OFFSHORE TECHNCLGGY CO~rERENCE PAPER

6200 North Central Expressway NUMBER 0 TC 2 r 86

Dallas, Texas 75206
THIS PRESENTATION IS SUBJECT TO CORRECTION

Computer Pred i ct ions for Ax i a I I y-Loaded Pi I es

with Non I i near Supports

By
Patrick L. Meyer, Darrel V. Holmquist and Hudson Matlock, U. of Texas at Austin

@Copyright 1975

Offshore Technology Conference on behalf of the American Institute of Mining, Metallurgical, and
Petroleu~ Engineers, Inc. (Society of Mining Engineers, The Metallurgical Society and Society of
Petroleum Engineers), American Association of Petroleu~ Geologists, American Institute of Chemi-
cal Engineers, American Society of Civil Engineers, American Society of Mechanical Engineers,
Institute of Electrical and Electronics Engineers, Marine Technology Society, Society of Explor-
ation Geo~hysicists, and Society of Naval Architects and Marine Engineers.

This paper was pre~ared for presentation at the Seventh Annual Offshore Technology
Conference to be held in Houston, Tex., May 5-8, 1975. Permission to copy is restricted
to an abstract of not more than 300 words. Illustrations may not be copied. Such use of
an abstract should contain conspicuous acknowledgment of where and by whom the paper is
presented.

Abstract (3) consideration of hysteretic response, and

(4) estimation of residual stresses after
Although techniques have been developed
loading or driving. These aspects are impor-
for prediction of lateral loading response of
tant both for static design and for analysis
foundation piles for offshore structures,
of dynamic response.
most of the attention to axial behavior has
Some phases of the axial soil-pile prob-
been focused only on consideration of ulti-
lem are under study at the present time in a
mate load capacity,' The field data to sup-
project supported by the American Petroleum
port predictive methods is inadequate, being
Institute entitled, "The Study of Piles Under
limited primarily to load settlement tests
Various Axial Loadings." An interim develop-
which report only the response at the head of
ment of this project is a computer program
the pile. In the vast majority of cases in
for static axial loading pile behavior.
offshore operations, the performance of load
The computer program uses a discrete-
tests is prohibitively expensive. A reason-
element model to produce predictions of the
ably satisfactory method to predict pile-head
settlement of a pile under static axial
response is critically needed. Various
loading. Nonlinear variations of soil reac-
aspects of the total problem include (1) de-
tion with axial displacement may be described
termination of soil properties, (2) predic-
along the length of the pile, and proper de-
tion of soil reaction versus pile displace-
velopment of such support curves is the only
ment in various soils at various depths,
significant uncertainty involved in the
References and illustrations at end of paper. analysis method.
 COMPUTER PREDICTIONS FOR AXIALLY-LOADED PILES WITH NONLINEAR SUPPORTS OTC 2186
Data from two pile-load tests are com- The discrete-element technique has been
pared to the results of analyses performed used extensively with good results for model-
using the computer program in relation to a ing the behavior of foundation piles under
parameter study involving simple support axial and lateral loading (Refs 1, 2, 3, 4).
curves. The sensitivity of the computer Computer programs have been developed for
prediction to changes in support curves is convenient and practical application of this
discussed in relation to the need for further analytical approach. One such FORTRAN com-
development of criteria for axial pile-soil puter program developed for the lateral solu-
behavior. Hysteretic soil response and re- tion of beam-columns on nonlinear supports is
sidual forces in the pile are also considered BMCOL 28 (Ref 3). Recently, an axial solu-

tion counterpart has been developed by modi-
Introduction
fying the BMCOL 28 program. This axial pro-
Foundation piles are typically designed
gram has been called AXCOL 1 (Ref 5). It was
for axial loading using ultimate capacity
produced to fulfill the need for a program
considerations. Often no prediction of the
which is convenient to use f~r obtaining
axial load-displacement characteristics is
directly the response of a pile to specified
made. However, such predictions, if reliable,
axial conditions.
would contribute to a more rational design of
piles, and would be useful as a support con- AXCOL 1

dition in frame analysis related to offshore AXCOL 1 utilizes a discrete-element

structures. This kind of prediction can be model as a basis in formulating a set of
made quickly with the use of ~ computer if simultaneous finite-difference equations
the axial stiffness and the soil support which are solved to produce a prediction of
conditions may be adequately described along the behavior of a pile under specified static
the length of the pile. loads and restraints. The support conditions
The displacement of a foundation pile may be considered linear or nonlinear but the
under axial load results from the elastic pile is considered to be linearly elastic.
shortening or elongation of the pile under If any nonlinear support is used in a par~

