Missouri University of Science and Technology

Scholars' Mine
International Conference on Case Histories in (1998) - Fourth International Conference on Case
Geotechnical Engineering Histories in Geotechnical Engineering

Mar 8th - Mar 15th

Non-linear Lateral Pile Deflection Predictions in

Clays
Shamsher Prakash
Missouri University of Science and Technology, prakash@mst.edu

M. H. Chen
University of Missouri-Rolla, Rolla, Missouri

Follow this and additional works at: https://scholarsmine.mst.edu/icchge

Part of the Geotechnical Engineering Commons

Recommended Citation
Prakash, Shamsher and Chen, M. H., "Non-linear Lateral Pile Deflection Predictions in Clays" (1998). International Conference on Case
Histories in Geotechnical Engineering. 58.
https://scholarsmine.mst.edu/icchge/4icchge/4icchge-session01/58

This Article - Conference proceedings is brought to you for free and open access by Scholars' Mine. It has been accepted for inclusion in International
Conference on Case Histories in Geotechnical Engineering by an authorized administrator of Scholars' Mine. This work is protected by U. S. Copyright
Law. Unauthorized use including reproduction for redistribution requires the permission of the copyright holder. For more information, please contact
scholarsmine@mst.edu.
 384
Proceedings: Fourth International Conference on Case Histories in Geotechnical Engineering, St. Louis, Missouri,
March 9-12, 1998.

NON-LINEAR LATERAL PILE DE}"LECTION PREDICTIONS IN CLAYS

Shamsher Prakash M.H. Chen Paper No 1.55

Professor in Civil Engineering Former Graduate Student
University of Missouri-Rolla University of Missouri-Rolla
Rolla, MO 65409-0030 USA Rolla, MO 65409-0030 USA

ABSTRACT

A method to predict lateral load-deflection relationships for single free-headed piles in clay, considering the non-linear modulus of the
sub grade reaction is proposed on the basis of the analysis of full-scale test data on ten piles. The modulus of the horizontal subgrade
reaction, kh, was obtained from the field load-deflection data of six piles in soft clay and four piles in stiff clay. The kh values depend on
lateral (shear) strain, which is related to pile diameter and deflection. Empirical equations of normalized modulus degradation with shear
strain have been proposed for both soft and stiff clays. The maximum value of the modulus of horizontal subgrade reaction, khmax• has
been related with the undrained shear strength of clays. J\ comparison of the results from the present method and the p-y method with
the measured load-deflection curves has also been made.

The proposed method predicts upper and lower bound load-deflection curves which are valuable guides to making informed decisions.

KEYWORDS

Lateral loads, pile, deflections, non-linear modulus, pile load tests.

INTRODUCTION curves and Qg-Yg curves, where Qg and yg are load and deflection
respectively at the ground level.
There are four categories of methods (Brinch Hansen, 1961;
Brorns, 1964; Reese and Matlock, 1956; and Poulos, 1971) for Mwindo ( 1992) analyzed piles of different materials embedded
the analysis of single piles under lateral loads. The first two in clays of different consistencies and developed fivt!" empirical
methods calculate the allowable lateral load by dividing the relations to obtain the strain-dependent modulus of the
ultimate (failure) load by an appropriate factor of safety. The horizontal subgrade reaction. These relationships have limited
other two methods, by Reese and Matlock (subgrade reaction application in practice; however~ this approach has been
method) aod Poulos (elastic method) obtain the allowable lateral extended in this study to predict the load-deflection relationship
loads corresponding to an acceptable lateral deflection. In for single free-head piles under lateral loads in clays.
design practice, the criterion for design in the majority of cases
is the maximum deflections of the piles. In this study, a step-by-step procedure has been developed for
non-linear prediction of the load-displacement relationship of
Lateral load deflection of piles in clay, using the subgrade single piles under lateral loads in clays of uniform properties
reaction method, is solved by considering a pile as a beam on an with depth. A unique relationship for normalized modulus
elastic foundation and replacing soil by a series of elastic, degradation with strain and a range of maximum values of the
closely-spaced, but independent, springs. This method has the horizontal subgrade reaction, k~>..,= at a shear strain of0.002 for
advantage of being relatively simple and factors like soil non- clays are recommended. The load-deflection by the proposed
linearity, variation ofsubgrade reaction with depth, and layered method has been compared with the solution obtained by the p-y
systems can be incorporated. The disadvantage of this method curve method also.
is that the modulus of the subgrade reaction is not a unique soil
property. Soil non-linearity has been considered (Reese 1984) The predictions of typical load-deflection relationships by the
by using the p-y curve approach, which is based on limited pile proposed method are similar to those obtained by the p-y curve
load tests and makes liberal use of empirical coefficients approach. However, the proposed method is more rational since
developed from these test data (Kumar 1993, Chen 1994). Also, upper and lower bound curves are obtained which are valuable
different values of the modulus are used for developing p-y guides to making informed engineering decisions.
Fourth International Conference on Case Histories in Geotechnical Engineering
Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu
 385

