Tunnelling and Underground Space Technology 45 (2015) 107–127

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology

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

Design of shield tunnel lining taking fluctuations of river

stage into account
Cungang Lin a,⇑, Shiming Wu b, Tangdai Xia c
a
Institute of Geotechnical Engineering, College of Civil, Construction and Environmental Engineering, Ningbo University, Ningbo 315211, China
b
Hangzhou Qing-chun Road Cross-river Tunnel Company Limited, Hangzhou 310002, China
c
Institute of Geotechnical Engineering, Key Laboratory of Soft Soils and Geoenvironmental Engineering, Ministry of Education, Zhejiang University, Hangzhou 310058, China

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

Article history: Fluctuations of river stage are expected to induce changes in loads acting on the tunnel linings and cause
Received 4 May 2013 readjustments of member forces in the segmental linings subsequently. Therefore, the evaluation of
Received in revised form 9 August 2014 impacts of time-dependent river levels on the loads acting on the tunnel linings is of great importance
Accepted 20 September 2014
in design of shield tunnel linings situated beneath the rivers. However, the loads acting the tunnel linings
are generally considered as constant in most design methods available, taking no account of the influ-
ences of constantly changing river stage. In this study, the influences of river stage on design of shield
Keywords:
tunnel linings are evaluated with respect to two common ground conditions: (a) impermeable overbur-
Design of shield tunnel lining
River stage
den strata of low permeability and (b) permeable overburden strata of high permeability. Two earth pres-
Earth pressure sure calculation models are correspondingly established. In addition, field observations in the Hangzhou
Reinforcement strain Qiantang River Tunnel are described in detail to present the responses of tunnel linings to fluctuations in
Field measurement river stage and validate the established design model for the former case.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction
The shield-driven tunnelling method is widely adopted for con-
struction of under-river tunnels in soft ground all across the world.
Behaviors of the tunnels beneath the rivers, unlike those beneath
the ground surface, are expected to be affected by fluctuations of
river levels. Nevertheless, there are few studies focusing on the
effect of time-dependent river stage on design of the shield tunnel
linings.
The following loads should always be considered in the design
of the linings: (1) ground pressure, (2) water pressure, (3) dead
load, (4) surcharge and (5) subgrade reaction (British Tunnelling
Society, 2004; ITA WG, 2000; Koyama, 2003; Mashimo and
Ishimura, 2003). Generally speaking, all of these loads are main-
tained stable for tunnels beneath the ground surface that have
been in service for a long period of time. So it is rational to take
these loads as constant in most design methods available
(Duddeck, 1981, 1989). But for tunnels beneath the rivers, imposed
loads constantly change as a result of fluctuations in river stage.
For cases like this, the time-dependent loads induced by changes
in river stage should be taken into consideration in tunnel lining
⇑ Corresponding author. Tel.: +86 150 8868 8010; fax: +86 0571 87169527.
E-mail address: cunganglin@163.com (C. Lin).
design. However, the impacts of changing river stage are seldom
counted in present practice.
Prediction of earth pressures acting on the tunnel lining is one
issue of great importance in design of a tunnel (Kim and
Eisenstein, 2006). In Japan, overburden earth pressure or reduced
earth pressure calculated by Terzaghi’s formula has generally been
adopted as vertical earth pressure acting on the tunnel lining for
the segment design on the basis of previous field measurements
(ITA WG, 2000; JSCE, 1996; Mashimo and Ishimura, 2003).
Murayama (1968) studied the vertical earth pressure in sandy lay-
ers by trapdoor tests. Mashimo and Ishimura (2003) evaluated the
loads on shield tunnel linings in gravel by field measurements at
two shield tunnels. Zhu et al. (2008) simulated relaxation effect
of vertical earth pressure during shield tunnelling. They observed
that the soil arching mainly occurred within 1–2 times the tunnel
diameter above the crown and the vertical earth pressure acting on
the tunnel lining was significantly decreased due to soil arching.
This problem has attracted considerable interest in the last dec-
ades. However, changes in earth pressures acting on the tunnel lin-
ings induced by fluctuations of river levels have rarely been paid
attention to. Moreover, there is a lack of calculation model for
loads acting on the tunnel linings taking time-dependent river
stage into account.
This paper evaluates the influences of river stage on design
of shield tunnel linings with respect to two common ground

http://dx.doi.org/10.1016/j.tust.2014.09.011
0886-7798/Ó 2014 Elsevier Ltd. All rights reserved.
108 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127

conditions: (a) impermeable overburden strata with low perme- p0 ¼ r w Hw ðtÞ ð1Þ
ability and (b) permeable overburden strata with high permeabil-
where p0 is the surcharge imposed by river water; rw is the unit
ity. Two corresponding earth pressure calculation models are
weight of water; Hw(t) is the river stage at time t, which constantly
established. The average uniform rigidity ring method (Koyama,
changes with time.
2003), which has been widely adopted in Japan for shield tunnel
The strata beneath the river are assumed to be completely sat-
lining design, is improved in this study taking stress relief prior
urated with groundwater table at the surface of the river bed. Ver-
to installation of the tunnel lining and time-dependent loads
tical water pressure acting on the crown of the tunnel lining is
induced by changing river stage into account.
figured out using Eq. (2).
In order to assess the impacts of fluctuations in river levels on
shield tunnels beneath the river, instrumentation was set up dur- pw1 ¼ r w C ð2Þ
ing construction of the Hangzhou Qiantang River Tunnel to mea-
where pw1 is the vertical water pressure at the crown of tunnel lin-
sure river stage, earth pressures acting on the tunnel linings,
ing; C is the depth of overburden.
strains of the reinforcing steel bars and convergence of the tunnel.
Vertical earth pressure acting on the tunnel crown is presented
The field observations are described in detail to present behaviors
in Eq. (3).
of the tunnel linings in responses to fluctuations in river stage and
 X 
validate the established design model for the former case. pe1 ¼ ð1  aÞ p0 þ r 0i Hi ð3Þ
This study aims to provide an initial insight into behaviors of
the shield tunnels subject to changing river stage and improve where pe1 is the vertical earth pressure at the crown of tunnel lin-
the average uniform rigidity ring method for design of shield tun- ing; a is the coefficient of stress reduction taking stress relief prior
nel lining taking fluctuations of river levels into account. to installation of the tunnel lining into account; Hi is the thickness
of Stratum No. i, which is located above the tunnel crown, note that
P
2. Design model for shield tunnel beneath river Hi ¼ C; r0 i is the submerged unit weight of soil of Stratum No. i.
The horizontal earth pressure and water pressure are simplified
According to differences in permeability of the tunnel overbur- to be uniformly varying loads that increase with depth acting on
den strata, the design models can be classified into two broad cat- the centroid of the tunnel lining from the crown to the bottom,
egories: (a) impermeable overburden strata with low permeability as shown in Fig. 1. They are calculated with Eqs. (4)–(7),
and (b) permeable overburden strata with high permeability, respectively.
which are referred to as Design model 1 and Design model 2,  X 
h
respectively, hereinafter. qe1 ¼ k p0 þ r 0i Hi þ r0j ð4Þ
2
2.1. Design model 1  
h
qw1 ¼ r w C þ ð5Þ
2.1.1. Computation of imposed loads 2
Fig. 1 depicts the calculation model of loads acting on the tunnel  X X 
lining for Design model 1, in which the overburden strata are qe2 ¼ k p0 þ r 0i Hi þ r 0j Hj ð6Þ
assumed to be impermeable. It is suited for shield tunnels beneath
the rivers with overburden strata of low permeability, such as clay  
h
and silty clay. qw2 ¼ r w C þ D  ð7Þ
2
For the overburden strata are very low in permeability, river
water is anticipated to apply a surcharge that is equal to its own where qe1 and qe2 are the horizontal earth pressure at the crown
weight to the strata below the river bed level. The river water and the bottom of tunnel lining, respectively; qw1 and qw2 are the
induced surcharge p0 is a function of river stage as horizontal water pressure at the crown and the bottom of tunnel

Fig. 1. Calculation model of loads acting on tunnel lining for Design model 1.
 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127 109

Notations: 1. M, N, Q are bending moment, axial force and shear force in the tunnel segment, respectively. 2. g is the effective ratio of bending rigidity of the circular ring. 3. E is modulus of elasticity of tunnel segment. 4. I is moment
lining, respectively; k is the coefficient of lateral earth pressure; D is

When p2 6 h 6 34p ; Q ¼ ð sin h cos h  0:7071 cos2 h sin hÞkdRc

When p4 6 h 6 p2 ; Q ¼ ð sin h cos h þ 0:7071 cos2 h sin hÞkdRc
the outer diameter of tunnel lining;
h is the thickness of tunnel lining; Hj, is the thickness of Stratum

N ¼ ðh  pÞ cos h þ p sin h cos h þ h sin h þ 16 sin h g 1 Rc

No. j, which is located between the crown and the bottom of tunnel
P
lining, note that Hj ¼ D  2h; r0 j is the submerged unit weight of


soil of Stratum No. j.
ðsin h þ 8 sin h cos h  4 sin h cos2 hÞ
As shown in Fig. 1, h is the angle measured with respect to ver-

When 0 6 h 6 p2 ; Q ¼ ðh cos h þ 16 sin hÞg 1 Rc

When 0 6 h 6 p4 ; Q ¼ 0:3536 sin hkdRc

When 34p 6 h 6 p; Q ¼ 0:3535 sin hkdRc

tical axis of the tunnel lining in a counter-clockwise direction. The
horizontal ground reaction is assumed to be triangularly distrib-
uted over a range of h = 45–135° and h = 225–315° (see Fig. 1). It
ðqe2 þ qw2  qe1  qw1 ÞRc
Q ¼ ðqe1 þ qw1 ÞRc sin h cos h

is a product of the coefficient of subgrade reaction and the dis-

Q ¼ ðpe1 þ pw1 ÞRc sin h cos h

placement of the tunnel lining, as shown in Eq. (8).

qr ¼ kd ð8Þ
When p2 6 h 6 p;
Shear force

where qr is the horizontal ground reaction at the spring line of the

Q ¼  16
1

tunnel lining; k is the coefficient of horizontal subgrade reaction; d



is the horizontal displacement of tunnel lining along its spring line.

