Pile Behavior—Consequences of Geological

and Construction Imperfections1

Harry G. Poulos, F.ASCE2

Abstract: This paper identifies some of the causes of imperfections in real pile foundations and then examines their possible effects on
pile behavior. The imperfections considered arise from two main sources, natural imperfections caused by the geological circumstances
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

and imperfections related to the construction of the piles. In each case, single piles are considered first, and then pile groups are addressed
separately. Case histories are presented to illustrate the effects on pile performance which are predicted by theoretical analyses. The
analyses with which these examinations are carried out will still inevitably involve some degree of idealization, but at least attempt to
incorporate the key “nonideal” aspects. The paper also discusses the particular problems involved in investigating and analyzing under-
performing pile foundations and reviews some methods by which the consequences of imperfections may be overcome or ameliorated.
DOI: 10.1061/共ASCE兲1090-0241共2005兲131:5共538兲
CE Database subject headings: Pile foundations; Foundation settlement; Construction; Geologic processes.

Introduction assumptions being invalid, at least to some degree. It is important

to understand how nonideal conditions can arise and what effects
In his Presidential Address to the 1st International Conference on they may have on pile behavior. The term “imperfection” will be
Soil Mechanics and Foundation Engineering, Terzaghi 共1936兲 used to describe a nonideal condition in which either the ground,
stated: “…the earth in its natural state is never uniform…Its or the pile as-installed, has a defect or flaw, as compared to the
properties are too complicated for rigorous theoretical ideal situation. It is also essential to be able to cope with the
treatment…Even an approximate mathematical solution of some consequences of imperfections when construction has been com-
menced, or after it is completed, and the performance of the foun-
of the most common problems is extremely difficult.” The sub-
dation system has not met requirements. In some circumstances,
stantial advances in numerical analysis made since that time have,
remedial works may be necessary to correct the foundation sys-
however, allowed theoretical treatment and solution of a wide
tem so that it remains both serviceable and safe.
range of problems to be obtained, and these advances have been
This paper will attempt to identify some of the causes of im-
particularly significant for pile foundations. perfections in real pile foundations, and will then examine their
Despite the availability of these methods to modern-day pile possible effects on pile behavior. In each case, single piles will be
analysts and designers, it is still very common to assume that considered first, and then pile groups will be addressed separately.
“ideal” circumstances prevail and that the earth is essentially uni- The analyses with which these examinations will be carried out
form. For example, it is usual to assume that all soil layers are will still inevitably involve some degree of idealization, but will
horizontal, the founding conditions below all piles are identical, at least attempt to incorporate the key “nonideal” aspects. The
and that all piles in a group are flawless, elastic, and of equal paper will also discuss the particular problems involved in ana-
dimensions and length. Indeed, some methods also make more lyzing and designing underperforming pile foundations, and will
radical assumptions with respect to ideal conditions, namely, that review some methods by which the consequences of imperfec-
the soil behavior is elastic, and that the soil profile is either uni- tions may be overcome or ameliorated.
form with depth or varies in a linear manner. The effects of con-
struction processes on the piles are also frequently ignored or
simplified. Such ideal conditions are rarely, if ever, encountered Process of Idealization
in real life. Both natural geological circumstances and the pro-
cesses involved in construction will generally lead to the above In any geotechnical analysis or design, it is necessary to simplify
the problem in hand and develop an idealized model that incor-
1
Presented at the GeoSupport 2004 Conference, Orlando, FL. porates the key geotechnical and structural features of the prob-
2
Senior Principal, Coffey Geosciences Sydney Australia & Emeritus lem. In particular, there are three aspects of idealization that may
Professor of Civil Engineering, Univ. of Sydney, 8/12 Mars Rd. Ln., be critical:
Cove West, NSW 2066, Australia. E-mail: [email protected] • The representation of the geology of the site;
Note. Discussion open until October 1, 2005. Separate discussions • The assignment of the relevant geotechnical and structural pa-
must be submitted for individual papers. To extend the closing date by
rameters; and
one month, a written request must be filed with the ASCE Managing
Editor. The manuscript for this paper was submitted for review and pos- • The consideration, as far as possible, of the “realities” of the
sible publication on July 7, 2004; approved on November 1, 2004. This problem. Depending on the nature of the problem, it may be
paper is part of the Journal of Geotechnical and Geoenvironmental essential to incorporate such factors as nonlinear pile-soil be-
Engineering, Vol. 131, No. 5, May 1, 2005. ©ASCE, ISSN 1090-0241/ havior, nonlinear pile material behavior, the presence of a
2005/5-538–563/$25.00. water table, layering of the soil profile, loading by structural

538 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 loads, and loading by ground movements.
In all cases, these latter realities should be recognized prior to the
analysis and allowed for. The assessment of appropriate geotech-
nical parameters is a major topic of research, and is often the key
ingredient in the success or failure of a design or analysis. Despite
its importance, this issue will not be the focus of this paper.
In many problems, there may exist “imperfections” that arise
because of natural geological or construction-related reasons, and
which are generally unanticipated deviations from the expected
circumstances. While it may be possible to handle these imper-
fections by appropriate analysis, the main difficulties are that the
designer/analyst may not recognize the existence of the imperfec-
tions during the design process, or else they may only become
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

manifest beyond the design process. Subsequent analysis may

then become either a remedial or a forensic exercise.
In this paper, it will be assumed that the “realities” of pile
analysis will have been recognized and will be incorporated into
the analyses as necessary, and that appropriate geotechnical pa-
rameters have been assessed for the problem. Hence attention will
be focused on the influence of “imperfections” which may not be
recognized at the time of design or construction, but which may
have an important influence on subsequent behavior of the foun-
dation system.

Sources of Imperfection
Fig. 1. Examples of natural 共geological兲 imperfections
The imperfections that may impact on pile foundation perfor-
mance may arise from a number of sources, including natural
sources, inadequate ground investigation, construction, pile load • A soft toe on bored piles due to inadequate base cleaning
testing, and loading during operation. Only the first three sources 共avoidable兲;
will be considered in this paper. However, the influence of other • Defects within the shaft of bored piles 共avoidable兲;
sources of imperfections has been studied by Poulos 共2000兲 in • Inadequate founding conditions 共avoidable兲;
relation to the effects of the testing setup on pile load-settlement • Ground movements developed due to drilling during the con-
performance, while the effects of loading arising from construc- struction process 共generally unavoidable兲;
tion operations during tunneling operations have been explored • Excavation and dewatering effects, especially with remedial
by Chen et al. 共1999兲; and Mroueh and Shahrour 共2002兲. piling projects 共generally unavoidable, but controllable兲; and
• Excessive driving of preformed piles 共avoidable兲.
Natural Geological Sources In general, construction-related imperfections in piles can be
broadly classified into two main categories, structural defects and
These imperfections arise from the geological processes at the
site, which have resulted in the present ground profile. They may
include 共among many others兲 layers which are not horizontal or
continuous, boulders within a soil layer, sloping bedrock, intru-
sions of rock over limited areas of the site, cavities in limestone
rock, or the presence of softer layers below what might be re-
garded as suitable founding strata for the piles. Fig. 1 illustrates
some of these situations.

Inadequate Ground Investigation

These imperfections are generally related to those that arise from
natural sources, but are exacerbated because the site is not prop-
erly characterized. Inadequacies are usually related to an insuffi-
cient number or depth of boreholes or probes to identify strati-
graphic variations across the site, or inadequate testing to quantify
the relevant geotechnical parameters. Fig. 2 illustrates a typical
example of such imperfections, and the inaccurate geotechnical
model that can result from them.

Construction
These imperfections arise from processes related to construction
of the piles, either from inadequate construction control or from
inevitable consequences of construction activities. They may in- Fig. 2. Examples of imperfections related to inadequate ground
clude: investigations

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005 / 539

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 Analyses Incorporating Imperfections

It is important to recognize the requirements for analyses of piles

which can take account of the imperfections inherent in practical
pile foundations. Such analyses should have the ability to handle
the following aspects:
1. Soil profiles involving nonhorizontal layers;
2. Different soil profiles along different piles within the same
group;
3. Piles of different length or diameter within the same group,
including consideration of interaction among dissimilar piles;
4. Piles containing structural defects or changes in diameter or
size along the length;
5. Piles that are added 共or activated兲 part-way through the load-
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

ing process 共to simulate the installation of remedial piles兲;

Fig. 3. Example of imperfections relation to construction techniques and
6. Loadings imposed from ground movements, as well as from
normal structural loadings.
In addition to these capabilities, the analyses should also be able
to cope with the realities of pile behavior mentioned previously,
geotechnical defects. Structural defects can result in the size, in particular, nonlinear pile-soil response, and also nonlinear pile
strength, and/or stiffness of the pile being less than assumed in material behavior.
design. Such defects have been discussed extensively in the lit- In principle, powerful contemporary numerical analyses, for
erature, particularly with respect to cast-in situ concrete piles, for example, FLAC 共Itasca 1999兲 and PLAXIS 共Brinkgreve and Ver-
example, Hobbs 共1957兲; Thorburn and Thorburn 共1977兲; and meer 1998兲, can take many, if not all, of the above factors into
Reese 共1978兲. Brown 共2004兲 discusses some aspects of construc- account. However, there are often limitations involved in using
tion techniques and materials that can lead to defects or subopti- such programs, related to the time involved in setting up the
mal pile performance. Examples of structural defects include the analyses of large pile groups, and the difficulty of discerning
following: “necking” of the shaft of bored piles, leading to a broad patterns of behavior from parametric studies. There are also
reduced cross-sectional area along part of the pile, poor quality potential problems with the accurate modeling of piles, particu-
control during the construction of bored piles, leading to some larly if a two-dimensional analysis is to be used and details of the
parts of the shaft having lower strength than assumed in design, internal forces and bending moments within the piles are re-
tensile cracking of large diameter bored piles under the influence quired. The writer believes that it may be preferable, at least from
of thermal strains, damage during the driving of timber, precast a practical viewpoint, to employ analyses which may involve a
concrete, or steel piles, leading to reduced strength and stiffness greater degree of approximation, but may be simpler to use and
of parts of the pile, especially near the top or tip of the pile, and more readily adapted to carrying out parametric studies, as well as
bending of slender driven piles. solving real-world problems. It is preferable to consider an im-
Geotechnical defects usually arise from either a misassessment perfection approximately, rather than to ignore it because it can-
of the in situ conditions during design, or else from construction- not be treated rigorously.
related problems, and may include reduced shaft friction and end Table 1 summarizes the capabilities of some of the computer
bearing resistance arising from localized softer or weaker geo- programs whose basis has been published 共although not all are
technical conditions in the vicinity of one or more of the piles in commercially available兲. The list in Table 1 is not meant to be
the group, reduced skin friction, and end bearing resistance aris- exhaustive, but rather, representative, and indicates that none of
ing from construction operations such as the use of bentonite the programs listed is able to incorporate all of the realities and
without due caution, and a “soft base” arising from inadequate imperfections that might arise. Indeed, the only aspect that is
cleaning of the base of bored piles. The latter is one of the most considered by all of the programs listed is axial behavior. How-
common concerns in bored pile construction and is likely to lead ever, a number are capable of considering at least some of the
to a reduction in the stiffness of the soil below the base of the realities and imperfections, albeit for a restricted range of prob-
pile. The ultimate base bearing capacity may not be affected sig- lems. In the following sections, three of these programs,
nificantly, but may require a large movement to be fully mobi- DAMPIG, GEPAN, and PIGS, will be employed to examine some
lized. Several writers have investigated the influence of construc- of the characteristics of pile group behavior when imperfections
tion defects on the shaft friction and end bearing resistance of are present. A more detailed outline of the basis of the PIGS
piles, for example, O’Neill and Hassan 共1994兲; O’Neill 共2001兲; program is given by Poulos 共2002兲.
and Abdrabbo and Abouseeda 共2002兲. O’Neill and Hassan sug-
gest bounds for the effects of construction-related factors and
present a framework for quantifying these parameters for design Imperfections Considered
purposes.
Fig. 3 illustrates an example of the usual idealization of a pile The effects on pile behavior of some of the imperfections arising
which may have both structural and geotechnical defects. In this from nature, inadequate investigation, or inadequate construction
paper, both types of defect will be considered, and it will be procedures will be examined in the following sections. For sim-
demonstrated that the type of defect may play a significant role in plicity, many of the cases examined will be idealized, and atten-
determining the subsequent behavior of that pile and of the pile tion will be focused on load-settlement behavior under axial load-
group containing it. ing. The following cases will be considered:

540 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 Table 1. Summary of Capabilities of Some Pile Group Analysis Programs
PILGP1 PIGLET FLPIER PIES DAMPIG GEPAN EMPIG PIGS
共O’Neill et al. 共Randolph 共Hoit and McVay 共Poulos 共Poulos 共Xu and Poulos 共Poulos 共Poulos
Factor 1982兲 2003兲 1996兲 1989兲 1997兲 2001兲 1999a兲 2002兲
Axial response 冑 冑 冑 冑 冑 冑 冑 冑
Foundation rotation 冑 冑 冑 — 冑 冑 冑 冑
Lateral response 冑 冑 冑 — 冑 冑 冑 —
Ground movements — — — 冑 — 冑 冑 冑
Nonlinear pile-soil 冑 — 冑 冑 冑 — 冑 冑
Nonlinear pile — — ? — 冑 — — 冑
Different pile length, etc. 冑 — — — — 冑 — 冑
Pile defects — — 冑 冑 i 冑 — i
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

Nonhorizontal soil layers — — — — — 冑 — i

Varying soil profiles — — — — — — — 冑
Piles added or removed — — — — — — — 冑
during loading process
Note: 冑⫽capability exists; i⫽factor can be incorporated, but indirectly; and—⫽capability does not exist.

