Tunnelling and Underground Space Technology 37 (2013) 1021

Contents lists available at SciVerse ScienceDirect

Tunnelling and Underground Space Technology

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

The effect of arching pressure on a vertical circular shaft

Kyoung-Yul Kim a, Dae-Soo Lee a, Jaeyeon Cho b, Sang-Seom Jeong b,, Sungjune Lee c
a
Korea Electric Power Research Institute, Daejeon, South Korea
b
Department of Civil and Environment Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul 120-749, South Korea
c
Department of Civil Engineering, Cheongju University, Cheongju 360-764, Republic of 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 of vertical circular shaft is investigated by using experimental tests and a the-
Received 15 August 2012 oretical analysis. The emphasis was quantifying on the magnitude and distribution of a lateral earth pres-
Received in revised form 15 January 2013 sure measured by three-dimensional arching effect. A framework for determining the earth pressure
Accepted 11 March 2013
distribution on the basis of both centrifuge model and full-scale eld test results is introduced. Through
Available online 13 April 2013
these studies, it is found that the lateral earth pressure decreases within active displacement (dh/H:
0.2  102, dh: horizontal displacement, H: Height of vertical shaft) of wall at deep excavation and the
Keywords:
arching effect is more signicant for deep excavation than for shallow excavation. It is also found that
Vertical circular shaft
Arching effect
the lateral earth pressure acting on a vertical circular shaft considering arching effect is 80% smaller than
Earth pressure that calculated by Rankine theory, and thus represents a signicant improvement in the prediction of
Centrifuge model test realistic earth pressure for vertical circular shaft subjected to arching pressure.
Full-scale eld test 2013 Elsevier Ltd. All rights reserved.

1. Introduction retaining wall in sands (Terzaghi, 1943; Steinfeld, 1958; Prater,

1977). Moreover, recent studies by Coates (1981) determined that
In South Korea, an increasing need for underground excavation the earth pressure acting on a vertical shaft lining. Also, Handy
and the application of the vertical circular shaft is becoming an (1983) and Wang (2000) reported that the friction between backll
important issue in engineering practice. Here the vertical shaft is and the retaining wall was caused by earth pressure in a curved
generally embedded in soft or hard rocks through weathered soil failure surface. However, most of previous studies were based on
varying thicknesses ranging up to 15 m. Since 1980, a number of two-dimensional plain strain condition and thus, it may not repre-
huge excavation projects such as underground space for construct- sent the behavior of actual vertical shaft which does not take into
ing power stations, subways, high-speed railways, and many life- account the actual lateral earth pressure acting on a vertical circu-
line constructions have been performed in urban areas (Jeong lar shaft.
and Seo, 2004). Although there have been considerable advance- To overcome this limitation, some research has been performed
ments in construction techniques involving vertical shaft, it is still by centrifuge model test for a vertical circular structure. Konig
very important for tunnel engineers to control the lateral earth et al., (1991) conducted a test to investigate the three-dimensional
pressure. Additionally, proper estimation of earth pressures is a earth pressure distribution, and Toyosawa et al., (2006) analyzed
key factor in design of vertical shafts. Because of three dimensional lateral earth pressure distribution for plane retaining wall using a
arching effects, i.e. convex arching and/or inverted arching, the centrifuge model test.
earth pressure acting on a circular type of vertical shaft is less than Handy (1983) analyzed soil arching action behind retaining
other types. Thus, the circular vertical shaft is generally used. walls, and Wang and Yen (1973) carried out this analysis for slopes.
Much work has been done to study the lateral earth pressure Nakai et al. (1997) performed a series of physical model tests under
acting on a rigid wall by many researchers. Terzaghi (1920) re- a 1 g conditions and carried out numerical analysis of these tests to
ported fundamental studies on characteristics of earth pressure investigate the arching effect. They found that the results obtained
acting on retaining wall. Berezantzev (1958) reported that the ac- from the model tests were in good agreement with those obtained
tive earth pressure acting on a gravel-backlled cylindrical retain- from the numerical analysis. Recently, Kim et al. (2009) reported
ing wall, however it could not consider the earth pressure changes that the earth pressure was reduced by horizontal/vertical three-
with varying wall shapes. In addition, a few previous studies have dimensional arching effect. However, these studies do not predict
been performed for the earth pressure acting on a cylindrical the characteristics of the earth pressure considering arching effect
at each excavation stage.
Corresponding author. Tel.: +82 19 362 4074. As discussed above, since Terzaghi (1920) started to examine
E-mail address: soj9081@yonsei.ac.kr (S.-S. Jeong).
the effect of lateral earth pressure acting on a plane wall, much

