EFFECTS OF CYCLIC LATERAL LOADS

ON PILES IN S A N D

By J. H. Long, t Member, ASCE, and Geert Vannestez

Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

ABSTRACT: The effect of repetitive lateral loads on deflections of two drilled piers
in Tampa Bay were significantly greater than predicted by a p-y procedure com-
monly used in practice. Reasons for the discrepancy between predicted and mea-
sured deflections are discussed. Two methods for predicting the effect of repetitive
lateral loads are developed using results of 34 cyclic lateral load tests to quantify
model parameters important to the behavior of piles subjected to repetitive lateral
loading. The two methods model cycliclateral load behavior of a pile by degrading
soil resistance as a function of number of cycles of load, method of pile installation,
soil density, and character of cyclic load. The two methods differ in the compu-
tational effort required to make the prediction. The first method is most suitable
for hand calculation and rule-of-thumb estimation and is based upon a beam-on-
an-elastic foundation model with a soil reaction modulus, Kh, increasing propor-
tionally with depth. The second method modifies nonlinear static p-y curves to
derive a cyclicp-y curve. The two methods provide a simple means for estimating
effects of cyclic lateral load.

INTRODUCTION

The effects of cyclic lateral loads on piles in sand are important to quantify
due to the occurrence of cyclic lateral loads in nature. W i n d , waves, earth
pressures, and water pressures, m a y subject cyclic lateral loads to pile sup-
ported structures. Methods c o m m o n l y used for predicting the response of
piles to cyclic lateral loads were evaluated for two piers in T a m p a Bay,
Florida, and found to predict poorly the behavior due to cyclic loading.
Using information from these tests and additional case histories, p a r a m e t e r s
that influence the behavior of piles subjected to cyclic lateral loads are
identified, and two methods for predicting load deflection behavior are
proposed.
On September 12, 1982, static and repetitive lateral load tests were con-
ducted on two offshore piers in T a m p a Bay adjacent to the Sunshine Skyway
Bridge. Predictions of the behavior of the piers were m a d e using a m e t h o d
proposed by Reese and his coworkers (Reese et al. 1974). This m e t h o d was
selected because it is based u p o n results of full-scale lateral load tests on
piles in sand and the m e t h o d provides reasonable agreement with results
from other uninstrumented load tests ( M e y e r and R e e s e 1979).
The two piers at T a m p a Bay experienced deflections due to cyclic loading
greater than predicted using a p-y m e t h o d outlined by Reese, et al. (1974).
Illustrated in Figs. 1 (a) and l ( b ) are m e a s u r e d and predicted load-deflection
(H-g) relationships for the first cycle of loading and for 50 cycles of loading.
Several possible explanations for differences between predicted and mea-
sured deflection were investigated. H o w e v e r , differences in construction,
soil properties, and loading characteristics for the tests conducted at T a m p a

1Assoc. Prof. of Civ. Engrg., Newmark Scholar, Univ. of Illinois, Urbana, IL

61801.
2Engr., lngelmunster, Belgium.
Note. Discussion open until June 1, 1994. To extend the closing date one month,
a written request must be filed with the ASCE Manager of Journals. The manuscript
for this paper was submitted for review and possible publication on March 5, 1990.
This paper is part of the Journal of Geotechnical Engineering, Vol. 120, No. 1,
January, 1994. 9 ISSN 0733-9410/94/0001-0225/$1.00 + $.15 per page. Paper
No. 27047.

225

J. Geotech. Engrg. 1994.120:225-244.

 Lateral Deflection of Shaft, m m
0 50 100 150 200
500 u I i i i i
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

East Shaft I .f" oeOe 9 00

400
9 Measured ~ / _ qN~

300

Z 200
Note: Delta ploffed for
100 1, 2. 5. 10, 2 0 ,
to a•d 40 cyclH

o n I J J I a n n n I I I I , I
..J 0-
to
k- 500
iD
.4.- West S h a f t /,, ~ 9
0 -- ~,,.d) I //
._1 400

300

200

[ NOte: Dola plofled for

100 ~ P I 1. z. s, lo. =o,
[ o.d ,0 oyo,.. ,,,
0
0 50 1 O0 150 200

Lateral Deflection of Shaft, m m

FIG. 1. Comparison of Measured and Predicted Response for Two Laterally Loaded
Piers at Tampa Bay, Fla.

Bay and Mustang Island (Reese et aL 1974) are too numerous to be quan-
tified with results from only these two sites; therefore, results of 34 repetitive
lateral load tests in sand were investigated. Two simple methods are pro-
posed herein to allow prediction of effects of cyclic lateral load on piles.
The two methods include effects due to the characteristics of cyclic load,
the number of cycles, the installation method, and soil density. The two
methods provide a means to estimate effects of cyclic lateral loads on piles
in sand.

B E H A V I O R O F PILES L O A D E D R E P E T I T I V E L Y

Qualitative Discussion of Cyclic Behavior

The behavior of a vertical pile subjected to repetitive lateral loads depends
upon characteristics of the lateral load, geometrical and structural properties
of the pile, the properties of the soil in which the pile is embedded, and
the change in soil properties as the pile is loaded repetitively.
A simplified soil-pile model is used to illustrate the behavior of a soil and
226

J. Geotech. Engrg. 1994.120:225-244.

 pile to the application of a cyclic, lateral load at the head of a pile, In this
model, the pile is assumed to be elastic, infinitely long, vertical, and embed-
ded in a cohesionless soil. A two-way lateral load applied at the head of
the pile varies sinusoidally with time at a frequency low enough that only
effects of repeated load are important (the effects of inertia are minimal)
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