1. Piles bearing on a rock stratum with compressible clay seams on the stiffness and ultimate axial capacity of the pile base;
共natural geological imperfections兲; and
2. Piles founded in a layer underlain by a compressible deposit 2. A pile-soil interaction analysis in which the base load-
共natural imperfections, and possibly imperfections from in- settlement characteristics are input to obtain the pile head
adequate investigations兲; load-settlement behavior.
3. A pile founded on a boulder within a soil layer 共natural geo- Fig. 4 shows an idealized case of a single seam below a pile base
logical imperfection兲; that has been analyzed by this approach. It is assumed here that
4. Piles founded in soil profiles of varying thickness 共natural the clay seam has a strength and Young’s modulus 1 / 100 times
geological imperfections兲; that of the rock into which the pile is founded. The finite element
5. Groups containing piles of different length 共imperfections program PLAXIS has been used to compute the reduction in base
from natural, investigation, and/or construction causes兲; capacity and stiffness as a function of the seam thickness and its
6. Piles with structural defects 共construction imperfections兲; depth below the pile, with the results being expressed in terms of
and base capacity and stiffness reduction factors, BCR and BSR, respec-
7. Piles with geotechnical defects 共construction imperfections兲, tively, where
in this case, piles with a “soft toe.”
In each case 共except numbers 4 and 5 for obvious reasons兲, the BCR = ultimate capacity of base with seam/
effects on single pile response will be considered, and then the ultimate capacity of base without seams 共1兲
case of pile groups will be examined.
BSR = stiffness of base with seam/stiffness of base without seam
共2兲
Effects of Thin Seams in Bedrock
The PLAXIS analyses indicate that, as might be expected, the
presence of the clay seam alters the failure mode of the pile base
Single Piles
An issue of considerable concern to pile foundation designers is
the presence of weak seams in the founding layer, below the pile
tip. Kulhawy 共1978兲 has given some consideration to the effects
of geological imperfections on rock mass stiffness and established
a relationship between the equivalent rock mass properties and
the individual rock mass properties of the rock material and the
discontinuities. The influence of discontinuities on bearing capac-
ity has been considered by Kulhawy and Carter 共1992兲. In many
cases, specifications for the construction of bored piles require
probing some distance below the founding level to avoid the pos-
sibility that discontinuities and clay seams underlie the pile base.
Nevertheless, the possibility remains that such seams may remain
undetected, and in the case of driven or jacked piles, the possi-
bilities may be even greater. The effects of such weak seams can
be investigated via numerical analysis procedures such as the fi-
nite element method. However, a more efficient approach is to
carry out a two-stage analysis process:
1. A finite-element analysis to assess the effect of a clay seam Fig. 4. Definition of clay seam imperfection

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005 / 541

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 and shale, where the rock grading depends on the proportion of
clay seam within the depth of influence of the pile base. Fig. 7
compares the reduction in capacity with increasing seam thick-
ness, for two different depths of the seam below the base, and for
sandstone and shale. It can be seen that the theoretical results for
a clay seam 0.2 base diameters below the base agree reasonably
well with the design recommendations from both the above-
mentioned sources. However, the design recommendations appear
to be unduly conservative if the clay seam is well below the pile
base, for example, more than about 0.4 base diameters. Similar
comments can be made in relation to the recommendations of the
Canadian Foundation Engineering Manual, where the thickness of
the clay seam is quantified explicitly.
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

From a design viewpoint, the result that is most critical is the

effect of the clay seam on the overall load-settlement behavior of
a pile. Using the computer program PIES, together with the re-
Fig. 5. Example of effect of clay seam on pile base load-settlement sults in Fig. 6, analyses have been carried out for a range of
behavior values of pile length-to-diameter, to assess the overall capacity
and head stiffness reduction due to the presence of a clay seam.
The results are shown in Fig. 8 for a single 1 m diameter pile
significantly. An examination of the displacement fields near fail- situated in a weak rock and founded on a stronger and stiffer rock
ure indicates that the seam provides a path for the failure surface containing a clay seam. Fig. 8共a兲 plots the ratio of ultimate pile
and can lead to a significant reduction in base capacity, as shown capacities with and without the clay seam below the base, while
in Fig. 5 which plots typical load-settlement curves for the pile tip Fig. 8共b兲 shows the corresponding plot for the initial pile head
with and without a clay seam. Fig. 6 shows the reduction factors stiffness. It can be seen that, as the pile length increases, the effect
BSR and BCR as functions of relative depth of the seam and as a of the seam becomes less than for the base alone, since a consid-
function of the seam thickness. It is clear that BCR reduces as the erable amount of load is carried by the pile shaft. For example,
seam thickness increases and as the distance of the seam from the even if a seam with a thickness of 0.3 times the pile diameter
pile base decreases. The effect is particularly dramatic if the clay exists 0.2 diameters below the pile tip, the reduction in pile head
seam is just below the pile base. The reduction in base stiffness is stiffness is only about 30% when the pile length is 20 diameters,
even more dramatic than the reduction in capacity. while the loss of capacity is less than 20%. As might be expected,
It is interesting to compare the theoretical results in Fig. 6 with the greater the proportion of load carried by the pile shaft, the
typical design requirements, for example, those recommended by smaller is the effect of the weak seam on the pile head stiffness
Pells et al. 共1978, 1998兲 for piles socketted into Sydney sandstone and ultimate capacity.

Fig. 6. Influence of clay seam on pile base stiffness and capacity

542 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

Fig. 7. Comparison between computed and design reduction factors 关adapted from Pells et al. 共1998兲兴

Fig. 8. Example of influence of clay seam on pile capacity and stiffness

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005 / 543

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 Pile Groups
It is important to note that a pile group may well be affected more
severely than a single pile by a clay seam. This is because, in
many cases, a group can be represented by an equivalent single
pier containing the piles and the intervening soil 共van Impe 1991;
Poulos 1993兲. In the case of a group, the diameter of the base of
the group is of course larger than that of a single pile, and more of
the applied load is carried by the base of the piles than is the case
for a single pile. Because of the greater equivalent diameter of the
group, the relative depth of a clay seam will be less than is the
case for a single pile, and consequently, the reduction in capacity
and stiffness of the base of the group will be greater than for a
single pile.
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

From the examples examined herein, it can be concluded that

the presence of weak seams within the founding layer can be of Fig. 9. Pile base founded in dense sand overlying clay 关after Matsui
共1993兲兴
considerable concern in relation to both the base load capacity
and the base stiffness. However, if the pile carries a significant
part of the load in skin friction, the decrease in pile head stiffness db⫽diameter of pile base. It is also possible to use elastic theory
and load capacity due to the seams is less dramatic, at least until to assess the effects of the underlying layer on the axial stiffness
the shaft capacity has been fully mobilized. When most or all of of the pile base. A series of such solutions is shown in Fig. 10, in
the load is carried by the pile base, the reduction in stiffness can terms of a base stiffness reduction factor, defined as the ratio of
be significant if the seam is relatively thick and located very near the base stiffness with the underlying layer to the base stiffness if
the base of the pile or pile group. Thus, as per the conventional there is no underlying layer. The Young’s moduli of the upper and
wisdom of good site investigation, it is essential to identify and lower layers are E1 and E2, respectively, and Poisson’s ratio for
take account of thin weak seams within the founding layers, pref- each layer is assumed to be 0.3. It can be seen that, as would be
erably via drilling to and beyond the depth of influence of the pile expected, the stiffness of the base decreases as the modulus E2 of
or group prior to determining the final founding level. the underlying layer decreases and as the distance from the pile
base to the underlying layer decreases.
The results in Fig. 10 and Eq. 共3兲 can be used to estimate the
Piles in Layer Underlain by Compressible Layer effects of the compressible underlying layer on the behavior of a
single pile. It is found that, as with the case of a clay seam, the
Single Piles overall load-settlement behavior of the pile becomes less sensitive
to the presence of the underlying layer as the proportion of load
One of the most insidious forms of imperfection arising from carried by the shaft increases. Thus, for relatively long slender
natural causes 共and which are not identified by the ground inves- piles, the reduction in ultimate capacity and stiffness may be rela-
tigation兲 is the presence of compressible layers below the pile tip tively small.
founding level. Such layers can have a significant effect on both
the ultimate capacity and stiffness of the pile, especially of the
pile base. Many methods of estimating the bearing capacity of Pile Groups
circular foundations on layered soils have been proposed 关for ex- The results for a single pile may be used to obtain an indication of
ample, Button 共1953兲; Meyerhof 共1976兲; Meyerhof and Sastry the importance of an underlying compressible layer on the group
共1978兲; Giroud et al. 共1973兲; Merifield et al. 共1999兲兴, and such load-settlement behavior. Poulos 共1993兲 and Randolph 共1994兲
solutions have been applied to the estimation of the bearing ca-
pacity of a pile base. A very useful study by Matsui 共1993兲 has
suggested a simplified approach to this problem of a pile tip
founded in a sand layer underlain by a clay layer, as shown in Fig.
9. This work indicates that the depth of influence of the clay layer
below the pile tip is less than had been previously suggested by
Meyerhof and Sastry 共1978兲, who indicated that the compressible
layer should be at least 6 pile base diameters below the pile base
before its effect could be neglected. Matsui’s results may be sum-
marized as follows:
fb = f2 共zc/db 艋 0.5兲 共3a兲

fb = f1 共zc/db 艌 3兲 共3b兲

f b = f 2 + 共0.4zc/db − 0.20兲 · 共f 1 − f 2兲 共3.0 艌 zc/db 艌 0.5兲

共3c兲
where f b⫽ultimate base bearing capacity; f 1⫽ultimate bearing
capacity of upper layer; f 2⫽ultimate bearing capacity of underly- Fig. 10. Stiffness reduction factor for pile base with compressible
ing layer; zc⫽depth of underlying layer below pile base; and underlying layer