0886-7798/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tust.2013.03.001
 K.-Y. Kim et al. / Tunnelling and Underground Space Technology 37 (2013) 1021 11

work has been done in the area of soilstructure interaction of

plane retaining walls. Relatively little work has been done on ex-
ible and circular walls such as vertical shaft, and hardly anything
on earth pressure distribution of vertical circular shaft. Soil-circu-
lar wall behavior is complicate since many complex construction
sequences, different soil conditions, initial conditions, and arching
effect are involved. Additionally, previous studies are not applica-
ble to the multi-layered and/or c/ soils.
The objective of this study is to investigate the actual magni-
tude and distribution of lateral earth pressure acting on the vertical
circular shaft considering the three-dimensional arching effect in
weathered soil. In order to achieve the objective, new theoretical
method is suggested for estimating the earth pressure in multi-lay-
ered c/ soils. Comparisons are made between the soil-vertical
shaft behavior observed by centrifuge test and full-scale eld tests
and that computed by the theoretical methods.

2. Theoretical earth pressure formula of vertical circular shaft

The behavior of a vertical shaft is affected and near the surface (a) Mode of yielding (Wong and Kaiser, 1988)
dominated by gravitational forces (Wong and Kaiser, 1988). It is a
truly three-dimensional problem and all three stress components
(r0t , r0m , r0r ). The stress concentrations near a vertical shaft would
cause yielding due to the stress difference between vertical (r0m )
and radial (r0r ) around a shaft (Fig. 1a). In order to accurately deter-
mine the lateral earth pressure on a vertical shaft, it is necessary to
consider an accurate shape for the slip surface in the formulation
for the earth pressure. In this study, it is assumed that the slip sur-
face for a vertical shaft is plane and slope at an angle of b = 45 + //
2 is used to calculate the earth pressure on the shaft (Fig. 1b).
The three-dimensional earth pressure acting on a vertical circu-
lar shaft in c/ soil is calculated by considering convex (lateral
direction) and inverted (vertical direction) arching effect (Fig. 2).
A comprehensive study of earth pressure acting on vertical shaft
has been reported by Kim et al. (2009). Therefore, only the coef-
cient of tangential and radial earth pressure will be briey summa-
rized here.

2.1. The coefcient of tangential earth pressure

The coefcient of tangential earth pressure (k) is dened to be

(b) Slip surface
the ratio of tangential stress to vertical stress. The earth pressure
on vertical shafts decreases with the increase of tangential stress Fig. 1. Behavior of the soil around vertical circular shaft: (a) mode of yielding and
acting on the failure surface, thus with increase of k value. Wong (b) slip surface.
and Kaiser (1988) suggested that k value is not 1.0 at elastic state;
but will reach 1.0 at plastic state. In this study, k is assumed to be
1 sin / (Lee et al., 2007). Z 2p Z 2p   Z 2p
dh cos b
Nw R dh dz 2r0t sin r dz sf r R dh dz
0 0 2 0 sin b
2.2. The coefcient of tangential earth pressure Z 2p
cos90  b
Nf r R dh dz 2
The coefcient of radial earth pressure is dened as the ratio of 0 sin b
radial stress to vertical stress. Rankine (1857) suggested the coef-
cient of active earth pressure (Ka) without considering wall friction
angle (d). Paik and Salgado (2003) assumed miner principle stress In Eq. (2), Nw is the lateral stress acting on wall surface of ele-
direction is changing in concave shape, such as elliptic, catenary ment and is same as rh, and sf is dened as shear strength at failure
and parabolic due to ctional resistances of wall, and the coef- surface. Nw and r0t will be expressed as
cient of earth pressure (Kwa) as follows: Nw r0h kwa r0m 3
2 2
3N cos h sin h
kwa 1 r0t 1  sin /r0m 4
3N  N  1 cos2 h
 p
where N tan2 45 /=2; h tan1
N1 N12 4N tan2 d
, d is the where rm is the vertical stress at certain depth z. If dh is very small,
0
2tand
wall friction angle (). then Eq. (5) can be obtained by substituting Eqs. (3) and (4) into Eq.
As shown in Fig. 3, the equilibrium condition of horizontal force (2):
acting on the differential soil element gives Rkwa r0m r1  sin /r0m c cos b
r R tan b
sin b
Nf  5
rR tan b  tan /
12 K.-Y. Kim et al. / Tunnelling and Underground Space Technology 37 (2013) 1021

(a) Lateral arching pressure

(b) Vertical arching pressure

Fig. 2. Lateral and vertical arching effect of vertical circular shaft: (a) lateral arching pressure and (b) vertical arching pressure.