MODULUS OF HORIZONTAL SUBGRADE REACTION -~ Initial Tangcnt

When a pile is subjected to horizontal loading, a typical plot of
soil reaction p with deflection y is shown in Figure I. The
secant modulus shown by the dashed lines is the modulus of the
\
~Secant Modulus
horizontal subgrade reaction (kh)- Tangent modulus sho\VTl in
Figure 1 is seldom used in practice. Since the soil reaction-
deflection plot is non-linear, kh is also a non-linear function of
deflection or strain in the soil around the pile, and degrades with
the increase in deflection of the pile (Prakash and Kumar 1996). Dellcction y (L)

The Winkler ( 1867) soil model assumes that the elastic wil
medium is replaced by a series of infinitely closely-spaced Fig. 1. Soil reaction vs deflection for soil around a pile
independent and elastic springs. The stiffness of the spring kh
can be expressed as
MODULUS OF SUBGRADE REACTION DEGRADATION
p WITH STRAIN
(I a)
y
Mwindo (1992) developed five empirical relationships of
modulus degradation with shear strain (y) for free-headed piles
where p is the soil reaction per unit length of pile (Ft ). of different materials embedded in clays of different relative
Therefore, units of kh are FL- 2 • If k sis the coefficient of the consistencies. The general form of the equations developed by
horizontal subgrade reaction (force per unit volume, FL-J) as Mwindo is:
used by Terzaghi (1955), then for a pile of wtdth B,

(lb) ay -b (3)

-For over-consolidated clays, for long term loading, kh can be

assumed to be constant with depth (Terzaghi 1955). Values of the constants a and b obtained by Mwindo are listed
in Table I.
In normally-consolidated clays, for long term loading, kh is
asswned to increase linearly with depth as for sands, according Table L VALUES OF COEFFICIENT a AND b
to the following relationship (Davisson 1960, Reese and FOR EQUATION (3) (AFTER MWINDO,
Matlock, 1956 Prakash and Kumar 1996 ): 1992)

(2)
Pile Soil a b

Timber Medium Clay 0.16 0.30

Where
x = considered depth Timber Stiff Clay 0.33 0.18
nh =constant of horizontal subgrade reaction
The p-y method of design incorporates non-linearity of soils and Steel Pipe Stiff Clay 0.05 0.48
uses (1) empirical coefficients from field tested piles and (2)
Prest. Cone. Soft Clay 0.08 0.42
different sets of equations for different soil conditions. Kumar
(1993) and Prakash and Kumar (1996) have developed a method Prest. Cone. Stiff Clay 0.19 0.27
to predict the load-deflection relationship for laterally-loaded
piles in sand based on the analysis of full-scale lateral pile load
For pile-head loading conditions, the shear strain y depends on
tests. There is, however, a need to develop a method to predict
the pile displacement, y at the pile head. Kagawa and Kraft
the load-deflection relationship for single piles under lateral
( 1980) have reported that more than 70 percent of the pile
loads in clays.
displacement is concentrated in the soil mass within two-pile
radius depth. Therefore, the increase in shear strain due to soil-
pile interaction is concentrated in this zone.

Fourth International Conference on Case Histories in Geotechnical Engineering

Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu
 386
The average normal strain {E) in the direction of pile movement Where
in the soil around the pile of radius r (or width B=2r) can be = deflection at ground level
approximated by ~ lateral load applied at the pile head at
ground level
y 5 r E 2.5BE (4) M, = moment applied at the pile head at ground
level
R =relative stiffness factor
The same relationship was used by Matlock (1970) following the El -= Flexural stiffness of pile
concept originally proposed by Skempton ( 1951) for footing = deflection coefficients for Q 8 and M 11
settlement problems. As the soil strain orthogonal toE is -vc, (Davisson 1960, Davisson and Gill 1963) and
the maximum shear strain y due toE and -vE is (l+v)E. Thus,
the average shear strain around a pile is given as
(8)
l+v
y = --y (5)
2.5B

It may be noted that values of R, the relative stiffness factor, are

The Poisson's v ratio of clay being approximately equal to 0.50, computed from the load-deflection values (Eq. 7) and not from
Eq. (5) becomes the sot! modulus (k,)(Eq. 8).