The dead load is the vertical load acting along the centroid of
the tunnel lining. It is calculated in accordance with Eq. (9).
ðqe2 þ qw2  qe1  qw1 ÞRc

W
g1 ¼ ð9Þ
N ¼ ð0:7071 cos h þ cos2 h þ 0:7071 sin h cos hÞkdRc

2pRc
When p2 6 h 6 34p ; N ¼ ðcos2 h þ 0:7071 cos3 hÞkdRc

N ¼ ðp sin h þ h sin h þ p sin h  16 cos hÞg 1 Rc

where g1 is the dead load of the tunnel lining imposed by its self-
When 0 6 h 6 p2 ; N ¼ ðh sin h  16 cos hÞg 1 Rc
2

weight; W is the self-weight of the tunnel lining per meter in the

When 34p 6 h 6 p; N ¼ 0:3535 cos hkdRc
When 0 6 h 6 p4 ; N ¼ 0:3536 cos hkdRc

longitudinal direction; Rc is the radius of controid of the tunnel

½2ðpe1 þpw1 Þðqe1 þqw1 Þðqe2 þqw2 Þþpg 1 R4c

lining.
ðcos h þ 8 cos2 h  4 cos3 hÞ

The vertical ground reaction can be obtained from Eqs. (10) and
2

(11) according to the balance of forces in the vertical direction.

24ðgEIþ0:0454kRc Þ
4
N ¼ ðqe1 þ qw1 ÞRc cos2 h
N ¼ ðpe1 þ pw1 ÞRc sin h
2

pr ¼ pw1 þ pe1 ð10Þ

When p4 6 h 6 p2 ;

When p 6 h 6 p;

pg1 ¼ pg 1 ð11Þ
Axial force

2
N ¼ 16
1

where pr is the vertical ground reaction in balance with the vertical

d¼

external loads acting on the tunnel lining; pg1 is the vertical ground
reaction in balance with the self-weight of the tunnel lining.
qw1 ÞR2c

Five of the loads mentioned above, pe1, qe1, qe2, qr and pr, are
When p4 6 h 6 p2 ; M ¼ ð0:3487 þ 0:5 sin2 h þ 0:2357 cos3 hÞkdR2c

When p2 6 h 6 34p ; M ¼ ð0:1513  0:5 cos2 h  0:2357 cos3 hÞkdR2c

partially determined by p0, which is a function of river stage.

ðqe2 þ qw2  qe1 

Hence, these five loads are expected to change with river stage.
of inertia of area of tunnel segment. 5. EI is flexural rigidity of tunnel segment.
When 34p 6 h 6 p; M ¼ ð0:2346 þ 0:3535 cos hÞkdR2c

M ¼  18 p þ ðp  hÞ sin h  56 cos h  12 p sin2 h g 1 R2c

When 0 6 h 6 p4 ; M ¼ ð0:2346  0:3536 cos hÞkdR2c

2.1.2. Computation of member forces

When 0 6 h 6 p2 ; M ¼ ð38 p  h sin h  56 cos hÞg 1 R2c

The uniform rigidity ring method, which was first put forward
i

in 1960, is one of the most widely adopted design methods of

 3 cos h  12 cos h þ 4 cos hÞ

shield tunnel lining in Japan. In this method, the flexural rigidity

3

of the circular ring is assumed to be uniform throughout the lining

M ¼ 14 ð1  2 cos2 hÞðqe1 þ qw1 ÞR2c

ring. It was improved later by taking into account the reduction of

M ¼ 14 ð1  2 sin hÞðpe1 þ pw1 ÞR2c
Computational formulas for member forces of tunnel lining.

rigidity due to the presence of joints and the increment of bending

2

moment in the joint area. The improved method, which is gener-

ally referred to as the average uniform rigidity method, is also
widely adopted in Japan. The computational formulas proposed
Bending moment

When p 6 h 6 p;
2

by these two methods (JSCE, 1996, 2010), as listed in Table 1, are

adopted for computation of member forces of the tunnel lining in
48 ð6

h 2

this study. The sign conventions adopted by the proposed methods

1
M¼

are as follows: the bending moment is taken as positive when the

tunnel inner surface is subject to tension; the axial force is taken as
positive when the cross-section of the lining is subject to compres-
(qe2 + qw2  qe1  qw1)

sion; and the shear force is considered as positive when it tends to

make the section rotate in a clockwise direction. As mentioned
above, some of the loads acting on the tunnel lining are time-
(pe1 + pw1)

(qe1 + qw1)

dependent with respect to river stage. Consequently, member

qr = kd

forces of the tunnel lining generated by these loads are affected

Loads
Table 1

by river stage. They are subject to continuous changes with fluctu-

g1

ations in river stage.

110 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127

Fig. 2. Calculation model of loads acting on tunnel lining for Design model 2.

2.2. Design model 2 As can be seen from Eqs. (12), (15) and (17), water pressure act-
ing on the tunnel lining is solely determined by the river stage for a
2.2.1. Computation of imposed loads fixed value of C, D and h. It is anticipated to change with fluctua-
Fig. 2 presents the calculation model of loads acting on the tun- tions in river stage.
nel lining for Design model 2. It is applicable to shield tunnels
beneath the rivers with overburden strata of high permeability, 2.2.2. Computation of member forces
such as sand and gravel. For high permeability of the overburden, Member forces of the tunnel lining for Design model 2 are also
it is assumed that there exists a complete hydraulic connection calculated with the computational formulas listed in Table 1.
between the river water and the groundwater. That is to say, river
water is supposed to generate pore water pressure that is identical
in magnitude to that generated by groundwater of the same height 3. Theoretical calculation
in the underlying aquifers. In this sense, the groundwater table can
assumed to be located at the surface of the river water for compu- The tidal bore at the estuary of the Qiantang River in China is
tation of water pressure acting on the tunnel lining. one of the most spectacular bores in the world. As a result of the
The loads acting on the tunnel lining, as shown in Fig. 2, are cal- tidal bore, water level of the Qiantang River is under constant fluc-
culated with Eqs. (12)–(20). In order to distinguish from previous tuation. Thus, behaviors of the shield tunnels beneath the Qiantang
calculations, loads of the same types to those mentioned in Design River are expected to be largely affected by fluctuations of the river
model 1 are overlined. The other variables have the same defini- stage. One issue of great concern in design of a shield tunnel
tions as before. beneath the Qiantang River is the evaluation of impacts of fluctua-
tions in river stage on the structural soundness of the tunnel over
w1 ¼ r w ½Hw ðtÞ þ C 
p ð12Þ its operational life. The Hangzhou Qiangtang River Tunnel, which
X was put into service in December of 2010, is the first shield tunnel
e1 ¼ ð1  aÞ
p r 0i Hi ð13Þ beneath the Qiantang River. Hence, a case study is conducted with
regard to this tunnel to explore behaviors of the tunnel lining in
X  responses to fluctuations in river stage and examine the above-
h
e1 ¼ k
q r 0i Hi þ r0j ð14Þ established design models.
2

h 3.1. Project overview and geology
w1 ¼ rw Hw ðtÞ þ C þ
q ð15Þ
2
The Hangzhou Qiantang River Tunnel was constructed to carry
X X 
the traffic across the Qiantang River in Hangzhou. It consisted of
e2
q ¼k r 0i Hi þ r 0j Hj ð16Þ
two parallel tubes designed for one-way traffic in opposite direc-
 tions. It was bored by two slurry shields with an outer diameter
h of 11.65 m and a length of 11.4 m. The shield tunnel was con-
w2 ¼ rw Hw ðtÞ þ C þ D 
q ð17Þ
2 structed with a precast reinforced concrete lining, which was com-
posed of nine segmental concrete pieces bolted together by steel
r ¼ kd
q ð18Þ bolts in both the circumferential and longitudinal directions.
The subsurface conditions along the tunnel alignment can be
r ¼ p
p w1 þ p
e1 ð19Þ generally described as a typical Qiantang River alluvial deposit.
Table 2 shows some of the basic soil properties at the site obtained
g1 ¼ pg 1
p ð20Þ from laboratory and in-situ soil tests. The geotechnical strata from
 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127 111

Table 2
Summary of soil properties at the site.

Stratum number Stratum name r (kN m3) K0 c (kPa) u (°) Kv (cm s1) k (104 kN m3) N
1 Made ground 19.0      
2–1 Sandy silt 19.5 0.214 35.9 34.8 1.49  104 2.0 10
2–2 Silt with mucky soil 19.2 0.242 37.0 35.5 1.78  104 1.4 7
3–1 Silty sand with silt 19.5 0.221 36.5 38.7 1.58  104 2.8 14
3–2 Sandy silt 19.8 0.196 42.0 21.7 2.24  104 1.8 9
3–3 Silty sand with silt 19.8 0.212 27.7 30.1 3.11  104 3.0 15
3–4 Clayey silt 19.5 0.177 21.6 18.0 2.46  104 1.0 5
4 Mucky silty clay 18.5 0.557 28.0 12.9 1.45  107 0.8 4
5–1 Silty clay 19.8 0.392 34.9 26.0 1.24  107 2.0 10
5–2 Silty clay 19.0 0.515 35.5 19.0 2.16  107 2.2 11
6–1 Clay 18.8 0.460 39.0 16.9 9.55  108 0.9 4
6–2 Silty clay 19.0 0.379 41.0 20.5 1.77  107 1.6 8
7–1 Silty clay with silty sand 20.1 0.367 55.7 25.7 3.92  107 2.6 13
7–2 Silty fine sand 20.5 0.206 54.7 23.8 2.74  104 4.8 24
8 Round gravel 21.8 0.492   7.50  102 8.0 28

Notations: 1. r is the natural unit weight of soil; 2. K0 is the coefficient of lateral earth pressure at rest; 3. c and u are the cohesion and internal friction angle of soil,
respectively, which are obtained from consolidated undrained triaxial shear test; 4. Kv is the coefficient of vertical permeability; 5. N is the blow counts of standard
penetration test.