and any excess porewater pressures generated in the soil are dissipated
quickly. Furthermore, the cohesionless soil is assumed to exhibit no signif-
icant cohesion or creep. The behavior of the pile is described for four phases
during one cycle of load. The effect of each quarter-cycle of load is described
and the influence of further cycles is mentioned.
During the first quarter-cycle, the magnitude of lateral load varies from
a value of zero to a maximum horizontal load Hmax in a direction to the
right. The head of the pile rotates and translates to the right in response to
the applied load. Resistance to pile deflection is provided by the soil along
the right side of the pile while the soil along the left side of the pile maintains
contact by flowing with the pile. The soil surrounding the pile may change
in volume depending on its initial density and state of stress (Chang and
Whitman 1988).
During the second quarter-cycle, the lateral load decreases from a value
of Hmaxto zero, and the head of the pile deflects toward its original position.
As the pile translates to the left, the soil resistance along the left side of
the pile increases while the soil resistance on the right side decreases. If the
soil pressure along the right side decreases to an active state, the cohesionless
soil will flow and prevent a gap, thus ensuring contact with the pile surface.
As with the first quarter-cycle of load, the cohesionless soil may change
volume depending on its density and change in state of stress.
The direction of the lateral load, Hmax, and the corresponding deflection
of the pile head are reversed for the third quarter-cycle. The magnitude of
horizontal load changes from zero to -H~ax causing the pile to deflect to
the left. The pile may resist significant lateral loads before reaching the
original location of the pile head because of the presence of cohesionless
sand that flowed with the back of the pile during the first quarter-cycle of
load. As the load approaches - H m a • the pile deflects to the left while the
soil maintains contact along the right side of the pile preventing any gap
between the pile and soil.
The response of pile and soil during the fourth quarter-cycle is similar,
but opposite in direction, to the response described during the second quarter-
cycle. Depending on the density of the soil and the stress state in the soil,
volume changes in the soil may occur.
Effect of further cycles on the maximum horizontal deflection of the pile
and on bending moments within the pile depend upon changes in mechanical
properties (strength, modulus) and the accumulation of permanent strains
in the soil.
Characteristics of the load can influence significantly the behavior of piles
subjected to cyclic lateral loads. If cyclic pile displacement is primarily in
one direction, then effects of cumulative deformations are more pro-
nounced. For example, one-way cyclic loading (load varies from 0 to Hma x
to 0 with no reversal of load direction) will induce more permanent strains
and greater cumulative deformations than piles subjected to two-way cyclic
loads (load varies from/-/max to - H m a x to Hma x to - H m a x ) .
Structural details of the pile may also play a role in the behavior of a pile
when subjected to cyclic lateral loads. As the pile is loaded, moments gen-
erated within the pile cause the pile to bend and mobilize tension and
227

J. Geotech. Engrg. 1994.120:225-244.

 compression within the pile cross section. Steel cross sections usually remain
elastic for magnitudes of maximum bending moment anticipated in design;
therefore, the flexural stiffness of the steel cross section remains unchanged
for all magnitudes and cycles of load. However, the flexural stiffness of
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

reinforced concrete sections decrease with increasing moment and load cycles
due to progressive cracking of the cross section (Little and Briaud 1988).
Changes in flexural stiffness are most pronounced for reinforced pile cross
sections that are neither post- nor pretensioned because tensile stresses are
greatest and crack formation is more prevalent. Since the deflection of a
pile is influenced by changes in flexural stiffness, the influence of these
changes on pile performance should be a design consideration. Often, effects
of degradation in flexural stiffness are insignificant because only a small
portion of the pile experiences large bending moments. Therefore, only a
small portion of the pile experiences significant degradation in flexural stiff-
ness (Long and Reese 1982; Kramer and Heavey 1988), and the deflection
of the pile at working loads is affected minimally.

History for Predicting Effects of Cyclic Lateral Loads

Attempts by others have been made to quantify the effect of cyclic lateral
loads on pile behavior by modeling the soil-pile system with a beam-on-an-
elastic-foundation (BOEF) analysis. A summary of previous attempts pro-
vides a perspective on parameters considered to be important and methods
employed for modeling effects of cyclic loading.
Reese and Matloek (1956) and Vesic (1977) suggest that solutions using
a linear, elastic soil response with a soil reaction modulus, Kh, increasing
proportionally with depth provides a reasonable model for determining the
lateral behavior of piles in sands. The coefficient of soil reaction, nh, is used
to identify the increase in soil reaction modulus, Kh, with depth as shown
here:
Kh = nh'Z ................................................. (1)
where z = depth below the ground surface.
Effects of cyclic loads are modeled by reducing the static value of n h.
Summarizing results from Prakash (1961), Davisson and Salley (1970), and
Alizadeh and Davisson (1970), and Davisson (1970) suggests reducing nh
to 30% of the static value if the load is cycled 50 times or more.
Broms (1964) suggests that deterioration of n h depends on density of the
cohesionless soil and recommends that for 40 cycles of load, the value of
nh should be reduced to one-fourth and one-half the original value of r/h
for low and high relative densities, respectively. Broms cautions that these
recommendations are based upon limited data.
Reese et al. (1974) used results of static and cyclic load tests on instru-
mented, full-scale piles to develop a semi-empirical, nonlinear p-y (soil
resistance-pile deflection) approach. Procedures developed to predict cyclic
p-y relationships are based upon degraded staticp-y curves, with degradation
factors determined empirically from results of the instrumented load tests.
The cyclic p-y curves proposed by Reese et al. (1974) were developed to
represent the resistance provided by the soil at a large number of cycles of
load. Therefore, recommendations for cyclic p-y curves are independent of
the number of cycles.
O'Neill and Murchison (1983) used a number of case histories of static
and some cyclic load tests to evaluate current procedures and develop im-
proved procedures for generating p-y curves. Although the main emphasis
228

J. Geotech. Engrg. 1994.120:225-244.

 was for static lateral loading, cyclic p - y curves are generated by reducing
the static soil resistance, p, for a given deflection, y. Like the method
proposed by Reese et al. (1974), the cyclic p - y curves are independent of
the number of cycles.
Little and Briaud (1988) model the deterioration of the soil reaction
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

modulus, Kh, due to cyclic loading as

KhN = K m " N - ~ . ............................................ (2)
where KhN = Kh at the Nth cycle of load, Khl = the value of the soil
reaction modulus for the first cycle of load, and a = a degradation param-
eter.
Little and Briaud use results from static pressuremeter tests to obtain
static, nonlinear p - y relationships. Cyclic p - y curves are constructed by
reducing the static soil resistance according to (3)
PN = pl"N -a 9 .............................................. (3)
where P N ~" the cyclic soil resistance for N cycles of load, pl -- the value
ofp for the first cycle of load, and a = a degradation parameter determined
from the results of cyclic pressurerneter tests. Little and Briaud found rea-
sonable agreement between values of a from full-scale tests on piles sub-
jected to one-way loading and from cyclic pressuremeter tests. Two advan-
tages of Little and Briaud's method are the ability to model the effect of
number of cycles, and the ability to determine parameters from in situ test
results.
Theoretical studies using a more sophisticated approach for modeling soil
behavior have been conducted by Swane and Poulos (1982), and Matlock
et al. (1978). These authors model each p - y curve for the complete load
history for every load cycle and solve for the response using a discrete
element model for the BOEF problem. While the two approaches are fun-
damental, the parameters required for analysis are usually unavailable for
typical soil exploration and site characterization studies.
Turner et al. (1987) collected results of 23 case histories on piles and
drilled shafts subjected to cyclic lateral loads. They suggest the cyclic be-
havior of piles is influenced by the tendency of the soil to dilate or contract
during cyclic loading.
The uncertainty associated with the aforementioned methods for pre-
dicting effects of cyclic lateral load remains unquantified as do the major
factors that affect cyclic behavior. Results of several full-scale load tests
were investigated to determine the effect of cyclic lateral loads on piles in
sand, and to assess uncertainties associated with the prediction method
proposed.

