Tunnelling and Underground Space Technology 45 (2015) 43–51

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology

journal homepage: www.elsevier.com/locate/tust

An analytical study of tunnel–pile interaction

Alec M. Marshall ⇑, Twana Haji
Department of Civil Engineering, Faculty of Engineering, University of Nottingham, University Park, Nottingham NG7 2RD, United Kingdom

a r t i c l e i n f o a b s t r a c t

Article history: Tunnelling under the foundations of structures is becoming more common because of the lack of avail-
Received 7 December 2013 able space for infrastructure, both above and below ground. The interaction between newly constructed
Received in revised form 8 August 2014 tunnels and existing piled foundations is an important issue. This paper presents results obtained using a
Accepted 1 September 2014
computationally efficient analytical approach which aims to estimate the effect that constructing a new
Available online 30 September 2014
tunnel will have on an existing pile. The method uses a spherical cavity expansion analysis to evaluate the
end-bearing capacity of the pile, and cylindrical cavity contraction to estimate the decrease in the
Keywords:
confining pressure and resulting reduction in pile capacity caused by tunnel volume loss. The paper
Tunnel
Pile
extends previously published work using this method by considering the effect of tunnel location on
Soil–structure interaction the tunnel–pile interactions, examining different possible assumptions of soil stiffness used in the anal-
Cavity expansion ysis, and by considering the effect that tunnel cavity contraction has on the friction along the pile shaft.
Ó 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).

1. Introduction site conditions, however in general it increases with depth due to

the additional cost of construction of associated excavations (e.g.
The need for effective Civil infrastructure in cities is paramount. shafts for tunnel boring machine launch, ventilation, and escala-
As populations grow and the demand on infrastructure systems tors/lifts). A careful decision must therefore be made at the design
increases, the need to further develop already congested under- stage with respect to the optimum depth for new tunnel
ground space in many urban areas will become unavoidable. This construction.
will result in new underground construction activities taking place Piled foundations are particularly sensitive to the effects of tun-
ever closer to existing structures and buried infrastructure. The nelling. Piles risk a reduction of their end-bearing capacity and
resulting interaction between construction activities and the shaft friction resistance due to the displacements and ground
affected Civil assets must be considered in the design process. stress redistributions that occur as a result of tunnelling. A variety
Tunnels are arguably the most popular medium to large-scale of research has been conducted on the subject, ranging from field
underground structures in crowded urban areas. They are used to studies (Kaalberg et al., 2005; Pang et al., 2006; Selemetas et al.,
minimise the volume of traffic on the surface and can also have 2006), experimental work (Bezuijen and Van der Schrier, 1994;
environmental benefits (e.g. traffic noise reduction). Tunnelling Loganathan et al., 2000; Jacobsz et al., 2004; Marshall and Mair,
inevitably causes some ground movements which can have detri- 2011), and numerical or analytical modelling (Chen et al., 1999;
mental effects on buried and above-ground infrastructure and Kitiyodom et al., 2005; Lee and Ng, 2005; Cheng et al., 2007). Ana-
buildings. There has been considerable research conducted on lytical methods provide an efficient way for studying soil–struc-
the subject of evaluating the shape of tunnelling induced ground ture interaction problems such as the effect of tunnelling on piles.
movements (Peck, 1969; O’Reilly and New, 1982; Mair et al., This paper studies the effect of a newly constructed tunnel on
1993; Marshall et al., 2012) and in determining the effects these an existing pile using the analytical approach introduced by
movements have on man-made assets (Attewell et al., 1986; Klar Marshall (2012, 2013). Some critical points from the original
et al., 2005; Vorster et al., 2005; Klar et al., 2008; Marshall et al., method are examined and some new ideas are presented which
2010; Zhang et al., 2012). In general, the potential for harmful are intended to achieve a more sensible and thorough analysis
interaction between tunnel construction and Civil assets is greatest approach. The paper includes data obtained using the original
for shallow tunnels, which suggests that a deeper tunnel is prefer- method of Marshall (2012) to elucidate the important effect that
able. The cost of tunnelling varies considerably depending on local depth and the relative horizontal and vertical distance between
the pile and tunnel have on results. Next, the selection and influ-
⇑ Corresponding author. ence of soil stiffness used in the tunnel–pile interaction analysis
E-mail address: [email protected] (A.M. Marshall).
are illustrated, and a new method for accounting for the effect of

http://dx.doi.org/10.1016/j.tust.2014.09.001
0886-7798/Ó 2014 The Authors. Published by Elsevier Ltd.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).
44 A.M. Marshall, T. Haji / Tunnelling and Underground Space Technology 45 (2015) 43–51