compressive or tensile loads, respectively, ticu1ar problem, that problem requires an

and the movement of the pile in the support- iterative type solution. Linear problems are
ing soil. A method of modeling the behavior solved directly. The axial solution of a
of piles under axial loading is to consider pile-soil system using AXCOL 1 is one-dimen-
the pile to be composed of a number of incre- sionai, therefore bending and buckling are
ments or "discrete elements." This method not involved. Although nonlinear, this axial
has been used by Smith (Ref 1) in relation to solution is elastic. However, special mea-
the dynamic response of axially loaded piles, sures may be taken to simulate inelastic be-
and by Coyle and Reese (Ref 2) in relation to havior as demonstrated in Example 3.
the static loading condition. Each of these The axial-column model used in AXCOL 1
pile elements has an axial stiffness and may is shown in Fig 1. The model is a mechanical
have specified loads and supports acting on analogy which has proven to be adequate in
it. Using this simple method of increments a representing the soil-structure interaction
computer program may be developed for rapid of an axially-loaded pile. When describing a
analysis of individual axially-loaded piles. pile in terms of the axial-column model, the
OTC 2186 PATRICK L. MEYER, DARREL V. HOLMQUIST AND HUDSON MATLOCK 377
pile is divided into a suitable number of in- problem is described by defining the axial
crements of equal length in the range of 50 stiffness, load, and support conditions along
to 200. the length of the column. The data input for-
The axial-column solution technique in- mat ofAXCOL 1 allows these parameters to be
volves the solution of a set of simultaneous automatically distributed between selected
finite-~ifference equations which are formu- input points in a linear variation.
lated from the model shown in Fig 1. A The resulting axial-column solution will
force-equilibrium relation can be written for be a mathematical solution of the axial-
each node (station) between increments. column model described by the program user in

o his input data. The program user makes the

-T"l + T l"+1 + Q"l - S"l u"~ (1)
approximations which will be involved in the
The axial thrust in each increment can be solution as he interprets the problem in cre-
expressed in terms of axial stiffness values ating the axial-column to be solved. There-
and axial displacements. fore, the solution is only as good as the
problem description.
(AE) "
~ (-u _ + u ) (2a) Some practical applications for which
i l i
AXCOL 1 might be used are the prediction of

= compressive load-settlement curves, pullout-

(2b)
displacement curves, and inelastic behavior
Then the governing equation of the axial- of a pile after loading is reduced or removed.
column solution may be formulated by substi-
Axial Soil-Support Curves
tuting equations 2a and 2b into equa-
tion 1 and rearranging terms. The only significant uncertainty involved
in this method of analysis is the development
of support curves along the length of the

where axial-column. These support curves are an

estimation of how axial support reaction
AE.l
force varies with displacement at a particu-
b.
l h
lar station on the axial-column. For soil
AE. AEi+l support considerations these curves fall into
_ .:l-
c.
l h - h S.
l two categories, side-shear and end-bearing.
The soil support curve description is
AEi+l
d. typically broken into two parts and such a
l h
separation has been found convenient to use
f.l Qi throughout this paper. The support curve
description is composed of:
All terms used in the above equations are
defined in Fig 1. During the solution the (1) an estimation of the peak soil
reaction which could be developed
simultaneous equations of the form of equa-
on a section of pile, and
tion 3 are solved for the unknown disp1ace-
(2) a curve shape which describBs the
ments u. by employing a direct elimination
l
variation of fractions of this
procedure described in Ref 5.
The axial-column model allows great peak soil reaction with pile dis-
placement.
flexibility in problem description. A
@ COMPUTER PREDICTIONS FOR AXIALLY-L ,DED PILES WITH NONLINEAR SUPPORTS OTC 2186

The peak soil reaction is used to refer to to provide any information describing the
the largest side-shear soil reaction which distribution of the side-shear transfer along
can act on one increment length of pile, or the length of the pile.
the largest end-bearing reaction which can Recommendations related to side-shear
act on the pile tip. The example problems curve shape for driven piles in clay have
included in this paper will provide some in- been made by Coyle and Reese in Ref 2. This
dication of how critical these two parts of a work is acknowledged as the first significant
support curve may be with regard to the pre- attempt to analyze the axial soil-structure
diction of static compressive load-settlement interaction problem. The paper contains
behavior. curve shape recommendations which have become

