Computers and Geotechnics 26 (2000) 113±144

www.elsevier.com/locate/compgeo

Visco-elastic consolidation subsequent to

pile installation
Wei Dong Guo*
Department of Civil Engineering, The National University of Singapore,
10 Kent Ridge Crescent, Singapore 119260, Singapore

Received 10 March 1999; received in revised form 20 August 1999; accepted 20 September 1999

Abstract
Driven piles normally generate excess pore pressures in the surrounding soil. Dissipation of
the pore pressures following driving is predicted currently by available elastic theory. How-
ever, viscosity is pronounced for many soft clays, therefore, its e€ect should be suitably
accounted for. The gradual increase in pile capacity is dominated by the dissipation of excess
pore pressure as has been widely explored both experimentally and theoretically. To predict
the load-settlement response, the variation of pile-soil sti€ness with the dissipation of pore
pressure must also be quanti®ed. This paper presents closed form solutions for the radial
consolidation of the soil around a driven pile, assuming that the soil skeleton deforms visco-
elastically. The solutions are theoretically valid for any initial distribution of excess pore
pressure. However, the current study is focussed on the logarithmic variation of the initial
pore pressure with radius, due to the expansion of a cylindrical cavity in an ideal elastic, per-
fectly plastic soil. The overall pile response measured from three cases has been used to
back®gure the time-dependent variation of the two key facets: shear modulus and strength.
The back analysis shows that variation of the two key facets can be well predicted with the
presented radial consolidation theory. # 2000 Elsevier Science Ltd. All rights reserved.

1. Introduction

Installation of a driven pile in a clay generally leads to a remoulding of the soil,

some loss in the strength and an increase in pore water pressure in the vicinity of the
pile. Increase in strength with time, subsequent to driving, results in the ®nal soil
strength being equal to, or greater than the initial value [1±4], accompanied by a

* Tel.: +61-3-9386-1819; fax: +61-3-9386-1819.

E-mail address: [email protected].

0266-352X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0266-352X(99)00028-2
114 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

Nomenclature

An ; Bn =coecients for predicting excess pore pressure;

cv =coecient of soil consolidation;
d…ro † =diameter (radius) of a pile;
F…t† =the creep compliance derived from the visco-elastic model;
G =elastic shear modulus;
G j =shear modulus at strain j for spring j (j=1, 2) within the visco-elastic
model;
Ji =Bessel functions of the ®rst kinds and of order i (i=0, 1);
k =permeability of soil;
L =embedded pile length;
m2 =ratio of shear moduli, G 1 =G 2 ;
N =visco-elastic time factor;
R =the radius beyond which the excess pore pressure is initially zero;
r =distance away from the pile axis.
ro =pile radius;
r =the radius at which the excess pore pressure, by the time they reach
there, are small and can be ignored;
s =argument of the Laplace transform;
su =undrained shear strength of soil;
t…t † =time elapsed;
t90 =time for …uo ÿ u†=uo =0.9;
T50 ; T90 =time factor, T for 50 and 90% degree of `consolidation' respectively;
T =rate of consolidation;
T2 =relaxation time,  2 =G 2 ;
Tn …t† =the time for the solution of the reconsolidation theory, also written
as T…t†;
u =pore water pressure;
uo =initial pore water pressure;
uo …r† =initial excess pore water pressure at radius r;
Vi =cylinder function of i-th order;
Yi =Bessel functions of the second kinds and of order i…i ˆ 0; 1†;
z =depth.
Greek
 =average pile-soil adhesion factor in terms of total stress;
c =non-dimensional creep parameter for standard linear model;
 =average pile-soil adhesion factor in terms of e€ective stress;
=shear strain;
w =the unit weight of water;
 =mean total stress;
r0 …0 † =increments of the e€ective stress during consolidation in radial and
circumferential directions;
 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144 115

z0 =increments of the e€ective stress during consolidation in depth

direction;
"r ; " ; "z =strain in the radial, circumferential and depth directions;
"v =the volumetric strain;
 2 …† =shear viscosity for the dash at strain 2 ;
l =relative sti€ness ratio between pile Young's modulus and the soil
shear modulus at just above the base level, Ep =GL ;
ln =the n-th root for the Bessel functions;
s =Poisson's ratio of soil;
 =outward radial displacement of the soil around a pile;
0
vo =e€ective overburden pressure;
 =shear stress;
fj =(maximum) undrained (pile-soil) adhesion (j=1, 2);

gradual decrease in water content in the clay adjacent to the pile, and increase in the
bearing capacity of the pile [5].
The maximum pore pressure occurs immediately following driving, and may
approximately equal, or exceed the total overburden pressure in overconsolidated
soil [2,6]. The magnitude of the pore pressures induced due to driving decreases
rapidly with distance from the pile wall, and becoming negligible at a distance of 5±
10 pile diameters. This distribution of excess pore pressure around a driven pile may
be simulated with sucient accuracy using the cylindrical cavity expansion analogy
[7] or the strain path method [8]. The former theory, though, is a one-dimensional
analysis, generally provides sucient accuracy, in comparison with the latter analy-
sis. A particular advantage of the approach is that it can be readily extended to the
case of visco-elastic soil response.
Using a radial consolidation theory, Soderberg [9], Randolph and Wroth [7] show
that the measured rate of development of pile capacity in soft clay appears to be
consistent with the rate of pore pressure dissipation. Therefore, with the assumption
of an impervious pile, the problem of predicting the variation of capacity becomes one
of predicting the hydrostatic excess pressures at the pile shaft as a function of time.
Dissipation of the excess pore pressures generated during driving leads to an
increase not only in shaft friction but also in the sti€ness of the surrounding soil [2,
10±13]. Accurate prediction of pile behaviour requires determination of the pro®le
of pile±soil interaction sti€ness and limiting shaft friction, which are generally trea-
ted as invariants with time. However, the soil strength is normally signi®cantly
altered by pile driving, which means that the overall pile-soil interaction should be
treated as a time-dependent problem. Many researchers have emphasised the impor-
tance of predicting the load-settlement behaviour [14±16], particularly where piles
act as settlement reducers. However, most research conducted to date has con-
centrated on the time-dependent bearing capacity, rather than how the overall
response is a€ected by soil reconsolidation following pile driving.
116 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

