Suite A, Reno, NV 89502. drained resistance of liquefiable soil 共saturated sands兲 and the
Note. Discussion open until December 1, 2002. Separate discussions effect of the expected developing liquefaction on the behavior of
must be submitted for individual papers. To extend the closing date by
laterally loaded piles as illustrated by Ashour and Norris 共1998,
one month, a written request must be filed with the ASCE Managing
Editor. The manuscript for this paper was submitted for review and pos- 1999, 2002兲 and Norris et al. 共1997兲. The profession still lacks a
sible publication on June 19, 1998; approved on September 25, realistic procedure for the design of pile foundations in liquefying
2001. This paper is part of the Journal of Bridge Engineering, Vol. 7, or liquefied soil. The most common practice employed is that
No. 4, July 1, 2002. ©ASCE, ISSN 1084-0702/2002/4- presented by Wang and Reese 共1998兲 in which a traditional p-y
245–254/$8.00⫹$.50 per page. curve shape is used but based on the undrained residual strength

JOURNAL OF BRIDGE ENGINEERING / JULY/AUGUST 2002 / 245

J. Bridge Eng., 2002, 7(4): 245-254

Downloaded from ascelibrary.org by Universite De Sherbrooke on 08/31/21. Copyright ASCE. For personal use only; all rights reserved.

Fig. 1. Configuration of strain wedge model

(S r ) of the sand. S r can be related to the standard penetration test lateral loading 共Norris 1986兲. The main concept associated with
corrected blow count (N 1 ) 60 共Seed and Harder 1990兲. However, a the SW model is that traditional one-dimensional beam on elastic
very large difference between values at the upper and lower limits foundation 共BEF兲 pile response parameters can be characterized
at a particular (N 1 ) 60 value affects the assessment of S r tremen- in terms of three-dimensional soil-pile interaction behavior 共Fig.
dously. Even if an accurate value of S r is available, S r occurs at a 1兲. In the last several years, the SW model has been improved and
large value of soil strain and higher peak undrained resistance is modified through additional series of research to accommodate a
ignored in such clay-type modeling. This is extremely conserva- laterally loaded pile with different head conditions that is embed-
tive. Furthermore, the p-y curve reflects soil-pile interaction, not ded in multiple soil layers 共sand and clay兲 and rock 共Ashour et al.
just soil behavior. Therefore, the effect of soil liquefaction 共i.e., 1998a, 2001a兲. The main objective behind the development of the
degradation in soil resistance兲 does not reflect a one-to-one SW model is to evaluate the modulus of subgrade reaction 共i.e.,
change in soil-pile or p-y curve response. Instead, the soil’s und- the secant slope of the p-y curve兲 E s in order to solve the BEF
rained stress-strain relationship should be used in a true soil-pile problem of a laterally loaded pile and pile group based on the
interaction model to assess the corresponding p-y curve behavior. envisioned soil-pile interaction and its dependence on both soil
Because the traditional p-y curve is based on field data, a very and pile properties in addition to the effect of interference among
large number of field tests for different pile types in liquefying the piles in a group. Based on the basic properties from the soil
sand would be required to develop a realistic, empirically based, profile 共effective unit weight of soil ␥, effective angle of internal
p-y characterization. friction ␸, axial strain at 50% of stress level ␧ 50 , and undrained
Moreover, the nonlinear behavior of pile material 共steel and/or shear strength of clay S u ), the SW model basic procedures pre-
concrete兲 based on soil-pile interaction is also analyzed using the sented by Ashour et al. 共1998a兲 can be summarized as follows.
SW model 共Ashour et al. 2001b兲.
1. For a particular value of lateral strain 共␧兲 in the developing
passive wedge of soil in front of the pile, the increase in
Overview of Laterally Loaded Pile and Drilled Shaft horizontal stress (⌬␴ h ), the stress level 共SL兲, and the asso-
Behavior based on the Strain Wedge Model ciated Young’s modulus (E⫽⌬␴ h /␧) are determined based
on the stress-strain relationship of soil 共Ashour et al.
The SW model is an approach that has been developed to predict 1998a,b兲 as assessed from conventional triaxial testing.
the response of a free-head flexible pile in uniform soil under 2. The associated geometry of the passive wedge of soil 共mo-

246 / JOURNAL OF BRIDGE ENGINEERING / JULY/AUGUST 2002

J. Bridge Eng., 2002, 7(4): 245-254

 bilized fan angle ␸ m , base angle ␤ m , and width of wedge
face BC兲 is assessed according to an assumed initial value
共h兲 of the passive wedge depth 共Fig. 1兲 which is related to
the depth (X 0 ) of the zero deflection point (y⫽0). The soil
layers within depth h are divided into thin sublayers, and
steps 1 and 2 are applied to each sublayer 关Fig. 1共b兲兴.
3. The current variation of soil-pile line load 共p兲 along depth h
关Fig. 1共c兲兴 is obtained as a function of soil and pile param-
eters (⌬␴ h BC, D, and ␶兲 and the pile cross-section shape. D
is the pile width and ␶ is the mobilized shear resistance along
the pile sides 关Fig. 1共a兲兴.
4. Pile deflection 共y兲 along the depth of the passive wedge is
determined as a function of ␧, Poisson’s ratio, the SL, and
Downloaded from ascelibrary.org by Universite De Sherbrooke on 08/31/21. Copyright ASCE. For personal use only; all rights reserved.