CASE HISTORIES ON CYCLICALLY, LATERALLY LOADED PILES

Results from 34 full-scale, cyclic lateral load tests in sand are used to
illustrate and quantify the influence of important parameters on pile be-
havior. Details of the cyclic lateral load tests are presented in Table 1. The
collection includes a wide range of soil density, pile type and material,
construction technique, and cyclic load characteristics. Load tests were con-
ducted by applying a specific load repetitively to the pile head, and reporting
the increase in deflection of the pile with number of cycles of load. Only
the load test by Morrison (1986) was conducted by controlling displacements
and recording the degradation of lateral load with cycles.
229

J. Geotech. Engrg. 1994.120:225-244.

 Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

TABLE 1. Summary of Cyclic Lateral Load Tests

Width Maximum Cyclic load
Pile diameter Method of number of ratio
Test type (ram) installation cycles (nmin/nmax) Reference
(1) (2) (3) (5) (7) (8) (9)
Timber pile 305 Driven 24 0 Alizadeh (1968) 1A
Timber pile 310 Driven 24 0 Alizadeh (1968) 1B
/-/-pile 145 Driven 100 0 Alizadeh and Davisson (1970)
6
Concrete pile 510 Sonic Vib 100 0 Alizadeh and Davisson (1970)
I1A
5 Drilled shaft 610 Drilled 10 0 Bhushan et al. (1981) 4
6 Drilled shaft 915 Drilled 10, 15 0 Bhushan et al. (1981) 5
7 Drilled shaft 915 Drilled 15 0 Bhushan et ai. (1981) 6
O 8 Drilled shaft 1,220 Drilled 15 0 Bhushan et al. (1981) 7
9 Drilled shaft 1,220 Backfilled 100 0 Davisson and Salley (1968) IN
10 Drilled shaft 1,430 Backfilled 80 0 Davisson and Salley (1968) 2N
11 Drilled shaft 1,220 Drilled 100 0 Davisson and Salley (1968) IS
12 Drilled shaft 1,430 Drilled 80 0 Davisson and Salley 1968) 2S
13 Specially fabricated 405 BackfiUed and com- 500 0 Heller (1964)

J. Geotech. Engrg. 1994.120:225-244.

wide flange beam pacted
14 Specially fabricated 405 Backfilled and com- 500 0 Heller (1964)
wide flange beam I pacted
15 Specially fabricated I 405 Backfilled and com- 500 0 Heller (1964)
wide flange beam I pacted
16 Drilled shaft I 915 Drilled 21 0, 0.5 Little and Briaud (1988) 1
17 Pipe pile I 610 Driven 21 0,0.5 Little and Briaud (1988) 2
18 Concrete pile I 510 Driven 21 0, 0.5 Little and Briaud (1988) 3
19 Drilled shaft 1,065 Drilled 21 0, 0.5 Little and Briaud (1988) 4
20 Drilled shaft 1,065 Drilled 21 0, 0.5 Little and Briaud (1988) 5
 Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

21 Drilled shaft 1,065 39.0 Drilled Loose 21 0, 0.5 Little and Briaud (1988) 6
22 Drilled shaft with 1,220 15.5 Vibrated Dense 40, 125 0 Long and Reese (1984) East
casing
23 Drilled shaft with 1,220 15.5 Vibrated Dense 40, 125 0 Long and Reese (1984) West
casing
24 H-pile 355 17.1 Driven Dense 25 0 Meyer and Reese (1979)
Bailley 1
25 /-/-pile 355 17.1 Driven Dense 25 0 Meyer and Reese (1979)
Bailley 2
26 Pipe pile 275 13.4 Backfilled and com- Mediu 100 -1.0 Morrison (1986)
pacted
27 /-/-pile 355 15.9 Driven Dense 23 - 1.0 O'Neill and Murchison (1983),
T3
28 Pipe pile 610 21.0 Driven Dense tOO - 0.25 Reese et at. (1974)
29 Timber pile 290 11.3 Driven Loose 5 0 Robinson (1979) 6
30 Timber pile 305 15.2 Driven Dense 5 0 Robinson (1979) 8
31 Timber pile 305 5.2 Driven Loose 5 0 Robinson (1979) 9
32 Timber pile 330 10.7 Driven Dense 23 0.1 Stevens et al. (1979)
33 /-/-pile 355 20.4 Driven MediuJ 25 0 Tucker and Briaud (1988) 1
H-pile 355 20.4 Driven Mediu__..~ 25 0 Tucker and Briaud (1988) 2

J. Geotech. Engrg. 1994.120:225-244.

 The response of a pile to cyclic lateral load depends on the change in
stiffness of the soil and pile with applied cycles of load. Changes in soil and
pile stiffness are influenced by the magnitude of load, the character of load
(one- versus two-way loading), the initial stiffness of the pile and soil, and
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

the change in stiffness of the pile and soil with number of cycles. The soil
is assumed to be insensitive to creep, and the pile material is assumed to
remain elastic.

Parameter for Quantifying Effect of Cyclic Lateral Load

Davisson (1970), Broms (1964), and Vesic (1977) have suggested that the
lateral response of a pile could be modeled simply as a BOEF with a linearly
increasing soil reaction modulus, nh. Furthermore, they suggested the effect
of cyclic loading could be modeled by reducing the static value of nh. The
model suggested by Davisson, Broms, and Vesic is simple and used often
for preliminary assessment of lateral load behavior. Therefore, the effect
of cyclic lateral loads for all 34 load tests was investigated using a BOEF
approach and a soil reaction nh. Parameters for this soil-structure interaction
model are the flexural stiffness of the pile (El), the length of the pile, L,
the coefficient of soil reaction (nh), and the magnitude of load (or moment).
The pile properties and the magnitude of maximum lateral load are assumed
to remain unchanged during cyclic loading; therefore, the increase in pile
deflection is affected only by a change in coefficient of soil reaction, nh.
The effects of cyclic lateral loads are studied by back-calculating values
of nh from measured load-deflection data, and comparing nh for the first
cycle of load with nh for subsequent cycles in terms of the ratio, Rn (Rn =
HhN/nhl). The parameter R. is selected because most cyclic lateral load tests
present only load-deflection data, deflection data are usually more reliable
than moment data for load tests conducted in which both moment and
deflection are measured, values of n h are more sensitive to changes in de-
flection than moment, and results of both instrumented and uninstrumented
pile tests can be compared with a common parameter (nh).

