Tunnelling and Underground Space Technology 48 (2015) 1119

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology

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

Analysis of lateral earth pressure on a vertical circular shaft considering

the 3D arching effect
Jaeyeon Cho a, Hyunsung Lim b, Sangseom Jeong b,, Kyoungyul Kim c,1
a
Department of Engineering, University of Cambridge, Trumpington Street, Cambridge CB2 1PZ, United Kingdom
b
Department of Civil and Environment Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul 120-749, South Korea
c
Korea Electric Power Research Institute, 105 Munji-ro, Yuseong-gu, Daejeon 305-760, South Korea

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

Article history: The lateral earth pressure on a vertical circular shaft is investigated using both experiments and
Received 20 March 2014 numerical analyses. The study focused on quantifying the magnitude and distribution of the lateral earth
Received in revised form 2 September 2014 pressure, which was measured by considering the three-dimensional arching effect. A framework for
Accepted 4 January 2015
determining distribution of the earth pressure based on centrifuge model tests and 2D FE analysis is
introduced. The FE modelling techniques and the constitutive relationships of the soil are presented in
detail. Parametric analyses showed that the arching effect on the lateral earth pressure is highly depen-
Keywords:
dent on the diameter and height of the shaft, the internal friction angle and the cohesion value of the soil,
Vertical circular shaft
Arching effect
the end-bearing conditions and the exural modulus of the shaft. The study found that when the arching
Earth pressure effect is considered, the lateral earth pressure on a vertical circular shaft is approximately 80% less than
Centrifuge model test that calculated using Rankines theory. The study also found that the arching effect of the soil is more
Numerical analysis signicant for exible vertical shafts than for rigid vertical shafts embedded in weathered soil.
2015 Elsevier Ltd. All rights reserved.

1. Introduction distribution is a result of arching effects (Handy, 1983), but these

Underground space and structures are under great demand.
Since 1980, many large excavation projects have been undertaken
in urban areas to create underground space for power stations,
subways, high-speed railways, and other lifelines (Jeong and Seo,
2004). Although the shape and size of underground structures vary
widely, they can generally be classied into two groups. The rst
group includes long or wide underground structures (horizontal
tunnels), and the other includes tall vertical underground struc-
tures (vertical shafts). These structures are deeper and larger than
typical existing foundations and tunnels. For this reason, a proper
estimation of the earth pressures is a key factor in the design of
underground structures, particularly vertical shafts.
Geotechnical engineers have traditionally estimated the earth
pressure that acts on structures using either Rankines (1857) or
Coulombs (1773) theories. Both theories assume that the distribu-
tion of lateral earth pressure that acts on a structure is triangular.
However, this is different from reality and these methods do not
consider the mode of structural movement. Many experimental
studies have shown that the non-linearity of the earth pressure
Corresponding author. Tel.: +82 2 2123 2807; fax: +82 2 364 5300.
E-mail addresses: soj9081@yonsei.ac.kr (S. Jeong), patric@kepco.co.kr (K. Kim).
1
Principal researcher.
methods are widely used in engineering practice with no sig-
nicant modications.
Arching is the process by which stress is transferred around a
region of soil, which then is subjected to lower stresses (Paik and
Salgado, 2003). Many studies have investigated the lateral earth
pressure that acts on rigid retaining walls by considering the arch-
ing effect. Handy (1983) analyzed soil arching behind retaining
walls, and Wang and Yen (1973) performed this analysis for slopes.
Nakai et al. (1997) performed a series of physical model tests under
1 g conditions and carried out numerical analyses of these tests to
investigate the arching effect. They results obtained from the model
tests agreed well with the results of the numerical analyses. Janssen
(1895) suggested a theoretical basis for understanding arching
effects in silos. Based on this theory Spangler and Handy (1984),
Harrop-Williams (1989) and Wang (2000) proposed equations
recently to estimate the non-linear distribution of active pressure
on retaining walls. More recently, Kim et al. (2013) conducted cen-
trifuge model tests and full-scale eld tests of vertical shafts and
concluded that the lateral earth pressure that acts on a circular ver-
tical shaft is less than other types of geotechnical structures because
of three dimensional arching effects (i.e., convex arching and/or
inverted arching). However, no practical solutions that consider
shaft displacements and movements of the surrounding soil are cur-
rently available for the determination of the lateral earth pressure.

