by
October 2013
Table of Contents
2-2 Derivation of Differential Equation for the Beam-Column and Methods of Solution....... 28
2-2-1 Derivation of the Differential Equation ...................................................................... 28
2-2-2 Solution of Reduced Form of Differential Equation................................................... 32
2-2-3 Solution by Finite Difference Equations..................................................................... 37
Chapter 3 Lateral Load-Transfer Curves for Soil and Rock......................................................... 45
3-1 Introduction ........................................................................................................................ 45
3-2 Experimental Measurements of p-y Curves........................................................................ 47
3-2-1 Direct Measurement of Soil Response ........................................................................ 47
3-2-2 Derivation of Soil Response from Moment Curves Obtained by Experiment............ 47
3-2-3 Nondimensional Methods for Obtaining Soil Response ............................................. 49
3-3 p-y Curves for Cohesive Soils ............................................................................................ 50
3-3-1 Initial Slope of Curves................................................................................................. 50
3-3-2 Analytical Solutions for Ultimate Lateral Resistance ................................................. 52
3-3-3 Influence of Diameter on p-y Curves .......................................................................... 58
3-3-4 Influence of Cyclic Loading........................................................................................ 59
3-3-5 Introduction to Procedures for p-y Curves in Clays.................................................... 61
3-3-5-1 Early Recommendations for p-y Curves in Clay ................................................. 61
3-3-5-2 Skempton (1951).................................................................................................. 61
3-3-5-3 Terzaghi (1955).................................................................................................... 63
3-3-5-4 McClelland and Focht (1956) .............................................................................. 63
3-3-6 Procedures for Computing p-y Curves in Clay ........................................................... 64
3-3-7 Response of Soft Clay in the Presence of Free Water................................................. 64
3-3-7-1 Description of Load Test Program....................................................................... 64
3-3-7-2 Procedure for Computing p-y Curves in Soft Clay for Static Loading................ 65
3-3-7-3 Procedure for Computing p-y Curves in Soft Clay for Cyclic Loading .............. 68
3-3-7-4 Recommended Soil Tests for Soft Clays ............................................................. 68
3-3-7-5 Examples.............................................................................................................. 68
3-3-8 Response of Stiff Clay in the Presence of Free Water ................................................ 70
3-3-8-1 Procedure for Computing p-y Curves for Static Loading .................................... 70
3-3-8-2 Procedure for Computing p-y Curves for Cyclic Loading................................... 73
3-3-8-3 Recommended Soil Tests..................................................................................... 74
3-3-8-4 Examples.............................................................................................................. 75
3-3-9 Response of Stiff Clay with No Free Water................................................................ 75
3-3-9-1 Procedure for Computing p-y Curves for Stiff Clay without Free Water for Static
Loading ............................................................................................................................. 76
3-3-9-2 Procedure for Computing p-y Curves for Stiff Clay without Free Water for Cyclic
Loading ............................................................................................................................. 78
3-3-9-3 Recommended Soil Tests for Stiff Clays............................................................. 79
3-3-9-4 Examples.............................................................................................................. 79
3-3-10 Modified p-y Criteria for Stiff Clay with No Free Water ......................................... 80
3-3-11 Other Recommendations for p-y Curves in Clays..................................................... 80
3-4 p-y Curves for Sands........................................................................................................... 81
3-4-1 Description of p-y Curves in Sands............................................................................. 81
3-4-1-1 Initial Portion of Curves....................................................................................... 81
3-4-1-2 Analytical Solutions for Ultimate Resistance ...................................................... 82
3-4-1-3 Influence of Diameter on p-y Curves................................................................... 83
iv
vii
List of Figures
Figure 3-2 p-y Curves from Static Load Test on 24-inch Diameter Pile (Reese, et al.
1975) .......................................................................................................................... 48
Figure 3-3 p-y Curves from Cyclic Load Tests on 24-inch Diameter Pile (Reese, et al.
1975) .......................................................................................................................... 49
Figure 3-4 Plot of Ratio of Initial Modulus to Undrained Shear Strength for Unconfinedcompression Tests on Clay ........................................................................................ 51
Figure 3-5 Variation of Initial Modulus with Depth.................................................................... 52
Figure 3-6 Assumed Passive Wedge Failure in Clay Soils, (a) Shape of Wedge, (b)
Forces Acting on Wedge ........................................................................................... 53
Figure 3-7 Measured Profiles of Ground Heave Near Piles Due to Static Loading, (a)
Heave at Maximum Load, (b) Residual Heave ......................................................... 54
Figure 3-8 Ultimate Lateral Resistance for Clay Soils ................................................................ 56
Figure 3-9 Assumed Mode of Soil Failure Around Pile in Clay, (a) Section Through
Pile, (b) Mohr-Coulomb Diagram, (c) Forces Acting on Section of Pile................. 57
Figure 3-10 Values of Ac and As................................................................................................... 58
Figure 3-11 Scour Around Pile in Clay During Cyclic Loading, (a) Profile View, (b)
Photograph of Turbulence Causing Erosion During Lateral Load Test .................... 60
Figure 3-12 p-y Curves in Soft Clay,(a) Static Loading, (b) Cyclic Loading.............................. 66
Figure 3-13 Example p-y Curves in Soft Clay Showing Effect of J............................................ 67
Figure 3-14 Shear Strength Profile Used for Example p-y Curves for Soft Clay........................ 69
Figure 3-15 Example p-y Curves for Soft Clay with the Presence of Free Water....................... 69
Figure 3-16 Characteristic Shape of p-y Curves for Static Loading in Stiff Clay with Free
Water.......................................................................................................................... 71
Figure 3-17 Characteristic Shape of p-y Curves for Cyclic Loading of Stiff Clay with
Free Water ................................................................................................................. 73
Figure 3-18 Example Shear Strength Profile for p-y Curves for Stiff Clay with No Free
Water.......................................................................................................................... 75
Figure 3-19 Example p-y Curves for Stiff Clay in Presence of Free Water for Cyclic
Loading ...................................................................................................................... 76
Figure 3-20 Characteristic Shape of p-y Curve for Static Loading in Stiff Clay without
Free Water ................................................................................................................. 77
Figure 3-21 Characteristic Shape of p-y Curves for Cyclic Loading in Stiff Clay with No
Free Water ................................................................................................................. 78
Figure 3-22 Example p-y Curves for Stiff Clay with No Free Water, Cyclic Loading .............. 79
Figure 3-23 Geometry Assumed for Passive Wedge Failure for Pile in Sand............................. 82
Figure 3-24 Assumed Mode of Soil Failure by Lateral Flow Around Pile in Sand, (a)
Section Though Pile, (b) Mohr-Coulomb Diagram................................................... 84
ix
Figure 3-25 Characteristic Shape of a Set of p-y Curves for Static and Cyclic Loading in
Sand ........................................................................................................................... 86
Figure 3-26 Values of Coefficients
and
........................................................................... 88
Figure 5-8 Interaction Diagram for Nominal Moment Capacity ............................................... 174
Figure 5-9 Example Pipe Section for Computation of Plastic Moment Capacity ..................... 175
Figure 5-10 Moment versus Curvature of Example Pipe Section ............................................. 175
Figure 5-11 Elasto-plastic Stress Distribution Computed by LPile........................................... 177
Figure 5-12 Stress-Strain Curves of Prestressing Strands Recommended by PCI Design
Handbook, 5th Edition.............................................................................................. 178
Figure 5-13 Sections for Prestressed Concrete Piles Modeled in LPile .................................... 180
Figure 6-1 Scheme for Installing Pile in a Slope Subject to Sliding.......................................... 185
Figure 6-2 Forces from Soil Acting Against a Pile in a Sliding Slope, (a) Pile, Slope, and
Slip Surface Geometry, (b) Distribution of Mobilized Forces, (c) Free-body
Diagram of Pile Below the Slip Surface.................................................................. 186
Figure 6-3 Influence of Stabilizing Pile on Factor of Safety Against Sliding ........................... 187
Figure 6-4 Matching of Computed and Assumed Values of hp ................................................. 189
Figure 6-5 Soil Conditions for Analysis of Slope for Low Water ............................................. 190
Figure 6-6 Preliminary Design of Stabilizing Piles ................................................................... 191
Figure 6-7 Load Distribution from Stabilizing Piles for Slope Stability Analysis .................... 192
xii
List of Tables
Table 3-1.
Stiff Clay (no longer recommended) ......................................................................... 63
Table 3-2. Representative Values of
50 .......................................................................................
65
50
Table 3-6. Representative Values of k for Submerged Sand for Static and Cyclic Loading ....... 89
Table 3-7. Representative Values of k for Sand Above Water Table for Static and Cyclic
Loading ...................................................................................................................... 89
Table 3-8. Results of Grout Plug Tests by Schmertmann (1977) .............................................. 115
Table 3-9. Values of Compressive Strength at San Francisco ................................................... 117
Table 5-1. LPile Output for Rectangular Concrete Section ....................................................... 166
Table 5-2. Comparison of Results from Hand Computation versus Computer Solution........... 173
xiii
Chapter 1
Introduction
Chapter 1 Introduction
Chapter 1 Introduction
6. Obtain curves showing maximum moment in the pile and lateral pile-head deflection versus
lateral shear loading and curves of lateral deflection, bending moment and shear force versus
depth along the pile.
7. Change the pile dimensions or pile type, if necessary and repeat the analyses until a range of
suitable pile types and sizes have been identified.
8. Identify the pile type and size for which the global factor of safety is adequate and the most
efficient cost of the pile and construction is estimate.
9. Compute behavior of pile under working loads.
Virtually none of the examples in this manual follow all steps indicated above. However,
in most cases, the examples do show the curves that are indicated in Step 6.
1-2-5 Suggestions for the Designing Engineer
As will be explained in some detail, there are five sets of boundary conditions that can be
employed; examples will be shown for the use of these different boundary conditions. However,
the manner in which the top of the pile is fastened to the pile cap or to the superstructure has a
significant influence on deflections and bending moments that are computed. The engineer may
be required to perform an analysis of the superstructure, or request that one be made, in order to
ensure that the boundary conditions at the top of the pile are satisfied as well as possible.
With regard to boundary conditions at the pile head, it is important to note the versatility
of LPile. For example, piles that are driven with an accidental batter or an accidental eccentricity
can be easily analyzed. It is merely necessary to define the appropriate conditions for the
analysis.
As noted earlier, selection of upper and lower bound values of soil properties is a
practical procedure. Parametric solutions are easily done and relatively inexpensive and such
solutions are recommended. With the range of maximum values of bending moment that result
from the parametric studies, for example, the insight and judgment of the engineer can be
improved and a design can probably be selected that is both safe and economical. Alternatively,
one may perform a first-order, second moment reliability analysis to evaluate variance in
performance for selected random variables. For further guidance on this topic, the reader is
referred to the textbook by Baecher and Christian (2003).
If the axial load is small or negligible, it is recommended to make solutions with piles of
various lengths. In the case of short piles, the mobilization shear force at the bottom of the pile
can be defined along with the soil properties. In most cases, the installation of a few extra feet of
pile length will add little cost to the project and, if there is doubt, a pile with a few feet of
additional length could possibly prevent a failure due to excessive deflection. If the base of the
pile is founded in rock, available evidence shows that often only a short socket will be necessary
to anchor the bottom of the pile. In all cases, the designer must assure that the pile has adequate
bending stiffness over its full length.
A useful activity for a designer is to use LPile to analyze piles for which experimental
results are available. It is, of course, necessary to know the appropriate details from the load
tests; pile geometry and bending stiffness, stratigraphy and soil properties, magnitude and point
of application of loading, and the type of loading (either static or cyclic). Many such experiments
have been run in the past. Comparison of the results from analysis and from experiment can yield
3
Chapter 1 Introduction
valuable information and insight to the designer. Some comparisons are provided in this
document, but those made by the user could be more site-specific and more valuable.
In some instances, the parametric studies may reveal that a field test is indicated. Such a
case occurs when a large project is planned and when the expected savings from an improved
design exceeds the cost of the testing. Savings in construction costs may be derived either by
proving a more economical foundation design is feasible, by permitting use of a lower factor of
safety or, in the case of a load and resistance factor design, use of an increased strength reduction
factor for the soil resistance.
There are two types of field tests. In one instance, the pile may be fully instrumented so
that experimental p-y curves are obtained. The second type of test requires no internal instrumentation in the pile but only the pile-head settlement, deflection, and rotation will be found as a
function of applied load. LPile can be used to analyze the experiment and the soil properties can
be adjusted until agreement is reached between the results from the computer and those from the
experiment. The adjusted soil properties can be used in the design of the production piles.
In performing the experiment, no attempt should be made to maintain the conditions at
the pile head identical to those in the design. Such a procedure could be virtually impossible.
Rather, the pile and the experiment should be designed so that the maximum amount of
deflection is achieved. Thus, the greatest amount of information can be obtained on soil
response.
The nature of the loading during testing; whether static, cyclic, or otherwise; should be
consistent for both the experimental pile and the production piles.
The two types of problems concerning the performance of pile groups of piles are
computation of the distribution of loading from the pile cap to a widely spaced group of piles and
the computation of the behavior of spaced-closely piles.
The first of these problems involves the solutions of the equations of structural mechanics
that govern the distribution of moments and forces to the piles in the pile group (Hrennikoff,
1950; Awoshika and Reese, 1971; Akinmusuru, 1980). For all but the most simple group
geometries, solution of this problem requires the use of a computer program developed for its
solution.
The second of the two problems is more difficult because less data from full-scale
experiments is available (and is often difficult to obtain). Some full-scale experiments have been
performed in recent years and have been reported (Brown, et al., 1987; Brown et al., 1988).
These and additional references are of assistance to the designer (Bogard and Matlock, 1983;
Focht and Koch, 1973;
, et al., 1977).
The technical literature includes significant findings from time to time on piles under
lateral loading. Ensoft will take advantage of the new information as it becomes available and
verified by loading testing and will issue new versions of LPile when appropriate. However, the
material that follows in the remaining sections of this document shows that there is an
opportunity for rewarding research on the topic of this document, and the user is urged to stay
current with the literature as much as possible.
Chapter 1 Introduction
Chapter 1 Introduction
a. Elevation View
Lumped masses
Foundation springs
b. Analytical Model
K33
K22
K33
K11
Rotation
c. Stiffness Matrix
Figure 1-1 Example of Modeling a Bridge Foundation
The stiffness K11 is the stiffness of the axial load-settlement curve for the axial load P.
This stiffness is obtained either from load test results or from a numerical analysis using an axial
capacity analysis program like Shaft or APile from Ensoft, Inc.
Chapter 1 Introduction
Axial
Py
x
Px
My
Mx
Axial Pile
Displacement, u
Mz P
z
Axial Soil
Reaction, q
Lateral
y
Torsional Pile
Displacement,
Lateral Soil
Reaction, p
Lateral Pile
Displacement, y
Torsional Soil
Reaction, t
(a) Three-dimensional
pile displacements
Torsional
Chapter 1 Introduction
P
M
M
V
K22 = V/
K23 = V/
and K32 = M/ .
and K33 = M/ .
Chapter 1 Introduction
Most analytical methods in structural mechanics can employ either the stiffness matrix or
the flexibility matrix to define the support condition at the pile head. If the user prefers to use the
stiffness matrix in the structural model, Figure 1-3 illustrates basic procedures used to compute a
stiffness matrix. The initial coefficients for the stiffness matrix may be defined based on the
magnitude of the service load. The user may need to make several iterations before achieving
acceptable agreement.
1-3-4 Concluding Comments
The correct modeling of the problem of the single pile to respond to axial and lateral
loading is challenging and complex, and the modeling of a group of piles is even more complex.
However, in spite of the fact that research is continuing, the following chapters will demonstrate
that usable solutions are at hand.
New developments in computer technology allow a complete solution to be readily
developed, including automatic generation of the nonlinear responses of the soil around a pile
and iteration to achieve force equilibrium and compatibility.
Chapter 1 Introduction
10
Chapter 2
Solution for Pile Response to Lateral Loading
2-1 Introduction
Many pile-supported structures will be subjected to horizontal loads during their
functional lifetime. If the loads are relatively small, a design can be made by building code
provisions that list allowable loads for vertical piles as a function of pile diameter and properties
of the soil. However, if the load per pile is large, the piles are frequently installed at a batter. The
analyst may assume that the horizontal load on the structure is resisted by components of the
axial loads on the battered piles. The implicit assumption in the procedure is that the piles do not
deflect laterally which, of course, is not true. Rational methods for the analysis of single piles
under lateral load, where the piles are vertical or battered, will be discussed herein, and methods
are given for investigating a wide variety of parameters. The problem of the analysis of a group
of piles is discussed in another publication.
As a foundation problem, the analysis of a pile under lateral loading is complicated
because the soil reaction (resistance) at any point along a pile is a function of pile deflection. The
pile deflection, on the other hand, is dependent on the soil resistance; therefore, solving for the
response of a pile under lateral loading is one of a class of soil-structure-interaction problems.
The conditions of compatibility and equilibrium must be satisfied between the pile and soil and
between the pile and the superstructure. Thus, the deformation and movement of the
superstructure, ranging from a concrete mat to an offshore platform, and the manner in which the
pile is attached to the superstructure, must be known or computed in order to obtain a correct
solution to most problems.
2-1-1 Influence of Pile Installation and Loading on Soil Characteristics
2-1-1-1 General Review
The most critical factor in solving for the response of a pile under lateral loading is the
prediction of the soil resistance at any point along a pile as a function of the pile deflection. Any
serious attempt to develop predictions of soil resistance must address the stress-deformation
characteristics of the soil. The properties to be considered, however, are those that exist after the
pile has been installed. Furthermore, the influence of lateral loading on soil behavior must be
taken into account.
The deformations of the soil from the driving of a pile into clay cause important and
significant changes in soil characteristics. Different but important effects are caused by driving
of piles into granular soils. Changes in soil properties are also associated with the installation of
bored piles. While definitive research is yet to be done, evidence clearly shows that the soil
immediately adjacent to a pile wall is most affected. Investigators (Malek, et al., 1989) have
suggested that the direct-simple-shear test can be used to predict the behavior of an axially
loaded pile, which suggests that the soil just next to the pile wall will control axial behavior.
However, the lateral deflection of a pile will cause strains and stresses to develop from the pile
11
wall to several diameters away. Therefore, the changes in soil characteristics due to pile
installation are less important for laterally loaded piles than for axially loaded piles.
The influence of the loading of the pile on soil response is another matter. Four classes of
lateral loading can be identified: short-term, repeated, sustained, and dynamic. The first three
classes are discussed herein, but the response of piles to dynamic loading is beyond the scope of
this document. The use of a pseudo-horizontal load as an approximation in making earthquakeresistant designs should be noted, however.
The influence of sustained or cyclic loading on the response of the soil will be discussed
in some detail in Chapter 3; however, some discussion is appropriate here to provide a basis for
evaluating the models that are presented in this chapter. If a pile is in granular soil or
overconsolidated clay, sustained loading, as from earth pressure, will likely cause only a
negligible amount of long-term lateral deflection. A pile in normally consolidated clay, on the
other hand, will experience long-term deflection, but, at present, the magnitude of such
deflection can only be approximated. A rigorous solution requires solution of the threedimensional consolidation equation stepwise with time. At some time, the pile-head will
experience an additional deflection that will cause a change in the horizontal stresses in the
continuum.
Methods have been developed, as reviewed later, for getting answers to the problem of
short-term loading by use of correlations between soil response and the in situ undrained strength
of clay and the inimportant because they can be used for sustained loading in some cases and because an initial
condition is provided for taking the influence of repeated loading into account. Experience has
shown that the loss of lateral resistance due to repeated loading is significant, especially if the
piles are installed in clay below free water. The clay can be pushed away from the pile wall and
the soil response can be significantly decreased. Predictions for the effect of cyclic loading are
given in Chapter 3.
Four general types of loading are recognized above and each of these types is further
discussed in the following sections. The importance of consideration and evaluation of loading
when analyzing a pile subjected to lateral loading cannot be overemphasized.
Many of the load tests described later in this chapter were performed by applying a lateral
load in increments, holding that load for a few minutes, and reading all the instruments that gave
the response of the pile. The data that were taken allowed p-y curves to be computed; analytical
expressions are developed from the experimental results and these expressions yield p-y curves
following section.
2-1-1-2 Static Loading
The static p-y curves can be thought of as backbone curves that can be correlated to some
extent with soil properties. Thus, the curves are useful for providing some theoretical basis to the
p-y method.
From the standpoint of design, the static p-y curves have application in the following
cases: where loadings are short-term and not repeated (probably not encountered); and for
sustained loadings, as in earth-pressure loadings, where the soil around the pile is not susceptible
to consolidation and creep (overconsolidated clays, clean sands, and rock).
12
As will be noted later in this chapter, the use of the p-y curves for repeated loading, a type
of loading that is frequently encountered in practice, will often yield significant increases in pile
deflection and bending moment. The engineer may wish to make computations with both the
static curves and with the repeated (cyclic) curves so that the influence of the loading on pile
response can be seen clearly.
2-1-1-3 Repeated Cyclic Loading
The full-scale field tests that were performed included repeated or cyclic loading as well as the
static loading described above. An increment of load was applied, the instruments were read, and
the load was repeated a number of times. In some instances, the load was forward and backward,
and in other cases only forward. The instruments were read after a given number of cycles and
the cycling was continued until there was no obvious increase in ground line deflection or in
bending moments. Another increment was applied and the procedure was repeated. The final
load that was applied brought the maximum bending moment close to the moment that would
cause the steel to yield plastically.
Four specific sets of recommendations for p-y curves for cyclic loading are described in
Chapter 3. For three of the sets, the recommendations that are given
case. That is, the data that were used to develop the p-y curves were from cases where the
ground-line deflection had substantially ceased with repetitions in loading. In the other case, for
stiff clay where there was no free water at the ground surface, the recommendations for p-y
curves are based on the number of cycles of load application, as well as other factors.
The presence of free water at the ground surface for clay soils can be significant in regard
to the loss of soil resistance due to cyclic loading (Long, 1984). After a deflection is exceeded
when the load is released. Free water moves into this space and on the next load application the
water is ejected bringing soil particles with it. This erosion causes a loss of soil resistance in
addition to the losses due to remolding of the soil as a result of the cyclic strains. At this point
the use of judgment in the design of the piles under lateral load should be emphasized. For
example, if the clay is below a layer of sand, or if provision could be made to supply sand around
the pile, the sand will settle around the pile, and probably restore the soil resistance that was lost
due to the cyclic loading.
Pile-supported structures are subjected to cyclic loading in many instances. Some
common cases are wind load against overhead signs and high-rise buildings, traffic loads on
bridge structures, wave loads against offshore structures, impact loads against docks and dolphin
structures, and ice loads against locks and dams. The nature of the loading must be considered
carefully. Factors to be considered are frequency, magnitude, duration, and direction. The
engineer will be required to use a considerable amount of judgment in the selection of the soil
parameters and response curves.
2-1-1-4 Sustained Loading
If the soil resisting the lateral deflection of a pile is overconsolidated clay, the influence
of sustained loading would probably be small. The maximum lateral stress from the pile against
the clay would probably be less than the previous lateral stress; thus, the additional deflection
due to consolidation and creep in the clay should be small or negligible.
13
If the soil that is effective in resisting lateral deflection of a pile is a granular material that
is freely-draining, the creep would be expected to be small in most cases. However, if the pile is
subjected to vibrations, there could be densification of the sand and a considerable amount of
additional deflection. Thus, the judgment of the engineer in making the design should be brought
into play.
If the soil resisting lateral deflection of a pile is soft, saturated clay, the stress applied by
the pile to the soil could cause a considerable amount of additional deflection due to
consolidation (if positive pore water pressures were generated) and creep. An initial solution
could be made, the properties of the clay could be employed, and an estimate could be made of
the additional deflection. The p-y curves could be modified to reflect the additional deflection
and a second solution obtained with the computer. In this manner, convergence could be
achieved. The writers know of no rational way to solve the three-dimensional, time-dependent
problem of the additional deflection that would occur so, again, the judgment and integrity of the
engineer will play an important role in obtaining an acceptable solution.