Fig. 3. Longitudinal profile of the tunnel and locations of the monitored sections.

Fig. 4. Geological profiles through the four monitored sections.

surface level along the route of the bored tunnels comprised vary-
ing thickness of made ground, sandy silt, clay, fine sand and round
gravel. The tunnel was mainly situated within stratum 3 to stratum
8, including silty sand, clay, fine sand and round gravel.
Four typical cross-sections of the tunnel lining were instru-
mented with total earth pressure cells and reinforcement strain
gauges to observe behaviors of the tunnel in service. Fig. 3 shows
the longitudinal profile of the tunnel and the approximate loca-
tions of the four monitored sections, and Fig. 4 depicts their geo-
logical profiles.
3.2. Computation of imposed loads and member forces
Section 1 as shown in Fig. 4 is selected for analysis. The loads
acting on the tunnel ling and the generated member forces are cal-
culated using above-proposed equations, with regard to Design
model 1 and 2, respectively.
Convergence of the tunnel lining was measured with instru-
mented sections spaced at 50 m intervals along the alignment dur-
ing construction of the Hangzhou Qiantang River Tunnel (Lin et al.,
2011). An increase of the horizontal diameter of the tunnel lining
112 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127

Table 3
Computational parameters.

Parameter Description Value Unit

Dimensions of segment
D 11.30 m
Rc 5.40 m
b Width of segment 2.00 m
h 0.50 m
Ground conditions
C 17.074 m
rw 10 kN m3
c Cohesion of soil in layer 5–2 35.5 kPa
u Internal friction angle of soil in stratum 5–2 19.0 °
K0 Coefficient of lateral earth pressure at rest of soil instratum 5–2 0.515 –
K a ¼ tan2 ðp4  u2 Þ Coefficient of lateral active earth pressure at rest of soil in stratum 5–2 0.509 –
k ¼ 12 ðK 0 þ K a Þ 0.512 –
a 0.2 –
Segmental lining
g1 12.50 kN m2
E 3.45  107 kN m2
I ¼ bh
3
0.020833333 m4
12
g 1 –

of no more than 2 mm was generally observed within one month Table 4

after it has been erected. In the following two months, the tunnel Computation results of imposed loads for Design model 1.

lining was generally observed to contract along its spring line. Load Result (kN m2)
The decrease of its horizontal diameter was generally within pe1 128.678 + 8.0 Hw(t)
8 mm three months after it has been erected. Hence, the horizontal pw1 170.740
ground reaction can be neglected when evaluating the long-term qe1 83.442 + 5.12 Hw(t)
loads acting on the tunnel lining in service. Therefore, the horizon- qw1 173.240
qe2 138.0860 + 5.12 Hw(t)
tal ground reaction is not taken into account in the following
qw2 281.240
computations. pr 299.418 + 8.0 Hw(t)
Table 3 gives the needed parameters for computation. The coef- pg1 39.270
ficient of stress reduction, a, is influenced by many factors, such as
Note: The unit of Hw(t) is m.
the stress-deformation-time characteristics of the surrounding soil,
the amount of deformation of the surrounding soil experienced, and
800
the time at which the actual contact is made between the soil and
the lining. In many instances it is further complicated by the process
700 pr
of filling the voids. Centrifuge model tests conducted by Nomoto
et al. (1999) find that the earth pressure acting on the tunnel is lar-
600
gely dominated by the differences of the construction processes.
Earth pressures acting on the shield segments were measured by pe1
500
Inokuma and Ishimura (1995) and Suzuki et al. (1996). Inokuma
Pressure (kPa)

and Ishimura (1995) concluded that the earth pressure on shield

400 qe2
tunnels was 80% of total cover pressure on shield driven tunnels
in sandy ground. Suzuki et al. (1996) showed that the measured qe1
300
earth pressure was about 40–60% of the total overburden earth qw2
pressure. In engineering practice a is usually determined by experi- 200 qw1
ence gained from full-scale tests. Here a is assumed to be 0.2 based pw1
on previous case studies in similar ground conditions. 100
Shield tests done by Nomoto et al. (1999) show that the value of
pg1
the coefficient of lateral earth pressure (k) largely depends on the 0
tail void thickness and the ground stress around the shield can 0 10 20 30 40 50
Hw(t) (m)
be assumed in an active state when the tail void thickness is large.
The observation demonstrates that the control of the tail void is
Fig. 5. Calculated imposed loads.
crucial in deciding k. The k values for different types of soils are
suggested by RTRI (1997) and JSCE (2010) according to the N val- proposed by JSCE (1996, 2010) that: (1) the value of K0 can be
ues. For clayey soils with N in the range of 8–25, k is suggested regarded as k when the horizontal ground reaction is difficult to
by RTRI (1997) and JSCE (2010) to be 0.40–0.50 and 0.45–0.55, be obtained, and (2) the value of Ka or a reduction of K0 can be used
respectively. ITA WG (2000) suggested that the k value to be used as k when the horizontal ground reaction is available. Following
in the design calculation should be between the value of the coef- these suggestions, the value of k is taken as half of the sum of K0
ficient of lateral earth pressure at rest (K0) and the value of the and Ka, as presented in Eq. (21).
coefficient of lateral active earth pressure (Ka). Zhong (2005) mea-
sured the earth pressures acting on the shield tunnel linings in dif- 1
k¼ ðK 0 þ K a Þ ð21Þ
ferent soil layers. It was generally observed that k is approximately 2
equal to K0 in clayey soils, and k is close to Ka in sandy soils. It was
where Ka is the coefficient of lateral active earth pressure.
 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127 113

1200

1000
θ=0 degree

800 180
600 160

400

M (kN⋅m/m)
200
40
0

-200

-400 120
-600
60

-800 70
80
-1000 90
-5 0 5 10 15 20 25 30 35 40 45 50 55 60
Hw(t) (m)

Fig. 8. Changes of bending moment with river stage at different cross-sections.

Fig. 6. Calculated bending moment of tunnel lining.

1000 50

40

30
500
20
M (kN⋅m/m)

10
5
0
Hw(t)=0

-500

-1000
-20 0 20 40 60 80 100 120 140 160 180
θ (degree)
Fig. 9. Calculated axial force of tunnel lining.
Fig. 7. Distribution of bending moment at different river stages (0–50 m).
4000

3800
Ka can be calculated using equation proposed by Rankine
3600
(Aysen, 2005), that is
Hw(t)=50 m
p u 3400
2
K a ¼ tan  ð22Þ 3200
4 2
40 m
3000
3.2.1. Computation for Design model 1
N (kN/m)

2800 30 m
3.2.1.1. Computation results of imposed loads. Table 4 lists the com-
2600
putation results of imposed loads, which are also presented in 20 m
2400
Fig. 5.
10 m
As can be seen from Fig. 5, pw1, qw1, qw2, and pg1 are not affected 2200
5m
by changes in river stage, but pe1, qe1, qe2 and pr are expected to 2000 0m
increase linearly with a rise of river stage.
1800

1600
3.2.1.2. Computation results of member forces.
0 20 40 60 80 100 120 140 160 180 200
(1) Bending moment
θ (degree)
Fig. 6 presents calculated bending moment of the tunnel lining.
As can be seen from Fig. 6, changes in river stage not only cause Fig. 10. Distribution of axial force at different river stages (0–50 m).
changes in magnitude of the bending moment, but also alter its
distribution along the circumference of the tunnel lining.
Fig. 7 shows calculated bending moment along the circumfer- Fig. 8 presents changes of bending moment with river stage at
ence of the tunnel lining at river stages in the range of 0–50 m. different cross-sections along the circumference of the tunnel lin-
As shown in Fig. 7, both the maximum positive bending moment ing. As obviously shown in Fig. 8, the bending moment at different
and the maximum negative bending moment are found to increase cross-sections is expected to increase linearly with different rates
with rising river stage. with the rise of river stage. For a given river stage, the magnitude
114 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127

4000 400
90
3800 Hw(t)=50 m
120
3600 60 300
40
150
3400
180 200 30
3200 30
20
3000 100
10
θ=0 dergee
N (kN/m)

Q (kN/m)
2800
0 0
2600
16.69 m

2400 -100
Q= 41.955 kN/m
2200
-200
when θ= 90 degree
2000

1800 -300
1600

-5 0 5 10 15 20 25 30 35 40 45 50 55 -400
0 20 40 60 80 100 120 140 160 180
Hw(t) (m) θ (degree)

Fig. 11. Changes of axial force with river stage at different cross-sections. Fig. 13. Distribution of shear force at different river stages (0–50 m).