the size of the passive wedge. As a result, the associated

profile of E s ⫽p/y can be predicted 关Fig. 1共c兲兴.
5. Based on the current profile of E s , the laterally loaded pile is
analyzed as a BEF under an arbitrary pile-head lateral load
( P 0 ). The values of pile-head deflection (Y 0 ) and X 0 共i.e., h兲
assessed using BEF analysis are compared to those of the
SW model analysis.
6. Through several iterative processes for the same value of soil
strain ␧, converged values of h and Y 0 are obtained. In ad-
dition, P 0 is modified as a function of the values of Y 0 from
both the BEF and SW model analyses 关( P 0 ) modified
⫽(Y 0 ) SW model( P 0 /Y 0 ) BEF].
7. For the next step of loading, a larger value for the horizontal
soil strain 共␧兲 is used, and steps 1 through 6 are repeated.

It should be noted that the deflection pattern 关Fig. 1共c兲兴 is

governed by pile-head fixity and pile bending stiffness (EI).
Consequently, the shape and size of the passive wedge of soil and
the associated E s are affected by these pile properties. Different
criteria for flow around failure, the stress-strain relationship of the Fig. 2. Comparison of results of strain wedge model, LPILE, and
soil, shear resistance along the pile sides, and the ultimate value field data for free- and fixed-head piles at Sabine River test 共Matlock
of soil-pile reaction ( p ult) are employed in the SW model analysis 1970兲
共Ashour et al. 1998a兲.
Soil response in the SW model is computed over the full
stress-strain (␴-␧) range of the soil 共down to 10⫺4 % strain兲 in-
stead of being projected from known empirical data 共often ob-
tained at larger strains兲 as in the COM624 and LPILE computer
programs 共Reese 1977; Reese and Sullivan 1980; Reese 1983兲.
This response exhibits very good agreement with the well known
Seed and Idriss 共1970兲 shear modulus reduction curve 共Ashour
and Norris 1999兲. The SW model approach has been employed in
different projects and field tests such as the retrofit of the San
Francisco Bay Bridge 共by California Department of Transporta-
tion, Caltrans兲, Cypress overpass 共Norris et al. 1993兲, and Oak-
land Outer Harbor Wharf 共Norris et al. 1996兲.
The SW model computer program has the efficiency to handle
the response of free- and fixed-head piles as seen in Fig. 2 共after
Matlock 1970兲. Also, it analyzes the behavior of laterally loaded
piles in soft, medium stiff, and stiff clay 共Matlock 1970; Reuss
et al. 1992兲 as shown by Figs. 2 and 3, respectively. The SW
model program is also capable of analyzing the behavior of long
drilled shafts although vertical side shear resistance due to lateral
deflection is not considered in the current version of the program.
In sand soils, the SW model approach assesses the behavior of
laterally loaded piles as seen with the Mustang Island test 共Fig. 4,
Reese et al. 1974兲 and the Arkansas River test 共Fig. 5, Alizadehh
and Davisson 1970兲. Fig. 5 features the capability of the SW
model to consider pile cross section shape 共H-shape and square Fig. 3. Measured pile-head response versus data obtained from strain
wedge model and LPILE for Pyramid Building, Memphis, test 共Reuss
piles as opposed to round兲. At the site of the Arkansas River test,
et al. 1992兲
the water table was approximately at 0.9 m below ground surface

JOURNAL OF BRIDGE ENGINEERING / JULY/AUGUST 2002 / 247

J. Bridge Eng., 2002, 7(4): 245-254

Downloaded from ascelibrary.org by Universite De Sherbrooke on 08/31/21. Copyright ASCE. For personal use only; all rights reserved.