http://dx.doi.org/10.1016/j.tust.2015.01.002
0886-7798/ 2015 Elsevier Ltd. All rights reserved.
12 J. Cho et al. / Tunnelling and Underground Space Technology 48 (2015) 1119
(a) Model vertical shaft
(b) Soil sample box
Fig. 1. Testing apparatus.
In engineering practice, the displacements are often controlled or
limited by the choice of a suitable factor of safety, and excessive
yielding is prevented by an appropriate construction sequence
(Wong and Kaiser, 1988).
The performance of the active earth pressure distribution
caused by the installation of a circular vertical shaft has been
investigated by several researchers using experimental or numeri-
cal analyses. Walz (1973) investigated the active earth pressure on
a circular vertical shaft using a model shaft equipped with a cutting
edge ring. Lade et al. (1981) and Konig et al. (1991) performed phy-
sical modelling in which the excavated soil was replaced by a ex-
ible rubber bag lled with liquid or gas. The liquid or gas pressure
was reduced in stages to simulate the excavation of the shaft. Addi-
tionally, a mechanical system was used to move the vertical shaft
to simulate the displacement of soil that may occur during the
excavation process. Several researchers (e.g., Fujii et al., 1994;
Imamura et al., 1999) have adopted simplied models to simulate
(a) Sectional view
(b) Vertical circular shaft model
Fig. 2. Sectional view of the soil sample box and vertical shaft model.
the radial displacement of the lining. Herten and Pulsfort (1999)
and Chun and Shin (2006) modied previous models by consider-
ing radial symmetry to model only a portion of the structure. More
recently, Tobar and Meguid (2009) developed a mechanical system
model both the full shaft geometry and the radial displacement of
the lining. In contrast, Herten and Pulsfort (1999) conducted
numerical analyses to simulate the construction of a laboratory
scale shaft in granular material using the discrete element method
(DEM). Tran et al. (2014) performed numerical studies to charac-
terize large deformations and particle movements. Relatively little
work has addressed the distribution of lateral earth pressure that
acts on exible and circular vertical shafts. And the effect of the
staged construction of a vertical shaft on the lateral pressure along
the shaft and the deformation of the shaft has not been investigat-
ed previously.
The objective of this study is to investigate the magnitude and
distribution of lateral earth pressure on a vertical circular shaft
according to the excavation sequence by considering the three-
dimensional arching effect in weathered soil. A series of centrifuge
model tests and FE analyses of a vertical shaft were performed. To
obtain enough information to cover a wide spectrum of vertical
shaft designs, comprehensive parametric studies were conducted
by varying the combinations of several key design factors. To verify
the accuracy of the numerical analyses, the results of the
experimental measurements from this study are compared with
those from the numerical analyses.
2. Centrifuge model test
2.1. Testing apparatus
A series of centrifuge model tests were performed to investigate
the distribution of earth pressure on a vertical circular shaft in
 J. Cho et al. / Tunnelling and Underground Space Technology 48 (2015) 1119 13

Table 1
Properties of the test soil.

c / Water G Dr D10 Cu Gs
(kN/m3) (deg.) content (kPa) (%) (mm)
(%)
Silica sand (SP) 12.66 36.95 2.0 75 8182 0.091 1.53 2.65

sand. The geotechnical centrifuge used in this study was located at

Daewoo Institute of Construction Technology (DICT). DICTs
geotechnical is an Actidyn system C65-2 and was manufactured
by Actidyn in France. The centrifuge has a platform radius of
3.0 m and a maximum capacity of 120 g-ton, and a centrifugal
acceleration factor of N = 75 g was used.
The model vertical shaft was made of aluminum alloy with a
hollow circular section and had an embedded height of 200 mm,
and an outer diameter of 80 mm (Fig. 1a). The model was equiva-
lent to a 15 m high vertical reinforced concrete shaft with an outer
diameter of 6 m. To investigate the lateral earth pressure at each
Fig. 4. Comparison of earth pressure distributions.
excavation stage, the test model was divided into two sections. A
small soil sample box 700  400  200 mm in size was constructed
with a steel frame (Fig. 1b).
A variety of instruments was installed to monitor the displace-
ments and earth pressures during the testing. The displacement of
the wall of the vertical shaft was measured by two linear variable
differential transformers (LVDTs) located on both sides of the shaft
surface as shown in Fig. 2(a). Pre-calibrated Kyowa earth pressure
transducers were installed on the external surface of the vertical
cylinder wall to measure the lateral earth pressures that acted on
the vertical shaft; their locations in the model are shown in
Fig. 2(b). Washers were installed between the earth pressure trans-
ducers and the wall to reduce the error of the earth pressure.