544 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 relatively high level, well above a deep layer of compressible
Leda Clay. Fig. 12 shows the stratigraphy of the site and some of
the key engineering properties revealed by the investigations. A
plan of the configuration of the pile group is also shown in Fig.
12. A number of the original boreholes extended to depths up to
236 ft 共72 m兲 without encountering bedrock.
A load test was carried out on a pile similar to that used for the
furnace foundations. At a typical working load of 75 US t 共668
kN兲, the measured settlement was about 0.04 in. 共1.0 mm兲. Ap-
plying normal pile group settlement theory to this result, it was
expected that the settlement of a 32-pile furnace group would be
of the order of 3–6 mm.
Fifteen years of settlement records were available for a bank of
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

five furnaces, and these measurements enabled some interesting

conclusions to be drawn regarding the sources of settlement of the
foundations. Fig. 13 reproduces the measured settlements over the
Fig. 11. Effect of compressible layer on group settlement bank of five furnaces, and reveals the following characteristics:
• The maximum settlement nearly 15 years after construction
was about 73 mm and was continuing to increase;
have shown that the behavior of pile groups can often be approxi- • The measured settlements were an order of magnitude greater
mated well by considering an equivalent pier of equal length and
than those which may have been expected simply on the basis
a diameter which gives a similar perimeter to that of the group.
of the pile load test; and
Using this approximation, it is clear that a compressible underly-
• The settlement of the end furnaces 共Furnace 1 and Furnace 5兲
ing layer will affect a pile group more than it affects a single pile.
was clearly affected by the other furnaces, and showed a sig-
The equivalent diameter of the group is obviously larger, and
more load is transferred to the base of the piles. Thus the relative nificant tilt.
depth of the underlying layer to the base diameter of the equiva- It was estimated by Golder and Osler that, taking into account the
lent pier is smaller, resulting in a greater reduction in base stiff- settlement of the compressible layers below the pile tips, the an-
ness and base capacity. Poulos et al. 共2002兲 have given an ex- ticipated final settlement of the end furnace 共No. 1兲 could be of
ample which shows that the increased settlement due to an the order of 87 mm, consisting of 10 mm of pile group settlement,
underlying layer increases with increasing group size. 13 mm consolidation of the silty clay layer below the pile tips,
It is also possible to examine the effect of a compressible and 64 mm from the deep Leda clay.
underlying layer via a pile group analysis, provided that due care This case clearly demonstrates the importance of taking ac-
is taken with the estimation of the interaction factors to make count of the compressibility of underlying compressible layers
some allowance for the effects of intervening piles between the below the pile tips, and of also considering the interaction among
influencing and influenced piles. The approximate approach sug- adjacent foundations. It also highlights the potential dangers of
gested by Poulos 共1988兲 has been used here, with the soil between relying solely on the results of a single pile load test to predict
the piles assumed to be three times stiffer than the soil near the pile group behavior, without a proper appreciation and character-
pile. For pile groups in a square configuration, with a center-to- ization of the ground conditions.
center spacing of 3 diameters, Fig. 11 shows the computed settle-
ment as a function of the number of piles and the Young’s modu-
lus of the underlying layer. The following observations may be
made:
• The effect of the underlying layer becomes more significant as
the size of the group increases;
• For a 122 group, the settlement may increase by a factor of
about 3.5 when the underlying layer has a modulus of 20 MPa,
compared to the case of a homogeneous founding layer with a
modulus of 100 MPa; and
• If the soil beneath the founding layer is very stiff 共1,000 MPa
in this case兲, the settlement may in be reduced by a factor of 2
compared to the case of a founding layer having a uniform
modulus of 100 MPa.
Fig. 11 thus indicates that the potential exists both to seriously
overestimate the settlement of pile groups if account is not taken
of the presence of stiffer layers at depth, and to seriously under-
estimate the settlement if underlying compressible layers are not
identified or are ignored.

Case History
Golder and Osler 共1968兲 have described an interesting case of a Fig. 12. Stratigraphy and pile group layout for furnace foundation
series of furnace foundations on piles, which were founded at a 关adapted from Golder and Osler 共1968兲兴

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005 / 545

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 1. Jacking of the pile to its final penetration, considering the
boulder to be an enlarged base at the tip of the pile;
2. Removal of the jacking force, thus unloading of the pile; and
3. Reloading of the pile, simulating a static load test on the pile.
In the PIES analysis, the connection between the enlarged base
共boulder兲 was assumed to have very small tensile resistance on
unloading.
Fig. 14 shows the computed load-settlement behavior for the
three stages above. The following points can be noted:
1. On initial jacking, a relatively large resistance is developed
because of the presence of the boulder;
2. On unloading, the pile tip separates from the boulder, and a
gap forms; and
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

3. On reloading in the static load test, the presence of the gap

formed during unloading leads to a reduced resistance until
the pile tip recontacts the boulder. Thus the pile appears to
have a reduced load capacity.

Pile Groups
If such a pile is part of a pile group, it will have a lower axial
stiffness and 共effectively兲 a lower axial capacity than other piles
which are properly founded on a hard stratum. Consequently, the
group will experience an uneven settlement as well as a lateral
displacement. The behavior of groups containing piles of different
lengths or stiffness is considered below.

Groups Containing Piles of Unequal Length

It is very common for the piles within a group to have different

lengths. Generally, the length differences are not great, and reflect
Fig. 13. Settlement observations along north-south line through
the variable nature of the founding stratum. However, there are
furnaces 1–5 关based on Golder and Osler 共1968兲兴
some circumstances in which significant length differences can
occur, in particular, if the group is founded on a steeply sloping
bedrock stratum 共Poulos 2003a,b兲, or if there are construction
Pile Founded on a Boulder inadequacies, or if one or more of the piles are founded on a
boulder.
An example of the latter source of imperfection occurred in a
Single Piles housing project in Hong Kong. Coring of some of the bored pile
In soil deposits of a colluvial or glacial origin, it is not unusual to foundations indicated that some of the piles were founded at a
have boulders present in soil strata. When piles are driven or higher level, and on less competent rock, than other piles. The
jacked into such strata, it is possible that they will be found on the writer was asked to investigate the consequences of having piles
boulder, rather than extending through to the underlying hard stra- of different length and founding conditions on a pile group con-
tum. Possible consequences of a pile founding on a boulder are as
follows:
1. With driven piles, there may be a false set because the pile
effectively has an enlarged base;
2. With jacked piles, there may be a false indication of the pile
capacity, as the jacking force will incorporate the effect of
the resistance of the boulder as well as the pile itself; and
3. With bored piles, the pile will tend to act as one with an
enlarged base, but having a compressible layer below. This
situation will then be similar to that discussed above for
compressible underlying layers.
To illustrate the possible consequences of a jacked pile being
founded on a boulder, the case of a pile jacked into a dense sand
layer containing boulders will be considered. This example is
based on a real case in the Gold Coast in Queensland Australia,
where the boulders originated from rockfill used to construct an
earlier revetment within a creek bed in which a new major struc-
ture was to be built. The program PIES was used to simulate the Fig. 14. Simulation of installation and subsequent unloading/
following phases of the pile history: reloading of jacked pile on boulder

546 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 Table 2. Summary of Computed Response of Bored Pile Group
Axial load Head moment Settlement
Pile number 共MN兲 共MNm兲 共mm兲
13 37.9 0.98 20.5
14 37.2 0.92 23.4
15 35.0 1.02 21.9
16 35.1 0.99 20.5
17 34.9 0.93 23.4
18 30.1 0.94 21.9

ing pressure, and ultimate lateral pile-soil pressure.

The parameters for the analyses 共soil Young’s modulus, ulti-
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

mate shaft friction, ultimate end bearing, and ultimate lateral pile-
soil pressure兲 were derived from the available SPT data using a
Fig. 15. Pile group layout in wing of high-rise building combination of local Hong Kong correlations and also some cor-
relations developed by Decourt 共1995兲. The upper section of the
piles was cased to a depth of 22 m to minimize possible negative
taining six bored piles, and supporting one of four wings of the skin friction.
high-rise building. Each pile had a shaft diameter of 2.5 m, with The results of the DAMPIG analysis of the 6-pile group are
the base belled to a diameter of 3.0 m. The pile layout for this summarized in Table 2 for an applied vertical load of 210 MN.
wing is shown in Fig. 15. The coring indicated that three of the The following observations may be made:
piles, Nos. 14, 17, and 18, were founded about 40 m below the 1. The maximum differential settlement across the group was
surface in conditions similar to those in BHB4, while the other only about 3 mm.
three piles 共Nos. 13, 15, and 16兲 were founded about 48 m below 2. The axial load distribution was nonuniform, with the longer
the surface, in conditions typified by BHB3. The geotechnical piles 共13,14,16兲 taking greater loads than the shorter piles.
models developed for each borehole are illustrated in Fig. 16, 3. Bending moments were induced in the piles, but these were
together with the average standard penetration test 共SPT兲 values found to be considerably less than the design moment capac-
for each layer. ity of the pile sections.
The consequent analysis was carried out in two stages: In addition, a lateral deflection was induced, but was only
1. Analyses of single pile response were undertaken using the about 1.5 mm.
computer program PIES for axial response and DEFPIG for The relatively small effects of the construction imperfections
lateral response. Analyses were carried out for each set of that resulted in piles of unequal length were largely due to the
three piles to obtain the axial and lateral stiffness values. piles being relatively long and the contrast in founding conditions
2. The computer program DAMPIG 共Poulos 1997兲 was then not being great. As a consequence of the analyses carried out, it
used to analyze the behavior of the 6-pile group, using the was decided that no additional remedial works would be neces-
single pile characteristics for each of the two sets of piles, sary in this case.
assuming that the pile cap was rigid.
In each case, account was taken of the realities of nonlinear pile
behavior by specifying limiting values of shaft friction, end bear- Influence of Structural Defects—Necking
of the Shaft

Single Piles
An example of the effect of structural damage on the load-
settlement behavior of a single floating pile is shown in Fig. 17.

Fig. 17. Example of effect of structural defect on load-settlement

Fig. 16. Details of piles analyzed curve for single end bearing pile

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005 / 547

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

Fig. 18. Load test on pile, showing structural failure 关adapted from Fig. 19. Influence of structural defects on pile head stiffness—
Abdrabbo 共1997兲兴 floating pile 关adapted from Poulos 共1997兲兴

stiffness of structurally defective pile

The structural defect consists of a “neck” in the pile from the pile RKS = 共5兲
stiffness of intact pile
head to a depth of 8 m, the necked area being 10% of the area of
the undamaged pile. Load-settlement curves are plotted for vari- Figs. 19 and 20 plot RKS as a function of the structural integrity
ous values of the compressive strength of the pile material. The factor, FSI, for friction piles and end-bearing piles, respectively. It
applied load P is expressed as a proportion of the ultimate geo- is assumed that the pile has a compressive strength sufficient to
technical capacity Pu of the pile. The following characteristics not fail structurally. In each figure, mean curves are plotted for a
may be noted: defect located at the pile top 共zd / L = 0兲 and at 0.4 times the pile
1. As would be expected, the presence of the defect leads to an length below the top 共zd / L = 0.4兲. The applied load level P / Pu is
increase in settlement; 0.5.
2. For pile strengths in excess of about 100 MPa, the pile re- These figures show that
mains intact structurally, and the settlement increase is rela- 1. Different types of defects give similar values of RKS, if the
tively small; value of FSI is the same;
3. As the pile compressive strength reduces, the pile fails struc- 2. The head stiffness reduction factor RKS decreases as FSI de-
turally, at progressively lower loads, relative to the geotech- creases 共i.e., as the defect becomes more extensive兲;
nical capacity Pu; 3. The reduction in stiffness is more severe when the defect is
4. The onset of structural failure results in a very sudden in- at or near the surface rather than at depth;
crease in settlement; this behavior is consistent with the re- 4. For a given value of FSI, the reduction in pile head stiffness
sults of model pile tests reported by Rao 共1996兲. is generally more severe for an end-bearing pile than a fric-
The latter conclusion is also confirmed in a pile load test on a 15 tion pile; and
m long bored pile reported by Abdrabbo 共1997兲. As shown in Fig.
18, a sudden failure occurred at a load of about 1,200 kN, which
is symptomatic of a fatal structural defect. Integrity testing of this
pile indicated the presence of a major defect at a depth of about 6
m.
In an attempt to provide a more convenient and general mea-
sure of the extent of pile structural defects, a “structural integrity
factor” FSI has been developed by Poulos 共1997, 1999b兲, in which

E dA dL
FSI = 共4兲
EALd
where E⫽Young’s modulus of intact pile; A⫽cross-sectional area
of intact pile; L⫽total pile length; Ed⫽Young’s modulus of dam-
aged portion of pile; Ad⫽cross-sectional area of defective portion
of pile; and Ld⫽length of defective portion of pile. If the pile is
intact, FSI is infinity; if the pile is seriously defective, FSI may be
small 共e.g., 0.1 or less兲, with a lower limit of zero.
To provide a measure of the pile performance, the “head stiff- Fig. 20. Influence of structural defects on pile head stiffness—end
ness reduction factor” RKS is introduced, where bearing pile 关adapted from Poulos 共1997兲兴

548 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 5. The effects of structural defects begin to be significant for
FSI less than about 1.
Some small-scale model tests on floating piles with diameter dis-
continuities, simulating shaft necking, have been reported by Xu
and Poulos 共2000a,b兲. In these tests, the load-settlement behavior
of piles with and without defects has been recorded, enabling the
reduction in pile head stiffness and capacity to be measured. Fig.
21 shows a comparison between the experimental results and the
theoretical curves in Fig. 19. The agreement is generally reason-
able, with the experimental results being reasonably consistent
with the lower bound curve expressing the relationship between
loss of stiffness and the structural integrity factor FSI.
There have been recent investigations of the effects of defects
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

on axial response using finite element analysis by Petek et al.