Effect of Number of Cycles of Load

Cyclic lateral loads applied to the head of a pile result in deformations
that may increase with every cycle. The increase in pile-head deformation
with continued cyclic loading may be modeled by reducing the coefficient
of soil reaction, nh. The effect of cyclic loading on the reduction of nh is
shown in Fig. 2 by plotting Rn versus the number of cycles, N, on a log-log
scale. A straight line relationship is expressed as

Rn- nhN_ N-' . ........................................... (4)

~hl
where - t = the slope of the straight line. The magnitude of t represents a
relative effect of cyclic loading on the deterioration of nh.
Results from full-scale tests have shown that one-way cyclic loading may
continue to influence peak deflections up to 500 cycles (HeUer 1964); how-
ever, the quantity of available data is limited. Most of the load-test data
summarized are for 50 cycles or less (Table 1). The effect of cyclic loading
is greatest for the first cycle of loading, with the effect of subsequent cycles
diminishing as cycling continues.
232

J. Geotech. Engrg. 1994.120:225-244.

 2 , i

L
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

c
1 r ~ A t = 0.02 -
RH = - 1
d
0 RH = -0.25
e~
0.5
"-I RH = 0.5
m
-I0
0
Legend
9 D a ~ * y t = 0.33 1 RH = 0 . 0
0
0.2 V Long and Reese (1984)
(3
(.3
9 Little and griaud (1988) 1
r] Reese, et, al. (1974)
9 klorrison (1986)
A O'Neill and Mui'chlson (1983)
0.1 i ~ i i i i iJ j ..... I

2 5 10 20 50 100 200

Number of C y c l e s , N

FIG. 2. Effect of Number of Cycles on Cyclic Modulus Ratio

Effect of Character of Cyclic Load

A cyclic load ratio, R~/, is used to quantify the character of cyclic load
and is defined as

,% -
•/min . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (5)
//max
where Hmin = the magnitude of minimum lateral load, and Hm,x = the
magnitude of maximum lateral load. A pile cycled from 0 (Hmi,) to Hma,
(one-way loading) is calculated to have a cyclic load ratio, Rn = 0. A pile
cycled with equal load magnitude in both directions (two-way loading) has
a ratio, Ru = - 1 , and a pile loaded statically would have a value of
R , = 1. Most cyclic lateral load tests reported in the literature and con-
ducted in practice use RH = 0 (one-way cyclic loading); however, some
notable lateral load tests have also been conducted using R/~ -- - 1 [two-
way cyclic loading (Morrison 1986)], R, = - 0 . 2 5 (Reese et al. 1974) and
Ru = 0.5 (Little and BriaNd 1988). The soil modulus ratio, R,,, versus
number of cycles is shown in Fig. 2 for different magnitudes of the cyclic
load ratio, RH. The corresponding degradation parameter, t, represents the
slope of a straight line fit and can be calculated as:
log(R,u)
t = log(N) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (6)
where Rnu = the value of R, at N cycles of load. Larger values of t cor-
respond to a greater effect of cyclic loading. Values of t shown in Fig. 2 are
0.03, 0.08, 0.27, and 0.13 for RH of - 1 , -0.25, 0, and 0.5, respectively.
Results of these full-scale tests show that one-way cyclic loading results in
greater degradation than two-way loading, and that intermediate cyclic load
ratios result in degradations between one- and two-way loading. Although
no comprehensive full-scale field tests have been conducted to confirm or
quantify the effect of RH specifically, results of model-scale pile tests suggest
233

J. Geotech. Engrg. 1994.120:225-244.

 one-way cyclic loading results in larger deflections than two-way cyclic load-
ing (Parry and Sanglamer 1977; Barton 1982). Thus, the degradation pa-
rameter, t, is assumed to be greater for one-way loading and less for two-
way loading. Finally, the degradation parameter, t, is 0 for R , = 1 since
this loading condition corresponds to a static load.
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

Effect of Magnitude of Deflection and Precycling

Results from the 34 case histories were used to investigate effects of
magnitude of deflection on the deterioration of soil modulus with number
of cycles, t. In Fig. 3 are shown values of t back-calculated from results of
each cyclic load test versus normalized deflection, ~/D (pile head deflection/
pile head diameter) for piles which had no previous significant cyclic loads
(virgin cycling). The effect of cyclic lateral loads on t appears to be relatively
unaffected by the magnitude of normalized deflection. Fig. 4 shows t values
for piles subjected previously to cyclic loads, at a lower magnitude of load.
The values of t exhibit considerable scatter. The value of t for virgin cycling

0.600
:, I
C
a) 0.400
E
P
o
O.
C
.o 0.200 ',, j : p ..........
"o
E .... ....... i , ,...i,i,,',, ",~' ........ ; . ....... ......~ ' o ; " " .i.................
ol
r-I
0.000
0.0002 0.001 0.01 0.1 0,2
Normalized Deflection, 8/D
l Nizodeh and Dc~luon (:1970) 46 Zl Meyer ond Roen (1979) #I
Nizod.ahand .Dgvfaeon(1970) 811A 9 Meyerand Rees.e(1979) #2
~izaoeh (1958) 1A 9 Nob'leon(1988)
Allzodeh (1968) 1B . 0 O'Nenland Murchiegn
Bhuehon,at el. (1981) #4 Rpaee. at 91. (1974).
Bhu=hon,et ale (1981) J5 =,1[evens,~ ~I. (1979)
Bhushon, et al. (1981) J6 .~- luckar one urlaud (1988) #I
Bhuehan,at ol. (1981} J7 Tucker and Brlo~d (1988) #2
DovLelonand Solley(1"96B)#IN <> Robinson(1979) #6
uavisson and Salley(1968) #2N 9 Robinson(1979) #8
uavlssonand Salley(1968) #1S ~> Robinson(1979) 49
uavia=onand Salley(1968) #25
i Heller (1964-) #I
#2
Hailer {1964-~43
Little and Bnoud (19B8) #1
Little and Bdoud (1988) #2
Uttle end Bdoud (1988) #3
Uttle end Brloud (1988) #4.
Little end Brloud (1988) r
Little and Bdoud.(1988) #6
Long and Reeee(1984.~-Eost
Long and Reese(1984) -West