2-1-1-5 Dynamic Loading
Two types of problems involving dynamic loading are frequently encountered in design:
machine foundations and earthquakes. The deflection from the vibratory loading from machine
foundations is usually quite small and the problem would be solved using the dynamic properties
of the soil. Equations yielding the response of the structure under dynamic loading would be
employed and the p-y method described herein would not be employed.
With regard to earthquakes, a rational solution should proceed from the definition of the
free-field motion of the near-surface soil due to the earthquake. Thus, the p-y method described
herein could not be used directly. In some cases, an approximate solution to the earthquake
problem has been made by applying a horizontal load to the superstructure that is assumed to
reflect the effect of the earthquake. In such a case, the p-y method can be used but such solutions
would plainly be approximate.
2-1-2 Models for Use in Analyses of Single Piles
A number of models have been proposed for the pile and soil system. The following are
brief descriptions for a few of them.
2-1-2-1 Elastic Pile and Soil
The model shown in Figure 2-1(a) depicts a pile in an elastic soil. A model of this sort
has been widely used in analysis. Terzaghi (1955) gave values of subgrade modulus that can be
used to solve for deflection and bending moment, but he went on to qualify his
recommendations. The standard equation for a beam was employed in a manner that had been
suggested earlier by such writers as Hetenyi (1946). Terzaghi stated that the tabulated values of
subgrade modulus could not be used for cases where the computed soil resistance was more than
one-half of the bearing capacity of the soil. However, recommendations were not included for
the computation of the bearing capacity under lateral load, nor were any comparisons given
between the results of computations and experiments.
The values of subgrade moduli published by Terzaghi have proved to be useful and
provide evidence that Terzaghi had excellent insight into the problem. However, in a private
conversation with the senior writer, Terzaghi said that he had not been enthusiastic about writing
14
the paper and only did so in response to numerous requests. The method illustrated by Figure 21(a) serves well in obtaining the response of a pile under small loads, in illustrating the various
interrelationships in the response, and in giving an overall insight into the nature of the problem.
The method cannot be employed without modification in solving for the loading at which a
plastic hinge will develop in the pile.
(a)
(b)
Mt
Mt
Pt
Pt
(c)
(d)
Figure 2-1 Models of Pile Under Lateral Loading, (a) 3-Dimensional Finite Element Mesh, and
(b) Cross-section of 3-D Finite Element Mesh,
15
16
p
y
p
y
p
y
p
y
p
y
x
Figure 2-2 Model of Pile Under Lateral Loading and p-y Curves
An axial load is indicated and is considered in the solution with respect to its effect on
bending and not in respect to axial settlement. However, as shown later in this manual, the
computational procedure is such that it allows for the determination of the axial load at which a
pile will buckle.
The soil around the pile is replaced by a set of mechanisms that indicate that the soil
resistance p is a nonlinear function of pile deflection y. The mechanisms, and the corresponding
curves that represent their behavior, are widely spaced in the sketch, but are close together in the
analysis. As may be seen, the p-y curves are nonlinear with respect to depth x along the pile and
pile deflection y. The top p-y curve is drawn to indicate that the pile may deflect a finite distance
with no soil resistance. The second curve from the top is drawn to show that the soil resistance is
deflection softening. There is no reasonable limit to the variations in the resistance of the soil to
the lateral deflection of a pile.
As will be shown later, the p-y method is versatile and provides a practical means for
design. The method was first suggested by McClelland and Focht (1956). Two developments
development of digital computer programs for
17
y
(a)
(b)
Figure 2-3 Distribution of Stresses Acting on a Pile, (a) Before Lateral Deflection and (b) After
Lateral Deflection y
18
Integration of the unit stresses results in the quantity p which acts opposite in direction to
y. The dimensions of p are load per unit length of the pile. These definitions of p and y are
convenient in the solution of the differential equation and are consistent with those used in the
solution of the ordinary beam equation.
2-1-2-7 Comments on the p-y method
The most common criticism of the p-y method is that the soil is not treated as a
continuum, but as a series of discrete springs (the Winkler model). Several comments can be
given in response to this valid criticism.
The recommendations for the prediction of p-y curves for use in the analysis of piles,
given in a subsequent chapter, are based for the most part on the results of full-scale experiments,
lock (1970) performed some
tests of a pile in soft clay where the pattern of pile deflection was varied along its length. The p-y
curves that were derived from each of the loading conditions were essentially the same. Thus,
Matlock found that experimental p-y curves from fully instrumented piles could predict, within
reasonable limits, the response of a pile whose head is free to rotate or is fixed against rotation.
The methods for computing p-y curves derived from correlations to the results of fullscale experiments have been used to make computations for the response of piles where only the
pile-head movements were recorded. These computations, some of which are shown in Chapter 6
of
, show reasonable to excellent agreement between computed predictions
and experimental measurements.
Finally, technology may advance so that the soil resistance for a given deflection at a
particular point along a pile can be modified quantitatively to reflect the influence of the
deflection of the pile above and below the point in question. In such a case, multi-valued p-y
curves can be developed at every point along the pile. The analytical solution that is presented
herein could be readily modified to deal with the multi-valued p-y curves.
In short, the p-y method has some limitations; however, there is much evidence to show
that the method yields information of considerable value to an analyst and designer.
2-1-3 Computational Approach for Single Piles
The general procedure to be used in computing the behavior of many piles under lateral
loading is illustrated in Figure 2-4. Figure 2-4 (a) shows a pile with a given geometry embedded
in a soil with known characteristics. A lateral load Pt, axial load Q, and moment M are acting at
the pile head. The loading presumably would have been found by considering the unfactored
loading on the superstructure. Each of the loads is decreased or increased by the same multiplier
and, for each combination of loads, a solution of the problem is found. A curve can be plotted,
such as shown by the solid line in Figure 2-4 (b), which will show the maximum bending
moment at some point along the pile as a function of the loading. With the value of the nominal
bending moment capacity Mnom for the section that takes into account the axial loading, the
ssumption is made that a plastic hinge at any point along
the length of the pile would not be tolerable. The failure loading is then divided by a global
factor of safety to find the allowable loading. The allowable loading is then compared to the
loading from the superstructure to determine if the pile that was selected was satisfactory.
19
Q
M
Loading
Pt
Loading at Failure
Mult
Allowable
Loading
(a)
(b)
Figure 2-4 Illustration of General Procedure for Selecting a Pile to Sustain a Given Set of Loads
An alternate approach makes use of the concept of partial safety factors. The parameters
that influence the resistance of the pile to lateral loading are factored and the curve shown by the
dashed line is computed. As shown in Figure 2-4, smaller values of the failure loading would be
found. The values of allowable loading would probably be about the same as before with the
loading being reduced by a smaller value of partial safety factor.
In the case of a very short pile, the performance failure might be due to excessive
design engineer can then employ a global
factor of safety or partial factors of safety to set the allowable load capacity.
As shown in Figure 2-4(b), the bending moment is a nonlinear function of load; therefore,
the use of allowable bending stresses, for example, is inappropriate and perhaps unsafe. A series
of solutions is necessary in order to obtain the allowable loading on a pile; therefore, the use of a
computer is required.
The next step in the computational process is to solve for the deflection of the pile under
the allowable loading. The tolerable deflection is frequently limited by special project
requirements and probably should not be dictated by building codes or standards. Among factors
to be considered are machinery that is sensitive to differential deflection and the comfort of
humans on structures that move a sensible amount under loading.
The computation of the load at failure requires values of the nominal bending moment
capacity and flexural rigidity of the section. Because the analyses require the structural section to
be stressed beyond the linear-elastic range, a computer program is required to compute the
nonlinear properties of the section. These capabilities are included in the LPile program.
General guidelines about making computations for the behavior of a pile under lateral
loading are presented in this manual. In addition, several examples are presented in detail.
20
However, it should be emphasized that the material presented herein is only a valuable tool for
the designer and that a complete design involves many other factors that are not addressed here.
2-1-3-1 Study of Pile Buckling
A second computational problem is shown in Figure 2-5. A pile that extends above the
ground line is subjected to a lateral load Pt and an axial load Q, as shown in Figure 2-5(a). The
engineer desires to solve for the axial load that will cause the pile to buckle. The lateral load is
held constant and the axial load is increased in increments. The deflection yt at the top of the pile
is plotted as a function of axial load, as shown in Figure 2-5(b). A value of axial load will be
approached at which the pile-head deflection will increase without limit. This load is selected for
the buckling load. It is important that the buckling load be found by starting the computer runs
with smaller values of axial load because the computer program fails to obtain a solution at axial
loads above the buckling load. An example analysis of pile buckling is presented in Section 4-4.
Q
yt
Pt
Buckling Load
yt
(b)
(a)
21
designer will normally select a pile for a particular application whose length is somewhat greater
than Lcrit.
Q
M
yt
Pt
Lcrit
L
Lcrit
Pile Length
22
h = 6.1 m
d = 838 mm
Ic = 5.876 x 10-3 m4
4m
V
V
d = 762 mm
Ip = 3.07 x 10-3 m4
E = 2 x 108 kPa
(a)
(b)
(c)
The velocity of the vessel is v and its energy on contact would be mv2. The deflection of the
pile could be computed by finding the area under the load-deflection curve that would equate to
the energy of the vessel.
The analyst would be concerned with a number of parameters in the problem. The level
of water could vary, requiring a number of solutions. The pile could be tapered to give it the
proper strength to sustain the computed bending moment while at the same time making it as
flexible as possible.
With the first impact of a vessel, the soil will behave as if it were under static loading
(assuming no inertia effects in the soil) and would be relatively stiff. With repeated loading on
the pile from berthing, the soil will behave as if under cyclic loading. The appropriate p-y curves
would need to be used, depending on the number of applications of load.
A single pile, or a group of piles, could support the primary fenders, but the exact types
and sizes of cushions or fenders to be used between the vessel and the pile need to be selected on
the basis of the vessel size and berthing velocity. It should be noted that fenders must be
mounted properly above the waterline to prevent damage to the berthing vessels.
m, v
Breasting
Dolphin
Deflection
24
The braking forces are shown properly in a plane parallel to the axis of the bridge and can
be large, if heavily loaded trucks are suddenly brought to a stop on a downward-sloping span.
The deflection that may be possible in the direction of the axis of the bridge is probably limited
to that allowed by the joints in the bridge deck. Thus, one of the boundary conditions for the
piles for such loading could be a limiting deflection.
If it is decided that significant loads can be acting simultaneously in perpendicular planes,
two independent solutions can be made, and the resulting bending moments can be added
algebraically. Such a procedure would not be perfectly rigorous but should yield results that will
be instructive to the designer.
Loads From Traffic
Loads From Braking
and Wind Forces
25
between the piles and the cap may be such that the pile heads are essentially free to rotate.
Alternatively, the design analysis may be made assuming that the pile heads are fixed against
rotation.
Wind
Load
Wind
Load
Column
Dead Load
Pile
Cap
Column
Dead Load
Two-Shaft
Foundation
(a)
Single-Shaft
Foundation
(b)
26
27
waterfront structures such as piers and quay walls; support for overhead pipes and for other
facilities found in industrial plants; and bridge abutments.
The method has the potential of analyzing the flexible bulkhead that is shown in Figure
2-12. The sheet piles (or tangent piles if bored piles are used) can be analyzed as a pile, if the p-y
curves are modified to reflect the soil resistance versus deflection for a wall, rather than for a
pile. Research on the topic has been undertaken (Wang, 1986) and has already been implemented
in the computer program PYWall from Ensoft, Inc.
Tie-back
Anchor Pile
(Dead Man)
28
y
x
Px
Vn
Vv
Vv
dx
Vv+dVv
y+dy
M+dM
Px
x
Figure 2-13 Element of Beam-Column (after Hetenyi, 1946)
. ....................................................(2-2)
Differentiating Equation 2-2 with respect to x, the following equation is obtained
.................................................(2-3)
The following definitions are noted:
29
Q sin S ..................................................(2-5)
Because S is usually small, we may assume the small angle relationships cos S = 1 and sin S =
tan S = dy/dx. Thus, Equation 2-6 is obtained.
..........................................................(2-6)
Vn will mostly be used in computations, but Vv can be computed from Equation 2-6 where
dy/dx is equal to the rotation S.
The ability to allow a distributed force W per unit of length along the upper portion of a
pile is convenient in the solution of a number of practical problems. The differential equation
then becomes as shown below.
..............................................(2-7)
where:
Q = axial thrust load in the pile,
y
= lateral deflection of the pile at a point x along the length of the pile,
...........................................................(2-9)
and,
30
.............................................................(2-10)
where
V = shear in the pile,
M = bending moment in the pile, and
S = slope of the elastic curve defined by the axis of the pile.
Except for the axial load Q, the sign conventions that are used in the differential equation
and in subsequent development are the same as those usually employed in the mechanics for
beams, with the axes for the pile rotated 90 degrees clockwise from the axes for the beam. The
axial load Q does not normally appear in the equations for beams. The sign conventions are
presented graphically in Figure 2-14. A solution of the differential equation yields a set of curves
such as shown in Figure 2-15. The mathematical relationships for the various curves that give the
response of the pile are shown in the figure for the case where no axial load is applied.
Slope (L/L)
Deflection (L)
y(+)
S (+)
Moment (F*L)
y
M (+)
x
Q (+)
Axial Force (F)
Shear (F)
V (+)
p (+)
5. The modulus of elasticity of the pile material is the same in tension and compression,
6. Transverse deflections of the pile are small,
7. The pile is not subjected to dynamic loading, and
8. Deflections due to shearing stresses are small.
Assumption 8 can be addressed by including more terms in the differential equation, but
errors associated with omission of these terms are usually small. The numerical method
presented later can deal with the behavior of a pile made of materials with nonlinear stress-strain
properties.
y
......................................................(2-12)
The solution to Equation 2-12 may be directly written as:
..........................................(2-13)
The coefficients C1, C2, C3, and C4 must be evaluated for the various boundary conditions that are
desired. A pile of any length is considered later but, if one considers a long pile, a simple set of
equations can be derived. An examination of Equation 2-13 shows that C1 and C2 must approach
zero because the term e x will increase without limit.
The boundary conditions for the top of the pile that are employed for the solution of the
reduced form of the differential equation are shown by the simple sketches in Figure 2-16. A
more complete discussion of boundary conditions for a pile is presented in the next section.
Spring (takes no shear, but
restrains pile head rotation)
Mt
Pt
Free-head
(a)
Pt
Fixed-Head
Pt
Partially Restrained
(b)
(c)
..........................................................(2-14)
33
...........................................................(2-15)
The differentiations of Equation 2-13 are made and the substitutions indicated by Equation 2-14
yield the following.
.........................................................(2-16)
The substitutions indicated by Equation 2-15 yield the following.
.....................................................(2-17)
Equations 2-16 and 2-17 are used and expressions for deflection y, slope S, bending moment M,
shear V, and soil resistance p can be written as shown in Equations 2-18 through 2-22.
................................(2-18)
............................(2-19)
.................................(2-20)
.................................(2-21)
.............................(2-22)
It is convenient to define some functions that make it easier to write the above equations.
These are:
A1 = e
B1 = e
( cos x
sin x) ..............................................(2-24)
C1 = e
cos x ......................................................(2-25)
D1 = e
sin x ......................................................(2-26)
34
................................................(2-27)
...............................................(2-28)
.....................................................(2-29)
V = PtB1
p =
2Mt D1 ..................................................(2-30)
Values for A1, B1, C1, and D1, are shown in Figure 2-17 as a function of the nondimensional
distance x along the pile.
A1, B1, C1, D1
-0.4
0.0
-0.2
0.2
0.4
0.6
0.8
0.5
1.0
1.5
2.0
2.5
A1
B1
3.0
C1
D1
3.5
4.0
4.5
5.0
5.5
6.0
For a pile whose head is fixed against rotation, as shown in Figure 2-16(b), the solution
may be obtained by employing the boundary conditions as given in Equations 2-32 and 2-33.
At x = 0,
.............................................................(2-32)
.........................................................(2-33)
Using the procedures as for the case where the boundary conditions were as shown in
Figure 2-4(a), the results are as follows.
.....................................................(2-34)
The solution for long piles is given in Equations 2-35 through 2-39.
..........................................................(2-35)
......................................................(2-36)
.........................................................(2-37)
V = Pt C1 ...........................................................(2-38)
p = Pt A1 ..........................................................(2-39)
It is sometimes convenient to have a solution for a third set of boundary conditions
describing the rotational restraint of the pile head, as shown in Figure 2-16(c). For this boundary
condition, the rotational spring does not take any shear, but does restrain the rotation of the pile
head. These boundary conditions are given in Equations 2-40 and 2-41. At the pile head, where x
= 0, the rotational restrain is controlled by
........................................................(2-40)
36
Employing these boundary conditions, the coefficients C3 and C4 can be evaluated, and the
results are shown in Equations 2-42 and 2-43. For convenience in writing, the rotational restraint
Mt /St is given the symbol k .
...................................................(2-42)
...................................................(2-43)
These expressions can be substituted into Equation 2-13, differentiation performed as
appropriate, and substitution of Equations 2-23 through 2-26 will yield a set of expressions for
the long pile similar to those in Equations 2-27 through 2-31 and 2-35 through 2-39.
Timoshenko (1941)
L is
greater than 4; however, there are occasions when the solution of the reduced differential
equation is desired for piles that have a nondimensional length less than 4. The solution can be
obtained by using the following boundary conditions at the tip of the pile. At x = L,
(M is zero at pile tip)...........................................(2-44)
and
(shear force, V, is zero at pile tip).................................(2-45)
When the above boundary conditions are used, along with a set for the top of the pile, the
four coefficients C1, C2, C3, and C4 can be evaluated. The solutions are not shown here, but new
values of the parameters A1, B1, C1, and D1 can be computed as a function of L. Such
computations, if carried out, will show readily the influence of the length of the pile.
The reduced form of the differential equation will not normally be used for the solution
of problems encountered in design; however, the influence of pile length and other parameters
can be illustrated with clarity. Furthermore, the closed-form solution can be used to check the
accuracy of the numerical solution shown in the next section.
2-2-3 Solution by Finite Difference Equations
The solution of Equation 2-7 is necessary for dealing with numerous problems that are
encountered in practice. The formulation of the differential equation in finite difference form and
a solution by iteration mandates a computer program. In addition, the following improvements in
the solutions shown in the previous section are then possible.
The effect of the axial load on deflection and bending moment can be considered and
problems of pile buckling can be solved.
The bending stiffness EI of the pile can be varied along the length of the pile.
37
Perhaps of more importance, the soil modulus Es can vary with pile deflection and with the
depth of the soil profile.
Soil displacements around the pile due to slope movements, seepage forces, or other causes
can be taken into account.
In the finite difference formulations, the derivative terms are replaced by algebraic
expressions. The following central difference expressions have errors proportional to the square
of the increment length h.
d4y
dx 4
ym
4 ym
6 ym 4 ym
h4
ym
If the pile is subdivided in increments of length h, as shown in Figure 2-18, the governing
differential equation, Equation 2-7, in difference form with collected terms for y is as follows:
y
ym+2
h
h
ym+1
ym
ym-1
ym-2
x
Figure 2-18 Representation of deflected pile
38
..................................(2-46)
where
Rm = EmIm (flexural rigidity of pile at point m) and
km = Esm.
The assumption is implicit in Equation 2-46 that the magnitude of Q is constant with
depth. Of course, that assumption is not strictly true. However, experience has shown that the
maximum bending moment usually occurs a relatively short distance below the ground line at a
point where the value of Q is undiminished. This fact plus the fact that Q, except in cases of
buckling, has little influence on the magnitudes of deflection and bending moment, leads to the
conclusion that the assumption of a constant Q is generally valid. For the reasons given, it is
thought to be unnecessary to vary Q in Equation 2-46; thus, a table of values of Q as a function
of x is not required.
If the pile is divided into n increments, n+1 equations of the sort as Equation 2-46 can be
written. There will be n+5 unknowns because two imaginary points will be introduced above the
top of the pile and two will be introduced below the bottom of the pile. If two equations giving
boundary conditions are written at the bottom and two at the top, there will be n+5 equations to
solve simultaneously for the n+5 unknowns. The set of algebraic equations can be solved by
matrix methods in any convenient way.
The two boundary conditions that are employed at the bottom of the pile involve the
moment and the shear. If the possible existence of an eccentric axial load that could produce a
moment at the bottom of the pile is discounted, the moment at the bottom of the pile is zero. The
assumption of a zero moment is believed to produce no error in all cases except for short rigid
piles that carry their loads in end bearing, and when the end bearing is applied eccentrically. (The
case where the moment at the bottom of a pile is not equal to zero is unusual and is not treated by
the procedure presented herein.) Thus, the boundary equation for zero moment at the bottom of
the pile requires
.....................................................(2-47)
where y0 denotes the lateral deflection at the bottom of the pile. Equation 2-47 is expressing the
condition that EI(d2y/dx2) = 0 at x = L (The numbering of the increments along the pile starts
with zero at the bottom for convenience).
The second boundary condition involves the shear force at the bottom of the pile. The
assumption is made that soil resistance due to shearing stress can develop at the bottom of a short
pile as deflection occurs. It is further assumed that information can be developed that will allow
39
V0, the shear at the bottom of the pile, to be known as a function of y0 Thus, the second equation
for the zero-shear boundary condition at the bottom of the pile is
...............................(2-48)
Equation 2-48 is expressing the condition that there is some shear at the bottom of the pile or that
EI(d3y/dx3) + Q(dy/dx) = V0 at x = L. The assumption is made in these equations that the pile
carries its axial load in end-bearing only, an assumption that is probably satisfactory for short
piles for which V0 would be important. The value of V0 should be set equal to zero for long piles
(2 or more points of zero deflection along the length of the pile).
As noted earlier, two boundary equations are needed at the top of the pile. Four sets of
boundary conditions, each with two equations, have been programmed. The engineer can select
the set that fits the physical problem.
Case 1 of the boundary conditions at the top of the pile is illustrated graphically in Fig 219. (The axial load Q is not shown in the sketches, but Q is assumed to be acting at the top of the
pile for each of the four cases of boundary conditions.). For the condition where the shear at the
top of the pile is equal to Pt, the following difference equation is employed.
Pt
+Mt
+Pt
yt+2
yt+1
yt
yt-1
yt-2
.........................(2-49)
For the condition where the moment at the top of the pile is equal to Mt, the following difference
equation is employed.
.............................................(2-50)
Case 2 of the boundary conditions at the top of the pile is illustrated graphically in Figure
2-20. The pile is assumed to be embedded in a concrete foundation for which the rotation is
known. In many cases, the rotation can be assumed to be zero, at least for the initial solutions.
40
Equation 2-49 is the first of the two equations that are needed. The second of the two needed
equations reflects the condition that the slope St at the top of the pile is known.
yt+2
yt+1
yt
+Pt
St
yt-1
yt-2
yt+2
yt+1
yt
+Pt
yt-1
yt-2
41
..............................................(2-52)
Case 4 of the boundary conditions at the top of the pile is illustrated in Figure 2-22. It is
assumed, for example, that a pile is embedded in a bridge abutment that moves laterally a given
amount; thus, the deflection yt at the top of the pile is known. It is further assumed that the
bending moment is known. If the embedment amount is small, the bending moment is frequently
assumed to be zero. The first of the two equations expresses the condition that the moment Mt at
the pile head is known, and Equation 2-50 can be employed. The second equation merely
expresses the fact that the pile-head deflection is known.
yt = Yt..............................................................(2-53)
Foundation
moves laterally
yt+2
Mt
yt+1
yt
yt-1
Pile-head moment is
known, may be zero
yt-2
42
in the model for the superstructure. Equation 2-53 can be used with a known value of yt and
Equation 2-51 can be used with a known value of St.
St
yt
yt+2
yt+1
yt
1
yt-1
yt-2
St
43
.............................................(2.54)
In building up the higher ordered terms by differentiation, the value of R is made to
correspond to the central term for y in the second-order expression. The errors that are involved
in using the above approximation where there is a change in the bending stiffness along the
length of a pile are thought to be small, but may be investigated as necessary.
44
Chapter 3
Lateral Load-Transfer Curves for Soil and Rock
3-1 Introduction
This chapter presents the formulation of expressions for p-y curves for soil and rock
under both static and cyclic loading. As part of this presentation, a number of fundamental
concepts are presented that are relevant to any method of analyzing deep foundations under
lateral loading. Chapter 1 presented the concept of the p-y method, and this chapter will present
details for the computation of load-transfer behavior for a pile under a variety of conditions.