2.2. Test soil

Silica sand classied as SP (according to the Unied Soil Classi-

cation System) with a specic gravity (Gs) of 2.65, an effective
grain size D10 = 0.091 mm and a uniformity coefcient Cu = 1.53
was used for the centrifuge tests. Unit weight tests were performed
using the water content, volume and weight of silica sand. The unit
weight of the silica sand was 12.6 kN/m3. The properties of the test
soils are summarized in Table 1.
(a) Upper section A dry sand sample with a relative density Dr of approximately
8182% (dense state) was prepared by the air pluvial method using
an automatic sand-rainer. The spreader was passed repeatedly
over the soil sample box until the sand layer was approximately
200 mm thick (equivalent to 15 m at the prototype scale). Addi-
tionally, aluminum-cans were installed in the soil box to investi-
gate the homogeneous soil. After the centrifuge tests, the unit
weight and water content of the silica sand were measured accord-
ing to soil depth.

2.3. Testing program

Centrifuge model tests of the lateral earth pressure on the ver-

tical shaft during excavation were performed on 1/75th scale mod-
els. After sample preparation, the sample box was mounted on the
platform of the centrifuge. A two-stage construction model system
was used to simulate a realistic excavation sequence for the verti-
cal circular shaft. In this model, the rst excavation is made and is
followed by the second excavation. First, the upper section moves
inward by a homogeneous radial displacement to simulate the rst
(b) Lower section excavation after the model was tested at a radial acceleration of
75 g. The lower section then moved in the same manner to simu-
Fig. 3. Earth pressure distributions. late the second excavation, which was simulated using a motor-
14 J. Cho et al. / Tunnelling and Underground Space Technology 48 (2015) 1119
Fig. 5. 2D axisymmetric FE model for a vertical shaft (D = 6 m and H = 15 m).
gear reducer system. The incremental active wall displacement
(dh/H, dh: horizontal displacement, H: height of the vertical shaft)
was 7.14  103 mm/s. Terzaghi (1920) suggested that the active
earth pressure occurs at an active wall displacement of
approximately 0.1%H, which is equivalent to a wall displacement
of 15 mm in this test. The diameters of the upper and lower vertical
shaft were gradually decreased in sequence, and the earth pressure
meters were monitored simultaneously.
2.4. Test results and discussion
Typical results from the centrifuge model tests are presented
here. The data are presented in terms of the prototype units scaled
up according to the relevant scaling relationships (Schoeld, 1980).
Fig. 3 shows the measured earth pressures under the conditions
of active wall displacement for a typical test. As shown in Fig. 3(a),
the lateral earth pressure on the upper section of the vertical shaft
(AP1-AP4) gradually decreases during the rst excavation step
until reaching a wall displacement of approximately 0.2  102.
This curve demonstrated the reductions in lateral earth pressure
with a gradual increase of wall displacement. When the active wall
displacement (dh/H: 0.2  102) is reached, the earth pressure was
approximately 30% less than the initial earth pressure (AP4). At this
time, the earth pressure on the lower section of the vertical shaft
(AP5AP8) was constant. During the second excavation step, large
decrease in the earth pressure on the lower section of the vertical
shaft occurred as shown in Fig. 3(b). When the active wall displace-
ment (dh/H: 0.2  102) is reached, the earth pressure was
approximately 80% (AP8) less than the initial earth pressure. The
earth pressure on the upper part of the vertical shaft (AP1AP4)
increased as the displacement of the lower part increased. Further-
more, the lateral earth pressure increased by approximately
1030% with a displacement from 0 to 0.2. A similar earth pressure
to the initial conditions (before the rst excavation) was measured
when the active wall displacement increased from 0 to 0.6.
The test results also showed that the lateral earth pressure at
deep depths decreased with increasing active displacement
(dh/H: 0.2  102) and then increased toward the initial conditions
as the displacement increased. This result is similar to that
described by Tatiana and Mohamed (2010) for the active displace-
ment of a wall. However, because the test vertical shaft was exca-
vated to a relatively shallow depth, there was very little decrease of
earth pressure. This difference in behavior between deep and shal-
low depths occurs because the redistribution of weight caused by
the arching effect will reduce the pressure more at deep depths
than at shallow depths (Kim et al., 2013).
Fig. 4 compares the test results with the results from Rankines
theory. All of the measured earth pressures are lower than the the-
oretical results. These results veried that the overall lateral earth
pressure on the vertical circular shaft was approximately 80% less
than the initial conditions and the theoretical solution. This is
because the Rankine theory is based on two-dimensional plane
strain and ignores the three-dimensional arching effect; thus, it
overestimates the lateral earth pressure with increasing depth.
This clearly demonstrates that a three-dimensional arching effect
occurs with vertical circular shafts, so this set of predictions and
measured results shows the reduction of lateral earth pressure.
3. Finite element analysis of a vertical shaft
A major objective of the numerical analysis was to investigate
the lateral earth pressure that acts on a vertical shaft in weathered
soil under controlled conditions. The main benet of the numerical
analysis versus laboratory testing was the ability to more closely
control the test conditions. Thus, a numerical method that can
accurately estimate the lateral earth pressure while considering
the 3-D arching effect is needed to analyze the vertical shafts.
The excavation procedure is considered by the numerical method.
Table 2
Summary of material parameters used for validation.
Type c E / c m Rinter Model
(kN/m3) (MPa) (deg.) (kPa)
Silica sand 12.6 10 36.95 4.50 0.3 0.67 M.C
Vertical shaft (Aluminum) 27 72,000 0.2 1.0 L.E
 J. Cho et al. / Tunnelling and Underground Space Technology 48 (2015) 1119 15