共2002兲, who have considered axial behavior, and their work has
confirmed the critical effect that the position of the defect has on
pile performance. Sarhan et al. 共2002兲 and Sarhan and O’Neill Fig. 22. Pile group analyzed 关adapted from Poulos 共1997兲兴
共2002兲 have considered the lateral response of piles with flaws.
Model tests and numerical analyses revealed that a concrete void
occupying 15% of the cross-sectional area of the pile, and pen- Fig. 23 shows the reduction in axial stiffness of the group as a
etrating inside the reinforcement cage, can reduce the flexural function of the proportion of piles which have structural defects.
resistance by about 30% compared to an intact shaft. The shaft This figure reveals the following, and expected, results:
ductility is also affected adversely. 1. The reduction in stiffness is more severe as the percentage of
defective piles increases;
Pile Groups 2. The reduction in stiffness is more severe as the applied load
level increases; and
For the purposes of examining the influence of defective piles on 3. The reduction in stiffness of the group is less than for the
group behavior, a simple hypothetical example has been consid- case of an isolated pile because of the ability of the stiffer
ered, as illustrated in Fig. 22. A group of six bored piles, 1.0 m in undamaged piles in the group to carry a greater proportion of
diameter and 20 m long, is assumed to be installed through a the load.
soft-medium clay layer onto a stiffer stratum. The pile cap is The location of the defective pile has some effect on the average
assumed to be rigid. The ultimate axial geotechnical capacity of axial stiffness of the group. For example, for an applied load level
each pile is 9,425 kN in compression, of which 6,280 kN is de- of 50% of ultimate, when Piles 1 and 2 have defects, the axial
rived from end bearing. The group is subjected to vertical loading stiffness reduction is about 30%, whereas if Piles 1 and 4 are
only, with loads up to 80% of the ultimate capacity of the group defective, the axial stiffness reduction is over 45%.
being applied. Changes in load distribution will occur when defective piles
The basic case of a group with no defective piles has been are present in the group. The load carried by the defective piles
used as the reference case. Consideration has then been given to a decreases, while the intact piles are forced to carry more load. For
group containing piles which have structural defects, with the pile the case of four defective piles, the two intact piles 共numbers 3
being severely “necked,” such that the structural integrity factor and 5兲 carry substantially larger loads than the other piles, and
FSI is 0.25, and the defects commencing from the pile head. For indeed the capacity of Pile 5 is fully mobilized. Nevertheless, in
each defective pile, the reduction in axial stiffness has been de- that case, the corresponding reduction in pile group stiffness is
rived from Fig. 20. only slightly more than 50%, demonstrating that the ability of the
group to redistribute the loads restricts the loss of stiffness.

Fig. 21. Influence of structural defects on pile head stiffness— Fig. 23. Effect of structurally defective piles on axial group stiffness
experiment versus theory 关adapted from Poulos 共1997兲兴

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005 / 549

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

Fig. 25. End bearing load-settlement relationships for piles on gravel

layer 关adapted from Woo and Moh 共1990兲兴

affected very significantly by the presence of a 25 mm layer of

slime at the pile tip. However, in the other two cases, the load-
settlement curves for the clean and “unclean” bases are quite
similar initially, and only separate significantly when the shaft
Fig. 24. Effect of structurally defective piles on lateral response of resistance is fully mobilized. Thereafter, large settlements are ex-
group under axial loading 关adapted from Poulos 共1999b兲兴 perienced by the piles with the unclean base. The deeper the pile,
the less significant is the effect of the unclean base on the load-
settlement behavior. This characteristic has also been observed
An important feature of group response with defective piles is
previously in relation to the effects of clay seams in rock below a
that, under vertical load, the group will deflect laterally and suffer
pile tip.
a rotation, with bending moments being developed at the pile
The latter two piles involved in the Woo and Moh study have
heads. Fig. 24 shows the computed maximum pile head moment,
been analyzed by the writer using the computer program PIES.
lateral deflection, and rotation as a function of the applied load
The presence of debris at the base has been simulated by a soft
level and the number of defective piles. It will be seen that, once
pile element at the base of the pile, with a Young’s modulus of 30
the applied load level exceeds about 0.35, substantial bending
MPa. The shaft friction and soil modulus values have been chosen
moments can be induced, perhaps exceeding the yield moment of
initially via judgment, with some later adjustment to obtain a
the piles. Significant lateral deflection and rotation are also devel-
better fit to the measured and calculated load-settlement curves
oped. The more asymmetric the location of the defective piles, the
for the piles with a clean base. The same parameters have then
more severe is the induced lateral response effects. For the cases
been used for the piles with the unclean base. Figs. 26 and 27
analyzed, the group with three defective piles 共pile numbers 1, 2,
show comparisons between the measured and calculated load-
and 4兲 appears to generate the most severe lateral response.
settlement curves. It can be seen that the analysis is capable of

Influence of Geotechnical Defects—Piles

with a “Soft Base”

Single Piles
A “soft base” at the toe of a bored pile is usually a result of debris
left from inadequate base cleaning during construction. Because
of its confinement by the surrounding soil, the debris is unlikely
to fail, but it will be expected to deform considerably. A very
instructive case history of the effects of inadequate base cleaning
has been presented by Woo and Moh 共1990兲. They have presented
load-settlement curves for three cases involving inadequate base
cleaning:
• End bearing piles on gravel;
• Piles about 32 m long through variable deposits, bearing in a
gravel layer which is underlain by clay; and
• Piles about 42 m long through variable deposits, bearing in a
fresh 共but fractured兲 sandstone.
The load-settlement curves measured for each of these three cases
are shown in Figs. 25–27. It can be seen from Fig. 25 that the Fig. 26. Effect of base cleanliness on pile performance 关adapted
load-settlement behavior of the piles bearing directly on gravel is from tests of Woo and Moh 共1990兲兴

550 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

Fig. 28. Observed floor elevation changes over 30 day period

关adapted from Milligan 共1997兲兴

Fig. 27. Effect of base cleanliness on pile performance 关adapted two shear walls near one side of the building. Fig. 28 shows the
from tests of Woo and Moh 共1990兲兴 measured changes in basement floor elevations. Subsequent in-
vestigations showed that not all the caissons were founded on
bedrock, and that the concrete shafts and/or bells had become
identifying the effects of the base debris in increasing the pile contaminated with fill debris, as indicated in Fig. 29. The location
settlement after mobilization of the shaft resistance. of the largest settlements was where a significant amount of fill
The computed load-settlement curves for the geotechnically debris was found at the base of the pile, and the bell had a
defective piles in Figs. 26 and 27 are also consistent with some relatively small enlargement. The structure was subsequently
other field observations. For example, Thorburn and Thorburn demolished.
共1977兲 quote the example of an 18 m long, 600 mm diameter This case demonstrates the vital importance of proper con-
bored pile, which performed satisfactory under a test load of 1.16 struction techniques in bored pile construction to avoid imperfec-
times the required working load, but appeared to fail 共settle ex- tions which may ultimately lead to destruction of the structure
cessively兲 at 1.5 times the working load. It was found subse- that the piles were intended to support. It was also perhaps unfor-
quently that the base was resting on about 0.6 m of natural debris tunate that the structure and the pile caps were not more rigid, as
共small lumps of soft clay兲 which had fallen from the unlined there appeared to be little ability for load redistribution to occur in
portion of the bore, thus giving a soft base condition. this case, and the highly imperfect pile thus appeared to be unable
to transfer load on to less imperfect piles.
Pile Groups
Case History of Two Buildings in Hong Kong
Groups of piles in which some contain geotechnical imperfections
behave in a similar way to those with piles containing structural Construction of two adjoining residential high-rise blocks in Sha-
defects. While the failure of geotechnically defective piles is less tin, Hong Kong commenced in mid-1999. The 41-story blocks
abrupt than for those with structural defects, the overall group were each supported on 18 bored piles, 2.3 m in shaft diameter,
behavior is more forgiving because of the redundancy provided with a base belled out top of 3.8 m. Fig. 30 shows a plan of the
by the group. The consequences of having defective piles will two blocks 共denoted as Blocks D and E兲, while a typical geotech-
again depend on their location, and if the defective piles are con- nical profile for a section at the site is shown in Fig. 31. The piles
centrated in a particular part of the group, there may be a signifi-
cant rotation induced in the group, with consequent lateral deflec-
tions and induced moments in the piles.
An interesting case history of the consequences of having geo-
technically defective piles has been reported by Milligan 共1997兲.
During construction of a 13-story building, settlement of the
foundations led to cracking and severe distress of the partially
completed structure. The building was approximately rectangular
in shape 共15 m by 60 m in plan兲, with the structural loads carried
by shear walls down to drilled shaft 共caisson兲 foundations. The
shafts were 750 mm in diameter, and extended through mixed
clayey fill and dense glacial till to shale bedrock, the upper sur-
face of which was fractured and weathered. The bases of the
caissons were belled by hand to diameters ranging between 1 and
2 m. The caissons were backfilled with concrete and were
unlined.
Settlement was observed about 6 months after completion of
the caissons, and as the building height increased, so did the Fig. 29. Partial section showing details observed during investigation
settlements. The excessive settlement was concentrated in one or 关adapted from Milligan 共1997兲兴

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005 / 551

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

Fig. 30. Foundation plan for high-rise buildings

were designed as end bearing piles, to be founded in Grade II 2000. The conclusions of the independent consultant carrying out
共slightly weathered兲 granite, at a depth ranging between about 35 the coring and the assessment of pile conditions are summarized
and 46 m. in Table 3.
When the buildings had reached the 17th story, a program of It can be seen that only four of the total of 36 piles were
settlement monitoring was commenced on the buildings, follow- compliant with the construction specifications which required
ing concerns about pile construction practices with some other founding on Grade II rock. The pile lengths were up to 14 m short
projects in Hong Kong. Two to three months later, it was ob- of bedrock level. In addition to the geotechnical deficiencies, six
served that both buildings were settling unevenly, with each block of the concrete cores that were taken from the piles and tested
tending to tilt as a rigid body. A program of investigatory coring showed low uniaxial compressive strengths, ranging between 22.5
through the bored piles was subsequently instigated in December and 2 MPa, compared to the nominal strength of 35 MPa.
1999, and this program revealed a number of deficiencies in both Fig. 32 shows typical time-settlement readings for one of the
the pile length and the founding conditions at the base of the blocks, the locations being as shown in Fig. 30. It is interesting to
piles, and construction was halted at the 34th floor, very early in note that the coring investigation works caused a temporary ac-

Fig. 31. Geotechnical profile along Section 1-1

552 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 Table 3. Summary of Inferred Pile Founding Conditions from Coring
Number of piles
Pile founding condition Block D Block E
Pile founded on Grade II rock 3 1
Thin sediment layer between pile base and rock 5 1
Sediment below pile base, underlain by inferior rock 5 4
Pile founded on inferior rock, without sediments 1 12
Thick layer of soft material below pile base 4 0

celeration in settlement of the buildings, and was an important

factor in the decision to suspend construction. Indeed, about 30
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

mm of additional settlement at Block E was attributed to the

coring process.
Fig. 32. Time-settlement records for four typical points—Block E
Because of concerns about the integrity of the existing foun-
dation system, extensive investigations were carried out by sev-
eral parties to assess whether it would be feasible to carry out
remedial work to the foundation system so that construction could mencement of settlement measurements after the 17th floor
be completed with an adequate level of confidence in the integrity had been reached. In this calibration, the observed additional
of the foundations. The writer carried out one of these investiga- settlements due to the coring process were excluded from the
tions, and took the following approach. total settlements, so that only settlements due to the building
• Geotechnical models were developed for each of the 36 piles, load were considered.
based on the site investigation data and the information de- • Reasonable fits between measured and calculated uneven
rived from the coring investigation. settlement were obtained, using soil modulus values of be-
• Analyses were carried out, using the program PIGS, to com- tween 2.8N and 4.2N MPa, where N = SPT value. Fig. 33
pute the relationship between building tilt and building height. shows the computed and measured settlements from Floors 17
The interaction between the two buildings was considered, and to 34, for Block E, and for three sections.
the geotechnical properties of the various layers were adjusted • Using the fitted model, PIGS was then used to investigate a
关according to the standard penetration test 共SPT兲 values兴 to number of alternative future options, including:
obtain a fit with the measured increases in tilt since the com- 1. Construction with the existing foundation system.

