Journal Pre-proofs

Research Paper

Analytical model regarding compression-bending capacity of segmental joint

reinforced by steel plate

Zhen Li, Xuezeng Liu

PII: S2467-9674(24)00045-X
DOI: https://doi.org/10.1016/j.undsp.2024.01.005
Reference: UNDSP 447

To appear in: Underground Space

Received Date: 27 April 2023

Revised Date: 11 January 2024
Accepted Date: 22 January 2024

Please cite this article as: Z. Li, X. Liu, Analytical model regarding compression-bending capacity of segmental
joint reinforced by steel plate, Underground Space (2024), doi: https://doi.org/10.1016/j.undsp.2024.01.005

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2024 Tongji University. Publishing Services by Elsevier B.V. on behalf of KeAi Communications Co. Ltd.
 Analytical model regarding compression-bending capacity of segmental joint

reinforced by steel plate

Zhen Lia, Xuezeng Liua,b*

aDepartment of Geotechnical Engineering, College of Civil Engineering, Tongji University, Shanghai 200092, China

bKey Laboratory of Geotechnical and Underground Engineering of Ministry of Education, Tongji University, Shanghai 200092, China

*Corresponding author: [email protected]

Abstract

The mechanical properties of the steel-plate-reinforced segmental lining are generally determined by the load-
bearing capacity of reinforced joints. However, there is a lack of valid calculation methods for compression-bending
bearing capacity, and researchers mainly rely on experience and analogy for the design of reinforced joints. This paper
proposes an analytical model based on the deformation and stress characteristics of the joint surface to calculate the
compression-bending capacity of the steel-plate-reinforced joint. After verifying the applicability of this analytical
model through finite element (FE) simulations, the evalution rules of the load-bearing capacity of the reinforced joint
were attained, followed by a quantitative investigation into the influence of joint parameters on it. The results show that:
(1) the bearing capacity curve of the reinforced joint under different axial forces can be separated into two parts, with
the maximum ultimate bending moment found at the demarcation point, where the steel plate yielding and joint failure
occur simultaneously; (2) the steel plate strength and cross-sectional area have a strong influence on the bearing
capacity of the reinforced joint when the axial force is under 0.15RFF, where RFF is the axial force at pure-
compression failure); (3) the concrete strength and segment width have a prominent influence on the curve when the
axial force is over 0.30RFF; (4) the impact of the fictitious strain, bolt strength, bolt diameter, and bolt location on the
bearing capacity is minimal in range and amplitude.

Keywords: Shield tunnel; Steel plate reinforcement; Segmental joint; Analytical model; Compression-bending bearing
capacity

1 Introduction

During long-term services of shield tunnels, due to material deterioration, surface surcharge, and other factors, the
load-bearing performance of lining structures may degrade, increasing the risks to their long-term safety. Subsequent
and timely maintenance is needed to reinforce the lining structure and enhance its load-bearing capacity (Shi & Li,
2015; Huang et al., 2017; Liu et al., 2018a). Compared with setting carbon fiber reinforced polymer (CFRP) and cover
arches (Elmalich & Rabinovitch, 2009; Liu et al., 2016, 2020; Li et al., 2021), steel plate reinforcement is convenient
due to its minimal space occupation. It also offers good tensile performance and has been widely applied in the field of
shield tunnel reinforcement. The segmental joint is a weak part of segmental linings, and its load-bearing capacity after
reinforcement will determine the mechanical performance of the integral segmental lining. Existing design methods are
mainly applicable to original joints and heavily rely on the experience and knowledge of engineers, which may result in
a waste of materials or insufficiency of bearing capacity. In such a context, a new analytical model is proposed to
calculate the load-bearing capacity of segmental joints that are reinforced by steel plates.
 Studies in recent years have focused on the load-bearing capacity of steel-plate-reinforced segmental linings (Zhao
et al., 2016; T. Liu et al., 2017; D. Zhang et al., 2019; Zhai et al., 2020), primarily using methods like tests and
numerical simulations that are labor-intensive and time-consuming. A few researchers have studied the calculation
theories concerning reinforced lining structures. De Lorenzis et al. (2006) and Liu et al. (2019) obtained multiple
calculation models of the interfacial stress distribution between the thin plate and curved beam to better understand the
interfacial failure mechanism, but this is unable to calculate the bearing capacity of the reinforced section; J. Zhang et
al. (2019) proposed a systematic analytical model to obtain the ultimate bending moment of segments reinforced by
steel-concrete composite (SCC) based on experimental data, but this model does not fit reinforced joints; She et al.
(2020) suggested four failure states of the reinforced segment section and deduced the corresponding calculation
method for ultimate bending moment, but this also is not applicable to reinforced joints; Zhai et al. (2022) proposed an
analytical solution for steel plate reinforced tunnel linings, taking various interface slip modes into account. This
method has positive significance for the rapid calculation of the bearing capacity of the whole segmental ring, but it has
limitations because it does not consider the structural details of the joints; Liu et al. (2023) proposed a method for
calculating the bearing capacity of reinforced segment considering steel plate corrosion, which can predict the
evolution rule of bearing capacity of reinforced lining. However, this method is only applicable to the segment body,
not the joint, and the existing structural damage before reinforcement is not considered. Some scholars have also
carried out research on the analytical solution for tunnels with double linings, which has certain reference values for the
solution of reinforced linings. For example, EL Naggar et al. (2008) proposed a closed-form analytical solution for a
composite lining of a circular tunnel in deep burial conditions, in which the stress and displacement findings of a
double-layered circular tunnel lining were presented using thick-wall cylinder theory.

To sum up, very few scholars have performed in-depth research on the calculation method concerning the value of
the bearing capacity of reinforced joints. Therefore, it is of significance to establish an analytical model to calculate the
compression-bending bearing capacity of reinforced joints, which helps quickly determine the ultimate bending
moment while providing guidance for reinforcement design. And L. Zhang et al. (2019) carried out the theoretical
research on the mechanical performance of segmental joints without reinforcement, which provided reference for the
related research of reinforced joints.

This paper built up a calculation model to analyze the bearing capacity of the steel-plate-reinforced joint and its
sensitivity to design parameters. First, based on the classical theory of concrete structures, a quantitative analytical
model was proposed for calculating the compression-bending bearing capacity of the reinforced joint. Then, the
reliability of this analytical model was verified. Finally, the load-bearing capacity curves of the reinforced joint under
different axial forces were attained, and the sensitivity of the ultimate bending moment to various parameters was
examined using the proposed model.