  
From Fig. 3a, the equilibrium of vertical stress gives 2p 1 1 tan b  tan /
T c cw cR r 1
A tan b tan b  tan /
r0m A A dzc r0m drm A 2pRsw dz 2pr Rfsf sin b
dz
N f sin90  bg 6   
sin b 2p 1 tan b  tan /
S kwa R tan d kwa R kr
A tan b  tan /
where A is the area of differential soil element, sf is the shear
strength at failure surface, and sw is the shear strength at wall
surface. where c is the soil unit weight (kN/m3), / is the soil friction angle
The rst order differential Eq. (7) can be obtained by substitut- (), c is the cohesion factor (kPa), R is the radius of vertical circular
ing Eq. (5) into Eq. (6). shaft (m), r is the width of plastic region in certain depth (m), A is
the horizontal area of loose area (m2), b is the angle of slip surface
dr0m
Sr0m T 7 (45 + //2), and q is the surcharge load.
dz Therefore, the total soil pressure (pi) can be expressed as
Eq. (7) can be solved with the boundary condition, r0m 0 when
z = 0. Then, r0m is
Pi kwa r0m 9
 
T sz T
r0m q e 8
S S
where kwa is a coefcient of radial earth pressure and r0m is the ver-
For parameter T and S can be written as tical effective overburden pressure.
 K.-Y. Kim et al. / Tunnelling and Underground Space Technology 37 (2013) 1021 13

(a) Section view

(b) 3D view
Fig. 3. Equilibrium of stress at soil: (a) section view and (b) 3D view.

3. Centrifuge model test shaft of length 15 m, and outer diameter 6 m at prototype scale. To
investigate the lateral earth pressure at each excavation stage, test
3.1. Testing program model was divided two sections.
The soil used in the model test had a specic gravity (Gs) of 2.65
A series of centrifuge model tests were performed to investigate and was classied as SP in the Unied Soil Classication System.
the soil pressure of vertical circular shaft in sand. Two model tests The sand layer was prepared from the sand hopper to form a uni-
were performed on 1/75th scale models to analyze characteristics formly dense sand bed with 8182% relative density approxi-
of earth pressure acting on the vertical circular shaft, as summa- mately. Table 2 presents the material properties used in the
rized in Table 1. The model tests were performed at a radial accel- centrifuge models. To observe the plastic region in soil sample, a
eration of 75g. different color was adopted at each soil layer.
Model vertical shaft was made of aluminum alloy with a hollow After completing sample preparation, the sample box was
circular section, 200 mm embedded length, and 80 mm outer mounted on the platform of the centrifuge. The lateral earth
diameter (Fig 4b). It was equivalent to reinforced concrete vertical pressure acting on the vertical shaft was monitored during the
14 K.-Y. Kim et al. / Tunnelling and Underground Space Technology 37 (2013) 1021

Table 1
Centrifuge model tests.

Soil Vertical shaft

3
c (kN/m ) / () Water content (%) G (kPa) Diameter (cm) H (cm) H/D
Test 1 12.66 36.95 2.0 75 8.0 20.0 2.5
Test 2 12.65 36.95 2.0 75 8.0 20.0 2.5

(a) Soil sample box (b) Model vertical shaft (c) Sand hopper
Fig. 4. Testing apparatus: (a) Soil sample box, (b) model vertical shaft and (c) sand hopper.