Nomenclature

a current radius of a cavity Q0 total load capacity of pile before tunnel volume loss
a0 original cavity radius Q Vl total load capacity of pile after tunnel volume loss
C cohesion rp pile radius
c current radius of the plastic zone around a pile or a tun- rt tunnel radius
nel Rqb pile end-bearing capacity reduction factor
c0 original distance from pile tip to elastic–plastic interface RQ pile capacity reduction factor
c1 parameter used to calculate G0 RQ;S pile capacity reduction factor including effect on pile
Dp pile diameter shaft
dtp distance from tunnel axis to pile tip S parameter used to calculate G0
dlp distance from tunnel lining to pile tip St ratio of radial effective stress near pile tip at failure to qb
E Young’s modulus Vl volume loss due to tunnelling, in %
G soil shear modulus z depth to any point below the ground surface
G0 small strain shear stiffness zp depth to pile tip
G0; mod modified shear stiffness due to the effect of pile instal- zt depth of tunnel axis
lation ac parameter used in calculation of qb
G0; tun shear stiffness calculated at tunnel depth bs ratio of shaft shear stress to vertical effective stress of
Id relative density soil
IR relative dilatancy index bmin , bmax minimum and maximum values of bs
K0 the coefficient of at-rest lateral earth pressure ds soil-shaft friction angle
k cavity expansion parameter: spherical k = 2; cylindrical h parameter used in calculation of ac
k¼1 / soil friction angle
L embedded pile length /0cv critical state friction angle
n parameter used to calculate G0 c soil unit weight
Nq bearing capacity factor ls a parameter to calculate bs
p0 mean effective stress or confining pressure m Poisson’s ratio
p00 initial isotropic stress at tunnel or pile tip r0 normal effective stress
p00; mod modified pressure r0r radial stress
p00; pile confining pressure at pile tip r0er radial stress in elastic zone
p00; tun confining pressure at tunnel depth r0pr radial stress in plastic zone
p0mid confining pressure half-way between pile tip and tunnel r0v vertical stress
lining r0h circumferential stress
p0V l confining pressure after tunnel volume loss r0eh circumferential stress in elastic zone
pa atmospheric pressure (100 kPa) r0ph circumferential stress in plastic zone
p0lim limiting stress for spherical cavity expansion ss shaft shear stress
P cavity pressure ss average shear stress on pile shaft
Pa current cavity pressure when cavity radius = a ss;0 average shear stress on pile shaft before tunnel volume
qb end-bearing capacity of pile loss
qb;0 end-bearing capacity of pile before tunnel volume loss ss;V l average shear stress on pile shaft after tunnel volume
qb;V l reduced end-bearing capacity of pile after tunnel vol- loss
ume loss w soil dilation angle
Q total load capacity of the pile

pile installation on soil stiffness is presented that gives a more sen-

sible approach than the original method from Marshall (2012). A
method for estimating the effect of tunnel cavity contraction on
pile shaft friction is also proposed. Finally, data obtained using
the original analysis method are compared against new results,
and a recommendation is provided, based on analysis of available
geotechnical centrifuge experiment data, on how to obtain a
conservative evaluation of tunnel–pile separation or safe tunnel
volume loss in order to avoid large pile displacements.

2. Cavity expansion methods

Fig. 1 shows the problem that is considered in this paper and

the main geometric parameters considered. A tunnel of radius rt
is constructed beneath the tip of an existing pile with radius rp .
Distance d is measured along the path connecting the pile tip to
the tunnel axis. The shortest distance from the centreline of the
pile to the axis of the tunnel is given by dtp ; the distance from pile
to the tunnel lining is given by dlp . The paper focuses on driven or
jacked piles which cause a significant impact on ground stresses
around the pile tip. The analysis could, however, be applied to Fig. 1. View of the analysis problem.
 A.M. Marshall, T. Haji / Tunnelling and Underground Space Technology 45 (2015) 43–51 45