400
40
30
300 60
20
200

10
100
80
Q (kN/m)

0 and 180
0
90
-100
100
-200 160
150
-300 120
θ =130 degree
-400
-5 0 5 10 15 20 25 30 35 40 45 50 55 60
Hw(t) (m)
Fig. 12. Calculated shear force of tunnel lining.
Fig. 14. Changes of shear force with river stage at different cross-sections.
and location of the anticipated maximum positive and negative
moment of the tunnel lining can be determined from Fig. 8.
Table 5
Computation results of imposed loads for Design model 2.
(2) Axial force
Load Result (kN m2)
Calculated axial force along the circumference of the tunnel lin- e1
p 128.678
ing at river stage of 0–50 m is shown in Fig. 9, as can be seen from w1
p 170.740 + 10 Hw(t)
which, the axial force increases with the rise of river stage at all e1
q 83.4418
w1
q 173.240 + 10 Hw(t)
cross-sections of the tunnel lining. e2
q 133.0860
Fig. 10 presents calculated axial force along the circumference w2
q 281.240 + 10 Hw(t)
of the tunnel lining at river stage of 0, 5, 10, 20, 30, 40 and 50 m, r
p 299.418 + 10 Hw(t)
respectively. Fig. 11 shows changes of axial force with river stage g1
p 39.270
at different degrees (0, 30, 60, 90, 120, 150 and 180 degree, respec-
tively) from the tunnel crown. As shown in Figs. 10 and 11, axial
force of different cross-sections is found to increase linearly at dif-
ferent rates with river stage. When the river stage is below 0–50 m. As can be seen from these figures, shear force of the tunnel
16.69 m, the maximum axial force generally occurs at the invert lining is significantly affected by changes in river stage. The calcu-
of the tunnel lining. As rising of the river stage, the cross-section lated shear force equals to zero at the crown (h = 0°) and the invert
where the maximum axial force appearing is gradually transferred (h = 180°). The shear force at the spring line (h = 90°) is also main-
to the spring line. tained constant, with a value of 41.955 kN per unit meter of seg-
ment. Shear force at the other sections along the circumference is
(3) Shear force found to increase linearly in the rage of h = 0–90° and decrease lin-
early in the range of h = 90–180° with rising river stage. It is noted
Figs. 12–14 present the calculated shear force at different loca- that the increase rate or decrease rate of shear force with river
tions along the circumference of the tunnel lining at a river stage of stage is different at different cross-sections.
 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127 115

800
pr
q w2
700 qw1

600
p w1

500
Pressure (kPa)

400

300 Lines for qw1 and pw1 almost overlap due to

small discrepancy in value between them.
200
qe2
pe1
100
qe1
p g1
0
0 20 40 60
Hw(t) (m)
Fig. 17. Calculated axial force of tunnel lining in Design model 2.
Fig. 15. Calculated imposed loads for Design model 2.

4400 Hw(t)= 50 m

4000
40 m
3600

30 m
N (kN/m)

3200

2800 20 m

2400
10 m
2000

0m
1600
0 20 40 60 80 100 120 140 160 180
θ (degree)

Fig. 18. Distribution of axial force at different river stages (0–50 m) in Design
model 2.

Fig. 16. Calculated bending moment of tunnel lining in Design model 2.

4800
θ =0 degree 180
θ =40 degree 140
3.2.2. Computation for Design model 2 4400
40
θ =140 degree
3.2.2.1. Computation results of imposed loads. Table 5 gives the com-
4000 θ =180 degree
putation results of imposed loads for Design model 2. Fig. 15 shows
their changes with river stage. As can be seen from Fig. 15, water 3600 θ =0 degree
pressure acting on the tunnel lining is expected to increase linearly
N (kN/m)

at the same rate with river stage. Nevertheless, the earth pressures 3200
e1 and q
e1 , q
(p e2 ) acting on the tunnel lining are not affected by
2800
river stage.
2400

3.2.2.2. Computation results of member forces.

2000
(1) Bending moment
Fig. 16 presents calculated bending moment along the circum- 1600
ference of the tunnel lining at river stage in the range of 0–50 m.
As demonstrated in Fig. 16, the bending moment is not dependent 1200
0 10 20 30 40 50
on the river stage. Hw(t) (m)

(2) Axial force Fig. 19. Changes of axial force with river stage at different cross-sections in Design
model 2.

Calculated axial force at cross-sections of different degrees from

the crown under different river stages is presented in Figs. 17–19,
as shown in which, axial force at all cross-sections is found to
increase linearly at the same rate with the rise of river stage. The
maximum axial force occurs at the invert (h = 180°) of the tunnel
lining.
(3) Shear force
The calculated shear force of the tunnel lining under river stage
of 0–50 m is shown in Fig. 20, as can be seen from which, changes
in river stage induce no changes in shear force.
116 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127

ments. Fig. 21 depicts their distributions at the instrumented

cross-sections. Each of the pressure cells and strain gauges is nom-
inated individually as labeled in Fig. 21. Convergence measure-
ment was carried out using a total station instrument to measure
the absolute displacements of the reflective targets mounted on
the inner surface of the tunnel lining along its spring line. For eval-
uation of influences of changes in river level on behaviors of the
tunnel, river stage above the tunnel in the entire year of 2011
was hourly measured. Table 6 lists the measurement items imple-
mented at the four instrumented sections.
The total earth pressure cells were installed at the exterior sur-
face of the tunnel segment. Pressures measured by them are a
combination of effective earth pressures and pore water pressures.
The strain gauges were welded to the main reinforcing steel bars as
sketched in Fig. 22.

Fig. 20. Calculated shear force of tunnel lining in Design model 2. 4.2. Measurement results

4.2.1. River stage

4. Observational verification
4.1. Instrumentation set-up
For verification of the design models established above, field
observations were carried out to measure the total earth pressures
acting on the tunnel linings, strains of the reinforcing steel bars
embedded within the segments and convergence of the tunnel lin-
ings. The instrumentation was set up at four cross-sections along
the alignment of the tunnel as shown in Fig. 3.
The total earth pressure cells and the reinforcement strain
gauges were installed during manufacture of the concrete seg-
4.2.1.1. Monthly changes in river stage. River stage of the Qiantang
River was automatically monitored once per hour throughout
2011 at location above the tunnel. Fig. 23(a) and (b) presents the
hourly monitored and daily averaged results, respectively. Peaks
and troughs of the river stage measured within each month are
referred to as long-term peaks and long-term troughs, respectively,
as labeled in Fig. 23(a) and (b). The observed amplitudes of river
stage between neighboring long-term peaks and troughs generally
fall in the range of 0.629–3.174 m, 1.375 m on average. Fig. 24
illustrates the time intervals between adjacent long-term peaks
and troughs throughout 2011. As shown in it, time intervals
between adjacent long-term peaks and troughs in 2011 mostly

Fig. 21. Arrangement of total earth pressure cells and reinforcement strain gauges.

Table 6
Items of measurements.

Items Description of instrument Number of instrument Recording mode and reading

frequency
Section 1, 2, and 3 Section 4
River stage Automatic monitoring system of Automatically recorded once per
river’s water level hour
Earth pressure Vibrating wire total earth pressure 4/ring, uniformly distributed 9/ring, uniformly distributed Automatically recorded once per
cells along the circumference along the circumference hour
Strains of reinforcing Vibrating wire strain gauges 8/ring, uniformly distributed 18/ring, uniformly distributed Automatically recorded once per
steel bars along the circumference along the circumference hour
Convergence of Total station instrument 1/ring 1/ring Manually recorded once in every
tunnel linings 15 or 30 days
 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127 117

12
Time intervals

10

Time interval (d)

6

2010/12/20 2011/2/20 2011/4/20 2011/6/20 2011/8/20 2011/10/20 2011/12/20

Fig. 22. Location of strain gauges welded to reinforcing steel bars.
range from 6 to 9 days, with an average of 7.40 days. Field observa-
tions reveal that river stage of the Qiantang River fluctuates period-
ically with an average period and amplitude of 14.80 days and
1.375 m, respectively, in the entire year of 2011.
4.2.1.2. Daily changes in river stage. Fig. 25(a) shows hourly-
recorded river stage from 13:00 27th July 2011 to 12:00 25th
August 2011, and Fig. 25(b) presents the corresponding time inter-
vals. Peaks and troughs of river stage observed in each day are
referred to as short-term peaks and short-term troughs, respec-
tively, as labeled in Fig. 25(a). The time required for the river stage
changing from one short-term peak to the next short-term peak,
from one short-term peak to the next short-term trough, and from
one short-term trough to the next short-term peak is referred to as
time interval from peak to peak, from peak to trough, and from
trough to peak, respectively, as shown in Fig. 25(b). As shown in
Fig. 25, river stage fluctuates periodically with an average cycle
time of about 12 h within each day. It is distinguished that the
decline of river stage from one short-term peak to the following
Fig. 24. Time intervals between adjacent long-term peaks and troughs of river
stage in 2011.
short-term trough consumes far more time than its rise from one
short-term trough to the next short-term peak.
Fig. 26 presents time intervals that characterize daily fluctua-
tions in river stage throughout 2011. As shown in it, it generally
takes an average time of about 10.74 h for river stage descending
from one short-term peak to the next trough, whereas only an
average of about 1.65 h for it rising from one short-term trough
to the following peak.
Above analysis demonstrates that river stage of the Qiantang
River fluctuates regularly and periodically within each day and
each month, with an average period of 12.38 h and 14.80 days,
respectively. This is exactly one of the distinguishing characteris-
tics of time dependent river stage of the Qiantang River governed
by the tidal bores.
4.2.2. Total earth pressures
The Hangzhou Qiantang River tunnel was put into service on
28th December 2010. Monitoring of the earth pressures started

9 River stage in 2011

8
Long-term peak
River stage (m)

4 Long-term trough
3
2010/12/20 2011/2/20 2011/4/20 2011/6/20 2011/8/20 2011/10/20 2011/12/20

(a)
9 River stage in 2011
8
River stage (m)

7
Long-term peak
6

4
Long-term trough
3
2010/12/20 2011/2/20 2011/4/20 2011/6/20 2011/8/20 2011/10/20 2011/12/20
(b)
Fig. 23. Field measured river stage of the Qiantang River throughout 2011: (a) hourly monitored; (b) daily averaged.
118 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127

Hourly measured river stage

9
from 13:00 ,7/27 to 12:00,8/25
Short-term peak
8

River stage (m)

7

5
Short-term trough
4

3
2011/7/26 2011/7/30 2011/8/3 2011/8/7 2011/8/11 2011/8/15 2011/8/19 2011/8/23 2011/8/27

(a)
20
From peak to peak
From peak to trough
16 From trough to peak
Time interval (h)

12

0
2011/7/26 2011/7/30 2011/8/3 2011/8/7 2011/8/11 2011/8/15 2011/8/19 2011/8/23 2011/8/27

(b)
Fig. 25. (a) Hourly-monitored river stage from 13:00 27th July 2011 to 12:00 25th August 2011; (b) corresponding time intervals.