Fig. 33. Measured and calculated settlements—Block E

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005 / 553

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 2. Construction with up to 66 new steel H-piles installed in existing structure and foundation system are likely to be
prebored holes, and extending 9 m into the Grade II granite. more immediate and severe than with a “greenfield” site.
An extended pile cap would be constructed to link the old For these reasons, it is worthwhile to give some attention to the
and new foundation elements, and this would require up to 5 problems of investigation, design, and construction within an ex-
m excavation and some dewatering. isting foundation system, and to examine in some detail the pos-
3. Truncation of the buildings so that the existing foundation sible consequences of inadequate control of the resulting ground
system would be adequate without any further enhancement. movements.
It was found that the first option was not feasible. In principle,
the second option, involving enhancement of the foundation sys-
Investigation Issues
tem with H-piles, would provide a satisfactory solution. However,
the risks involved in carrying out the remedial foundation works Investigations to identify possible causes of foundation perfor-
beneath heavily loaded buildings were considered to be unaccept- mance irregularity may involve some form of drilling adjacent to,
able, particularly in view of the potential for significant ground or through, the existing pile foundations. Both forms of drilling
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

movements to be generated during the enhancement works. Trun- may have deleterious effects on the piles being investigated. As
cation of the buildings was assessed to be technically feasible, but previously mentioned, the drilling of holes adjacent to piles will
was not considered to be economically feasible by the owner generally cause vertical and lateral ground movements and these
because of the stigma of having two lower-rise buildings 共12 or will act upon the nearby piles, inducing additional stresses and
13 stories high兲 within a larger group of much taller buildings. movements. These effects may be particularly severe if the
In a decision that was widely publicized in Hong Kong, the ground is highly stressed. Coring through the pile itself may also
decision was taken to demolish both buildings, and this was car- create difficulties for the existing foundation, via the following
ried out in 2001–2002. This was therefore a rather extreme case mechanisms:
of the consequences of a series of construction imperfections, • Unbalanced fluid pressures inside the core hole and outside the
leading to three undesirable conditions: pile. These may cause loosening or even piping of the soil
• Piles with a soft base; beneath the pile toe when “breakthrough” is achieved and the
• Piles with variable founding conditions; and underlying soil is soft or loose. In turn, the settlement of the
• Piles of variable length within a group. pile may then be exacerbated.
• SPT testing of the soil below the pile base may cause further
disturbance if it is not carried out carefully, and the SPT rods
Coping with the Consequences of Imperfections are withdrawn too quickly, thus causing a suction within the
during and after Construction soil surrounding the hole.
Thus the investigation process itself may help to accentuate the
problem being investigated. Such a mechanism was thought to
Introduction have contributed to the additional settlements of up to 30 mm
It would appear essential that monitoring of the foundation be- during the previously mentioned investigation of defective bored
havior be carried out during and after construction, so that any piles in a high-rise building in Hong Kong. Clearly, it is impera-
irregularities can be identified at the earliest possible opportunity, tive that such possible “side effects” of the investigation are an-
and appropriate investigation works, and if necessary remedial ticipated and that appropriate cautionary measures are adopted to
works, can be undertaken. The later problems are uncovered, the minimize the negative impact of these side effects.
more difficult and costly remedial action is likely to be.
In new foundation construction, there is usually relatively un- Design Issues
impeded access to the site and to the areas in which the new
foundation system is to be constructed. However, the environment There are at least three key design issues that may need to be
around or within an existing foundation system which is being addressed when designing remedial works for pile foundations
investigated and/or upgraded may pose several constraints and which have been demonstrated to be inadequate or are not per-
problems, and the following characteristics can be anticipated: forming to expectations:
1. Access to the area may be very difficult and may limit the • Correction of uneven settlements, if the foundation has already
range of construction methods that can be employed; undergone excessive tilting or differential settlement, or is
2. The ground will often be highly stressed, and thus changes in likely to do so during or after the remedial works;
the stress regime due to investigation or construction may • Design of remedial or enhancement works, which may include
result in larger ground movements than would be the case in repair of defective piles, the installation of additional piling, or
a “greenfield” situation; extension of the pile cap to obtain additional capacity and
3. The foundation system will generally be loaded, and it is stiffness; and
therefore necessary to try and assess these existing loads so • Consideration of the load sharing between the existing piles
that a more realistic design can be carried out. One cannot and the additional foundation elements. It is possible that ex-
assume 共as so often happens in new design兲 that the founda- cessive load may be carried by the additional elements, unless
tion system is stress-free and load-free; the design can incorporate means of controlling the distribu-
4. Existing piles will generally be subjected to some measure of tion of loads in the upgraded system.
restraint from the building which they are supporting, via These issues are discussed in more detail below.
attachment to pile caps and the overall foundation system;
5. Strict control of investigation and construction processes are Correction of Uneven Settlement
likely to be more critical, but more difficult to achieve, than
with “greenfield” situations; and Methods of correcting uneven settlements of buildings can be
6. The consequences of uncontrolled ground movements on the divided broadly into two categories:

554 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 1. “Hard” methods, which rely on the application of some form
of direct force to the building; and
2. “Soft” methods, which rely on processes which produce cor-
rective foundation movements by inducing appropriate
ground movements.
In both cases, the treatment can be carried out to lower the build-
ing on the “high” side, or alternatively, to raise the low side of the
building. In both cases, the treatment may be accompanied by
some form of foundation strengthening or remediation on the
“low” side. Amirsoleymani 共1991a,b兲 lists six different methods
that have been used to reduce or eliminate differential settlement
and tilting. Three methods involve lifting of the low side of the
structure, while three involve lowering of the high side.
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

“Hard” Methods
1. Application of Force by Anchor Stressing. This method in-
volves the installation of a series of strategically located an-
chors within the foundation system. The anchors are grouted
into a suitable hard stratum at depth below the building. The
anchors are then stressed 共typically to 60–75% of their ulti- Fig. 34. Details of soil extraction scheme for grain silo on mat
mate capacity兲 to obtain a corrective tilt to the foundation. In foundation 关adapted from Amirsoleymani 共1991b兲兴
some cases, repeated stressing and destressing may have a
beneficial effect in developing additional corrective tilting.
2. Application of Additional Loading. This method involves the plumb position. Amirsoleymani 共1991a兲 describes the use of
application of additional loading on the high side of the foun- mechanical jacking to correct the settlement of a storage
dation by water or other means. There are often limitations tank, and the use of chemical grouting to restore a piled
on the amount of loading that can be applied because of foundation that tilted and failed after a deep excavation
height limitations within the structure, and the limitations nearby caused piping of sand near the pile tips. An expansive
that may be imposed by the limited strength of the structure admixture was used in the grout to promote uplift of the
itself. columns, which were raised by 28 mm. Maffei et al. 共2001兲
3. Cutting of Piles. This method involves the cutting of some describe a case of a tall building in Sao Paulo in which the
piles supporting the high side of the building to promote load uneven settlement causing a tilt of 2.2° was corrected by
transfer to other piles on the high side and hence promote constructing a new pile foundation and jacking the low side
beneficial settlement. When the process has been completed up against these piles to transfer load from the old to the new
and the settlements have ceased, the piles may be reattached foundation system.
so that they may carry part of the future loadings. The cutting 6. Fracture Grouting. This method involves the use of a grout
of the piles is in itself a process that requires considerable under controlled high pressure to fracture the soil and cause
care so that the cut pile is not destroyed in the process. A uplift of the foundation. Amirsoleymani 共1991a兲 has de-
method that has been developed in Hong Kong involves the scribed a case of a five-story warehouse in which 210 mm of
partial cutting of both sides of the pile, the placement of differential settlement was corrected by hydraulic fracturing
jacks to support the upper and lower parts of the pile and through 24 tubes installed into rock below the foundation.
carry the load existing in the pile, and then the cutting away Cement lenses 50–100 mm in thickness were found to have
of the remaining central portion of the pile which is freed of been formed by the fracture grouting.
load prior to the cutting by the jacks. This process has been
used in conjunction with anchor stressing in correcting a
“Soft” Methods
building in Hong Kong. Unfortunately, because of legal con-
straints, it is not possible to present details of this case. 1. Soil Extraction. In this method, soil is excavated from be-
4. Jetting of the Soil Beneath Pile Tips. There have been anec- neath or between the piles on the high side. This process
dotal reports of the use in China of high pressure water jet- causes the ground to settle and thus induces a settlement of
ting applied below the tips of piles to reduce their stiffness the pile foundation also. Soil extraction was used to arrest the
and capacity and promote settlement of the high side of the tilt of the famous Pisa Tower in Italy 共which was supported
building. This process is in some ways similar to the cutting by a shallow foundation兲 共Jamiolkowski 2001; Burland
of piles, but is a less controlled procedure whose effects ap- 2004兲. Amirsoleymani 共1991b兲 described the use of a similar
pear to be difficult to predict. process to correct the tilt of a grain storage silo. In that case,
5. Jacking of the Foundations on the “Low” Side. This method thin layers of soil were extracted via specially constructed
generally employs compaction grouting to push the low side chains which cut through the soil 共Fig. 34兲. Brandl 共1989兲
up and at the same time strengthen the foundation. An ex- has described the use of soil extraction to correct uneven
ample of this approach is described by Tsai et al. 共1991兲 and settlement of piles supporting bridge piers, while the use of
was used to correct a building on a raft foundation that had soil extraction to correct uneven foundation settlements has
tilted by about 0.74° after being subjected to hurricane load- been described by Tamez et al. 共1997兲. In this case, involving
ing together with seismic forces. A process of staged grout- two historic churches in Mexico City, soil extraction was
ing allied with careful monitoring restored the building to a carried out via 32 shafts 3 m in diameter, from tubes 100 mm

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005 / 555

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 in diameter, inserted a maximum distance of 22 m into the
soil. After 4 years of treatment, more than 2 , 600 m3 of soil
had been removed, and corrective settlements of about 800
mm were achieved. It was estimated by the writers that about
65% of the settlement could be attributed to the soil extrac-
tion, while the remaining 35% was due to consolidation aris-
ing from pumping from the deep wells at the bottom of the
shafts. Poulos 共2003b兲 developed a method of analysis for
pile foundations subjected to ground movements from soil
extraction, in which the process of drilling subhorizontal
holes is represented by the formation of small-diameter tun-
nels. The resulting ground movements are used in a pile-soil
interaction analysis via the program PIGS to assess the effect
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

of the underexcavation on the pile group response.

