Response of Laterally Loaded Rigid Piles

Respuesta de Pilotes Rígidos Cargados Lateralmente

Wei Dong Guo
Griffith University, Gold Coast 9726, Queensland, Australia

Abstract
Design of laterally loaded rigid piles generally lies in selection of soil parameters that may be obtained from
experimental tests. However, the parameters may be significantly affected by the pile dimension, and pressure
distribution around and down the pile. Capacity of a laterally loaded rigid pile has generally been estimated using
some assumed soil pressure profiles, which however, should depend on the magnitude of the pile displacement.
In this paper, elastic-plastic solution for laterally loaded rigid piles has been established considering yield from
pile top and base. The solution can duly overcome the abovementioned limitations using load transfer approach.
With the solution, lateral pile response is examined and compared with relevant measured data.

Resumen
El diseño de pilotes rígidos cargados lateralmente se basa generalmente en la selección de parámetros del suelo
que se obtienen de ensayos experimentales. Sin embargo, los parámetros pueden ser afectados significativamente
por la dimensión del pilote, y por la distribución de la presión alrededor y debajo del pilote. La capacidad de un
pilote rígido cargado lateralmente se ha evaluado generalmente usando perfiles estimados de la presión del suelo,
los cuales dependen a su vez de la magnitud del desplazamiento de los pilotes.
En este artículo, se ha establecido una solución elastoplástica para los pilotes rígidos cargados lateralmente que
considera el momento en el que el tope y la punta del pilote ceden. La solución supera las limitaciones
mencionadas anteriormente usando el método de transferencia de carga. Con la solución, la respuesta lateral del
pilote se examina y se compara con datos medidos relevantes.

Experimental results are often used to back-

1 INTRODUCTION estimate the soil parameters. However, little
justification has been made on the approaches of
Design of laterally loaded rigid piles (including the back-estimation. In fact, the approach is
piers and drilled shaft etc.) has been based on significantly affected by the (rigid) pile dimension
empirical information gathered from full-scale (Guo 2001), which has not been considered so far.
tests in field or model studies in laboratory Expressions for estimating capacity of a
(Mayne et al. 1995), which were generally laterally loaded rigid pile were widely developed
presented as empirical expressions. This design based on some assumed soil pressure profiles.
was further developed using centrifuge and These pressure profiles (Brinch Hansen 1961;
numerical (FEM, FBM) modeling (Laman et al. Broms 1964; Petrasovits and Award 1972;
1999) to account for the non-linear pile-soil Meyerhof et al. 1981; Prasad and Chari 1999) are
interaction. The modeling, however, does not generally different from each other and even from
allow the mechanism of the pile-soil interaction to the measured ones (Prasad and Chari 1999). The
be readily explored. Therefore, the modeling has measured profiles obtained by various tests are the
not been widely adopted in practice. The key of results of the interaction between pile
the design generally lies in selection of soil displacement and limiting force profiles along the
parameters, such as soil modulus, etc. pile-soil interface. Given a definite limiting force
profile such as a linearly increase profile (Broms made, which will be uniquely determined via a
1964), the pressure profile can be different as the new approach established herein.
displacement (Fig. 1 (a)) develops such as those For laterally loaded piles, the interaction
shown Fig. 1(b) for the pre-tip yield state that is between a pile and soil is generally in a three-
similar to that by Prasad and Chari (1999), and dimensional (3-D) manner, such that a realistic
Fig. 1(d) for the fully plastic state that by Brinch exploration generally requires using complex
Hansen (1961); and Petrasovits and Award numerical methods. However, using a simplified
(1972). Therefore, the existing solutions may be soil displacement model, the 3-D interaction may
corresponding to different states of the pile-soil be readily modelled using a load transfer approach
interaction. The compatibility between an (Guo and Lee 2001). In this paper, using the load
assumed pressure profile and pile displacement transfer approach, a rigorous elastic-plastic
profile in the above-mentioned solutions has not solution will be established, assuming the limiting
been (or cannot be) checked, neither the non- force increasing linearly with depth. The solution
uniform circumferential pressure distribution will be compared with relevant measured data.
around the pile diameter (Prasad and Chari 1999), Focusing on yield initiated from pile- top and tip,
which may affect value of modulus of subgrade the solution offers non-linear predictions, in
reaction (Guo 2003). explicit closed form expressions, in particular, of
the pile head-load and displacement relationship
Tt that allows the capacity and the corresponding
e
displacement to be readily estimated. The solution
zo Plastic also permits the limiting force profile and shear
modulus to be uniquely determined.
zr zone
l zo 2 SIMPLE SOLUTIONS
Elastic zr
zone 2.1 Coupled load transfer model
For a laterally loaded pile as illustrated in Fig.
1(a), the laterally loaded pile may be sufficiently
accurately simulated by a load transfer approach
(a) Pile - soil system (b) Pre-tip yield state (Guo and Lee 2001) that essentially offers a
governing equation for the pile similar to that
obtained by modeling the pile-soil interaction as a
series of independent (uncoupled) springs
distributed along the pile shaft and at the tip. The
z o = zr = z 1 springs may be assumed as an elastic-perfectly
zr
zo plastic medium (Guo 2002). If the gradient of a
z1 lateral force, p (simply referred to as force late) is
less than a limiting value py described late, then
the force, p at any depth is proportional to the
local pile displacement, u [L] with a gradient of
(c) Elastic-plastic state (d) Fully plastic state modulus of subgrade reaction, kd [FL-2]:
p = kdu (Elastic state)(1a)
Figure 1 Schematic analysis for a rigid pile
where d = pile diameter or width [L]; py =
The ultimate capacity is generally determined gradient of limiting force at depth z [FL-1] for
as a load corresponding to a definite displacement plastic zone; and p = force per unit depth [FL-1]
(e.g. 0.2d), once a load-displacement curve is for elastic zone, respectively. Following Guo and
available. This practice is certainly inconsistent Lee (2001), the following conclusions will be
with the existing expressions for the capacity directly adopted: (1) The modulus, kd (written as
derived using the aforementioned soil pressure k previously) may be estimated by
profiles. To resolve this problem, a state (say tip
3πGs   
2
2   K1 (γ b ) 
yield) parameter may be introduced, to correlate K1 (γ b )
the effect of a pressure profile to that of the kd = 2γ b − γ b   − 1  (2)
2  K o (γ b )   K o (γ b )  
magnitude of displacement. This parameter will   
allow a consistent definition of the capacity to be
where Gs = soil shear modulus. The ‘k’ is a • The pile rotates against a point, zr at which the
constant with depth. (2) The factor ‘γb’ may be displacement equal zero, thus
estimated by zr = − ut θ (6)
γ b = k1 (ro l ) (3)
Soil pressure, kPa
where l = the pile length; ro = radius of an -400 -200 0 200
equivalent solid cylinder pile. k1 = 2.14, and 3.8 0
for lateral loading (e = 0), and moment loading (e