Two basic approaches are commonly used for analysing consolidation problems.
One was developed from the di€usion theory by e.g. Terzaghi [17] and Rendulic [18]
(reported by Murray [19]). The other was developed from elastic theory by e.g. Biot
[20] and more recently by Randolph and Wroth [7] for dissipation of pore pressure
generated due to pile driving.
The di€usion theory is generally less rigorous than the elastic theory. However, the
di€usion theory is mathematically much simpler to apply, and can be readily extended
to account for complex conditions. e.g. soil visco-elasticity, soil shear modulus non-
homogeneity. In fact, the di€usion theory is di€erent from the elastic theory in that
(1) the mean total stress is assumed constant; (2) the coecients of consolidation
derived for the two theories are generally di€erent [19]. If the mean total stress
change happens to be zero, the only di€erence between the two theories is in the
coecients of consolidation. Therefore, a coecient from elastic theory may be used
to replace the coecient in the solution of the di€usion theory, then the solution is
converted into a rigorous solution.
In a word, simulation of the behaviour of a driven pile may be divided into two
stages: stage 1, the generation of the pore water pressure due to installation of a pile,
which may be modelled for example by the cylindrical cavity expansion for an elastic-
plastic medium; stage 2, the radial reconsolidation, following installation of the pile,
which was simulated by elastic theory [7], using the initial conditions resulting from
stage 1. However, the soil generally behaves as a visco-elastic material as reported by
many researchers [21]. The viscosity e€ect becomes important particularly for pre-
dicting load-settlement response [22]. Therefore, the previous reconsolidation [7] has
been extended to account for visco-elastic response in the current paper.
In this paper:
(1) Volumetric strain is ®rst generated and expressed as the sole variable of excess
pore water pressure.
(2) A governing equation is established for radial reconsolidation of a visco-elastic
medium, for which a general solution is established.
(3) Equations for radial consolidation for a given logarithmic variation of initial
pore pressure are provided.
(4) Three case studies are described to illustrate the time variation of pore pressure,
pile capacity, average pile shaft cohesion, and average shear modulus.

2. Governing equation for reconsolidation

The e€ect of driving a pile into clay can be simulated by expansion of a long
cylindrical cavity under undrained conditions in an ideal elastic, perfectly plastic
material, characterised by the shear moduli and the undrained shear strength.
Experiment shows that the expansion is a plane strain problem for the middle part
of the pile [23]. The soil properties and the stress state immediately following pile
driving have been simpli®ed and illustrated in Fig. 1, which o€ers the initial stress
and boundary conditions for the reconsolidation investigated herein.
 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144 117

2.1. Visco-elastic stress±strain model

In the process of reconsolidation, the soil is assumed to be a visco-elastical med-

ium, and is described by a model (also called Mechant's model) as illustrated in Fig.
2. For a prolonged constant loading, the creep compliance, F…t† for the model is [24]

1
F…t† ˆ …1 ‡ m2 …1 ÿ exp…ÿt=T2 ††† …1†
G 1

Fig. 1. Diagram of radial consolidation around a driven pile.

Fig. 2. Visco-elastic model for radial consolidation analysis.

118 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

where m2 ˆ G 1 =G 2 ; 1=T2 ˆ G 2 = 2 ;  2 is the shear viscosity at visco element 2;

G j is an average shear modulus, at the outset of the reconsolidation, for each of the
elastic spring, j for the concerned domain as described later. More generally, in the
model shown in Fig. 2, a Voigt element may be added to simulate some special soil
behaviour [21, 25]. It must stress that (1) at the outset of the reconsolidation, and
just after installation of a pile, the soil has already undergone substantial deforma-
tion. Thus the soil parameters adopted herein should correspond to this deformation
level. To stress this point, the subscript ` j ' indicating the strain level, j for spring j, is
added in the notations for the parameters. (2) During the process of reconsolidation,
the modulus, G j and shear viscosity,  2 remain constants. (3) Subsequent to com-
pletion of the reconsolidation, once a load is added at the pile top, the soil parameters
should be modi®ed to account for the e€ect of the stress (strain) level on pile-soil
interaction, as elaborated previously by Guo [22] and shown in the Appendix.

2.2. Volumetric stress±strain relation of soil skeleton

In this section, the volumetric e€ective stress±strain relationship is ®rst given for
an elastic medium and then the relationship is converted into that for a visco-elastic
medium. The plane strain version of Hooke's law

1   @
"r ˆ …1 ÿ s †r0 ÿ s 0 ˆ ÿ
2G @r
1   
" ˆ ÿs r0 ‡ …1 ÿ s †0 ˆ ÿ
2G r
"z ˆ 0
@"
also "r ÿ " ˆ r …2†
@r

where  is the outward radial displacement of the soil around a pile; r is distance
away from the pile axis; s is the Poisson's ratio of the soil; "r ; " ; "z are strains in the
radial, circumferential and depth directions, respectively; G is the elastic soil shear
modulus; r0 ; 0 ; z0 are the increments of the e€ective stresses during consolidation
in the radial, circumferential and depth directions, with z0 ˆ s …r0 ‡ 0 †. Com-
bining Eq. (2) and the e€ective stress principle, the volumetric e€ective stress±strain
relationship for plain strain cases may be written as

1 ÿ 2s
" ˆ … ÿ …u ÿ uo †† …3†
G

where " is the volumetric strain;  is the total mean stress change,  ˆ 0:5…r ‡
 †; u is the excess pore pressure; uo is the initial value following driving [7]. Radial
equilibrium leads to:

@r0 @r0 ÿ @0 @…u ÿ uo †

‡ ‡ ˆ0 …4†
@r r @r
 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144 119

In a similar way to that by Mckinley [26], from Eqs. (2) and (4), it follows,

1
r0 ‡ 0 ˆ ÿ …u ÿ uo † …5†
1 ‡ s

Eq. (5) is valid, as long as the outer boundary is suciently far way that beyond
the boundary, (i.e r > r ; r is a critical radius as detailed later) the pore pressure, u
and the stresses, r0 ; 0 ; z0 are all zero at any time. Thus, the total mean stress may
be expressed as

1 ÿ 2s
 ˆ …u ÿ uo † …6†
2…1 ÿ s †

and Eq. (3) may be rewritten as

1 ÿ 2s
" ˆ ÿ …u ÿ uo † …7†
2…1 ÿ s †G

Eq. (7) is derived for an elastic medium. Similar volumetric expression for visco-
elastic media may be directly transformed from Eq. (7), using the correspondence
principle [27], by the following procedures.
(1) Applying the Laplace transform to Eq. (1), allowing the transformed shear
modulus G, to be related to the transformed compliance, F…t† by

F…t† ˆ 1=…sG† …8†

where s is the argument of the Laplace transform.
(2) Applying the Laplace transform to Eq. (7), and using F…t† to replace the
transformed modulus, G, to give

…1 ÿ 2s †
"v ˆ ÿ sF…t† …u ÿ uo † …9†
2…1 ÿ s †
(3) Applying the inverse Laplace transform to Eq. (9), to obtain the ®nal
expression of the volumetric strain for visco-elastic media as
 …t 
1 ÿ 2s dF…t ÿ †
" ˆ ÿ …u ÿ uo † ‡ G 1 …u ÿ uo † j d …10†
2…1 ÿ s †G 1 0 d…t ÿ †

At t ˆ 0; v ˆ 0. The Poisson's ratio was regarded as a constant in the above

derivation, and the e€ect of this assumption is generally ignored even for numerical
analysis [28]. In fact, considering the viscous e€ect on Poisson's ratio would lead to a
formidable inverse Laplace transform.
In terms of Eq. (10), the changing rate of volumetric strain may then be expressed by
120 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

 …t 
@" …1 ÿ 2s † 1 @u @u dF…t ÿ †
ˆÿ ‡ G 1 d …11†
@t 2…1 ÿ s † G 1 @t 0 @ d…t ÿ †

It is worth noting that Eq. (11) satis®es both radial equilibrium and Hooke's law.