FIG. 3. Effect of Normalized Deflection on Degradation Parameter, t, for Virgin

Cycling

234

J. Geotech. Engrg. 1994.120:225-244.

 ,4-/
0.600 9 . .i , t i

I
L ..................................... PRE-CYCLED I
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

E
P 0.400 ........ ............ i--~.i-.i..i 9 ................. ......... ~...... 9 ........... ....... ..---...-.i.! ...........
0
Q.
r ! i!!i 9 ioi ~ ........
0
~4-,-
0
"10
P 0.200
......... ........ j . . : ! ! :.............
01
rt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

O.O00
0.0002 0.001 0.01 0.1 0.2
Normalized Deflection, 4/D

l Alizadeh and Dov~eeon(1970) 86 A Meyer and Reeel (1979) 41

Allzodeh end Day[seen (1970) 811A 9 Meyer and Reeee (lg79) 82
Nizadeh (1968) 1A 9 Mondeon (1986)
Alizndeh (1968) 1B O O'Neill end Murchilan
Bhulhan, et el. (1981) 84 "<P Reeee, et ol. (lg74).
Bhushan, et eL (1981) 85 Stwens, etal. (197.9)
Bhushan, et el. (1981) 86 Tur and Briaud (1988) 81
Bhushon~et el. (1981) 87 Tucker and Briaud (1988) 82
Davisian and Solley (1968) 81N ~> Robineon (1979) #6
~r Davis=on and Solley (1968) 82N 9 Robinson (1979) 88
&L Dovissonand Solley (1988) 81S ~> Robinson (1979) 89
Davieeon end Salley (1968) 42S
~ Hener (1964) 81
Hailer (19w 82
Heller (196~) 83
Little and Bnoud (19B8) 01
Little and Brloud (1988) 02
~ Little end Briaud (1988) 83
Little and Briaud (1988) Q4
Little and Brioud (1988) #5
Little end Brloud [1988) # 6
Long and Reel= (1984~ -East
Long and Reese (1984) -West

FIG. 4. Effect of N o r m a l i z e d D e f l e c t i o n o n D e g r a d a t i o n P a r a m e t e r , t, for Precycled

Piles

ranges between 0.52 and 0.00 with an average near 0.22. The values of t
for precycled piles are smaller with a range between 0.0 and 0.35 and an
average of 0.18.

DEVELOPMENT OF MODELS FOR CYCLIC LOADING EFFECTS

Both methods developed in this paper employ results of 34 cyclic lateral

load tests and use the same parameters for determining the effect of cyclic
loading. The two methods differ in the computational effort required to
solve the soil-structure interaction problem to obtain deflection. The first
method is based upon a closed-form solution for a B O E F with a linearly
increasing soil reaction modulus (LISM) that varies proportionally with
depth. The LISM method is simple to apply and uses parameters obtained
directly from the results of the cyclic lateral load tests.
The second method is based upon the degradation of static p-y (DSPY)
235

J. Geotech. Engrg. 1994.120:225-244.

 curves. A p-y curve for static loading is modified to a cyclic p-y curve by
reducing the static soil resistance, p, while increasing the static displacement,
y. The DSPY method can accommodate nonlinear p-y curves and different
procedures for generating static p-y curves; however, the solution for the
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

soil-structure interaction equations requires the use of a computer.

Linearly Increasing Soil Modulus (LISM)

The LISM method employs a BOEF analysis. The soil reaction modulus
is assumed to increase linearly with depth according to (1). Hetenyi (1946)
developed solutions for deflection of the pile head in the form of infinite
series, but the solution can be simplified by calculating separately the con-
tribution to deflection by lateral load, H, and bending moment, M, as

A .H B.M
ON -- it?10.4.~0.6 "~- l t 2 / 0 . 6 . . 0 . 4 .................................. (7)
JL~t S ~ hN ~t~Jt r ~ hN

where gN = the lateral deflection at the Nth cycle of load, E1 = the flexural
rigidity (product of pile modulus and the moment of inertia), nhN the =

coefficient of soil reaction at N cycles of load. A and B = constants deter-

mined from the length of the pile, L, and the relative stiffness ([EI/nhn] ~
according to beam-on-elastic-foundation theory (Hetenyi 1946). All the var-
iables are taken as known except the coefficient of subgrade reaction for
cyclic loading, nhN, which is calculated as

nhS = nhl"N-' .............................................. (8)

where nhl = the coefficient of subgrade reaction for the first cycle of load.
All 34 load tests were evaluated to determine appropriate values of t.
Figs. 5 and 6 illustrate the effect of load, installation, and soil density for
piles subjected to cyclic loads for the first time (virgin cycling), and for piles
previously subjected to cyclic loads at a lower level of load, respectively.
Frequency histograms are superimposed on the figure when necessary. The
effect of load ratio is shown in Figs. 5(a) and 6(a). Figs. 5(b) and 6(b)
illustrate the effect of pile installation for virgin cycling and for precycled
piles. Figs. 5(c) and 6(c) show effects of soil density. Figs. 5(b), 5(c), 6(a),
and 6(c) use only results from tests conducted with a cyclic load ratio greater
than zero. Based upon trends in Figs. 5(a)-5(c), and 6(a)-6(c), the follow-
ing observations are made.

Virgin Cycling (Fig. 5)

Piles installed by driving or by vibration exhibit the least effect of cyclic
loading (smallest values of t), while drilled shafts and shafts backfilled with-
out compaction exhibit the greatest effect of cyclic loading and the greatest
variation in t. Installation methods listed in order of increasing average t
values are as follows: vibrated, backfilled and compacted, driven, sonic
vibrated, drilled, and backfilled. Foundations in dense soil exhibit less effect
of cyclic loading than foundations in loose soil.

Tests that Have Been Cycled Previously

The average value of t is smaller than values of t for virgin cyclic loading,
and t values are much less sensitive to differences in soil density and in-
stallation procedure.
236

J. Geotech. Engrg. 1994.120:225-244.

 L.

I I
) CtfcIlc Load l 9
fects ~"25 0025
f--.
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

g~ f
.~ 0.2
,~ 0.1
~00 *t ~ 1 7 6

-1.0
I

-0.5
0.0 0.5
If'
I I

1.0
Cyclic Load Ratio, R H

0.6 , , , ,
~E I Ib) Installation
0.sj " 0"2, E,,.ots

e 0, i--.--

9~ o.o-- , , , , ; ;

9"=- -- .- o

g o ~

L.
i
9 0.6--
T T------~c ) So|I De nsity'l
E 0.5 ~ - - o ~ s -" [ Effects I
no 0.4
0.3 - ~ ~
.T. 0.2
0.1
0.0

FIG. 5. Value of Degradation Parameter, t, for Virgin Cycling as Affected by: (a)
Cyclic Load Ratio; (b) Installation; and (c) Soil Density

For Both Precycled and Virgin Cycfic Loading

The cyclic load ratio appears to affect the cyclic behavior most signifi-
cantly. The least effect of cyclic loading is seen for RH = --1 (two-way
loading) while the greatest effect of cyclic loading is seen when R , = 0
(one-way loading).