2 Analytical model

A shield tunnel comprises multiple segments, each including a joint surface, bolts, and other components. The
joint surface is usually not designed to be flat for waterproofing and positioning purposes, which often causes
incomplete contact (discontinuity) between adjacent concrete pieces under external loads. Such discontinuity results in
a lower load-bearing capacity compared with a similar-sized concrete section. Figure 1 indicates the discontinuous
features found on the joint surface, where h refers to the cross-section height of the joint, d refers to the distance

between the bolt and the surface's outer edge, d1 refers to the outer edge of the joint, d 2 refers to the outside concrete,

2
d3 refers to the waterproof zone, d 4 refers to the core concrete, and d5 refers to the inner edge of the joint. Note that

d 2 and d 4 are the actual compression zones of the joint surface. In contrast to the general methods that simplify the

joint surface as a flat section (Zhu et al., 2014; Zhang & Li, 2020), the approach adopted in this paper takes into
account the discontinuous features, which is crucial in the quantitative calculation of the ultimate bending moment.

(a)

(b)

Fig. 1. Discontinuous features of the joint surface. (a) Without reinforcement, and (b) after reinforcement.

2.1 Basic assumptions

To simplify the theoretical derivation process and analytical formulas, a series of assumptions were made.

(1) The plane-section assumption, the basis of the classical theory of concrete structures, was selected to describe
joint surface deformation (Li et al., 2015; Xiao et al., 2019; L. Zhang et al., 2019; Y. Zhang et al., 2020).

(2) The actual compression zones were assumed to be in close contact, and the influence of the prestress to bolts
on the load-bearing capacity of the joint was neglected (X. Liu et al., 2017; Xiao et al., 2019; L. Zhang et al., 2019).
 (3) The ultimate state was determined at the point when the joint surface reaches the ultimate strain of concrete (Li
et al., 2015; L. Zhang et al., 2019; Guo et al., 2023).

(4) For the purpose of predicting the flexural capacity of the joint with a safety margin, the slippage and failure of
the bond surface between the concrete and steel plate were not considered (Liu et al., 2023; Y. Zhang et al., 2020; She
et al., 2020; J. Zhang et al., 2019).

(5) The bolts and steel plate were assumed to bear the tensile stress only, that was, the contribution to the
compressive capacity of the joint was neglected (Li et al., 2015; Xiao et al., 2019; L. Zhang et al., 2019; She et al.,
2020).

(6) The joint opening before reinforcement was moderate, that was, the height of the compression zone was no

less than d1 +d 2 +d3 .

2.2 Constitutive model

The constitutive model is vital to the accuracy of the calculation results. This paper employs the constitutive
model recommended by Hongnestad, which consists of a parabolic section and a linear-descending section, as shown in
Eq. (1) and Fig. 2.

ì é æ e ö2 ù
ï f c ê1 - ç1 - c ÷ ú , e c £ e c0 ,
ï ê è e c0 ø úû
s c e c  = í ë (1)
ï é æ e c - e c0 ö ù
ï f c ê1 - 0.15 ç e - e ÷ ú , e c0 < e c £ e cu ,
î ë è cu c0 ø û

where sc and e c are the concrete stress and strain, respectively; f c refers to the concrete's yield stress; e c0 = 0.002
refers to concrete's yield strain; e cu = 0.0033 refers to the concrete's ultimate strain.

Fig. 2. Constitutive model of concrete.

4
 The constitutive models adopted for the bolts and steel plate are considered bilinear, which can be described as
Eqs. (2) and (3), respectively.

ïì f my × e m e my , 0 £ e m £ e my ,
s m e m  = í (2)
ïî f my , e my < e m ,

ïì fspy × e sp e spy , 0 £ e sp £ e spy ,

s sp = í (3)
ïî fspy , e spy < e sp ,

where s m and s sp are the bolt and steel plate stress, respectively; e m and e sp are the bolt and steel plate strain,

respectively; e my and e spy are the yield strain of the bolts and steel plate, respectively; f my and f spy are the yield stress

of the bolts and steel plate, respectively.

2.3 Analytical model and solution algorithm

2.3.1 Determining the fictitious strain e sp,0

The segmental joint subjected to the initial compression-bending loads (defined as N 0 , M 0 ) has already rotated to

a certain degree before being reinforced, causing the outer edge of the segment to be compressed and the inner edge to
be open, as shown in Fig. 3(a). The joint surface satisfies the plane-section assumption, so the strain distribution before

reinforcement can be illustrated in Fig. 3(b), where e c,0 refers to the concrete strain at the outer edge and e m,0 refers to

the bolt strain. For the tension zone, due to the joint surface's discontinuity, the tensile strain of the concrete is
fictitious. This simplified strain distribution of segmental joint surface has been used by previous studies (Li et al.,
2015; J. Zhang et al., 2019; L. Zhang et al., 2020). In this paper, the tensile strain of the concrete at the inner edge is

defined as e sp,0 .
 (a) (b)

Fig. 3. Deformation diagram of the joint surface just before steel plate reinforcement. (a) Bending deformation, and (b)
strain distribution.

The segmental joint experiences three major stages in the entire stress process: partial core concrete compression,
joint compression of the core and the outside concrete, and outside concrete compression (see Fig. 4). In the first stage,
the joint is mainly subjected to axial force and slightly affected by bending moment; in the second stage, the inner edge
of joint undergoes a certain opening due to bending moment, and the stress curve of concrete in the compression zone
is an approximately straight line; and in the third stage, excessive opening of the joint occurs due to high-level bending
moment, which must be avoided in the service of tunnels. From this, we can determine that the second stage is the most
appropriate time to conduct steel plate reinforcement. Figure 5 shows the concrete stress distribution of the joint

surface just before reinforcement, where s c,0 represents the concrete stress at the outer edge, s m,0 represents the bolt

stress, and x0 represents the height of the compressive strain distribution range of the concrete. Due to the large

compression area of the concrete and low bending moment, the concrete constitutive curve (indicated by Eq. (1)) can
be considered to be in the parabolic section just before reinforcement. Based on the discontinuity of the compression

zone and equation å N = 0 , we can find the axial force equilibrium equation with Eq. (4), while s m,0 can be
obtained according to Eq. (5), which is established based on the plane-section assumption. The bending moment

equilibrium equation of the joint surface subjected to a certain axial force can be represented as Eq. (6). Thus, x0 and

e c,0 can be obtained according to Eqs. (4)–(6). Based on the plane-section assumption, the fictitious strain e sp,0 can be
obtained from Eq. (7).