y 3. Check if the pile is long, i.e., LIR > 4,

y (6)
1.667B
4. For every R, determined in (2) above, the constant of the
modulus of horizontal subgrade reaction (kh) was calculated
Blaney and O'Neill (1986) used Eq. 6 to relate deflections and from the following equation: (see Eqn 8 also)
shear strains around piles in clays.
k = EI (9)
h
Although the original stipulation of strain by Kagawa and Kraft R'
(1980) was for small deflections, tt has been demonstrated by
Prakash and Kumar ( 1996) that it is appropriate to use the same
description of strain even with relatively large pile deflections. 5. Compute the shear strain by using Eq. (6). The kh versus
surface deflection (y) and kh versus surface shear strain (y) were
plotted.
Mwindo's 0992) Analysis
6. The kh versus shear strain plot was then normalized with
Mwindo (1992) developed five empirical relationships or regard to fhe maximum value of k, (k,m~l at 0.002 sttain
modulus degradation with shear strain for piles of different (Mwindo 1992, Kumar 1993, Chen 1994, Prakash and Kumar
materials embedded in clays of different consistency assuming 1996).
k, as constant wifh depfh for all clays. Steps used to develop
these relationships are listed below. 7. Normalized kh versus strain plots of several p11es of the same
material embedded in the same type of soil were then averaged
1. Measured load deflection curves from the lateral pile load test and an equation (Eq. 3) was fitted.
data on free-headed piles available in the literature were
replotted and the loads and corresponding deflections were 8. The above procedure was repeated for other pile load tests to
tabulated. get 5 sets of equations of modulus degradation with strain.

2. For every load-deflection set, the relative stiffness factor R The values of the constants a and b (Mwindo, 1992) are sho\Vll
was calculated using the following equation: in Table l. Figure 2 shows k1/khmax versus shear strain plots
using five equations. These five equations are based on analyses
M R2 of different types of pile materials in clays of different
g
A (7) consistencies, and suffer from the following shortcomings;
Y' EI EI
1. The assumption of kh constant with depth for both soft and
stiff clays is not appropriate.

Fourth International Conference on Case Histories in Geotechnical Engineering

Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu
 387
Table 2. ORIGINAL INFORMATION OF PILES

*Pile Pile Soil Length (m) Width(m) EI (kN-m 2) e(m)' Undrained 1 Author
No. shear (kPa)

39 Steel pipe Stiff clay 13.11 0.152 2118.24 0 57.0 Reese et al., 1975

40 Steel pipe Stiff clay 18.07 0.610 135983.52 0 140.0 Reese et al., 197 5

63 Steel pipe Stiff clay 10.00 0.273 13212.05 0.305 65.0 Brown et al., 1987

64 Steel pipe Stiff clay 11.40 1.220 1025047.0 0 95.0 Dunnavant, 1986

31 Steel Soft Clay 12.80 0.324 12176.82 () 14.4 Long and Reese, 1987

34 Timber Soft peat 21.34 0.318 6889.46 0 14.4 Robinson, 1979

36 Timber Soft clay 15.24 0.305 5851.53 0 28.7 Robinson, 1979

37 Timber Soft clay 9.45 0.305 5851.53 0 14.4 Robinson, 1979

53 Timber Soft clay 13.41 0.311 6007.91 0 28.7 Alizadeh, 1969

54 Timber Soft clay 13.41 0.305 5842.20 0 28.7 Alizadeh, 1969

1 2
Averaged up to 5-diameters, vertical distance above the ground where load is applied
*same pile numbers as used by Mwindo ( 1992)

2. No generalized value or range of khmax (or nhma:J has been below the water table. Reese and Welch (1975) developed a
reconunended which can be used for practical purposes. procedure for p-y curves in stiff clays above the water table.
The results of these experiments were used to develop
3. The modulus degradation equations suggested are of very recommendations for developing p-y curves, and are presented
limited use in practice because these are related lo pile materials. elsewhere (Reese 1984, Chen 1994), and will not be repeated
here.

o-yMethod The current approach of incorporating the factor of soil non-

linearity (p-y curve metbod) involves liberal use of empirical
Matlock (1970) proposed a procedure for determination of the coefficients developed from only a few pile load tests.
p-y curve in soft clays. Reese et al ( 1975) extended this Therefore a simple method to predict a realistic lateral load-
procedure for the determination of p-y curves for stiff clays deflection curve is needed.

PRESENT ANALYSIS

This study may be divided into the following steps:

1. Developing the modulus degradation equation with strain for

clays from analysis of field data from published literature.

2. Determining the range ofkhmux

3. Predicting the load-deflection curves.

4. Comparing the predictions with measured curves and the p-y

method.

The method described below was adopted to analyze the actual

Fig. 2 kh/khrnn versus shear strain with different pile pile test data from the literature in order to obtain the
materials in clays (l\1windo, 1992) relationship between kh and shear strain ( y ). This relationship

Fourth International Conference on Case Histories in Geotechnical Engineering

Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu
 388

can be expressed by a modulus degradation equation. To obtain factor (Eq. 8).