A typical p-y curve is shown in Figure 3-1a. The p-y curve is just one of a family of p-y
curves that describe the lateral-load transfer along the pile as a function of depth and of lateral
deflection. It would be desirable if soil reaction could be found analytically at any depth below
the ground surface and for any value of pile deflection. Factors that might be considered are pile
geometry, soil properties, and whether the type of loading, static is cyclic, sustained, or dynamic.
Unfortunately, common methods of analysis are currently inadequate for solving all possible
problems. However, principles of geotechnical engineering can be helpful in gaining insight into
the evaluation of two characteristic portions of a p-y curve.
b
p
c
Pile Deflection, y
y
(b)
(a)
y
(c)
The p-y curve in Figure 3-1(a) is meant to represent the case where a short-term
convenience and will seldom, if ever, be encountered in practice. However, the static loading
curve is useful because analytical procedures can be used to develop expressions to correlate
with some portions of the curve, and the static curve serves as a baseline for demonstrating the
effects of other types of loading.
The three curves in Figure 3-1 show a straight-line relationship between p and y from the
origin to point a. If it can be reasonably assumed that for small strains in soil there is a linear
relationship between p and y for small values of y. Analytical methods for computing the slopes
of the initial portion of the p-y curves, Esi, are discussed later.
Recommendations will be given in this chapter for the selection of the slope of the initial
portion of p-y curves for the various cases of soils and loadings that are addressed. The point
should be made, however, that the recommendations for the slope of the initial portion are meant
to be somewhat conservative because the deflection and bending moment of a pile under light
loads will probably be somewhat less than computed by use of the recommendations. There are
some cases in the design of piles under lateral loading when it will be unconservative to compute
more deflection than will actually occur; in such cases, a field load test must be made.
The portion of the curve in Figure 3-1(a) from points a to b shows that the value of p is
strain softening with respect to y. This behavior is reflecting the nonlinear portion of the stressstrain curve for natural soil. Currently, there are no accepted analytical procedures that can be
used to compute the a-b portion of a p-y curve. Rather, that portion of the curves is empirical and
based on results of full-scale tests of piles in a variety of soils with both monotonic and cyclic
loading.
The horizontal, straight-line portion of the p-y curve in Figure 3-1(a) implies that the soil
is behaving plastically with no loss of shear strength with increasing strain. Using this
assumption, some analytical models can be used to compute the ultimate resistance pu as a
function of pile dimensions, soil properties, and depth below the ground surface. One part of a
model is for soil resistance near the ground surface and assumes that at failure the soil mass
moves vertically and horizontally. The other part of the model is for the soil resistance deep
below the ground surface and assumes only horizontal movement of the soil mass around the
pile.
Figure 3-1(b) shows a shaded portion of the curve in Figure 3-1(a). The decreasing values
of p from point c to point d reflect the effects of cyclic loading. The curves in Figures 3-1(a) and
3-1(b) are identical up to point c, which implies that the soil behaves identically for both type of
loading at small deflections. The loss of resistance shown by the shaded area depends on the
number of cycles of loading.
A possible effect of sustained, long-term loading is shown in Figure 3-1(c). This figure
shows that there is a time-dependent increase in deflection with sustained loading. The
decreasing value of p implies that the resistance is shifted to other elements of soil along the pile
as the deflection occurs at some particular point. The effect of sustained loading should be
negligible for heavily overconsolidated clays and for granular soils. The effect for soft clays
must be approximated at present.
46
The experimental p-y curves can be plotted once multiple of curves showing the
distribution of deflection and soil resistance for multiple levels of loading have been developed.
A check can be made of the accuracy of the analyses by using the experimental p-y curves to
compute bending-moment curves versus depth. The computed bending moments should agree
closely with those measured in the load test. In addition, computed values of pile-head slope and
deflection can be compared to the values measured during the load test. Usually, it is more
difficult to obtain agreement between computations and measurement of pile-head deflection and
slope over the full range of loading than for bending moment.
Examples of p-y curves that were obtained from a full-scale experiment with pipe piles
with a diameter of 641 mm (24 in.) and a penetration of 15.2 m (50 ft) are shown in Figures 3-2
and 3-3 (Reese et al., 1975) . The piles were instrumented for measurement of bending moment
at close spacing along the length and were tested in overconsolidated clay.
3,000
x = 12"
x = 24"
x = 36"
2,500
x = 48"
x = 60"
x = 72"
x = 96"
2,000
x = 120"
1,500
1,000
500
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Deflection, y, inches
Figure 3-2 p-y Curves from Static Load Test on 24-inch Diameter Pile (Reese, et al. 1975)
48
3,000
x = 12"
x = 24"
x = 36"
x = 48"
2,500
x = 60"
x = 72"
x = 84"
x = 96"
x = 108"
2,000
x = 120"
1,500
1,000
500
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Deflection, y, inches
Figure 3-3 p-y Curves from Cyclic Load Tests on 24-inch Diameter Pile (Reese, et al. 1975)
3-2-3 Nondimensional Methods for Obtaining Soil Response
Reese and Cox (1968) described a method for obtaining p-y curves for cases where only
pile-head measurements are made during lateral loading. They noted that nondimensional curves
could be obtained for many variations of soil modulus with depth. Equations for the soil modulus
involving two parameters were employed, such as shown in Equations 3-1 and 3-2.
Es = k1 + k2 x, .........................................................(3-1)
or
Es = k1 xn .............................................................(3-2)
Measurements of pile-head deflection and rotation at the ground line are necessary. Then,
either of the equations is selected and the two parameters are computed for a given applied load
and moment. With an expression for soil modulus for a particular load, the soil resistance and
deflection along the pile are computed.
49
The procedure is repeated for each of the applied loadings. While the method is
approximate, the p-y curves computed in this fashion do reflect the measured behavior of the pile
head. Soil response derived from a sizable number of such experiments can add significantly to
the existing information.
As previously indicated, the major field experiments that have led to the development of
the current criteria for p-y curves have involved the acquisition of experimental moment curves.
However, nondimensional methods of analyses, as indicated above, have assisted in the
development of p-y curves in some instances.
In the remaining portion of this chapter, details are presented for developing p-y curves
for clays and for sands. In addition, some discussion is presented for producing p-y curves for
other types of soil.
0.65
b
Ei b 4
Ep I p
50
1 / 12
Ei
1
..........................................(3-3)
Ei /c
0
100
200
300
Manor Road
Lake Austin
12
51
200
400
600
800
Pile 1 Static
0.6
1.2
1.8
2.4
Pile 2 (Cyclic)
3.0
Two analyses are used to gain some insight into the ultimate lateral resistance pu that
develop near the ground surface in one case and at depth in the other case. The first analysis is
for values of ultimate lateral resistance near the ground surface and considers the resistance a
passive wedge of soil displaced by the pile. The second analysis is for values of lateral resistance
well beneath the ground surface and models the plane-strain (flow-around) behavior of the soil.
52
The first analytical model for clay near the ground surface is shown in Figure 3-6. Some
justification can be presented for making use of a model that assumes that the ground surface
will move upward. Contours of the measured rise of the ground surface during a lateral load test
are shown in Figure 3-7. The p-y curves for the overconsolidated clay in which the pile was
tested are shown in Figures 3-3 and 3-4. As shown in Figure 3-7(a) for a load of 596 kN (134
kips), the ground-surface moved upward out to a distance of about 4 meters (13 ft) from the axis
of the pile. After the load was removed from the pile, the ground surface subsided to the profile
as shown in Figure 3-7 (b).
Ft
Ft
Ff
Fn
Fp
Fs
(a)
(b)
Figure 3-6 Assumed Passive Wedge Failure in Clay Soils, (a) Shape of Wedge,
(b) Forces Acting on Wedge
The use of plane sliding surfaces, shown in Figure 3-6, will obviously not model the
movement that is indicated by the contours in Figure 3-7; however, a solution with the simplified
model should give some insight into the variation of the ultimate lateral resistance pu with depth.
Summing the forces in the vertical direction yields
Fn sin
= W + Fs cos
+ 2 Ft cos
+ Ff
.................................(3-4)
where
= angle of the inclined plane with the vertical, and
W = the weight of the wedge.
An expression for W is
........................................................(3-5)
53
25 mm
19 mm
3 mm
6 mm
596 kN
13 mm
3 mm
6 mm
0 kN
13 mm
Scale, meters
Figure 3-7 Measured Profiles of Ground Heave Near Piles Due to Static Loading,
(a) Heave at Maximum Load, (b) Residual Heave
where
= unit weight of soil,
b = width (diameter) of pile, and
H = depth of wedge.
The resultant shear force on the inclined plane Fs is
........................................................(3-6)
where
ca = average undrained shear strength of the clay over depth H.
The resultant shear force on a side plane is
54
........................................................(3-7)
The frictional force between the wedge and the pile is
...........................................................(3-8)
where
= a reduction factor.
The above equations are solved for Fp, and Fp is differentiated with respect to H to solve
for the soil resistance pc1 per unit length of the pile.
.................(3-9)
The value of can be set to zero with some logic for the case of cyclic loading because
one can reason that the relative movement between pile and soil would be small under repeated
loads. The value of can be taken as 45 degrees, if the soil is assumed to behave in an undrained
mode. With these assumptions, Equation 3-9 becomes
............................................(3-10)
However, Thompson (1977) differentiated Equation 3-9 with respect to H and evaluated
the integrals numerically. His results are shown in Figure 3-8 with the assumption that the value
of the term /ca is negligible. The cases where is assumed to be zero and where is assumed
1.0 are shown in the figure. Also shown in Figure 3-8 is a plot of Equation 3-10 with the same
assumption with respect to /ca. As shown, the differences in the plots are not great. The curve in
Figure 3-8 from Hansen (1961a, 1961b) is discussed on page 56.
The equations developed above do not address the case of tension in the pile. If piles are
designed for a permanent uplift force, the equation for ultimate soil resistance should be
modified to reflect the effect of an uplift force at the face of the pile (Darr, et al., 1990).
The second of the two models for computing the ultimate resistance pu is shown in the
plan view in Figure 3-9(a). At some point below the ground surface, the maximum value of soil
resistance will occur with the soil moving horizontally. Movement in only one side of the pile is
indicated; but movement, of course, will be around both sides of the pile. Again, planes are
assumed for the sliding surfaces with the acceptance of some approximation in the results.
A cylindrical pile is indicated in the figure, but for ease in computation, a prismatic block
of soil is subjected to horizontal movement. Block 5 is moved laterally as shown and stress of
sufficient magnitude is generated in that block to cause failure. Stress is transmitted to Block 4
and on around the pile to Block 1, with the assumed movements indicated by the dotted lines.
Block 3 is assumed not to distort, but failure stresses develop on the sides of the block as it
slides.
55
p u /cb
0
10
15
20
25
30
0
Hansen
K = 0 Thompson
K = 0.5 Thompson
K = 1.0 Thompson
Eq. 3-10
1
2
3
4
H/b 5
6
7
8
9
10
56
(a)
2c
(b)
cb/2
6b
pu
1b
cb/2
(c)
Figure 3-9 Assumed Mode of Soil Failure Around Pile in Clay, (a) Section Through Pile,
(b) Mohr-Coulomb Diagram, (c) Forces Acting on Section of Pile
................................................(3-12)
Equations 3-10 and 3-11 are similar to Equations 3-20 and 3-21, shown later, that are
used in the recommendations for two of the sets of p-y curves. However, the emphasis was
placed directly on experimental results. The values of pu obtained in the full-scale experiments
were compared to the analytical values, and empirical factors were found by which Equations 310 and 3-11 could be modified. The adjustment factors that were found are shown in Figure 3-10
(see Section 3-3-7 on page 64 for more discussion), and it can be seen that the experimental
values of ultimate resistance for overconsolidated clay below the water table were far smaller
than the computed values. The recommended method of computing the p-y curves for such clays
is demonstrated later.
Ac and As
0
0.2
0.4
0.6
0.8
1.0
Ac
As
performed on piles with diameters of 273 mm (10.75 in.), 1,220 mm (48 in.), and 1,830 mm (72
in.). They found that the site-specific response of the soil could best be characterized by a
nonlinear function of the diameter.
There is good reason to believe that the diameter of the pile should not appear as a linear
function when piles in clays below the water table are subjected to cyclic loading. However, data
from experiments are insufficient at present to allow general recommendations to be made. The
influence of cyclic loading on p-y curves is discussed in the next section.
3-3-4 Influence of Cyclic Loading
Cyclic loading is specified in a number of the examples presented in Chapter 1; a notable
example is an offshore platform. Therefore, a number of the field tests employing fully
instrumented piles have employed cyclic loading in the experimental procedures. Cyclic loading
has invariably resulted in increased deflection and bending moment above the respective values
obtained in short-term loading. A dramatic example of the loss of soil resistance due to cyclic
loading may be seen by comparing the two sets of p-y curves in Figures 3-2 and 3-3.
Wang (1982) and Long (1984) did extensive studies of the influence of cyclic loading
on p-y curves for clays. Some of the results of those studies were reported by Reese, et al.
(1989). The following two reasons can be suggested for the reduction in soil resistance from
cyclic loading: the subjection of the clay to repeated strains of large magnitude, and scour from
the enforced flow of water near the pile. Long (1984) studied the first of these factors by
performing some triaxial tests with repeated loading using specimens from sites where piles had
been tested. The second of the effects is present when water is above the ground surface, and its
influence can be severe.
Welch and Reese (1972) report some experiments with a bored pile under repeated lateral
loading in overconsolidated clay with no free water present. During the cyclic loading, the
deflection of the pile at the ground line was in the order of 25 mm (1 in.). After a load was
released, a gap was revealed at the face of the pile where the soil had been pushed back. In
addition, cracks a few millimeters in width radiated away from the front of the pile. Had water
covered the ground surface, it is evident that water would have penetrated the gap and the cracks.
With the application of a load, the gap would have closed and the water carrying soil particles
would have been forced to the ground surface. This process was dramatically revealed during the
soil testing in overconsolidated clay at Manor (Reese, et al., 1975) and at Houston (
and
Dunnavant, 1984) .
The phenomenon of scour is illustrated in Figure 3-11. A gap has opened in the
overconsolidated clay in front of the pile and it has filled with water as load is released. With the
next cycle of loading on the pile, the water is forced upward from the space. The water exits
from the gap with turbulence and the clay is eroded from around the pile.
Wang (1982) constructed a laboratory device to investigate the scouring process. A
specimen of undisturbed soil from the site of a pile test was brought to the laboratory, placed in a
mold, and a vertical hole about 25 mm (1 in.) in diameter was cut in the specimen. A rod was
carefully fitted into the hole and hinged at its base. Water a few millimeters deep was kept over
the surface of the specimen and the rod was pushed and pulled by a machine at a given period
and a given deflection for a measured period. The soil that was scoured to the surface of the
specimen was carefully collected, dried, and weighed. The deflection was increased, and the
59
process was repeated. A curve was plotted showing the weight of soil that was removed as a
function of the imposed deflection. The characteristics of the curve were used to define the scour
potential of that particular clay.
Boiling and turbulence
as space closes
(a)
(b)
Figure 3-11 Scour Around Pile in Clay During Cyclic Loading, (a) Profile View,
(b) Photograph of Turbulence Causing Erosion During Lateral Load Test
The device developed by Wang was far more discriminating about scour potential of a
clay than was the pinhole test (Sherard, et al., 1976), but the results of the test could not explain
fully the differences in the loss of resistance experienced at different sites where lateral-load tests
were performed in clay with water above the ground surface. At one site where the loss of
resistance due to cyclic loading was relatively small, it was observed that the clay included some
seams of sand. It was reasoned that the sand would not have been scoured readily and that
particles of sand could have partially filled the space that was developed around the pile. In this
respect, one experiment showed that pea gravel placed around a pile during cyclic loading was
effective in restoring most of the loss of resistance. However,
and Dunnavant (1984)
-soil gap formed during previous cyclic loading did
not produce a significant regain in lateral pile-head stiffness
While both Long (1984) and Wang (1982) developed considerable information about
the factors that influence the loss of resistance in clays under free water due to cyclic loading,
their work did not produce a definitive method for predicting the loss of resistance. Thus, the
analyst should be cautious when making use of the numerical results presented here with regard
to the behavior of piles in clay under cyclic loading. Full-scale experiments with instrumented
piles at a particular site are recommended for those cases where behavior under cyclic loading is
a critical design requirement.
60
4c q
..........................................................(3-14)
E qf
61
Skempton noted that the influence value I decreases with depth below the ground surface and
the bearing capacity factor increases; therefore, as a first approximation Equation 3-14 is valid at
any depth.
In an undrained compression test, the axial strain is given by
....................................................(3-15)
Where E is Yo
3).
For saturated clays with no change in water content, Equation 3-15 may be rewritten as
.................................................... (3-16)
Where
Equations 3-14 and 3-16 show that, for the same ratio of applied stress to ultimate stress,
the strain in the footing test (or pile under lateral loading) is related to the strain in the laboratory
compression test by the following equation.
62
Table 3-1.
for Laterally Loaded Piles in Stiff Clay (no longer recommended)
Consistency of Clay
Stiff
Very Stiff
Hard
qu, kPa
100-200
200-400
> 400
qu, tsf
1-2
2-4
>4
3.2-6.4
6.4-12.8
> 12.8
460-925
925-1,850
> 1,850
Soil Modulus,
Soil Modulus,
T,
MPa
T,
psi
63
loading. The load was applied in two directions, with the load in the forward direction being
more than twice as large as the load in the backward direction. After a significant number of
cycles, the deflection at the top of the pile was either stable or creeping slowly, so an equilibrium
condition was assumed. The p-y curves for cyclic loading are intended to represent a lowerbound condition. Thus, a designer might possibly be computing an overly conservative response
of a pile, if the cyclic p-y curves are used and if there are only a small number of applications of
the design load (the factored load).
3-3-7-2 Procedure for Computing p-y Curves in Soft Clay for Static Loading
The following procedure is for short-term static loading and is illustrated by Figure 312(a). As noted earlier, the curves for static loading constitute the basis for indicating the
influence of cyclic loading and would be rarely used in design if cyclic loading is of concern.
1.
Obtain the best possible estimates of the variation of undrained shear strength c and
effective unit weight with depth. Also, obtain the value of 50, the strain corresponding to
one-half the maximum principal stress difference. If no stress-strain curves are available,
typical values of 50 are given in Table 3-2.
Table 3-2. Representative Values of
Consistency of Clay
2.
50
50
Soft
0.020
Medium
0.010
Stiff
0.005
Compute the ultimate soil resistance per unit length of pile, using the smaller of the values
given by the equations below.
.............................................. (3-20)
.......................................................... (3-21)
where
= average effective unit weight from ground surface to p-y curve,1
x = depth from the ground surface to p-y curve,
c = shear strength at depth x, and
b = width of pile.
1
Matlock did not specify in his original paper whether the unit weight was total unit weight or
effective unit weight. However, API RP2A specifies that effective unit weight be used. Most
users have adopted the recommendation by API and this is the implementation chosen for LPile.
65
0.5
0 1
8.0
(a)
1
0.72
0.5
15
1
(b)
Figure 3-12 p-y Curves in Soft Clay,(a) Static Loading, (b) Cyclic Loading
Matlock (1970) stated that the value of J was determined experimentally to be 0.5 for soft
clay and about 0.25 for a medium clay. A value of 0.5 is frequently used for J for offshore
soils in the Gulf of Mexico. The value of pu is computed at each depth where a p-y curve is
desired, based on shear strength at that depth.
Equations 3-20 and 3-21 are solved simultaneously to find the transition depth, xr, where
the transition in definition of pu by Equation 3-20 to 3-21 occurs. In general, the minimum
value of xr should be 2.5 pile diameters (see API RP2A, 2010, Section 6.8.2). If the unit
weight and shear strength are constant in the soil layer, then xr is computed using
66
.................................................. (3-22)
LPile has two versions of the soft clay criteria. One version uses a value of J equal to 0.5
by default. This is the version used by most users. The second version is identical in
computations as the first, but the user may enter the value of J at the top and bottom of the
soil layer. LPile does not perform error checking on the input value of J. If the p-y curve
with variable J (API soft clay with user-defined J) is selected, the user should consider the
advice by Matlock for selecting the J value discussed on page 66.
The net effect of using a J value less than 0.5 is to reduce the strength of the p-y curve. An
example of the effect of J on a p-y curve at a depth of 5 feet for a 36-inch diameter pile in
soft clay with c = 1,000 psf and = 55 pcf is shown in Figure 3-13.
1,200
1,000
800
600
400
J = 0.5
J = 0.25
200
0
0
y, inches
Figure 3-13 Example p-y Curves in Soft Clay Showing Effect of J
3.
Compute deflection at one-half the ultimate soil resistance, y50, from the following
equation:
y50 = 2.5
4.
50b
....................................................... (3-23)
Compute points describing the p-y curve from the origin up to 8 y50 using
...................................................... (3-24)
67
Construct the p-y curve in the same manner as for short-term static loading for values of p
less than 0.72pu. For lateral displacements in this range, there is not significant degradation
of the p-y curve during cyclic loading.
2.
If the depth to the p-y curve is greater than or equal to xr (Equation 3-22), select p as 0.72pu
for y equal to 3y50 (Note that the number 0.72 is computed using Equation 3-24 as 1/2 *
31/3 = 0.721124785 ~ 0.72).
3.
If the depth of the p-y curve is less than xr, note that the value of p decreases from 0.72pu at
y = 3y50 down to the value given by Equation 3-25 at y = 15y50.
..................................................... (3-25)
The value of p remains constant beyond y = 15y50.
for the full depth of the soil profile. The loading was assumed to be static. The p-y curves were
computed for the following depths below the ground surface: 1.5 m (5 ft), 3 m (10 ft), 6 m (20
ft), and 12 m (40 ft). The plotted curves are shown in Figure 3-15.
0
2
4
6
8
10
12
14
16
0
10
20
30
40
50
Figure 3-14 Shear Strength Profile Used for Example p-y Curves for Soft Clay
250
200
150
Depth = 2.00 m
Depth = 3.00 m
Depth = 6.00 m
Depth = 12.00 m
100
50
0
0
0.1
0.2
0.3
0.4
0.5
0.6
Figure 3-15 Example p-y Curves for Soft Clay with the Presence of Free Water
69
Obtain values of undrained shear strength c, effective unit weight , and pile diameter b at
depth x.
2.
3.
Compute the soil resistance per unit length of pile, pc, using the smaller of the pct or pcd
from Equations 3-26 and 3-27.
............................................. (3-26)
70
0.5pc
y50
6y50
As y50
18y50
Figure 3-16 Characteristic Shape of p-y Curves for Static Loading in Stiff Clay with Free Water
........................................................ (3-27)
4.
Choose the appropriate value of As from Figure 3-10 on page 58 for modifying pct and pcd
and for shaping the p-y curves or compute As using
............................................ (3-28)
5.
Establish the initial linear portion of the p-y curve, using the appropriate value of ks for
static loading or kc for cyclic loading from Table 3-3 for k.
p = (kx) y.......................................................... (3-29)
Table 3-3. Representative Values of k for Stiff Clays
Average Undrained Shear Strength*
50-100 kPa
1,000-2,000 psf
100-200 kPa
2,000-4,000 psf
200-400 kPa
4,000-6,000 psf
ks (static)
135 MN/m3
(500 pci)
270 MN/m3
(1,000 pci)
540 MN/m3
(2,000 pci)
kc (cyclic)
55 MN/m3
(200 pci)
110 MN/m3
(400 pci)
220 MN/m3
(800 pci)
*The average shear strength should be computed as the average of shear strength of the soil from the ground surface to a
depth of 5 pile diameters. It should be defined as one-half the maximum principal stress difference in an unconsolidatedundrained triaxial test. Note: Conversions of stress ranges are approximate in this table.
6. Compute y50 as
71
.......................................................... (3-30)
Using an appropriate value of
laboratory tests, from Table 3-4.
50
50
7.