28
(1) From peak to peak clude that the annual changes in measured total earth pressures
24 (2) From peak to trough are dominated by the annual changes in temperature. Note that
(3) From trough to peak annual changes in total earth pressures measured by the pressure
20 cells, which are mainly generated by temperature changes, are not
a true representation of the actual changes in total earth pressures
Time interval (h)

16 acting on the tunnel lining. Hence, the components of measured

12 (1)
(2)
8
4 demonstrates daily-averaged total earth pressures measured by
(3)
0 31st December 2011, respectively. It is noted that the earth pres-
2010/12/10 2011/2/10 2011/4/10 2011/6/10
Fig. 26. Time interval for daily fluctuations of river stage.
from 24th February 2011, with a reading frequency of once per
hour. Total earth pressures measured by TB1, TC1, TA2 and TC2 are
analyzed as a typical representation of changes of total earth pres-
sures acting on the tunnel lining with time.
4.2.2.1. Annual changes in total earth pressures. Fig. 27(a) shows
hourly measured earth pressures by TB1, TC1, TA2 and TC2 from
24th February 2011 to 14th December 2012, and Fig. 27(b) shows
daily averaged temperature monitored by a thermometer mounted
on the inner surface of the tunnel lining throughout 2011 and
2012. In fact, temperature changes cause changes in readings of
the total earth pressure cells. Changes in readings for each degree
change in temperature depend on the intrinsic property of each
pressure cell. As shown in Fig. 27, annual changes in total earth
pressures and temperatures follow the almost same tendency
throughout 2011 and 2012. From this point, it is possible to con-
earth pressures caused by temperature changes should be elimi-
nated when evaluating the total earth pressures acting on the tun-
nel lining.
4.2.2.2. Monthly changes in total earth pressures. Fig. 28(a) and (b)
TB1, TC1, TA2 and TC2 and river stage from 24th February 2011 to
2011/8/10 2011/10/10 2011/12/10
sure fluctuate periodically in each month with fluctuations of river
levels at the same time. To better study monthly changes of total
earth pressures, the measurements from two periods (24th Febru-
ary to 10th May 2011, and 1st August 2011 to 31st October 2011)
are shown in Figs. 29 and 30. As can be seen from Figs. 29 and 30,
total earth pressures and river stage fluctuate synchronously in
each month with an average period of about 14.80 days. In this
sense, it can be concluded that monthly fluctuations of total earth
pressures acting on the tunnel lining are mainly governed by
monthly changes in river stage.
4.2.2.3. Daily changes in total earth pressures. Measurements from
two periods (0:00 10th October 2011 to 23:00 21st October
2011, and 0:00 24th October 2011 to 23:00 3rd November 2011)
are presented in Figs. 31 and 32, respectively, to study daily
changes in total earth pressures and their correlations with river
stage. As obviously shown in Figs. 31 and 32, a good synchronism
is observed between changes in earth pressures and river stage.
A statistical analysis of the time intervals for earth pressures
changing from one short-term peak to the next short-term trough
(as shown in Fig. 31) finds that it generally takes about an average
 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127 119

1000
TB1
TC1
800

Earth pressure (kPa)

TA2 TC1
TC2
600

TB1
400 TC2

200 TA2

/20 1/3/
20
1/6/
20
1/9/
20 /20 2/3/
20
2/6/
20
2/9/
20 /20
0/12 201 201 201 1/12 201 201 201 2/12
201 201 201
(a)
40
Average temperature

30
Temperature (°C)

20

10

-10
2/2 0 20 20 20 /20 20 20 20 /20
10/1 1/3/ 1/6/ 1/9/ 1/12 2/3/ 2/6/ 2/9/ 2/12
20 201 201 201 201 201 201 201 201
(b)
Fig. 27. (a) Hourly measured total earth pressures by TB1, TC1, TA2 and TC2 from 24th February 2011 to 14th December 2012; (b) daily changes in temperature of Hangzhou
throughout 2011 and 2012.

800 TB1
TC1 TC1
700
TA2
Earth pressure (kPa)

600 TC2
500
TB1
400
TA2
300

200
TC2

2011/2/20 2011/4/20 2011/6/20 2011/8/20 2011/10/20 2011/12/20

(a)
10
Daily-averaged river stage
8
River stage (m)

0
2011/2/20 2011/4/20 2011/6/20 2011/8/20 2011/10/20 2011/12/20
(b)
Fig. 28. Daily-averaged (a) total earth pressures by TB1, TC1, TA2 and TC2 and (b) river stage from 24th February 2011 to 31st December 2011.

of 1.70 h for earth pressure rising from one short-term trough to the tunnel lining are predominantly governed by daily fluctuations
the following short-term peak. While the decline of earth pressure of river stage.
from one short-term peak to the next short-term trough generally
consumes an average time of 10.70 h. This is exactly consistent 4.2.3. Strains of the reinforcing steel bars
with river stage changes in each day. Judging from this, it can be Measurement of reinforcement strains started from 24th Febru-
concluded that daily changes of total earth pressures acting on ary 2011 with a reading frequency of once per hour. Despite
120 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127

20 35
TC1
650
450 Earth pressure measured by T B1 18 30
Earth pressure measured by T C1 600
Earth pressure by TB1
16 25
Earth pressure measured by T A2 Earth pressure by TC1
400
Earth pressure measured by T C2 14 550 Earth pressure by TA2 20
River stage Earth pressure by TC2

Earth pressure (kPa)

Earth pressure (kPa)

12 500 15
350

River stage (m)

River stage (m)
River stage
10 450 River stage 10
300 8
400 5

6
TB1
0
250 350
4
-5
300
2 TA2
200 -10
250 TC2
0
2011/2/20 2011/3/6 2011/3/20 2011/4/3 2011/4/17 2011/5/1 2011/5/15 -15
2011/10/23 2011/10/25 2011/10/27 2011/10/29 2011/10/31 2011/11/2 2011/11/4

Fig. 29. Daily-averaged total earth pressures and river stage from 24th February
Fig. 32. Hourly measured total earth pressures and river stage from 0:00 24th
2011 to 10th May 2011.
October 2011 to 23:00 3rd November 2011.

20 20
50 YWA1
800
18
15
25
Earth pressure by TB1 16
700 YNA1 10
Earth pressure by TC1 14
0
Earth pressure by TA2
Earth pressure (kPa)

River stage (m)

600 12 -25
River stage (m)

Earth pressure by TC2

Strain (με)

River stage
River stage 10 0
-50
500
8 -5
-75
400 6
-100 -10
4 Strains measured by YWA1
300
-125 Strains measured by YNA1 -15
2
River stage
200 0 -150 -20
/25 /8/8 /22 /9/5 /19 0/3 /17 /31 /14 /14 /14 /14 2/1/
14
2/4/
14
2/7/
14 /14
201
1/7 201
1
201
1/8 201
1
201
1/9
201
1/1 1/10 1/10 201
1/1
201
1/4
201
1/7 1/10 201 201 201 2/10
201 201 201 201

Fig. 30. Daily-averaged total earth pressures and river stage from 1st August 2011 Fig. 33. Hourly monitored strains by YWA1 and YNA1 and river stage.
to 31st October 2011.

700 35 50 25
TC1
650 30
Earth pressure by TB1 YWB1
20
600 Earth pressure by TC1 25 25
YNB1
Earth pressure by TA2
550 20 15
Earth pressure (kPa)

Earth pressure by TC2

River stage (m)

500 15
River stage 0
River stage (m)

10
Strain (με)

Short-term peak of earth pressure

450 10
TB1
400 5 5
-25
Short-term trough of earth pressure River stage
350 0 River stage
0
300 TA2 -5
-50
Strains measured by YWB1
250 TC2 -10 -5
Strains measured by YNB1
200 -15 River stage
1/10
/9 /11 /13 /15 /17 /19 /21 /23 -75 -10
201 1/10 1/10 1/10 1/10 1/10 1/10 1/10
201 201 201 201 201 201 201 14 14 14 /14 14 14 14 /14
1/1/ 1/4/ 1/7/ 1/10 2/1/ 2/4/ 2/7/ 2/10
201 201 201 201 201 201 201 201
Fig. 31. Hourly measured total earth pressures and river stage from 0:00 10th
October 2011 to 23:00 21st October 2011. Fig. 34. Hourly monitored strains by YWB1 and YNB1 and river stage.
 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127 121

40 25 7.5

YNB2 5.0
20
20 2.5

15 0.0

Convergence (mm)
0
YWB2 -2.5

River stage (m)

10
Strain (με)

-5.0
-20
-7.5
5
-10.0
-40 River stage Section 1
0 -12.5 Section 2
-15.0 Section 3
-60 Strains measured by YWB2
-5 Section 4
Strains measured by YNB2 -17.5
/1 /1 /1 /1 /1 /1 /1 /1 /1
River stage 0/12 2011/3 2011/6 2011/9 011/12 2012/3 2012/6 2012/9 012/12
201 2 2
-80 -10
/1/1
4
/4/1
4
/7/1
4 /14 2/1/
14
2/4/
14
2/7/
14 /14
201
1
201
1
201
1 1/10 201 201 201 2/10 Fig. 37. Convergence of tunnel lining from 24th December 2010 to 1st December
201 201
2012.
Fig. 35. Hourly monitored strains by YWB2 and YNB2 and river stage.
trends. That is to say, reinforcement strains tend to increase as
river stage declines, and they tend to decrease as river stage rises.
20 25
River stage 4.2.4. Convergence of the tunnel linings
YWA1 Convergence of the tunnel lining along its spring line was mon-
10 20
YNA1 itored at the four instrumented cross-sections from 24th December
YNB1
2010, with a reading frequency of once in every 15 or 30 days. The
0
YWB1 15
convergence is taken as positive when horizontal diameter of the
tunnel lining increases. The measurement results from 24th
River stage (m)