the modulus degradation equation, two assumptions were made
(Davisson 1970). Four selected lateral load-deflection curves from the lateral pile
load test data available in the literature were replotted and the
1. Kh is constant with depth in stiff clay. loads and corresponding deflections were tabulated (Chen 1994).
TI1e original load-deflection plot for Pile 40 is shown in
2. Kb increases linearly with depth in soft clay (including Figure 4.
medium stiff clay).
~,lk.hmax versus strain values were obtained as explained earlier.
Tn Figure 5, kb versus Yg has been plotted for Pile 40. In Figure
Development of Modulus Degradation with Strain 6, kh versus shear strain y and in Figure 7 normalized, kh/k~
versus y for pile 40 have been plotted. In a similar manner,
Ten lateral load-deflection curves in clays were obtained from normalized modulus degradation with shear strain for the other
the published literature. Table 2 presents the original three piles (Pile 39, 63, 64) in stiff clays were computed and are
information on pile material, their dimensions, flexural plotted in Figure 8, along with that of pile 40.
stiffnesses, soil conditions, average undrained shear strength
(computed up to 5 pile diameters) and references. Since the variation in the degradation of the curve is not
significant (maximum variation being 0.4 to 0.25 at a strain of
0.02), all these curves can be represented by one curve of
Determination of Soil Conditions at Test Site. The soil modulus degradation. The best-fit exponential curve relating
conditions at the pile test sites can be defmed by undrained shear khlkhma~ to the shear strain y was fitted by the method of least
strength and soil density. Reese, eta] (1975) suggested that the squares. Figure 8 shows this curve using Equation (10) with
average undrained shear strength should be computed from the respect to all the data points.
shear strength of the soil to a depth of 5 pile diameters. This
depth is considered as the significant zone controlling the 4
= 0.052y' 0 · ' (10)
deflections for lateral loads.

Consistency of Clays. Table 3 is a guide for consistency of clays

in this analysis.
This equation gives the degradation of the normalized modulus
Table 3. UNDRAINED SHEAR STRENGTH AND of horizontal subgrade reaction in stiff clays with shear strain
CONSISTENCY OF CLAY (AFTER lfrespective of the pile material and undrained shear strength of
REESE, 1984) clay.

2. Soft Clay. A pile is considered ]ong if LIT is greater than or

Consistency of Clay S., (kPa)
equals 5. T (Eqn. 12) is the relative stiffness factor in soils for
Soft 12-50 which the modulus of subgrade reaction increases linearly with
depth.
Medium 25-50
Piles 31, 34, 36, 37, 53 and 54 were tested in soft clays (Table
Stiff 50-400 2). The procedure to obtain the kb/k.hmax versus strain plot are
described as following (Kumar 1993, Prakash and Kumar 1996),
Figure 3 shows the soil profile of the site for Pile 39 (0.5 ft in
diameter) and Pile 40 (2ft in diameter) which were tested at the 1. Compute the relative stiffness factor T from the following
same site (Reese et al., 1975). The average undrained shear equation using field data for Os and Yg·
strength is computed to a depth of 5 pile diameters (2.5 ft and I 0
ft, respectively) which is 57 kPa (0.6 Tsf) and 140 kPa (1.5 Tsf), Q T' M T2
respectively. As per Table 3, it is classified as stiff clay. A - • - + B --'-"- (11)
Y EI y EI

Modulus Degradation Equation.

Where
1. Stiff Clay. A pile is considered long and will behave as a Ay, By = deflection coefficients for Q gand M g for soils in
flexible member if the ratio LIR is greater than or equals 4. If which the modulus of subgrade reaction
L/R is smaller than 2, the pile behaves as a rigid member, where increases linearly with depth.
L is the embedded pile length and R is the relative stiffness
Fourth International Conference on Case Histories in Geotechnical Engineering
Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu
 389
UNDRAINED
PLASTIC LIMIT, WATER CONTENT SHEAR STRENGTH, ,.,n~/tt 1
AND LIQUlD LlMIT, .,.,. T =relative stiffness factor
0 20 -40 60 80 100
=
= 5~
i·-------+
~
+----------··
-+:--------+
... uoculrot~
T EI (12)
1 ...... ,
n,
...
0.

+---------+ ~
u
ro ;;_
~
0

.
~

• 0 15 i 2. Compute depth coefficient Z ~ L(f to check if the pile is a

.
··----1" -
.. 20
0
0
"
~
long pile.
3. Compute nh, the constant of the horizontal subgrade reaction,
s
~
Equation (13).
e • lhotr <"110!
2.5 ':
............
.,.._.,. ,lUll<-·

~o
~ ~
•
~ (13)
L___J __ __L_ __ L_ _~----~~30°
'
Fig. 3. Composite soil Profile of pile 30,40 (Reese et al., 1975)
In Eq. (13) 'x' of one meter. is adopted as the standard depth for
500 r---- analysis (Davisson 1995, Kumar 1993, and Prakash and Kumar
! 1996) also recommended and used a standard depth as one meter
I below the ground surface.
500:------- 4. Compute the shear strain by using Eq. (6).
/
5. From a plot of kh vs y, determine khmax @ y = 0.002
""r·- - 6. K/khmax values are computed for specified strain values and
! I the normalized kh/khma~ versus strain for six piles in soft clays
o" f
and peat are plotted in Figure 9. The best-fit equation is
determined by the method of least squares and is given by:

0.054y -0.477 (14)

khma:x
""r----
The values for this Equation are also plotted in Figure 9.