50
50-100 kPa
1,000-2,000 psf
0.007
100-200 kPa
2,000-4,000 psf
0.005
200-400 kPa
4,000-6,000 psf
0.004
Establish the first parabolic portion of the p-y curve, using the following equation and
obtaining pc from Equations 3-26 or 3-27.
.................................................... (3-31)
Equation 3-31 should define the portion of the p-y curve from the point of the intersection
with Equation 3-29 to a point where y is equal to Asy50 (see note in Step 10).
8.
9.
0.5 pc 6 As
0.411 pc
0.0625
p c y 6 As y 50 ........................ (3-33)
y 50
Equation 3-33 should define the portion of the p-y curve from the point where y is equal to
6Asy50 to a point where y is equal to 18Asy50 (see note in Step 10).
10.
Equation 3-34 should define the portion of the p-y curve from the point where y is equal to
18Asy50 and for all larger values of y, see the following note.
Note: The p-y curve shown in Figure 3-16 is drawn, as if there is an intersection
between Equation 3-29 and 3-31. However, for small values of k there may be no
intersection of Equation 3-29 with any of the other equations defining the p-y curve.
Equation 3-29 defines the p-y curve until it intersects with one of the other equations
or, if no intersection occurs, Equation 3-29 defines the full p-y curve.
3-3-8-2 Procedure for Computing p-y Curves for Cyclic Loading
A second pile, identical to the pile used for the static loading, was tested under cyclic
loading. The following procedure is for cyclic loading and is illustrated in Figure 3-17. As may
be seen from a study of the p-y curves that are recommended, the results of load tests performed
at the Manor site showed a very large loss of soil resistance. The data from the tests have been
studied carefully and the recommended p-y curves for cyclic loading accurately reflect the
behavior of the soil present at the site. Nevertheless, the loss of resistance due to cyclic loading
for the soils at Manor is much more than has been observed elsewhere. Therefore, the use of the
recommendations in this section for cyclic loading will yield conservative results for many clays.
Long (1984) was unable to show precisely why the loss of resistance occurred during cyclic
loading. One clue was that the clay from Manor was found to lose volume by slaking when a
specimen was placed in fresh water; thus, the clay was quite susceptible to erosion from the
hydraulic action of the free water flushing from the annular gap around the pile as the pile was
pushed back and forth during cyclic loading.
Ac pc
0.45yp 0.6yp
1.8yp
Figure 3-17 Characteristic Shape of p-y Curves for Cyclic Loading of Stiff Clay with Free Water
1. Obtain values of undrained shear strength c, effective unit weight , and pile diameter b.
2. Compute the average undrained shear strength ca over the depth x.
73
3. Compute the soil resistance per unit length of pile, pc, using the smaller of the pct or pcd from
Equations 3-26 and 3-27.
4. Choose the appropriate value of Ac from Figure 3-10 on page 58 or compute Ac using
............................................. (3-36)
5. Compute yp using
....................................................... (3-37)
6. Establish the initial linear portion of the p-y curve, using the appropriate value of ks for static
loading or kc for cyclic loading from Table 3-3 for k. and compute p using Equation 3-29.
7. Compute y50 using Equation 3-30.
8. Establish the parabolic portion of the p-y curve,
........................................... (3-38)
Equation 3-38 should define the portion of the p-y curve from the point of the intersection
with Equation 3-29 to where y is equal to 0.6yp (see note in step 9).
8. Establish the next straight-line portion of the p-y curve,
................................... (3-39)
Equation 3-39 should define the portion of the p-y curve from the point where y is equal to
0.6yp to the point where y is equal to 1.8yp (see note on Step 9).
9. Establish the final straight-line portion of the p-y curve,
........................................... (3-40)
Equation 3-40 defines the p-y curve from the point where y equals 1.8yp and all larger values
of y (see following note).
Note: Figure 3-17 is drawn, as if there is an intersection between Equation 3-29 and
Equation 3-38. There may be no intersection of Equation 3-29 with any of the other
equations defining the p-y curve. If there is no intersection, the equation should be employed
that gives the smallest value of p for any value of y.
3-3-8-3 Recommended Soil Tests
Triaxial compression tests of the unconsolidated-undrained type with confining pressures
conforming to in situ pressures are recommended for determining the shear strength of the soil.
74
The value of 50 should be taken as the strain during the test corresponding to the stress equal to
one-half the maximum total-principal-stress difference. The shear strength, c, should be
interpreted as one-half of the maximum total-principal-stress difference. Values obtained from
triaxial tests might be somewhat conservative but would represent more realistic strength values
than other tests. The unit weight of the soil must be determined.
3-3-8-4 Examples
Example p-y curves were computed for stiff clay for a pile with a diameter of 610 mm
(24 in.). The soil profile that was used is shown in Figure 3-18. The submerged unit weight of
the soil was 7.9 kN/m3 (50 pcf) over the full depth.
In the absence of a stress-strain curve, 50 was taken as 0.005 for the full depth of the soil
profile. The slope of the initial portion of the p-y curve was established by assuming a value of k
of 135 MN/m3 (500 pci). The loading was assumed to be cyclic. The p-y curves were computed
for the following depths below the ground surface: 0.6 m (0.2 ft), 1.5 m (5 ft), 3 m (10 ft), and 12
m (40 ft). The plotted curves are shown in Figure 3-19.
3-3-9 Response of Stiff Clay with No Free Water
A lateral-load test was performed at a site in Houston, Texas on a drilled shaft (bored
pile), with a diameter of 915 mm (36 in.). A 254-mm (10 in)-diameter steel pipe instrumented
with strain gages was positioned at the central axis of the pile before concrete was placed. The
embedded length of the pile was 12.8 m (42 ft). The average undrained shear strength of the clay
in the upper 6 m (20 ft) was approximately 105 kPa (2,200 psf). The experiments and their
interpretation were reported in the papers by Welch and Reese (1972) and Reese and Welch
(1975).
0
2
4
6
8
10
12
14
16
0
50
100
150
200
Figure 3-18 Example Shear Strength Profile for p-y Curves for Stiff Clay with No Free Water
75
250
Depth = 1.00 m
Depth = 2.00 m
Depth = 3.00 m
Depth = 12.00 m
200
150
100
50
0
0.0
0.005
0.01
0.015
0.02
0.025
Lateral Deflection y, meters
0.03
0.035
Figure 3-19 Example p-y Curves for Stiff Clay in Presence of Free Water for Cyclic Loading
The same experimental setup was used to develop both the static and the cyclic p-y
curves, contrary to the procedures employed for the two other experiments with piles in clays.
The load was applied in only one direction rather than in two directions, also in variance with the
other experiments.
A load was applied and maintained until the strain gages were read with a high-speed
data-acquisition system. The same load was then cycled for a number of times and held constant
while the strain gages were read at specific numbers of cycles of loading. The load was then
increased and the procedure was repeated. The difference in the magnitude of successive loads
was relatively large and the assumption was made that cycling at the previous load did not
influence the readings for the first cycle at the new higher load.
The p-y curves obtained for these load tests were relatively consistent in shape and
showed the increase in lateral deflection during cyclic loading. This permitted the expressions of
lateral deflection to be formulated in terms of the stress level and the number of cycles of
loading. Thus, the engineer can specify a number of cycles of loading (up to a maximum of
5,000 cycles of loading) in doing the computations for a particular design.
3-3-9-1 Procedure for Computing p-y Curves for Stiff Clay without Free Water for Static
Loading
The following procedure is for short-term static loading and the p-y curve for stiff clay
without free water is illustrated in Figure 3-20.
76
y
16y50
Figure 3-20 Characteristic Shape of p-y Curve for Static Loading in Stiff Clay without Free
Water
1.
Obtain values for undrained shear strength c, effective unit weight , and pile diameter b.
Also, obtain the values of 50 from stress-strain curves. If no stress-strain curves are
available, use a value of 50 of 0.010 or 0.005 as given in Table 3-2, the larger value being
more conservative.
2.
Compute the ultimate soil resistance, pu, per unit length of pile using the smaller of the
values given by Equations 3-20 and 3-21. (In the use of Equation 3-20, the shear strength
is taken as the average from the ground surface to the depth being considered and J is
taken as 0.5. The unit weight of the soil should reflect the position of the water table.)
...............................................(3-20)
...........................................................(3-21)
3.
Compute the deflection, y50, at one-half the ultimate soil resistance from Equation 3-23.
y50 = 2.5
4.
(3-23)
Compute points describing the p-y curve from the relationship below.
p
5.
50b ........................................................
pu y
2 y50
0.25
..................................................... (3-41)
77
3-3-9-2 Procedure for Computing p-y Curves for Stiff Clay without Free Water for Cyclic
Loading
The following procedure is for cyclic loading and the p-y curve for stiff clay without free
water is illustrated in Figure 3-21.
pu
N1
N3
N2
yc = ys + y50 C log N3
yc = ys + y50 C log N2
yc = ys + y50 C log N1
16y50+9.6(y50)logN1
y
16y50+9.6(y50)logN3
16y50+9.6(y50)logN2
Figure 3-21 Characteristic Shape of p-y Curves for Cyclic Loading in Stiff Clay with No Free
Water
1.
Determine the p-y curve for short-term static loading by the procedure previously given.
2.
Determine the number of times the lateral load will be applied to the pile.
3.
Obtain the value of C for several values of p/pu, where C is the parameter describing the
effect of repeated loading on deformation. The value of C is found from a relationship
developed by laboratory tests, (Welch and Reese, 1972), or in the absence of tests, from
....................................................... (3-42)
4.
At the value of p corresponding to the values of p/pu selected in Step 3, compute new
values of y for cyclic loading from
................................................. (3-43)
where
yc = deflection under N-cycles of load,
ys = deflection under short-term static load,
y50 = deflection under short-term static load at one-half the ultimate resistance, and
78
The p-y curve defines the soil response after N-cycles of loading.
300
200
Depth = 0.60 m
Depth = 1.50 m
Depth = 3.00 m
Depth = 12.00 m
100
0
0.0
0.05
0.1
0.15
0.2
Lateral Deflection y, meters
0.25
0.3
Figure 3-22 Example p-y Curves for Stiff Clay with No Free Water,
Cyclic Loading
79
3-3-10 Modified p-y Criteria for Stiff Clay with No Free Water
The p-y criteria for stiff clay with no free water were described in Section 3-3-9. The p-y
curve for stiff clay with no free water is based on Equation 3-41, which does not contain an
initial stiffness parameter k. Although the criteria for stiff clay without free water has been used
successfully for many year, there have been some reported cases from the Southeastern United
States where load tests have found that the initial load-deformation response is modeled too
stiffly.
The ultimate load-transfer resistance pu used in the p-y criteria is consistent with the
theory of plasticity and has also correlated well with the results of load tests. However, the soil
resistance at small deflections is influenced by factors such as soil moisture content, clay
mineralogy, clay structure, possible desiccation, and pile diameter. Brown (2002) has
recommended the use of a k value to modify the initial portion of the p-y curves if one has the
results of lateral load test for local calibration of the initial stiffness k. Judicious use of this
modified p-y criteria enables one to obtain improved predictions with experimental readings that
may be used later for design computations.
The user may select an initial stiffness k based on Table 3-3 or from a site-specific lateral
load test. LPile will use the lower of the values computed using Equation 3-29 or Equation 3-41
for pile response as a function of lateral pile displacement.
3-3-11 Other Recommendations for p-y Curves in Clays
As noted earlier in this chapter, the selection of the set of p-y curves for a particular field
application is a critical feature of the method of analysis. The presentation of three particular
methods for clays does not mean the other recommendations are not worthy of consideration.
Some of these methods are mentioned here for consideration and their existence is an indication
of the level of activity with regard to the response of soil to lateral deflection.
Sullivan, et al. (1980) studied data from tests of piles in clay when water was above the
ground surface and proposed a procedure that unified the results from those tests. While the
proposed method was able to predict the behavior of the experimental piles with excellent
accuracy, two parameters were included in the method that could not be found by any rational
procedures. Further work could develop means of determining those two parameters.
Stevens and Audibert (1979) reexamined the available experimental data and suggested
specific procedures for formulating p-y curves. Bhushan, et al. (1979) described field tests on
drilled shafts under lateral load and recommended procedures for formulating p-y curves for stiff
clays. Briaud, et al. (1982) suggested a procedure for use of the pressuremeter in developing p-y
curves. A number of other authors have also presented proposals for the use of results of
pressuremeter tests for obtaining p-y curves.
and Gazioglu (1984) reviewed all of the data that were available on p-y curves
for clay and presented a summary report to the American Petroleum Institute. The research
conducted by
and his co-workers (
and Dunnavant, 1984; Dunnavant and
, 1985) at the test site on the campus of the University of Houston developed a large
volume of data on p-y curves. This work will most likely result in specific recommendations in
due course.
80
Medium
Dense
0.95 - 2.8
(3.5 - 10.4)
3.5 - 10.9
(13.0 - 40.0)
13.8 - 27.7
(51.0 - 102.0)
0.53 - 1.7
(2.1 - 6.4)
2.2 - 7.3
(8.0 - 27.0)
8.3 - 17.9
(32.0 - 64.0)
81
Fs
y
Ff
Fs
C
Fn
D
W
Fp
Fn
Ft
Ff
Fp
(b)
Pile of
Diameter b
Fs
Fn
Fp
(a)
Fpt
Fa
(c)
Figure 3-23 Geometry Assumed for Passive Wedge Failure for Pile in Sand
The force Fpt may be computed by following a procedure similar to that used to solve the
equation in the clay model (Figure 3-6). The resulting equation is
............... (3-45)
where:
= the angle of the wedge in the horizontal direction
82
( pu ) sa
K 0 tan sin
tan(
) cos
H K 0 H tan
tan sin
tan
tan(
tan
b H tan tan
................ (3-46)
K Ab
Bowman (1958) performed some laboratory experiments with careful measurements and
suggested values of from /3 to /2 for loose sand and up to for dense sand. The value of is
approximated by the following equation.
........................................................ (3-47)
The model for computing the ultimate soil resistance at some distance below the ground
surface is shown in Figure 3-24(a). The stress 1 at the back of the pile must be equal or larger
than the minimum active earth pressure; if not, the soil could fail by slumping. The assumption is
based on two-dimensional behavior; thus, it is subject to some uncertainty. If the states of stress
shown in Figure 3-24(b) are assumed, the ultimate soil resistance for horizontal movement of the
soil is
............................ (3-48)
The equations for (pu)sa and (pu)sb are approximate because of the elementary nature of
the models that were used in the computations. However, the equations serve a useful purpose in
indicating the form, if not the magnitude, of the ultimate soil resistance.
3-4-1-3 Influence of Diameter on p-y Curves
No studies have been reported on the influence of pile diameter on p-y curves in sand.
The reported case studies of piles in sand, some of which are of large diameter, do not reveal any
particular influence of the pile diameter. However, virtually all of the reported lateral-load tests,
except the ones described herein, have used only static loading.
83
(a)
(b)
Figure 3-24 Assumed Mode of Soil Failure by Lateral Flow Around Pile in Sand,
(a) Section Though Pile, (b) Mohr-Coulomb Diagram
3-4-1-4 Influence of Cyclic Loading
As noted above, very few reports of tests of piles subjected to cyclic lateral loading have
been reported. There is evidence that the repeated loading on a pile in predominantly one
direction will result in a permanent deflection in the direction of loading. It has been observed
that when a relatively large cyclic load is applied in one direction, the top of the pile will deflect
a significant amount, allowing grains of cohesionless soil to fall into the open gap at the back of
the pile. Thus in such a case, the pile cannot return to its initial position after cyclic loading
ceases.
84
Observations of the behavior of sand near the ground surface during cyclic loading
support the idea that the void ratio of sand is approaching a critical value. That is, dense sand
will loosens and loose sand will densify under cyclic loading.
A careful study of the two phenomena mentioned above should provide information of
use to engineers. Full-scale experiments with detailed studies of the nature of the sand around the
top of a pile, both before and after loading, would be a welcome contribution.
3-4-1-5 Early Recommendations
The values of subgrade moduli recommended by Terzaghi (1955) provided some basis
for computation o
to
practice until the digital computer and the required programs became widely available. There
was a period of a few years
when engineers were solving the difference equations
using mechanical calculators. The piles for some early offshore platforms were designed using
this method.
Parker and Reese (1971)
performed some small-scale experiments, examined
unpublished data, and recommended procedures for predicting p-y curves for sand. The method
of Parker and Reese received little use in practice because the method of Cox, et al. (1974)
described later, was based on a comprehensive load testing program on full-sized piles and
became available shortly afterward.
3-4-1-6 Field Experiments
An extensive series of field tests were performed at a site on Mustang Island, near Corpus
Christi, Texas (Cox, et al., 1974). Two steel-pipe piles, 610 mm (24 in.) in diameter, were driven
into sand in a manner to simulate the driving of an open-ended pipe and were subjected to lateral
loading. The embedded length of the piles was 21 meters (69 feet). One of the piles was
subjected to short-term loading and the other to cyclic loading.
The soil at the test site was classified as SP using the Unified Soil Classification System,.
The sand was poorly graded, fine sand with an angle of internal friction of 39 degrees. The
effective unit weight was 10.4 kN/m3 (66 pcf). The water surface was maintained at 150 mm (6
in.) above the ground surface throughout the test program.
3-4-1-7 Response of Sand Above and Below the Water Table
The procedure for developing p-y curves for piles in sand is shown in detail in the next
section. The piles that were used in the experiments, described briefly below, were the ones used
at Manor, except that the piles at Manor had an extra wrap of steel plate.
3-4-2 Response of Sand
The following procedure is for both short-term static loading and for cyclic loading for a
flat ground surface and a vertical pile. The procedure is illustrated in Figure 3-25 (Reese, et al.,
1974).
85
p
x = x4
x = x3
x = x2
pu
m
k
pk
x = x1
pm
ym
yu
b/60
3b/80
yk
ksx
Figure 3-25 Characteristic Shape of a Set of p-y Curves for Static and Cyclic Loading in Sand
3-4-2-1 Procedure for Computing p-y Curves in Sand
1.
Obtain values for the depth of the p-y curve x, the angle of internal friction , effective
unit weight of soil , and pile diameter b (Note: use effective unit weight for sand below
the water table and total unit weight for sand above the water table).
2.
3.
, and
..................... (3-49)
Compute the ultimate soil resistance per unit length of pile using the smaller of the values
given by
,
where
.................... (3-50)
............................. (3-51)
86
4.
Establish
Compute pu using:
or
............................................... (3-52)
87
and
B
2
Bs (static)
Bc (cyclic)
88
Table 3-6. Representative Values of k for Submerged Sand for Static and Cyclic Loading
Recommended k
MN/m3
(pci)
Relative Density
Loose
Medium
Dense
5.4
16.3
34
(20.0)
(60.0)
(125.0)
Table 3-7. Representative Values of k for Sand Above Water Table for Static and Cyclic
Loading
Recommended k
MN/m3
(pci)
Loose
6.8
(25.0)
Relative Density
Medium
Dense
24.4
61.0
(90.0)
(225.0)
If the input value of k is left equal to zero, a default value will be computed by LPile
using the curves shown in Figure 3-31 on page 94. Whether the sand is above or below the water
table will be determined from the input value of effective unit weight. If the effective unit weight
is less than 77.76 pcf (12.225 kN/m3) the sand is considered below the water table. If the input
value of is greater than 40 degrees, a k value corresponding to 40 degrees is used by LPile.
7.
as follows:
........................................................... (3-59)
89
........................................................ (3-60)
e. Compute appropriate number of points on the parabola by using Equation 3-56.
Note: The curve in Figure 3-25 is drawn as if there is an intersection between the initial
straight-line portion of the p-y curve and the parabolic portion of the curve at point k. However,
in some instances there may be no intersection with the parabola. Equation 3-55 defines the p-y
curve until there is an intersection with another portion of the p-y curve or if no intersection
occurs, Equation 3-55 defines the complete p-y curve. If yk is in between points ym and yu, the
curve is tri-linear and if yk is greater than yu, the curve is bi-linear as shown in Figure 3-28.
3-4-2-2 Recommended Soil Tests
Fully drained triaxial compression tests are recommended for obtaining the angle of
internal friction of the sand. Confining pressures should be used which are close or equal to those
at the depths being considered in the analysis. Tests must be performed to determine the unit
weight of the sand. However, it may be impossible to obtain undisturbed samples and frequently
the angle of internal friction is estimated from results of some type of in-situ test.
The procedure above can be used for sand above the water table if appropriate
adjustments are made in the unit weight and angle of internal friction of the sand. Some smallscale experiments were performed by Parker and Reese (1971) , and recommendations for the py curves for dry sand were developed from those experiments. The results from the Parker and
Reese experiments should be useful in checking solutions from results of experiments with fullscale piles.
p
Lower k x
kx
Higher k x
kx
90
4,000
3,000
2,000
1,000
0
0.0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
Lateral Deflection y, m
Depth = 1.50 m
Depth = 3.00 m
Depth = 6.00 m
Depth = 12.00 m
Figure 3-29 Example p-y Curves for Sand Below the Water Table, Static Loading
3-4-3 API RP 2A Recommendation for Response of Sand Above and Below the
Water Table
3-4-3-1 Background of API Method for Sand
This method is recommended by the American Petroleum Institute in its manual for
recommended practice for designing fixed offshore platforms (API RP 2A). Thus, the method
has official recognition. The API procedure for p-y curves in sand was based on a number of
field experiments. There is no difference for ultimate resistance (pu) between the Reese et al.
criteria and the API criteria. The API method uses a hyperbolic tangent function for computation.
The main difference between those two criteria will be the initial modulus of subgrade reaction
and the shape of the curves.
91
3-4-3-2 Procedure for Computing p-y Curves Using the API Sand Method
The following procedure is for both short-term static loading and for cyclic loading as
described in API RP2A (2010) .
1.
Obtain values for the angle of internal friction , the effective unit weight of soil, , and
the pile diameter b.
2.
Compute the ultimate soil resistance at a selected depth x. The ultimate lateral bearing
capacity (ultimate lateral resistance pu) for sand has been found to vary from a value at
shallow depths determined by Equation 3-61 to a value at deep depths determined by
Equation 3-62. At a given depth, the equation giving the smallest value of pu should be
used as the ultimate bearing capacity. The value of pu is the lesser of pu at shallow depths,
pus, or pu at great depth, pud , where:
................................................... (3-61)
........................................................ (3-62)
where:
pu = ultimate resistance (force/unit length), lb./in. (kN/m),
= effective unit weight, pci (kN/m3),
x = depth, in. (m),
= angle of internal friction of sand, degrees,
C1, C2, C3 = coefficients determined from Figure 3-30 as a function of , or
where
and
92
100
100
5.0
5
90
80
80
4
4.0
70
3
3.0
60
60
C2
50
40
40
2
2.0
C1
30
C3
1
1.0
20
20
10
0
0.0
15
15
00
20
20
25
25
30
30
35
35
40
40
Compute the load-deflection curve based on the ultimate soil resistance pu which is the
minimum value of pu calculated in Step 2. The lateral soil resistance-deflection (p-y)
relationships for sand are nonlinear and, in the absence of more definitive information,
may be approximated at any specific depth x by the following expression:
................................................ (3-63)
where
A = factor to account for cyclic or static loading. Evaluated by:
A = 0.9 for cyclic loading.
for static loading,
pu = smaller of values computed from Equation 3-61 or 3-62, lb./in. (kN/m),
93
k = initial modulus of subgrade reaction, pci (kN/m3). Determine k from Figure 3-31 as
function of angle of internal friction, ,
y = lateral deflection, in. (m), and
x = depth, inches (m).
, Friction Angle, degrees
28
29
Very
Loose
300
36
30
Loose
Medium Dense
40
Dense
45
Very
Dense
Sand above
the water
table
250
200
150
Sand below
the water
table
100
50
0
0
20
40
60
80
100
Relative Density, %
94
A hand calculation for p-y curves at a depth of 20 in. was made to check the computer
solution, as shown in the following.
1.
2.
3.
Compute coefficient A
A = 3.0
5.
A pu tanh
If y = 1.35 in.
kx
y
A pu
A pu tanh
(140)(20 in.)
(1.35 in.)
(2.55)(255 lb/in. 3 )
The check by hand computations yielded exact values for the two values of deflection that
were considered.
The computed curves are presented in Figure 3-32.