2.3. Flow of pore water and continuity of volume strain rate

The volumetric strain rate may also be obtained by considering the ¯ow of pore
water and continuity of volume. The pore water velocity may be related to the
pressure gradient by Darcy's law. For continuity, the rate of volumetric strain may
be related to the ¯ow of pore water into and out of any region by [7]
 
@" k 1@ @u
ˆÿ r …12†
@t w r @r @r

where k is the permeability of the soil; and w is the unit weight of water. Eqs. (11)
and (12) may be combined to yield
  …t
1@ @u @u du dF…t ÿ †
c r ˆ ‡ G 1 d …13†
r @r @r @t 0 d d…t ÿ †

where cv is coecient of soil consolidation, under plain strain condition [7], given by

kG 1 2…1 ÿ s †
cv ˆ …14†
w 1 ÿ 2s

In the following parts of this paper, the subscript `` 1 '' in G 1 will be dropped, for
convenience. Eq. (13) is the governing equation for radial consolidation. If the soil is
treated as an elastic medium, then from Eq. (1), it follows: dF…t ÿ †=d…t ÿ † ˆ 0;
which in turn allows Eq. (13) to reduce to elastic case as
 
1@ @u @u
cv r ˆ …15†
r @r @r @t

2.4. Comments and di€usion equation

Generally, from Eq. (6), it may be noted that the total mean stress,  is dependent
on the pore pressure, u. Thus, it is not a constant during the process of consolida-
tion, which is similar to other kinds of consolidation [19, 29, 30]. However, in many
cases, it is dicult to achieve a simple expression for  by u. A common treatment is
that the  is simply taken as a constant or zero [17±19]. With this treatment, a new,
similar equation to Eq. (13) will be derived, albeit that a di€erent coecient cv , is
obtained [21]. Thus, the new equation does not necessarily satisfy the radial equilibrium,
 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144 121

so it is normally referred to as a di€usion equation [19]. The solution of a di€usion

equation is much simpler, and compares well with the corresponding rigorous solu-
tion [19, 31, 32]. In fact, as noted by Murray [19], many of the currently popular
theories are based on this assumption, for instance, the sand drain problem solved
by Barron [33].

3. Boundary conditions

Boundary conditions for radial consolidation of an elastic medium around a rigid,

impermeable pile were detailed previously by Randolph and Wroth [7]. These con-
ditions as described below are generally valid for the visco-elastic case as well:

u jtˆ0 ˆ uo …r†…t ˆ 0; r5ro † …16a†


@u
ˆ 0 …t50† …16b†
@r rˆro

u jr5r ˆ 0 …t50† …16c†

u ˆ 0 as t ! 1 …r5ro † …16d†

where ro is the pile radius; r is some radius beyond which the excess pore pressures
are zero at any time. Initially, u=0 for r5R (R is the width of the plastic zone).
However, during consolidation, outward ¯ow of pore water will give rise to excess
pore pressures in the region r > R, and generally it is necessary to take r as 5±10
times R.

4. General solution to the governing equation

The general solution to Eq. (13) may be obtained by separating the variables for
time dependent and independent parts, i.e.

u ˆ w…r†T…t† …17†

With a separation constant of l2n , it follows

@2 w 1 @w
‡ ‡ l2n w ˆ 0 …18†
@r2 r @r
…t
dT…t† dT…t† dF…t ÿ †
‡ G 1 d ‡ 2n T…t† ˆ 0 …19†
dt o d d…t ÿ †

where
122 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

2n ˆ c l2n …20†

The parameter, ln , is one of the in®nite roots satisfying Eq. (18), which may be
expressed in terms of Bessel functions as

wn …r† ˆ An Jo …ln r† ‡ Bn Yo …ln r† …21†

where An is dependent on the boundary conditions. The functions Jo ; Yo ; J1 ; Y1 , are

Bessel functions of zero order and ®rst order, with Ji being Bessel functions of the
®rst kind, and the Yi being Bessel functions of the second kind.
Cylinder functions, Vi …ln ro † of i-th order may be expressed as [34]

J1 …ln ro †
Vi …ln r† ˆ Ji …ln r† ÿ Yi …ln r† …22†
Y1 …ln ro †

Based on the boundary condition of Eq. (16b), Bn ˆ ÿAn J1 …ln ro †=Y1 …ln ro †. Thus,
from Eq. (21),

wn …r† ˆ An Vo …ln r† …23†


dwn …r†
ˆ An V1 …ln r† jrˆro ˆ 0 …24†
dr rˆro

Also, with Eq. (16c), u ˆ 0 for r5r , it follows

J1 …ln ro †
Vo …ln r † ˆ Jo …ln r † ÿ Yo …ln r † ˆ 0 …25†
Y1 …ln ro †

Eqs. (24) and (25) render the cylinder functions to be de®ned. There is an in®nite
number of roots of ln satisfying these equations.
The time-dependant solution for the standard linear visco-elastic model (Fig. 2) is

…!1 …n† ÿ C † exp…ÿ!1 …n†t† ÿ …!2 …n† ÿ C † exp…ÿ!2 …n†t†

Tn …t† ˆ …26†
!1 …n† ÿ !2 …n†

where
q
c ‡ 2n 1 ÿ
2
!1 …n† ˆ ‡ c ‡ 2n ÿ42n =T2 …27†
2 2
q
c ‡ 2n 1 ÿ
2
!2 …n† ˆ ÿ c ‡ 2n ÿ42n =T2 …28†
2 2

c ˆ …1 ‡ m2 †=T2 …29†
 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144 123

This time factor, Tn …t† is essentially identical to that given by Christie [35] using a
similar model for one-dimensional consolidation. Thus, the correctness of the cur-
rent derivation is veri®ed. For the elastic case of c ˆ 0, and m2 ˆ 0, in terms of Eqs.
(26), (27) and (28), it follows: !1 …n† ˆ 2n ; !2 …n† ˆ 0, thus Eq. (26) reduces to
2
Tn …t† ˆ eÿn t …30†

The full expression for pore pressure, u, will be a summation of all the possible
solutions
X1
uˆ An Vo …ln r†Tn …t† …31†
nˆ1

Normally the ®rst 50 roots of the Bessel functions are found to give sucient
accuracy. With Eqs. (16a) and (31), it follows
… r … r
An ˆ uo …r†Vo …rln †rdr= V2o …rln †rdr …32†
ro ro

The above visco-elastic solutions may be readily obtained, in terms of the elastic
solutions, by the correspondence principle [21].

5. Consolidation for logarithmic variation of uo

The initial stress state for radial consolidation of a visco-elastic medium around a
rigid, impermeable pile is similar to that of an elastic medium [22]. Therefore, rele-
vant equations for elastic soil [7] are adopted herein with suitable shear modulus, as
described below.
(1) For a cavity expanded from zero radius to a radius of ro (pile radius), the stress
change,  within the plastic zone (ro 4r4R as shown in Fig. 1) is given by
ÿ ÿ


 ˆ su ln G 1 =su ÿ 2ln…r=ro † …33†

where su is the undrained shear strength of the soil.