Table 2 geotechnical centrifuge was manufactured by Actidyn company

Properties of the test soil. in France and corresponds to an Actidyn system C65-2, with a plat-
D10 (mm) 0.091 form radius of 3.0 m and a maximum capacity of 120 g ton. The
D20 (mm) 0.098 specication of DICTs centrifuge is summarized in Table 3.
D60 (mm) 0.140 A small soil sample box 700 mm long, 400 mm wide, and
D90 (mm) 0.175
Coefcient of uniformity, Cu 1.53
200 mm high was made with a steel frame (Fig. 4a). Model vertical
Specic gravity, Gs 2.65 shaft was made of aluminum alloy with a hollow circular section,
Unied soil classication system SP 400 mm embedded length, 20 mm outer diameter, and 3 mm
thickness. They are equivalent to reinforced concrete piles of
length 15 m, outer diameter 6 m at prototype scale.
A sand hopper was designed to form a uniform soil bed with the
excavation. To simulate the excavation step, incremental active required density (Fig. 4c). It is automatically controlled by a com-
wall displacement (dh/H, dh: horizontal displacement, H: Height puter system according to the traveling velocity of the hopper and
of vertical shaft) was chosen to be 1.0 mm/140 s using a multi-gear the height and quantity of pluviation.
motor. Terzaghi (1920) suggested that active earth pressure occurs A variety of instrumentations were installed to monitor dis-
at about 0.1%H of active wall displacement. This value is applicable placements and earth pressures during the testing. The displace-
to 15 mm wall displacement in this test. ment of vertical shaft wall was measured by two linear variable
differential transformers (LVDTs) located on both sides of the shaft
3.2. Testing apparatus surface shown in Fig. 5. Pre-calibrated Kyowa earth pressure trans-
ducers were installed on the external surface of vertical cylinder
The geotechnical centrifuge used in this study was located at wall to measure the lateral earth pressures acting on the vertical
Daewoo Institute of Construction Technology (DICT). DICTs shaft, and their locations in the model are also shown in Fig. 5.

Table 3
Main properties of geotechnical centrifuge equipment.

Centrifuge radius Platform radius (m) 3.0

Nominal radius (m) 2.7
Usable payload dimension Platform width (m) 0.8
Nominal depth (m) 1.0
Usable height (m) 0.8
Performance Maximum payload (kN) 12
Acceleration at max. payload (g) 100
Acceleration range (g) 10230
Speed range (rpm) 25265
Electrical slipring Signal, power, and video lines 46, 12, 2
Rotary joint Oil 20 MPa working pressure Each two ports
Water 2 MPa working pressure
Air
 K.-Y. Kim et al. / Tunnelling and Underground Space Technology 37 (2013) 1021 15

(a) Section view

(b) Plan view

(c) Vertical circular shaft model

Fig. 5. Sectional and plan views of the soil sample box and vertical shaft model.

3.3. Test results and discussion according to the relevant scaling relationships (Schoeld, 1980).
The validation of initial condition (displacement of vertical
Some typical results from centrifuge model tests are presented shaft = 0) for earth pressure was examined by a comparison with
here. The data are presented in terms of prototype units scaled up a K0 condition. Fig. 6a and b shows a comparison of the measured
16 K.-Y. Kim et al. / Tunnelling and Underground Space Technology 37 (2013) 1021

Lateral earth pressure (kPa) 80

0 20 40 60 80
0
K0=0.39

Lateral earth pressure (kPa)

60
0mm

40
Depth (m)

20

12

0
0 0.2 0.4 0.6
-2
16 h/H (x10 )
(a) Test 1 (a) Upper section
Lateral earth pressure (kPa)
80
0 20 40 60 80
AP1
0 AP2
K0=0.39 AP3
Lateral earth pressure (kPa)

0mm 60 AP4
AP5
4 AP6
AP7
AP8
40
Depth (m)

20

12

0
0 0.2 0.4 0.6
16
-2
h/H (x10 )
(b) Test 2
Fig. 6. Comparison of measured and calculated earth pressure at rest (at acceler-
(b) Lower section
ation 75g).
Fig. 7. Earth pressure distributions in Test 1: (a) upper section and (b) lower
section.
and calculated earth pressures. It is shown that agreement be-
tween the K0 and initial condition of test is generally good (coef-
cient of determination, R2 = 0.96, 0.99). to increase as the active displacement of lower part increases. Fur-
Figs. 7 and 8 show the results from two typical tests with mea- thermore, an increase of lateral earth pressure in the active dis-
sured earth pressure under the condition of active wall displace- placement from 0 to 0.2 results in about 1030%. Similar
ment. As shown in Figs. 7a and 8a, it is noted that the lateral magnitude of earth pressure as initial condition (before the rst
earth pressure on upper section of the vertical shaft (AP1AP4) excavation) is measured when the active wall displacement was
at rst excavation step gradually decreases, until a wall displace- increased from 0 to 0.6.
ment is reached at about 0.2. This curve demonstrated the reduc- Fig. 9 shows the distribution of plastic deformation in soil sam-
tions in lateral earth pressure with gradual increase of wall ple. It is noted that plastic region in the soil sample is about 1D
displacement. When the active wall displacement (dh/H: (D = diameter of vertical shaft) and then, the settlement of soil
0.2  102) is occurred, the earth pressure was about 30% (AP4) de- sample was about 13 mm (97.5 cm in full scale). Fig. 10 shows a
creases in initial earth pressure. At this time, the earth pressure on comparison of the test result and theoretical solution proposed
lower section (AP5AP8) of vertical shaft is measured in a constant. by Kim et al. (2009). All the measured data have a relatively smal-
In the second excavation step, where the earth pressure in lower ler earth pressure than the result of the proposed solution. How-
section of the vertical shaft, large reductions of earth pressure ever, earth pressure at the bottom, 14 m in depth, occurred
are occurred in Figs. 7b and 8b. When the active wall displacement relatively high, because soil container used in test is the particu-
(dh/H: 0.2  102) is occurred, the earth pressure was about 80% larly strong steel structure, earth pressure caused by soil loosening
(AP8) decreases of initial earth pressure. Additionally, an earth could not be dissipated at the bottom section. In most of previous
pressure acting on the upper part (AP1AP4) of vertical shaft tends studies, it is veried that the overall lateral earth pressure acting
 K.-Y. Kim et al. / Tunnelling and Underground Space Technology 37 (2013) 1021 17