(a) (b) (c)

6
 Fig. 4. Entire stress process of segmental joint with the sagging moment. (a) Partial core concrete compression, (b)
joint compression, and (c) outside concrete compression.

Fig. 5. Compressive stress distribution of joint surface just before steel plate reinforcement.

x0 - d1 - d 2 - d3 æ p ö x0 - d1 æ p ö
N0 = bò sc ç × e c,0 ÷ dp + b ò s c ç × e c,0 ÷ dp - s m,0 Am , (4)
0
è x0 ø x0 - d1 - d 2
è x0 ø

where b is the segment width, p is the integral variable along the compression height, and Am is the cross-sectional

area of the bolt.

e c,0  d - x0  Em
s m,0 = , (5)
x0

where Em is the elastic modulus of the bolt.

x0 - d1 - d 2 - d3 æ p ö x0 - d1 æ p ö
M 0 = bò sc ç × e c,0 ÷  h - x0 + p  dp + b ò s c ç × e c,0 ÷  h - x0 + p  dp -
0
è x0 ø x0 - d1 - d 2
è x0 ø (6)
h
s m,0 Am  h - d  - N 0 × ,
2

e c,0 h
e sp,0 = - e c,0 . (7)
x0
 xcb
2.3.2 Determining the critical height of the concrete compression zone

Failure of reinforced joints, similar to original joints, is deemed to occur at the moment when the outer edge of
concrete reaches the ultimate compressive strain (Li et al., 2015; L. Zhang et al., 2019, 2020). As the main tensile
components of a reinforced joint, the bolts and steel plate may have already yielded before joint failure. Herein, the
height of the concrete compression zone corresponding to the simultaneous joint failure and bolt yielding is defined as

xcb1 , while that corresponding to the simultaneous joint failure and steel plate yielding is defined as xcb2 . Based on the

strain distribution characteristics, the calculation for xcb1 and xcb2 can be expressed as Eqs. (8) and (9), respectively.

e cu d
xcb1 = , (8)
e cu + f my Em

e cu h
xcb2 = , (9)
e cu + e sp,0 + fspy Esp

where Esp represents the elastic modulus of the steel plate.

2.3.3 Determining the ultimate state of the reinforced joint

There are four ultimate states of the reinforced joint based on the stress level of the bolts and steel plate:

(1) S-a, where the bolts have yielded and the steel plate has not;

(2) S-b, where both the bolts and the steel plate have yielded;

(3) S-c, where the steel plate has yielded and bolts have not;

(4) S-d, where neither the bolts nor steel plate has yielded.

The height of the compression zone corresponding to each failure state is successively defined as xc1 , xc2 , xc3 ,

and xc4 , and needs to satisfy the following conditions:

(1) for S-a, xcb2 < xc1 £ xcb1 ;

(2) for S-b, xc2 £ xcb1 and xc2 £ xcb2 ;

(3) for S-c, xcb1 < xc3 £ xcb2 ;

8
 (4) for S-d, xc4 > xcb1 and xc4 > xcb2 .

To establish the equilibriums of the axial force and bending moment corresponding to each failure state, the
simplified distribution of strain and stress in ultimate bearing states is illustrated in Fig. 6(a) and (b). Figure 6(a) shows
the ultimate state where the outside concrete is compressed jointly with the core concrete, and Fig. 6(b) shows the

ultimate state where only the outside concrete is compressed, where xc refers to the height of the concrete compression

zone corresponding to any failure state, and M and N refer to the bending moment and axial force at any failure
state, respectively.

(a)

(b)

Fig. 6. Simplified distribution of strain and stress in several ultimate bearing states. (a) Joint compression of core
and outside concrete, and (b) outside concrete compression.

If the steel plate cross-sectional area is large or the tensile strength is high to a certain degree, the failure state S-a
may occur. The axial force equilibrium equation is written as Eq. (10). The steel plate stress s sp can be obtained by Eq.

(11). Additionally, the bending moment equilibrium equation is written as Eq. (12), where the influence of steel plate
thickness on the bending moment is ignored due to the huge difference in thickness between steel plate and segment.

xc1 can be obtained by substituting N1 into Eqs. (10) and (11). M b1 can be obtained by substituting xc1 and N1 into

Eq. (12).

n
æ p ö
N1 = bå ò
di +1
sc ç × e cu ÷ dp - f my Am - s sp Asp , (10)
i =1
di
è xc1 ø

where Asp is the cross-sectional area of the steel plate.

s sp = e sp Esp =  e cu h xc1 - e cu - e sp,0  Esp , (11)

n
æ p ö h
M b1 = bå ò
di +1
sc ç × e cu ÷  h - xc1 + p  dp - f my Am  h - d  - N1 × . (12)
i =1
di
è xc1 ø 2

For the failure state S-b, the equilibrium equations can be established as Eqs. (13) and (14).

n
æ p ö
N 2 = bå ò
di +1
sc ç × e cu ÷ dp - f my Am - fspy Asp , (13)
i =1
di
è xc2 ø

n
æ p ö h
M b2 = bå ò
di +1
sc ç × e cu ÷  h - xc2 + p  dp - f my Am  h - d  - N 2 × . (14)
i =1
di
è xc2 ø 2

Generally positioned at the inner edge of the joint, the steel plate may suffer obvious deformation when co-bearing
loads with concrete segments. If the steel plate cross-sectional area is small or the tensile strength is weak, the failure
state S-c may occur, that is, the bolts have not yielded and the steel plate has yielded when the concrete reaches
ultimate strain. The equilibrium equations of axial force and bending moment are written as Eqs. (15) and (17),

respectively. Accordingly, the bolt stress s m can be obtained by Eq. (16), which can be established following the
plane-section assumption.

n
æ p ö
N 3 = bå ò
di +1
sc ç × e cu ÷ dp - s m Am - fspy Asp , (15)
i =1
di
è xc3 ø

æe d ö
s m = e m Em = ç cu - e cu ÷ Em , (16)
è xc3 ø
10
 n
æ p ö h
M b3 = bå ò
di +1
sc ç × e cu ÷  h - xc3 + p  dp - s m Am  h - d  - N 3 × . (17)
i =1
di
è xc3 ø 2