Fig. 6. Effect of pile stiffness on p-y curve of free-head pile in sand

at 1.22 m depth

1988兲, and FLPIER 共University of Florida 1996兲, are a function

of soil properties and only pile width. The problem of a laterally
loaded pile is often solved as BEF involving nonlinear modeling
of the soil-pile interaction response ( p-y curve兲. Currently em-
ployed p-y curve models were established/verified based on the
results of field tests in uniform soils such as the Mustang Island
共Reese et al. 1974兲, Sabine River 共Matlock 1970兲, and Houston
共Reese and Welch 1975兲 tests, and adjusted mathematically using
empirical parameters to extrapolate beyond the soil’s specific field
test conditions.
Fig. 4. Measured and predicted response of laterally loaded pile in
The traditional p-y curve models developed by Matlock
sand at Mustang Island test 共Reese et al. 1974兲
共1970兲 and Reese et al. 共1974兲 are semiemipirical models in
which soil response is characterized as independent nonlinear
and the capillary tension influence above the water table, without springs 共Winkler springs兲 at discrete locations. Therefore, the ef-
doubt, must be considered in the analysis as seen in Fig. 5. fect of a change in soil type of one layer on the response 共p-y
curve兲 of another is not specifically considered. In addition, the
formulations for these p-y curve models do not account for a
Soil-Pile Modeling „p - y curve… change in pile properties such as pile bending stiffness, pile cross-
sectional shape, pile-head fixity, and pile-head embedment below
Matlock-Reese p-y curves, which are employed in the computer the ground surface. Soil-pile interaction or p-y curve behavior is
programs COM624/GROUP 共Reese 1977; Reese and Sullivan not unique but a function of both soil and pile properties 共Ashour
1980; Reese and Wang 1996兲, LPILE1 共Reese 1987兲, PAR 共PMB and Norris 2000a兲. It would be prohibitively expensive to system-
atically evaluate all such effects through additional field tests;
hence it behooves us to consider such influences based on avail-
able theoretical means 共SW model formulation兲 that allow trans-
formation of the envisioned three-dimensional soil-pile interac-
tion response to one-dimensional BEF parameters. As Terzaghi
共1955兲 and Vesic 共1961兲 stated, the subgrade modulus E s 共and,
therefore, the p-y curve兲 is not just a soil but, rather, a soil-pile
interaction 共and, therefore, a pile property dependent兲 response.
Unlike the Matlock-Reese p-y curves, a soil-pile interaction
p-y curve should be affected by pile bending stiffness 共EI兲 as
seen in Fig. 6. Here, the stiffer pile yields the stiffer associated
p-y curve. The two piles analyzed by using the SW model in Fig.
6 are a steel pipe pile 共stiff pile兲 and a timber pile 共flexible pile兲.
Both piles have the same shape, size, and head conditions 共free-
head pile兲.
The significant influence of pile cross sectional shape on the
nature of the p-y curve is seen in Fig. 7. The two piles employed
Fig. 5. Comparison of measured and predicted response of Pile 6 of
共square pile and circular pile兲 are assumed to have the same
Arkansas River test 共Alizadeh and Davisson 1970兲
width, bending stiffness, and pile-head conditions, and are driven

248 / JOURNAL OF BRIDGE ENGINEERING / JULY/AUGUST 2002

J. Bridge Eng., 2002, 7(4): 245-254

Downloaded from ascelibrary.org by Universite De Sherbrooke on 08/31/21. Copyright ASCE. For personal use only; all rights reserved.

Fig. 7. Effect of pile cross-section shape on p-y curve of free-head Fig. 9. Pile-head response at Treasure Island test site under develop-
pile in sand at 1.22 m depth ing liquefaction 共CISS, 0.324 m兲

in the same soil. More details on the nonuniqueness of the p-y equivalent uniform cycles produced over the full duration of the
curve are presented by Ashour and Norris 共2000a兲. earthquake. Thereafter, the lateral load 共from the superstructure兲
is applied at the pile head, which generates additional porewater
pressure (u xs,nf) in the soil immediately around the pile, given the
Undrained Response of PilesÕShafts in Liquefiable degradation in soil strength already caused by u xs,ff . Thereafter,
Soil „Saturated Sands… the undrained behavior due to an inertial induced lateral load is
assessed using undrained stress-strain formulation in the SW
Due to the shaking from an earthquake and the associated lateral model 共Ashour and Norris 2002兲. Thus, the procedure accounts
load from the superstructure, excess porewater pressure in the for both u xs,ff and u xs,nf . It should be noted that these procedures
free and near fields develops and reduces the strength of loose to incorporate the whole undrained stress-strain curve 共at any level
medium sand around a pile. The degradation in soil resistance and of loading兲, not only the residual strength of the sand. The SW
the induced excess porewater pressure in the free field (u xs,ff) is model analysis characterizes the reduction in E s 共i.e., the p-y
based on the procedures proposed by Seed et al. 共1983兲. This is curve兲 and pile response due to a drop in sand strength and
followed by the assessment of the excess porewater pressure Young’s modulus as a result of developing liquefaction in the
(u x,nf) in the near-field soil region 共adjacent to the pile, Fig. 8兲 sand.
induced by the lateral load from the superstructure. The variation The SW model analysis of laterally loaded piles in liquefiable
in soil resistance 共undrained stress-strain relationship兲 around the soil utilizes the basic properties of the sand such as its relative
pile in the near-field zone is evaluated based on the undrained density (D r ), particle shape 共roundness兲, and percentage of fines
formulation for saturated sand presented by Ashour and Norris in addition to the characteristics of the earthquake 共such as maxi-
共1999兲. mum ground acceleration a max and earthquake magnitude兲. The
The assessed value of the free-field excess porewater pressure undrained stress-strain relationship, the effective stress path, and
ratio r u induced by the earthquake is obtained using Seed’s the potential of sand to liquefy or dilate can be predicted.
method 共Seed et al. 1983兲. u xs,ff is calculated conservatively at the The full-scale load tests on the postliquefaction lateral re-
end of earthquake shaking corresponding to the number of sponse of piles that were performed at Treasure Island 共Ashford