Depth, z (mm)
= ∞), respectively. And (3) a pile is defined as
‘rigid’, should the pile-soil relative stiffness,
(Ep/G*) be greater than a critical stiffness ratio, CF 200
(Ep/G*)c that is equal to 0.052(l/ro)4/(1+ 0.75νs),
G* = (1+3νs/4)Gs, νs = Poisson’s ratio of the soil; 0.25
and Ep = Young’s modulus of an equivalent solid 0.5
400
pile. 0.75
2.2 The problem
The problem addressed here is previously 600
illustrated in Fig. 1(a), where a rigid pile is
subjected to a lateral load, Tt at an eccentricity of
‘e’ above the ground surface. Should the pile be Figure 2 The predicted (CF) versus the measured
flexible, relevant solutions developed previously data at various ID as shown
(Guo 2002) may be adopted directly. However,
the possible tip yield (i.e. p = py) for a rigid pile • For the pile in cohensionless soil, the gradient
renders the solutions invalid. The soil pressure of net limiting force varies linearly with depth:
profiles mobilised along the pile as illustrated in p y = Ar zd (Plastic state) (1b)
the Fig. 1 (b-d) are different from those along a
flexible pile. Fig. 2 shows a typical measured where Arz = pressure on pile surface [FL-2]; py =
profile (as discussed late in the case study), which gradient of limiting force at depth z [FL-1] for
corresponds to the pre-tip yield case of Fig. 1(b). plastic zone. Within plastic state, the force
between pile-soil interface at any depth equals the
2.3 Assumptions and critical states limiting force. The gradient of the linearly
Using a load transfer model similar to the increasing pressure may be taken as (Fleming et
current model for a pile in a homogenous soil, al. 1992):
solutions were derived previously for limiting soil
pressure constant with depth (Scott 1981), and Ar = γk p2 (7)
linearly increase with depth (Motta 1997). Both
solutions are useful, however, the latter solution where kp = the coefficient of net passive earth
seems to offer an unexpected expression for the pressure on pile surface; γ = unit weight of the
‘depth of rotation’. Therefore, solutions are re- soil. Eq. (7) offers prediction in good comparison
derived for this case. For a rigid pile, it follows: with the measured data by Prasad and Chari
• The displacement, u(z) may be expressed as (1999) (Fig. 2).
• The limiting force by Eq. (1b) should equal that
u = θz + u t (4)
by Eq. (1a) at the critical depth, zo, thus the
threshold displacement, u* may also be given as
where ut, θ = pile-head displacement and rotation
angle, respectively. u * = Ar z o k (5b)
• As shown in Fig. 1, there exists a depth, zo
above which the pile-soil interface may have • Eq. (5a) is valid in pre- and post- tip yield
undergone plastic deformation; below which the states. To distinguish the states, the parameters θ,
pile may be still in elastic state. At this depth ut, and zo will be rewritten as θy, and uty, and zy for
(referred to as critical depth), the pile the latter state.
displacement reaches a threshold value of u* • The tip yield starts at a depth of l, and then
given by moves up to a depth of z1 (Fig. 1(c)), the
corresponding tip displacement is -u*l/zy,
u * = θzo + ut (5a)
therefore from Eq. (4), it follows,
− u *l / z y = θ y z1 + u ty the further downwards extension of the existing
(8) yielding zone initiated from the pile top. To
Given that z1 equals l, zy equals z o . simulate this behaviour, the horizontal force and
the moment equilibrium, against the pile-head
2.4 Solutions for the pre-tip yield state using the limting force distribution illustrated in
Applying a lateral load, Tt at a height of ‘e’ above Fig.1(c), allow two governing equations to be
the ground level, the force, and the moment established. Solution of the equations leads to the
equilibrium against the head give two expressions, following expressions (Guo 2003):
• The normalised head-load, Tt is given by
which allow the following expressions (before the
tip yields) to be derived (Guo 2003): Tt  2  zr  2 
=  − 1
 1 − C 2  l 
• The nomarlised head load, Tt may be given by 0.5 (12)
Ar dl 2 
Tt zo l  
2
= (9)
Ar dl 2(2 + z o l + 3 e l ) • The pile-head displacement, ut is given by
A z 1
• The normalised head displacement, ut is given ut = r r l (13)
by k l C
ut k (3 e l + 2) z o l • The depth, zr is given by one of the following
= (10)
Ar l (2 + z o l + 3 e l )(1 − z o l )2 expressions
z r l = 0.5(1 − C 2 )(δ 2 − e l )
• The rotation angle, θ of the pile is expressed as
Ar z o ( z o l ) − 3 + 3( z o l − 2) e l
2
( )(
z r l = 0.5 1 − C 2 − 0.5δ 2 + 0.5δ 1 − 3 − e l ) (14)
θ= (11)
k l (2 + z o l + 3 e l )(1 − z o l )2
( )(
z r l = 0.5 1 − C 2 − 0.5δ 2 − 0.5δ 1 − 3 − e l)
The ratio of e/l is a constant for a given where δ j = 3 A1 + (−1) j 3 A2 (j = 1, 2)
problem. As the value of zo/l increases, waxing
pile-head load, displacement and rotation may be 3
(
A j = −(e l ) + 2 D − 2(−1) j D 0.5 − (e l ) + D
3
)
1/ 2