2. Dewatering. In principle, lowering of the water table can be
used to promote settlement of the high side and thus correct
uneven settlement. Amirsoleymani 共1991a兲 describes the use
of this approach for apartment building supported by raft
foundations. However, this method is fraught with difficulty
as the effects of ground water lowering are highly dependent
on the local hydrogeology, are time-dependent, and may ex-
tend considerably beyond the building being treated. Amir- Fig. 35. Principle of RSS method 关adapted from Poulos et al. 共2003兴
soleymani 共1991a兲 summarized this method succinctly as fol-
lows: “lowering water table to eliminate differential
settlement is one of the most unreliable methods.” Despite 3. Extension of existing pile caps or rafts to provide additional
these misgivings, Liu 共2004兲 has described a successful ap- bearing capacity and stiffness.
plication of dewatering in conjunction with grouting to re- In each case, attention needs to be given to the transfer of load
duce the tilt of a building adjacent to a deep excavation. from the existing to the enhanced foundation system.
3. Compensation Grouting. van der Stoel et al. 共2003兲 described Repair of Existing Piles. Methods of repair of existing piles
the use of compensation grouting via fracturing to control the are relatively limited because access to the pile is frequently lim-
vertical movement of timber pile foundations affected by ited, the nature of the defect may be uncertain, and the repair
tunneling operations. The system involved the use of 22 sub- process may itself result in deleterious side effects, such as addi-
horizontal tubes-a-manchette 共TAMs兲 in two levels. About 5 tional ground movements, which may adversely affect the re-
mm of heave was observed on a structure after postgrouting mainder of the foundation system. Structural defects, especially
was performed following the tunneling. those found near the head of a pile, have been repaired with
4. Removal of Soil Support (RSS). Poulos et al. 共2003兲 de- epoxy jackets. Geotechnical defects, especially soft bases as a
scribed an approach which involves the drilling of a series of result of inadequate base cleaning, have been treated successfully
vertical or subvertical holes just outside the high side of the by base grouting techniques. Examples have been described by
structure 共see Fig. 35兲. The removal of soil reduces the lat- Teparaksa et al. 共1999兲 and Moh 共1994兲. Fig. 36 shows the dra-
eral support of the ground and therefore promotes settlement matic improvement in load-settlement performance arising from
of the high side of the structure. The greater the stress im- base grouting a bored pile 1.5 m in diameter and 22 m long 共Moh
posed in the ground by the building, the greater will be the 1994兲. At the working load of 6.6 MN, the settlement was re-
settlement. Excavation of a continuous trench can, in prin- duced from 80 mm to less than 10 mm.
ciple, cause significant settlement to occur. The advantage of Addition of New Piles. The addition of piles to an existing
this approach is that it can be applied without intruding into foundation system is a remedial measure of long standing, espe-
the building footprint. While no cases of the use of this cially in relation to underpinning operations, and has been carried
method for pile foundations appear to have been reported in out using a variety of systems for the new piles. Because of con-
the literature, it does appear to have been used successfully straints with access in most cases, the methods of installation
for buildings on shallow foundations 共Zou 1996兲. Settle- must allow for limited headroom; as a consequence, most addi-
ments of the order of 150 mm appear to have been developed tional piles tend to be relatively small in diameter, although their
by this approach. Preliminary centrifuge tests on a model length can be significant 共e.g., Bruce 1994兲. The use of jet grouted
shallow footing 共C. W. Ng, personal communication, 2002兲 piles has been described by Popa et al. 共2001兲, while the use of
have confirmed the potential of this approach. compaction grouted piles to remediate an extensive warehouse
raft foundation has been described by Hayward Baker 共2003兲.
In analyzing the effectiveness of a foundation with additional
Foundation Enhancement Works remedial piles, it is necessary to take account of several factors,
including the capacity and stiffness of the existing foundation
Among the options that may be considered for foundation en- system, the time at which the additional piling is installed, and the
hancement works are the following: interaction between the original system and the new components,
1. Repair of the existing piles which contain imperfections or including the subsequent load sharing. Makarchian and Poulos
defects; 共1996兲 developed an approximate method of analysis of a reme-
2. The addition of new piles to strengthen and/or stiffen the diated foundation system which uses concepts of pile-raft inter-
existing foundation; and action. This method has been applied to model pile tests and

556 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

Fig. 36. Effects of base grouting on pile behavior 关adapted from Moh 共1994兲兴

reasonable agreement has been found between the measured and pile load test results is shown in Fig. 38 and indicates an ultimate
computed ratio of settlement reduction arising from the installa- capacity of only about 2,500 kN. Many of the piles had already
tion of piles beneath a shallow footing. been installed and the pile caps cast. For the particular pile tested,
An issue of concern with the installation of additional piles is the working load on the column was about 2,000 kN, and various
the disturbance of the ground caused by the installation process. options for remedial action were considered. The conventional
Installation of displacement piles will cause both vertical and lat- solution was to add an additional two piles to the group 共one on
eral ground movements which will interact with the existing foun- either side of the original pile, to avoid asymmetric loading兲, thus
dations system. Such ground movements have the potential to
cause additional vertical and lateral forces and bending moments
in existing piles, which may compromise their integrity. The ef-
fects of such ground movements may be particularly severe for
the case where the existing piles are restrained from moving lat-
erally or vertically, as mentioned previously. For piles which are
installed by drilling, there may also be potentially damaging
ground movements as a consequence of the release of lateral
ground stress and changes in water pressure arising from lack of
control of water levels during drilling. The effects may be of
particular concern if the existing foundation is heavily loaded, as
the release of stress due to drilling can then be large. These side
effects of foundation remediation are thought to have been sig-
nificant in a recent case in Hong Kong, where an existing building
adjacent to a building being remediated experienced some addi-
tional settlements during the installation of over 50 remedial
piles.
Extension of Existing Pile Caps. A remedial option which can
be useful is to extend an existing pile cap and make use of its
capacity and stiffness. Clearly, this option may be limited to those
cases in which the near-surface soils themselves have reasonable
strength and stiffness. The performance of such a remediated sys-
tem can again be assessed readily using pile raft analysis concepts
共Randolph 1994; Poulos 2001兲.
The above approach was recommended in a major project in
Queensland Australia, involving jacked piles in sand which had
not developed their anticipated ultimate capacity of 4,000 kN
共this being the jacking force employed for installation兲. Fig. 37
shows the geotechnical conditions at the site and reveals reason-
able ground conditions above the pile founding level. One of the Fig. 37. Geotechnical profile for jacked pile project

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005 / 557

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 Control of Load Distribution in the Foundation
System

An issue of concern is the transfer or sharing of load between the

original and the remediated foundation system. It should be ex-
pected that, following the laws of mechanics, a greater proportion
of any additional load that is imposed on a system would be
carried by the stiffer components. This may in turn mean that the
new and stiffer piles may become overstressed or overloaded. It is
possible to control the load distribution within a pile foundation
system by means of inserts that are attached to the head of se-
lected piles. Fig. 39共a兲 shows a schematic of these inserts, which
may be neoprene or a similar semicompressible material. These
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

serve to decrease the pile head stiffness in a controlled manner,

and are referred to herein as “controlled stiffness inserts” 共CSIs兲.
Fig. 39共b兲 shows an example of the use of CSIs at the heads of
remedial prebored H-piles in a housing project in Hong Kong.
Fig. 38. Load-settlement performance of inadequate pile and of
alternative remedial measures
It is possible to examine the benefits to be gained by the use of
CSIs via an analysis using the program PIGS. This program has
the facility of incorporating a CSI at the head of selected piles,
and of “activating” remedial piles at a predetermined stage in the
necessitating an extension of the pile cap. Another solution inves- loading history of the foundation. Clearly, the earlier the remedial
tigated was to simply extend the pile cap and make use of its piles are installed, the more favorable is the subsequent founda-
capacity and stiffness to complement the existing pile. Calcula- tion performance likely to be. Poulos 共2004兲 has demonstrated
tions were carried out for both these options. The following ap- that the required stiffness, KIi, of an insert for a pile i, to develop
proach was adopted: a specified settlement and load, can be computed from the follow-
1. Estimate the geotechnical parameters for the existing pile ing expression:
and for a pile cap founded near the surface;
2. Calibrate the computed pile behavior with the measured 1/KIi = STi/Pi − 1/KGi 共6兲
load-settlement behavior from the pile load test; and where STi⫽“target” settlement of pile i; Pi⫽required load on pile
3. Carry out calculations for the two remedial options and as- i, and KGi⫽original stiffness of pile i within group environment.
sess their relative suitability. If all piles are required to carry equal load and settle equally, then
The computer program DEFPIG was used to carry out the calcu- Pi will be the average load on the group, and STi will be the same
lations. The ultimate shaft and base capacity of the pile were for all piles within the group. KGi can be computed from a pile
estimated on the basis of the SPT 共N兲 data, using the following
correlations:
• Ultimate shaft friction in the upper cohesive layers⫽20–25
kPa;
• Ultimate shaft friction in the sands⫽2.8N + 10 kPa 共Decourt
1995兲;
• Ultimate end bearing capacity in the indurated sand⫽12 MPa;
• Ultimate bearing capacity of the pile cap⫽210 kPa 共based on
the shear strength of the uppermost layers兲; and
• Young’s modulus of the soils along the shaft⫽2.5N MPa, and
below the pile tip⫽5N MPa 共to allow for compaction during
the installation兲, while an average value of 10 MPa was used
for the soil below the pile cap.
Fig. 38 shows the computed load settlement behavior of the test
pile, which agrees well with the measured behavior. Also shown
are the computed load-settlement curves for the three-pile option
and the extended pile cap option, with a cap 2.5 m square. In this
case, the three-pile option is an overdesign, while the pile plus
enlarged cap appears to perform satisfactorily. At the working
load, the settlement is about 10 mm, which is less than the allow-
able of 15 mm, while the overall factor of safety is in excess of
1.8. It should be noted that for the three-pile option, the pile cap
would have had to be extended to accommodate the two extra
piles, and its area would have been almost 4 m2, which is a con-
siderable proportion of the cap area in the extended cap option.
Clearly, the latter had some economic, as well as logistical, ad-
vantages in this case, in that the additional piles would have been
very difficult to install because part of the structure had already
been constructed. Fig. 39. Bored pile cases analyzed

558 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

Fig. 40. Controlled stiffness inserts

group analysis such as DEFPIG, PIGS, or PIGLET in which all 1. What was the anticipated load-settlement behavior of the
the piles are assumed to be equally loaded. bored pile, as-constructed, versus the behavior as-designed?
From Eq. 共6兲, to avoid a negative insert stiffness, there is a 2. What was the load-settlement behavior of the remediated pile
lower limit for the target settlement 共STi lim兲 which is as follows: group, upon the application of the remaining dead plus live
load?
STi lim = Pi/KGi 共7兲 Two main steps were involved in the subsequent analysis:
Alternatively, there will be a lower limit 共KGi lim兲 to the pile stiff- 1. An analysis to compute the load-settlement behavior of the
ness for which stiffness inserts are required, and this limit is given original pile and the remedial piles; and
by 2. The development of an interaction analysis, which combined
the computed load-settlement behavior of the original pile
KGi lim = Pi/STi lim 共8兲 with the computed behavior of the remedial piles, in order to
obtain the load-settlement behavior of the remediated group.
By adjusting the target value of settlement for each pile, and the
This analysis took into account the existing load acting on
load on that pile, it is possible 共at least in principle兲 to design the
the pile prior to the addition of the remedial piles, and was
stiffness inserts so that the pile group settles in a predetermined
manner. implemented via the computer program PIGS.
An example of an inadequate foundation that was remediated The required geotechnical parameters were estimated, based on
with additional piles is described below. Fig. 40 shows a simpli- correlations used in Hong Kong, as shown in Table 4. The fol-
fied and idealized example of a case involving a large-diameter lowing analyses were carried out.
bored pile which was judged to be founded on an inadequate • The as-designed pile was analyzed to obtain the load-
bearing stratum 共Grade III/IV rock, rather than the intended Grade settlement behavior to failure. This pile was meant to be
II rock兲. The alleged deficiency in the pile was discovered during founded on Grade II rock, and would have had an ultimate
construction of the building, when the dead loading was about load capacity in excess of 250 MN.
one-quarter of the final value. Subsequently, construction was • The as-constructed pile was analyzed to obtain the estimated
halted, and a remedial piling system was designed to compensate load-settlement behavior of the deficient pile. This pile was
for the alleged pile deficiency. The actual remedial system founded on Grade IV/III rock, and had an estimated ultimate
adopted was via the use of eight remedial prebored H-piles, i.e., load capacity of about 104 MN.
H-pile sections which were grouted into prebored holes 0.5 m in • The remedial H-piles were analyzed to obtain the load-
diameter. The new remedial piles were then joined to the original settlement curve to failure. The computed ultimate capacity of
bored pile by means of a pile cap. However, the system consid- each pile was about 8 MN.
ered herein is an alternative scheme consisting of an additional 16 • The as-constructed pile was analyzed with the additional re-
prebored H-piles, as shown in Fig. 40. medial piles being part of the group. The remedial piles were
Two questions needed to be answered in relation to this case: “activated” once the load of 20 MN had been reached 共1 / 4 of