bored piles in which the effect of pile installation on the ground is the stresses within the ground between the tunnel and the pile.
minimal. The volume loss induced by tunnelling affects the pile (4) The reduction in pile end-bearing capacity is evaluated based
and may reduce its load-carrying capacity. The analysis aims to on the altered stress conditions within the ground (due to (3)) at
evaluate the distance between the pile and the tunnel, dtp , that is the tip of the pile. Stage (3) of the analysis incorporates an estima-
required to ensure the pile does not suffer a significant reduction tion of the effect of pile installation on soil stiffness; this aspect of
of its capacity (which could result in large pile displacements). the analysis is examined closely in this paper.
The cavity expansion method has been used for the study of a The method involves superposition of results from two separate
wide variety of geotechnical problems since its early application elastic–plastic analyses, and therefore can only be regarded as pro-
to pressuremeter test interpretation by Gibson and Anderson viding an approximation of the real interaction between the tunnel
(1961). These include the study of in situ soil testing (Salgado and pile. Also, it is assumed that the installation of the pile has lit-
and Prezzi, 2007; Mo et al., 2014), deep foundations (Randolph tle effect on the confining stress at the location of the tunnel. This
et al., 1994), and tunnels and underground excavations (Mair and is likely to be adequate for most practical scenarios however it
Taylor, 1993; Yu and Rowe, 1999). Yu (2000) provided a thorough should be recognised as a feature of the method. A summary of
review of the method and its various applications. the analysis and the relevant equations is provided as Appendix
The adopted analysis considers an enlarging or contracting cav- A. Readers should also refer to Marshall (2012, 2013) for further
ity of initial radius a0 in an infinite soil mass, as illustrated in Fig. 2. details. This type of analysis could be used to evaluate the effect
The cavity can be either cylindrical or spherical in shape. Various of tunnelling on bored piles (where pile installation has little effect
assumptions can be applied in the method; the description here on the ground) by omitting stage (2) of the analysis. In this case, if
applies to the analysis undertaken by Marshall (2012). The soil is an estimation of the effect of tunnelling on pile shaft friction is
assumed to be elastic–perfectly plastic with a non-associated required then an appropriate relationship between radial and ver-
Mohr–Coulomb yield criterion. Prior to the formation of the cavity, tical stress along the pile shaft should be adopted, as described by
isotropic stress conditions are assumed and given by p00 . Initially, Fleming et al. (1992).
the cavity pressure P is equal to this isotropic stress. As the cavity
pressure increases to Pa and the cavity expands to radius a, a plas-
4. Effect of tunnel and pile depth
tic zone forms around it that extends to radius c from the cavity
centre. The yielded soil mass is surrounded by elastic soil that
The results presented in Marshall (2012) illustrate the sensitiv-
extends to infinity. The radial and circumferential stresses within
ity of the analysis to the tunnel–pile separation, given by dtp . The
the ground are given by r0r and r0h , respectively. The cavity contrac-
analysis results are also sensitive to the depth at which the tunnel
tion problem, where the cavity size decreases, can be considered in
and pile are located; the variation of results with dtp will vary
a similar manner.
depending on this depth. To illustrate this feature of the analysis,
the cases presented in Fig. 3 are examined using the original
3. Basic analysis procedure Marshall (2012) analysis method. In Case 1, the relative tunnel–
pile separation is increased by moving the tunnel laterally away
The analyses undertaken as part of this work followed the gen- from the pile so that the depth of the pile and tunnel remain con-
eral approach set out in Marshall (2012, 2013). The cavity expan- stant but dtp increases. In Case 2, the pile tip depth remains con-
sion analysis for the interaction between tunnel construction and stant and the tunnel–pile separation is increased by increasing
an existing driven or jacked pile can be summarised as follows. the depth of the tunnel. In Case 3, the tunnel–pile separation is
The analysis consists of 4 stages. (1) The end-bearing capacity of kept constant while the depth of both the tunnel and pile are
the pile is evaluated following the method of Randolph et al. increased. All cases considered the following parameters: tunnel
(1994) whereby a spherical cavity expansion analysis is used to radius rt = 3 m, pile radius r p = 0.5 m, tunnel volume loss V l = 5%,
evaluate the limiting cavity pressure, p0lim , and the end-bearing critical state friction angle /0cv = 30 , soil unit weight c = 18 kN/
pressure of the pile, qb . (2) The change in ground stress around m3, relative density Id = 0.8, at rest earth pressure coefficient K 0
the installed pile and the effect of pile installation on the ground = 0.5, Poisson’s ratio m = 0.2. All other parameters, including soil
stress profile is evaluated from the spherical cavity expansion anal- stiffness, friction angle, and dilation angle were determined using
ysis in (1). (3) A cylindrical cavity contraction analysis is used to the methods outlined in Marshall (2012, 2013) and in Appendix A.
evaluate the effect of tunnel volume loss (cavity contraction) on The results of the Marshall (2012) analysis of the 3 cases are
presented in Fig. 4. The results relate to the pile end-bearing capac-
ity reduction factor, Rqb , which was defined by Marshall (2012) as:
qb;V l
Rqb ¼ ð1Þ
qb;0

where qb is the end-bearing capacity of the pile calculated in stage

(1) of the analysis and qb;V l is the reduced end-bearing capacity
caused by tunnel cavity contraction calculated in stage (4). Note
that a value of Rqb = 1 indicates that the tunnel has no effect on
the pile end-bearing capacity; a lower value of Rqb indicates that
the tunnel causes a reduction in the pile end bearing capacity. Com-
paring Case 1 with Case 2, for a given tunnel–pile separation, Case 1
results in a lower value of Rqb than Case 2. The effect of moving the
tunnel deeper in Case 2 has a beneficial effect on the interaction
analysis. This beneficial effect of depth is also demonstrated by con-
sidering Case 3, where the tunnel–pile separation is kept constant
but the depth of both the tunnel and pile are increased. Fig. 4 shows
that as depth increases for Case 3, the value of Rqb increases. This
Fig. 2. Cavity expansion in yielding soil. beneficial effect of depth on the results of the analysis is due to
46 A.M. Marshall, T. Haji / Tunnelling and Underground Space Technology 45 (2015) 43–51

Fig. 3. Cases considered in evaluating the effect of tunnel and pile depth.

Fig. 5. Variation of confining pressure and modified stiffness with tunnel–pile

separation.
Fig. 4. Results of analysis of cases presented in Fig. 3.