obtained readily using Eqs. (9), (10) and (11).

Only the Tt is dependent of the pile diameter D = (1.5e / l + 1) / (1 − C 2 )
(width). Clearly plastic zone is developed (zo/l >
0) wherever a load is applied. Therefore, the The rotation depth, zr may be estimated for each
validity of the available elastic analysis (Scott constant of C that must be lower than a limiting
1981; Sogge 1981) should be checked against the value, Cy to ensure the tip yield have occurred.
current elastic-plastic solution. The height of ‘e’ While the Cy, for instance, may be obtained by
may be the real distance from the ground level to taking the zr/l as 0.764, or 0.667 (as further
the loading point of the Tt; it may also be a discussed late) at e/l = 0 or ∞, respectively.
fictitious length introduced to accommodate the
case when a moment (M = Tte) is applied. It is 2.6 Full plastic-state (Rotation point yield)
worth to stress that as the value of ‘e’ approaches As pile-head displacement approaches an
infinite, expressions (9) through (11) reduce to infinite value (by taking C → 0), Eq. (12) reduces
those for pure moment loading. However, the to the form presented previously by Pestrasovits
subgrade modulus due to lateral loading should be and Award (1972). The rotation depth, the slip
different from that due to moment loading as depths from pile-head, and tip become identical.
treated elsewhere (Guo 2003). In the late case For instance, the zo/l (= zr/l) is estimated to be
study, the modulus was taken as that in between 0.794, and 0.707 for e/l = 0, and ∞, respectively
those for lateral load and moment, which (Guo 2003). The rotation depth itself depends on
fortunately seems to offer little difference from the e/l and the C, or the normalised displacement,
the previous estimation (Guo 2003) in the (Arl)/kut as inferred from Eq. (13).
obtained moduli.
2.7 Complete elastic-plastic response
• Value of zo at tip yield
2.5 Solutions for the post-tip yield state
After the pile tip yields, a new yielding zone At the moment of the tip yielding, the depth, zo is
appears around the pile tip, which will gradually rewritten as a critical depth, z o which has been
extend up towards the rotation point in contrast to obtained from Eq. (8) as
zo  e  e e either from Eq. (7) or a measured force profile;
= −1.5 + 0.5  + 0.5 5 + 12 + 9( ) 2 (15) the modulus ‘k’ thus can be back-estimated by
l  l  l l matching the predicted versus the measured load-
Table 1 Parameters for the tip yield state displacement (or rotation) curve.
e/l z o /l zr/l [utk/(Arl)]y Cy
0.01

Normalised lateral load, Tt /(Ardl )

0 0.618 0.764 3.235 0.236

2
∞ 0.50 0.667 4.0 0.167
0.008
ID = 0.25
0.006
2500 0.5
ID = 0.75
0.004 0.75
2000 Lines: CF predicted
Dotted points:
0.002
Load, Tt (N)