Table 4
Material properties used in the analyses (parametric studies).

Type c E / c m Rinter Model

(kN/m3) (MPa) (deg.) (kPa)
Weathered soil 19 50 35 15 0.32 0.7 M.C
Weathered rock 21 300 39 34 0.3 1.0
Soft rock 23 500 40 50 0.25 1.0
Vertical shaft 5 250 0.4 1.0 L.E
Depth (m)

elements, and the soil consists of 15-node triangle elements. The

model provides a fourth order interpolation for displacements
and the numerical integration involves twelve Gauss points (stress
points). The interfaces are composed of 10-node interface elements
with ve pairs of nodes, which are compatible with the 15-node
soil elements.
The model extends a distance ten times the diameter of the ver-
tical shaft (D) from the center of the shaft and has a total height
equal to the height of the vertical shaft (H) plus an additional
(a) Upper section
0.7H below the level of the shaft toe. These dimensions were ade-
quate to eliminate the inuence of boundary effects on the perfor-
mance of the vertical shaft. A relatively ne mesh was used near the
interface between the vertical shaft and the surrounding soil, and
Depth (m)

(b) Lower section

Fig. 6. Comparison of the results (dh/H: 0.2  102).

Table 3 (a) Lateral earth pressure distribution

Summary of numerical analyses conducted.

Parameters Case
Vertical Diameter (D) D = 3, 6, 9 and 12 m
shaft (H = 15 m)
Height (H) H = 10, 15 and 20 m
(D = 6 m)
Elastic modulus (E) E = 250, 1500, 3000, 5000, 10,000 and
28,000 MPa
(D = 6 m, H = 15 m)
Soil Internal friction angle / = 10, 20, 30 and 40
(/)
Cohesion (c) c = 10, 20, 40, 60 and 80 kPa
Bearing layer Weathered soil
Weathered rock
Soft rock