ESTABLISHING THE RANGE OF k,m,. VALUES

Fig. 4. Original load-deflection plot for pile 40 (after Reese
et al., 1975) · - For developing the load-deflection curve, the values of khmax
. '
which are related to the undrained shear strength of the clays are

·-
i
' ' needed.
'
.
I
'
' 1. Range of kh.n:o.x in Stiff Clay. As noted earlier, khrnax is
"'"" '
'
' defmed as the value ofkh at a shear strain of0.002. The average
I i
I
,..,. undrained shear strength Su is estimated to a depth of 5 pile
. diameters. Since undrained shear strength is one of the most

-
'

.
~
\
''
\
I
I
'

i II I common properties of clay, a correlation was developed between

khmax and the undrained shear strength (S 11) of clays. For Piles 39,
J
..,., •'' '
I 40, 63 and 64 in stiff clays, all these values are listed in Table 4 .
In this I able values of k11,....a./Su are also listed. These values vary

~"''
I
i from 357 to 587. The reconunended value ofkhmu is 300 8 11 to
I 600 Sll' A value of 300 is a conservative reconunendation.
"""'
-
! i

' 2. Range of khrnax in Soft Clay. In soft clays kn increases

'
linearly with depth. A standard depth of one meter was adopted
OeftBdion. y, (mm}
" " in Eq. {13) for soft clay as explained before. The values ofkhmax•
Fig. 5. Kh versus deflection for Pile 40 in stiff clay Su and ktrm,./Su were determined for soft clays in a similar

Fourth International Conference on Case Histories in Geotechnical Engineering

Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu
 390

manner as for stiff clays, Table 5. The undrained shear strength ; I

I

of pile 34 was not reported in the literature. Therefore, Table 5 """' '
I

I
'

shows only five sets of data. I I

'""" i
In Table 5, k.,,.)S, varies from 793 to 1761. The recommended :
I
value ofkhma~ is 800 Su to 1800 Su.
,....,

~ -\
J\:hlno ..

Table 4.

Pile No.
Khmax VALUES RELATED

STIFF CLAY (Chen 1994)

Komu (kPa)
WITH
UNDRAINED SHEAR STRENGTH IN

S, (kPa) khmm/Su
~

-.... '
I
I

~·
I

~
I
'

....I
- .....
i ,'----
39 21591.2 57 378.8 """'"" 0.0111 0.0'15
""
40 82219.1 140 587.3
Fig. 6. k 1, versus shear strain for Pile 40 in stiff clay
63 23208.8 65 357.1 ,
64 43783.3 95 460.0
i
' ' i

Table 5. Khmax VALUES RELATED WITH . \! I

'

I I
UNDRAINED SHEAR STRENGTH IN •• I '

SOFT CLAY
'
1'\ I'
Pile No. Khm.u. (kPa) s" (kPa) khrnaJSu ••
"' ~
I
I
31 16692.3 14.4 1159.2 • 1---- j-...._
'

I
36 22777.8 28.7 793.7 ••
37 15199.9 14.4 1055.5 •""'0002 OQDII
I i
0.1.11&
..
ll-ODB O.Q1
'
,, . !-i--- Q_Q14 0016 ~Q1· Q_

__
~,)
53 48887.3 28.7 1703.4
Fig. 7. Normalized kJkhmn versus shear strain for Pile 40 in
54 50541.9 28.7 1761.0 stiff clay
, ,.
-Pie<40
' --PM63
DEVELOPING TH~; LOAD-DEFLECTION CURVES
• 1\ I ---PM64

Stiff Clay. 'lbe range ofkhmax has been determined from 300 S" I~ i I I
' I
I
to 600 Su. A procedure to predict Qg-y g curves is described as • '
:
~\
' ' I '
follows:
1. Select a value of y g for which Qg will be computed
,,_, r i
I i - !' i
I ! '
2. For the above deflection, calculate strain m.ing N
!' ~I
'

~~

~
Equation ( 6) • ' .
'I '
3. Calculate the soil modulus kh from Equation (10) '
I
''
using khmax = 300 Su •
4. Calculate the relative stiffness factor R using . ~ ~ b.. t---
k
1 sesttt0yv~
Equation (8) '
5. Calculate lateral load from Equation (7) 1"'::
6. Repeat steps 1-5 for all the deflections at which "
lateral loads are desired ~ SlraiCT)

7. Repeat steps 1-6 using khmW< = 600 Su Fig. 8_ Normalized kJk..mu versus shear strain for four piles
8. Then plot the load-deflection curves as in fig. 10 for in stiff clays

Fourth International Conference on Case Histories in Geotechnical Engineering

Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu
 391
,_, I

.. f\ I I
! -Pile31
-PM34
--Pile36
•

•• ~ -Pie37
--PWtSJ •
~'\ [\_ --Pie 54
I

,_,
•
••

••
Rest Fit
' "'
l\
~~ I '
culve.) "'
~ ~ ['... ~
!