YNB2 December 2010 to 1st December 2012 are presented in Fig. 37,
Strain (με)

-10
10 as can be seen from which, the tunnel lining is generally observed
to contract along its spring line. The horizontal ground reaction is
-20 neglected in previous theoretical calculations based on tunnel con-
5
vergence measurements during construction. Long-term observa-
River stage
-30
tions of the tunnel in service manifest that it is rational to
YWA1 YNA1 0
neglect the horizontal ground reaction in computation of imposed
YWB1 YNB1 loads and member forces.
-40 YWB2
YWB2 YNB2
-5 4.3. Verification of established design model
2011/10/7 2011/10/13 2011/10/19 2011/10/25 2011/10/31 2011/11/6

Fig. 36. Hourly monitored reinforcement strains and river stage from 10th October The field observations at Section 1 are used to examine the
2011 to 3rd November 2011. above-established design models. As shown in Table 2, permeabil-
ity of the overburden strata is very low, so it is deemed that Design
changes in external loads applied to the tunnel lining, temperature model 1 is suited for this case study.
changes induce significant strains in the reinforcing steel bars.
Similarly, the temperature effect should be eliminated for evalua- 4.3.1. Verification of calculation model of imposed loads
tion of strains in the steel bars induced by river stage changes. Pressure measured by TB1, which will be referred to as pc here-
Figs. 33–35 show reinforcement strains measured by some of the inafter, is the sum of pe1 and pw1. That is
strain gauges welded to the reinforcement steel bars in the tunnel pc ¼ pw1 þ pe1 ð23Þ
segments located at Section 1 and 2 from 24th February 2011 to
A combination of Eqs. (1), (2), (3) and (23) yields
14th December 2012. As shown in these three figures, strains of X
the steel bars imbedded in the tunnel segments typically repeat pc ¼ ð1  aÞr w Hw ðtÞ þ ð1  aÞ r 0i Hi þ rw C ð24Þ
annual cyclic changes in accordance with changes in annual aver- In practice, a is assumed as a constant in the design stage. So the
age temperature. This is consistent with the observations reported second and the third terms on the right-hand side of Eq. (24) for
by Tuchiya et al. (2009) at the Seikan Tunnel in Japan. one certain tunnel section located beneath the river are constants.
A close correspondence is observed between changes in rein- Therefore, it is expected to observe a linear correlation between pc
forcement strains and river stage during the same period. But it and p0 if the relationship described by Eq. (24) exists.
is noted that the strains descend with the rise of river stage. The Pressure measured by TC1, which will be referred to as ps here-
reinforcement strains fluctuate periodically with an average cycle inafter, is the total earth pressure acting on the spring line of the
time of about 14.80 days in each month. This is highly consistent tunnel lining. It can be calculated by Eq. (25).
with monthly fluctuations of river stage.
1
Hourly measurements from 10th October 2011 to 3rd Novem- ps ¼ ½ðq þ qw1 Þ þ ðqe2 þ qw2 Þ ð25Þ
2 e1
ber 2011 are presented in Fig. 36 as a typical example of daily
A combination of Eqs. (1), (4), (5), (6), (7) and (25) gives
changes of reinforcement strains. As can be seen from it, reinforce-
X 1 X 0 1 1
ment strains measured by different strain gauges all fluctuate syn- ps ¼ kr w Hw ðtÞ þ k r 0i Hi þ k rj Hj þ kr 0j h þ rw C þ r w D
chronously with river stage in each day. Reinforcement strains and 2 4 2
river stage fluctuate synchronously, but with opposite changing ð26Þ
122 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127

475 445
pc versus corresponding p0 pc versus corresponding p0
470 recorded at the same time 440 recorded at the same time
Best linear fitting line: R2=0.85971, Best linear fitting line: R2=0.79458,
465
Fitted α=0.17 435 Fitted α=0.16
460

pc (kPa)
430
pc (kPa)

455
425
450

420
445 from 2011/9/26 to 2011/9/28

440 415
from 2011/10/10 to 2011/10/12

435 410
45 50 55 60 65 70 75 80 85 45 50 55 60 65 70 75

p0 (kPa) p0 (kPa)

440 420
pc versus corresponding p0 pc versus corresponding p0
recorded at the same time 415 recorded at the same time
435
Best linear fitting line: Best linear fitting line: R2=0.79929,
410
R2=0.79224, Fitted α=0.24
430 Fitted α=0.18 405
pc (kPa)
pc (kPa)

425 400

395
420
390
415
385
from 2011/10/30 to 2011/11/1
from 2011/10/16 to 2011/10/18
410 380
45 50 55 60 65 70 75 50 55 60 65 70 75 80
p0 (kPa) p0 (kPa)

Fig. 38. pc versus p0 and presentation of the best linear fitting line.

Table 7 calculation model of imposed loads can be carried out through

Fitted results of a and k. examination of the correlations between pc and p0, and between
ps and p0.
Time Fitted a Fitted k K0 Ka
Scatter diagrams presenting pc versus p0 recorded at the same
Value R2 Value R2
time are plotted to determine their relationships (see Fig. 38). In
3.07–3.09 0.22 0.76911 0.45 0.57964 Fig. 38, R2 stands for the adjusted R-square, which is generally
3.21–3.23 0.19 0.69684 0.47 0.55277
the best indicator of the fit quality. Fig. 38 shows that the relation-
4.05–4.07 0.07 0.91076 0.47 0.83723
4.20–4.22 0.26 0.63775 0.46 0.68757 ships between pc and p0 could be satisfactorily described by linear
5.06–5.08 0.1 0.63315 0.43 0.73511 equations with R2 from 0.79224 to 0.85971, indicating a high linear
5.19–5.21 0.09 0.89625 0.48 0.55189 correlation between pc and p0. This is consistent with the relation-
6.21–6.23 0.15 0.76270 0.58 0.87173 ship established by Eq. (24). Fitting results at some other periods
7.16–7.18 0.1 0.57808 0.41 0.71069
are listed in Table 7.
7.19–7.20 0.01 0.78715 0.50 0.82060
8.02–8.04 0.02 0.63139 0.49 0.84981 Similarly, a high linear relationship is observed between ps and
9.01–9.04 0.23 0.66823 0.40 0.63794 p0 (see Fig. 39 and Table 7).
9.14–9.16 0.12 0.68586 0.47 0.85986 It should be noted that p0 is solely determined by the river stage
9.26–9.28 0.17 0.85971 0.53 0.93719
(see Eq. (1)). To conclude, a high linear correlation is generally
9.29–10.01 0.29 0.69468 0.47 0.75159
10.10–10.12 0.16 0.79458 0.54 0.68515 observed between the field-measured total earth pressures by
10.13–10.15 0.26 0.74445 0.51 0.78481 TB1 and the river stage at Section 1. And a high linear correlation
10.16–10.18 0.18 0.79224 0.55 0.71919 also exists between the field-measured total earth pressures by
10.27–10.29 0.36 0.64520 0.52 0.92594 TC1 and the river stage. This validates the established calculation
10.30–11.01 0.24 0.79929 0.50 0.91296
model of imposed loads in Design model 1.
11.14–11.16 0.31 0.78108 0.48 0.91397
11.25–11.28 0.29 0.76014 0.50 0.86439 In this case, the average back-calculated a and k is 0.18 and
0.49, respectively. They are very close to the values suggested in
Average 0.18 0.49 0.515 0.509
Table 3. Wood (1975) recommended a 50% reduction of the over-
burden pressures to account for support delay. The 50% stress
reduction is an arbitrary value, and various suggestions have been
In practice, k is also assumed to be a constant in the design given by others, e.g. about a 33% stress reduction as suggested by
stage. Similarly, there exists a linear relationship between ps and Panet (Kim and Eisenstein, 2006). Einstein and Schwartz (1980)
p0 if Eq. (26) is validated. Therefore, verification of the established suggested that the stress reduction coefficient could be between
 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127 123

700 750
ps versus corresponding p0 ps versus corresponding p0
recorded at the same time recorded at the same time
745
695 Best linear fitting line: R2=0.82060, Best linear fitting line: R2=0.75159,
Fitted λ=0.50 Fitted λ=0.47
740
690

ps (kPa)
ps (kPa)

735

685
730

680
725

from 2011/7/19 to 2011/7/20 from 2011/9/29 to 2011/10/1

675 720
45 50 55 60 65 70 75 80 50 55 60 65 70 75 80 85 90

p0 (kPa) p0 (kPa)

700 695
ps versus corresponding p0 ps versus corresponding p0
recorded at the same time recorded at the same time
695 690
Best linear fitting line: R2=0.78481, Best linear fitting line: R2=0.71919
Fitted λ=0.51 Fitted λ=0.55
690 685
ps (kPa)

685 ps (kPa) 680

680 675

675 670

from 2011/10/13 to 2011/10/15 from 2011/10/16 to 2011/10/18

670 665
50 55 60 65 70 75 80 45 50 55 60 65 70 75
p0 (kPa) p0 (kPa)

Fig. 39. ps versus p0 and presentation of the best linear fitting line.