Fig. 8. Schematic figure of excess porewater pressure zones around Fig. 10. Pile-head response at Treasure Island test site under devel-
laterally loaded pile oping liquefaction 共H shape, 0.310 m兲

JOURNAL OF BRIDGE ENGINEERING / JULY/AUGUST 2002 / 249

J. Bridge Eng., 2002, 7(4): 245-254

Downloaded from ascelibrary.org by Universite De Sherbrooke on 08/31/21. Copyright ASCE. For personal use only; all rights reserved.

Fig. 11. Pile-head response in banding sand under developing

liquefaction (u xs,ff⫹uxs,hf) Fig. 13. Modified p-y curve on individual pile in pile group

and Rollins 1999兲 are the most significant related tests. Figs. 9 共Brown et al. 1988兲. Consequently, a full-scale field test 共which is
and 10 show a comparison between the Treasure Island test re- costly兲 is strongly recommended in order to determine the value
sults 共liquefied soil兲 and the data predicted using the SW model of f m of the soil profile at the site under consideration. Moreover,
analysis. This methodology also covers the behavior of piles in a the suggested value of the multiplier ( f m ) is taken to be constant
group. More details are presented by Ashour and Norris 共1998, for each soil layer at all levels of loading 共Fig. 13兲.
2000b, 2002兲. In reality, the p multiplier should vary in accord with the pile
Figs. 11 and 12 show the effect of different values of maxi- and soil properties, the level of load or deflection 共and thus the
mum ground surface acceleration a max on the pile-head response current size or depth of the wedges兲, and the location of the pile
and the associated p-y curves of a laterally loaded pile in satu- in the pile group. The value of this multiplier is very difficult to
rated medium loose Banding sand of D r ⫽37% 共Ashour and Nor- evaluate from experiment and much more research is clearly
ris 1999, 2002兲. Varying values of (u xs,ff⫹u xs,nf) develop at dif- needed to establish a methodology for empirically predicting the
ferent levels of a max . variation of f m . The interference among the piles in a group
decreases with depth 共Fig. 14兲 thus generating lower values of f m
near the ground surface 共or pile head兲 and greater ones at deeper
Strain Wedge Analysis of Pile Group points. Consequently, the values of f m increase with depth below
the pile head in the same soil and it will be more complex in
At present there is no particular technique that assesses the re- layered soils. The p multipliers should be developed based on site
sponse of piles in a group using completely reasonable assump- specific field tests.
tions. The most common design procedure is to reduce the stiff- As presented by Ashour et al. 共2001c兲, the SW model ap-
ness of the traditional 共Matlock-Reese兲 p-y curve by using a proach calculates the ever changing geometry of the mobilized
multiplier ( f m ⬍1), as seen in Fig. 13. The value of the p-y curve passive wedge of soil in front of the pile. The shape of the passive
multiplier ( f m ) should be assumed and is based on the data col- wedge of soil 共Fig. 1兲 varies according to the associated level of
lected from full-scale field tests on pile groups, which are few

Fig. 12. Drained and undrained 共liquefaction兲 p-y curves at 3 m Fig. 14. Interference among piles in group based on strain wedge
depth in banding sand model

250 / JOURNAL OF BRIDGE ENGINEERING / JULY/AUGUST 2002

J. Bridge Eng., 2002, 7(4): 245-254

Downloaded from ascelibrary.org by Universite De Sherbrooke on 08/31/21. Copyright ASCE. For personal use only; all rights reserved.