80 Lateral earth pressure (kPa)

0 20 40 60 80
0
Lateral earth pressure (kPa)

K0=0.39
60 0mm (Test1)
theory (Kim)
4 h/B=0.2x10 2 (Test1)
h/B=0.2x10 2 (Test2)
40

Depth (m)
8

20

12

0
0 0.2 0.4 0.6
h/H (x10-2 ) 16
(a) Upper section Fig. 10. Comparison of earth pressure distributions.
80
on the vertical circular shaft was reduced approximately 80% com-
AP1
AP2 pared with the initial condition.
AP3 From the test results, it is note that the lateral earth pressure at
Lateral earth pressure (kPa)

AP4 deep depth decreased within active displacement (dh/H:

60 0.2  102), beyond which it continue to increase toward the initial
AP5
AP6 condition as displacement increase. This nding is similar from
AP7 what Tatiana and Mohamed (2010) discussed about active dis-
AP8 placement for a wall. On the other hand, test vertical shaft is exca-
40 vated into a relatively shallow excavation, there is very little
reduction of earth pressure. The reason for this behavioral differ-
ence between deep and shallow excavation step is explained by
the fact that the weight re-distribution caused by the arching effect
20 will produce more pressure reduction in the deep depth than shal-
low depth. Conclusively, the arching effect is generally more signif-
icant for deep excavation than for shallow excavation and the
observed values by the centrifuge model have a smaller earth pres-
0 sure when compared with the calculated by existing method (Kim
0 0.2 0.4 0.6 et al., 2009).

h/H (x10 - 2 )
4. Field test
(b) Lower section
Full-scale eld test was conducted at the test site in order to
Fig. 8. Earth pressure distributions in Test 2: (a) upper section and (b) lower
section.
investigate the lateral earth pressure acting on vertical circular
shafts. This test was performed to provide a basis for a design

Fig. 9. Failure surface (model height 200 mm, 75 g).

18 K.-Y. Kim et al. / Tunnelling and Underground Space Technology 37 (2013) 1021

(a) Configuration of test shaft

(b) Shaft embedment

Fig. 11. Conguration of test shaft and shaft embedment.

guideline that would be applicable to a vertical shaft installed in three-dimensional arching effect. Special attention is given to the
weathered soils. Based on the soil condition, the applicable limiting analyzing lateral earth pressure acting on full-scale vertical shaft
depth of vertical shaft is conned as up to 15 m to identify the (Fig. 11a) as excavation proceeds.
 K.-Y. Kim et al. / Tunnelling and Underground Space Technology 37 (2013) 1021 19

Table 4
Summary of eld and laboratory test results.