' '
!_,.
!.
•
~
., I
- ~ ~~ •
~
I

.., I
I I I
0.000 0.002 Q_OIM 0 000 0_0011 0 010 0.012 0 014 0.016 0.01! 0 U20
• 0

s.- S&ni!(' )
Fig. 9. Normalized kh/khmax versus shear strain for six piles
in soft clays (Chen 1994) Fig. 10. Measured curve compared to lower bound (khmu=
300S.) and upper bound (k,~ 600S.) for Pile 39 in stiff clay
Pile No 39 in stiff clay. Similarly, in Figs. 11 through 13, the
computed load detlection curves tor Pile No 40, 63 and 64,
respectively are plotted both for the lower bound (khm.•x = 300 S,J ""
and upper bound (khmax = 600 Su) ofkhmax·

Soft Clay. The range of khmax is from 800 Su to 1800 Su. The
.. Uppcr Bound

M~

following method is used to predict Qg-Yg curves.

1. Select a value of y g for which Qg will be computed
2. For the above deflection, calculate strain using
...
...
~·;·

Equation 6
3. Calculate the soil modulus kn from Equation (14) z
"
..
using knm'" = 800 Su
4. Determine the value ofnh from the known value of c!
kh using Equation (13) where x = I m !
5. Calculate the relative stiffness factor T for selected
detlection using Equation ( 12)
6. Calculate the lateral load from Equation (II) for the ""
selected deflection.
7. Repeat steps 1-6 for other deflection values.
8. Repeat step 1-7 using k 11 max = 1800 Su
9. Then plot the load-deflectiOn curves as in Fig. 14 for
Pile No 31. Figures 15 through 18 show similar plots fOr Piles oL---~----~--~~--~--~
0 10 Hi 211
No 36, 37, 53 and 54, respectively.

It should be noted that the recommendations correlating kh and

Su are averaged over the whole range of values for piles in both Fig. 1 1. Measured curve compared to lower bound (khmn=
stiff or soft clays. l11e predictions have been then made with 300SJ and upper bound (k,m ..~ 600SJ for Pile 40 in stiff clay
these average kh values for each pile in Table 2.

Soil parameters required to develop the p-y curve were not

COMPARISON WITH MEASURED CURVES AND THE available for all the piles rl->ported in this paper. Figures 10, II,
p-yMETHOD 12 and 13 show a comparison of Qg-Yg curves from the p-y
method and the present method with the measured load-
The procedure to establish the Qg-yg curve using the p-y curve deflection curve for Piles 39, 40, 63 and 64, respectively, in stiff
method is described elsewhere and will not be repeated here clays. Figures 14, 15,16 17 and 18 show similar plots for Piles
(Reese 1984, Prakash and Sharma 1990, Chen 1994). 31, 36 37, 53 and 54, respectively, in soft clays.

Fourth International Conference on Case Histories in Geotechnical Engineering

Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu
 392

In Table 6 are presented the maximum range of variation with the maximum variation of the measured values with the upper
the measured values by the present and the p-y methods at bound curve with khmax = 1800 Sll, the variation is minimum with
different deflections. respect to the lower bound curve with khmax = 800 Su. This is
because the lower bound prediction almost matches the
measured curve.
Table 6. TOTAL VARIATION IN THE Q,-y, COMPARED
WITH p-y METHOD AND k,ik,m,. APPROACH The advantage of this method is that we obtain the upper and the
WITH MEASURED VALUES lower bound load-deflection curves. Therefore, an engineer has
more confidence in load-deflection predictions because the
Pile Soil Deflection Percent Variation
maximum probable range of variation is known in advance.
No. Typo (mm)
p-y Present Method
Method
CONCLUSIONS
Maximum Minimum

39 Stiff 6 83.3% 27.5% -25.4% The proposed method for predicting the load-deflection
Clay relationship of single piles under lateral loads in clay is based on
12 51.2% 37.6% -18.1% the analysis of six piles in soft clay and four piles in stiff clay.
The fOllowing conclusions have been drawn from this study.
40 5 26.2% -0.1% -40.7%

15 -11.3% 3.0°AJ -38.5%

Stiff Clay
63 10 26.2% 55.9% -7.2%
I. The modulus of the subgrade reaction is constant with depth.
20 1.8% 36.1% -18.9% 2. The knmax has been related to the undrained shear strength.
The range of~max values varies from 300 Su to 600 Su, where Su
64 15 7.5% 30.1% -22.6%
is average undrained shear strength up to 5 pile diameters below
20 -7.2% 47.6% -12.1% the ground.
3. A procedure to predict load-deflection curves of single piles
31 Soft 20 -19.8% 54.9% -7.8% in clay considering soil non-linearity has been proposed.
Clay 4. Maximum variation in the load-deflection plot using this
50 -24.9% 71.8% 2.5o/o method is 55.9% and that by using p-y approach is 83.3%. This
proposed method is simple and more realistic than the current
36 6 -8.4% 31.7% -19.2%
design method (p-y method).
12 -13.7% 36.5% -16.2%