Fig. 15. Interference among piles in group at given depth

loading and becomes deeper and wider 共larger fanning angle兲

with increasing lateral load and deflection. This characterization
allows the SW model to evaluate the interference among neigh-
boring piles in the pile group at different depths and increasing
pile-head deflection.
The changing value of E s as determined by the SW model
analysis will account for the additional strains 共i.e., stresses兲 in
the adjacent soil due to pile interference within the group 共Figs.
14 and 15兲. Thus the E s 共i.e., the secant slope of the p-y curve兲 of
an individual pile in a group will be reduced in a mobilized fash-
ion according to pile and soil properties, pile spacing and posi-
tion, the level of loading, and depth x 共Fig. 16兲. No single reduc- Fig. 17. Lateral response of 3⫻3 free-head pile group in layered
tion factor ( f m ) for the p-y curve 共commonly assumed to be a soil 共sand and clay兲 共Brown and Reese 1985兲
constant value with depth and level of loading兲 is needed or ad-
vised.
As seen in Fig. 15, the soil around the piles in the group parison with observed data 共Brown and Reese 1985; McVay et al.
interferes horizontally with that of adjacent piles by an amount 1995兲 for a pile group in layered clay soil and uniform sand,
that varies with depth. Therefore, the varying overlap of the respectively.
wedges of neighboring piles in different sublayers over the depth
of the interference and the associated increase in soil stress/strain
can be determined as a function of the amount of overlap 共Fig. Material Modeling of Piles and Drilled Shafts
14兲. As pile lateral load increases, the wedges grow deeper and
Modeling the actual 共i.e., nonlinear兲 behavior of pile material in a
fan out horizontally, thus causing a further change in overlap and
SW model analysis of piles enables the designer to evaluate the
group interference, all of which vary with a change in soil and
pile properties. Such analysis is incorporated in the SWM-3.2
computer program 共Ashour et al. 1998b兲.
Figs. 17 and 18 show the capability of the technique in com-

Fig. 16. E s variations for pile in group and isolated pile in strain Fig. 18. 3⫻3 free-head pile group model (3D and 5D spacing兲 in
wedge model analysis medium to loose sand 共McVay et al. 1995兲

JOURNAL OF BRIDGE ENGINEERING / JULY/AUGUST 2002 / 251

J. Bridge Eng., 2002, 7(4): 245-254

Downloaded from ascelibrary.org by Universite De Sherbrooke on 08/31/21. Copyright ASCE. For personal use only; all rights reserved.

Fig. 21. Comparison of measured and calculated pile-head deflec-

Fig. 19. Behavior of laterally loaded pile divided into segments
tions for Houston test 共Reese and Welch 1975兲

ultimate lateral resistance of piles and shafts in a realistic fashion (M -␸) relationship of the pile/shaft cross section. The ultimate
based upon soil-pile interaction concepts 共Fig. 19兲. The problem bending moment (M ultimate) carried by the pile cross section is
of material nonlinearity in the analysis of laterally loaded piles obtained and employed to locate the point of failure in the loaded
and drilled shaft behavior was originally considered by Reese pile. An equivalent pile or shaft with the same dimensions as the
共1984兲. Because the Matlock-Reese p-y curves are not a function original pile and a reduced bending stiffness (EI) r , constant over
of the bending stiffness 共EI兲 of the pile, a simplified procedure the entire length, is employed in the COM624 program analysis
was suggested by which the EI of the studied pile or drilled shaft 共Reese 1977; Reese and Sullivan 1980兲. These procedures have
is reduced by a specific value based on the moment-curvature been modified by Reese and Wang 共1996兲 to use the cracked EI of
the reinforced concrete pile/shaft for the upper deflected portion
of the pile/shaft. The current practice of the p-y method
共COM624, LPILE, and FLPIER兲 accounts for the continuous
variation of EI on the pile/shaft resistance but with no corre-
sponding consideration of such an effect on the associated p-y
curve 共the traditional p-y curve is a function only of the pile
width兲.
Contrary to the p-y method, the variation in the value of EI
has a significant effect on the nature of the p-y curve and the
modulus of the subgrade reaction 共Terzaghi 1955; Vesic 1961;
Ashour and Norris 2000a兲 especially in the case of drilled shafts.
In the same fashion, the pile response computed by COM624 at
any level of loading represents the behavior of a fictitious pile of
reduced stiffness (EI) r over its entire length 共Fig. 20兲. However,
the comparison to COM624, shown in Fig. 20, does not reflect the
most recent practice of COM624/LPILE.

Fig. 20. Response of laterally loaded free-head drilled shaft based on Fig. 22. Effect of nonlinear behavior of drilled shaft material (R/C)
strain wedge material modeling and Reese 共1984兲 on nature of p-y curve

252 / JOURNAL OF BRIDGE ENGINEERING / JULY/AUGUST 2002

J. Bridge Eng., 2002, 7(4): 245-254

 and near fields due to the shake of an earthquake and imposition
of superstructure load. Also, the nonlinear behavior of pile/shaft
material can be accounted for in the SW model to predict the
ultimate capacity and the reduction in lateral resistance of piles/
shafts and the associated p-y curves.

Acknowledgments

Financial support from the California Department of Transporta-

tion 共Caltrans兲 for the work presented is greatly acknowledged.
The writers would like to thank Dr. Saad Azazy, Mr. Anoosh
Shamsabadi, Mr. Steeve McBride, Mr. Thomas Shantz, Dr. Abbas
Downloaded from ascelibrary.org by Universite De Sherbrooke on 08/31/21. Copyright ASCE. For personal use only; all rights reserved.