Under a high axial force, the extent of the joint opening is small, resulting in low stress on the bolts and steel
plate. This may lead to failure state S-d, where both the bolts and the steel plate do not yield despite the ultimate strain
of concrete. The equilibrium equations are established as Eqs. (18) and (19).

n
æ p ö
N 4 = bå ò
di +1
sc ç × e cu ÷ dp - s m Am - s sp Asp , (18)
i =1
di
è xc4 ø

n
æ p ö h
M b4 = bå ò
di +1
sc ç × e cu ÷  h - xc4 + p  dp - s m Am  h - d  - N 4 × . (19)
i =1
di
è xc4 ø 2

Since the bolts and the steel plate are deemed tension-only in this analytical model, the value(s) of s m Am and

s m Am  h - d  in Eqs. (15)–(19) should be removed when xc > d . Additionally, the value(s) of s sp Asp in Eqs. (10)

e cu h
and (18) should also be removed when xc ³ .
e cu + e sp,0

2.4 Approximate solution

The analytical model requires complex integral calculation, which hinders efficiency for solutions to the ultimate
bending moment. To facilitate the solution process, the concrete's stress distribution can be transformed from nonlinear
distribution to uniform distribution following the equivalent principle (see Fig. 7). Equations (20) and (21) should be
satisfied to ensure that the total axial force value and the force line of the joint surface remain constant before and after
this processing.

Fig. 7. Equivalent distribution of the concrete stress.

xc æ p ö
ò 0
sc ç
è xc
× e cu ÷ dp = a f c b xc ,
ø
(20)
 xc æ p ö bx
× e cu ÷  h - xc + p  dp = a f c b xc æç h - c ö÷ .
ò
0
sc ç
è xc ø è 2 ø
(21)

In Eqs. (20) and (21), a and b are the equivalent stress-correction coefficients, which are 0.91 and 0.85 after
solution (Xiao et al., 2019). Considering the discontinuous distribution of concrete stress due to the non-flat nature of
the joint surface, the integral formulas involved in this analytical model can be approximately represented, as shown in
Table 1.

Table 1 Approximate representation of integral formulas.

Approximate representation
Approximate representation of
of
Height of the concrete n
æ p ö
åò
di +1
compression zone æ p ö × e cu ÷  h - xc + p  dp
n
sc ç
åò
di +1
s c ç × e cu ÷ dp
i =1
di
è xc ø i =1
di
è xc ø

2
b xc b d1 ö
d1 £ xc £ å di a f c b  xc - d1  (22) a f c b  xc - d1  çæ h - - ÷ (23)
i =1 è 2 2 ø

b d2 ö
a f c b d 2 æç h - b d1 -
2 3

å di £ xc £ å di
i =1 i =1
a fc b d2 (24)
è
÷
2 ø
(25)

b d2 ö
a f c b d 2 æç h - b d1 - ÷+
3
è 2 ø
xc > å di a f c b  xc - d1 - d3  (26) (27)
1 æ 3
öæ 3
ö
i =1
a f c b ç xc - å di ÷ ç 2h - b xc - å b di ÷
2 è i =1 øè i =1 ø

Based on the aforementioned theoretical analysis, the calculation process for the ultimate bending moment of a
steel-plate-reinforced joint can be determined (see flowchart in Fig. 8). First, the configuration parameters, material

parameters, and initial loads are input. Second, e sp,0 , xcb1 , and xcb2 are determined according to Eqs. (1)–(9). Third,

the axial force N is substituted into the equilibrium equation of a random failure state, i.e., Eq. (10), (13), (15), or
(18), and a random approximate formula of the integral calculation, i.e., Eq. (22), (24), or (26), is used to obtain the

height of compression zone xc . Fourth, the value xc is checked to see whether it falls into the required range of the

selected formulas. If it does, the selected joint failure state and approximate formulas are correct; otherwise, the third

step should be repeated and another formula of other failure states should be adopted. Fifth, xc is substituted into the

corresponding bending moment equilibrium equation, i.e., Eq. (12), (14), (17), or (19), and approximate formula,

i.e., Eq. (23), (25), or (27), to solve for M b . This process is used to obtain the compression-bending capacity of a

12
 steel plate-reinforced joint, which is crucial for the design and bearing state evaluation of such joints.

Start

Step 1 Input configuration parameters, material parameters,

initial loads, etc.

Step 2 Calculate virtual strain ɛsp,0 and critical height of the

concrete compression zone xcb

Step 3 Substitute N into the axial force equilibrium equation of a random

failure state, and adopt a random approximate formula to calculate the
height of the concrete compression zone xc

Step 4 Does xc satisfy the value

requirements of the selected equilibrium equation and No
approximate formula?

Yes
Step 5 Calculate the ultimate bending moment Mb by
substituting xc into the corresponding bending moment
equilibrium equation

End

Fig. 8. Ultimate bending moment calculation process of a steel-plate-reinforced joint.

3 Model verification

To ensure that the aforementioned calculation model is both accurate and can be applied extensively, ABAQUS, a
commercial finite element (FE) software, was used to analyse the compression-bending capacity.

3.1 FE modelling

The simulation model is detailed in Fig. 9. The tunnel segments and bolts were modelled using C3D8R elements,
steel plate was modelled using C3D8I elements, rebars were modelled using T3D2 elements, and the loading pads used
the rigid body. Considering the computational efficiency and accuracy of this model, the mesh size ranged from 5 to 50
mm, and the established model contained a total of 48 646 elements. Surface-to-surface contacts were assigned to the
interface between segments and between the lining and bolts. The normal behavior of the interface was defined by a
hard contact allowing separation. The tangential behavior was determined by Coulomb friction based on the penalty
function. The friction coefficients between the lining and bolts and between segments were 0.3 and 0.6, respectively
(Tassios & Vintzēleou, 1987; Xia et al., 2021). For boundary conditions, a fixed hinge at one end and a hinge with
horizontal displacement freedom at the other end were used. Kinematic coupling points were set at the left and right
end surfaces to avoid stress concentration.