1500 Measured data

0
0.5
1000 0 1 2 3
0.25 kut /(Arl )
500 Tip yield
Figure 4 Predicted normalised relationship
0.2d
0 between Tt and ut at different ID
0 50 100 3 CASE STUDY
Displacement, ut (mm)
3.1 Model tests by Parasad and Chari (1999)
A total of 15 steel model pile tests were
Figure 3 The (CF) predicted versus the measured performed in dry sand at three different relative
Tt and ut relationships at various ID densities, ID of 0.25, 0.5 and 0.75. The
corresponding soil bulk density γ and frictional
• Constructing load-displacement curve angles φ are: 16.5 kN/m3 and 35o; 17.3 kN/m3 and
An expression correlating the head load and 41o; and 18.3 kN/m3 and 45.5o. For all the tests,
displacement may be derived, for instance, using each pile was 102 mm in an outside diameter,
Eqs. (9), and (10) for the pre-tip yield state. 1135 mm in length, and 5.6 mm in a wall
However, it looks complicated, thus a simple thickness. The load was applied at a normalised
procedure is suggested here to construct the curve: eccentricity, e/l of 0.245. The measured soil
(a) For a given ratio of e/l, the normalized critical pressure profiles were provided in Fig. 2.
depth, z o /l is estimated using Eq. (15); the critical Following the procedure described previously, a
value, [utk/(Arl)]y using Eq. (10) by replacing the full pile-head load and displacement relationship
zo/l with the z o /l; the zr/l using Eq. (6); and finally is predicted and shown together with the measured
the Cy using Eq. (13) with the estimated zr/l. Two data in Fig. 3 for each density, ID. The ‘ultimate’
critical cases were examined in this manner and load corresponding to tip yield is found to be
presented in Table 1. (b) Assuming a series of different from that to a displacement of 0.2d.
increasing values of zo/l (< z o /l), the waxing load Thus, a full load-displacement curve is preferred.
and displacement may be determined using Eqs. The shear moduli provided in Table 2 for the three
(9), and (10), respectively. (c) Given the ‘C’ (< different ID are only about 4.5~7.5% the derived
Cy), the rotation depth zr/l may be evaluated using soil modului at pile tip level, which are 3.78, 6.19
Eq. (14). (d) Each zr/l is then used to calculate the and 9.22 MPa (taking νs = 0.3), respectively
load using Eq. (12), and displacement using Eq. (Parasad and Chari 1999). This difference shows
(13), respectively. the significant effect of the approaches used in
Steps (a) through (d) allow a complete pile-head back-estimating the modulus. The normalised
load and displacement curve to be constructed and relationship for each ID, as provided in Fig. 4,
plotted in Fig. 3 as ‘CF predicted’ with only two shows that the curves are only dependent of the
parameters (k and Ar). The Ar can be determined zo/l as also inferred from Eqs. (9), and (10).
 Fleming, W. G. K., Weltman, A. J., Randolph, M. F.
Table 2 Parameters for the case study and Elson, W. K. (1992). Piling engineering. New
York, Surrey University Press, Glasgow, Halst
ID 0.25 0.50 0.75 Press.
Guo, W. D. (2001). Subgrade modulus for laterally
Ar (kN/m3) 244.9 340. 739. loaded piles. Proceedings, 8th Internationl
Conference Civil and Structural Engineering
k (MN/m3) 3.88 12.05 16.96 Computing, CIVIL-COMP2001, Eisenstadt, nr
Vienna, Austria.
Gs (MPa) 1.68d 4.45d 6.74d Guo, W. D. (2002). "On critical depth and lateral pile
response." Journal of Geotechnical and
Geoenvironmental Engineering Division ASCE,