Method for Determining Degradation Parameter, t

A specific value of t for a cyclically laterally loaded pile depends on
characteristics of the cyclic load ratio, RH, the installation method, the soil
density, and if the pile has been precycled. The value of the degradation
parameter, t, that includes effects of load direction, installation procedure,
and soil density can be calculated as
237

J. Geotech. Engrg. 1994.120:225-244.

 ~a

0.6
iE 0.5
a) Cyclic L o a d ]
Effects J-
'
o~0 o~o
'
o~o
o
0.4
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

o.
0.3
o
0.2
o
"o
o 0.1
o~
J
o 0.0 i
0
-1.0 -0.5 0.0 0.5 1.0
Cyclic Load Ratio, R H

0.6,
@ ' ' ' ]b) Installation J
E 0.5 Effects 1
0.4 0 ",0 o ", ,0
c 0.3
o
=o
"o
0.2

P 0.1
o~
@ 0.0
o

o "~o J~

=o

C
o
0.6 1 1 1-[c) Soll Denslt
0.5 ...... ~ ......................... - t ~ Effects
o
o 3o o 3o
0.4 .......... .:~ .............. -~- ................
o.
0.5
i
..........................................
c
o

"o
P 0.1~ _T2-
0.0
._~

FIG. 6. Value of Degradation Parameter, t, for Precycled Piles as Affected by: (a)
Cyclic Load Ratio; (b) Installation; and (c) Soil Density

t = 0.17.FL.Fz'Fo .......................................... (9)

where FL, F~, and Fo, are factors based upon details of the cyclic load ratio,
pile installation, and soil density, respectively. Suggested values for FL, Fz,
and Fo, based on results of the 34 load tests, are given in Tables 2-4. Values
for FL in Table 2 reflect the effect of cyclic load with the greatest value for
R , = 0 and 0.5, decreasing nonlinearly to 0.2 for R , = - 1 , and to 0.0
for Rn = 1 (static load). Linear interpolation can be used to determine FL
at load ratios between specified points. The effect of cyclic load ratio appears
to be the most important factor.
A value of 0.24 for t can be calculated using (9) for a drilled shaft (F~ =
1.3) subjected to one-way cyclic loading (FL = 1.0) in loose soil (Fo =
1.1). For comparison, recommendations by Davisson (1970) suggest t equals
238

J. Geotech. Engrg. 1994.120:225-244.

 TABLE 2. Effect of Cyclic Load Ratio on Parameter Fr
Load ratio RH FL
(1) (2)
- 1.0 (two-way loading) 0.2
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

- 0.25 0.4
0.0 (one-way loading) 1.0
0.5 1.0
1.0 (static loading) 0.0

TABLE 3. Effect of Installation on Parameter F1

Method of installation FI
(1) (2)
Driven 1.0
Vibrated 0.9
Backfilled 1.4
Backfilled and compacted 1.0
Drilled 1.3
Precycled (regardless of installation) 1.0

TABLE 4. Effect of Soil Density on Parameter FD

Soil density F,~
(1) (2)
Loose (contractive) 1.1
Medium 1.0
Dense 0.8
Precycled (regardless of density) 1.0

0.31, while Broms recommends 0.38 and 0.19 for loose and dense soils,
respectively.
The uncertainty for estimating t was determined by comparing the values
of t measured from the 34 lateral load tests with values of t predicted using
(9). Fifty percent of the measured values of t exceeded values of t predicted
using (9). If the predicted value of t is multiplied by 1.4, only 16% of the
measured values exceed those predicted. Similarly, if the predicted value
of t is multiplied by 2, a mere 3% of the measured t values exceed the
predicted value.
The LISM method provides a simple procedure for predicting the effect
of cyclic lateral loading; however, the method is restricted to analyses that
employ a linearly increasing soil reaction modulus. Additionally, the LISM
method cannot explicitly account for effects of nonlinear soil response,
layered soil, and many fundamental parameters (soil unit weight, soil strength,
and so forth) that affect lateral load response. Because p-y methods can
include effects of nonlinearity, soil layering, and other soil properties, rec-
ommendations for using results from the 34 load tests with nonlinear p-y
curve analyses are provided next.

Deterioration of Static p - y Curve (DSPY)

The use of nonlinear p-y curves to represent the static soil resistance
provided by piles is a common approach for analyzing the response of
239

J. Geotech. Engrg. 1994.120:225-244.

 laterally loaded piles. Recommendations for developing p - y curves for both
static and cyclic loading have been given by (Reese et al. 1974; Briaud 1988;
O'Neill and Murchison 1983). The DSPY method presented herein modifies
a static p - y curve to a p - y curve for cyclic loading.
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

The DSPY approach is similar to the LISM approach; namely, the static
soil reaction modulus is reduced to account for effects of cyclic loading
according to (10)
KhN = K h , ' N - ' . ........................................... (10)

The degradation parameter, t, is determined using (9). Since staticp-y curves

are nonlinear, additional recommendations for the reduction in soil resis-
tance, p, must be provided, otherwise, any combination of decrease in p
and increase in y could be used to satisfy (10).
Little and Briaud (1988) suggest constructing the nonlinear cyclic p - y
curve by maintaining deflection, y , constant for both static and cyclic p - y
curves while reducing the soil resistance, Per = Pl" N-t. Thus, the reduction
in soil reaction modulus prescribed in (10) is satisfied. Values of t were
back-calculated for the 34 case histories using nonlinear, p - y curves modified
by reducing soil resistance. Values of t for the Little and Briaud method
were significantly lower in magnitude than values of t backcalculated using
the LISM method.
Requirements for reducing the cyclic soil reaction modulus according to
(10) can also be satisfied by modifying the soil deflection according to the
relationship, YN = Yl" N ~. This approach gives the cyclic p - y curve a larger
deflection, y, while maintaining the same soil resistance, p, for both static
and cyclic load. Values of t back-calculated for the 34 load tests using this
method were much greater than t values back-calculated using the LISM
method.
The aforementioned methods for modifying nonlinear p - y curves should
yield t values different from those obtained with the LISM method, because
the load is held constant during a cyclic lateral load test while the pile head
progressively increases in deformation. Thus, the decrease in soil reaction
modulus is caused by both a reduction in the soil resistance, p, and an
increase in pile deflection, y.
The DSPY method specifies a change in both soil resistance and deflection
by including a term, et, into (10) as shown here:

PN = P l "N(~'-I)' ............................................ (11)

YN = Yl "N~'t ............................................... (12)
where PN = the soil resistance for N cycles of load and p~ = the soil
resistance for the first cycle of load.
The factor, o~, controls the relative contribution of soil resistance and
deflection to decrease the soil reaction modulus. The value of eLvaries from
0 (to produce a change in p only) to 1 (to produce a change in y only). For
et values between 0 and 1, both p and y are changed. A value of 0.6 for
can be derived by using the same B O E F analysis used for the LISM method,
where 5 = the groundline deflection of a pile subjected to a lateral load,
H , at the ground surface. From (7), the ratio of groundline deflection of a
pile for the first cycle of load, 5~, and the deflection at N cycles of load,
8N, is shown here
240

J. Geotech. Engrg. 1994.120:225-244.

 A'H
EIO.a.nOi6 -
/ "(nhN~ ................................. (13a)
~N A "H \nhx/
EiO.4..0.6
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

t~hN

continuing

~N = 81"
( o6
g/hi = ~ l ' ( N t ) 0"6 = ~I'N ~ . .................... (13b)
\nhu/

and since near the ground surface, y is approximately 8

nhl ~ yl'(Nt) ~ = yl'N ~ ..................... (13c)

YN = YI" \ n h N /

Good agreement between t values back-calculated with the DSPY method

and t values back-calculated with the LISM method is observed when a =
0.6. Theoretically, the value of a varies with depth, however, numerical
investigations using e~varying with depth provided no better agreement with
the LISM method than with a constant value of 0.6. The DSPY method
can be applied to nonlinear p - y curves to account for effects of cyclic lateral
loading by modifying the soil resistance, p, and soil deflection, y, according
to
PN = Pl "N-O'4t ............................................. (14)
YN = Yl "NO'6t 9 ............................................. (15)

Values of parameter, t, can be determined using values listed in Tables 2 -

4 and the formula provided in (9).

LIMITATIONS
The method proposed is an empirical approach intended to provide the
designer with a simple and expedient means to estimate effects of cyclic
lateral load on piles in sand. Because the methods are empirical, efforts
should continue to include more results to verify or modify the recommen-
dations herein. Approaches based on more fundamental soil properties are
needed.
The proposed method may overpredict the effect of cyclic load for piles
subjected to cyclic load when the load ratio is less than 0. While the deg-
radation using (9) is predicted to be minimal, the stiffness at the pile head
may actually increase due to soil densification in the proximity of the pile.
Results of this study are believed to be valid for long piles in sand sub-
jected to 50-100 cycles of nondynamic lateral loads. Although one test
showed continued effects of cyclic load at 500 cycles, the majority of load
tests were for 50 cycles or less. Caution should be exercised when predicting
effects of cyclic loading beyond 50 cycles of load. Finally, the flexural rigidity
of the pile is assumed to remain constant.

CONCLUSIONS
Two simple methods for determining the effect of cyclic lateral loads on
piles in sand are presented. The two methods use parameters derived from
241

J. Geotech. Engrg. 1994.120:225-244.

 the results of 34 full-scale, cyclic, lateral load tests to determine the effects
of cyclic loading. The LISM method employs a closed-form solution using
a B O E F analysis with a linearly increasing soil reaction modulus degraded
from a static modulus. The DSPY method deteriorates the resistance pro-
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

vided by a given static p-y curve to account for the effects of cyclic lateral
load. The DSPY procedure can be applied to nonlinear static p-y curves;
therefore, current procedures for predicting the static p-y curves can still
be employed, while the DSPY method can be applied to estimate the effect
of cyclic loads.
The most important p a r a m e t e r found to govern the behavior of piles
during cyclic loading are the characteristics of the cyclic load. Other factors
include the method of installation and the soil density.

ACKNOWLEDGMENTS
This study was supported by a research initiation grant from the University
of Illinois and by a fellowship from Shell Oil C o m p a n y Foundation. Their
financial support is gratefully acknowledged.

APPENDIX I. REFERENCES
Alizadeh, M. (1969). "Lateral load tests on instrumented timber piles." Performance
of deep foundations, ASTM STP 444, American Society for Testing and Materials,
379-394.
Alizadeh, M., and Davisson, M. T. (1970). "Lateral load test on piles--Arkansas
River project." J. Soil Mech. and Found. Engrg. Div., 96(5), 1583-1604.
Barton, Y. O. (1982). "Laterally loaded model piles in sand, centrifuge tests and
finite element analyses," PhD thesis, University of Cambridge.
Bhushan, K., Lee, L. J., and Grime, D. B. (1981). "Lateral load tests on drilled
piers in sands." Drilled piers and caissons, Proc. Geotech. Engrg. Div. at ASCE
Nat. Convention, St. Louis, Mo., Oct. 28.
Broms, B. (1964). "Lateral resistance of piles in cohesionless soils." J. Soil Mech.
and Found. Engrg., ASCE, 90(3), 123-156.
Chang, C. S., and Whitman, R. V. (1988). "Drained permanent deformation of
sand due to cyclic loading." J. Geotech. Engrg. Div., ASCE, 114(10), 1164-1180.
Davisson, M. T. (1970). "Lateral load capacity of piles." Highway Res. Record, 333,
104-112.
Davisson, M. T., and Salley, J. R. (1968). "Lateral load tests on drilled piers."
ASTM Symp. on Deep Foundations, San Francisco, June 24.
Davisson, M. T., and Salley, J. R. (1970). "Model study of laterally loaded piles."
J. Soil Mech. and Found. Engrg., ASCE, 96(5), 1605-1627.
HeUer, L. W. (1965). "Lateral thrust on piles." Technical Report R283, U.S. Civ.
Engrg. Lab., Port Hueneme, Calif., June 15.
Hetenyi, M. (1946). Beams on elastic foundations. University of Michigan Press,
Ann Arbor, Mich.
Kramer, S. L., and Heavey, E. J. (1988). "Lateral load analysis of nonlinear piles."
J. Geotech. Engrg., ASCE, 114(9), 1045-1049.
Little, R. L., and Briaud, J-L. (1988). "Full scale cyclic lateral load tests on six single
piles in sand." Miscellaneous Paper GL-88-27, Geoteehnical Div., Texas A&M
Univ., College Station, Tex.
Lock and dam No. 4, Arkansas River and tributaries, Arkansas and Oklahoma.
(1964). U.S. Army Engineer District, Little Rock, Corps of Engineers.
Long, J. H., and Reese, L. C. (1982). "Prediction of lateral load behavior for
reinforced concrete pile." letter report to L. Johnson, WES, Mar.
Long, J. H., and Reese, L. C. (1984). "Testing and analysis of two offshore piles
subjected to lateral loads." Laterally loaded deep foundations: analysis and per-
formance, ASTM STP 835, J. A. Langer, E. Mosely, and C. Thompson, eds.,
American Society for Testing Materials, Philadelphia, Pa., 214-228.
242