37 6 -13.3% 40.3% -17.1% Soft Clay

12 -21.2% 38.6% -14.6% 1. The modulus of the subgrade reaction increases linearly with
depth as in non-cohesive soils and k 11 has been defmed at one
53 20 NIA 16.7% -29.0%
meter depth.
40 NIA 34.5% -17.4% 2. The modulus degradation with shear strain is independent of
the pile material.
54 20 NA! 17.8% -27.TYr. 3. The khmax has been related to the undrained shear strength.
The range of khmax values varies from 800 Su to 1800 S11, where
40 N/A 36.7% -16.1% Su is the average undrained shear strength up to 5-pile diameters
N/A: Not Available below the ground.
4. A procedure to predict load-deflection curves of a single pile
in clay considering soil non-linearly has been proposed.
It will be seen from Table 6 that the maximum variation in the 5. The maximum variation in the load-deflection plot using this
loads at corresponding deflection by the present method in stiff method is 71.8% and that by using the p-y approach is 24.5%.
clay is 55.9% and by the p·y method is 83.3%. In this case, the 6. The range of variation ofkhma~ may be improved as and when
present method has a better prediction than the p-y method. The more field data becomes available for similar analysis.
maximum variation of the present method in soft clay is 71.8%
and the p-y method is 24. 9%. It may seem that the p-y method A practical advantage of this method, both for stiff and soft
has a better prediction. But the present method gives two clays, is that the present method predicts the upper and lower
predicted boundary values. For Pile 31 (Figure 14), which has bound load-deflection relationship. The one which has the
Fourth International Conference on Case Histories in Geotechnical Engineering
Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu
 393
Upper Bound \kasurcd

! .. ~--t--+--.L-+----7f'=-
!" . ~-+--++-+-1-----c..l=~

~~.~~~~~~_L~~,~-.L,~~,.~~,.
Olledlon. y9 (nwr~)

~y,(mm)

Fig. 12. Measured curve compared to lower bound Fig. 15. :Uusured curve compared to lower bound
(k, ..=300S.) and upper bound (k,..,,=600SJ for Pile 63 in (k,..,=800SJ and upper bound (k.,...=l800SJ for Pile 36 in
stiff day
- I
-
I I
sort clay

-
i L'ppcr Dmlnd /

i i } I
!
d'-
I
I
I
v Mo~
i I
p-y !·~+-~~-¥~~~

! ...
!
I ~~ ~/~woL, "!·~--+~~~~~~
ll, 1/ v
' ' I
: I

- ~v /
)
I
I

.f ! '
I
'' • w
" ~

Fig. 13. Measured curve compared to lower bound Fig. 16. Measured curve compared to lower bound
(k,.. =300S,) and upper bound (k,. • .,=600S,J for Pile 64 in (k,..,=800S,) and upper bound (k,....=ISOOS.) for Pile 37 in
stiff day sort clay
. 1Jpper Boll11!1
'
v
-
llpp<> Round

/ /
·~r-----------~~----------~
/
v Mv-
•

• v
;!/v
/
v
/
v
Lo ..·erBound

.I v
•

•!,--~,~-~!-~~~-~~~~~-~=-~ro~~M
o.lltdlon, y9 (nwn)
•
w " ..
Fig. 14. l\leasured cunre compared to lower bound Fig. 17. )1easured curve -~-compared to ]ower bound
(k,. .. =SOOS,) and upper bound (k,. ... =lSOOS,) for Pile 31 in (k• .,=SOOSJ and upper bound (k,....=1800SJ for Pile 53 in
soft day soft clay

Fourth International Conference on Case Histories in Geotechnical Engineering

Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu
 394

Brown, D.A., L.C. Reese, and M.W. O'Neill (1987), "Cyclic

~
Lateral Loading of a Large-Scale Pile Group," Journal of

. Uppor fiouod

v
/
Geotechnical Engineering, ASCE, Vol. 113, No. II, pp 1326-
1343 .

I
- /
[7
/

::t
- Chen, M.H. (1994), "Lateral Load Deflection Prediction of
Single Piles in Clays using Non-Linear Subgrade Reaction," MS
.r • Thesis, University of Missouri-Rolla .
! r~ / /
l.-/1
Lo"er Buomd I
• Davisson, M.T. (1960), "Behavior of Flexural Vertical Piles

•
v
v ~ I Subjected to Moment, Shear and Axial Load," PhD Thesis,
University of Illinois, Urbana .

• ! I
'

Davisson, M.T. (1970), "Lateral Load Capacity of Piles,"

'// I r
Highway Research Record No. 333, Highway Research Board,
, •
Washington D.C., pp 104-112.
" • •
Davisson, M.T. (1995), Personal Communication.

Fig. 18. Measured curve compared to lower bound (khmu= Davisson, M.T. and H.L. Gill (1963), "Laterally Loaded Piles In
8008,) and upper bound (k,,..,.~ 18008,) for Pile 54 in soft a Layered Soil System," Journal of Geotechnical Engineering,
clay ASCE, Vol. 89, No. SM3, May, pp 63-94.