The cohesive elements were employed to simulate potential interface damage at the interface between the segment
and steel plate, with their mechanical behavior defined by a damage evolution model based on the traction-separation
response (Ding et al., 2023). Figure 10 depicts the technical issues associated with simulating the mechanical behavior

of the bonding interface, where t represents the interfacial stress, t peak is the bond strength, s is the relative

displacement between the steel plate and the lining, is the ultimate relative displacement, K is the interfacial

stiffness, and G C is the interfacial fracture energy. In addition, the variant standard discrete element group (SDEG)

ranging from 0 to 1.0 is used to quantitatively evaluate the interface damage degree. The more serious the interface
damage is, the larger the value of SDEG. The specific parameter values are listed in Table 2 (Sun et al., 2022; Zhao et
al., 2016). In this simulation, the implementation of steel plate reinforcement was delayed until the internal force within
the segment reached a certain threshold. To maintain consistency between the shape of the initially activated steel plate
and the deformed segment, a method was employed using the trace elements with minimal elastic modulus and weight.
This allows for accurate positioning of the steel plate elements.

Fig. 9. Geometric model and meshing effect.

14
 Fig. 10. Technical issues for simulating the mechanical behaivour of the bonding interface.

Table 2 Parameters of the damage evolution model.

Direction t peak (MPa) sult (mm) K (MPa/mm) GC (N/m)

In the normal direction 2.4 0.8 144 960

In the tangential direction 2.5 4.5 40 5625

3.2 Loading method

The horizontal load N s was applied to the two end faces of the curved beams, and the vertical load Ps was

applied to the two loading pads fixed on the outer surface of the segments, as shown in Fig. 11. According to the
classical theory of structural mechanics, the bending moment of the joint surface can be derived as Eq. (28).

M = Ps L - N s H , (28)

where L refers to the horizontal distance between the functional point of Ps and the adjacent end face, and H is the
vertical distance between the functional point of N s and the geometric center of the joint surface.

The secondary loading mode was used in the FE simulation. During the loading process before reinforcement, N s

was constant and Ps was continuously increased. When the joint surface's bending moment reached M 0 , the steel

plate was activated. Thereafter, Ps was increased until joint failure occurred.
 Fig. 11. Schematic diagram of loading method.

3.3 Results comparison

A comparison between FE simulation and theoretical calculation results was performed in terms of the ultimate
bearing capacity and failure states of reinforced joints to verify the accuracy of the analytical model. The parameter

values were listed as follows: h is 350 mm, d is 200 mm, d1 is 5 mm, d 2 is 35 mm, d3 is 74 mm, d 4 is 211 mm, b

is 1500 mm, f c is 25.3 MPa, f spy is 200 MPa, f my is 400 MPa, Esp is 210 GPa, E m is 206 GPa, e sp，0 is 0.0015, Am

is 1413.5 mm2, and Asp is 3000 mm2. The simulation results are shown in Fig. 12.

(a) (b)

Fig. 12. FE simulation results. (a) Structural deflection, and (b) structural stress and interface damage.

As shown in Fig. 12, in the initial stage, the development rate of the reinforced joint deflection was basically
stable. As M exceeded 270 kNm, the deflection development rate increased significantly. During the loading process,
the following four damage behaviors occurred in turn: A. Interface damage occurred (M = 262 kNm), Hereafter, the
joint deflection developed rapidly; B. Steel plate yielded (M = 319 kNm); C. Strain of the outside concrete basically

reached e cu (M = 363 kNm), which is generally taken as a sign of structural ultimate state in the design theory of
16
concrete structure; D. Reinforced joint reached ultimate state (M = 415 kNm), the criterion of which was that the bolts
had yielded. The discrepancy of the ultimate bending moment obtained by FE simulation and theoretical analysis (415
and 334 kNm, respectively) was about 19.5%, which can be attributed primarily to the difference in the structural
ultimate state criterion employed by each method. The criterion utilized in the theoretical method was intentionally
conservative, as it prioritized engineering safety considerations. When the reinforced joint in simulations reached the
ultimate state employed in theoretical analysis, the bending moment was found to be only 8.6% higher than the
prediction made by the analytical model. The prediction accuracy of the analytical model could be further improved by
the introduction of the correction coefficient or the consideration of the interface slip, which are beyond the scope of
this paper.

4 Evolution rules of load-bearing capacity of reinforced joint

From the analytical model proposed in Section 2, it can be known that the load-bearing capacity and failure state
of a reinforced joint vary as the axial force N increases. To visually reflect the joint failure state and reinforcing
benefits under different axial forces, this section presents the load-bearing capacity curves of reinforced and original
joints obtained by the analytical model (see Fig. 13). Herein, the indicator of the original joint failure is consistent with
that of the reinforced joint.

(a) (b)

Fig. 13. Comparison of load-bearing capacity curves of reinforced and original joints. (a) Reinforced joint, and (b)
original joint.

In Fig. 13(a), the load-bearing capacity curve of the reinforced joint contains three failure states: (1) S-b, where
both the bolts and steel plate have yielded; (2) S-c, where the steel plate has yielded but not the bolts; (3) S-d, where
both the bolts and steel plate have not yielded. Six characteristic points can be found on the compression-bending

capacity curve: RAM stands for M b at pure-bending failure; RBF stands for N above which bolt yielding will not

occur; RCF stands for N over which the bolts will not be in tension; RDM stands for the maximum value of M b ;

REF stands for N above which the steel plate will not be in tension; and RFF stands for N at pure-compression
failure. When N is small, the apparent joint opening can lead to the yielding of the bolts and steel plate before joint
failure (S-b). As N increases, the total joint opening degree decreases, and bolt yielding will not occur (S-c). In
particular, after N exceeds RCF, the bolts will no longer be stressed. When M b reaches RDM, the steel plate yielding

and joint failure coincide. With a further increase of N , both bolts and the steel plate cannot yield (S-d), and the
bearing capacity decreases continuously. When N reaches RFF, the reinforced joint experiences a pure-compression
failure, i.e., the ultimate bending moment is zero. In terms of its trend, the curve can be separated into two parts. In the

first part (from RAM to RDM), M b develops gradually as N increases, and the failure mode is a large eccentric

compression failure because the steel plate eventually yields. In the second part (from RDM to RFF), M b decreases as

N increases, and the failure mode is a small eccentric compression failure because joint failure occurs before the steel
plate yields. The demarcation point between these two parts indicates the simultaneous occurrence of steel plate
yielding and joint failure.