J. Geotech. Engrg. 1994.120:225-244.

 Matlock, H. (1974). "Correlations for design of laterally loaded piles in soft clay."
Second Annual Offshore Technology Conf., Paper No. 1204, Houston, Texas,
May, 577-594.
Matlock, H., Foo, S. H. C., and Bryant, L. M. (1978). "Simulation of lateral pile
behavior under earthquake motion." Proc. ASCE Specialty Conf. on Earthquake
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

Engrg. and Soil Dynamics, June, 600-619.

Meyer, B., and Reese, L. C. (1979). "Analysis of single piles under lateral loading."
Research Report 244-1, Ctr. for Transp. Res., Dec.
Morrison, C. S. (1986). "A lateral load test of a full-scale pile group in sand," PhD
thesis, University of Texas, Austin, Tex.
O'Neill, M. W., and Murchison, J. M. (1983). "An evaluation of p-y relationships
in sands." Research Report No. GT-DF02-83, University of Houston, Houston,
Tex., May.
Parry, R. H. G., and Sanglamer, A. (1977). "Lateral load tests on single model piles
with radiographic observations." Cambridge University Interim Report, CUED~C,
Soils TR 36.
Poulos, H. G. (1982). "Single pile response to cyclic lateral load." J. Geotech. Engrg.,
ASCE, 108(3), 355-375.
Prakash, S. (1962). "Behavior of pile groups subjected to lateral loads," thesis,
University of Illinois, Urbana, II1.
Reese, L. C. (1984). "Handbook on design of piles and drilled shafts under lateral
load." FHWA-1P-84-11, U.S. Dept. of Transp., Federal Highway Administration,
July.
Reese, L. C., Cox, W. R., and Koop, F. D. (1974). "Analysis of laterally loaded
piles in sand." Paper No. 2080, Sixth Annual Offshore Tech. Conf., Vol. 2, Hous-
ton, Texas, May.
Reese, L. C., and Matlock, H. (1956). "Non-dimensional solutions for laterally
loaded piles with soil modulus assumed proportional to depth." 8th Texas Conf.
on Soil Mech. and Foundation Engrg., Sep. 14.
Robinson, K. E. (1979). "Horizontal subgrade reaction estimated from lateral load
tests on timber piles." Behavior of deep foundations, ASTM STP 670, Raymond
Lungren, ed., American Society for Testing Materials, Philadelphia, Pa., 520-
536.
Sanglamer, A. (1979). "Model study of laterally loaded single piles." Proc. Seventh
European Conf. on Soil Mech. and Foundation Engrg., Brighton, England, Sep.,
Vol. 2, 115-120.
Stevens, J. B., Holloway, M. D., Moriwaki, Y,, and Demsky, E. C. (1979). "Pile
group response to axial and lateral loading." Proc. Symp. Sponsored by Geotech.
Engrg. Div., ASCE, Oct. 25.
Swane, I. C., and Poulos, H. G. (1982). "A theoretical study of the cyclic shakedown
of laterally loaded piles." Research Report No. R415, School of Civ. and Mining
Engrg., Univ. of Sydney, July.
Tucker, L. M., and Briaud, J. L. (1988). Analysis of the pile load test program at
the lock and dam 26 replacement project. U.S. Army Engineer District, St. Louis,
Mo., June.
Turner, J. P., Kulhawy, F. H., and Charlie, W. A. (1987). "Review of load tests
on deep foundations subjected to repeated loading." Report EL-5375, Electrical
Power Res. Inst., Palo Alto, Calif., Aug.
Vesic, A. (1977). "Design of pile foundations." National cooperative highway re-
search program synthesis of highway practice, Report No. 42, Transportation Re-
search Board, Washington, D,C.

APPENDIX II. NOTATION

The following symbols are used in this paper:

A = deflection constant for B O E F analysis;

a = degradation parameter for method by Little and Briaud;
243

J. Geotech. Engrg. 1994.120:225-244.

 n deflection constant for BOEF analysis;
D= diameter of pile (L);
E/= bending stiffness of pile, modulus times moment of inertia (FL2);
H= lateral load applied to top of pile (F);
Downloaded from ascelibrary.org by ELYACHAR CENTRAL LIBRARY on 05/08/14. Copyright ASCE. For personal use only; all rights reserved.

//min = minimum horizontal load during cyclic loading (F);

/max z maximum horizontal load during cyclic loading (F);
K,= soil reaction modulus ( F / L 2 ) ;
KhN = soil reaction modulus at N cycles of load ( F / L 2 ) ;
Kha = soil reaction modulus for first cycle of load ( F / L 2 ) ;
L = length of pile (L);
M = bending moment in pile (FL);
N= number of cycles of load;
nh = coefficient of soil reaction (F/L3);
nhN coefficient of soil reaction at N cycles of load (F/L3);
nhl = coefficient of soil reaction for first cycle of load (F/L3);
p = resistance provided by soil ( F / L ) ;
P~v = soil resistance N cycles of load ( F / L ) ;
Pl = soil resistance for first cycle of load ( F / L ) ;
R14 = cyclic load ratio, ratio of minimum to maximum horizontal load
(=nminlHmax);
Rn = cyclic modulus ratio, ratio of cyclic to static nh;
RnN = value of Rn at N cycles;
Rnl = value of Rn on first cycle of load;
t = degradation parameter for elastic soil modulus, nh;
y = pile deflection corresponding to specific soil resistance at specific
depth (L);
Yl = pile deflection corresponding to specific soil resistance at specific
depth for first cycle of load (L);
YN = pile deflection corresponding to specific soil resistance at specific
depth for N cycles of load (L);
Ot = degradation parameter for soil resistance, p;
= lateral deflection at top of the pile (L); and
~3N = on Nth cycle of load (L).

244

J. Geotech. Engrg. 1994.120:225-244.