(2) The width of the plastic zone is given by

R ˆ ro …G 1 =su †1=2 …34†

(3) Under undrained conditions, if the mean e€ective stress remains constant, the
initial excess pore pressure distribution away from pile wall varies according to

uo …r† ˆ 2su ln…R=r† ro 4r4R

…35†
uo ˆ 0 R < r < r

where R is the radius, beyond which the excess pore pressure is initially zero.
124 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

In light of the initial pore pressure distribution of Eq. (35), the coecients, An may
be simpli®ed as

4su Vo …ln ro † ÿ Vo …ln R†

An ˆ …36†
l2n r 2 V21 …ln r † ÿ r2o V2o …ln ro †


With these values of An , the pore pressure is readily estimated using Eq. (31). The
estimations were undertaken in a spreadsheet program operating in Microsoft ExcelTM
The rate of consolidation may be expressed by the non-dimensional time factor [9],

T ˆ cv t=r2o …37†

The visco-elastic e€ect may be represented by the factor [35],

r2o w
Nˆ …38†
T2 G 1 k

A parametric study was undertaken using Eq. (31), as illustrated in Fig. 3 for the
dissipation of pore water pressure expressed as …uo ÿ u†=uo (uo ; u are the pore pressure
on the pile-soil interface). Fig. 3(a) illustrates the e€ect of viscosity,  by the factor, N on
the dissipation process. It may be noted that for N ˆ 0, and 1, the curves all behave as
Terzaghi's consolidation, except for the time factor being as a multiplier of the fac-
tor, T [35]. As the factor, N increases, the e€ect on the dissipation process will start
earlier. Fig. 3(b) shows that ratios of primary and secondary shear moduli, G 1 =G 2 ,
can only shift the dissipation curve of pore pressure at a later stage of the process.
The time factor, T for the dissipation of pore pressure, …uo ÿ u†=uo at 50 and 90% is
denoted as T50 ; T90 . It must be stressed that the degree of consolidation de®ned herein
based on dissipation of pore pressure is not equal to the degree based on settlement [24,
35], because it is a visco-elastic medium. Fig. 4 shows a set of plots of the factor, T50 and
T90 at di€erent values of uo …ro †=su and G 1 =G 2 …uo …ro † is the initial pore pressure on
pile-soil interface immediately following pile installation). Increase in the ratio of
G 1 =G 2 , (i.e. including the secondary consolidation), higher values of T50 and T90 are
expected, thus, longer consolidation times and higher displacements result, in com-
parison with those for elastic analysis …G 1 =G 2 ˆ 0†. The rules shown in Figs. 3 and 4
are applicable for both cases of constant total stress and plane strain deformation,
except that the corresponding coecient of consolidation, cv , should be adopted.

6. Visco-elastic behaviour

6.1. Parameters for the creep model

The magnitude of the relaxation time, T2 is provided in Table 1, based on the

relevant publications [36±38]. In particular, as reviewed previously by Guo [21], the
experiment by Lo [25] showed that
 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144 125

Fig. 3. In¯uence of creep parameters on the excess pore pressure.

(1) The rate factor, G 2 = 2 , is generally a constant for a given clay. For the clays
tested, it lies between 0.2 and 0.4 (10ÿ5sÿ1).
(2) The compressibility index ratio, G 1 =G 2 , is only in¯uenced by the soil water
content, and generally lies between 0.05 and 0.2, except for a soil of loose
structure.
(3) The individual values of G 1 ; G 2 and  2 , however, vary with load (stress)
level.

6.2. Back-estimation of the soil modulus and limiting shaft stress

Experimental results by Clark and Meyerhof [23] show that:

126 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

Fig. 4. Variation of times T50 ; T90 , with the ratio uo …ro †=su .

Table 1
Summary of relaxation factors for creep analysis

Authors Lo [25] Edil and Mochtar [36] Qian et al. [37] Ramalho OrtigaÄo
and Randolph [38]

G 2 = 2 …10ÿ5 sÿ1 † 0.2±0.4 0.5±2.67 1.71±3.29 0.36±0.664

Description Oedometer test Creep test on model piles Vacuum preloading Field pile test

(1) during a loading test, the change in pore water pressure along the shaft of a
pile is insigni®cant;
(2) the magnitude of the total and e€ective radial stress surrounding a pile is
primarily related to the stress changes brought about when the pile is driven
and during subsequent consolidation. Changes with time due to loading are
insigni®cant relative to the initial values.
 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144 127

Therefore, the ratio of G 1 =f1 , may generally be assumed as a constant during a

loading test, and it can be back-estimated by matching the measured load-settlement
curve with the theoretical solution [39, 40]. Soil stress±strain non-linearity has only
limited e€ect on such a back-analysis [21], as further demonstrated in the subsequent
example.

6.2.1. Example study

Recovery of the soil strength and modulus with reconsolidation has normally been
investigated by a series of loading tests performed at di€erent time intervals follow-
ing installation of the piles [13]. With each of the measured load-settlement curves,
the soil strength and modulus may be back-®gured by matching with a theoretical
solution. Thus the variations of the strength and the corresponding shear modulus
with time may be obtained, as illustrated later in this section. The theoretical solu-
tion in this investigation was based on the GASPILE program [22,39,40], which is a
load transfer based numerical program, operating in Microsoft ExcelTM, as detailed
in the Appendix. Analysis shows that only the initial shear modulus, G 1 needs to be
changed, in order to match a measured load-settlement curve with GASPILE ana-
lysis by selecting di€erent (1) Young's modulus of a pile, (2) ratio of creep moduli,
G 1 =G 2 , and even (3) the load transfer factor, j . As long as identical values for (1),
(2) and (3) were adopted for each pile analyses at di€erent times in the back-analy-
sis, the obtained variations of normalised values (e.g. modulus) versus time (e.g.
Table 8) are not a€ected by the selected values.

6.3. Test reported by Trenter and Burt (1981)

Load tests were performed on four driven open-ended piles in Indonesia, mainly
by maintained load procedure [13]. The basic pile properties are shown in Table 2;
Young's modulus of the pile is assumed to be 29,430 MPa (the e€ect of this
assumption is explored later). The undrained shear strength of the subsoil at the site
varies basically according to su =1.5 z (su , kPa; z, depth, m). The initial shear mod-
ulus, G 1 is taken as a multiple of the undrained shear strength, su . The ratio, G 1 =su
is back-analysed from the test data. No information about the creep properties is
available. However, based on previous publications [21,22,25], the shear modulus
ratio, G 1 =G 2 is taken as 0.15, and the rate factor, Gy2 = 2 is taken as, 0.510ÿ5
(sÿ1). The e€ect of taking these assumed values is justi®ed as minor as shown later.