pile installation (based on p0lim from the cavity expansion analysis),

the increase in the isotropic confining pressure, p00 , which is based normalised by the original value of p00 at the tip of the pile. The dis-
on the depth of the tunnel and pile tip at different stages of the tance, d, is measured from the location of the pile tip along the path
analysis. The method calculates a friction and dilation angle based of dlp (refer to Fig. 1); d is normalised by dlp to indicate the relative
on the relative dilatancy term, IR , (Bolton, 1986, 1987) which is a distance between the pile tip and the tunnel lining. The minimum
function of the confining pressure (friction and dilation angles value of d=dlp plotted in Fig. 5 corresponds to a value just greater
increase with an increase in p00 ). The shear stiffness of the soil, G, than the radius of the pile (which is the final radius of the cavity
also increases with confining stress in the analysis, but this has considered in the expansion analysis). The data plotted in Fig. 5
the effect of decreasing the value of Rqb . The increase in strength considers Case 2 from Fig. 3 with zp = 15 m and ztp = 10 m. The
and dilation with p00 evidently has a greater effect on the resulting mean effective stress reduces at an exponential rate and the
value of Rqb than the effect of increased stiffness. The results in normalised value reaches unity at the plastic–elastic interface.
Fig. 4 illustrate the importance of considering the specific geometry For the parameters considered, this occurs at d=dlp = 0.7.
of the interaction problem when using the Marshall (2012) analysis Also plotted in Fig. 5 is the modified value of small strain shear
approach. stiffness, G0; mod , used in Marshall (2012), calculated using the value
of p0 illustrated in Fig. 5 at a given distance d from the pile tip, and
5. Effect of modified soil stiffness normalised by the value of G0 at the depth of the tunnel (G0; tun ).
The G0; mod term was used in the original analysis to account for
The small strain shear stiffness of the soil, G0 , is used in the pile the effect of pile installation on soil stiffness between the tunnel
end-bearing capacity analysis as a representative value of soil stiff- and the pile. Due to the way that G0; mod was calculated, the norma-
ness (Randolph et al., 1994) (see Appendix A). In order to evaluate lised term does not go to unity if it is calculated at the plastic–elas-
G0 , the mean effective stress or confining pressure, p0 , is needed. tic interface. In the Marshall (2012) method, G0; mod was calculated
Fig. 5 shows the variation of mean effective stress in the soil after based on a modified mean effective stress: p00; mod ¼ p00; tun þ p0mid ,
 A.M. Marshall, T. Haji / Tunnelling and Underground Space Technology 45 (2015) 43–51 47

where p00; tun is the isotropic confining pressure at the depth of the and the tunnel lining (d=dlp = 0.5). Fig. 6b shows that for the case
tunnel and p0mid is the value of p0 due to pile installation at a dis- considered, there is very little difference between the values of
tance equivalent to half-way between the pile tip and the tunnel Rqb between the two methods of modifying G0 (based on d=dlp =
lining (d=dlp = 0.5 in Fig. 5). 0.5 in Fig. 6b). In Fig. 6b, the ’new method’ is noted to agree with
There may be a more sensible and conservative approach to the unmodified method at the plastic–elastic interface, which
calculate G0; mod than that proposed in Marshall (2012). The seems more reasonable than the Marshall (2012) trend. The use
Marshall (2012) method added the value of p0 at the mid-point of this ’new method’ of evaluating soil stiffness to determine the
between the pile tip and the tunnel lining to the value of confining effect of tunnelling on pile capacity is compared against the
stress at the depth of the tunnel axis in order to evaluate p00; mod and original results of Marshall (2012) in a later section.
G0; mod in stage (3) of the analysis. It is suggested here that a more
rational approach is to normalise the calculated confining pressure
at the mid-point using the initial confining pressure at the pile tip 6. The effect of tunnel volume loss on pile shaft friction
(p00; pile Þ and then to factor the confining pressure at the tunnel axis
by this value. This ‘new method’ of calculating p00; mod is represented The original analysis of Marshall (2012) did not consider the
by Eq. (2) and results in a more rational trend of G0; mod =G0; tun , as effect of tunnelling on shaft friction but did include the
illustrated in Fig. 6a where the value for this ‘new method’ is 1.0 contribution of shaft friction to the total pile capacity. The b-method
when G0; mod is calculated based on p0 at the plastic–elastic interface described in Randolph et al. (1994) was used to estimate the distri-
(the Marshall (2012) line from Fig. 5 is presented again in Fig. 6a bution of shear stress along the pile shaft, as defined by Eq. (3).
for comparison). ss  
bs ðzÞ ¼ ¼ bmin þ ðbmax  bmin Þ exp ls ðL  zÞ=Dp ð3Þ
r0v
p0
p00; mod ¼ 0 mid  p00; tun ð2Þ
p0; pile where bmin = 0.2, bmax = St N q tanðds Þ; N q ¼ qb =r0v (r0v is vertical effec-
tive stress at pile tip), St ¼ 2exp½7tanð/0cv Þ; Dp = pile diameter, ls =
In Fig. 6b, the values of Rqb calculated based on the modified 0.05, qb is calculated using the cavity expansion analysis outlined in
stiffness values in Fig. 6a are compared against results obtained Randolph et al. (1994), L is the embedded pile length, and z is mea-
when no modification to G0 is made (i.e. G0 is calculated based sured from the surface. The profile of shear stress (ss ) can be
on the confining pressure at the depth of the tunnel axis). The integrated along the pile length in order to calculate the total shaft
’no modification’ line represents an example of results which could friction contribution to the pile capacity. It should be noted that Eq.
apply to bored piles in which pile installation does not have a sig- (3) is based on driven piles. If this type of analysis were to be used
nificant effect on the in situ soil conditions. In Marshall (2012), the for bored piles then a modified profile of horizontal stress along the
value of p00; mod is calculated at the mid-point between the pile tip pile length should be adopted (refer to Fleming et al. (1992) for
guidance on this topic).
A pile capacity reduction factor, RQ , was defined in Marshall
(2012) as:

Q V l qb;V l Dp þ 4ss;0 L
RQ ¼ ¼ ð4Þ
Q0 qb;0 Dp þ 4ss;0 L

where ss is the equivalent average shear stress which provides the

same total shaft load as the distributed shaft shear stress, and the
subscripts 0 and V l indicate the initial and post tunnel volume loss
values, respectively. Note that the ss;0 term is included in the
numerator, indicating that the value of Q V l does not include for
the effect of tunnel volume loss on shaft friction.
A method for evaluating the effect of tunnelling on shaft friction
using the cavity contraction analysis is now proposed. In stage (3)
of the basic analysis procedure, a distribution of radial and circum-
ferential stresses within an initially isotropic stress field is calcu-
lated. These stresses are used to evaluate the change in mean
effective stress at the location of the pile tip so that a reduced
end-bearing capacity of the pile can be calculated in stage (4). In
a similar fashion, the change in mean effective stress along the
length of the pile axis may be used to estimate the effect of tunnel
volume loss on pile shaft friction. Fig. 7 shows contours of mean
effective stress after tunnel volume loss (p0V l ) normalised by the ini-
tial isotropic stress calculated at the depth of the tunnel axis
(p00;tun ). The plotted data were obtained using the same parameters
as the cases in Fig. 4 but with zp = 15 m, ztp = 10 m, and using the
’new method’ for evaluating p00; mod (Eq. (2)). Two cases are shown,
where the lateral offset between the tunnel and pile are (a) 0 m
and (b) 15 m. Fig. 7b shows that the length of the pile falls outside
the elastic–plastic interface where there is no change in the value
of mean effective stress (p0V l =p00;tun ¼ 1). In Fig. 7a, however, a large
portion of the pile shaft is shown to be within the zone where the
value of mean effective stress is reduced due to the tunnel cavity
Fig. 6. Comparison of results for (a) modified soil stiffness and (b) Rqb . contraction.
48 A.M. Marshall, T. Haji / Tunnelling and Underground Space Technology 45 (2015) 43–51

Fig. 8. Profiles along pile length: (a) p0V l =p00;tun and (b) bs .

The data shows that as the lateral offset between the tunnel and
pile increases, the length of the pile affected by the tunnel
decreases; the bs profile is unaffected when the offset is 15 m.
The modified profile of bs can be used to determine a new distribu-
tion of shear stress along the pile shaft using Eq. (3). The contribu-
tion of pile shaft shear stress to the total pile capacity after tunnel
volume loss, Q V l , can then be calculated. A new pile capacity reduc-
Fig. 7. Contours of p0V l =p00;tun . tion factor which accounts for the effect of the tunnel cavity con-
traction on both pile end-bearing capacity and shaft friction is
defined by Eq. (5). The term ss;V l that now appears in the numerator
(compared to Eq. (4)) accounts for the effect of tunnel cavity con-
As discussed in Randolph et al. (1994), the bs function essen-
traction on shaft friction. The following section compares results
tially describes a profile of horizontal effective stress along the pile
obtained using this method for evaluating the effect of tunnelling
length. A given change in horizontal stress will therefore result in a
on pile capacity against the original method from Marshall (2012).
proportional change in the value of bs . It is suggested here that in
order to obtain an estimate of the effect of tunnel cavity contrac- Q V l qb;V l Dp þ 4ss;V l L
tion on the pile shaft shear stress, the profile of bs may be scaled RQ ;S ¼ ¼ ð5Þ
Q0 qb;0 Dp þ 4ss;0 L
by the ratio of p0V l =p00;tun along the pile axis (as illustrated in
Fig. 7). This is an approximation since it involves the superposition
of the mean effective stresses from the isotropic tunnel cavity 7. Comparison of results with Marshall (2012)
contraction analysis on the original horizontal effective stress state
along the length of the pile, which would realistically not be isotro- This section compares analysis results from the original
pic (K 0 is assumed to be 0.5 in the current analysis). Marshall (2012) method against those obtained by considering
The profiles of p0V l =p00;tun and bs for different values of tunnel–pile the ’new method’ for evaluating p00;mod (Eq. (2)) as well as the effect
offsets are shown in Fig. 8 (all other parameters are consistent with of tunnelling on shaft friction (Eq. (5)). The geotechnical centrifuge
cases from Fig. 7). The data in Fig. 8a illustrate the sections of the experiments of tunnel–pile interaction provided by Marshall
pile in which mean effective stress is affected by the tunnel cavity (2009) and Jacobsz (2002) which were analysed in Marshall
contraction (i.e inside the elastic–plastic interface where (2012) are considered again here. Fig. 9 shows the results obtained
p0V l =p00;tun < 1). When the original profile of bs is factored by the from the three analyses. Note that the shaft analysis (RQ ;S ) also
values in Fig. 8a, the new profiles shown in Fig. 8b are obtained. incorporates the ’new method’ for evaluating p00;mod . Each set of
 A.M. Marshall, T. Haji / Tunnelling and Underground Space Technology 45 (2015) 43–51 49

was also proposed. Results were compared against the experimen-

tal data used in Marshall (2012) and it was again found that a value
of RQ = 0.85 is a conservative threshold for determining the safe
tunnel volume loss or relative tunnel–pile separation.
The analytical approach presented makes various
simplifications with regard to the real tunnel–pile interaction
problem. The analysis is aimed at providing a quick and relatively
straightforward method for evaluating if tunnel construction is
likely to have significant adverse effects on existing piles. The eval-
uation of a safe value of RQ or RQ ;S is based on data at pile failure
(i.e. significant pile displacements initiated). The analysis does
not attempt to evaluate the pre-failure displacements that would
occur within the soil or to the pile. Whilst these limitations of
Fig. 9. Comparison of results with Marshall (2012).
the analysis should not be ignored, it is felt that the approach does
provide a valuable tool for tunnel–pile interaction analysis.