Abghari, Mr. Angel Perez-Cobo, and Mr. Ken Jackura for their
interest and encouragement.

Fig. 23. Measured and computed results for Gill test Pile 9 共Gill References
1969兲
Alizadeh, M., and Davisson, M. T. 共1970兲. ‘‘Lateral load test on piles—
Arkansas river project.’’ J. Soil Mech. Found. Div., 1583–1604.
Ashford, S. A., and Rollins, K. 共1999兲. ‘‘Full-scale behavior of laterally
In reality, the laterally loaded pile continues to deflect based loaded deep foundations in liquefied sands.’’ Rep. No. TR-99/03,
on its initial bending stiffness 共EI兲, i.e., it exhibits linear behavior, Structural Engineering Dept., Univ. of California, San Diego.
until some part of the pile cross section starts to behave nonlin- Ashour, M., and Norris, G. 共1998兲. ‘‘Undrained laterally loaded pile re-
early, maintaining a reduced value of bending stiffness. The re- sponse in sand.’’ Proc., Geotechnical Earthquake Engineering and
duction of EI along the deflected portion of the pile is reflective of Soil Dynamics Conf., Special Publication, Paper No. 63 ASCE, New
a combination of pile and soil properties and the M -␸ relation- York.
Ashour, M., and Norris, G., 共1999兲. ‘‘Liquefaction and undrained re-
ship at any level of loading, which is considered in SW model
sponse evaluation of sands from drained formulation.’’ J. Geotech.
analysis 共Ashour et al. 2001b兲. A stress-strain model for confined Geoenviron. Eng., 125共8兲, 649– 658.
concrete 共Mander et al. 1988; Paulay and Priestly 1992兲 and a Ashour, M., and Norris, G. M., 共2000a兲. ‘‘Modeling lateral soil-pile re-
bilinear stress-strain model for steel are employed in the analysis. sponse based on soil-pile interaction.’’ J. Geotech. Geoenviron. Eng.,
These procedures are implemented in the consideration of the 126共5兲, 420-428.
behavior of the laterally loaded pile group to evaluate the pile- Ashour, M., and Norris, G. M. 共2000b兲. ‘‘Undrained lateral pile and pile
head ductility and the sequence of failure of the piles in the group. group response in saturated sand.’’ Rep. No. CCEER-00-01, Civil En-
Fig. 21 provides a comparison between observed and predicted gineering Dept., Univ. of Nevada, Reno, Nev.
responses. Note the improved SW model capability provided by Ashour, M., and Norris, G. M. 共2002兲. ‘‘Lateral load pile response in
realistic nonlinear material modeling. As seen in Fig. 22, such liquefied soil.’’ J. Geotech. Geoenviron. Eng., in press.
Ashour, M., Norris, G., and Pilling, P. 共1998a兲. ‘‘Lateral loading of a pile
nonlinear material behavior causes a change in p-y response, par-
in layered soil using the strain wedge model.’’ J. Geotech. Geoenvi-
ticularly in the zone of maximum moment. As expected, the dif- ron. Eng., 124共4兲, 303–315.
ferences between the linear and nonlinear model responses of Ashour, M., Norris, G. M., Bowman, S., Beeston, H., Billing, P., and
Figs. 21 and 22 are small at the lower levels of load/deflection. Shamsabadi, A. 共2001a兲. ‘‘Modeling pile lateral response in weathered
Fig. 23 presents another comparison for the response of a laterally rock.’’ Proc., 36th Engineering Geology and Geotechnical Engineer-
loaded a steel pipe pile from a field test 共Gill 1969兲 using the SW ing Symposium, Univ. of Nevada, Las Vegas.
model nonlinear analysis and COM624. Ashour, M., Norris, G. M., and Shamsabadi, A. 共2001b兲. ‘‘Effect of the
non-linear behavior of pile material on the response of laterally loaded
piles.’’ Proc., 4th Int. Conf. on Recent Advances in Geotechnical
Earthquake Engineering and Soil Dynamics, Univ. of Missouri–
Conclusions
Rolla, Mo., Paper 6.10.
Ashour, M., Pilling, P., and Norris, G. 共1998b兲. ‘‘Updated documentation
The SW model provides a realistic demonstration of soil-pile in- of the strain wedge model program for analyzing laterally loaded piles
teraction by which the effect of such pile properties as pile size, and pile groups.’’ Proc., 33rd Engineering Geology and Geotechnical
pile cross-sectional shape, bending stiffness, pile-head conditions, Engineering Symposium, Univ. of Nevada, Reno, 177–178.
and pile interference effects are taken into account in the devel- Ashour, M., Pilling, P., and Norris, G. M. 共2001c兲. ‘‘Asssessment of pile
opment of the p-y curves. Such p-y curves are a product of, not group response under lateral load.’’ Proc., 4th Int. Conf. on Recent
the input to, the SW model. The SW model has been contrasted Advances in Geotechnical Earthquake Engineering and Soil Dynam-
with the BEF analysis using Matlock-Reese type p-y curves. The ics, Univ. of Missouri–Rolla, Mo., Paper 6.11.
SW model approach provides a three-dimensional analysis of a Briaud, J. L., Smith, T., and Mayer, B. 共1984兲. ‘‘Laterally loaded piles
laterally loaded pile group linked to the BEF analysis based on and the pressuremeter: Comparison of existing methods.’’ Laterally
loaded deep foundations, ASTM STP 835, West Conshohocken, Pa.,
both soil and pile properties 共soil-pile interaction兲 and pile spac-
97–111.
ing. Therefore, unlike the current procedures, no reduction factors Broms, Bengt B. 共1964a兲. ‘‘Lateral resistance of piles in cohesive soils.’’
or multipliers are required to reflect the interference among the J. Soil Mech. Found. Div., Am. Soc. Civ. Eng., 90共SM2兲, 26 – 63.
piles in the group. The SW model assesses the response of later- Broms, Bengt B. 共1964b兲. ‘‘Lateral resistance of piles in cohesionless
ally loaded piles in liquefiable soils 共postliquefaction兲 with the soils.’’ J. Soil Mech. Found. Div., Am. Soc. Civ. Eng., 90共SM3兲, 123–
incorporation of the induced excess porewater pressure in the free 156.