3.1. FE mesh and boundary conditions

Two dimensional (2D) axisymmetric conditions are used to

model the vertical shaft as a cylinder and the surrounding soils
as concentric hollow cylinders. The commercial FE computer
program PLAXIS 2D (2012) is used in the numerical analysis. (b) Shaft displacement
Fig. 5 shows the typical 2D axisymmetric FE model used in this
study. The vertical shaft was modelled using 5-node plate Fig. 7. Effect of shaft diameter.
16 J. Cho et al. / Tunnelling and Underground Space Technology 48 (2015) 1119
(a) Lateral earth pressure distribution
(b) Shaft displacement
Fig. 8. Effect of shaft height.
the mesh became coarser farther from the vertical shaft. The typical
vertical shaft had a diameter D of 6.0 m and a height H of 15 m. The
vertical shaft extends through the layer of weathered soil, and the
bearing end is located on weathered rock (or soft rock). The mesh
consists of a total of 39,261 fteen-node triangular elements.
The vertical boundaries are allowed to move only in the vertical
direction, and the bottom boundary is xed in the horizontal and
vertical directions, which assumes a stiff undeformable stratum
such as a rock layer. The specied initial stress distributions should
match a calculation based on the weight of the material. After the
initial step, excavation of the soil was simulated by removing the
soil in lifts. The soil was removed in 3, 4 and 6 steps for the 10,
15 and 20 m vertical shafts, respectively. The vertical shaft was
installed when the excavation was completed.
3.2. Constitutive model and material parameters
The vertical shaft is modelled as linear-elastic material, while a
MohrCoulomb non-associated ow rule is adopted for the sur-
rounding soil layer. MohrCoulomb constitutive models of soils
have commonly been used in FE modelling of retaining structures
(Potts and Fourie, 1984; Day and Potts, 1993; Grande et al., 2002;
(a) Lateral earth pressure distribution
(b) Shaft displacement
Fig. 9. Effect of internal friction angle.
Jeong and Seo, 2004; Karlsrud and Andresen, 2005; Krabbenhoft
et al., 2005; Tan and Paikowsky, 2008), and were also used to simu-
late the arching response of soil layers in this study.
The shaft-soil interface is modelled with an interface element
using the bilinear MohrCoulomb model (Kim and Jeong, 2011).
The interface element is treated as a zone of virtual thickness. It
behaves as an element with the same material properties as the
adjacent soil elements before slip occurs. A lower value of shear
modulus is assigned to the interface element when slip occurs on
it. The decrease of the strength of the interface element is repre-
sented by a strength reduction factor Rinter in PLAXIS. The interface
properties are calculated from following equation:
cinter Rinter csoil ; tan /inter Rinter tan /soil 1
where cinter and /inter are the cohesion and friction angle of the
interface, respectively, and csoil and /soil are the cohesion and fric-
tion angle of the soil mass, respectively.
3.3. Validation with centrifuge model tests
This section discusses the validation of the numerical modelling
technique using the results of the FE analysis for the centrifuge
model tests. The properties of the vertical shaft and soil used in
 J. Cho et al. / Tunnelling and Underground Space Technology 48 (2015) 1119
(a) Lateral earth press
(b) Shaft displacement
Fig. 10. Effect of cohesion.
the centrifuge model test, which were described in Section 2.1,
were employed for this validation. The material properties of the
silica sand and the vertical shaft are summarized in Table 2. A
2D axisymmetric model was also used, and lateral displacements
(dh/H: 0.2  102) were applied to simulate the excavation, as
described in Section 2.3.
Representative results of the FE analysis and centrifuge model
tests are shown in Fig. 6. The results show that the numerical pre-
dictions match the experimental measurements quite well for each
section of the vertical shaft. However, the computed lateral earth
pressures near the upper section of the vertical shaft are slightly
less than the experimental results because the FE numerical model
(i.e., continuum analysis) did not consider the separation between
the vertical shaft and the soil that occurred in the experiments.
4. Parametric studies
An extensive parametric study performed to investigate the
inuence of the arching behavior on the vertical shaft that could
not be constrained in the centrifuge model tests. A series of FE ana-
lyses on a vertical shaft in weathered soil was performed using the