Soil Depth (m) N value Gs W.C. (%) E (MPa) / () c (kPa) Sieve analysis (%)
#4 #10 #40 #100 #200
Top soil 06.2 17/3050/18 2.59 8.3 20.96 98.0 82.1 23.1 8.7 5.9
Weathered soil 6.213.1 16/3050/10 2.62 18.3 23.3 25.31 126.4 99.1 84.5 37.7 15.9 8.6
Weathered rock 13.123.6 50/150/0 33.8 (66.9) 31.8 294.0
Soft rock 23.625.5

Lateral earth pressure (kPa)

0 20 40 60 80 100
0
Measured
Theory (Kim)
2 Rankine

4
Depth (m)

8
Fig. 13. Stress applied on circular cylinder.

10
to determine the particle size distribution, water content in rela-
tion to Atterberg limits.
12 A generalized soil prole for this site is presented in Fig. 11b.
The soil prole near the surface consists of layers of top soil,
14 weathered soil and weathered rock underlain by soft rock. The
water table was located 9.7 m below the natural ground surface
(a) P-1 during the tests. The surface soils consist of poorly graded sand
Lateral earth pressure (kPa) and silt sand classied as SP or SM according to the Unied Soil
0 20 40 60 80 100
Classication System (USCS). The undrained shear strength was
typically 126.4294 kPa. The underlying weathered soil layer con-
0
Measured sists of a well graded residual soil classied as SW or SM (USCS).
Theory (Kim) Table 4 shows the summary of eld and laboratory test results.
2 Rankine
4.2. Test program and setup
4
A schematic representation of an instrumented vertical shaft is
also shown in Fig. 11b. The circular shaft had an outer diameter of
Depth (m)

6
6 m and a wall thickness of 80 mm. The depth of the shaft below
ground level was 13.6 m. Excavation was performed by the back-
8 hoe shovel. The reinforcing structure consisted of 12 H-pile
(300  200  9  14 mm) with a ring wale (300  200  9 
10 14 mm). During the construction of H-pile, S.G.R. (Space Grouting
Rocket) method was adopted for the stability of the soil.
Many instrumentation was used to monitor the response of the
12
vertical shaft-soil system. Automated equipment including soil
pressure gauge, vibrating wire-stress meter, inclination, void water
14 pressure and ground water level measurement device were in-
stalled. Earth pressure was measured at 1 m below the top of
(b) P-2
H-pile using 10 soil pressure gauges, which were attached at two
Fig. 12. Measured and calculated earth pressure: (a) P-1 and (b) P-2. locations (P-1 and P-2). Two vibrating wire-stress meter was
installed along the ring wales to measure the earth pressure. The
4.1. Geotechnical conditions at test site gauges, shown in Fig. 11, were sealed with a JIG to prevent
mechanical damage. Data from the various sensors were collected
A comprehensive geotechnical investigation was carried out to simultaneously using a computer controlled data acquisition
dene the soil prole and properties at the test site as accurately system.
as possible. The in situ eld tests included the standard penetration
test (SPT), pressure meter testing (PMT) and borehole shear test. 4.3. Test results and discussion
Undisturbed soil samples were retrieved for laboratory testing
from the boring. Relatively disturbed specimens were also re- The observed lateral earth pressures are compared with the
trieved from the SPT sampler. Laboratory testing was performed predicted values of theoretical methods. Fig. 12a and b shows
20 K.-Y. Kim et al. / Tunnelling and Underground Space Technology 37 (2013) 1021

Lateral earth pressure (kPa) Lateral earth pressure (kPa)

0 20 40 60 80 100 0 20 40 60 80 100
0 0
Converted Measured (P-1)
Theory (Kim) Measured (P-2)
2 Rankine 2 Converted (P-1)
Converted (P-2)
Theory (Kim)
4 4
Rankine
Depth (m)

Depth (m)
6 6

8 8

10 10

12 12

14 14
(a) P-1 Fig. 15. Comparison of test result for vertical circular shaft.

Lateral earth pressure (kPa)

0 20 40 60 80 100 2pi R 2F 10
0
Converted where pi is the soil pressure, R is the radius of vertical shaft, F is the
Theory (Kim) ring beam axial stress. In this analysis (+) stress value means active
2 Rankine
earth pressure and () value means passive earth pressure.
Fig. 14 shows the converted and predicted earth pressures dis-
4 tribution. The calculated results of existing method (Kim et al.,
2009) closely approach the converted data from the measured
Depth (m)