95
3,000
2,500
2,000
1,500
1,000
500
0
0.0
0.25
0.5
0.75
1.0
1.25
1.5
1.75
2.0
96
When sand is liquefied under undrained conditions, some suggest that it behaves in a
manner similar to the behavior of soft clay. Wang and Reese (1998) have studied the behavior of
piles in liquefied soil by modeling the liquefied sand as soft clay. The p-y curves were generated
using the model for soft clay by equating the cohesive strength equal to the residual strength of
liquefied sand. The strain factor 50 was set equal to 0.05 in their study.
Laboratory procedures cannot measure the residual shear strength of liquefied sand with
reasonable accuracy due to the unstable nature of the soil. Some case histories must be evaluated
to gather information on the behavior of liquefied deposit. Recognizing the need to use case
studies, Seed and Harder (1990) examined cases reported where major lateral spreading has
occurred due to liquefaction and where some conclusions can be drawn concerning the strength
and deformation of liquefied soil.
Unfortunately, cases are rare where data are available on strength and deformation of
liquefied soils. However, a limited number of such cases do exist, for which the residual
strengths of liquefied sand and silty sand can be determined with a reasonable accuracy. Seed
and Harder found that a residual strength of about 10 percent of the effective overburden stress
can be used for liquefied sand.
Although simplified methods based on engineering judgment have been used for design,
full-scale field tests are needed to develop a full range of p-y curves for liquefied sand. Rollins et
al. (2005b) have performed full scale load tests on a pile group in liquefied sand with an initial
relative density between 45 and 55 percent. The p-y curves developed on the basis of these
studies have a concave upward shape, as shown in Figure 3-33. This characteristic shape appears
to result primarily from dilative behavior during shearing, although gapping effects may also
contribute to the observed load-transfer response. Rollins and his co-workers also found that p-y
curves for liquefied sand stiffen with depth (or initial confining stress). With increasing depth,
small displacement is required to develop significant resistance and the rate at which resistance
develops as a function of lateral pile displacement also increases.
y
150 mm
97
Following liquefaction, p-y curves in sand become progressively stiffer with the passage
of time as excess pore water pressures dissipate. The shape of a p-y curve appears to transition
from concave up to concave down as pore water pressure decreases. An equation based on the
results of the load tests has been developed by Rollins et al. (2003) to describe the observed loaddisplacement response of liquefied sand as a function depth.
3-5-2 Procedure for Computing p-y Curves in Liquefied Sand
The expression developed by Rollins et al. (2005a) for p-y curves in liquefied sands at
different depths is shown below is based on their fully-instrumented load tests. Coefficients for
these equations were fit to the test data using a trial and error process in which the errors between
the target p-y curves and those predicted by the equations were minimized. The resulting
equations were then compared, and the equation that produced the most consistent fit was
selected.
........................................................(3-64)
...................................................(3-65)
.....................................................(3-66)
.....................................................(3-67)
where p is the soil resistance in kN/meter, y is the lateral deflection of the pile in millimeters, z is
the depth in meters (see note in last paragraph of this section), and Pd is the diameter correction
discussed below.
Rollins et al. (2005a) studied the diameter effects for different sizes of piles and
recommended a modification factor for correcting Equation 3-64, as shown below.
...................................................(3-68)
where b is the diameter or width of the pile or drilled shaft in meters. The p-y curves for liquefied
sand can be multiplied by Pd to obtain values for p-y curves for deep foundations of varying
diameters.
Note that use of the diameter correction is limited to foundations between 0.3 and 2.6
meters in diameter. This limitation on diameter prevents implementation of the above relations to
micropiles because their diameters are generally less than 0.3 meters.
Application of Equation 3-64 should generally be limited to conditions comparable to
those from which it was derived. These conditions are:
Relative density between 45 and 55 percent
Lateral soil resistance less than 15 kN/meter
Lateral pile deflection less than 150 mm (0.15 m),
Depths of 6 meters or less, and
98
99
There were
-inch diameter
cyclic test shafts, with ten load cycles (N = 1 through 10) performed per load increment. The
lateral load for each load cycle were sustained for only a few seconds with the exception of load
cycles 1 and 10 which were sustained for approximately 15 to 20 minutes to allow time for the
inclinometer readings to be performed. For load cycles 2 through 9, the duration for each load
cycle was approximately 1 minute, 2 minutes, 3.5 minutes, and 6.5 minutes for load increments
A though D, respectively, as a greater time was required to reach the larger loads. The load was
reversed after each load cycle to return the top of pile to approximately the same location.
3-6-1-2 Soil Profile from Cone Penetration Testing
A back-fit model of the pile behavior using the available soil strength data obtained (from
both in-situ and laboratory tests) to the measured pile performance led to the conclusion that the
CPT testing provided the best correlation. Furthermore, CPT testing can be easily performed in
the loess soils being modeled and has become readily widely available.
Three cone penetration tests were performed by the Kansas Department of Transportation
at the test site location. A preliminary cone penetration test was performed in the general vicinity
of the test shafts (designated as CPT-1). Two additional cone penetration tests were performed
subsequent to the lateral load testing. A cone penetration test was performed between the 42-inch
diameter static test shafts (Shafts 1 and 2) shortly after on the same day the lateral load test was
performed on these shafts. A cone penetration test was performed between the 30-inch diameter
static test shafts (Shafts 3 and 4) two days after the completion of the load test performed on
these shafts. The locations of the cone penetration tests were a few feet from the test shafts.
Given the nature of the soil conditions and the absence of a ground water table, it is reasonable to
assume that the cone penetration tests were unaffected by any pore water pressure effects that
may have been induced by the load testing.
An idealized profile of cone tip resistance with depth interpreted as an average from the
cone penetration tests performed between the static test shafts is shown in Figure 3-34. This
profile is considered representative of the subsurface conditions for all the test shaft locations.
Note that it is most useful to break the idealized soil profile into layers wherein the cone tip
resistance is either constant with depth or linearly varies with depth as these two conditions are
easily accommodated by most lateral pile analyses software.
The cone tip resistance is reduced by 50% at the soil surface, and allowed to increase
linearly with depth to the full value at a depth of two pile diameters, as shown in Figure 3-34.
This is done to account for the passive wedge failure mechanism exhibited at the ground surface
that reduces the lateral resistance of the soil between the ground surface and a lower depth
(assumed at two shaft diameters). Below a depth of two shaft diameters, the lateral resistance is
considered as a flow around bearing failure mechanism.
The idealized cone tip resistance values were correlated with depth with the ultimate
lateral soil resistance (pu0) at corresponding depths.
100
15
20
Used For Model
25
Between 30"
A.L.T. (6/9/2005)
30
Between 42"
A.L.T. (6/8/2005)
CPT-1
(8/12/2004)
35
40
0
20
40
60
80
100
120
140
160
180
200
qc, ksf
Figure 3-34 Idealized Tip Resistance Profile from CPT Testing Used for Analyses.
3-6-2 Procedure for Computing p-y Curves in Loess
3-6-2-1 General Description of p-y Curves in Loess
Procedures are provided to produce a p-y curve for loess, shown generically in Figure 335. The ultimate soil resistance (pu0) that can be provided by the soil is correlated to the cone tip
resistance at any given elevation. Note that to account for the passive wedge failure mechanism
exhibited at the ground surface, the cone tip resistance is reduced by 50% at the soil surface and
allowed to return to the full value at a depth equal to two pile diameters. The initial modulus of
the p-y curve, Ei, is determined from the ultimate lateral soil reaction expressed on a per unit
length of pile basis, pu, for the specified pile diameter, and specified reference displacement, yref.
A hyperbolic relationship is used to compute the secant modulus of the p-y curve, Es, at any
given pile displacement, y. The lateral soil reaction per unit pile length, p, for any given pile
displacement is determined by the secant modulus at that displacement. Provisions for the
degradation of the p-y curve as a function of the number of cycles loading, N, are incorporated
into the relationship for ultimate soil reaction.
The model is of a p-y curve that is smooth and continuous. This model is similar to the
lateral behavior of pile in loess soil measured in load tests.
3-6-2-2 Equations of p-y Model for Loess
The ultimate unit lateral soil resistance, pu0, is computed from the cone tip resistance
multiplied by the cone bearing capacity factor, NCPT using
........................................................(3-69)
101
p
pu
Ei
Es
yref
Figure 3-35. Generic p-y curve for Drilled Shafts in Loess Soils
where NCPT is dimensionless, and pu0 and qc are in consistent units of (force/length2)
The value of NCPT was determined from a best fit to the load test data. It is believed that
NCPT is relatively insensitive to soil type as this is a geotechnical property determined by in-situ
testing. The value of NCPT derived from the load test data is
........................................................(3-70)
The ultimate lateral soil reaction, pu, is computed by multiplying the ultimate unit lateral
soil resistance by the pile diameter, b, and dividing by an adjustment term to account for cyclic
loading. The adjustment term for cyclic loading takes into account the number of cycles of
loading, N, and a dimensionless constant, CN.
....................................................(3-71)
where:
b is the pile diameter in any consistent unit of length,
CN is a dimensionless constant,
N is the number of cycles of loading (1 to 10), and
pu is in units of (force/length).
102
CN was determined from a best fit of cyclic degradation for two 30-inch diameter test
shafts subjected to cyclic loading. CN is
...........................................................(3-72)
The cyclic degradation term (the denominator of Equation 3-71) equals 1 for N = 1
(initial cycle, or static load) and equals 1.24 for N = 10. The value of CN has a direct effect on the
amount of cyclic degradation to the p-y curve (i.e., a greater value of CN will allow greater
degradation of the p-y curve, resulting in a smaller pu). Note that the degradation of the ultimate
soil resistance per unit length of shaft parameter will also have the desired degradation effect
built into the computation of the p-y modulus values.
A parameter is needed to define the rate at which the strength develops towards its
ultimate value (pu0). The reference displacement, yref, is defined as the displacement at which the
tangent to the p-y curve at zero displacement intersects the ultimate soil resistance asymptote
(pu), as shown in Figure 3-35. The best fit to the load test data was obtained with the following
value for reference displacement.
yref = 0.117 inches = 0.0029718 meters .................................. (3-73)
Note that the suggested value for the reference displacement provided the best fit to the
piles tested at a single test site in Kansas for a particular loess formation. Unlike the ultimate unit
lateral resistance (pu0), it is believed that the rate at which the strength is mobilized may be
sensitive to soil type. Thus, re-evaluation of the reference displacement parameter is
recommended when performing lateral analyses for piles in different soil conditions because this
parameter is likely to have a substantial effect on the resulting pile deflections. The effect of the
reference displacement is proportional to pile performance that is a larger value of yref will allow
for larger pile head displacements at a given lateral load.
The initial modulus, Ei, is defined as the ratio of the ultimate lateral resistance expressed
on a per unit length of pile basis over the reference displacement.
........................................................... (3-74)
A secant modulus, Es, is determined for any given displacement, y, by the following
hyperbolic relationship of the initial modulus expressed on a per unit length of pile basis and a
hyperbolic term ( ) which is in turn a function of the given displacement (y), the reference
displacement (yref), and a dimensionless correlation constant (a).
......................................................... (3-75)
.............................................. (3-76)
103
............................................................(3-77)
where Es and Ei are in units of force/length2, and a and
are dimensionless.
The constant a was found from a best fit to the load test data. Note that the constant a
primarily affects the secant modulus at small displacements (say within approximately 1 inch or
25 mm), and is inversely proportional to the stiffness response of the p-y curve (i.e., a larger
value of a will reduce the mobilization of soil resistance with displacement). Combining the two
equations above, one obtains
.............................................(3-78)
The modulus ratio (secant modulus over initial modulus, Es/Ei) versus displacement used
for p-y curves in loess is shown in Figure 3-36. Note that the modulus ratio is only a function of
the hyperbolic parameters of the constant (a) and the reference displacement (yref), thus the curve
presented is valid for all pile diameters and cone tip bearing values tested.
1.0
0.9
a = 0.1
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.001
0.01
0.1
1.0
10
100
104
........................................................... (3-79)
where:
Es is the secant modulus in units of force/length2, and
y is the lateral pile displacement.
Several p-y curves obtained from the model described above is presented in Figure 3-37
for the 30-inch diameter shafts, and Figure 3-38 for the 42-inch diameters shafts. Note that there
are three sets of curves presented for each shaft diameter which correspond to the cone tip
resistance values of 11 ksf, 22 ksf, and 100 ksf (as was shown in Figure 3-34). These p-y curves
were used in the LPile analyses presented later.
9,000
8,000
7,000
6,000
11 ksf
5,000
22 ksf
100 ksf
4,000
3,000
2,000
1,000
0
0
y , inches
8,000
22 ksf
100 ksf
6,000
4,000
2,000
0
0
y , inches
Figure 3-38 p-y Curves and Secant Modulus for the 42-inch Diameter Shafts.
105
The static p-y curves shown in Figure 3-37 and 3-38 were degraded with load cycle
number (N) for use in the cyclic load analyses. Figure 3-39 presents the cyclic p-y curve
generated for the analyses of the 30-inch diameter shafts at the cone tip resistance value of 22
ksf.
2,000
1,800
1,600
1,400
N= 1
1,200
N= 5
1,000
N = 10
800
600
400
200
0
0
y , inches
9. Determine the soil resistance per unit length of pile (p) for each of the displacements selected
in Step 7 in accordance with Equation 3-79.
3-6-2-4 Limitations on Conditions for Validity of Model
The p-y curve for static loading was based on best fits of data from full scale load tests on
30-inch and 42-inch diameter shafts installed in a loess soil formation with average cone tip
resistance values ranging from 20 to 105 ksf (960 to 5,000 kPa).
Caution is advised when extrapolating the static model formulation for shaft diameters or
soil types and/or strengths outside these limits. In addition, the formulation for the cyclic
degradation model parameters are based on load tests with only ten cycles of loading (N = 1 to
10) obtained at four different load increments on an additional two 30-inch diameter shafts.
Caution is thus also warranted when extrapolating the cyclic model to predict results beyond 10
cycles of load (N > 10), particularly as the magnitude of loading increases.
3-7 p-y Curves in Soils with Both Cohesion and Internal Friction
3-7-1 Background
The previous methods that were presented were for soils that can be characterized as
either cohesive or cohesionless (clay or sand, for example). There are currently no generally
accepted recommendations for developing p-y curves for c- soils.
Among the reasons for the limitation on soil characteristics are the following. Firstly, in
foundation design, where the p-y analysis has been used mostly, the characterization of the soil
by either a value of c or , but not both, has been used. Secondly, the major experiments on
which the p-y predictions have been based have been performed in soils that can be described by
either c or . However, there are now numerous occasions when it is desirable, and perhaps
necessary, to describe the characteristics of the soil more carefully.
An example of the need to have predictions for p-y curves for c- soils is when piles are
used to stabilize a slope. A detailed explanation of the analysis procedure is presented in Chapter
6. It is well known that most of the currently accepted methods of analysis of slope stability
characterize the soils in terms of c and for long-term or drained analysis. Therefore, it is
inconsistent, and either unsafe or unconservative, to assume the pile to be in soil that is
characterized either by c or alone.
There are other instances in the design of piles under lateral loading where it is desirable
to have methods of prediction for p-y curves for c- soils. The shear strength of unsaturated,
cohesive soils generally is represented by strength components of both c and . In many practical
cases, however, there is the likelihood that the soil deposit might become saturated because of
rainfall and rise of the ground water table. However, there could well be times when the ability
to design for dry seasons is critical.
Cemented soils are frequently found in subsurface investigations. Some comments for the
response of laterally loaded piles in calcareous soils were presented by Reese (1988). It is
apparent that cohesion from the cementation will increase soil resistance significantly, especially
for soils near the ground surface.
107
p
m
pm
pk
yk
ym
pu
yu
ks
y
b/60
3b/80
b = Cp
b.................................................... (3-80)
where
p
= passive pressure including the three-dimensional effect of the passive wedge (F/L2)
.................... (3-84)
.............................. (3-85)
The cohesion component (puc) will be the smaller of the two equations below.
.............................................. (3-86)
.......................................................... (3-87)
3-7-3 Procedure for Computing p-y Curves in Soils with Both Cohesion and
Internal Friction
To develop the p-y curves, the procedures described earlier for sand by Reese et al (1974)
will be used because the stress-strain behavior of c- soils are believed to be closer to the stressstrain curve of cohesionless soil than for cohesive soil. The following procedures are used to
develop the p-y curves for soils with both cohesion and internal friction.
109
1.
2.
.................................... (3-89)
Compute ym as
........................................................... (3-90)
Compute pm by the following equation:
or
.............................................. (3-91)
Use the appropriate value of Bs or Bc from Figure 3-27 on page 88 for the particular nondimensional depth, and for either the static or cyclic case. Use the appropriate equation for
ps. The two straight-line portions of the p-y curve, beyond the point where y is equal to
b/60, can now be established.
3.
4.
110
........................................................ (3-95)
2,000
500,000
1,500
kc (static)
400,000
kc (cyclic)
1,000
300,000
200,000
k (submerged)
500
100,000
0
0
deg.
28
32
36
40
c kPa
96
192
287
383
c tsf
as follows:
........................................................ (3-97)
....................................................... (3-98)
e. Compute appropriate number of points on the parabola by using Equation 3-94.
111
Depth = 2.00 m
Depth = 3.00 m
1,200
1,000
800
600
400
200
0
0.0
0.005
0.01
0.015
0.02
0.025
Lateral Deflection y, m
the response of piles under lateral loading in such soils where response is given principally by
deflection of the pile at the point of loading.
Data from one such experiment, however, was available and the writers have elected to
use that data in an example to demonstrate the use of this criterion. A comparison was made
there between results from experiment and results from computations.
The reader will note that the procedure presented above does not reflect a severe loss of
soil resistance under cyclic loading that is a characteristic for clays below a free-water surface.
Rather, the procedures described above are for a material that is primarily granular in nature,
which does not reflect such loss of resistance. Therefore, if a c- soil has a very low value of
and a relatively large value of c, the user is advised to ignore the
and to use the
recommendations for p-y curves for clay. Further, a relatively large factor of safety is
recommended in any case, and a field program of testing of prototype piles is certainly in order
for jobs that involve any large number of piles.
Furthermore, full-scale testing may be economical if a large number of piles are to be installed at
a particular site. Such field testing will add to the data bank and lead to improvements in the
recommendations shown below, which are to considered as preliminary because of the meager
amount of experimental data that is available.
In most cases of design, the deflection of the drilled shaft (or other kind of pile) will be so
small that the ultimate strength pur of the rock is not developed. However, the ultimate resistance
of the rock should be predicted in order to allow the computation of the lateral loading that
causes the failure of the pile. Contrary to the predictions of p-y curves for soil, where the unit
weight is a significant parameter, the unit weight of rock is neglected in developing the
prediction equations that follow. While a pile may move laterally only a small amount under the
working loads, the prediction of the early portion of the p-y curve is important because the small
deflections may be critical in some designs.
Most intact rocks are brittle and will develop shear planes at low shear strains. This fact
leads to an important concept about intact rock. The rock is assumed to fracture and lose strength
under small values of deflection of a pile. If the RQD of a stratum of rock is zero, or has a low
value, the rock is assumed to have already fractured and, thus, will deflect without significant
loss of strength. The above concept leads to the recommendation of two sets of criteria for rock,
one for strong rock and the other for weak rock. For the purposes of the presentations herein,
strong rock is assumed to have a compressive strength of 6.9 MPa (1,000 psi) or above.
The methods of predicting the response of rock is based strongly on a limited number of
experiments and on correlations that have been presented in technical literature. Some of the
correlations are inexact; for example, if the engineer enters the figure for correlation between
stiffness and strength with a value of stiffness from the pressuremeter, the resulting strength can
vary by an order of magnitude, depending on the curve that is selected. The inexactness of the
necessary correlations, plus the limited amount of data from controlled experiments, mean that
the methods for the analysis of piles in rock must be used with a good deal of both judgment and
caution. For major projects, full-scale load testing is recommended to verify foundation
performance and to evaluate the efficiency of proposed construction methods.
3-8-2 Descriptions of Two Field Experiments
3-8-2-1 Islamorada, Florida
An instrumented drilled shaft (bored pile) was installed in vuggy limestone in the Florida
Keys (Reese and Nyman, 1978) and was tested under lateral loads. The test was performed for
gaining information for the design of foundations for highway bridges.
Considerable difficulty was encountered in obtaining properties of the intact rock. Cores
broke during excavation and penetrometer tests were misleading because of the presence of vugs
or could not be performed. It was possible to test two cores from the site. The small
discontinuities in the outside surface of the specimens were covered with a thin layer of gypsum
cement in an effort to minimize stress concentrations. The ends of the specimens were cut with a
rock saw and lapped flat and parallel. The specimens were 149 mm (5.88 in.) in diameter and
with heights of 302 mm (11.88 in.) for Specimen 1 and 265 mm (10.44 in.) for Specimen 2. The
undrained shear strength values of the specimens were taken as one-half the unconfined
compressive strength and were 1.67 MPa (17.4 tsf) and 1.30 MPa (13.6 tsf) for Specimens 1 and
2, respectively.
114
The rock at the site was also investigated by in-situ-grout-plug tests (Schmertmann,
1977). In these tests, a 140-mm (5.5 in.) hole was drilled into the limestone, a high-strength steel
bar was placed to the bottom of the hole, and a grout plug was cast over the lower end of the bar.
The bar was pulled until failure occurred, and the grout was examined to see that failure occurred
at the interface of the grout and limestone. Tests were performed at three borings, and the results
shown in Table 3-8 were obtained. The average of the eight tests was 1.56 MPa (226 psi or 16.3
tsf). However, the rock was stronger in the zone where the deflections of the drilled shaft were
greatest and a shear strength of 1.72 MPa (250 psi or 18.0 tsf) was selected for correlation.
Table 3-8. Results of Grout Plug Tests by Schmertmann (1977)
Depth Range
meters
0.76-1.52
2.44-3.05
feet
2.5-5.0
8.0-10.0
5.49-6.10 18.0-20.0
Ultimate Resistance
MPa
psf
tsf
2.27
331
23.8
1.31
190
13.7
1.15
167
12.0
1.74
253
18.2
2.08
301
21.7
2.54
368
26.5
1.31
190
13.7
1.02
149
10.7
The bored pile was 1,220 mm (48 in.) in diameter and penetrated 13.3 m (43.7 ft) into the
limestone. The overburden of fill was 4.3 m (14 ft) thick and was cased. The load was applied at
3.51 m (11.5 ft) above the limestone. A maximum horizontal load of 667 kN (75 tons) was
applied to the pile. The maximum deflection at the point of load application was 18.0 mm (0.71
in.) and at the top of the rock (bottom of casing) it was 0.54 mm (0.0213 in.). While the curve of
load versus deflection was nonlinear, there was no indication of failure of the rock. Other details
about the experiment are shown in the Case Studies that follow.
3-8-2-2 San Francisco, California
The California Department of Transportation (Caltrans) performed lateral-load tests of
two drilled shafts near San Francisco (Speer, 1992). The results of these unpublished tests have
been provided by courtesy of Caltrans.
Two exploratory borings were made into the rock and sampling was done with a NWD4
core barrel in a cased hole with a diameter of 102 mm (4 in.). A 98-mm (3.88-in.) tri-cone roller
bit was used in drilling. The sandstone was medium to fine grained with grain sizes from 0.1 to
0.5 mm (0.004 to 0.02 in.), well sorted, and thinly bedded with thickness of 25 to 75 mm (1 to 3
in.). Core recovery was generally 100%. The reported values of RQD ranged from zero to 80,
with an average of 45. The sandstone was described by Speer (1992) as moderately to very
intensely fractured with bedding joints, joints, and fracture zones.
115
Pressuremeter tests were performed and the results were scattered. The results for moduli
values of the rock are plotted in Figure 3-43. The dashed lines in the figure show the average
values that were used for analysis. Correlations of RQD to modulus reduction ratio shown in
Figure 3-44 and the correlation of rock strength and modulus shown in Figure 3-45 were
employed in developing the correlation between the initial stiffness from Figure 3-43 and the
compressive strength, and the values were obtained as shown in Table 3-9.