JOURNAL OF BRIDGE ENGINEERING / JULY/AUGUST 2002 / 253

J. Bridge Eng., 2002, 7(4): 245-254

 Brown, D. A., Morrison, C., and Reese, L. C. 共1988兲. ‘‘Lateral load under lateral load.’’ Rep., Geotechnical Engineering Center, Bureau of
behavior of pile load in sand.’’ J. Geotech. Eng., 114共11兲, 1261–1276. Engineering Research, Univ. of Texas at Austin, Tex.
Brown, D. A., and Reese, L. C. 共1985兲. ‘‘Behavior of a large-scale pile Reese, L. C. 共1987兲. ‘‘Documentation of the computer program LPILE1.’’
group subjected to cyclic lateral loading.’’ Geotech. Eng. Rep. GR85- Ensoft Inc., P. O. Box 180348, Austin, Tex.
12, Geotech. Eng. Center, Bureau of Eng. Research, Austin, Tex. Reese, L. C., Cox, W. R., and Koop, F. D. 共1974兲. ‘‘Analysis of laterally
Gill, H. L. 共1969兲. ‘‘Soil pile interaction under lateral loading.’’ Proc., loaded piles in sand.’’ Proc., 6th Annual Offshore Technology Conf.,
Conf. on In-Situ Testing of Soils and Rocks, London, 141–147. Dallas, Paper No. OTC-2080, 473– 483.
Mander, J. B., Briestley, M. J. N., and Park, R., 共1988兲. ‘‘Theoretical Reese, L. C., and Sullivan, W. R. 共1980兲. ‘‘Documentation of computer
stress-strain model for confined concrete.’’ J. Struct. Eng., 114共8兲, program COM624, parts I and II: analysis of stresses and deflections
1804 –1826. for laterally loaded piles including generation of p-y curves.’’ Geo-
Matlock, H. 共1970兲. ‘‘Correlations for design of laterally loaded piles in technical engineering software GS80-1, Geotechnical Engineering
soft clay.’’ Proc., 2nd Annual Offshore Technology Conf., Dallas, Center, Bureau of Engineering Research, Univ. of Texas at Austin,
Paper No. OTC-1204, 577– 607. Tex.
McVay, M., Casper, R., and Shang, T., 共1995兲. ‘‘Lateral response of three- Reese, L., and Wang, S. T. 共1996兲. ‘‘Documentation of the computer
Downloaded from ascelibrary.org by Universite De Sherbrooke on 08/31/21. Copyright ASCE. For personal use only; all rights reserved.