6 one. Although a reasonably good agreement between the theoret-

ical and the measured one was obtained, the existing method
(Kim et al., 2009) has a larger lateral earth pressure those of the
8 measured values at the same depth. In this case, it is also shown
that maximum lateral earth pressure acting on the vertical shaft
10 occurred within the depth of 46 m below the ground level. How-
ever, the calculated results by Rankine theory have a signicantly
70% lager lateral earth pressure at the deep depth than the results
12
of previous theory (Kim et al., 2009) and converted one. The three-
dimensional arching effect is generally more signicant for deep
14 depth than for shallow depth and the previous method considering
(b) P-2 arching effect has a considerably smaller earth pressure when
compared with the results by Rankine theory.
Fig. 14. Converted and calculated earth pressure: (a) P-1 and (b) P-2.
Fig. 15 shows the lateral earth pressure distributions obtained
representatively the predicted and measured earth pressure at two from the measured and predicted data. It is shown that the predic-
different locations (P-1 and P-2). A reasonably good agreement tions by considering three-dimensional arching effect are in good
between the measured ones and the solution presented by the agreement with the general trend observed by eld measurements
existing theories are obtained. In addition, maximum lateral earth and has a considerably smaller earth pressure than the earth pres-
pressure acting on the vertical shaft occurred within the depth of sure calculated by the Rankine theory. This is because the Rankine
shaft diameter below the ground level. However, the calculated theory based on the two-dimensional plane stress ignores three-
results by Rankine theory have a relatively 80% lager earth pres- dimensional arching effect and, thus, overestimates the lateral
sure as the depth increased than the results of previous theory earth pressure with increasing depth. This clearly demonstrates
(Kim et al., 2009) and measured one. that for vertical circular shaft, there exists three-dimensional
These results clearly demonstrate that the distribution pattern arching effect, so that this set of prediction and measured results
of the earth pressure acting on the vertical shaft is signicantly provides the reducing of lateral earth pressure.
dominated by three-dimensional arching effect. Therefore, the
Rankine theory may not reect the nature of the arching effect of 4.4. Reduction ratio of lateral earth pressure
soil for the general case, and thus it requires validation before
being used in practice. In the present design of vertical shaft, geotechnical engineers
As mentioned above, two vibrating wire-stress meter was in- have traditionally calculated the active earth pressure against
stalled along the ring wales to investigate the earth pressure acting walls using Rankine theory. It was assumed that the distribution
on the vertical shaft. As shown in Fig. 13, the ring wale axial stress of active earth pressure exerted against the wall is triangular. How-
as pressure acting on the circular cylinder can be converted to ever, many experimental results (Tsagareli, 1965; Matsuo et al.,
earth pressure. The earth pressure is calculated from the following 1978; Fang and Ishibashi, 1986) show that the distribution of
equation; active pressure acting on a wall depends on the mode of wall
 K.-Y. Kim et al. / Tunnelling and Underground Space Technology 37 (2013) 1021 21

Table 5
Summary of Rf values for vertical circular shaft.

Soil Values z/D

0.2 0.5 0.7 1 1.7
c/ soil Measured earth pressure (kPa) 18.16 24.26 19.80 13.71 10.89
Rankine earth pressure (kPa) 21.09 37.5 45.7 51.56 70.31
Rf 13.89 35.31 56.67 73.41 84.5