Two drilled shafts, each with diameters of 2.25 m (7.38 ft), and with penetrations of 12.5
m (41 ft) and 13.8 m (45 ft), were tested simultaneously by pulling the shafts together. Lateral
loading was applied using hydraulic rams acting on high-strength steel bars that were passed
through tubes, transverse and perpendicular to the axes of the shafts. Lateral load was measured
using electronic load cells. Lateral deflections of the shaft heads were measured using
displacement transducers. The slope and deflection of the shaft heads were obtained by readings
from slope indicators.
The load was applied in increments at 1.41 m (4.6 ft) above the ground line for Pile A
and 1.24 m (4.1 ft) for Pile B. The pile-head deflection was measured at slightly different points
above the rock line, but the results were adjusted slightly to yield equivalent values for each of
the piles. Other details about the loading-test program are shown in the case studies that follow.
Initial Modulus, Eir, MPa
0
800
400
1,200
1,600
2,000
2
186 MPa
4
3.9 m
645 MPa
6
8
8.8 m
10
1,600 MPa
12
Figure 3-43 Initial Moduli of Rock Measured by Pressuremeter for San Francisco Load Test
116
1.2
1.0
0.8
0.6
0.4
0.2
?
?
?
0.0
0%
25%
50%
75%
100%
Compressive Strength
ft
MPa
psi
0.0 to 3.9
0.0 to 12.8
1.86
270
3.9 to 8.8
12.8 to 28.9
6.45
936
below 8.8
below 28.9
16.0
2,320
The rock below 8.8 m (28.9 ft) is in the range of strong rock, but the rock
above that depth will control the lateral behavior of the drilled shaft.
117
(MPa)
1
10
Rock Strength
Classification
(Deere)
100
1,000
Very Low
Low
Medium
High
Very High
100
Upper and
Middle Chalk
(Hobbs)
Concrete
10
(MPa)
Steel
100,000
Gneiss
1.0
Grades
of Chalk
(Ward et al.)
I
II
III
0.1
Limestone,
Dolomite
Basalt and other
Flow Rocks
Lower
Chalk
(Hobbs)
Deere
10,000
Sandstone
1,000
Trias (Hobbs)
IV
V
Keuper
100
Black Shale
0.01
Grey Shale
Hendron, et al.
10
Medium
0.001
Stiff
Very Stiff
Hard
0.01
0.1
Clay
1
1.0
100
10
psi
103
118
1,000
3-8-3 Procedure for Computing p-y Curves for Strong Rock (Vuggy Limestone)
The p-y curve recommended for strong rock (vuggy limestone), with compressive
strength of intact specimens larger than 6.9 MPa (1,000 psi), shown in Figure 3-46. If the rock
increases in strength with depth, the strength at the top of the stratum will normally control.
Cyclic loading is assumed to cause no loss of resistance.
As shown in the Figure 3-46, load tests are recommended if deflection of the rock (and
pile) is greater than 0.0004b and brittle fracture is assumed if the lateral stress (force per unit
length) against the rock becomes greater than half the diameter times the compressive strength of
the rock.
The p-y curve shown in Figure 3-46 should be employed with caution because of the
limited amount of experimental data and because of the great variability in rock. The behavior of
rock at a site could be controlled by joints, cracks, and secondary structure and not by the
strength of intact specimens.
Perform proof test if
deflection is in this range
pu = b su
Assume brittle fracture if
deflection is in this range
Es = 100su
Es = 2000su
NOT TO SCALE
y
0.0004b
0.0024b
119
p
Mir
pur
yA
Figure 3-47 Sketch of p-y Curve for Weak Rock (after Reese, 1997)
for xr
3b ...................................... (3-99)
........................................... (3-100)
where:
qur = compressive strength of the rock, usually lower-bound as a function of depth,
r
b =
xr =
The assumption is made that fracturing will occur at the surface of the rock under small
deflections, therefore, the compressive strength of intact specimens is reduced by multiplication
by r to account for the fracturing. The value of r is assumed to be 1.0 at RQD of zero and to
decrease linearly to a value of one-third for an RQD value of 100%. If RQD is zero, the
compressive strength may be obtained directly from a pressuremeter curve, or approximately
from Figure 3-45, by entering with the value of the pressuremeter modulus.
................................................ (3-101)
120
If one were to consider a strip from a beam resting on an elastic, homogeneous, and
isotropic solid, the initial modulus Mir (pi divided by yi) in Figure 3-47 may be shown to have the
following value (using the symbols for rock). 2
Mir
where
Eir = the initial modulus of the rock, and
kir = dimensionless constant defined by Equation 3-103.
Equations 3-102 and 3-103 for the dimensionless constant kir are derived from data available
from experiment and reflect the assumption that the presence of the rock surface will have a
similar effect on kir as was shown for pur for ultimate resistance.
.................................... (3-103)
kir = 500 for xr > 3b................................................ (3-104)
With guidelines for computing pur and Mir, the equations for the three branches of the
family of p-y curves for rock in Figure 3-46 can be presented. The equation for the straight-line,
initial portion of the curves is given by Equation 3-105 and for the other branches by Equations
3-106 through 3-108.
for
for
...............................................(3-105)
...............................(3-106)
rm
b.........................................................(3-108)
where
rm
= a constant, typically ranging from 0.0005 to 0.00005 that serves to establish the upper
limit of the elastic range of the curves using Equation 3-108. rm is analogous to 50
used for p-y curves in clays. The stress-strain curve for the uniaxial compressive test
may be used to determine rm in a similar manner to that used to determined 50.
The value of yA is found by solving for the intersection of Equations 3-105 and 3-106, and the
solution is presented in Equation 3-109.
The notation used here for Mir and rm differs from that used in Reese (1997). The notation was
changed to improve the clarity of the presentation.
121
.............................................(3-109)
As shown in the case studies that follow, the equations from weak rock predict with
reasonable accuracy the behavior of single piles under lateral loading for the two cases that are
available. An adequate factor of safety should be employed in all cases.
The equations are based on the assumption that p is a function only of y. This assumption
appears to be valid if loading is static and resistance is only due to lateral stresses. However,
(1996) noted
-pull shear
produced by the axial shears caused by the rotation of the pile. In rock, this effect could be
significant, especially for small deflections, if the diameter of the pile is large
3-8-5 Case Histories for Drilled Shafts in Weak Rock
3-8-5-1 Islamorada
The drilled shaft was 1.22 m (48 in.) diameter and penetrated 13.3 m (43.7 ft) into
limestone. A layer of sand over the limestone was retained by a steel casing, and the lateral load
was applied at 3.51 m (11.5 ft) above the surface of the rock. A maximum lateral load of 667 kN
(150 kips) was applied and the measured curve of load versus deflection was nonlinear.
Values of the strengths of the concrete and steel were unavailable and the bending
stiffness of the gross section was used for the initial solutions. The following values were used to
compute the p-y curves:
qur = 3.45 MPa (500 psi),
r
106 psi),
= 0.0005,
109 ksi).
A comparison of pile-head deflection curves from experiment and from analysis is shown
in Figure 3-48. Excellent agreement between the elastic EI and experiment and is found for
loading levels up to about 350 kN (78.7 kips), where sharp change in the load-deflection curve
occurs. Above that level of loading, nonlinear EI is required to match the experimental values
reasonably well.
Curves giving deflection and bending moment as a function of depth were computed for a
lateral load of 334 kN (75 kips), one-half of the ultimate lateral load, and are shown in Figure 349. The plotting is shown for limited depths because the values to the full length are too small to
plot. The stiffness of the rock, compared to the stiffness of the pile, is reflected by a total of 13
points of zero deflection over the length of the pile of 15.2 meters (50 ft). However, for the data
employed here, the pile will behave as a long pile through the full range of loading.
122
400
800
1,200
M
2
y
Rock Surface
8
1
Lateral Deflection, y, mm
Figure 3-49 Computed Curves of Lateral Deflection and Bending Moment versus Depth,
Islamorada Test, Lateral Load of 334 kN (after Reese, 1997)
The data on deflection as a function of loads showed that the two piles behaved about the
same for the beginning loads but the curve for Pile B exhibited a large increase in pile-head
deflection at the largest load. The experimental curve for Pile B shown by the heavy solid line in
Figure 3-50 suggests that a plastic hinge developed at the ultimate bending moment of 17,740 mkN (157,012 in-kips).
Consideration was given to the probable reduction in the values of EI with increasing
load and three methods were used to predict the reduced values. The three methods were: the
analytical method as presented in Chapter 4, the approximate method of the American Concrete
Institute (ACI 318) which does not account for axial load and may be used here; and the
experimental method in which EI is found by trial-and-error computations that match computed
and observed deflections. The plots of the three methods are shown in Figure 3-51 and all three
curves show a sharp decrease in EI with increase in bending moment. For convenience in the
computations, the value of EI was changed for the entire length of the pile but errors in using
constant values of EI in the regions of low values of M are thought to be small.
The computed and measured lateral load versus pile-head deflection curves are shown in
Figure 3-50. The computed load-deflection curve computed using EI values derived from the
load test agrees well with the load test curve, but the computed load-deflection curves using
of 2.0 and higher are selected, the computed deflections would be about 2 or 3 mm (0.078 to
0.118 in.) with the experiment showing about 4 mm (0.157 in.). Thus, the differences are
probably not very important in the range of the service loading.
124
10,000
8,000
Pile B
6,000
4,000
Unmodified EI
Analytical
ACI
Experimental
2,000
0
0
10
20
30
40
50
Groundline Deflection, mm
Figure 3-50 Comparison of Experimental and Computed Values of Pile-Head Deflection for
Different Values of EI, San Francisco Test
Also shown in Figure 3-50 is a curve showing deflection as a function of lateral load with
no reduction in the values of EI. The need to reduce EI as a function of bending moment is
apparent.
Values of bending stiffness in Figure 3-51 along with EI of the gross section were used to
compute the maximum bending moment mobilized in the shaft as a function of the applied load
are shown in Figure 3-52. The close agreement between computations from all the methods is
striking. The curve based on the gross value of EI is reasonably close to the curves based on
adjusted values of EI, indicating that the computation of bending moment for this particular
example is not very sensitive to the selected values of bending stiffness.
125
40
Analytical
Experimental
ACI
30
20
10
0
5,000
10,000
15,000
20,000
7,500
5,000
Unmodified EI
Analytical
ACI
Experimental
2,500
0
0
5,000
10,000
15,000
20,000
126
where pu is the ultimate lateral resistance of the rock mass and Ki is the initial slope of the p-y
curve. A drawing of the p-y curve for massive rock is presented in Figure 3-53.
p
pu
Ki
y
Figure 3-53 p-y Curve in Massive Rock
3-9-1 Determination of pu Near Ground Surface
For a passive wedge type failure near the ground surface, as shown in Figure 3-54, the
ultimate lateral resistance per unit length, pu of the drilled shaft at depth H is
.............................(3-111)
where
, c = effective cohesion,
effective unit weight respectively of the rock mass and the following equations are used to
compute parameters C1 through C5:
127
Fs
Fp
Fn
, and
, with the condition that
128
Equation 3-111 is valid for homogeneous rock mass. For layered rock mass,
representative properties can be computed by a weighted method based on the volume of the
failure wedge. Methods for obtaining the rock properties c and are given on page 129.
3-9-2 Rock Mass Failure at Great Depth
The passive wedge failure mechanism is not likely to form if the overburden pressure is
sufficiently large. Studies of rock sockets using three-dimensional stress analysis using the finite
element method have concluded that at depth the rock failure first in tension, followed by failure
in friction between the shaft and rock, followed finally by failure of the rock in compression.
Therefore, the expression for ultimate resistance at depth is a function of the limiting pressure,
pL, and the peak frictional resistance max. The ultimate resistance at depth can be computed
using
...........................................(3-112)
where pa is the active horizontal active earth pressure given by
with the condition that
........................(3-113)
= effective overburden pressure at the depth under consideration including the pressure from
overburden soils, pL is the limiting normal pressure of the rock mass (discussed later), and max is
the axial side resistance of the rock-shaft interface, proposed by Kulhawy and Phoon (1993)
V
.....................................................(3-114)
where both
max
and
ci
............................................(3-116)
129
where 1 and 3 are the major and minor principal stresses at failure, ci is the uniaxial
compressive strength of intact rock, and mb, s, and a are material constants that depend on the
characteristics of the rock mass; s = 1 for intact rock, and a = 0.5 for most rock types.
and
Hoek (1990) provided a method for estimating the Mohr-Coulomb failure parameters c
of the rock mass from the principal stresses at failure. These parameters are:
.............................................(3-117)
....................................................(3-118)
1
...........................................(3-120)
The parameters mb and s can be determined for many types of rock using the
recommendations of Marinos and Hoek (2000).3
Two methods for evaluating rock mass modulus are recommended by Liang et al. One
method is to compute rock mass modulus by multiplying the intact rock modulus measured in
the laboratory by the modulus reduction ratio, Em/Ei, computed using the geological strength
index, GSI., using Equation 3-121
..................................................(3-121)
The modulus reduction ratio and is shown as a function of GSI in Figure 3-55.
The second method recommended for determining rock mass modulus is to perform an
in-situ rock pressuremeter test. The difficulty in using this approach is that many pressuremeter
testing devices are not capable of reaching large pressures, so difficulties might arise during their
use. In addition, interpretation of test results may be difficult because of the limited range of
expansion pressures possible.
130
100
Bieniawski (1978)
Serafin and Pereira (1983)
Ironton-Russell
Regression Line
80
60
40
20
20
40
60
80
100
ci
2. Obtain values for the rock mass modulus, Em, by use of Equation 3-121 if pressuremeter data
are unavailable. If Equation 3-121 is used, obtain values of GSI and mi according to the
recommendations of Marinos and Hoek (2000) .
3. Select a shaft diameter and reinforcing detail.
4. Compute the bending stiffness and nominal moment capacity of the drilled shaft. Set the
value of bending stiffness equal to the cracked section bending stiffness at a level of loading
where the reinforcement is in the elastic range.
5. Compute Ki using Equation 3-115.
6. Compute pu at shallow depth using Equation 3-111 with 3 equal to the vertical effective
stress at H/3 when computing the values of and c using Equations 3-117 and 3-118.
7. Compute pu at great depth using Equation 3-112 with pL taken as
Equation 3-116 and equating 3 equal to v.
computed using
8. Compute pu as the smaller of the values computed by Equations 3-111 and 3-112.
9. The values of the p-y curve can then be computed using 3-110 for selected values of pile
movement y.
131
y/b
0.0375..........................(3-123)
.................................................(3-124)
132
pu
y
0.001b
0.0375b
modifications would be needed in the method to compute the ultimate soil resistance pu, and
consequently modifications would be needed in the p-y curves.
The problem of the layered soil has been given intensive study by Allen (1985); however,
the methods
developed by Allen with the methods shown herein must be delayed until a later date when this
research can be put in a readily usable form.
3-11-1 Layering Correction Method of Georgiadis
The method of Georgiadis (1983) is ba
of all the layers existing below the upper layer. The p-y curves of the upper layer are determined
according to the methods for homogeneous soils. To compute the p-y curves of the second layer,
the equivalent depth H2 to the top of the second layer has to be determined by summing the
ultimate resistances of the upper layer and equating that value to the summation as if the upper
layer had been composed of the same material as in the second layer. The values of pu are
computed using the equations for homogeneous soils. Thus, the following two equations are
solved simultaneously for H2.
.................................................... (3-125)
and
......................................................(3-126)
The equivalent thickness H2 of the upper layer along with the soil properties of the second layer,
are used to compute the p-y curves for the second layer.
The concepts presented above can be used to get the equivalent thickness of two or more
dissimilar layers of soil overlying the layer for whom the p-y curves are desired. One possible
consequence is that the equivalent depths may be either smaller or greater than the actual depths
of the soil layers, depending on the relative strengths of the layers of the soil profiles. This is
illustrated in Figure 3-58.
3-11-2 Example p-y Curves in Layered Soils
The example problem to demonstrate the manner in which layered soils are modeled is
shown in Figure 3-59. As seen in the sketch, a pile with a diameter of 610 mm (24 in.) is
embedded in soil consisting of an upper layer of soft clay, overlying a layer of loose sand, which
in turn overlays a layer of stiff clay. The water table is at the ground surface, and the loading is
static.
134
Groundline
h3
h1
F1
F2
Fi
1.73 m
Soft Clay
1.32 m
Loose Sand
6.1 m
Stiff Clay
c = 23.9 kPa
50 = 0.02
= 7.9 kN/m3
= 30 deg.
= 7.9 kN/m3
c = 95.8 kPa
50 = 0.005
= 9.4 kN/m3
k = 20,400 kPa
Static Loading
0.61 m
135
Four p-y curves for the case of layered soil are shown in Figure 3-60. The curves are for
points A, B, C and D as shown in the sketch in Figure 3-61, at depths of 0.92 m (36 in.), 1.83 m
(72 in.), 3.66 m (144 in.), and 7.32 m (288 in.), respectively. The curve at a depth of 0.92 m (36
in.) falls in the upper zone of soft clay; the curve for the depth of 1.83 m (72 in.) falls in the sand
just below the soft clay; and the curves for depths of 3.66 m (144 in.) and 7.32 m (288 in.) fall in
the lower zone of stiff clay.
400
350
300
Sof t Clay, x = 0.92 m
Sand, x = 1.83 m
250
200
150
100
50
0
0.0
0.01
0.02
0.03
Lateral Deflection y, meters
0.04
0.05
A
xEQ = 2.057 m
B
1.73 m
Loose
Sand
3.05 m
xEQ = 1.816 m
Stiff
Clay
D
Actual
Depth, m
Equivalent
Depth, m
0.92
0.92
1.83
2.057
3.66
1.816
7.32
5.476
9.14 m
0.61 m
Figure 3-61 Equivalent Depths of Soil Layers Used for Computing p-y Curves
136
Following the method suggested by Georgiadis, the p-y curve for soft clay can be
computed as if the profile consists altogether of that soil. When dealing with the sand, an
equivalent depth of sand is found such that the integrals of the ultimate soil resistance of an
equivalent sand layer and for the soft clay are equal at the interface. The equivalent thickness of
loose sand to replace the 1.73 m (68 in.) of soft clay was found to be 1.88 meters (74 in.). Thus,
the equivalent depth to point B in loose sand is 1.98 meters (78 in.). A plot of the integrals of
ultimate soil resistance and equivalent depths is presented in Figure 3-61.
An equivalent depth of stiff clay was found such that the sum of the ultimate soil
resistance for the stiff clay is equal to the sum of the ultimate soil resistance of the loose sand
and soft clay. In making the computation, the equivalent and actual thicknesses of the loose sand,
1.88 m (74 in.) and 1.32 m (52 in.), respectively, were replaced by 1.14 m (45 in.) of stiff clay.
Thus, the actual thicknesses of the soft clay and loose sand of 3.05 m (120 in.) were reduced by
1.91 m (75 in.), leading to equivalent depths in the stiff clay of points C and D of 1.75 m (69 in.)
and 5.41 m (213 in.), respectively (Figure 3-61).
Another point of considerable interest is that the recommendations for p-y curves for stiff
clay in the presence of no free water were used for the stiff clay. This decision is based on the
assumption that the sand above the stiff clay can move downward and fill any gap that develops
between the clay and the pile. Furthermore, in the stiff-clay experiment where free water was
present, the free water moved upward along the face of the pile with each cycle of loading. The
presence of soft clay and sand to a depth of 3.05 m (120 in.) above the stiff clay is believed to
suppress the hydraulic action of free water even though the sand did not serve to close the
potential gaps in the stiff clay.
The equations used to compute lateral load transfer at failure are the ultimate values.
Soft Clay static loading
.............................................. (3-20)
.......................................................... (3-21)
Soft Clay cyclic loading
....................................................... (3-24)
..................................................... (3-25)
Stiff Clay with Free Water Static
pct = 2cab + bx + 2.83 cax ............................................ (3-26)
pcd = 11cb ......................................................... (3-27)
137
...................................... (3-35)
Stiff Clay with Free Water Cyclic
........................................... (3-40)
Stiff Clay without Free Water static and cyclic loading
...............................................(3-20)
...........................................................(3-21)
Sand
..................... (3-50)
............................... (3-51)
or
............................................... (3-52)
API Sand
................................................... (3-61)
........................................................ (3-62)
138
3-12 Modifications to p-y Curves for Pile Batter and Ground Slope
3-12-1 Piles in Sloping Ground
The formulations for p-y curves presented to this manual were developed for a horizontal
ground surface. In order to allow designs to be made if a pile is installed on a slope,
modifications must be made to the p-y curves. The modifications involve revisions in the manner
in which the ultimate soil resistance is computed. In this regard, the assumption is made that the
flow-around failure that occurs at depth will not be influenced by sloping ground; therefore, only
the equations for the wedge-type failures near the ground surface need modification.
The modifications to p-y curves presented here are based on earth pressure theory and
should be considered as preliminary. Future changes may be needed once laboratory and field
study are completed.
3-12-1-1 Equations for Ultimate Resistance in Clay in Sloping Ground
The ultimate soil resistance near the ground surface for saturated clay where the pile was
installed in ground with a horizontal slope was derived by Reese (1958) and is shown in
Equation 3-127.
....................................... (3-127)
If the ground surface has a slope angle as shown in Figure 3-62, the soil resistance at the front
of the pile, following the Reese approach is:
139
....................................... (3-128)
The soil resistance at the back of the pile is:
......................... (3-129)
where:
(pu)ca = ultimate soil resistance near ground surface,
ca =
b =
pile diameter,
=
H =
=
A comparison of Equations 3-127 and 3-128 shows that the equations are identical except for the
terms at the right side of the parenthesis. If is equal to zero, the equations become equal to the
original equation.
3-12-1-2 Equations for Ultimate Resistance in Sand
The ultimate soil resistance near the ground surface for sand where the pile was installed
in ground with a horizontal slope was derived earlier and is:
.............. (3-130)
If the ground surface has a slope angle , the ultimate soil resistance in the front of the pile is:
( pu ) sa
K 0 H tan sin
(4 D13 3D12 1)
tan(
) cos
tan
tan(
bD2
140
............... (3-131)
................................................. (3-132)
D2
where
Note that the denominator of Equation 3-132 for D1 will equal zero when the sum of the
slope and friction angles is 90 degrees. This occurs when the inclination of the failure wedge is
parallel to the ground surface. In computations, the lower value of (pu)sa or to pu from Equation
3-51 is used, so no computational problem arises.
The ultimate soil resistance in the back of the pile is:
( pu ) sa
K 0 H tan sin
(4 D33
tan(
) cos
tan
tan(
bD4
3D32
1)
............. (3-135)
3D32
1)
K Ab
where
................................................ (3-136)
and
D4 = 1 + D3....................................................... (3-137)
This completes the necessary derivations for modifying the equations for clay and sand to
analyze a pile under lateral load in sloping ground.
3-12-1-3 Effect of Direction of Loading on Output p-y Curves
The equations for computing maximum soil resistance for p-y curves in sand depend on
whether the pile is being pushed up or down the slope. LPile determines which case to compute
by using the values of lateral pile deflection and slope angle. Whenever, p-y curves are generated
for output, the curve that is output by the program is based on the lateral deflection computed for
loading case 1. If the user desires output of both sides of an unsymmetrical p-y curve it is
necessary to run an analysis twice, with the pile-head loadings for shear, moment, rotation, or
displacement reversed for the two analyses, while keeping the axial thrust force unchanged. The
user may then combine the two output curves together.
141
30
20
10
10
20
30
Load
1.0
0
30
20
10
10
20
30
143
Chapter 4
Special Analyses
4-1 Introduction
LPile has several options for making special analyses. This chapter provides explanations
about the various options and guidance for using the optional features for making special
analyses.
144
Sand, 9 m
M=0
c = 12 to 24 kPa
= 8.95 kN/m3
= 38 to 40
= 9.50 kN/m3
Figure 4-2 Variation of Top Deflection versus Depth for Example Problem
145
200
150
100
50
0
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
0.016
0.018
0.02
0.022
0.024
Top Deflection, m
pile tip is sufficiently embedded in a strong layer of soil or rock. In this case, the designer must
decide how reliably the depth of the strong layer can be predicted. In such a case, the designer
may wish to specify the length for a drilled foundation to be long enough to penetrate into the
strong layer after considering the variability of the depth to the strong layer and add a
requirement for the construction inspector to notify the design engineer if the strong layer is not
encountered in the field after drilling to the full depth. In the case of a driven pile foundation, the
design engineer can set the pile length to be long enough to reach a specified driving resistance
that is based a pile driving analysis that is based on the presence of the strong layer.