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005 / 559

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 Table 4. Summary of Correlations Used for Hong Kong Pile Group Analysis
Young’s Ultimate skin Ultimate base
modulus Es friction f s pressure f b
Layer 共MPa兲 共kPa兲 共MPa兲 Remarks
Soil layers 共alluvium, completely 1.0N 1.0N — N = SPT value
decomposed granite兲
Grade IV/III rock 500 200 9.0 f b based on factor of safety of 3 applied to allowable value
Grade II rock 20,000 500 22.5 f b based on factor of safety of 3 applied to allowable value

the dead load, the stage at which the remedial piles were in-
stalled兲.
The results of the calculations are shown in Figs. 41 and 42.
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

Fig. 41 shows the load-settlement curves for the three cases de-
scribed above. It can be seen that the pile, as-designed, would
have had a factor of safety in excess of 3 and would have settled
about 8 mm under the design load of 80 MN. The short as-
constructed pile would barely have had adequate capacity to sus-
tain the design load and would have settled about 140 mm. The
remediated pile 共group兲 with 16 prebored H-piles develops suffi-
cient load capacity to bring the capacity back to that of the as-
designed bored pile. However, the settlement of the remediated
pile exceeds that of the as-designed pile, although the beneficial
effects of the increased stiffness of the remediated pile group can
be seen clearly after the loading is resumed beyond 1 / 4 of the
dead load.
Fig. 42 shows the computed loads within the original short pile
and all the additional remedial piles as a function of the applied
Fig. 41. Load-settlement curves for bored pile and effects of reme-
load. Clearly, the original short pile carries all the load until the
dial actions
additional piles are added and the construction is resumed. There-
after, the remedial piles tend to take the majority of the load until
the full capacity of the additional piles is developed. The original
pile then continues to take all the additional load until the ultimate
capacity of the group is mobilized.
Analyses were also carried out to investigate the effects of
including CSIs at the top of the remedial piles. By application of
Eq. 共6兲, it was possible to develop a system that developed equal
loads in the remedial piles. Table 5 summarizes the computed pile
loads and settlement at the design load of 80 MN, without and
with CSIs. It can be seen that, without CSIs, there is a significant
inequality of load in the remedial piles, with the corner piles
carrying about 71% more than the least-loaded piles. With the use
of CSIs, the load distribution is almost uniform in the remedial
piles. As would be expected, the settlement increases with the
CSIs in place, but only by a small amount. The reduction in load
inequality is particularly important in places such as Hong Kong,
where local building regulations require that all piles should have
a prescribed factor of safety against individual pile failure of at
least 2 at the working load. While this requirement has little logi-
Fig. 42. Load carried by original bored pile in remediated system cal basis, it is nevertheless enforced. The use of CSIs in this case

Table 5. Summary of Computed Performance of Remediated Bored Pile 共at Design Load of 80 MN兲
Without inserts With inserts
Pile load Settlement Pile load Settlement Insert stiffness
Pile 共MN兲 共mm兲 共MN兲 共mm兲 共MN/ m兲
Original bored pile 21.24 47.5 21.72 51.3 —
Remedial piles A 4.79 47.5 3.66 51.3 992
Remedial piles B 3.91 47.5 3.62 51.3 1,865
Remedial piles C 3.19 47.5 3.66 51.3 6,021
Remedial piles D 2.80 47.5 3.64 51.3 —

560 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 facilitates the acceptance of the remedial design. ally more than compensates for the small settlement increase that
This example demonstrates the importance of considering the occurs, and may facilitate acceptance of a remedial scheme which
soil-structure interaction processes and modeling the correct se- might otherwise be required to have more piles added to satisfy
quence of applied loading. It also shows the importance of being local design requirements related to the factor of safety of indi-
able to rationally analyze the behavior of pile groups which con- vidual piles within a group.
tain piles which are not all identical, and which may contain In his paper to the 1951 Building Research Congress, Terzaghi
either structural defects or 共as in this case兲 geotechnical defects. It 共1951兲 stated: “Modern methods for the design of pile founda-
also demonstrates the potential for using CSIs to control the dis- tions … can be considered reliable, provided the subsoil of the
tribution of load within a group of piles. proposed foundation has been adequately explored. On the other
hand, if the subsoil conditions have been misjudged, for instance,
on account of inadequate sampling operations, erroneous interpo-
Conclusions lation between boreholes, or lack of care in the examination of the
soil samples, the difference between anticipated and real settle-
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

Natural imperfections, arising from the geology of the site, can ment can be distressingly important.” It is hoped that this paper
create significant problems for a piled foundation and lead to will make some contribution to the understanding of the possible
reductions in both capacity and stiffness of a single pile or pile effects which complex real soil conditions can have on the behav-
group. Among the more significant imperfections are ior of pile foundations, and an appreciation of the many difficul-
• Clay seams below the pile toe; ties which, more than 50 years later, geotechnical engineers still
• Compressible layers below the founding levels of the piles; face in forecasting pile foundation performance.
• Soil layers of uneven thickness; and
• Differences in founding conditions, which can give rise to
piles of different length within a group. Acknowledgments
The first two sources have been shown to have the potential to
reduce axial pile load capacity and decrease the stiffness. The The writer is grateful to John P. Carter, John C. Small, Patrick
effects on pile groups are generally more significant than on a Wong, and Masoud Makarchian for their constructive comments,
single pile. The latter two sources of imperfection are of particu- and to Charles C. W. Ng for providing details of the RSS centri-
lar concern, as they generate uneven settlements and can induce fuge tests.
unexpected bending moments and shears in the piles under axial
loading.
Construction-related imperfections include both “geotechni- References
cal” and structural defects. Geotechnical defects, such as a “soft
base” at the pile tip, may cause a reduction in pile head stiffness Abdrabbo, F. M. 共1997兲. “Mis-use of soils and foundation causes disas-
that depends on the soil modulus at the pile tip, and the applied ter.” Proc., Int. Conf. on Foundation Failures, Singapore, T. W.
load level. Even a modest softening of the soil at or near the pile Hulme and Y. S. Lau, eds., Inst. of Engs. Singapore, 121–130.
base can lead to significant reductions in pile head stiffness. Abdrabbo, F. M., and Abouseeda, H. 共2002兲. “Effect of construction pro-
Structural defects may lead to only modest reductions in single cedures on the performance of bored piles.” Deep Foundations 2002,
pile stiffness, as long as structural failure of the pile does not ASCE Geot. Spec. Pub. No. 116, Vol. 2, ASCE, Reston, Va., 1438–
occur. Structural failure of a pile is more abrupt than for a pile 1454.
with geotechnical defects. Amirsoleymani, T. 共1991a兲. “Elimination of excessive differential settle-
Within a pile group, the ability of the stiffer undamaged piles ment by different methods.” Proc., 9th Asian Reg. Conf. Soil Mechan-
in a group to carry additional load reduces the potential conse- ics Foundation Engineering, Bangkok, 1, 351–354.
quences of imperfections and defective piles, as compared with Amirsoleymani, T. 共1991b兲. “Removing soil layers under foundations to
an isolated pile. However, careful consideration needs to be given rotate buildings”. Proc., 9th Asian Reg. Conf. Soil Mechanichs Foun-
to the additional bending moments induced in the piles by the dation Engineering, Bangkok, 1, 201–204.
presence of the defective piles. A critical aspect of the group Brandl, H. 共1989兲. “Underpinning.” Special Lecture D, Proc., 12th Int.
response is that the presence of defective piles can result in in- Conf. Soil Mechanics Foundation Engineering, Rio de Janeiro,
duced lateral deflection and cap rotation of the group, and addi- Balkema, Rotterdam, 4, 2227–2258.
tional bending moments in the piles. This induced lateral response Brinkgreve, R. B. J., and Vermeer, P. A. 共1998兲. PLAXIS finite element
can become more severe as the location of the defective piles code for soil and rock analyses, Balkema, Rotterdam.
Brown, D. A. 共2004兲. “Zen and the art of drilled shaft construction: The
becomes more asymmetric.
pursuit of quality.” Proc. Geosupport 2004, J. P. Turner and P. W.
It is highly desirable that methods be developed to account for
Mayne, eds., ASCE Geot. Spec. Pub. No. 124, ASCE, Reston, Va.,
such imperfections so that more realistic designs can be executed,
19–33.
especially if remedial design is necessary. Remedial design poses Bruce, D. A. 共1994兲. “Small-diameter cast-in-place elements for load-
interesting and important challenges to the foundation designer in bearing and in situ earth reinforcement.” Ground control and improve-
that both the realities and the imperfections inherent in real prob- ment, P. P. Xanthakos, L. W. Abramson, and D. A. Bruce, eds., Wiley
lems need to be properly taken into account. Factors such as the Interscience, New York.
addition of new pile into a foundation system after loading has Burland, J. B. 共2004兲. “The leaning tower of Pisa revisited.” State of the
commenced, or the removal of piles, also need to be given proper Art and Practice Paper No. 2, Proc., 5th Int. Conf. Case Histories in
consideration in order to obtain realistic assessments of the sub- Geotechnical Engineering, New York, S. Prakash, ed., Univ. of Mis-
sequent behavior of the foundation system. It has been demon- souri, Rolla, CD volume.
strated that the load distribution within a pile group can be con- Button, S. J. 共1953兲. “The bearing capacity of footings on a two layer
trolled by the use of controlled stiffness inserts 共CSIs兲 at the pile cohesive subsoil.” Proc., 3rd Int. Conf. Soil Mechanics Foundation
head. The reduction of unevenness of load within a group gener- Engineering, Zurich, 1, 332–335.