row groups in loose to dense sands at 3D and 5D pile spacing.’’ J. program GROUP.’’ Ensoft Inc., P. O. Box 180348, Austin, Tex.
Geotech. Eng., 121共5兲, 436 – 441. Reese, L. C., and Welch, R. C. 共1975兲. ‘‘Lateral loading of deep founda-
McVay, M., Zhang, L., Molnit, T., and Lai, P. 共1998兲. ‘‘Centrifuge testing tions in stiff clay.’’ J. Geotech. Eng. Div., Am. Soc. Civ. Eng., 101共7兲,
of large laterally loaded pile groups in sand.’’ J. Geotech. Eng. 633– 649.
124共10兲, 1016 –1026.
Reuss, R., Wang, S. T., and Reese, L. C. 共1992兲. ‘‘Tests of piles under
Murchison, J. M., and O’Neil, M. W. 共1984兲. ‘‘Evaluation of p-y rela-
lateral loading at the Pyramid Building, Memphis, Tennessee.’’ Geo-
tionships in cohesionless soils.’’ Analysis and design of pile founda-
tech. News, December, 44 – 46.
tions, J. R. Meyer, ed., ASCE, New York, 174 –191.
Robertson, P. K., Davies, M. P., and Campanella, R. G. 共1989兲. ‘‘Design
Norris, G. M. 共1986兲. ‘‘Theoretically based BEF laterally loaded pile
of laterally loaded driven piles using the flat dilatometer.’’ Geotech.
analysis.’’ Proc., 3rd Int. Conf. on Numerical Methods in Offshore
Testing J., ASTM, 12共1兲, 30–39.
Piling, TECHNIP Ed., Paris, 361–386.
Rollins, K. M., Peterson, K. T., and Weaver, T. J. 共1998兲. ‘‘Lateral load
Norris, G. M., Siddharthan, R., Zafir, Z., and Gowda, P. 共1993兲. ‘‘Soil and
behavior of full-scale pile group in clay.’’ J. Geotech. Geoenviron.
foundation conditions and ground motions at Cypress.’’ Transp. Res.
Eng., 124共6兲, 468 – 478.
Rec., 1411, 61– 69.
Norris, G. M., Siddharthan, R., Zafir, Z., Abdel Ghaffar, S., and Gowda, Seed, R. B., and Harder, L. F. 共1990兲. ‘‘SPT-based analysis of cyclic pore
P. 共1996兲. ‘‘Soil-foundation-structure behavior at the Oakland outer pressure generation and undrained residual strength.’’ H. Bolton Seed
harbor wharf.’’ Transp. Res. Rec., 1546, 100–111. Memorial Symposium Proc., BiTech, Vancouver, B.C., Vol. 2.
Norris, G., Siddharthan, R., Zafir, Z., and Madhu, R. 共1997兲. ‘‘Liquefac- Seed, H. B., and Idriss, I. M. 共1970兲. ‘‘Soil moduli and damping factors
tion and residual strength of sands from drained triaxial tests.’’ J. for dynamic response analyses.’’ Rep. No. EERC 70-10, College of
Geotech. Eng., 123共3兲, 220–228. Engineering, Univ. of California, Berkeley, Calif.
Paulay, T., and Priestley, M. J. N. 共1992兲. ‘‘Seismic design of reinforced Seed, H. B., Idriss, I. M., and Arango, I. 共1983兲. ‘‘Evaluation of liquefac-
concrete and masonry.’’ Wiley, New York, 95–157. tion potential using field performance data.’’ J. Geotech. Eng., 109共3兲,
PMB Engineering Inc. 共1988兲. Pile analysis routine (PAR), San Fran- 458 – 482.
cisco. Smith, T. D. 共1983兲. ‘‘Pressuremeter design method for single piles sub-
Poulos, H. G. 共1971a兲. ‘‘Behavior of laterally loaded piles. I: Single jected to static lateral load.’’ PhD dissertation, Texas A&M Univ.,
piles.’’ J. Soil Mech. Found. Div., Am. Soc. Civ. Eng. 97共5兲, 711–731. College Station, Tex.
Poulos, H. G. 共1971b兲. ‘‘Behavior of laterally loaded piles. II: Pile Terzaghi, K. 共1955兲. ‘‘Evaluation of coefficients of subgrade reaction.’’
groups.’’ J. Soil Mech. Found. Div., Am. Soc. Civ. Eng., 97共5兲, 733– Geotechnique, 5共4兲, 297–326.
751. Univ. of Florida. 共1996兲. User manual for FLORIDA-PIER program,
Reese, L. C. 共1977兲. ‘‘Laterally loaded Piles: Program documentation.’’ J. Dept. of Civil Engineering, Univ. of Florida, Gainesville, Fla.
Geotech. Eng. Div., Am. Soc. Civ. Eng., 103共4兲, 287–305. Vesic, A. 共1961兲. ‘‘Bending of beams resting on isotropic elastic solid.’’ J.
Reese, L. C. 共1983兲. ‘‘Behavior of piles and pile groups under lateral Eng. Mech. Div., Am. Soc. Civ. Eng., 87共2兲, 35–53.
load.’’ Rep. to the U.S. Dept. of Transportation, Federal Highway Wang, S. T., and Reese, L. C. 共1998兲. ‘‘Design of pile foundations in
Administration, Office of Research, Development, and Technology, liquefied soils.’’ ASCE, Proc., Geotechnical Earthquake Engineering
Washington, D.C. and Soil Dynamics Conf., Publication No. 75, ASCE, New York, Vol.
Reese, L. C. 共1984兲. ‘‘Handbook on design of piles and drilled shafts II, 1331–1343.

254 / JOURNAL OF BRIDGE ENGINEERING / JULY/AUGUST 2002

J. Bridge Eng., 2002, 7(4): 245-254