Table 2
Parameters for the analysis of the tests by Trenter and Burt [13]
!
Pile No. Diameter (mm) Wall thickness (mm) Penetration (m) 1 ( o1 ˆ 0) 1 ( o1 ˆ 0:5)
b
2 400 12 24/30.3 4.5/4.73 5.2/5.4 1.0/2.0
3 400 12 53.5/54.5 5.31 6.0 1.0/2.0
4 400 12 43.3 4.4 5.08 1.0/2.0
128 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

In the estimation of the load transfer factor, 1 (Appendix), the slenderness ratio,
L=ro was limited to 180, and the accuracy of the estimation has been tested against
rigorous numerical solutions [21, 41]. In pthe
 case that a slenderness ratio exceeds a
critical pile slenderness ratio of about 3 l (l is the ratio of pile Young's modulus to
the soil shear modulus at pile tip level [42]), the ratio may be replaced with the cri-
tical pile slenderness ratio. However, this estimation of the critical pile slenderness
ratio is recursive. In the current analysis, the load transfer factor is simply estimated
using the real pile slenderness ratios, since the accuracy of the load transfer factor
becomes relatively unimportant. With the input parameters tabulated in Table 2,
using GASPILE analysis, the relevant average values are back-®gured and shown in
Tables 3, 4 and 5 for pile 4, 3 and 2, respectively. When the piles reach their ultimate
capacities, the measured load-settlement curves show that pile-head displacements
have reached 4 and 6.3% of the diameter for pile 4 and 3 respectively. Therefore,
according to Kuwabara [43], the value of G 1 =f1 , for pile 3 (Table 4) should be
lower than that for pile 4 (Table 3).
Various theoretical solutions using GASPILE were adopted to match the mea-
sured data, as illustrated in Fig. 5 (pile 2), Fig. 6 (pile 4) and Fig. 7 (pile 3) individu-
ally. A list of the abbreviations used in the ®gures is detailed in Table 6. For pile 4 at

Table 3
Parameters for analysis of pile 4 tested by Trenter and Burt [13]

G 1 G 2 G 1
Time (days) G 1 (MPa) f1 (kPa)
f 1  2 t G 2

1.7 6.11 20.18 303 12.96 0.15

10.5 7.64 25.28 302 12.96 0.15
20.5 8.43 27.12 311 12.96 0.15
32.5 8.43 27.5 306 12.96 0.15

Table 4
Parameters for analysis of pile 3 tested by Trenter and Burt [13]

G 1 G 1 G 2
Time (days) G 1 (MPa) f1 (kPa)
f1 G 2  2 t
2.3 3.62 20.717 175 0.15 12.96
3.0 3.69 21.13 175 0.15 12.96
4.2 3.9 22.308 175 0.15 12.96

Table 5
Parameters for analysis of pile 2 tested by Trenter and Burt [13]

G 1 G 1 G 2 ÿ1
Length (m) G 1 (MPa) f1 (kPa) …s †
f1 G 2  2
24.0 8.46 20.225 418 0.15 0.510ÿ5
30.3 7.75 23.865 308 0.15 0.510ÿ5
 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144 129

Fig. 5. Comparisons between the calculated results and the measurements of Trenter and Burt [13] for pile 2.
130 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

Fig. 6. Comparisons between the calculated results and the measurements of Trenter and Burt [13] for pile 4.
 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144 131

Fig. 7. Comparisons between the calculated results and the measurements of Trenter and Burt [13] for pile 3.
132 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

Table 6
Explanation of the abbreviations used in Figs. 5±7

Abbreviations Meanings

NLVE Non-linear visco-elastic analysis, using oj ˆ 0:5 for estimating the load transfer
factor, j , and relevant parameters in the corresponding tables. The ax is referred
to the time for creep (e.g. 2.5 h means a 2.5 h has been adopted in the estimation).
LVE Linear visco-elastic analysis. Every parameter is exactly the same as that used in
NLVE except choosing oj ˆ 0 for estimating 1 .
NLE Non-linear elastic analysis. Every parameter is exactly the same as that used in
NLVE except choosing G 1 =G 2 ˆ 0.
LE Linear elastic analysis. Every parameters is exactly the same as that used in NLVE
except choosing G 1 =G 2 ˆ 0 and oj ˆ 0 for estimating 1 .
Pb Calculated base load-settlement relationship
Mea Measured pile-head load-settlement relationship.

1.7 and 10.5 days, the following analyses were undertaken: non-linear elastic (NLE),
non-linear visco-elastic (NLVE), linear elastic (LE), and linear visco-elastic (LVE).
As illustrated in Fig. 5, the di€erence amongst these analyses (NLE, NLVE, LVE and
LE) are negligibly small, thus only non-linear visco-elastic (NLVE) and linear elastic
(LE) analyses are shown in the rest of the ®gures. Actually as shown previously [21, 22],
the maximum settlement under a given load is determined by the value of G 1 =G 2 ;
while the developing rate of the settlement is controlled by the factor, G 2 = 2 . It
may be seen from Fig. 5 that increase in time from 2.5 h to 1 month would not cause
signi®cant increase in the settlement, since the value of G 1 =G 2 =0.15 is rather low.
The shaft resistance was analysed in terms of total and e€ective stress using the
following expressions

f1 ˆ su …39†
0
f1 ˆ vo …40†

0
where vo is the e€ective overburden pressure; f1 , is the limiting shaft stress,  is the
average pile-soil adhesion factor in terms of total stress;  is the average pile-soil
adhesion factor in terms of e€ective stress. The corresponding parameters …; † for
piles 3 and 4 are tabulated in Table 7, as reported by Trenter and Burt [13]. Normal-
ising the parameters by those at 1.7 days, the normalised variation of the parameters
versus time is found quite consistent with that of the shear modulus (Table 8). More
generally, strength increases logarithmically with time [12, 44], but obviously the
increase should be limited. Otherwise, the strength becomes unrealistically high.
This example demonstrates that (1) the pile-soil interaction sti€ness increases
simultaneously as soil strength regains; (2) secondary compression of clay only
accounts for a small fraction of the settlement of the pile, when the value of G 1 =G 2
is low, and particularly given a short period of consolidation time.
 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144 133

Table 7
Parameters for empirical formulas (from Trenter and Burt [13])

Pile No. 4 3

Time (days) 1.7 10.5 20.5 32.5 2.3 3.0 4.2

a 0.63 0.81 0.87 0.87 0.51 0.53 0.55

b 0.16 0.20 0.22 0.22 0.13 0.13 0.14

Table 8
Comparison of the parameters for bearing capacity predictions

Pile No. 4 3

Time (days) 1.7 10.5 20.5 32.5 2.3 3.0 4.2

=o a 1.0 1.286 1.381 1.381 1.0 1.039 1.0784

=o a 1.0 1.25 1.375 1.375 1.0 1.0 1.077
Gi =Gio a 1.0 1.25 1.38 1.38 1.0 1.02 1.077
a
o ; o ; Gio are the values of ; ; Gi at 1.7 days.

7. Case study

Pore pressure, u was estimated using Eq. (31) for plane strain deformation for
each case study discussed below. The dissipated pressure, uo ÿ u, (uo initial value
given by Eq. (35) at r ˆ ro ) is then normalised by the initial value, uo and compared
with the following non-dimensional parameters:
(1) the di€erence of the measured (if available) pore pressure, uo ÿ u, normalised
by the initial pressure, uo ;
(2) the back-®gured shear modulus normalised by the modulus at t90;
(3) the back-®gured limiting shear strength normalised by the strength at t90 ;
(4) the measured time-dependent pile capacity normalised by the capacity at t90 .
The shear modulus and limiting strength with time were back-analysed in a similar
manner to that described in Section 6.2, using the measured load-settlement curves
at di€erent times following pile driving. Also the values at t90 were obtained through
linear interpolation.