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005 / 561

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 Chen, L., Poulos, H. G., and Loganathan, N. 共1999兲. “Pile responses ney region.” Austral. Geomechs., G8, 31–39.
caused by tunnelling.” J. Geotech. Geoenviron. Eng., 125共3兲, 207– Pells, P. J. N., Mostyn, G., and Walker, B. F. 共1998兲. “Foundations on
215. sandstone and shale in the Sydney region.” Austral. Geomechs.,
Decourt, L. 共1995兲. “Prediction of load-settlement relationships for foun- 33共3兲, 17–29.
dations on the basis of the SPT-T.” Ciclo de Conferencias Intern. Petek, K., Felice, C. W., and Holtz, R. D. 共2002兲. “Capacity analysis of
“Leonardo Zeevaert,” UNAM, Mexico, 85–104. drilled shafts with defects.” Deep Foundations 2002, Geot. Spec. Pub.
Giroud, J-P., Vo-Nhiem, T., and Obin, J. P.共1973兲. Tables pour le calcul No. 116, Vol. 2, ASCE, Reston, Va., 1120–1135.
des foundations. Tome 3, Dunod, Paris 共in French兲. Popa, A., Lacatus, F., Rebeleanu, V., and Taria, O. 共2001兲. “Underpinning
Golder, H. Q., and Osler, J. C. 共1968兲. “Settlement of a furnace founda- of buildings by means of jet grouted piles.” Proc., 15th Int. Conf. Soil
tion, Sorel, Quebec.” Can. Geotech. J., 5共1兲, 46–56. Mechanics Geotechnical Engineering, Istanbul, 3, 1835–1838.
Hayward Baker. 共2003兲. Project summary on warehouse remediation, Poulos, H. G. 共1988兲. “Modified calculation of pile group settlement in-
promotional literature. teraction.” J. Geotech. Eng., 114共6兲, 697–706.
Hobbs, N. B. 共1957兲. “Unusual necking of cast-in-situ concrete piles.” Poulos, H. G. 共1989兲. “Pile behaviour—Theory and application.” Geo-
Proc. 4th Int. Conf. Soil Mechanics Foundation Engineering, London, technique, 39共3兲, 65–415.
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

3, 40–42. Poulos, H. G. 共1993兲. “Settlement of bored pile groups.” Proc., BAP II,
Hoit, M. I., and McVay, M. C. 共1996兲. FLPIER User’s Manual, Univer- Ghent, Balkema, Rotterdam, 103–117.
sity of Florida, Gainesville, Fla. Poulos, H. G. 共1997兲. “Behaviour of pile groups with defective piles.”
Itasca. 共1999兲. FLAC Users’Manual, Itasca Consulting Group Inc., Min- Proc., 14th Int. Conf. Soil Mechanics Foundation Engineering, Ham-
nesota. burg, 2, 871–876.
Jamiolkowski, M. 共2001兲. “The leaning tower of Pisa: End of an odys- Poulos, H. G. 共1999a兲. “Approximate computer analysis of pile groups
sey.” Terzaghi Oration, Proc., 15th Int. Conf. Soil Mechanics Foun- subjected to loads and ground movements.” Int. J. Numer. Analyt.
dation Engineering, Istanbul, 4, 2979–2996. Meth. Geomech., 23, 1021–1041.
Kulhawy, F. H. 共1978兲. “Geomechanical model for rock foundation Poulos, H. G. 共1999b兲. “Pile defects—Influence on foundation perfor-
settlement.” J. Geotech. Eng. Div., Am. Soc. Civ. Eng., 104共2兲, 211– mance.” Keynote Lecture, Proc., 4th Int. Conf. on Deep Foundations,
227. Singapore, CI Premier, 57–69.
Kulhawy, F. H., and Carter, J. P. 共1992兲. “Settlement and bearing capacity Poulos, H. G. 共2000兲. “Pile testing—From the designer’s viewpoint.”
of foundations on rock masses.” Engineering in Rock Masses, F. G. Keynote Lecture, Statnamic Loading Test ’98, O. Kusakabe, F. Kuwa-
Bell, ed., Butterworth Heinemann, London, 231-245. bara, and T. Matsumoto, eds., Balkema, Rotterdam, 3–21.
Liu, J. 共2004兲. “Grouting and dewatering in balancing settlement of a Poulos, H. G. 共2001兲. “Piled raft foundations—Design and applications.”
building.” Proc., 5th Int. Conf. Case Histories, New York, S. Prakash, Geotechnique, 51共2兲, 95–113.
ed., CD volume, Paper No. 1.15. Poulos, H. G. 共2002兲. “Prediction of behaviour of piled building founda-
Maffei, C. E., Goncalves, H. H. S., Pimenta, P. M., and Murakami, C. A. tions due to tunnelling operations.” Proc., 3rd Int. Symp. on Geotech-
共2001兲. “The plumbing of 2.2° inclined tall building.” Proc., 15th Int. nical Aspects of Tunnelling in Soft Ground, Toulouse, Preprint Vol-
Conf. Soil Mechanics Geotechnical Engineering, Istanbul, 3, 1799– ume, 4.55–4.61.
1802. Poulos, H. G. 共2003a兲. “The significance of ground characterisation for
Makarchian, M., and Poulos, H. G. 共1996兲. “Simplified method for design foundation deformation prediction.” Geotech. Eng., 3, 125–145.
of underpinning piles.” J. Geotech. Eng., 122共9兲, 745–751. Poulos, H. G. 共2003b兲. “Analysis of soil extraction for correcting uneven
Matsui, T. 共1993兲. “Case studies on cast-in-place bored piles and some settlement of pile foundations.” Proc., 12th Asian Reg. Conf. Soil
considerations for design.” Proc. BAP II, Ghent, Balkema, Rotterdam, Mechanics Geotechnical Engineering, Singapore, C. F. Leung et al.,
77–102. eds., 1, 653–656.
Merifield, R. S., Sloan, S. W., and Yu, H. S. 共1999兲. “Rigorous plasticity Poulos, H. G. 共2004兲. “Control of settlement and load distribution in pile
solutions for the bearing capacity of two-layered clays.”Geotechnique groups via stiffness inserts.” Proc., Symposium on Recent Develop-
49共4兲, 471–490. ments in Foundation Practice, Center for Research and Professional
Meyerhof, G. G. 共1976兲. “Bearing capacity and settlement of pile foun- Development, Hong Kong.
dations.” J. Geotech. Eng. Div., Am. Soc. Civ. Eng., 102共3兲, 195–228. Poulos, H. G., Badelow, F., and Powell, G. 共2003兲. “A theoretical study of
Meyerhof, G. G., and Sastry, V. V. R. N. 共1978兲. “Bearing capacity of constructive application of excavation for foundation correction.”
piles in layered soils: Part I & Part II.” Can. Geotech. J., 15共2兲, Proc. Int. Conf. on Response of Buildings to Excavation-Induced
171–189. Ground Movements, London, CIRIA Spec. Pub. 199, F. M. Jardine,
Milligan, V. 共1997兲. “Lessons from foundation failures.” Proc., Int. Conf. ed., 469–484.
on Foundation Failures, Singapore, T. W. Hulme and Y. S. Lau, eds., Poulos, H. G., Carter, J. P., and Small, J. C. 共2002兲. “Foundations and
43–52. retaining structures—Research and practice.” State of the Art Lecture,
Moh, Z. C. 共1994兲. “Current deep foundation practice in Taiwan and Proc., 15th Int. Conf. Soil Mechanics Foundation Engineering, Istan-
Southeast Asia.” Proc., Int. Conf. on Design and Construction of bul, 4, 2527–2606.
Deep Foundations, Orlando, FHWA, 1, 236–259. Randolph, M. F. 共1994兲. “Design methods for pile groups and piled rafts.”
Mroueh, H., and Shahrour, I. 共2002兲. “Three-dimensional finite element Proc., 13th Int. Conf. S.M. & Foundation Engineering, 5, 61–82.
analysis of the interaction between tunneling and pile foundation.” Randolph, M. F. 共2003兲. PIGLET. Analysis and design of pile groups,
Int. J. Numer. Analyt. Meth. Geomech., 26, 217–230. Users’ manual, University of Western Australia, Perth, Australia.
O’Neill, M. W. 共2001兲. “Side resistance in piles and drilled shafts.” 34th Rao, N. 共1996兲. “Defective piles in clay.” Proc., 6th Int. Conf. & Exhi-
Terzaghi Lecture, J. Geotech. Geoenviron. Eng., 127共1兲, 1-16. bition on Piling & Deep Foundations, Bombay, paper 4.4.
O’Neill, M. W., and Hassan, K. M. 共1994兲. “Drilled shafts: Effects of Reese, L. C. 共1978兲. “Design and construction of drilled shafts.” J. Geo-
construction on performance and design criteria.” Proc., Int. Conf. tech. Eng. Div., Am. Soc. Civ. Eng., 104共1兲, 95–116.
Des. Constr. Deep Founds., Orlando, FHWA, 1, 137–187. Sarhan, H. A., and O’Neill, M. W. 共2002兲. “Aspects of structural design
O’Neill, M. W., Hawkins, R. A., and Mahar, L. J. 共1982兲. “Load transfer of drilled shafts for flexure”. Deep Foundations 2002, Geot. Spec.
mechanisms in piles and pile groups.” Geotech. Eng., 108共12兲, 1605– Pub. No. 116, Vol. 2, ASCE, Reston, Va., 1151–1165.
1623. Sarhan, H., Tabsh, S. W., O’Neill, M. W., Ata, A., and Ealy, C. 共2002兲.
Pells, P. J. N., Douglas, D. J., Rodway, B., Thorne, C. P., and McMahon, “Flexural behavior of drilled shafts with minor flaws.” Deep Founda-
B. R. 共1978兲. “Design loadings for shales and sandstones in the Syd- tions 2002, ASCE Geot. Spec. Pub. No. 116, Vol. 2, ASCE, Reston,

562 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005

J. Geotech. Geoenviron. Eng. 2005.131:538-563.

 Va., 1136–1150. sibility of compensation grouting of timber pile foundations to miti-
Tamez, E., Santoyo, E., and Ovando, E. 共1997兲. “Underexcavation of gate TBM settlements.” Proc. Int. Conf. on Response of Buildings to
Mexico City’s Metropolitan Cathedral and Sagrarion Church.” Proc., Excavation-Induced Ground Movements, London, CIRIA Spec. Pub.
14th Int. Conf. Soil Mechanics Geotechnical Engineering, Hamburg, 199, F. M. Jardine, ed., 613–620.
Balkema, Rotterdam, 4, 2105–2126. van Impe, W. F. 共1991兲. “Deformations of deep foundations.” Proc. 10th
Teparaksa, W., Thasninapan, N., and Anwar, M. A. 共1999兲. “Base grout- European Conf. SM & Foundation Engineering, Florence, 3, 1031–
ing of wet process bored piles in Bangkok subsoils.” Proc., 11th Asian 1062.
Reg. Conf. Soil Mechanics & Geotechnical Engineering, Seoul, 1, Woo, S. M., and Moh, Z-C. 共1990兲. “Geotechnical characteristics of soils
269–272. in the Taipei basin.” Proc., 10th Southeast Asian Geotechnical Conf.,
Terzaghi, K. 共1936兲. “Presidential Address.” Proc., 1st Int. Cong. Soil Taipei, 2, 51–65.
Mechanics Foundation Engineering, Cambridge, Mass., 3, 13–18. Xu, K. J., and P oulos, H. G. 共2000a兲. “Measured and predicted axial
Terzaghi, K. 共1951兲. “The influence of modern soil studies on the design response of piles with diameter discontinuities.” Geotech. Eng.,
and construction of foundations.” Building Res. Congress, London, 31共3兲, 171–191.
Div. 1, Part III, 139–145. Xu, K. J., and Poulos, H. G. 共2000b兲. “General elastic analysis of piles
Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 01/22/15. Copyright ASCE. For personal use only; all rights reserved.

Thorburn, S., and Thorburn, J. Q. 共1977兲. “Review of problems associ- and pile groups.” Int. J. Numer. Analyt. Meth. Geomech., 24, 1109–
ated with construction of cast-in-place concrete piles.” Report of DOE 1138.
and CIRIA Piling Devel. Group, CIRIA, London. Xu, K. J., and Poulos, H. G. 共2001兲.“A general load-settlement analysis of
Tsai, K. W., Chao, C. S., and Chou, K. T. 共1991兲. “Tilted high-rise build- pile groups.” Proc., 5th Int. Conf. on Deep Foundation Practice In-
ing corrected by grouting.” Proc., 11th European Conf. Soil Mechan- corporating Piletalk, Singapore, CI Premier, Singapore.
ics Foundation Engineering, Florence, 2, 623–624. Zou, Y. 共1996兲. “Ein neues verfahren zum aufrichten geneigter gebaude.”
van der Stoel, A. E. C., Haasnoot, J. K., and Essler, R. D. 共2003兲. “Fea- Bautechnik, 73共7兲, 437–442.

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MAY 2005 / 563

J. Geotech. Geoenviron. Eng. 2005.131:538-563.