Acknowledgements
data points represents the analysis of an individual centrifuge
experiment and includes an input of material properties,
The authors would like to thank Mr. Tiago Dias and Professor
geometrical conditions, and the known tunnel volume loss at which
Adam Bezuijen of the University of Ghent for their insightful sug-
the pile failed (defined as the moment when the rate of change of
gestions during the preparation of this paper.
the pile displacement showed a significant increase). Including
the ’new method’ is shown to give a slightly more conservative
(lower) evaluation of RQ than the original method (maximum value Appendix A. Summary of analytical tunnel–pile interaction
of RQ reduced from 0.78 to 0.75). As illustrated in Fig. 6, the ’new method
method’ tends to give a higher value of soil stiffness, resulting in
a lower prediction of base capacity and therefore a lower evaluation This appendix provides a summary of the analytical tunnel–pile
of RQ . Likewise, incorporating the effect of shaft friction (and using interaction analysis presented in Marshall (2012, 2013). The cavity
the ’new method’) gives an even lower result, where RQ ;S < RQ and expansion analyses are based on Yu (2000), with common param-
the maximum value of RQ ;S is 0.72. Fig. 8 showed that the shaft anal- eter definitions provided in Eq. (A.12).
ysis only reduces the value of bs therefore resulting in a lower value
of Q V l and reducing the value of RQ ;S below that of RQ . (1) Predict the end-bearing pressure, qb , using spherical cavity
The conclusion from Marshall (2012) was that, based on the expansion analysis to evaluate p0lim in Eq. (A.1) (Randolph
experimental data available and the analysis proposed, a value of et al., 1994) where ac ¼ h þ /0cv =2 and h = 45 or the
RQ = 0.85 provides a conservative approximation of a safe volume penetrometer cone tip angle.
loss or tunnel–pile separation to avoid pile stability issues and
   
qb ¼ p0lim 1 þ tan /0cv tan ðac Þ ðA:1Þ
potentially large displacements. This conclusion is still valid when
adopting the ’new method’ proposed here since RQ = 0.85 is an The limit pressure, p0lim , is found using Eqs. (A.3)–(A.5) by
even more conservative threshold for this analysis. Including the varying the value of p0lim in Eq. (A.3) until the left and right
effect of tunnelling on shaft friction in the analysis provides an sides of Eq. (A.4) are equal. Soil stiffness may be evaluated
even more conservative evaluation of pile capacity reduction. The using a variety of methods; in this analysis the small strain
proposed shaft analysis method involves an approximation shear stiffness, G0 , was used based on the method suggested
whereby the change in mean effective stresses from the tunnel by Randolph et al. (1994):
cavity contraction analysis are used to evaluate the change in hor-  0 n
G0 p
izontal stresses along the pile shaft. Given the higher level of con- ¼ S expðc1 Id Þ 0 ðA:2Þ
pa pa
servatism obtained, it is suggested that the validity of this
approximation should be evaluated by any individual conducting where S = 600, c1 = 0.7, n = 0.43 (Lo Presti, 1987), and pa is
such an analysis, with appropriate consideration of the various atmospheric pressure (100 kPa).
project-specific conditions and risks.  
ðk þ aÞ Y þ ða  1Þp0lim
Rlim ¼   ðA:3Þ
að1 þ kÞ Y þ ða  1Þp00
8. Conclusions X
1
v bþk
An ðRlim ; lÞ ¼ ð1  dÞ b ðA:4Þ
n¼0
c
This paper deals with the problem of tunnel–pile interaction ( ln
and presented results obtained using analytical cavity expansion/ n!
ln R if n ¼ c
contraction methods. The analysis aims to provide an efficient An ðR; lÞ ¼ ln ðA:5Þ
n!ðncÞ
ðRnc  1Þ otherwise
means of assessing the effect of a newly constructed tunnel on
an existing pile. The results presented were obtained using an anal- (2) Assuming an infinite ground mass with an isotropic stress p00
ysis which generally followed the approach set out by Marshall based on the pile tip depth, and using the determined value
(2012, 2013), as summarised in Appendix A. The paper illustrated of p0lim from stage 1, use spherical cavity expansion to calcu-
the importance of considering the specific geometry of each case late the location of the elastic–plastic interface, c (Eq. (A.6)),
due to the sensitivity of results to the depth of the pile and tunnel. and distribution of confining stress, p0 , resulting from pile
A new method of evaluating the soil stiffness and modified confin- installation, where p0 ¼ ðr0r þ 2r0h Þ=3. Ground stresses result-
ing stress used in the analysis was proposed. The new method ing from pile installation, where the superscripts e and p
gives a more rational approach since the value of G0; mod =G0; tun goes refer to elastic and plastic, respectively, are given by Eq.
to unity at the plastic–elastic interface. A method for approximat- (A.7) (refer also to Fig. 2).
ing the effect of tunnel cavity contraction on pile shaft shear stress
50 A.M. Marshall, T. Haji / Tunnelling and Underground Space Technology 45 (2015) 43–51