20
(a) Measured strain by YWA1
(b) Initial calculated results (d) River stage
15 4
(c) Calculated results with Rm=0.1

10 0

5 (a)
-4
Strain (με)

(c)
0

Hw(t) (m)
Fig. 40. Section of segment and arrangement of steel bars. -8

-5
-12
15% and 100% according to simple analytical and numerical tech-
niques and case study. The back-calculated a in this case study falls -10 (b)

in the ranges suggested by Einstein and Schwartz, but it is close to -16

the lower limit. And it is smaller than the values suggested by -15
Wood and Panet. Compared to former studies, the stress relief is -20
rather smaller, which is most probably due to the superior tunnel- -20
ling technique of slurry shield. 2011/10/24 2011/10/26 2011/10/28 2011/10/30 2011/11/1 2011/11/3 2011/11/5

Note that the fitted k at different periods is very close to K0 and

Fig. 41. Field measured reinforcement strains by YWA1 and calculated results from
Ka. This also gives a strong support to the calculation model. 25th October 2011 to 3rd November 2011.

4.3.2. Verification of calculation model of member forces

The formulas listed in Table 1 are adopted for computation of
member forces in this case study. Computation results for Design
model 1 reveal that M, N and Q at different cross-sections along
the circumference are all expected to change linearly with river
stage. The calculation model of member forces will be verified by
comparison of field observed reinforcement strains by YWA1,
YNA1 and YWB1 and their results from theoretical calculations.
Fig. 40 depicts the arrangement of reinforcement steel bars in
the segment. Where, As and A0 s stands for cross-sectional area of
steel bars at the inner and outer surface of segment, respectively.
Strains of the steel bars are calculated following methods sug-
gested by JSCE (2010) based on the member forces obtained from
above equations. As only changes of reinforcement strains in
response to changes of river stage are concerned, the values of
124 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127

15 40 8
Field-measured strains by YWA1 versus river stage
(a) Measured strain by YNA1
at the same time from 2011/10/5-2011/11/3 35
(b) Initial calculated results (d) River stage
Best linear fitting line: R2=0.93188, slope=-2.173 4
10
30 (c) Calculated results with Rm=0.1
Initial calculated strains versus river stage
Calculated strains with Rm=0.1 versus river stage Slope=8.913
25 0
5
20
Strain (με)

15 (b) -4
0

Strain (με)
10

Hw(t) (m)
Slope= 1.297 (c) -8
-5 5

Slope= 2.173 0 -12

-10
-5 (a)
-16
-10
-15
-0.25 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75
-15
Hw(t) (m) -20
-20
Fig. 42. Scatter diagram presenting measured and calculated strains by YWA1 2011/10/24 2011/10/26 2011/10/28 2011/10/30 2011/11/1 2011/11/3 2011/11/5
versus river stage from 2011/10/30–2011/10/31.
Fig. 43. Field measured reinforcement strains by YNA1 and calculated results from
25th October 2011 to 3rd November 2011.

monitored strains and river stage on 24th February 2011 are both
set as datum. 30
Fig. 41 shows the measured reinforcement strains from 25th Field-measured strains by YNA1 versus river stage
October 2011 to 3rd November 2011 by YWA1 and the correspond- 25 at the same time from 2011/10/25-2011/11/3
Best linear fitting line: R2=0.93353, slope=-3.232
ing calculated results. Fig. 42 is a corresponding scatter diagram Initial calculated strains versus river stage
20
presenting measured and calculated strains by YWA1 versus river Calculated strains with Rm=0.1 versus river stage
stage at the same time. The initial calculated results, as shown in 15
Fig. 41, are reinforcement strains computed under member forces Slope=14.277
10
Strain (με)

obtained using formulas in Table 1. As shown in Figs. 41 and 42,

strains of the reinforcement steel bars are observed to decrease lin- 5
early with an increase in river stage. The initial calculated results, Slope=3.616
though, reflect the linear correction between changes in reinforce- 0

ment strains and river stage. However, they deviate considerably -5

from the observed values. Thus, the initial calculated results fail Slope=3.232
to give a reasonable prediction of changes in reinforcement strains -10
with river stage. This is mostly probably attributed to the fact that
-15
the bending moment of the lining segment is over-estimated. -0.25 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75
Peck (1969) pointed out that the bending moments produced in Hw(t) (m)
the tunnel linings largely depend on their rigidity. For shield-dri-
ven tunnel linings used in practice, the deduced bending moments Fig. 44. Scatter diagram presenting measured and calculated strains by YNA1 versus
are significantly weakened by the presence of joints. The develop- river stage from 2011/10/30–2011/10/31.

ment of the bending moment in a jointed tunnel lining is signifi-

cantly affected by the joint stiffness and the number of segments
20 8
in each ring (Lee and Ge, 2001; Lee et al., 2001; (a) Measured strain by YWB1
Teachavorasinskun and Chub-uppakarn, 2010). The jointed ring 15 (b) Initial calculated results (d) River stage
carries smaller value of the bending moment as compared with a 4
(c) Calculated results with Rm=0.1
10
continuous ring. The reduction of the bending moment generated
in a jointed lining due to the existence of segment joints was taken 5 0
into account by introducing a coefficient called bending moment
0
ratio Rm, which was defined as (Lee et al., 2001) -4
(b)
Strain (με)

-5
Hw(t) (m)

(c) -8
jMaximum bending moment of the jointed tunnelj -10
Rm ¼
jMaximum bending moment of the continuous tunnelj -15
-12
ð27Þ
-20 (a)
Numerical studies show that the number and the orientation of -16
-25
joints have a significant influence on the maximum bending
moment in a segmental lining (Do et al., 2013; Hefny and Chua, -30
-20
2006). Generally, the higher the joint number, the lower the max-
-35
imum bending moment. Parametric studies show that the maxi- 2011/10/24 2011/10/26 2011/10/28 2011/10/30 2011/11/1 2011/11/3 2011/11/5
mum bending moment induced in the lining becomes negligible
small when the joint number exceeds 8 (Hefny and Chua, 2006). Fig. 45. Field measured reinforcement strains by YWB1 and calculated results from
25th October 2011 to 3rd November 2011.
The maximum bending moment produced in a 9-joint tunnel lin-
ing is no more than 10% of that in a 3-joint lining.
 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127 125

20 200000
Field-measured strains by YWB1 versus river stage (a) Inner bars
15 at the same time from 2011/10/25-2011/11/3 175000 (b) Outer bars (b)
Best linear fitting line: R2=0.95354, slope=2.590
10 Initial calculated strains versus river stage (c) Inner surface of segment
150000
Calculated strains with Rm=0.1 versus river stage (d) Outer surface of segment
5
125000 (e) Allowable compressive
Slope=12.797
0 strength of concrete:

Stress (kN⋅m-2)
Strain (με)

100000 fc= 23100 kN⋅m-2

-5

-10 75000

-15 Slope=2.587 50000

194.303 m (a)
-20
Slope=2.590 25000 (e)
-25 (d)
0 (c)
-30
-0.25 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 -25000
Hw(t) (m) 0 50 100 150 200 250 300
Hw (t)
Fig. 46. Scatter diagram presenting measured and calculated strains by YWB1
versus river stage from 2011/10/30–2011/10/31. Fig. 48. Calculated member stresses at h = 0° with river stage (Rm = 0.1).

As mentioned previously, the proposed methods by JSCE (1996,

2010) for computation of member forces are based on the assump- safety of the tunnel lining (with regarding to Section 1 as shown in
tion that the flexural rigidity of the circular ring is uniform Fig. 4) are evaluated with respect to Design model 1 and Design
throughout the lining. In other words, the tunnel lining is simpli- model 2, respectively.
fied as a continuous ring. According to the above-mentioned stud-
ies, the calculated bending moment in a jointed tunnel lining using
formulas in Table 1 will be larger than its actual value. To account
4.4.1. Design model 1
the effect of segment joints on bending moment in the tunnel lin-
Fig. 47 presents the predicted stresses of the concrete and the
ing, strains of the reinforcement steel bars are recalculated under
steel bars in the cross-section of 90 degree from the tunnel crown.
different values of Rm. As can be seen from Figs. 41 and 42, the cal-
The member stresses are expected to increase linearly with an
culated strains with Rm = 0.1 are highly consistent with the
increase in river stage. When the river stage reaches a height of
observed results.
146.256 m, the stress of the concrete of the inner surface reaches
Analysis of the reinforcement strains measured by YNA1 and
the allowable compressive strength (fc = 23,100 kN m2).
YWB1 (see Figs. 43–46) also demonstrates that the calculated
Fig. 48 presents the predicted stresses in the cross-section at the
results with Rm = 0.1 can give a reasonable estimate of the rein-
tunnel crown. Similarly, member stresses increase linearly with
forcement strains induced by changes in river stage. Hence, it
river stage. The allowable compressive strength of concrete is
can be concluded that: (1) the bending moment calculated by for-
reached under the river stage of 194.303 m. Therefore, for struc-
mulas in Table 1 is over-estimated due to ignorance of the effect of
tural safety of the tunnel lining, the limit river stage is 146.256 m.
joints, and (2) after an appropriate reduction of the bending
It should be noted that the calculated stresses are significantly
moment, the member forces calculated using formulas in Table 1
affected by the value of Rm. Here it is assumed to be 0.1 based on
can be used for stress and strain analysis.
field observations. In design of similar shield tunnel linings
beneath the river, the value of Rm should be cautiously determined
4.4. Structural safety evaluation
according to the joint stiffness and the number of segments in each
ring.
Based on the above-established calculation models of imposed
loads and member forces, influences of river stage on the structural
275000 250000
(a) Inner bars Allowable strength of (a) Inner bars
250000 225000
(b) Outer bars reinforcement: (b) Outer bars
(a) (b)
225000 fy= 300000 kN⋅m-2 200000 (c) Inner surface of segment
(c) Inner surface of segment
(a)
200000 (d) Outer surface of segment 175000 (d) Outer surface of segment
175000 (e) Allowable compressive (e) Allowable compressive
150000
strength of concrete: fc= 23100 kN⋅m-2
Stress (kN⋅m-2)

strength of concrete:
Stress (kN⋅m-2)