inuential parameters, including the shaft diameter (D), the shaft
17
(a) Lateral earth pressure distribution
(b) Shaft displacement
Fig. 11. Effect of bearing layer.
height (H), the internal friction angle of the soil (/), the cohesion
of the soil (c), the end-bearing layer and the elastic modulus of
the shaft (E), which are summarized in Table 3. The material prop-
erties used in the FE analyses are shown in Table 4.
4.1. Effect of shaft diameter
Fig. 7(a) shows the lateral earth pressure versus the soil depth
for several shaft diameters (D = 3.0, 6.0, 9.0 and 12.0 m) and the
same shaft height (H = 15 m). The soil depth was normalized by
the shaft height. As shown in the gure, the distribution of the lat-
eral earth pressure is not triangular, as is proposed by Rankines
theory (1857). The results show that as the shaft diameter increas-
es, the lateral earth pressure in the upper zone of the shaft increas-
es. This is because the shaft diameter has a slight impact on the
arching-induced lateral load transfer; the earth pressure on the
shaft increases as the shaft diameter increases, but the arching
effect from the tangential stress of the soil decreases. In other
words, for large shaft diameters, the distribution of the lateral
earth pressure is similar to plane strain conditions. This result is
similar to the description of the effect of the shape of drilled shafts
given by Liang and Zeng (2002).
18 J. Cho et al. / Tunnelling and Underground Space Technology 48 (2015) 1119
(a) Lateral earth pressure distribution
(b) Shaft displacement
Fig. 12. Effect of shaft elastic modulus.
Fig. 7(b) shows the nal displacements at the end of excavation.
The deformation shows that most of the shaft moved away from
the soil toward the excavation, which indicates active conditions
behind the shaft. With an increase in the lateral earth pressure,
the deformation of shaft gradually is increased.
4.2. Effect of shaft height
The distributions of the computed lateral earth pressure for dif-
ferent heights are shown in Fig. 8(a) and illustrate the effects of
shaft height. The results from Rankines theory are included for
comparison. Rankines theory predicts signicantly greater earth
pressures than the FE results at large depths. As described in the
previous section, Rankines theory is based on two-dimensional
plane strain and thus overestimates the lateral earth pressures
with increasing depth. As shown in Fig. 8(a), the magnitude and
distribution of lateral earth pressure is clearly related to the shaft
height, and the pressure increases signicantly with increasing
shaft height.
Fig. 8(b) shows the deformed shape of the shaft at the end of
excavation. The displacement of shaft increased with an increase
in the lateral earth pressure. This result was different from the dis-
cussion of Paik and Salgado (2003) of the arching effect on rigid
retaining walls. Most previous studies were based on two-dimen-
sional plain strain conditions, which do not consider the actual lat-
eral earth pressure that acts on a vertical circular shaft and thus
may not accurately represent the behavior of actual vertical shafts.
4.3. Effect of the strength parameters of the soil
The effect of the internal friction angle of the soil was investi-
gated. The friction angle of typical Korean weathered soil ranges
from 0 to 40. Fig. 9(a) shows the distribution of lateral earth pres-
sure with different internal friction angles. The results show a
reduction in the lateral earth pressure with a gradual increase of
internal friction. Soil with a higher friction angle is more likely to
mobilize granular interlocking and stronger arching. Fig. 9(b)
shows the variation of shaft deformation with increasing friction
angle. These results show that as the internal friction angle of
the soil increases, the earth pressure and displacement of the shaft
decrease. The results also indicate that Rankines theory may not
reect the nature of the arching effect of soil for the general case
and thus requires validation before being used in practice.
Fig. 10 shows the lateral earth pressure and deformation of the
shaft for different soil cohesions (c = 10, 20, 40, 60 and 80 kPa). As
the cohesion of the soil increases, the earth pressure and displace-
ment decrease. These results are similar to those of the effect of the
internal friction angle of the soil. When the cohesion reaches a cer-
tain value, e.g., 40 kPa, the arching effect is fully developed. This
behavior is explained by fact that the residual forces remained
unchanged with a further increase of cohesion (Liang and Zeng,