147
ps
y
y ys
ys
Epy
148
7. Increase the magnitude of axial thrust force in even increments for the subsequent load
cases. An initial increment size may be 5 percent of the axial structural capacity. Up to
100 load steps may be specified.
8. Perform the analysis with the option for pile buckling analysis.
9. Examine the output report and pile buckling graph.
An example buckling study was performed. The pile head is at the elevation of the
ground surface. The soil profile is sand from 0 to 2 meters (API sand, = 18 kN/m3, = 30
degrees, and k = 13,550 kN/m3), soft clay from 2 to 8.5 meters ( = 7.19 kN/m3, c = 1 kPa, 50 =
0.06), and sand below 8.5 meters (API sand, = 10 kN/m3, = 40 degrees, k = 60,000 kN/m3).
The pile has a diameter of 0.15 meters, a length of 18 meters, a cross-sectional area of 0.0177
m2, a moment of inertia of 1.678 10-7
s modulus of 200 GPa. Two curves are
plotted in Figure 3-6. For one curve, the specified shear force is 0.1 kN and buckling failure
occurs for thrust values above 218 kN. For the second curve, the specified shear force is 1.0 kN
and buckling failure occurs for thrust values above 121 kN. This graph illustrates that the
buckling capacity is a function of the pile head loading conditions, with a lower capacity
associated with a greater loading condition.
250
V = 0.1 kN
200
V = 1.0 kN
150
100
50
0
0
0.002
0.004
0.006
0.008
0.01
Figure 4-6 Examples of Pile Buckling Curves for Different Shear Force Values
These curves illustrate that the axial buckling capacity is a function of the specified
lateral shear force used in the analysis and that the buckling capacity is reduced as the lateral
shear force is increased. Thus, it is important to use the maximum expected load condition, if it is
known, since a range of computed buckling capacities is possible.
149
Incorrect
350
300
250
200
150
100
50
0
-0.5
-0.4
-0.3
-0.2
-0.1
0.1
0.2
150
y0
Pile-head Deflection, y
151
a
1
b
y
y0
152
153
Formation of
plastic hinge
Figure 4-11 Pile-head Shear Force versus Displacement from Pushover Analysis
Formation of
plastic hinge
Figure 4-12 Maximum Moment Developed in Pile versus Displacement from Pushover Analysis
In general, it is not possible to develop more than one plastic hinge in a pile if the pilehead condition is pinned. It is sometimes possible to develop two plastic hinges in the pile if the
pile-head condition is fixed head and the axial load is zero.
154
155
156
Chapter 5
Computation of Nonlinear Bending Stiffness
and Moment Capacity
5-1 Introduction
5-1-1 Application
The designer of deep foundations under lateral loading must make computations to
ascertain that three factors of performance are within tolerable limits: combined axial and
bending stress, shear stress, and pile-head deflection. The flexural rigidity, EI, of the deep
foundation (bending stiffness) is an important parameter that influences the computations (Reese
and Wang, 1988; Isenhower, 1994).
In general, flexural rigidity of reinforced concrete varies nonlinearly with the level of
applied bending moment, and to employ a constant value of EI in the p-y analysis for a concrete
pile will result in some degree of inaccuracy in the computations.
The response of a pile is nonlinear with respect to load because the soil has nonlinear
stress-strain characteristics. Consequently, the load and resistance factor design (LRFD) method
is recommended when evaluating piles as structural members. This requires evaluation of the
nominal (i.e. unfactored) bending moment of the deep foundation.
Special features in LPile have been developed to compute the nominal-moment capacity
of a reinforced-concrete drilled shaft, prestressed concrete pile, or steel-pipe pile and to compute
the bending stiffness of such piles as a function of applied moment or bending curvature. The
designer can utilize this information to make a correct judgment in the selection of a
representative EI value in accordance with the loading range and can compute the ultimate lateral
load for a given cross-section.
5-1-2 Assumptions
The program computes the behavior of a beam or beam-column. It is of interest to note
that the EI of the concrete member will undergo a significant change in EI when tensile cracking
occurs. In the coding used herein, the assumption is made that the tensile strength of concrete is
minimal and that cracking will be closely spaced when it appears. Actually, such cracks will
initially be spaced at some distance apart and the change in the EI will not be so drastic. In
respect to the cracking of concrete, therefore, the EI for a beam will change more gradually than
is given by the coding.
The nominal bending moment of a reinforced-concrete section in compression is
computed at a compression-control strain limit in concrete of 0.003 and is not affected by the
crack spacing. The ultimate bending moment for steel, because of the large amount of
deformation of steel when stressed about the proportional limit, is taken at a maximum strain of
0.015 which is five times that of concrete.
157
fc
0.15 f c
Ec
0.0038
fr
Figure 5-1 Stress-Strain Relationship for Concrete Used by LPile
The following equations are used to compute concrete stress. The value of concrete
compressive strength, f c, in these equations is specified by the engineer.
for
for
.......................................(5-1)
............................(5-2)
The modulus of rupture, fr, is the tensile strength of concrete in bending. The modulus of
rupture for drilled shafts and bored piles is computed using
.............................................(5-3)
.............................................(5-4)
0,
is computed using
............................................................(5-6)
The tensile strain at fracture for concrete, t, is computed using
...................................................(5-7)
The stress-strain ( - ) curve for steel is shown in Figure 5-2. There is no practical limit to
plastic deformation in tension or compression. The stress-strain curves for tension and
compression are assumed identical in shape.
fy
159
The yield strength of the steel, fy, is selected according to the material being used, and the
following equations apply.
..............................................................(5-8)
where Es = 200,000 MPa (29,000,000 psi).
The models and the equations shown here are employed in the derivations that are shown
subsequently.
5-1-4 Cross Sectional Shape Types
The following types of cross sections can be analyzed:
1. Square or rectangular, reinforced concrete,
2. Circular, reinforced concrete,
3. Circular, reinforced concrete, with permanent steel casing,
4. Circular, reinforced concrete, with permanent steel casing and tubular core,
5. Circular, steel pipe,
6. Round prestressed concrete
7. Round prestressed concrete with hollow circular core,
8. Square prestressed concrete,
9. Square prestressed concrete with hollow circular core,
10. Octagonal prestressed concrete,
11. Octagonal prestressed concrete with hollow circular core,
12. Elastic shapes with rectangular, round, tubular, strong H-sections, or weak H-sections,
and
13. Elastic-plastic shapes with rectangular, round, tubular, strong H-sections, or weak Hsections.
The computed output consists of a set of values for bending moment M versus bending
stiffness EI for different axial loads ranging from zero to the axial-load capacity for the column.
160
d
a
dx
M
c
Figure 5-3 Element of Beam Subjected to Pure Bending
The following equality is derived from the geometry of similar triangles
161
.............................................................(5-10)
where:
= distance from the neutral axis, and
= radius of curvature.
Equation 5-11 is obtained from Equations 5-9 and 5-10, as follows:
.................................................(5-11)
...........................................................(5-12)
where:
x
E=
s modulus.
162
d and
.........................................................(5-17)
For convenience here, the symbol is substituted for the curvature 1/ . The following equation
is developed from this substitution and Equations 5-16 and 5-17
............................................................(5-18)
and because
d and
.............................................................(5-19)
163
Common design practice in North America and Europe is to restrict the steel
reinforcement to be between 1 and 8 percent of the gross cross-sectional area for drilled shafts
without permanent casing. Usually, reinforcement percentages higher than 3.5 to 4 percent are
attainable only by a combination of bundling of bars and by reducing the maximum aggregate
size to be small enough to pass through the reinforcement cage. LPile has features that help the
user to identify the combinations of reinforcement details that satisfy requirement for
constructability.
For prestressed concrete piles, the equations for the nominal axial structural capacity
differ depending on the cross-sectional shape and the level of prestressing. As for uncased
reinforced concrete sections, the concrete stress at failure is assumed to be 0.85 f c. With axial
loading, the effective prestress in the section is lowered. At a compressive strain of 0.003, only
about 60 percent of the prestressing remains in the member. Thus, the nominal strength can be
computed as
...............................................(5-21)
where fpc is the effective prestress.
The service load capacity for short column piles established by the Portland Cement
Association is based on a factor of safety between 2 and 3 is
...............................................(5-22)
Conventional construction practice in North American is to use effective prestressing of
600 to 1,200 psi (4.15 to 8.3 MPa) for driven piling. The level of prestressed used varies with the
overall length of the pile and local practice. Usually, the designing engineer obtains the value of
prestress and fraction of losses from the pile supplier.
164
increments.4
The fifth column of the output shows the value of the position of the neutral axis, as
measured from the compression side of the member. Other columns in the output, for each value
of , give the bending moment, the EI, and the maximum compressive strain in the concrete. For
the validation that follows, only one line of output was selected.
0.510 m
0.076 m
0.203 m
0.203 m
0.760 m
0.203 m
0.076 m
LPile uses an algorithm to compute the initial increment of curvature that is based on the depth of the pile section. This algorithm is designed to
obtain initial values of curvature small enough to capture the uncracked behavior for all pile sizes.
165
=
=
=
=
=
=
=
=
=
=
=
=
15.24000000
0.76000000
0.51000000
10
413686.
199948000.
27600.
-39.40177573
0.38760000
0.00500000
1.28998971
9093.096
m
m
m
bars
kPa
kPa
kPa
kPa
sq. m
sq. m
percent
kN
Bar
Index
-----------16
16
16
16
16
16
16
16
16
16
Bar Diam.
m
-----------0.025200
0.025200
0.025200
0.025200
0.025200
0.025200
0.025200
0.025200
0.025200
0.025200
Bar Area
sq. m
-----------0.000500
0.000500
0.000500
0.000500
0.000500
0.000500
0.000500
0.000500
0.000500
0.000500
Bar X
m
------------0.167500
0.000000
0.167500
-0.167500
0.167500
-0.167500
0.167500
-0.167500
0.000000
0.167500
Bar Y
m
-----------0.304800
0.304800
0.304800
0.101600
0.101600
-0.101600
-0.101600
-0.304800
-0.304800
-0.304800
Concrete Properties:
Compressive Strength of Concrete
Modulus of Elasticity of Concrete
Modulus of Rupture of Concrete
Compression Strain at Peak Stress
Tensile Strain at Fracture
Maximum Coarse Aggregate Size
=
=
=
=
=
=
27600.
24865024.
-3271.7136591
0.0018870
-0.0001154
0.0190500
kPa
kPa
kPa
m
900.000 kN
Bending
Bending
Bending
Depth to
Max Comp
Max Tens
Max Concrete
Max Steel
Curvature
Moment
Stiffness
N Axis
Strain
Strain
Stress
Stress
rad/m
kN-m
kN-m2
m
m/m
m/m
kPa
kPa
------------- ------------- ------------- ------------- ------------- ------------- ------------- ------------0.0000492
28.3173948
575409.
1.9085538
0.0000939
0.0000565 2674.0029283
18743.
0.0000984
56.6333321
575395.
1.1451716
0.0001127
0.0000379 3188.4483827
22462.
.
. (deleted lines)
166
Run
Msg
---
0.0004429
0.0004921
0.0005413
0.0005906
253.1619332
280.6180646
280.6180646
280.6180646
.
. (deleted lines)
.
0.0038878
651.6508321
0.0039862
663.0531399
0.0040846
674.4235902
0.0041831
685.7618089
.
. (deleted lines)
.
0.0176673
907.1915259
.
. (deleted lines)
.
0.0239665
913.9027316
571583.
570216.
518378.
475180.
0.5542915
0.5375669
0.4727569
0.4548249
0.0002455
0.0002646
0.0002559
0.0002686
-0.0000911
-0.0001095
-0.0001555
-0.0001802
6671.6631466
7149.3433542
6926.7437852
7241.7196541
48751.
52522.
50760.
53257.
167614.
166336.
165112.
163937.
0.2450564
0.2440064
0.2430210
0.2420960
0.0009527
0.0009727
0.0009927
0.0010127
-0.0020020
-0.0020569
-0.0021117
-0.0021664
20619.
20904.
21183.
21458.
-397341.
-408237.
-413686.
-413686.
C
C
CY
CY
51349.
0.1701205
0.0030056
-0.0104216
27596.
413686.
CY
38132.
0.1658249
0.0039742
-0.0142403
27600.
413686.
CY
C
C
Axial Thrust
kN
---------------900.000
Max. Comp.
Strain
-----------0.00300000
Note note that the values of moment capacity in the table above are not
factored by a strength reduction factor (phi-factor).
In ACI 318-08, the value of the strength reduction factor depends on whether the
transverse reinforcing steel bars are spirals or tied hoops.
The above values should be multiplied by the appropriate strength reduction
factor to compute ultimate moment capacity according to ACI 318-08, Section 9.3.2.2
or the value required by the design standard being followed.
0.0755 m) = +0.001672
Similarly,
2
= 0.001915
= 0.005501
= 0.009088
In order to obtain the forces in the steel at each level, it is necessary to know if the steel is
in the elastic or plastic range. Thus, it is required to compute the value of yield strain y using
Equation 5-8.
..........................................(5-23)
167
This computation shows that the bars in rows 1 and 2 are in the elastic range and the bars in the
other two rows are in the plastic range. Thus, the forces in each row of bars are:
F1 = (3 bars) (5
108 kPa) =
501.51 kN
F2 = (2 bars) (5
382.95 kN
F3 = (2 bars) (5
413.00 kN
F4 = (3 bars) (5
619.50 kN
913.95 kN.
The second value in the parentheses is the distance from the neutral axis to the mid-height of the
first slice. Similarly, the strains at the centers of the other slices are:
2
= 0.002554
= 0.002254
= 0.001954
= 0.001653
= 0.001353
= 0.001052
= 0.000751
= 0.000451
10
= 0.000150
The forces in the concrete are computed by employing Figure 5-4 and Equations 5-1
through 5-8. The first step is to compute the value of 0 from Equation 5-6 and to see the strains
are lower or greater than the strain for the peak stress.
The strain in the top two slices show that stress can be found by use of the second branch
of the compressive portion of the curve in Figure 5-1 and the stress in the other slices can be
computed using Equation 5-1. From Figure 5-4, the following quantity is computed
168
Then, the following equation can be used to compute the stress along the descending section of
the stress-strain curve corresponding to 1 and 2.
f c3
27,600 2
0.001870
0.001870
169
Fc9 = 101.53 kN
Fc10 =
36.93 kN
There is a small section of concrete in tension. The depth of the tensile section is
determined by the strains up to the strain developed at the modulus of rupture (Equation 5-3).
In this zone, it is assumed that the stress-stain curve in tension is defined by the average concrete
modulus (Equation 5-5).
The modulus of elasticity of concrete, Ec, is computed using
The force in tension is the product of average tensile stress is and the area in tension and is
A reduction in the computed concrete force is needed because the top row of steel bars is
in compression zone. The compressive force computed in concrete for the area occupied by the
steel bars must be subtracted from the computed value. The compressive strain at the location of
the top row of bars is 0.001447, the area of the bars is 0.0015 m2, the concrete stress is 27,289
kPa, and the force is 40.93 kN.
Thus, the total force carried in the concrete is sum of the computed compressive forces
plus the tensile concrete force minus the correction for the area of concrete occupied by the top
row of reinforce is 1814.10 kN.
5-3-1-5 Computation of Balance of Axial Thrust Forces
The summation of the internal forces yields the following expression for the sum of axial
thrust forces:
170
F = 1814.10 kN
Taking into account the applied axial load in compression of 900 kN, the section is out of
balance by only 0.15 kN (33.7 lbs).
This hand computation confirms the validity of the computations made by LPile. The
selection of a thickness of the increments of concrete of 0.01701 m is thicker than that used in
LPile. LPile uses 100 slices of the full section depth in its computations, so the slice thickness
used by LPile is 0.0076 m for this example problem. Also, some error was introduced by the
reduced precision in the hand computations, whereas LPile uses 64-bit precision in all
computations.
5-3-1-6 Computation of Bending Moment and EI
Bending moment is computed by summing the products of the slice forces about the
centroid of the section. The axial thrust load does not cause a moment because it is applied with
no eccentricity. The moments in the steel bars and concrete can be added together because the
bending strains are compatible in the two materials.
The moments due to forces in the steel bars are computed by multiplying the forces in the
steel bars times the distances from the centroid of the section. The values of moment in the steel
bars are:
Moment due to bar row 1: (479.1 kN) (0.3045) =
152.71 kN-m
38.87 kN-m
41.92 kN-m
188.64 kN-m
344.40 kN-m
The moments due to forces in the concrete are computed by multiplying the forces in the
concrete times the distances from the centroid of the section. The values of moments in the
concrete slices are:
Moment in slice 1: (241.37 kN) (0.3728 m) =
82.15 kN-m
80.37 kN-m
78.40 kN-m
76.24 kN-m
71.68 kN-m
63.33 kN-m
52.16 kN-m
38.81 kN-m
23.90 kN-m
8.07 kN-m
Moment correction for top row of steel bars = ( 40.93 kN) (0.3045 m) = 12.46 kN-m
171
561.32 kN-m.
0.076 m
501.51 kN
0.203 m
382.95 kN
0.760 m
0.203 m
413 kN
0.203 m
619.5 kN
0.076 m
Figure 5-5 Free Body Diagram Used for Computing Nominal Moment Capacity of Reinforced
Concrete Section
The value of bending stiffness is computed using Equation 5-18.
A comparison of results from hand versus computer solutions is summarized in Table 52. The moment computed by LPile was 907.19 kN-m. Thus, the hand calculation is within 0.16%
of the computer solution. The value of the EI is computed by LPile is 51,348.62 kN-m2. The
hand solution is within 0.16% of the computer solution. The hand solution for axial thrust is
within 0.0-2% of the computer solution
The agreement is close between the values computed by hand using only a small number
of slices and the values from the computer solution computed using 100 slices. This example
hand computation serves to confirm of the accuracy of the computer solution for the problem
that was examined.
172
Table 5-2. Comparison of Results from Hand Computation versus Computer Solution
Parameter
By LPile
By Hand
Hand Error, %
907.19
905.71
0.16%
51,348.62
51,265.02
0.16%
Axial Thrust, kN
900.00*
900.15
+0.02%
* Input value
The rectangular section used for above example solution was chosen because the
geometric shapes of the slices are easy to visualize and their areas and centroid positions are easy
to compute. In reality, the algorithms used in LPile for the geometrical computation are much
more powerful because of the circular and non-circular shapes considered in the computations.
For example, when a large number of slices are used in computations, individual bars are divided
by the slice boundaries. So, in the computations made by LPile, the areas and positions of
centroids in each circular segment of the bars are computed. In addition, the areas of bars and
strands in a slice are subtracted from the area of concrete in a slice.
The two following graphs are examples of the output from LPile for curves of moment
versus curvature and ending stiffness versus bending moment. These graphs are examples of the
output from the presentation graphics utility that is part of LPile. Both of these graphs were
exported as enhanced Windows metafiles, which were then pasted into this document.
Moment vs. Curvature - All Sections
1,000
950
900
850
800
750
700
650
600
550
500
450
400
350
300
250
200
150
100
50
0
0.0
0.005
0.01
0.015
Curvature, radians/m
173
0.02
9,000
8,500
8,000
7,500
7,000
6,500
6,000
5,500
5,000
4,500
4,000
3,500
3,000
2,500
2,000
1,500
1,000
500
0
0
200
400
600
800
1,000
1,200
174
0.838 m
0.7817 m
Figure 5-9 Example Pipe Section for Computation of Plastic Moment Capacity
8,000
7,500
7,000
6,500
6,000
5,500
5,000
4,500
4,000
3,500
3,000
2,500
2,000
1,500
1,000
500
0
0.0
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0.009
0.01
0.011
0.012
0.013
0.014
0.015
Curvature, radians/m
practical point of view, it is unrealistic to assume that the bending strains developed in a section
can be large enough to yield the condition that is assumed; however, the computation should
result in a value that is larger than 7,488 kN-m (5,863 ft-kips) but in the appropriate range.
The expression for the plastic moment capacity Mp is the product of the yield stress fy and
plastic modulus Z.
..........................................................(5-24)
Referring to the dimensions shown in Figure 5-9, the plastic modulus Z of the pipe is
As expected, the value of Mp computed from the plastic modulus is slightly larger than
the 7,488 kN-m from the computed solution at a strain of 0.0149 rad/m. However, the close
agreement and the slight over-estimation provide confidence that the computer code computes
the plastic moment capacity accurately.
Another check on the accuracy of the computations is to examine the computed bending
stiffness in the elastic range. From elastic theory, the bending stiffness for the example problem
is
EI
d o4 d i4
64
8
2 10 kPa
0.838 m
0.7817 m
64
1,175,726 kN - m
The value computed by LPile is 1,175,686 kN-m2. The error in bending stiffness for the
computed solution is 0.0035 percent, which is amazingly accurate for a numerical computation.
Please note that the fifth through seventh digits in the above values are shown to be able to
illustrate the comparison and are not indicative of the precision possible in normal computations.
Often, engineers use specified material strengths that are usually exceeded in reality.
The reason that the bending stiffness value computed by LPile is slightly smaller than the
full plastic yield value is that the stresses and strains near the neutral axis remain in the elastic
range. The stress distribution for a curvature of 0.015 rad/m is shown in Figure 5-11.
Approximately, the middle third of this section is in the elastic range.
176
414,000 kPa
0.838 m
0.138 m
0.7817 m
= 0.015 rad/m
manufacturer from region to region and will also vary with the shape, size, and compressive of
the concrete. For most commercially obtained prestressed piles, Fps can be estimated by
assuming some level of initial prestressing in the concrete. Given a value of Fps the program
solves the statically indeterminate problem of balancing the prestressing forces in the concrete
and reinforcement using the nonlinear stress-strain relationships selected for both concrete and
reinforcing steel.
The stress-strain relationships used in prestressed concrete is defined using the stressstrain curves of concrete recommended by the Design Handbook of the Prestressed Concrete
Institute (PCI), as shown in Figure 5-12 and in equation form in Equations 5-25 to 5-28.
270
270 ksi
250
250 ksi
Minimum yield strength = 243 ksi at 1%
Elongation for 270 ksi (ASTM A 416)
230
Minimum yield strength = 225 ksi at 1%
Elongation for 250 ksi (ASTM A 416)
210
190
170
150
0
0.005
0.01
0.015
0.02
0.025
0.03
Strain, in/in
Figure 5-12 Stress-Strain Curves of Prestressing Strands Recommended by PCI Design
Handbook, 5th Edition.
For 250 ksi 7-wire low-relaxation strands:
.......................................(5-25)
ps
0.0076; f ps
250
ps
178
0.04
(ksi) ................................(5-26)
0.0064
is the prestress in the concrete and Ac is the cross-sectional area of the concrete.
The user should check the output report from the program to see if the computed level of
prestressed force in the concrete at the initial stage is in the desired range. The computation
procedures for stresses of concrete for a specific curvature of the cross section are the same as
that for ordinary concrete, described in a previous section, except the current state of stresses of
concrete and strands should take into account the initial stress conditions. The stress levels for
both concrete and strands under loading conditions should be checked to ensure that the stresses
are in the desired range.
Elementary considerations show that a distance from the end of a pile is necessary for the
full transfer of stresses from reinforcing steel to concrete. The development length of the strand
is not computed in LPile. Usually the zone of development is about 50
the axial strand
diameter from the end of the pile.
Typical cross sections of prestressed piles are square solid, square hollow, octagonal
solid, octagonal hollow, round solid, or round hollow, are shown in Figure 5-13.
179
5-4 Discussion
Use of the mechanistic method of analysis of moment-curvature relations by hand is
relatively straightforward for cases of simple cross sections. Use of this method becomes
significantly more laborious when using geometrical values for complex cross sections and
nonlinear stress-strain relationships of concrete and steel or when including the effect of
prestressing in the case of prestressed concrete piles. Thus, use of a computer program is a
necessary feature of the method of analysis presented here.