7.1. Tests reported by Seed and Reese [5]

To assess the change in pile bearing capacity with reconsolidation of soil following
pile installation, Seed and Reese [5] performed instrumented pile loading tests at
intervals after driving. The pile, of radius 0.0762 m, was installed through a sleeve,
penetrating the silty clay from a depth of 2.75 to 7 m. The Young's modulus is
2.07105 MPa and the cross-sectional area is 9.032 cm2 . Therefore, the equivalent
pile modulus can be inferred as 10,250 MPa. Through ®tting the measured load-
134 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

settlement response by the analysis using the GASPILE program (Fig. 8), values
of G 1 ; f1 were back®gured from each load-settlement curves. These values are
tabulated in Table 9.
In terms of Poisson's ratio, s =0.49, permeability k=210ÿ6 m/s, cv =0.0529 m2/
day, the 90% degree of reconsolidation (elastic case) is estimated to occur at
t90 =8.76 days (T90 =74.43, G 1 =f1 =350). From Table 9, at the time of t90 , the
shear modulus is about 3.55 MPa, which is less than 90% of the maximum value of
4.5 MPa at 33 days. The shaft friction is estimated as about 11.6 kPa compared with
the ®nal pile-soil friction of 12.6 kPa, which is a fraction of the initial soil strength of
18 kPa, due to soil sensitivity. The back-®gured shear modulus and the limiting
strength were normalised by the values at t90 and plotted in Fig. 9(a) together with

Fig. 8. Comparison between the calculated and measured [5] load-settlement curves at di€erent time
intervals after driving.

Table 9
Back-®gured parameters from the measured by Seed and Reese [5]

Time (days) 0.125 1 3 7 14 33

f1 (kPa) 2.26 5.71 8.4 11.3 12.52 12.68

G 1 (MPa) 0.6 1.6 2.1 3.4 4 4.5
 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144 135

Fig. 9. The normalised measured time-dependent properties [5] compared with the theoretical decay of
excess pore pressure.
136 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

the normalised pile-capacity and elastic prediction of dissipation of pore water

pressure by Eq. (31).
Assuming G 1 =G 2 ˆ 1, with the parameters for elastic analysis, the visco-elastic
analysis leads to a new t90 =16.35 days (T90 =148.85, G 1 =f1 =350). Also from
Table 9, at this new t90 , the corresponding shear modulus is 4.06 MPa, which is
about 90% of the ®nal value, 4.5 MPa at 33 days. The shaft friction is 12.54 kPa.
The normalised back-®gured shear modulus and the limiting strength by values at
this new t90 , are shown in Fig. 9(b), together with the normalised pile-capacity (by
value at this new t90 ), and the viscoelastic prediction.
This case study shows that both elastic and visco-elastic analysis can well simulate
variation of shear strength or pile capacity. However, the di€erence in consolidation
time is signi®cant.

7.2. Tests reported by Konrad and Roy [45]

Konrad and Roy [45] reported the results of an instrumented pile, loaded to fail-
ure at intervals after driving. The pile, of outside radius 0.219 m, and wall thickness
8.0 mm, was jacked closed-ended to a depth of 7.6 m below ground level. The
Young's modulus was 2.07105 MPa and the cross sectional area was about 53.03
cm2. Therefore, the equivalent pile modulus is inferred as 29,663 MPa.
The increase in pile (shaft) capacity with consolidation was reported by Konrad
and Roy [45]. The normalised increase by that at 2 years after installation is shown
to be generally consistent with the normalised dissipation of pore pressure measured
at three depths of 3.05, 4.6, and 6.1 m, as illustrated in Fig. 10(a) and (b).
With the initial load-settlement response measured at di€erent time intervals after
pile installation, an initial value of G 1 =su =270 was back-analysed using elastic
analysis [21]. From the ®nal load-settlement relationships at di€erent time intervals,
the values of G 1 =su were found to be almost a constant of 210±230, with a value of
G 1 =G 2 =2 and using visco-elastic GASPILE analysis. Assuming shear modulus
and strength increase linearly with depth, the back-®gured values are tabulated in
Table 10.
The visco-elastic analysis gives a better agreement with the measured response at
low load levels of about 70% ultimate load (Fig. 11) than elastic analysis, although
the di€erence between the two analyses is not signi®cant (Fig. 12). The prediction
for high load levels may be improved, if the non-linear base response and the var-
iation of 1 , with load level is accounted for. However, the current analysis is su-
ciently accurate for assessing the interested values: the variation of shear strength
and modulus with reconsolidation.
The value of the coecient of consolidation was estimated as cv =0.0423 m2/day
(s =0.45) [45]. Using elastic analysis, the time factor for 90% degree of consolida-
tion, T90 , is about 65, with G 1 =su =230 from Fig. 4; hence, t90 18 days. From Table
10, at the time of t90 the shear strength at the pile base level is estimated to be 22.4
kPa, which is in good agreement with the value of 23.0 kPa obtained as 90% of
limiting stress, f1 (=25.61 kPa); while the corresponding modulus is 4.79 MPa,
which is slightly lower than 5.08MPa obtained as 90% of the maximum shear
 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144 137

Fig. 10. The normalised measured time-dependent properties [45] compared with the theoretical decay of
excess pore pressure.
138 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

modulus (=5.64 MPa). Using visco-elastic analysis, with G 1 =G 2 =2, T90 =205.1,
t90 is estimated to be about 57 days, the corresponding values are f1 =23.99 kPa and
G 1 =5.16 MPa by interpolation from the data given in Table 10. With these esti-
mated values at t90 , the normalised variations are plotted in Fig. 10(a) and (b)

Table 10
Back-®gured paramers from the measured by Konrad and Roy [45]

Time (days) 4 8 20 33 730

f1 (kPa) 5.58a/12.93b 6.56/19.49 7.75/23.0 8.06/23.93 8.63/25.61

G 1 (MPa) 1.07/3.19 1.44/4.29 1.65/4.9 1.73/5.15 1.9/5.64
a
Numerators for ground level.
b
Denominators for the pile base level.