2002). Cohesive soil with high cohesion may not need to experi-
ence yielding to develop arching, by which the vertical and radial
stresses induced from the gravitational forces can be transferred
to the vertical shaft.
4.4. Effect of the bearing layer
To investigate the effect of the bearing layer, additional FE ana-
lyses were conducted under the multi-layered soil conditions. The
vertical shaft extends through up to 15 m of weathered soil and is
embedded in weathered or soft rocks. The results based on the end
bearing conditions are shown in Fig. 11. As expected, the results
show smaller lateral earth pressures and displacements for the
embedded shaft compared to the oating shaft (i.e., embedded in
a weathered soil layer).
4.5. Effect of the elastic modulus of the shaft
The effect of the shaft-soil stiffness was investigated by changing
the elastic modulus of shaft, and the results are shown in Fig. 12. As
expected, the lateral earth pressure that acts on the shaft and the
behavior of the shaft are greatly inuenced by the stiffness. As the
elastic modulus of the shaft increases, the arching effect of the soil
decreases. When the elastic modulus of the shaft is greater than
28,000 MPa, no arching effect occurs, and the vertical shaft behaves
like a rigid retaining wall. As the elastic modulus of the shaft
increases, the difference between the results of Rankines theory
and those of this study decrease, and the pressures approach the
distribution of the lateral earth pressure at rest (i.e., K0 condition).
5. Conclusions
The primary objective of this study was to investigate the lateral
earth pressure that acts on a vertical circular shaft embedded in
weathered soil. A limited study of the response of a centrifuge
 J. Cho et al. / Tunnelling and Underground Space Technology 48 (2015) 1119
model test was performed. In addition, a series of 2D nite element
analyses was conducted to examine the three-dimensional arching
effect of the lateral earth pressure. A comprehensive parametric
study was performed using the validated FE modelling techniques.
The information obtained from this study provides a qualitative
understanding of the arching effect for the design of vertical shafts.
Based on the results of this study, the following conclusions can be
drawn:
1. The experimental results provide a realistic distribution of the
lateral earth pressure on a vertical circular shaft, which is not
linear with increasing excavation depth. Consequently, a com-
monly used theoretical method (i.e., Rankines equation) for cal-
culating the lateral pressure can substantially overestimate the
lateral pressure (by approximately 80%) in real situations.
2. The parametric study found that the magnitude and distribu-
tion of lateral earth pressure was related to the soil arching
effect due to the combination of inuencing factors. Of these
factors, the lateral earth pressure is affected more by the soil
properties and shaft stiffness (particularly the friction angle soil
and elastic modulus of the shaft) than by the geometry of the
vertical shaft (i.e., the diameter and length of the shaft).
3. The arching effect is generally more signicant for exible ver-
tical shafts than for rigid vertical shafts, and relatively exible
vertical shafts have lower earth pressures than rigid vertical
shafts. When a vertical shaft is designed under limited displace-
ment conditions (i.e., rigid vertical shaft), the lateral earth pres-
sure can be overestimated because the arching effect of the soil
is ignored.
Acknowledgement
This work was supported by the National Research Foundation
of Korea (NRF) grant funded by the Korea government (MSIP) (No.
2011-0030040). This work also was supported by the Power
Generation & Electricity Delivery of the Korea Institute of Energy
Technology Evaluation and Planning (KETEP) grant funded by the
Korea government Ministry of Trade, Industry & Energy (No.
20131010501790).
References
Chun, B.S., Shin, Y.W., 2006. Active earth pressure acting on the cylindrical retaining
wall of a shaft. South Korea Ground Environ. Eng. J. 7 (4), 1524.
Coulomb, C.A., 1773. Essai sur une application des rgles de maximis & minimis
quelques problmes de statique, relatifs larchitecture. Mmoires de
mathmatique & de physique prsents lAcadmie Royale des Sciences par
divers savans & ls dans ses assembles, 7, 343382, Paris (printed in 1776).
Day, R.A., Potts, D.M., 1993. Modelling sheet pile retaining walls. Comput. Geotech.
15 (3), 125143.
Fujii, T., Hagiwara, T., Ueno, K., Taguchi, A., 1994. Experiment and analysis of earth