Most of the published recommendations widely known as t-z curves.

concerning axially-loaded piles have been A curve shape recommendation for end-
oriented toward the estimation of total pile earing in clay has been made by Skempton in
capacity. The American Petroleum Institute Ref 10.
recommends methods of estimating side-shear Coyle and Sulaiman have investigated
capacity and end-bearing capacity of piles in side-shear for steel piles in sand and report
cohesive and cohesionless soils in Ref 6. field load test data and curve shape recom-
The concept of correlating the ratio of mendations in Ref 11.
maximum side-shear to undrained shear strength The end-bearing curve shape for piles
for piles driven through cohesive soil as driven into sand remains relatively undefined.
discussed in Refs 7 and 8 has become popular. These kinds of criteria can be developed
More recently Vijayvergiya and Focht have from pile load test data in which piles have
suggested that the total side-shear caPgcitY _ been instrumented for measurement of axial
of an axially-loaded pile is related to the thrust variation along the length of the
effective stress and the undrained shear pile. Other methods of obtaining useful in-
strength (Ref 9). formation are laboratory tests and computer
Before using a particular recommendation studies using finite-element techniques.
for estimating peak soil reactions acting Clearly, no single set of criteria could
along the length of the axial-column, the be used to resolve the soil %upport descrip-

AXCOL 1 program user must consider the use- tion problem in general since real soil sup- ,

fulness of the particular recommendation. In port certainly depends on several conditions

the opinion of the authors the axial-column such as type of soil, effective stresses,

simulations of pile behavior should represent pile type, procedure of installation, and

the best available estimate of the real be- kind of loading. Considerable discretion and
havior. Conservatism and safety factors judgment must be used when developing a set
should be applied only to the finished of support curves,for analysis of.each par-
product. Some of the published recommenda- ticular problem.
tions related to axially-loaded pile capacity
Example Problems
cannot be used for an axial-column analysis
TWO of the three example problems are
because the inherent conservatism defeats the
real pile-soil systems taken from the litera-
purpose of the analysis. Also some recommen-
ture in which pile load test data have been
dations, while apparently useful in predicting
reported. These two problems are presented
total side-shear capacity, are not intended
fITf!
-.
?1$lA
...
DATT?TOi7
.-A...*..
T
L1.
?mV17w
*W *U,,,
llAPRl?T.
JJ41L..U.IU
Vv . U(IT Mf)TTT~T
Lv J.l. Ly A J.
AN7T
u
.Z..
?TTTTlSf3N
..u..
MA
---------
TT.fWK ?7
-,,

to demonstrate the use of AXCOL 1, and more c is the undrained shear

importantly to show how variations in support strength, and

curve description affect the resulting pre- As is the shaft area corre-
dicted behavior. A reported field load test spending to an incremental
curve is shown plotted together with predict- section of pile.
ed behavior only in order to provide a refer-
For each of these example problems the
ence base for the reader. The reader should
shape of the side-shear support curves is
be aware that these example problems are used
assumed to be elasto-plastic. Actually the
here to present a parameter study on support How-
curve shape is usually more nonlinear.
curves and are not intended to recommend a ever, the elasto-plastic form is convenient
method of developing axial soil support
for the study presented herein. The side-
curves. shear curve shape is constant along the
The third example problem demonstrates axial-column for each set of solutions of the
how an axial-column analysis can be used to example problems.
simulate the inelastic axial behavior of a
Example 1 - Intermediate-Length Pile
foundtitionpile during unloading and reload-
ing. This problem is purely hypothetical but One of the field load tests reported by
demonstrates a particular capability. NcCammon and Golder in Ref 12 was performed

The peak end-bearing reaction was esti- on a 24-inch diameter circular steel pipe
mated in each problem using the following: pile having an embedment of 158 feet in co-
hesive material. An interpretation of the
Qpeak = NccAt
data is shown in Fig 2, The length of pile

where extending above the groundline was assumed to

be 10 feet. The pile was closed at the tip.
N the assumed bearing capacity
c
This field load test was performed after the
factor, is equal to 9,
pile had remained undisturbed for a period of
c is the undrained shear strength
170 days.
of the soil at the pile tip, The assumed support curve shapes and re-
and
suiting computer predictions of compressive