Fig. 11. Comparison of load-settlement relationship predicted by visco-elastic GASPILE analysis with
the measured [45].
 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144 139

Fig. 12. Comparison of elastic and visco-elastic load-settlement relationships predicted by GASPILE.

respectively for elastic and visco-elastic analyses, together with the theoretical curves
of dissipation of pore water pressure.
The prediction by Eq. (31) at the initial stage are lower than the measured data,
probably due to the radial non-homogeneity of shear modulus [21] or soil stress±
strain non-linearity [46]. Radial non-homogeneity of modulus can reduce the ratio
of …u ÿ uo †=uo ; it can also retard the regain in the average shear modulus (at some
distance away from the pile axis). Thus, a comparatively higher new value of t90 for
the non-homogenous modulus regain is expected than the current predicted t90 . If
using values at the new t90 to normalise the rest mesured data or back-estimated
values, a lower trend of the curves is expected than the current presentation in the
®gure. Thus, a better comparison between those normalised values and the prediction
by Eq. (31) should result.
Soil strength may increase due to reconsolidation, but it may also decrease due to
creep (creep causes soil strength to reduce, until ®nally the strength approaches to a
140 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

long term strength, which is about 70% of the soil strength [21,47,48]). The e€ect of
reconsolidation and creep on the soil strength may o€set in this particular case. How-
ever, since the creep leads to an increase in settlement, only with a visco-elastic analysis,
(e.g. with a value of G 1 =G 2 ˆ 2 for the current example), can a good prediction be
made as shown in Fig. 11 for ®nal settlement and for the time-dependent process [21].

7.3. Comments on the current predictions

The back-®gured values of G 1 =G 2 for the two case studies are higher than those
reported by Lo [25]. The reason for this may be that the former are based on ®eld
tests, while the latter are based on con®ned compression (oedometer) tests. The
current radial consolidation theory is based on a homogeneous medium. However,
as just argued, radial non-homogeneity can alter the shape of the time-dependent
curve at the initial stage and increase the time for regain of shear modulus.

8. Conclusions

The research outlined here was attempted to o€er a prediction of the overall
response of a pile following driving, rather than just the pile capacity. A number of
important conclusions can be drawn:
(1) Visco-elastic solutions can be obtained by the available elastic solutions using
the correspondence principle.
(2) The viscosity of a soil can signi®cantly increase the consolidation time, hence
increase the pile-head settlement. However, it has negligible e€ect on soil
strength or pile capacity.
(3) Almost all the case studies show that the variation of the normalised pile-soil
interaction sti€ness (or soil shear modulus) due to reconsolidation is con-
sistent with that of the pore pressure dissipation on the pile-soil interface and
that of increase in soil strength. Therefore, the time-dependent properties
following pile installation can be suciently accurately predicted by the radial
consolidation theory. With the predicted time-dependent parameters, it is also
straightforward to obtain load-settlement response at any times following driving
by either GASPILE analysis or the previous closed form solutions [21,39,40].

Acknowledgements

The work reported here was undertaken during doctoral studies by the author at
the University of Western Australia, under Professor Mark F. Randolph. During the
period, the author was supported by an Australian Overseas Postgraduate Research
Scholarship and by scholarships from the University of Western Australia. This
®nancial assistance is gratefully acknowledged. The anonymous referees' comments
are also greatly acknowledged.
 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144 141

Appendix. GASPILE Ð a spreadsheet program

The GASPILE program was based on the load transfer approach [21, 22, 40]. The
approach simulates pile-soil interaction by a series of independent linear (or non-
linear visco-elastic) springs distributed down the pile shaft and at pile base. The
spring sti€ness is evaluated by the shaft and base models described herein.
The shaft model for the visco-elastic case (for step loading) is
 
 o ro G 1 t=T2
ws ˆ 1 ‡ 2 …1 ÿ e † …A:1†
G 1 G 2

where G j and T2 are as de®ned in the paper; ws is shaft displacement; o is shear

stress on pile-soil interface; The radial shear in¯uence, j may be expressed by
 
rm =ro ÿ oj
j ˆ ln …A:2†
1 ÿ oj

where j is a load transfer factor for spring j…j ˆ 1; 2†, in the current analysis, 1 is
simply taken as 2 . Normally 1 is gradually higher than 2 , as stress level, oj
increases, since o2 is gradually higher than o1 , if the failure stress, ult2 is lower
than ult1 [21,47,48]. oj ˆ …oj Rfj =ultj †, non-linear stress level on the pile-soil inter-
face for the spring j; Rfj is a parameter, which controls the degree of non-linearity
for the spring j; oj is shear stress on pile-soil interface at elastic spring j; ultj is ulti-
mate (soil) shear stress for the spring j: rm is the maximum radius of in¯uence of the
pile beyond which the shear stress becomes negligible, and may be expressed in
terms of the pile length, L, as [22,39,41]

1 ÿ vs
rm ˆ A L ‡ Bro …A:3†
1‡n

The values of parameter A and B are detailed previously [21,41]. For the current
research, A is generally taken as 2, and B as unity. n is the power of depth, z for
shear modulus distribution.
Using the Mechant's model, the conclusions described below have been directly
adopted in this program [21,22,40]:

(1) The limiting shaft stress, ultj and the failure stress, fj , on the pile-soil inter-
face are assumed identical.
(2) An appropriate value of ultl can be correlated with the shear strength of the
soil, or with the e€ective overburden stress [49±51]. The failure stress of f2 ,
may be assessed through experiments, and is generally correlated with the f1 ,
e.g. f2 was reported to be approximately equal to 0.7 f1 [47,48]. In the current
analysis, f1 is simply taken as f2 .
(3) With the model, two types of responses to stress are re¯ected: instantaneous
elasticity …G 1 † and delayed elasticity …G 2 †. At any given time, e.g. at the
142 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

onset of loading, the stress-strain response may be modelled as a non-linear

hyperbolic curve. Under a speci®ed stress, the displacement develops as a
creep process.
(4) With the model, secondary deformation is generally suciently accurately
modelled for all remoulded and undisturbed clays. However, as reported by
Lo [25], for a soil of loose structure, a generalised model [21] may be used.

The base settlement is estimated through a rigid punch acting on an elastic half-
space using a hyperbolic model for simulating the base load-settlement relationship
[22,40]

Pb …1 ÿ s †! 1
wb ˆ …A:4†
4ro Gb …t† …1 ÿ Rfb Pb =Pfb †2

where wb ; Pb are the mobilised base settlement and load, respectively; ! is the pile-
base shape and depth factor, which is taken as 1.0 [21, 40]; Pfb is the limiting base
load; Rfb is a parameter which controls the degree of non-linearity. The time-
dependent shear modulus may be estimated by

GL1 =b
Gb …t† ˆ …A:5†
1 ‡ GL1 =GL2 …1 ÿ et=T †

where b is the ratio of GL …t†=Gb …t†; GLj is the shear modulus just above the pile-tip
level for spring j. The ratio of GL1 over GL2 is taken as that of G 1 =G 2 .