pressure on an axisymmetric shaft in sand. In: Proceedings of the 1994
International Conference on Centrifuge. Singapore, pp. 791796.
Grande, L., Soreide, O.K., Tefera, T.H., 2002. Large scale model testing on the moment
distribution and deformation behaviour of a sheet pile wall. In: Proc., 2nd Int.
19
Conf. on Soil Structure Interaction in Urban Civil Engineering, Swill Federal
Institute of Technology, Zurich, Switzerland, pp. 389394.
Handy, R.I., 1983. The arch in soil arching. J. Geotech. Eng., ASCE 111 (3), 302318.
Harrop-Williams, K.O., 1989. Geostatic wall pressures. J. Geotech. Eng, ASCE 115 (9),
13211325.
Herten, M., Pulsfort, M., 1999. Determination of spatial earth pressure on circular
shaft constructions. Granular Matter 2 (1), 17.
Imamura, S., Nomoto, T., Fujii, T., Hagiwara, T., 1999. Earth pressures acting on a
deep shaft and the movements of adjacent ground in sand. In: Proceedings of
the International Symposium on Geotechnical Aspects of Underground
Construction in Soft Ground, Tokyo, Japan, pp. 647652.
Janssen, H.A., 1895. Versuche ber getreidedruck in silozellen. Zeitschrift, Verein
Deutscher Ingenieure 39, 10451049 (partial English translation in Proc. Inst.
Civ. Engrs, 1986, 553.
Jeong, S.S., Seo, D.H., 2004. Analysis of tieback walls using proposed py curves for
coupled soil springs. Comput. Geotech. 31 (6), 443456.
Karlsrud, K., Andresen, L., 2005. Loads on braced excavation in soft clay. Int. J.
Geomech. 5 (2), 107113.
Kim, Y.H., Jeong, S.S., 2011. Analysis of soil resistance on laterally loaded piles based
on 3D soil-pile interaction. Comput. Geotech. 38 (2), 248257.
Kim, K.Y., Lee, D.S., Cho, J.Y., Jeong, S.S., 2013. The effect of arching pressure on a
vertical circular shaft. Tunn. Undergr. Space Technol. 37, 1021.
Konig, D., Guettler, U., Jessberger, H.L., 1991. Stress redistributions during tunnel
and shaft constructions. In: Proceedings of the International Conference
Centrifuge 1991, Boulder, Colorado, pp. 129135.
Krabbenhoft, K., Damkilde, L., Krabbenhoft, S., 2005. Ultimate limit state design of
sheet pile walls by nite elements and nonlinear programming. Comput. Struc.
83 (45), 383393.
Lade, P.V., Jessberger, H.L., Makowski, E., Jordan, P., 1981. Modeling of deep shafts in
centrifuge test. In: Proceedings of the International Conference on Soil
Mechanics and Foundation Engineering, Stockholm, Sweden, vol. 1, pp. 683
691.
Liang, R., Zeng, S., 2002. Numerical study of soil arching mechanism in drilled shafts
for slope stabilization. Soils Found. 42 (2), 8392.
Nakai, T., Xu, L., Yamazaki, H., 1997. 3D and 2D model tests and numerical analyses
of settlements and earth pressures due to tunnel excavation. Soils Found. 37 (3),
3141.
Paik, K.H., Salgado, R., 2003. Estimation of active earth pressure against rigid
retaining walls considering arching effects. Geotechnique 53 (7), 643653.
Potts, D.M., Fourie, A.B., 1984. The behavior of a propped retaining wall: results of a
numerical experiment. Geotechnique 34 (3), 383404.
Rankine, W.J.M., 1857. On the stability of loose earth. Phil. Trans. Royal. Society,
London 147 (1), 927.
Schoeld, A.N., 1980. Cambridge geotechnical centrifuge operations. Gotechnique
30 (3), 227268.
Spangler, M.G., Handy, R.L., 1984. Soil Engineering. Harper & Row, New York.
Tan, Y., Paikowsky, S.G., 2008. Performance of sheet pile wall in peat. J. Geotech.
Geoenviron. Eng. 134 (4), 445458.
Tatiana, T., Mohamed, A.M., 2010. Comparative evaluation of methods to determine
the earth pressure distribution on cylindrical shaft: a review. Tunn. Undergr.
Space Technol. 25, 188197.
Terzaghi, K., 1920. Old earth-pressure theories and new test results. Eng. News-
Record 85 (13), 632637.
Tobar, T., Meguid, M.A., 2009. Distribution of earth pressure on vertical shafts. In:
62nd Canadian Geotechnical Conference, Halifax, September 2009, CD, 6 pages.
Tran, V.D.H., Meguid, M.A., Chouinard, L.E., 2014. Discrete element and
experimental investigations of the earth pressure distribution on cylindrical
shafts. Int. J. Geomech., ASCE 14 (1), 8091.
Walz, B., 1973. Left bracket apparatus for measuring the three-dimensional active
soil pressure on a round model caisson right bracket. Baumaschine Bautechnik
20 (9), 339344.
Wang, Y.Z., 2000. Distribution of earth pressure on a retaining wall. Geotechnique
50 (1), 8588.
Wang, W.L., Yen, B.C., 1973. Soil arching in slopes. J. Geotech. Eng., ASCE 100 (1), 61
78.
Wong, R.C.K., Kaiser, P.K., 1988. Design and performance evaluation of vertical
shaft: rational shaft design method and verication of design method. Can.
Geotech. J. 25, 320337.