In Fig. 13(b), the load-bearing capacity curve of the original joint contains two failure states: (1) S-m, the bolts
have yielded, and (2) S-n, the bolts have not yielded. Four characteristic points can be obtained on the curve: OAM

stands for M b at pure-bending failure; OBF stands for N above which bolt will not yield; OCM stands for the

maximum value of M b ; and ODF stands for N at pure-compression failure. Also, in terms of trend, the curve can be

separated into two parts. The first part is from OAM to OCM, where M b increases with the increase of N , and the

second part is from OCM to ODM, where M b gradually drops. At the demarcation point of these two parts, the bolts

are not stressed at joint failure.

In Fig. 13(a) and (b), the load-bearing capacity curves of the reinforced and original joints both can be split into

two parts in terms of the development trend of M b , and the maximum M b can be obtained at the demarcation point,

which corresponds to different failure states in these two curves. Adopting steel plate reinforcement in part Ⅰ can
appreciably enhance the bearing capacity of the original joint.

5 Parametric analysis

In the aforementioned analytical model, the compression-bending bearing capacity of a steel plate-reinforced joint
is decided by various parameters from the joint surface, bolts, concrete, and steel plate. The existing design methods
are only for original segmental joints and mainly rely on experiences (Y. Zhang et al., 2020). These methods are
insufficient in minimizing resource use and meeting project requirements. Therefore, the analytical method proposed in
this paper is used to clarify the importance of various parameters on the bearing capacity of the reinforced joint and
provide a reference for reinforcement design. In practical engineering, the heights of the joint surface functional zone

( ) must meet load, waterproofing, durability, and other functional requirements, and as such, the analysis

of the heights is not conducted here. Based on the control variable method, this section measures the influence of the

steel plate cross-sectional area ( Asp ), steel plate strength ( f spy ), fictitious strain ( e sp,0 ), concrete strength( f c ),

18
segment width ( b ), bolt strength ( f my ), bolt diameter ( d m ), and bolt location ( d h ), on the ultimate bending moment

of the reinforced joint under different axial force levels. Table 3 displays the calculation cases for each parameter. The
calculation results for eight parameters are shown in Figs. 14–21.

Table 3 Calculation cases for the parameters.

Parameter Value under each case Standard value

Steel plate cross-sectional area Asp 3000, 4000, 5000 mm2 3000 mm2

Steel plate strength fspy 200, 300, 400 MPa 200 MPa

Fictitious strain ɛsp,0 0.0012, 0.0015, 0.0018 0.0015

Concrete strength fc 23.1, 25.3, 27.5 MPa 25.3 MPa

Segment width b 1.2, 1.5, 1.8 m 1.5 m

Bolt strength fmy 320, 400, 480 MPa 400 MPa

Bolt diameter dm 30, 35, 40 mm 30 mm

Bolt location d/h 0.52, 0.57, 0.62 0.57

Figures 14 and 15 show the load-bearing capacity of reinforced joints that have different steel plate cross-sectional

areas and strengths, respectively. Herein, the eccentricity corresponding to the point RDM is defined as eb . When N

is so small that the eccentricity exceeds eb , M b is strongly affected by changes in Asp and f spy , that is, a significant

rise of M b can be seen along with the increase of Asp and f spy under a certain axial force. When eccentricity is under

eb , the increase in Asp and fspy has little impact on M b , because the steel plate cannot fully exert tensile properties.

The main difference between these two parameters is that Asp has a larger influence than f spy on the curves.

Considering the uncertainty of N during the service life of tunnels, it is better to increase the steel plate cross-
sectional area rather than enhance the steel plate strength to improve the load-bearing capacity of the reinforced joint.

Figure 16 shows the load-bearing capacity curves of reinforced joints with different fictitious strains. It can be

found that the influence of e sp,0 on the curve is limited to the range between the characteristic points RDM and REF.

This is in line with the fact that only the equilibrium equations of failure state S-d contain e sp,0 in the analytical model.
Within the affected range, M b under a certain axial force decreases as e sp,0 increases. This indicates that increasing

e sp,0 is not conducive to improving the bearing capacity of the reinforced joint.

Figures 17 and 18 show the load-bearing capacity curves of reinforced joints with different concrete strengths and

segment widths, respectively. It can be seen that f c and b have an integral effect on the curve, which is indicated by

the expansion of the entire curve along with the increase in the two parameters. Within the range where the eccentricity

is greater than eb , the variation amplitude of M b caused by the changes in f c and b develops with the increase of

N , while outside the range, the variation amplitude is stable. Unlike the results of changing Asp , fspy , and e sp,0 , the

bearing capacity of the reinforced joint under a higher N can be greatly enhanced by increasing f c and b .

Considering that f c and b cannot be adjusted after the tunnel is put into service, the level and distribution of the loads

to be borne by the lining structure should be determined before construction to ensure that the values of these two
parameters are reasonable.

Figure 19 represents the load-bearing capacity curves of reinforced joints with different bolt strengths. The bolt

strength, f my , impacts the curve only in the low range of N . Within the impacted range, the M b under a certain N

increases in line with increasing f my , but the variation amplitude is small. From this, it can be concluded that

optimizing the bolt strength has minimal improvement on the bearing capacity of the reinforced joint.

Figures 20 and 21 show the load-bearing capacity curves of reinforced joints with different bolt diameters and bolt

locations. The impact of d m and d h on the curve is similar both in range and amplitude. When the eccentricity

exceeds eb , the M b under a certain N increases as d m and d h go up. When the eccentricity is less than eb , d m ,

and d h have almost no impact on M b .

The above analyses indicate that different parameters exert different influences on the bearing capacity curve of
the reinforced joint. Specifically, (1) the steel plate strength and cross-sectional area significantly influence the bearing

capacity of the reinforced joint, which is reflected in the large variation amplitude of M b with changes in these two

parameters, but considering that the affected range of the cross-sectional area of the steel plate is larger than that of the
steel plate strength, it is better to improve the bearing capacity of the reinforced joint by optimizing cross-section areas
of the steel plate; (2) the fictitious strain has minimal influence on the curve both in range and amplitude; (3) the
concrete strength and segment width are highly influential on the bearing capacity curve, especially when N is large;
and (4) the impact of bolt diameter and the location on the bearing capacity of the reinforced joint are limited in
amplitude, and the bolt strength has a negligible effect on the bearing capacity curve.
20
Fig. 14. Load-bearing capacity curves of reinforced Fig. 15. Load-bearing capacity curves of reinforced
joints with different Asp. joints with different fspy.