At is the end area of the pile load-settlement behavior are shown in Figs 3
tip. and 4. The end-bearing support curve remains
unchanged for each set of solutions, However,
The peak side-shear reaction acting on
the soil support yield displacement is 0.05
each station of the axial-column was esti-
inches for side-shear curve shape A and 0.30
mated in each problem by using an a
inches for side-shear curve shape B. A corn-
value method as recommended in Ref 8. The
parison of the computer predictions of load-
relation follows:
settlement behavior is shown in Fig 4 and
Q peak =a cAS
shows that such variations in yeild displace-
ment do not produce greatly differing results
where
for this problem.
a is a multiplier which is In order to demonstrate the effect of
correlated with undrained shear overly conservative estimates of peak side-
strength,
shear, all of the peak side-shear reactions
380 COMPUTER PREDICTIONS FOR AXIALLY-LOADED PILES WITH NONLI@AR SUPPORTS OTC 2186

acting along the length of the axial-column support yield displacement is 0.05 inches for
were reduced by 50 percent. A set of axial- side-shear curve shape A and 0.30 inches for
column solutions were then performed using side-shear curve shape B. A comparison of
side-shear curve shape B of Fig 3. The re- the load-settlement predictions performed
suiting axial load-settlement behavior is using these two curve shapes indicates that a
shown in Fig 4 for comparison with the origi- six to one variation in side-shear yield dis-

nal prediction performed using the same side- placement has even less effect on the slope of
shear curve shape. Clearly, if such an un- the early portion of the load-settlement
derestimation of the peak soil reactions was curve for this example than for the previous
made, the predicted load-settlement curve example. This is because the longer pile ex-
would be quite different from the actual be- hibits more axial shortening than in the pre-
havior. vious example and this tends to mask the
It is often desired to have an estimate effect of the yield displacements assumed.
of settlement of a pile under design loading Next, all peak side-shear support reac-
which typically is about one-third to one- tions acting along the axial-column were uni-
half of the ultimate capacity. In this formly reduced by 18 percent. The 18 pefcent
range, Fig 4 shows that for the example con- reduction of the original estimates.of peak
sidered, a change in yield displacement in side-shear would make the total estimated
the ratio of one to six has about the same axial load capacity the same as measured

effect on the slope of the load settlement during the field test. A set of axial-column
curve as a reduction of one-half of the peak solutions were then performed using curve
side-shear resistance. shape A of Fig 6 forside-shear support. The
resulting prediction of load-settlement be-
Example 2 - Deep Penetration Offshore Pile
havior is shown in Fig 7 for comparison with
A series of field load tests reported by
the field curve.
McClelland and Lipscomb in Ref 13 were per- Even though the predicted load-se~tlement
formed on a 14-inch diameter circular steel curve using reduced side-shear agreeswell
pipe pile having an embedment of 333 feet in with the field curve, some consideration could
a cohesive underconsolidated material. The be directed to the fact that it passes beneath
test location was in the South Pass area off Such a condition, though
the field curve.
the coast of Louisiana. An interpretation of
not pronounced in this particular example,
the data is shown in Fig 5. The length of could be the result of too much axial short-
pile extending above the mudline was assumed ening effects in the axial-column prediction.
to be 74 feet. The field load test curve
This could be the result of error in estimat-
shown in Fig 7 is of a reloading test which
ing the appropriate distribution of peak sup-
was conducted immediately after the pile was
port reactions along the length of the pile.
loaded to plunging and then unloaded. Before
Example 3 - Simulation of Inelastic Behavior
the first plunging load test the pile had been
left undisturbed for a period of 23 days. The hypothetical pile-soil system shown

Assumed support curve shapes and re- in Fig 8 will be used in demonstrating the
suiting computer predictions of compressive use of AXCOL 1 for simulating the behavior of
load-settlement behavior are shown in Figs 6 axially-loaded piles.
and 7. The end-bearing support curve remains First, a set of axial-column solutions
unchanged in the various solutions. The soil were performed using the curve shapes shown
.
OTC 2186 PATRICKL. MEYER. DARREL-V. HOLMCIUISTAND HUDSON MATLOCK 381