movement and is non-linear, different from the assumption made 3. Based on Rf value, the arching effect is generally more signi-
by Rankine. The non-linearity of the active earth pressure distribu- cant for deep excavation depth than for shallow excavation
tion results from arching effects (Handy, 1983). Especially, proper depth and the measured data has a smaller earth pressure when
estimation of earth pressures is a key factor in design of vertical compared with the results by existing theories. When test shaft
circular shafts. Because of three dimensional arching effects, i.e. were excavated into a relatively shallow depth, there was very
convex arching and/or inverted arching, the earth pressure acting little arching effect caused by small active earth pressure acting
on a circular type of vertical shaft is less than other types. This sec- on vertical shaft and weight re-distribution.
tion presents a correlation between Rankine earth pressure and re-
duced lateral earth pressure based on the results of full-scale eld
test performed in this study. References
As mentioned above, in the case of the vertical circular shaft,
there is no criterion to evaluate earth pressure for vertical shaft. Berezantzev, V.G., 1958. Earth pressure on the cylindrical retaining walls. In:
Conference on Earth Pressure Problems, Brussels, 1958.
Therefore, the measured versus the predicted by Rankine theory Coates, D.F., 1981. Rock Mechanics Principles: Energy, Mines and Resources, Canada
were compared using an Rf value (reduction factor) dened as (Chapter 2 and 3).
Fang, Y., Ishibashi, I., 1986. Static earth pressures with various wall movements. J.
Predicted earth pressure  Measured earth pressure Geotech. Eng., ASCE 112 (3), 317333.
Rf Handy, R.I., 1983. The arch in soil arching. J. Geotech. Eng., ASCE 111 (3), 302318.
Predicted earth pressure
Jeong, S.S., Seo, D.H., 2004. Analysis of tieback walls using proposed py curves for
 100% 11 coupled soil springs. Comput. Geotech. 31 (6), 443456.
Kim, D.H., Cha, M.H., Lee, D.S., Kim, K.Y., Lee, I.M., 2009. Earth pressure acting on
A conservative Rf value ranges from 0% to 100% and Table 5 vertical circular shafts considering arching effects in soils. Tunell. Technol. 12
summarizes the magnitude of Rf with varying excavation depths (2), 129144.
Konig, D., Guettler, U., Jessberger, H.L., 1991. Stress redistributions during tunnel
(z/D). Based on the Rf value results of Table 5, it is found that the and shaft constructions. In: Proceedings of the International Conference
Rf values show a signicant reduction in the prediction of real earth Centrifuge 1991, Boulder, Colorado.
pressure. Lee, I.M., Moon, H.P., Lee, D.S., Kim, K.R., Choi, M.S., 2007. Earth pressure of vertical
shaft considering arching effect in layered soils. Tunnell. Technol. 9 (1), 4962.
Matsuo, M., Kenmochi, S., Yagi, H., 1978. Experimental study on earth pressure of
5. Conclusions retaining wall by eld tests. Soils Found. 18 (3), 2741.
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),
The main objective of this study was to investigate a lateral 3141.
earth pressure acting on a vertical circular shaft embedded in Paik, K.H., Salgado, R., 2003. Estimation of active earth pressure against rigid
retaining walls considering arching effects. Geotechnique 53 (7), 643653.
weathered soil. A limited study of the response of centrifuge and
Prater, E.G., 1977. An examination of some theories of earth pressure on shaft
eld tests was performed to examine the three-dimensional arch- linings. Can. Geotech. J. 14 (1), 91106.
ing effect for lateral earth pressure. When applied to vertical circu- Rankine, W.J.M., 1857. On the stability of loose earth. Philos. Trans. Royal. Soc.,
lar shaft constructed in weathered soil, comparative studies have Lond. 147 (1), 927.
Schoeld, A.N., 1980. Cambridge geotechnical centrifuge operations. Gotechnique
clearly demonstrated the important arching effect of vertical circu- 30 (3), 227268.
lar shaft. Based on the ndings of this study, the following conclu- Steinfeld, K., 1958. Uber den erddruck auf schacht und brunnenwandungen. In:
sions can be drawn: Contribution to the Foundation Engineering Meering, Hambrug, German Soc. of,
Soil Mech. Found. Eng., pp. 111126.
Tatiana, T., Mohamed, A.M., 2010. Comparative evaluation of methods to determine
1. The test results show that the magnitude and distribution of a the earth pressure distribution on cylindrical shafts: a review. Tunn. Undergr.
lateral earth pressure acting on a vertical circular shaft is not Space Technol. 25, 188197.
Terzaghi, K., 1920. Old earth-pressure theories and new test results. Eng. News-Rec.
linear with increasing excavation depth. Experimental observa- 85 (13), 632637.
tions conrm that more realistic distribution of lateral earth Terzaghi, K., 1943. Theoretical Soil Mechanics. Wiley, New York.
pressure is introduced for a vertical shaft embedded in weath- Toyosawa, Y. Itoh, Tamrakar, S.B., Suemasa, N., 2006. Redistribution of active earth
pressures using movable earth support apparatus in centrifuge. Phys. Model.
ered soil.
Geotech. 6th ICPMG., 11131118.
2. Through a comparison of theoretical calculation and full-scale Tsagareli, Z.V., 1965. Experimental investigation of the pressure of a loose medium
test with eld measurements, it is found that for test shaft a on retaining walls with a vertical back face and horizontal backll surface. J. Soil
Mech. Found. Eng., ASCE 91 (4), 197200.
coupled three-dimensional arching effect is important. The lat-
Wang, Y.Z., 2000. Distribution of earth pressure on a retaining wall. Geotechnique
eral earth pressure acting on the vertical circular shaft was 50 (1), 8588.
maximum 80% reduction in a certain depth. Therefore, a com- Wang, W.L., Yen, B.C., 1973. Soil arching in slopes. J. Geotech. Eng., ASCE 100 (1), 61
monly used method like the Rankine equation for calculating 78.
Wong, R.C.K., Kaiser, P.K., 1988. Design and performance evaluation of vertical
the lateral pressure can substantially overestimate the lateral shafts: rational shaft design method and verication of design method. Can.
pressure in real situations. Geotech. J. 25, 320337.