Fig. 16. Load-bearing capacity curves of reinforced Fig. 17. Load-bearing capacity curves of reinforced
joints with different ɛsp,0. joints with different fc.

Fig. 18. Load-bearing capacity for reinforced joints Fig. 19. Load-bearing capacity for reinforced joints
 with different b. with different fmy.

Fig. 20. Load-bearing capacity curves of reinforced Fig. 21. Load-bearing capacity for reinforced joints
joints with different dm. with different d/h.

6 Sensitivities of ultimate bending moment of reinforced joints

The analyses in previous sections indicate that the steel plate cross-sectional area ( Asp ), steel plate strength

( f spy ), concrete strength ( f c ), and segment width ( b ) have the largest impact on the bearing capacity of the

reinforced joint, but the impacts under different axial forces are not constant. In this section, the sensitivities of these

four parameters on M b are quantitatively analyzed. According to the development trends of M b with increasing N

described in Figs. 13–20, four representative axial force levels (0.15RFF, 0.30RFF, 0.45RFF, and 0.60RFF) are used to
reflect the impact of axial force level on the sensitivities of various parameters. To ensure the comparability of the
calculation results of each condition, the variation amplitude of parameters in each condition is 25.0% of the standard
value (listed in Table 3). The calculation conditions are shown in Table 4. Then, Eq. (29) can be adopted to normalize
the ultimate bending moment values in all conditions (Liu et al., 2018b):

xi - min  xi 
Z= , (29)
max  xi  - min  xi 

where Z is the normalized value of M b , being between 0 and 1; xi is the value of M b ; max  xi  and min  xi  are

the maximum and minimum values of xi , taken as 502 and 0 kN × m , respectively.

Table 4 Calculation conditions.

22
 Axial force level Parameter

Asp = 1500, 2250, 3000, 3750, 4500 mm2

fspy = 100, 150, 200, 250, 300 MPa

N = 0.15RFF, 0.30RFF, 0.45RFF, and 0.60RFF
fc = 12.6, 19.0, 25.3, 31.6, 38.0 MPa

b = 0.750, 1.125, 1.500, 1.875, 2.250 m

Figure 22 shows the normalized results of ultimate bending moment values of reinforced joints under different
axial forces. When N reaches 0.15RFF, the slopes of the Z-value curves corresponding to these four parameters are
close, and the total increase of Z-value caused by the changes in these parameters is 0.25, 0.25, 0.35, and 0.35,
respectively, indicating that the sensitivities of these parameters are close under low axial forces. With 0.30RFF, the
slopes of the Z-value curves corresponding to the concrete strength and segment width are much larger than those
corresponding to the steel plate strength and cross-sectional area, and the total increase of Z-value caused by the
changes in these parameters is 0.23, 0.18, 0.65, and 0.65, respectively. This indicates that the sensitivities of the
concrete strength and segment width are appreciably higher than those of the other two parameters. With N increasing
to 0.45RFF, the slopes of the Z-value curves corresponding to the steel plate cross-sectional area and steel plate
strength decrease, while those corresponding to the concrete strength and segment width increase, and the total increase
of Z-value caused by the changes in these parameters is 0.09, 0.03, 0.93, and 0.93, respectively.

The above analysis indicates that, when N does not exceed 0.15RFF, the sensitivities of these four parameters

with regard to M b are similar. When N is between 0.30RFF and 0.45RFF, the sensitivities of the concrete strength

and segment width with regard to M b are significantly higher than the other two parameters, and the difference

increases in line with N . When N exceeds 0.45RFF, the concrete strength and segment width have prominent effects

on M b , while the other two parameters have almost no impact.

(a) N = 0.15RFF

(b) N = 0.30RFF

24
 (c) N = 0.45RFF

(d) N = 0.60RFF

Fig. 22. Normalized results of ultimate bending moment values of reinforced joints under different axial forces.

7 Conclusions

This paper proposed an analytical model that serves as a reference for the design of steel plate reinforcement in
segmental joints under sagging moment, helping determine the load-bearing capacity of a steel-plate-reinforced joint
with a safety margin. After verifying the validity of this model, the evolution rule of load-bearing capacity of the
reinforced joint was attained, and corresponding parametric analysis was conducted. The following conclusions can be
drawn:

(1) Based on the deformation and stress characteristics of the steel-plate-reinforced joint, an analytical model was
proposed to calculate the compression-bending capacity of the joint. The proposed analytical model can provide a
valuable reference for the design of reinforcement in shield tunnels and offer significant advantages over the FE
simulation in efficiency.
 (2) The load-bearing capacity curve of the reinforced joint under different axial forces is nonlinear, featured with
six special points that are determined based on the external force characteristics of the joint surface and the stress level

of the bolts and steel plate. The curve comprises two parts: in part Ⅰ, M b goes up with the increase of axial force ( N ),

and the failure mode is a large eccentric compression failure; in part Ⅱ, M b goes down with the increase of N , and

the failure mode is a small eccentric compression failure. The demarcation point between these parts represents the

simultaneous steel plate yielding and joint failure, from which the maximum M b can be attained. Comparing the

bearing capacity curves of reinforced and original joints reveals an obvious reinforcement effect in part I, on which the
N at each point is relatively low.

(3) When N is no more than 0.15RFF, the influence of the steel plate strength and cross-sectional area on the
load-bearing capacity curve of the reinforced joint is significant, with the affected range of the former is larger than that
of the latter. The fictitious strain has minimal impact on the curve both in range and amplitude. The influence of
concrete strength and segment width is prominent after N exceeds 0.30RFF. The influence of the bolt diameter and
location is limited in amplitude, while the bolt strength has a negligible impact on the curve.

CRediT authorship contribution statement

Zhen Li: Conceptualization, Methodology, Investigation, Formal analysis, Writing - original draft. Xuezeng Liu:
Conceptualization, Methodology, Funding acquisition, Project administration, Writing - review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have
appeared to influence the work reported in this paper.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Acknowledgement

The authors acknowledge the support provided by National Natural Science Foundation of China (Grant No.
52278409).

References

De Lorenzis, L., Teng, J. G., & Zhang, L. (2006). Interfacial stresses in curved members bonded with a thin plate.
International Journal of Solids and Structures, 43(25–26), 7501–7517.

Ding, W. Q., Ma, C., Guo, Y. J., Li, X. R., & Li, S. B. (2023). Numerical analysis of new stainless-steel corrugated-
plate reinforcement of shield-tunnel segmental joints based on virtual-tracking-element technology. Applied
Sciences, 13(10), 5904.