150000
fc= 23100 kN⋅m-2 125000
125000
100000
100000 (b)
75000
75000
50000 (d)
50000 146.256 m 180.155 m
(c)
25000 (c)
25000 (e)
(d)
(e)
0 0

-25000 -25000
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Hw (t) Hw (t)

Fig. 47. Calculated member stresses at h = 90° with river stage (Rm = 0.1). Fig. 49. Calculated member stresses at h = 90° with river stage (Design model 2).
126 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127
250000
(a) Inner bars
225000
(b) Outer bars
200000 (c) Inner surface of segment
175000 (d) Outer surface of segment
(e) Allowable compressive
150000
strength of concrete: fc= 23100 kN⋅m-2
Stress (kN⋅m-2)
125000
100000
75000
50000
194.165 m
25000
0 to a height of 146.256 m. If Design model 2 is applicable,
-25000
0 50 100 150 200
Hw (t)
Fig. 50. Calculated member stresses at h = 0° with river stage (Design model 2).
4.4.2. Design model 2
Figs. 49 and 50 show the predicted member stresses in cross-
sections at the spring line and the crown, respectively. The limit
river stage for Design model 2 is 180.155 m.
5. Conclusions
Two design models for shield tunnel linings beneath the rivers
are established taking fluctuations of river sage into account. The
former model is validated by field observations. On the basis of
the field-observed and calculated results, the following conclusions
can be drawn:
(1) The loads acting on the tunnel lining and the generated
member forces in the tunnel lining are significantly affected
by river stage. When the overburden strata are very low in
permeability, the earth pressures acting on the tunnel lining
are expected to change linearly with river stage. In addition,
member forces of the tunnel lining also change linearly with
river stage. When permeability of the overburden strata is
very high, water pressure acting on the tunnel lining is antic-
ipated to change linearly with river stage. In this case, the
axial forces of the tunnel lining change linearly with river
stage, leaving the bending moments and shear forces
unchanged.
(2) River stage of the Qiantang River fluctuates regularly and
periodically within each day and each month, with an aver-
age period of 12.38 h and 14.80 days, respectively.
(3) There is a close correspondence between fluctuations in
total earth pressures acting the tunnel lining and river stage
of the Qiantang River. The monthly and daily changes in
total earth pressures are predominantly governed by fluctu-
ations in river stage.
(4) A good synchronism is observed between changes in
observed strains of the reinforcing steel bars embedded in
the segments and river stage of the Qiantang River. The rein-
forcement strains change linearly with river stage.
(5) The established calculation model of loads acting on the tun-
nel lining for overburden strata of low permeability is vali-
dated by field-observations. The average back-calculated a
and k is 0.18 and 0.49, respectively. The value of k is very
close to K0 and Ka.
(6) The bending moment in the tunnel lining calculated using
formulas proposed by JSCE (1996, 2010) is over-estimated
due to neglecting the effects of segment joints. After an
appropriate reduction of the bending moment, the member
forces obtained using formulas proposed by JSCE can be
(b) applied for stress and strain analysis of the tunnel lining
with great accuracy. In this case study, the value of bending
(a) moment ratio Rm is assumed to be 0.1 based on field obser-
vations. It should be noted that the calculated stresses of the
tunnel lining are significantly affected by the value of Rm. In
design of similar shield tunnel linings beneath rivers, the
value of Rm should be determined with great caution accord-
ing to the joint stiffness and the number of segments in each
ring.
(d) (7) For Section 1 in Hangzhou Qiantang River tunnel, stress of
(c) the concrete of the inner surface is expected to reach the
(e) allowable compressive strength when the river stage comes
when the overburden strata are assumed to be completely
250 300
hydraulically connected to the river water, the limit river
stage is 180.155 m.
(8) The theoretical design models established in this study pro-
vides an easy means to assess structural responses of shield
tunnel linings to changing river stages. The advantage of
these simple theoretical models is that they can be applied
for a first order covering of the dominant mechanisms and
for preliminary design. However, to take into account the
complexity of geometry, geology and construction process
in practice, numerical models are necessary for a final
design. For a deeper insight into structural design of shield
tunnel lining taking fluctuations of river stage into account,
numerical analyses are to be conducted in the further study.
Acknowledgements
The authors appreciate the help from the staff of Hangzhou
Qing-chun Road Cross-river Tunnel Company Ltd. during instru-
mentation set-up and data acquisition. This study is sponsored
by K.C. Wong Magna Fund in Ningbo University.
References
Aysen, A., 2005. Soil Mechanics: Basic Concepts and Engineering Applications.
Taylor & Francis.
British Tunnelling Society, 2004. Tunnel Lining Design Guide. Thomas Telford
Services Limited, London.
Do, N.A., Dias, D., Oreste, P., Djeran-Maigre, I., 2013. 2D numerical investigation of
segmental tunnel lining behavior. Tunn. Undergr. Space Technol. 37, 115–127.
Duddeck, H., 1981. Structural design models of tunnels in soft soil in use at present
in different countries. In: Proceedings of the International Symposium on Weak
Rock. Japan Society of Civil Engineers, Tokyo.
Duddeck, H., 1989. Guidelines for tunnel design. Tunn. Tunn. Int. 21 (12), 42–44.
Einstein, H.H., Schwartz, C., 1980. Discussion on simplified analysis for tunnel
supports. J. Geotech. Eng. Div. ASCE GT7, 835–838.
Hefny, A.M., Chua, H.C., 2006. An investigation into the behaviour of jointed tunnel
lining. Tunn. Undergr. Space Technol. 21 (3), 428-428.
Inokuma, A., Ishimura, T., 1995. Earth pressure acting on shield driven tunnels in
soft ground. In: Fujita, Kusakabe (Eds.), Proc., Underground Constr. in Soft
Ground, Balkema, Rotterdam, The Netherlands, pp. 221–224.
ITA, Wg, 2000. Guidelines for the design of shield tunnel lining. Tunn. Undergr.
Space Technol. 15 (3), 303–331.
Japanese Society of Civil Engineers (JSCE), 1996. Japanese Standard for Shield
Tunnelling. third ed.
Japanese Society of Civil Engineers (JSCE), 2010. Design of Shield Tunnel Lining.
Revised ed. (in Japanese).
Kim, H.J., Eisenstein, Z., 2006. Prediction of tunnel lining loads using correction
factors. Eng. Geol. 85 (3), 302–312.
Koyama, Y., 2003. Present status and technology of shield tunneling method in
Japan. Tunn. Undergr. Space Technol. 18 (2), 145–159.
Lee, K.M., Ge, X.W., 2001. The equivalence of a jointed shield-driven tunnel lining to
a continuous ring structure. Can. Geotech. J. 38 (3), 461–483.
Lee, K.M., Hou, X.Y., Ge, X.W., Tang, Y., 2001. An analytical solution for a jointed
shield-driven tunnel lining. Int. J. Num. Anal. Met. 25 (4), 365–390.
Lin, C.G., Wu, S.M., Zhang, Z.M., Li, Z.L., 2011. Monitoring and analysis of
deformations and movements of slurry shield tunnel linings. In: Proceedings
 C. Lin et al. / Tunnelling and Underground Space Technology 45 (2015) 107–127
of the International Conference on Electric Technology and Civil Engineering,
Lushan, China, Institute of Electrical and Electronics Engineers. pp. 2379–2383.
Mashimo, M., Ishimura, T., 2003. Evaluation of the load on shield tunnel lining in
gravel. Tunn. Undergr. Space Technol. 18 (2), 233–241.
Murayama, S., 1968. Earth pressure on vertically yielding section in sand layer. DPRI
Annuals, No. 11 (B), Kyoto University, pp. 549–565.
Nomoto, T., Imamura, S., Hagiwara, T., Kusakabe, O., Fujii, N., 1999. Shield tunnel
construction in centrifuge. J. Geotech. Geoenviron. Eng. 125 (4), 289–300.
Peck, R.B., 1969. Deep excavation and tunnelling in soft ground. In: Proc. 7th Int.
Conf. Soil Mech. & Found. Eng., Mexico City, State of the Art Volume, pp. 225–
290.
Railway Technical Research Institute (RTRI), 1997. Design Standard for Railway
Structures (Shield-Driven Tunnel). Maruzen, pp. 47–61 (in Japanese).
Suzuki, M., Komada, T., Nakagawa, H., Hashimoto, T., Satsukawa, Y., 1996.
Measurement of earth and water pressures acting on the great depth tunnel
127
segment. In: Mair, R.J., Talor, R.N. (Eds.), Proc., Int. Symp. on Geotech. Aspects of
Underground Constr. in Soft Ground, Balkema, Rotterdam, the Netherlands, pp.
613–619.
Teachavorasinskun, S., Chubuppakarn, T., 2010. Influence of segmental joints on
tunnel lining. Tunn. Undergr. Space Technol. 25 (4), 490–494.
Tuchiya, Y., Kurakawa, T., Matsunaga, T., Kudo, T., 2009. Research on the long-term
behavior and evaluation of lining concrete of the Seikan Tunnel. Soils Found. 49
(6), 969–980.
Wood, A.M.M., 1975. The circular tunnel in elastic ground. Geotechnique 25 (1),
115–127.
Zhong, X.C., 2005. Research on earth pressure for shield tunnel lining. Doctoral
dissertation, HoHai University (in Chinese).
Zhu, W., Zhong, X.C., Jia, R., 2008. Simulation on relaxation effect of vertical earth
pressure for shield tunnels by particle flow code. Chin. J. Geotech. Eng. 30 (5),
750–754 (in Chinese).