( ) a
E E 2C cos / 1 þ sin /
kða1Þ
c ðk þ aÞ½Y þ ða  1Þp G¼ M¼ Y¼ a¼
¼   ðA:6Þ 2ð1 þ mÞ 1  m2 ð2  kÞ 1  sin / 1  sin /
a0 að1 þ kÞ Y þ ða  1Þp00
1 þ sin w aðb þ kÞ Y þ ða  1Þp00
b¼ c¼ d¼
1  sin w kða  1Þb 2ðk þ aÞG
Plastic zone : r < c   
ð1 þ kÞd 1  m2 ð2  kÞ kmða þ bÞ
r0pr ¼ a1
Y
þ Ar 
kða1Þ
l¼ ab þ kð1  2mÞ þ 2m 
a
ð1 þ mÞða  1Þb 1  mð2  kÞ
 
kða1Þ ðb þ kÞð1  2mÞð1 þ ð2  kÞmÞ Y þ ða  1Þp00
r0ph ¼ a1
Y
þ Aa r  a v ¼ exp
Eða  1Þb
ð1þkÞa½Yþða1Þp00  kða1Þ ðA:12Þ
A2 ¼  ða1ÞðkþaÞ
c a
ðA:7Þ
Elastic zone : r > c

r0er ¼ p00  Brð1þkÞ

References
r0eh ¼ p00 þ Bk rð1þkÞ
Attewell, P.B., Yeates, J., Selby, A.R., 1986. Soil Movements Induced by Tunnelling
k½Yþða1Þp00  and Their Effects on Pipelines and Structures. Blackie and Son Ltd, UK.
B2 ¼ kþa
c1þk Bezuijen, A., Van der Schrier, J., 1994. The influence of a bored tunnel on pile
foundations. In: Leung, C.F., Lee, F.H., Tan, T.S. (Eds.), Centrifuge ’94: Proceedings
(3) Use a cylindrical cavity contraction analysis to evaluate the of the international conference. Balkema, Singapore, pp. 681–686.
Bolton, M.D., 1986. The strength and dilatancy of sands. Geotechnique 36 (1), 65–
effect of tunnel volume loss on ground stresses. An isotropic
78.
stress is assumed based on the depth of the tunnel axis. A Bolton, M.D., 1987. Discussion on The strength and dilatancy of sands.
modified (increased) value of shear stiffness, Gmod (based Geotechnique 37 (2), 219–226.
Chen, L.T., Poulos, H.G., Loganathan, N., 1999. Pile responses caused by tunneling. J.
on p0mod ), is used to account for the effect of pile installation
Geotech. Geoenviron. Eng. 125 (3), 207–215.
on the soil between the tunnel and pile. For an assumed Cheng, C.Y., Dasari, G.R., Chow, Y.K., Leung, C.F., 2007. Finite element analysis of
value of volume loss, V l , and assuming concentric contrac- tunnel–soil–pile interaction using displacement controlled model. Tunn.
tion of the tunnel p boundary, the final tunnel radius can be Undergr. Space Technol. 22 (4), 450–466.
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Fleming, W.G.K., Weltman, A.J., Randolph, M.F., Elson, W.K., 1992. Piling
found using a ¼ a0 ð1  V l =100Þ. For cases where a plastic Engineering, 2nd editio edn. Blackier and Son Ltd, Glasgow.
zone forms around the tunnel (most relevant scenario), the Gibson, R., Anderson, W., 1961. In situ measurement of soil properties with the
cavity pressure P when r ¼ a is required. Using Eq. (A.8) to pressuremeter. Civil Eng. Pub. Works Rev. 56 (658), 615–618.
Jacobsz, S.W., 2002. The effects of tunnelling on piled foundations.
determine c0 =c, the value of P in Eq. (A.9) is iterated until Jacobsz, S.W., Standing, J.R., Mair, R.J., Hagiwara, T., Sugiyama, T., 2004. Centrifuge
the desired value of a is obtained. Eq. (A.10) is then used modelling of tunnelling near driven piles. Soils Found. 44 (1), 49–56.
to evaluate c and Eq. (A.11) is used to determine the ground Kaalberg, F.J., Teunissen, E.A.H., van Tol, A.F., Bosch, J.W., 2005. Dutch research on
the impact of shield tunnelling on pile foundations. In: Bakker, K.J., Bezuijen, A.,
stresses. Broere, W., Kwast, E.A. (Eds.), 5th International Symposium on Geotechnical
Aspects of Underground Construction in Soft Ground, Amsterdam, The
c0 ð1  aÞp00 þ Y
¼1 ðA:8Þ Netherlands, pp. 123–131.
c 2ð1 þ kaÞGmod Kitiyodom, P., Matsumoto, T., Kawaguchi, K., 2005. A simplified analysis method for
piled raft foundations subjected to ground movements induced by tunnelling.
h  1þkb i1þkb
1 Int. J. Numer. Anal. Meth. Geomech. 29 (15), 1485–1507, ISSN 1096-9853.
a0
a
¼ 1  X 1  cc0 Klar, A., Vorster, T.E.B., Soga, K., Mair, R.J., 2005. Soil-pipe interaction due to
ðA:9Þ tunnelling: comparison between winkler and elastic continuum solutions.
n o 1þkb Geotechnique 55 (6), 461–466.
ð1þkaÞ½Yþða1ÞP kð1aÞ
X ¼ ð1þkÞ½Yþða1Þp00  Klar, A., Marshall, A.M., Soga, K., Mair, R.J., 2008. Tunneling effects on jointed
pipelines. Can. Geotech. J. 45 (1), 131–139.
Lee, G.T.K., Ng, C.W.W., 2005. Effects of advancing open face tunneling on an
1
c ð1 þ kaÞ½Y þ ða  1ÞP kð1aÞ existing loaded pile. J. Geotech. Geoenviron. Eng. 131 (2), 193–201.
¼ ðA:10Þ Lo Presti, D., 1987. Mechanical behaviour of Ticino sand from resonant column tests.
a ð1 þ kÞ½Y þ ða  1Þp00  Ph.D. thesis. Politecnico di Torino.
Loganathan, N., Poulos, H.G., Stewart, D.P., 2000. Centrifuge model testing of
tunnelling-induced ground and pile deformations. Geotechnique 50 (3), 283–
Plastic zone : r < c
294.
r0pr ¼ a1
Y
þ Ar kða1Þ Mair, R.J., Taylor, R.N., 1993. Prediction of clay behaviour around tunnels using
plasticity solutions. In: Predictive Soil Mechanics – Proceedings of the Wroth
r0ph ¼ a1
Y
þ Aar kða1Þ Memorial Symposium. Thomas Telford, Oxford, UK, pp. 449–463.
ð1þkÞ½Yþða1Þp00  Mair, R.J., Taylor, R.N., Bracegirdle, A., 1993. Subsurface settlement profiles above
A3 ¼  ða1Þð1þkaÞ
cð1aÞk tunnels in clays. Geotechnique 43 (2), 315–320.
ðA:11Þ Marshall, A.M., 2009. Tunnelling in sand and its effect on pipelines and piles. Phd
Elastic zone : r > c
thesis. Cambridge University.
ð1þkÞ
r0er ¼ p0  Br Marshall, A.M., 2012. Tunnel–pile interaction analysis using cavity expansion
methods. ASCE J. Geotech. Geoenviron. Eng. 138 (10), 1237–1246.
r0eh ¼ p0 þ Bk rð1þkÞ Marshall, A.M., 2013. Closure to tunnel–pile interaction analysis using cavity
k½ð1aÞp00 Y  expansion methods by Alec M. Marshall. ASCE J. Geotech. Geoenviron. Eng. 139
B3 ¼ 1þka
c1þk (11), 2002–2004.
Marshall, A.M., Mair, R.J., 2011. Tunneling beneath driven or jacked end-bearing
(4) A reduced end-bearing capacity based on the change in piles in sand. Can. Geotech. J. 48 (12), 1757–1771.
Marshall, A.M., Klar, A., Mair, R.J., 2010. Tunneling beneath buried pipes – a view of
stresses at the base of the pile due to tunnel volume loss soil strain and its effect on pipeline behavior. ASCE J. Geotech. Geoenviron. Eng.
(Dr0r ; Dr0h ) from stage 3 is calculated using the same process 136 (12), 1664–1672.
as in stage 1. The confining stress at the pile tip is factored by Marshall, A.M., Farrell, R.P., Klar, A., Mair, R.J., 2012. Tunnels in sands the effect of
size, depth, and volume loss on greenfield displacements. Geotechnique 62 (5),
Rp ¼ 1  Dp0 =p00; tun , where Dp0 ¼ ½ð1 þ mÞðDr0r þ Dr0h Þ=3.
385–399.
Mo, P.Q., Marshall, A.M., Yu, H.S., 2014. Elastic–plastic solutions for expanding
Common parameters used for cavity expansion and contraction cavities embedded in two different cohesive-frictional materials. Int. J. Numer.
analyses: Anal. Method. Geomech. 38, 961–977.
 A.M. Marshall, T. Haji / Tunnelling and Underground Space Technology 45 (2015) 43–51 51