Dunnavant, T. W. (1986), "Experimental and Analytical

maximwn variation may also have the minimum variation, if the Investigation of the Behavior of Single Piles in Overconsolidated
predicted curve matches with either the lower bound or the Clay Subjected to Cyclic Lateral Loads," UMI Dissertation
upper bound curve. Information Service, Ann Arbor, Michigan.

Kumar, S. (1993), "Non-Linear Load-Deflection Prediction of

ACKNOWLEDGMENT Smgle Piles in Sands, Using Subgrade Reaction Approach,"
Master of Science Thesis, University of Missouri-Rolla.
Comments were o±Iered by Sanjeev Kumar, and Sami Arsoy.
Alex Wu assisted in drafting some figures. Charlcna Ousley Kagawa, T. And L.M. Kraft (1980), "Lateral Load-Deflection
typed the manuscript with great care. All assistance is gratefully Relationship for Piles Subjected to Dynamic Loadings," Soils
appreciated. and Foundations, J., SM &FE, Vol20, No 4, pp 19.34.

Long, J.H. and L.C. Reese (1987), "Behavior of2 Piles in Clay
REFERENCES Subjected to Cyclic Lateral Loads," Proceedings of International
Symposium on Prediction and Performance in Geotechnical
Blaney, G.W. and M.W. O'Neill (1986), "Measured Lateral Engineering, Calgary, June, pp 97-104.
Response of Mass on Single Pile in Clay," Journal of
Geotechnical Engineering, ASCE, Vol. 112, April, pp 443-457. Matlock, H. (1970), "Correlations for Design of Laterally
Loaded Piles in Soft Clay," No. OTC 1204, Offshore
Bowles, J.E. (1988), "Geotechnical Properties; Laboratory Technology Conference, Houston, pp 577-594.
Testing; Index Settlement and Strength Correlations,"
Foundation Analysis and Design, 4th Edition, McGraw-Hill Mwindo, J.M. (1992), "Strain Dependent Soil Modules of
Book Co., pp 24-26. Horizontal Subgrade Reaction," Master of Science Thesis,
University of Missouri-Rolla.
Brinch Hansen, J. (1961), "The Ultimate Resistance of Rigid
Piles Against Transversal Forces," Danish Geotechnical Institute Poulos, H. G. (1971), "Behavior of Laterally Loaded Piles: !-
Bulletin, No. 12, Copenhagen, pp 5-9. Single Piles," Journal of the Soil Mechanics and Foundations
Division, ASCE, Vol. 997, No. SM5, Proc. Paper 8092, May, pp
Broms, B.B. (1964), "Lateral Resistance of Piles in Cohesive 711-731.
Soils," Journal of the Soil Mechanics and Foundations Division,
ASCE, Vol. 90, No. SM#, Proc. Paper 3909, May, pp 123-156. Prakash, S. And S. Kumar (1996), "Non-linear Lateral Pile
Deflection Prediction in Sands," J. Geot. Engg. ASCE, Voll22
Fourth International Conference on Case Histories in Geotechnical Engineering
Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu
 395
No 2 Paper No 9346 February, pp 130-138.

Prakash, S. And H.D. Sharma (1990), "Analysis and Design of

the Pile Foundations Under Lateral Loads," Pile Foundation in
Engineering Practice, John Wiley & Sons, Inc., pp 322-472.

Reese, L.C. ( 1984), "Handbook on Design of Piles and Drilled

Shafts Under Lateral Loads," U.S. Department of
Transportation, Federal Highway Administration.

Reese, L.C., W.R. Cox and W.D. Koop (1975), "Field Testing
and Analysis of Laterally Loaded Piles in Stiff Clay," Proc. 7th
OTC, Paper No. 2312, Houston, Texas, May, VoL II, pp 671-
689.

Reese, L.C. and H. Matlock (1956), "Non-Dimensional

Solutions for Laterally Loaded Piles with Soil Modules Assumed
Proportional to Depth," Proceedings, Eighth Texas Conference
on Soil Mechanics and Foundation Engineering.

Reese, L.C. and R.C. Welch (1975), "Lateral Loading of Deep

Foundation in Stiff Clay," Journal ofGeoteclmical Engineering,
ASCE, Vol. 101, No. GT7, July, pp 633-649.

Robinson, K.E. (1979), "Horizontal Subgrade Reaction

Estimated From Lateral Loading Tests on Timber Piles,"
Behavior of Deep Foundations, ASTM STP 670, Raymond
Lundgren, Ed. American Society for Testing and Materials, pp
520-536.

Skempton, A.W. (1951), "The Bearing Capacity of Clays"

Building Research congress, Division I, Part 3, London, pp 180-
189.

Terzaghi, K. (1955), "Evaluation of Coefficient of Subgrade

Reaction," Geotechnique, London, England, Vol. 5, pp 297-336.

Winkler, A.B. (1867), Die Lchrc Von Elastiziat und Festigkeit,

Fourth International Conference on Case Histories in Geotechnical Engineering
Publisher,
Missouri University of Prague.
Science and Technology
http://ICCHGE1984-2013.mst.edu