Note: l = 621 mm, d =102 mm tentatively accepted.
Guo, W. D. (2003). "Laterally loaded rigid piers in
3.2 Comments on current predictions cohesionless soil." submitted to 'Journal of
Only the net lateral resistance was taken into Geotechnical and Geoenvironmental Engineering
account in the current solutions. For very short Division'.
Guo, W. D. and Lee, F. H. (2001). "Load transfer
piers such as pole foundations, the vertical approach for laterally loaded piles." Internationl
resistance along the pier, and the soil resistance Journal for Numerical and Analytical Methods in
from the pier base may become significant. Geomechanics 25(11): 1101-1129.
Ignoring these resistances, the solutions may be Laman, M., King, G. J. W. and Dickin, E. A. (1999).
conservative (Vallabhan and Alikhanlou 1982). "Three-dimensional finite element studies of the
moment-carrying capacity of short pier foundations
The stiffness was assumed as a constant, and in cohesionless soil." Computers and Geotechnics
limiting force profile as linearly increase with 25: 141-155.
depth. Otherwise, the back-figured modulus can Mayne, P. W., Kulhawy, F. H. and Trautmann, C. H.
be significantly different from what was obtained (1995). "Laboratory modeling of laterally-loaded
here. drilled shafts in clay." Journal of Geotechnical and
Geoenvironmental Engineering, ASCE 121(12):
827-835.
4 CONCLUSIONS Meyerhof, G. G., Mathur, S. K. and Valsangkar, A. J.
(1981). "Lateral resistance and deflection of rigid
In this paper, elastic-plastic solution for wall and piles in layered soils." Can. Geotech. J. 18:
laterally loaded rigid piles was established 159-170.
considering yield from pile top and base. The Motta, E. (1997). "Discussion on "Laboratory
modeling of laterally-loaded drilled shafts in clay"
solution, based on load transfer approach, allows a by Mayne, P. W., Kulhawy, F. H. and Trautmann,
full load and displacement response to be readily C. H." Journal of Geotechnical and
estimated, which compares well with relevant Geoenvironmental Engineering, ASCE 123?(5):
measured data, so as to avoid the inconsistency in 489-490.
the definition of capacity occurred when using the Petrasovits, G. and Award, A. (1972). Ultimate lateral
available expressions. The estimation of resistance of a rigid pile in cohesionless soil. Proc.
5th European Conf. on SMFE, Madrid.
parameters for the proposed solution, however, Prasad, Y. V. S. N. and Chari, T. R. (1999). "Lateral
was explored and presented in other publications. capacity of model rigid piles in cohesionless soils."
Soils and Foundations 39(2): 21-29.
5 ACKNOWLEDGEMENTS Scott, R. F. (1981). Foundation analysis. N. J.,
Prentice Hall, Englewood Cliffs.
Sogge, R. L. (1981). "Laterally Loaded Pile Design."
The work reported here is currently sponsored Journal of Geotechnical Engrg, ASCE 107(9): 1179-
by Australian Research Council through the 1199.
Discovery project, grant number: DP0209027. Vallabhan, C. V. and Alikhanlou, F. (1982). "Short
This financial assistance is gratefully rigid piers in clays." Journal of Geotechnical
acknowledged. Engrg., ASCE 108(10): 1255-1272.

REFERENCES
Brinch Hansen, J. (1961). "The ultimate resistance of
rigid piles against tranversal forces." Danish
Geotechnical Institute, Copenhagen(Bulletin No.
12): 5-9.
Broms, B. (1964). "The lateral response of piles in
cohesionless soils." Journal of Soil Mechanics and
Foundation Engineering Division, ASCE 90(3):
123-56.