A new user to the program may wish to practice using LPile by repeating the solutions
for the example problems. When LPile is employed for any problem being addressed by the user,
some procedure should be employed to obtain an approximate solution of the section properties
in order to verify the results and to detect gross input errors.
180
LPile
Index No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
D, in
Area, in2
Wt/ft
D, mm
Area, mm2
Kg/m
0.375
0.500
0.625
0.750
0.875
1.000
1.128
1.270
1.410
1.693
2.257
0.445
0.630
0.768
0.992
1.177
1.406
1.720
2.220
0.236
0.315
0.394
0.472
0.551
0.630
0.787
0.984
1.260
1.575
0.250
0.315
0.375
0.500
0.626
0.752
0.874
1.000
1.126
1.252
1.374
1.504
1.626
0.11
0.20
0.31
0.44
0.60
0.79
1.00
1.27
1.56
2.25
4.00
0.155
0.310
0.466
0.777
1.088
1.554
2.332
3.886
0.043
0.078
0.122
0.175
0.239
0.312
0.487
0.761
1.246
1.947
0.049
0.078
0.111
0.196
0.308
0.444
0.600
0.785
0.996
1.231
1.483
1.767
2.077
0.376
0.668
1.043
1.502
2.044
2.670
3.400
4.303
5.313
7.650
13.600
0.526
1.052
1.578
2.629
3.681
5.259
7.880
13.150
0.147
0.263
0.415
0.594
0.810
1.057
1.651
2.581
4.227
6.604
0.167
0.263
0.375
0.666
1.044
1.506
2.035
2.664
3.377
4.176
5.029
5.994
7.045
9.5
12.7
15.9
19.1
22.2
25.4
28.7
32.3
35.8
43.0
57.3
11.3
16.0
19.5
25.2
29.9
35.7
43.7
56.4
6.0
8.0
10.0
12.0
14.0
16.0
20.0
25.0
32.0
40.0
6.35
8.0
9.53
12.7
15.9
19.1
22.2
25.4
28.6
31.8
34.9
38.2
41.3
71.3
126.7
198.6
286.5
387.1
506.7
646.9
819.4
1006
1452
2579
100
200
300
500
700
1000
1500
2500
28
50
79
113
154
201
314
491
804
1256
31.67
50
71.33
126.7
198.6
286.5
387.1
506.7
642.4
794.2
956.6
1140
1340
0.559
0.993
1.557
2.246
3.035
3.973
5.072
6.424
7.887
11.384
20.219
0.784
1.568
2.352
3.920
5.488
7.840
11.76
19.60
0.220
0.392
0.619
0.886
1.207
1.576
2.462
3.849
6.303
9.847
0.248
0.392
0.559
0.993
1.557
2.246
3.035
3.973
5.036
6.227
7.500
8.938
10.506
181
LPile
Index
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Grade,
ksi
D, in
Area, in
Wt/ft
D, mm
Area,
mm2
Kg/m
250
250
250
250
250
250
250
270
270
270
270
270
270
270
270
300
300
300
300
300
145
145
145
145
145
145
160
160
160
160
160
160
157
150
160
150
160
160
0.340
0.250
0.3125
0.375
0.4375
0.500
0.600
0.34
0.375
0.4375
0.500
0.500
0.5625
0.600
0.700
0.375
0.438
0.500
0.500
0.600
0.750
0.875
1.000
1.125
1.250
1.375
0.75
0.875
1
1.125
1.25
1.375
0.625
1
1
1.25
1.25
1.375
0.058
0.036
0.058
0.080
0.108
0.144
0.216
0.058
0.085
0.115
0.153
0.167
0.192
0.217
0.294
0.085
0.115
0.153
0.167
0.217
0.442
0.601
0.785
0.994
1.227
1.485
0.442
0.601
0.785
0.994
1.227
1.485
0.28
0.85
0.85
1.25
1.25
1.58
0.2
0.122
0.197
0.272
0.367
0.49
0.737
0.2
0.29
0.39
0.52
0.58
0.65
0.74
1.01
0.29
0.39
0.52
0.58
0.74
1.5
2.04
2.67
3.38
4.17
5.05
1.5
2.04
2.67
3.38
4.17
5.05
0.98
3.01
3.01
4.39
4.39
5.56
8.6
6.4
7.9
9.5
11.1
12.7
15.2
8.6
9.5
11.1
12.7
12.7
14.3
15.2
17.8
9.5
11.1
12.7
12.7
15.2
19.1
22.2
25.4
28.6
31.8
34.9
19.1
22.2
25.4
28.6
31.8
34.9
15.9
25.4
25.4
31.8
31.8
34.9
37.4
23.2
37.4
51.6
69.7
92.9
138.7
37.4
54.8
74.2
98.7
107.7
123.9
138.7
189.7
54.8
74.2
98.7
107.7
140.0
285.2
387.7
506.5
641.3
791.6
958.1
285.2
387.7
506.5
641.3
791.6
958.1
180.6
548.4
548.4
806.5
806.5
1019.4
0.298
0.182
0.293
0.405
0.546
0.729
1.096
0.298
0.431
0.580
0.774
0.863
0.967
1.101
1.505
0.431
0.580
0.774
0.863
1.101
2.232
3.035
3.972
5.029
6.204
7.513
2.232
3.035
3.972
5.029
6.204
7.513
1.458
4.478
4.478
6.531
6.531
8.272
182
Section
HP 14
HP 360
HP 13
HP 330
HP 12
HP 310
HP10
HP 250
HP 8
HP 200
Weight
Area, A
lb/ft
kg/m
in
2
cm
in
mm
117
34.4
175
222
102
30
Depth, d
Thickness
Flange
Width, b
Ixx
4
Iyy
4
Compact
Section
Criteria
F'y
ksi
MPa
in
mm
Flange, tf
in.
mm
Web, tw
in.
mm
14.21
14.885
0.805
0.805
1220
443
361
378
20.4
20.4
50800
18400
341
14.01
14.785
0.705
0.705
1050
380
38.4
in
4
cm
in
4
cm
49.4
153
194
356
376
17.9
17.9
43700
15800
265
89
26.1
13.83
14.695
0.615
0.615
904
326
29.6
133
168
351
373
15.6
15.6
37600
13600
204
20.3
73
21.4
13.61
14.585
0.505
0.505
729
261
109
138
346
370
12.8
12.8
30300
10900
140
100
29.4
13.15
13.205
0.765
0.765
886
294
56.7
150
190
334
335
19.4
19.4
36878
12237
391
43.5
87
25.5
12.95
13.105
0.665
0.665
755
250
130
165
329
333
16.9
16.9
31425
10406
300
73
21.6
12.75
13.005
0.565
0.565
630
207
31.9
109
139
324
330
14.4
14.4
26223
8616
220
60
17.5
12.54
12.9
0.46
0.46
503
165
21.5
90
113
319
328
11.7
11.7
20936
6868
148
52.5
84
24.6
12.28
12.295
0.685
0.685
650
213
126
159
312
312
17.4
17.4
27100
8870
362
74
21.8
12.13
12.215
0.61
0.61
569
186
42.1
111
141
308
310
15.5
15.5
23700
7740
290
63
18.4
11.94
12.125
0.515
0.515
472
153
30.5
94
119
303
308
13.1
13.1
19600
6370
210
53
15.5
11.78
12.045
0.435
0.435
393
127
22
79
100
299
306
11
11
16400
5290
152
57
16.8
9.99
10.225
0.565
0.565
294
101
51.6
85
108
254
260
14.4
14.4
12200
4200
356
42
12.4
9.7
10.075
0.42
0.42
210
71.7
29.4
63
80
246
256
10.7
10.7
8740
2980
203
36
10.6
8.02
8.155
0.445
0.445
119
40.3
50.3
54
68.4
204
207
11.3
11.3
4950
1680
347
183
Chapter 6
Use of Vertical Piles in Stabilizing a Slope
6-1 Introduction
The computation of slope stability is a problem often faced by geotechnical engineers.
Numerous methods have been presented for making the necessary analyses; one of the first of
these available as a computer solution was the simplified method of slices developed by Bishop
(1955). Over the years, there have been additional developments for analyzing slope stability.
For example, the method of Morgenstern and Price (1965) was the first method of analysis that
was capable of solving all equations of equilibrium for a limit analysis of slope stability. The
widely used computer programs UTexas4, Slope/W, and Slide implement modern developments
in computation of slope stability. In view of advances in methods of analysis, the availability of
computer programs, and numerous comparisons of results of analysis and observed slope
failures, many engineers will obtain approximately identical factors of safety for a particular
problem of slope stability. This chapter is written with the assumption that the user is familiar
with the theory of slope stability computations and has a computer program available for use.
In spite of the ability to make reasonable computations, there are occasions when
engineering judgment may indicate the need to increase the factor of safety for a particular slope.
There are a large number of methods for accomplishing such a purpose. For example, the factor
of safety may be increased by flattening the slope, if possible, or by providing subsurface
drainage to lower the water table in the slope.
The method proposed in this chapter presents the engineer with additional option that
might prove useful in some cases. Piles have been used in the past to increase the stability of a
slope, but without an analysis to judge their effectiveness. Thus, a method of analysis to
investigate the benefits of using piles for this purpose is a useful tool for engineers.
stratum that was not found earlier, or changes in environmental conditions could have caused a
weakening of the soils in the slope. The use of drilled shaft foundations to strengthen the slope
might then be considered.
185
in.) in diameter were installed in the slide over a period of three years. Computations indicated
that the presence of the piles increased the factor of safety against sliding by about 0.18, which
was sufficient to prevent further movement. Strain gages were installed on five of the piles and
these piles were recovered after some time. At least two of the piles were fractured due to
excessive bending moment.
Hassiotis and Chameau (1984) and Oakland and Chameau (1986) present brief
descriptions of a large number of cases where piles have been used to stabilize slopes. The
authors present a detailed discussion of the use of piles and drilled piers in the stabilization of
slopes.
M
hp
P
(a)
(b)
(c)
Figure 6-2 Forces from Soil Acting Against a Pile in a Sliding Slope, (a) Pile, Slope, and Slip
Surface Geometry, (b) Distribution of Mobilized Forces, (c) Free-body Diagram of Pile Below
the Slip Surface
The principles of limit equilibrium are usually employed in slope stability analysis. The
influence of stabilizing piles on the factor of safety against sliding is illustrated in Figure 6-3.
The resultant of the resistance of the pile, T can be included in the analysis of slope stability.
Therefore, a consistent assumption is that the sliding soil has moved a sufficient amount that the
186
peak resistance from the soil has developed against the pile. If one considers the force acting on a
pile from a wedge of soil with a sloping surface, the force parallel to the soil surface is larger
than if the surface were horizontal. However, a reasonable assumption is that the peak resistance
acting perpendicular to the pile can be found from the p-y curve formations presented in Chapter
3.
Safety factor for moment equilibrium considering the same forces as above,
plus the effect of the stabilizing pile is expressed as:
......................................(6-1)
Where T is the average total force per unit length horizontally resisting soil
movement and z is the distance from the centroid of resisting pressure to
center of rotation.
Figure 6-3 Influence of Stabilizing Pile on Factor of Safety Against Sliding
The discussion above leads to the following step-by-step procedure:
1. Find the factor of safety against sliding for the slope using an appropriate computer
program.
2. At the proposed position for the stabilizing pile, tabulate the relevant soil properties with
depth.
3. Select a pile with a selected diameter and structural properties and compute the bending
stiffness and nominal moment capacity. Compute the ultimate moment capacity (i.e.
factored moment capacity) by multiplying by an appropriate strength reduction factor
(typically around 0.65)
4. Assume that the sliding surface is the same as found in Step 1, then use LPile to compute
the p-y curves at selected depths above the sliding surface. Employ the peak soil reaction
187
versus depth as a distributed lateral force for depths above the sliding surface as shown in
Figure 6-2(b) and analyze the pile again using LPile.
5. Compare the maximum bending moment found in Step 4 with the nominal moment
capacity from Step 3. At this point, an adjustment of the size or geometry of the pile may
or may not be made, depending on the results of the comparison. Note that in general, the
presence of the piles may change the position of the sliding surface, which will also
change the maximum bending moment developed in the pile. However, in some cases,
the position of the sliding surface will be known because of the location of a weak soil
layer, and, in any case, it is unlikely that the position of the sliding surface will be
changed significantly by the presence of the piles.
6. Employ the resisting shear and moment in the slope stability analysis used in Step 1 and
find the new position of the sliding surface. While only one pile is shown in Figure 6-3,
one or more rows of piles are most likely to be used. In such a case, the forces due to a
single pile should be divided by the center-to-center spacing along the row of piles prior
to input to the slope stability analysis program because the two-dimensional slope
stability analysis is written assuming that the thickness of the third dimension is unity.
Some programs for slope stability analysis can use the profile of distributed loads in the
computation of the new sliding surface.
7. Change the depth of sliding, hp, to the depth of sliding employed in Step 4, obtain new
values of M and P, and repeat the analyses until agreement is found between that surface
and the resisting forces for the piles. Also, the geometry of the piles should be adjusted so
that the maximum bending moment found in the analyses is close to the ultimate moment
capacity of the piles.
8. Finally, compare the factor of safety against sliding of the slope with no piles to that with
piles in place and determine whether or not the improvement in factor of safety justifies
the use of the piles.
188
maximum value. The positive conclusion from this field test is that the bending-moment curve
given by Fukuoka had the general shape that would be expected.
At another site at the Higashi-tono landslide, Fukuoka described an experiment where a
number of steel-pipe piles were used in a sliding soil. Some of them were removed after a
considerable period of time and found to have failed in bending. One of them had a diameter of
318.5 mm and a wall thickness of 10.3 mm. The collapse moment for the pipe was computed to
be 241 kN-m. Assuming a triangular distribution of earth pressure on the pile from the sliding
mass of soil, which had a thickness of 5 m, the undrained shear strength that was required to
cause the pile to fail was 10.7 kPa. The author merely stated that the soil had a NSPT that was less
than 10 bpf. That value of NSPT probably reflects an undrained shear strength that encompasses
the computed strength to cause the pile to fail.
6-6-2 Example Computation
The example that was selected for analysis is shown in Figure 6-5. The slope exists along
the bank of a river where sudden drawdown is possible. Slides had been observed along the river
at numerous places and it was desirable to stabilize the slope to allow a bridge to be constructed.
Elevation, m
80
75
Fill
c = 47.9 kPa
= 19.6 kN/m3
70
Silt
c = 23.9 kPa
cresidual =12.4 kPa
= 17.3 kN/3m3
65
60
Clay
c = 36.3 kPa
= 17.3 kN/m3
Sand
= 19.6 kN/m3
= 30 to 40 deg.
55
Figure 6-5 Soil Conditions for Analysis of Slope for Low Water
The undrained analysis for the sudden-drawdown case was made based on the Spencer's
method, and the factor of safety was found to be 1.06, a value that is in reasonable agreement
with observations. Plainly, some method of design and construction would be necessary in order
for bridge piers to be placed at the site. The method described herein was employed to select
sizes and spacing of drilled shafts that could be used to achieve stability.
190
A preliminary design is shown in Figure 6-6, but not shown in the figure is the distance
along the river for which the slope was to be stabilized. Drilled shafts were selected that were
915 mm (3 ft) in diameter and penetrated well below the sliding surface, as shown in the figure.
Further, as shown in the figure, it was found that the tops of the shafts had to be restrained with
grade beam anchored in stable soil. The use of the grade beam was required because of the depth
of the slide. The results of the analysis, for each of the groups perpendicular to the river, gave the
following loads at the top of the drilled shafts: Shafts 1, 2, and 3, +1,090 kN; Shaft 4, 1,310 kN;
and Shaft 5, 1,690 kN. The member connecting the tops of the 5 piles would be designed to
sustain the indicated loading. The maximum bending moment for Shaft 5 was about 6,250 kN-m,
which would require heavy reinforcement. The computed bending moments for the other drilled
shafts was much smaller.
With the piles in place and with the restraining forces of the piles against the sliding soil,
shown Figure 6-7, a second analysis was performed to find the new factor of safety against
sliding. The value that was obtained was 1.82. This result was sufficient to show that the
technique was feasible. However, in a practical design, a series of analyses would have been
performed to find the most economical geometry and spacing for the piles in the group.
Pile Row 1
5.5 m
Pile diameter 915 mm
Grade Beam
30 m
4.6 m 4.6 m
15.2 m
15.2 m
191
Elevation, m
80
48 kPa
48 kPa
75
70
108 kPa
108 kPa
65
71 kPa
71 kPa
60
55
Figure 6-7 Load Distribution from Stabilizing Piles for Slope Stability Analysis
6-6-3 Conclusions
The results predicted by the proposed design method are compared with results from
available full-scale experiments. The case studies yield information on the applicability of the
proposed method of analysis.
A complete analysis for the stability of slopes with drilled shafts in place is presented.
The method of analysis is considered to be practical and can be implemented by engineers by
using readily available methods of analysis. The benefits of using the method is that rationality
and convenience are indicated that have not been previously available.
192
193
References
Akinmusuru, J. O.,
n Piled Footings, Journal of the
Geotechnical Engineering Division, ASCE, Vol. 106, No. GT11, November, pp. 1263-1268.
Allen, J., 1985. p-y
Austin, May.
American Petroleum Institute, 2010. Recommended Practice for Planning, Designing and
Constructing Fixed Offshore Platforms - Working Stress Design, API RP 2A-WSD, 21st
Edition, Errata and Supplement, 2010.
Awoshika, K., and Reese, L. C., 1971.
-Spaced Batter
Research Report 11 7-3F, Center for Highway Research, The University of Texas at
Austin, February.
Awoshika, K., and Reese, L. C., 1971.
Proceedings, Intl. Symposium on the Engineering Properties of Sea-Floor Soils and
Their Geophysical Identification, University of Washington, Seattle.
Baecher, G. B., and Christian, J. T., 2003. Reliability and Statistics in Geotechnical Engineering,
Wiley, New York, 605 p.
Baguelin, F.; Jezequel, J. F.; and Shields, D. H., 1978. The Pressuremeter and Foundation
Engineering, Trans Tech Publications.
Bhushan, K.; Haley, S. C.; and Fong, P. T., 1979.
Journal of the Geotechnical Engineering Division, ASCE, Vol. 105, No. GT8, pp.
969-985.
Bhushan, K.; Lee, L. J.; and Grime, D. B., 1981.
Preprint, ASCE Annual Meeting, St. Louis, Missouri.
Bishop, A.W.
The Use of the Slip Circle in the Stability Analysis of Slopes,
Gotechnique, Vol. 5, No. 1, pp. 7-17
Bogard, D., and Matlock, H., 1983.
Proceedings, Geotechnical Practice in Offshore Engineering, ASCE.
Bowman, E. R., 1959.
M.S. thesis, The University of Texas at Austin, January, 84 p.
Briaud, J.-L.; Smith, T. D.; and Meyer, B. J., 1982.
Loaded Piles Using
Symposium, The Pressuremeter and Marine Applications, Paris.
Broms, B. B., 1964a.
Journal of the Soil
Mechanics and Foundations Engineering Division, ASCE, Vol. 90, No. SM3, pp. 123-156.
Broms, B. B., 1964b.
Journal of the Soil
Mechanics and Foundations Engineering Division, ASCE, Vol. 90, No. SM2, pp. 27-63.
194
References
195
References
Proceedings,
Transactions, ASCE,
Japanese Society for Architectural Engineering, Committee for the Study of the Behavior of
Piles During Earthquakes, 1965. Research Concerning Horizontal Resistance and Dynamic
Response of Pile Foundations, (in Japanese).
Johnson, G. W., 1982.
-Boring Pressuremeter in Obtaining In-Situ Shear Moduli
M.S. thesis, The University of Texas, Austin, Texas, August, 156 p.
Johnson, R. M., Parsons, R. L., Dapp, S. D., and Brown, D. A., 2006. Soil Characterization and
p-y Curve Development for Loess, Kansas Department of Transportation, Bureau of
Materials and Research, July.
Kooijman, A. P., 1989.
196
References
Journal of
Lieng, J. T., 1988. Behavior of Laterally Loaded Piles in Sand - Large Scale Model Tests,
Ph.D. Thesis, Department of Civil Engineering, Norwegian Institute of Technology, 206 p.
Long, J. H., 1984.
etitive
Ph.D. dissertation, The University of Texas, Austin, Texas, 332 p.
Malek, A. M.; Azzouz, A. S.; Baligh, M. M.; and Germaine, J. T., 1989.
Journal of Geotechnical
Engineering Division, ASCE, Vol. 115, No. 5, pp. 615-637.
Marinos, P., and Hoek, E.,
A Geologically Friendly Tool for Rock Mass Strength
Proceedings, GeoEngineering 2000 Conference, Melbourne, pp.1,422-1,442.
Matlock, H., 1970.
Proceedings, 2nd Offshore Technology Conference, Vol. I, pp. 577-594.
Matlock, H., and Ripperger, E. A., 1958.
Proceedings, ASTM, Vol. 58, pp. 1245-1259.
Matlock, H.; Ripperger, E. A.; and Fitzgibbon, D. P. 1956.
(unpublished), 167 p.
-Loading of
197
References
-Dimensional Pile
Proceedings, 9th
havior
Texas Civil Engineer, May, pp. 17-23.
198
References
and
Florida, (unpublished).
Seed, R. B. and Harder, L. F., 1990.
and U
Publishers Ltd., pp. 351-376
Journal
Gotechnique, Vol. 5,
p-y
hesis, The University of Texas at Austin.
Timoshenko, S. P., 1956. Strength of Materials, Part II, Advanced Theory and Problems, 3rd
Edition, Van Nostrand, New York.
Bending of Beams Resting on Isotropic Elastic Sol
Engineering Mechanics Division, ASCE, Vol. 87, No. EN2, pp. 35-53.
199
Journal of the
References
Ph.D.
-Scale Lateral
200
References
201
Name Index
Akinmusuru, J. O. ....................................... 4
Dunnigan, L. P. ......................................... 60
Fenske, C. W............................................. 80
Fitzgibbon, D. P. ....................................... 50
Audibert, J. M. E....................................... 80
Awoshika, K. .............................................. 4
Fong, P. T.................................................. 80
Azzouz, A. S. ............................................ 11
Baecher, G. B.............................................. 3
Gazioglu, S. M. ......................................... 80
George, P................................................... 18
Baligh, M. M............................................. 11
Gerber, T. M. ............................................ 97
Germaine, J. T........................................... 11
Bogard, D.................................................... 4
Grime, D. B............................................... 96
Bowman, E. R. .......................................... 83
Hales, L. J. ................................................ 98
Briaud, J. L,............................................... 80
Haley, S. C. ............................................... 80
Broms, B. B............................................... 16
Hansen, J. B. ............................................. 55
Harder, L. F............................................... 97
Bryant, L. M................................................ 7
Christian, J. T.............................................. 3
Cox, W. R. .... 48, 49, 50, 58, 59, 70, 85, 109
Dapp, S. D................................................. 99
Hrennikoff, A.............................................. 4
Darr, K. ..................................................... 55
Davis, E. H................................................ 18
Jamiolkowski, M....................................... 18
Decker, R. S. ............................................. 60
Johnson, G. W........................................... 50
DiGiola, A. M. .......................................... 16
Johnson, R. M. .......................................... 99
Koch, K. J. .................................................. 4
202
References
Kooijman, A. P. ........................................ 51
Reese, L. C.4, 18, 48, 49, 50, 51, 55, 58, 59,
70, 75, 78, 80, 81, 85, 90, 97, 109, 113,
114, 120, 123, 124, 139, 157
Rojas-Gonzalez, L..................................... 16
Lane, J. D. ................................................. 97
Lee, L. J..................................................... 96
Seed, R. B. ................................................ 97
Sherard, J. L. ............................................. 60
Malek, A. M.............................................. 11
Skempton, A. W........................................ 61
Smith, T. D................................................ 80
Meyer, B. J................................................ 80
Stevens, J. B.............................................. 80
Morrison, C. M. ........................................ 51
Stokoe, K. H.............................................. 50
Murchison, J. M. ....................................... 96
Sullivan, W. R........................................... 80
Newman, F. B. .......................................... 16
Timoshenko, S. P. ..................................... 37
............................................... 50
Parsons, R. L. ............................................ 99
Wood, D.................................................... 18
Poulos, H. G.............................................. 18
203