in Fig 9 for successively larger pile-head Summary

displacements until plunging load was reached.
The axial-column method of analysis can
The resulting load-settlement curve is shown
be used, together with reliable pile-soil
.
in Fig 10.
system data, as a rational method of modeling
Next the support curves were altered to
the behavior of foundation piles under static
reflect the inelastic supporting characteris-
axial-loading. Work is continuing which will
tics for a following series of solutions
extend this analysis method for more convenient
obtained in simulation of unloading. Fig 11
use in relation to inelastic behavior. A more
describes how support curves were altered
general and convenient inelastic capability
when loading was reversed. The axial-column will lead to an improved method of analysis of
was unloaded to zero applied axial load as
pile behavior under transient and dynamic
shown by the unloading curve in Fig 10. The loading.
solution shows the residual forces that are
This paper is an interim report intended
in effect along the length of the axial-
only to demonstrate the static loading analy-
column after loading has been removed. The
sis methods and to spot-check the sensitivity
predicted variation of axial thrust (internal
of typical computer solutions to soil support
compressive force) along the length of the
variations. No attempt has been made to
pile after complete unloading is shown in
recommend soil support criteria.
Fig 12, with the orientation of the support
The solutions to the axial problems
reactions shown at the left. The maximum selected appear to suggest that axial short-
residual thrust in this example is only about ening effects are dominant over soil-deforma-
one-sixth of the pile capacity but might be
tion effects in contributing to the load-
considerably greater for a longer pile.
settlement behavior of intermediate length
Finally, the support curves were changed and long friction piles. As pile length in-
again to accommodate reloading (Fig 11), and
creases, this effect is more pronounced. The
a series of solutions were performed to simu- predicted behavior of long friction piles is
late the pile behavior. The difference
nme sensitive to the estimation of peak
between the predicted path of the unloading
side-shear reactions and the distribution of
curve and the reloading curve, as shown in these reactions along the length of the axial-
Fig 10, is only slight for this problem. column than to the side-shear yield displace-
AXCOL 1 is not structured conveniently ment.
for use in simulating inelastic behavior.
Acknowledgements
The curve shape changes shown in Fig 11 must
bemade for each axial-column station when The present computer program, AXCOL 1,
the direction of loading reverses. Chan and is documented for general use in a report to
Matlock (Ref 4) have developed computer rou- the American Petroleum Institute (Ref 5).
tines which can be used to make an inelastic The authors wish to express their apprecia-
analysis more convenient. tion to the American Petroleum Institute for
their sponsorship of a research project en-
titled The Study of Piles Under Various
Axial Loadings. This research project in-
volves both laboratory and analytical studies
and is partly responsible for the work re-
ported in this paper.
382 COMPUTER PREDICTIONS FOR AXIALLY-LOADED PILES WITH NONLINEAR SUPPORTS OTC 2186

References International Conference on Soil Mechan-

ics and Foundation Engineering, London
1. Smith, E. A. L.: Pile-Driving Analysts (1957), Vol II, 66-71.
by the Wave Equation, Journal of the
8. Woodward, R. J., Lundgren, R., and
Soil Mechanics and Foundation Division,
ASCE (August 1960), Vol 86, SM4, 35-61. Boitano, J. D., Jr.: Pile Loading
Tests in Stiff Clays, Proceedings, 5th
2. Coyle, Harry M., and Reese, Lymon C.: International Conference on Soil Me-
Load Transfer for Axially Loaded Piles chanics and Foundation Engineering,
in Clay, Journal of the Soil Mechanics Paris, France (July 1961), Vol 2,
and Foundations Division, ASCE (March 177-184.
1966), Vol 2, No. SM2, 1-25.
9. Vijayvergia, V. N., and Focht, J. A.,
3. Matlock, Hudson, and Halliburton, T. Jr.: A New Way to Predict Capacity of
Allan: llBMCOL28, A Program for Finite- Piles in Clay, Paper No. OTC 1718 pre-
Element Solution of Beam-Columns on Non- sented at the Fourth Annual Offshore
Linear Supports,Austin, Texas (1964). Technology Conference, Houston, Texas,
May 1-3, 1972.
4. Chan, Jack H. C., and Matlock,Hudson:
A Discrete-Element Method for Trans- 10 Skempton, A. W.: The Bearing Capacity
verse Vibrations of Beam-Columns Resting. of Clays, Building Research Congress,
on Linearly Elastic or Inelastic Sup- Division 1, London.
ports, ~tpaper No. OTC 1841, presented at
the Fifth Annual Offshore Technology 1,1. Coyle, Harry M., and Sulaiman, Abrahirn
Conference, Houston, Texas, April 30- H .: Skin Friction for Steel Piles in
May 2, 1973. Sand, Journal of the Soil Mechanics and
5. Holmquist, Darrel V., Meyer, Patrick L., Foundations Division, ASCE (November
and Matlock, Hudson: A Program for 1967), Vol 93, No. SM6, 261-278.
Discrete-Element Solution of Axial- 12 McCammon, N. R., and Golder, H. Q.:
Columns with Non-Linear Supports, Some Loading Tests on Long Pipe Piles,
report to the American Petroleum In- Geotechnique (1970), VO1 20, No. 2,
stitute, Austin, Texas (1975). 171-184.
6. API Recommended Practice for Planning, 13 McClelland, Bramlette, and Lipscomb,
Designing, and Constructing Fixed Off- Lindsey J.: Load Test of a 333-Foot
shore Platforms, American Petroleum Friction Pile in Deep Under-Consolidated
Institute, Washington, D.C. (January Clay, paper presented to the National
1974). Meeting of the ASCE in Dallas, Texas
(February 1956).
7. Tomlinson, M. J.: l!TheAdhesion of piles
Driven in Clay Soils, Proceedings, 4th