O’Reilly, M.P., New, B.M., 1982. Settlements above tunnels in the United Kingdom - Salgado, R., Prezzi, M., 2007. Computation of cavity expansion pressure and
their magnitude and prediction. In: Tunnelling ’82, Papers Presented at the 3rd penetration resistance in sands. Int. J. Geomech. 7 (4), 251–265.
International Symposium., Inst of Mining and Metallurgy, London, England, Selemetas, D., Standing, J.R., Mair, R.J., 2006. The response of full-scale piles to
Brighton, England, pp. 173–181. tunnelling. In: Bakker, K.J., Bezuijen, A., Broere, W., Kwast, E.A. (Eds.), 5th
Pang, C.H., Yong, K.Y., Chow, Y.K., Wang, J., 2006. The response of pile foundations International Symposium on Geotechnical Aspects of Underground
subjected to shield tunnelling. In: Bakker, K.J., Bezuijen, A., Broere, W., Kwast, Construction in Soft Ground. Taylor & Francis, pp. 763–769.
E.A. (Eds.), 5th International Symposium on Geotechnical Aspects of Vorster, T.E.B., Klar, A., Soga, K., Mair, R.J., 2005. Estimating the effects of tunneling
Underground Construction in Soft Ground. Taylor & Francis, pp. 737–743, on existing pipelines. J. Geotechn. Geoenvironm. Eng. 131 (11), 1399–1410.
ISBN 978-0-415-39124-5. Yu, H.S., 2000. Cavity expansion methods in geomechanics. Kluwer Academic
Peck, R.B., 1969. Deep excavations and tunnelling in soft ground. In: 7th Publishers.
International Conference on Soil Mechanics and Foundation Engineering, Yu, H.S., Rowe, R.K., 1999. Plasticity solutions for soil behaviour around contracting
Mexico City, pp. 225–290. cavities and tunnels. Int. J. Numer. Anal. Meth. Geomech. 23 (12), 1245–1279.
Randolph, M.F., Dolwin, J., Beck, R., 1994. Design of driven piles in sand. Zhang, C., Yu, J., Huang, M., 2012. Effects of tunnelling on existing pipelines in
Geotechnique 44 (3), 427–448. layered soils. Comput. Geotech. 43 (0), 12–25, ISSN 0266-352X.