References

[1] Orrje O, Broms B. E€ect of pile driving on soil properties.. J of Soil Mech and Found Engrg Div
1967;93(5):59±73.
[2] Flaate K. E€ects of pile driving in clay. Can Geotech J 1972;9(1):81±8.
[3] Fellenius BH, Samson L. Testing of drivability of concrete piles and disturbance to sensitive clay.
Can Geotech J 1976;13(2):139±60.
[4] Bozozuk M, Fellenius BH, Samson L. Soil disturbance from pile driving in sensitive clay. Can Geo-
tech J 1978;15(3):346±61.
[5] Seed HB, Reese LC. The action of soft clay along friction piles. Proceedings, ASCE 1955;122:731±54.
[6] Koizumi Y, Ito K. Field tests with regard to pile driving and bearing capacity of piled foundations.
Soils and Foundations 1967;7(3):30±53.
[7] Randolph MF, Wroth CP. An analytical solution for the consolidation around a driven pile. Int J
Numer and Analy Methods in Geomechanics 1979;3:217±29.
[8] Baligh MM. Strain path method. J of Geotech Engrg Div ASCE 1985;111(9):1108±36.
[9] Soderberg LG. Consolidation theory applied to foundation pile time e€ects. Geotechnique
1962;12:217±25.
[10] Eide O, Hutchinson JN, Landva A. Short and long term test loading of a friction pile in clay. Proc
5th ICSMFE, Paris 1961;2:45±53.
[11] Flaate K, Selnes P. Side friction of piles in clay. Proc of 9th ICSMFE, Tokyo, Japan 1977;1:517±22.
[12] Bergdahl U, Hult G. Load tests on friction piles in clay. Proc of the XICSMFE, Stockhom: Balkema
1981;2:625±30.
 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144 143

[13] Trenter NA, Burt NJ. Steel pipe piles in silty clay soils at Belavan, Indonesia. Proc of the XICSMFE,
Stockhom: Balkema 1981;3:873±80.
[14] Olson RE. Axial load capacity of steel pipe piles in clay. Proc. 22nd Annual OTC, Houston,
1990.
[15] Fleming WGK. A new method for single pile settlement prediction and analysis. Geotechnique
1992;42(3):411±25.
[16] Randolph MF. Design methods for pile groups and piled rafts. XIII ICSMFE, New Delhi, India
1994;5:61±82.
[17] Terzaghi KV. Theoretical soil mechanics. New York: John Wiley and Sons, 1943.
[18] Rendulic L. Porenzi€er und porenwasser druck in tonen. Der Bauingenieur 1936;17(51/53):559±64.
[19] Murray RT. Developments in two- and three-dimensional consolidation theory. In: Scott CR, editor.
Developments in soil mechanics. UK: Applied Science Publishers, 1978. p. 103±47.
[20] Biot MA. General theory of three-dimensional consolidation. J of Appl Phys 1941;12:155±64.
[21] Guo WD. Analytical and numerical analyses for pile foundations. Thesis presented to the University
of Western Australia for the degree of Doctor of Philosophy, 1997.
[22] Guo WD. Visco-elastic load transfer models for axially loaded piles. International J Num & Analy
methods in Geomechanics, in press.
[23] Clark JI, Meyerhof GG. The behaviour of piles driven clay. I. An investigation of soil stress and pore
water pressure as related to soil properties. Can Geotech J 1972;9(4):351±73.
[24] Booker JR, Small JC. Finite element analysis of primary and secondary consolidation. Int J of Solid
Structure 1977;13:137±49.
[25] Lo KY. Secondary compression of clays. J of Soil Mech and Found Engrg Div 1961;87(4):61±87.
[26] McKinley JD. Coupled consolidation of a solid, in®nite cylinder using a Terzaghi formulation.
Computers and Geotechnics 1998;23:193±204.
[27] Mase GE. Theory and problems of continuum mechanics. Schaum's Outline Series, McGram-Hill
Book Company, 1970.
[28] Booker JR, Poulos HG. Analysis of creep settlement of pile foundations. J of Geotech Engrg Div,
ASCE 1976;102(1):1±14.
[29] Mandel J. Consolidation des Couches D'argiles. Proc 4th ICSMFE, England 1957;1:360±7.
[30] Cryer CW. A comparison of the three-dimensional consolidation theories of Biot and Terzaghi.
Quart J Mech and Math 1963;16:401±12.
[31] Davis EH, Poulos HG. Rate of settlement under two- and three-dimensional conditions. Geo-
technique 1972;22(1):95±114.
[32] Christian JT, Boehmer JW. Plane strain consolidation by ®nite elements. J of Soil Mech and Found
Engrg Div 1972;96(4):1435±57.
[33] Barron R. Consolidation of ®ne grained soils by drain wells. Trans Amer Soc Civ Engrs
1948;113:718±42.
[34] McLachlan NW. Bessel functions for engineers. New York: Oxford University Press, 1955.
[35] Christie, IF. A re-appraisal of Mechant's contribution to the theory of consolidation, Geotechnique,
1964; 14(2), 309±20
[36] Edil TB, Mochtar IB. Creep response of model pile in clay. J of Geotech Engrg Div, ASCE,
1988;114(11):1245±59.
[37] Qian JH, Zhao WB, Cheung YK, Lee PKK. The theory and practice of vacuum preloading. Com-
puters and Geotechnics 1992;13(2):103±18.
[38] Ramalho OrtigaÄo JA, Randolph MF. Creep e€ects on tension piles for the design of buoyant o€-
shore structures. Int Sympo on O€shore Engrg 12/16, September, 1983.
[39] Guo WD, Randolph MF. Vertically loaded piles in nonhomogeneous media. Int J Num & Analy
methods in Geornechanics 1997;21(8):507±32.
[40] Guo WD, Randolph MF. Non-linear visco-elastic analysis of piles though spreadsheet program. 9th
Int. Conf of the Association for Computer Methods and Advances in Geomechanics. IACMAG 97,
Wuhan, China 1997; 3: 2105±2110.
[41] Guo WD, Randolph MF. Rationality of load transfer approach for pile analysis. Computers and
Geotechnics 1998;23(1±2):85±112.
144 Wei Dong Guo / Computers and Geotechnics 26 (2000) 113±144

[42] Fleming WGK, Weltman AJ, Randolph MF, Elson WK. Piling engineering (2nd ed.). Glasgow;
Halst Press, New York: Surrey University Press, 1992.
[43] Kuwabara F. Settlement behaviour of nonlinear soil around single piles subjected to vertical loads.
Soils and Foundations, JSSMFE 1991;31(1):39±46.
[44] Sen GS, Zhen DT. Soft soil foundation and underground engineering. Beijing; China Construction
Industry Press House, 1984. pp. 516±521, in Chinese.
[45] Konrad J-M, Roy M. Bearing capacity of friction piles in marine clay. Geotechnique 1987;37(2):163±75.
[46] Davis EH, Raymond GP. A non-linear theory of consolidation. Geotechnique 1965;15(2):161±73.
[47] Geuze ECWA, Tan TK. The mechanical behaviour of clays. Proc. 2nd Int. Congress on Rheology,
1953.
[48] Murayama S, Shibata T. Rheological properties of clays. Proc. of 5th ICSMFE, 6, Paris, France
1961, 269±73.
[49] API RP2A: Recommended practice for planning, designing and constructing ®xed o€shore plat-
forms. Washington, DC: American Petroleum Institute (API), 1987.
[50] Tomlinson MJ. Some e€ects of pile driving on skin friction. Behaviour of piles. London: The Insti-
tution of Civil Engineers, 1970: 107±14.
[51] Randolph MR, Murphy BS. Shaft capacity of driven piles in clay. 17th Annual OTC, Houston,
Texas, May 6±9, Paper OTC 4883, 1985, pp. 371±378.