 . .
kH.l+, i_,

:$mT;$+l<;$+2
Station i-1 Stiffness + i Stiffness ~ i+l,

i = external station number

Fig. 1 - Axial-column modeL. x = length along column from
station zero (top)
ui = displacement at station i
hi = length of one increment
AEi = productof area and mo&lus
of elasticity for the bar
between station i and i-l.

PEAK SIDE
SHEAR STRENGTH SHEAR
Sta O .: c, psf Qpeak, kips/sta

o 5 10.15
Sta 5-- ?&.@,A\~,~~

20 15
Steel
Pipe Pile
24 Diameter, 40 25
1/2 Wall
Thickness
60 g 35
t
aJ z
J-2 s
. = 45
s 80
m i- 0
m
CL
w i=
n 100 ; 55

20 65

40 75

Sta
I 84 158 84

Increment Length =2ft Peak End Bearing =56.5 kips

Fig.2 - Example 1, an interpretation of the pile-soil system tested by

McCammon and Golder (Ref 11) .
 1.0 ---~
;A B

Q
Q peak
0.5 Y
:

;
I

O 0.5 1.0 1.5 2.0 2.5 3.0-

U, AXIAL DISPLACEMENT, inches
Fig. 3 - Assumed curve shapes for side-shear and end-bearing used in
Example 1.

AXIAL LOAD AT STATION ZERO, kips

w
N
z
o
1.0 -
Using
Using Curve Curve
2.0 Shape B, Qpeak Shape B
\
~ Reduced by

z 3.0 50 o -+
Ld
1 Field
v
Load
4.0 Test ~

I
Fig. 4 - Load-settlement curves for the pile head
of Example 1.
 Sta O
PEAK SIDE
SHEAR STRENGTH SHEAR
c, $Sf Qpeak, kipshta
~0 200400600
Sta 27
~~
Steel
Pipe Pile 50 -
14 Diameter,
0.437 Wall I 00
Thickness h

-i
%

Sta 148 I

Increment Length =33 inches Peak End Bearing =5.8 kips

Fig. 5 - Example 2, an interpretation
of the pile-soilsystemtestedby
McClellandand Lipscomb12

N.

o
 PEAK SIDE
SHEAR STRENGTH SHEAR
c, psf Q LxyJk,kips/sta
500 1000 002: 5075 *
Sta O Ks===m30 I

Steel
Pipe Pile
24 Diameter, 20
1/2 Wall
Thickness

40

80

Sta 10 100 \ 10
Increment Length =lOft Peak End Bearing =30.5 kips
Fig. 8 - Example 3, a hypothetical
pile-soilsystemuse~ to demonstrate
elasticbehavior.

AXIAL LOAD AT STATION ZERO, kips

100 200 300 400
o

;1,0

0-
az
I.Ll
N

..

4.0 .

Fig. 10 - Inelasticload-displacement
behaviorof the pile
heaa for Example3.
 0H7----L--
Q /Aij >
Qpeak 0
1 u, AXIAL DISPLACEMENT
, /
I

-1.0 1-
t
EL
t#!
1 I

Fig. 11 - The method of representing inelastic support

for Example 3.

THRUST, kips
1 203040 50 60 70

20 -
-i-
:
- 40 -
z
1-
~ 60 -

80 -

100

Fig. 12 - Prediction for Example 3 afterunloading

showingvaristion in axial thrust along the length of
the pile and orientation of the support reactions.