26
Elmalich, D., & Rabinovitch, O. (2009). Stress analysis of monolithic circular arches strengthened with composite
materials. Journal of Composites for Construction, 13(5), 431–441.

EL Naggar, H., Hinchberger, S. D., & Lo, K. Y. (2008). A closed-form solution for composite tunnel linings in a
homogeneous infinite isotropic elastic medium. Canadian Geotechnical Journal, 45(2), 266–287.

Guo, W. Q., Feng, K., Zhou, Y. L., Yang, W. Q., Lu, X. Y., Xiao, M. Q., & He, C. (2023). Full-scale test and
numerical modeling on deformation and damage behavior of segmental joints under ultimate compression-
bending load. Engineering Structures, 279, 115648.

Huang, H. W., Shao, H., Zhang, D. M., & Wang, F. (2017). Deformational responses of operated shield tunnel to
extreme surcharge: a case study. Structure and Infrastructure Engineering, 13(3), 345–360.

Li, X. Q., Zhang, T., Ding, Z. D., Yang, X., & Wen, J. C. (2021). Numerical analysis of normal concrete lining
strengthening methods under different damage levels. Structure and Infrastructure Engineering, 17(12), 1597–
1611.

Liu, D. J., Wang, F., Zhang, D. M., & Duan, K. (2019). Interfacial stresses of shield tunnel strengthened by a thin plate
at inner surface. Tunnelling and Underground Space Technology, 91, 103021.

Liu, D. J., Zhang, Y., Zuo, J. P., & Dai, Q. Q. (2023). Strengthening evaluation and evolution law of the bearing
capacity of the normal section of the shield tunnel based on M-N curve. Journal of China University of Mining &
Technology, 52(1), 76–85 (in Chinese).

Liu, T. J., Huang, H. H., Xu, R., & Yang, X. P. (2017). Research on load-bearing capacity of metro shield tunnel lining
strengthened by bonded steel plates. China Journal of Highway and Transport, 30(8), 91–99 (in Chinese).

Liu, X., Jiang, Z. J., Yuan, Y., & Mang, H. A. (2018a). Experimental investigation of the ultimate bearing capacity of
deformed segmental tunnel linings strengthened by epoxy-bonded steel plates. Structure and Infrastructure
Engineering, 14(6), 685–700.

Liu, X., Jiang, Z. J., Yuan, Y., & Mang, H. A. (2018b). Numerical investigation of the mechanical behavior of
deformed segmental tunnel linings, strengthened by epoxy-bonded filament wound profiles. Tunnelling and
Underground Space Technology, 78, 231–244.

Liu, X., Zhang, C. G., Zhang, C., & Jiang, Z. J. (2016). Experimental study on the longitudinal joint in shield tunnel
reinforced with FRP material. Journal of Railway Science and Engineering, 13(2), 316–324 (in Chinese).

Liu, X., Zhang, C., Zhang C. G., & Yuan, Y. (2017). Ultimate load-carrying capacity of the longitudinal joints in
segmental tunnel linings. Structural Concrete, 18(5), 693–709.

Li, X. J., Yan, Z. G., Wang, Z., & Zhu, H. H. (2015). A progressive model to simulate the full mechanical behavior of
concrete segmental lining longitudinal joints. Engineering Structures, 93, 91–113.

Liu, X. Z., Sang, Y. L., Ding, S., You, G. L., Zhu, W. X., Zhou, R. Y., Wei, Q., & Jiang, L. (2020). Experimental
Study on the Mechanics Characteristics of CFRP Strengthening of Highway Tunnels at Different Damage States.
Geofluids, 2020, 6665996.
She, C. G., Zhang, B. L., Zheng, J., & Li, X. X. (2020). Theoretical analysis of bearing capacity for shield tunnel
reinforcement structure. In Proceedings of the 6th International Conference on Environmental Science and Civil
Engineering (pp. 012171).

Shi, P. X., & Li, P. (2015). Mechanism of soft ground tunnel defect generation and functional degradation. Tunnelling
and Underground Space Technology, 50, 334–344.

Sun, Y. Z., Yu, Y., Wang J. C., Ye, Y. L., & Tan, Q. Y. (2022). Mechanical properties of linings of shield tunnel
strengthened by steel plates considering interface effects. Chinese Journal of Geotechnical Engineering, 44(2),
343–351 (in Chinese).

Tassios, T. P., & Vintzēleou, E. N. (1987). Concrete-to-Concrete Friction. Journal of Structural Engineering, 113(4),
832–849.

Xia, J., Shan, K. Y., Wu, X. H., Gan, R. L., & Jin, W. L. (2021). Shear-friction behavior of concrete-to-concrete
interface under direct shear load. Engineering Structures, 238, 112211.

Xiao, M. Q., Feng, K., Zhang, L., & He, C. (2019). A calculation model of flexural bearing capacity of segmental joint
for shield tunnels. China Civil Engineering Journal, 52(11), 108–119 (in Chinese).

Zhai, W. Z., Chapman, D., Zhang, D. M., & Huang, H. W. (2020). Experimental study on the effectiveness of
strengthening over-deformed segmental tunnel lining by steel plates. Tunnelling and Underground Space
Technology, 104, 103530.

Zhai, W. Z., Sun, K. G., Zhang, D. M., Huang, H. W., & Chapman, D. (2022). An analytical solution for steel plate
strengthened circular tunnels with various interface slip modes. Buildings, 12(8), 1172.

Zhang, D. M., Zhai, W. Z., Huang, H. W., & Chapman, D. (2019). Robust retrofitting design for rehabilitation of
segmental tunnel linings: Using the example of steel plates. Tunnelling and Underground Space Technology, 83,
231–242.

Zhang, J. G., & Li, W. (2020). Research on key issues of improved strip algorithm for segmental joints. Journal of the
China Railway Society, 42(3), 147–152 (in Chinese).

Zhang, J. L., Liu, X., Ren, T. Y., Yuan, Y., & Mang, H. A. (2019). Structural behavior of reinforced concrete segments
of tunnel linings strengthened by a steel-concrete composite. Composites Part B: Engineering, 178, 107444.

Zhang, L., Feng, K., Li, M. R., Sun, Y. P., He, C., & Xiao, M. Q. (2019). Analytical method regarding compression-
bending capacity of segmental joints: Theoretical model and verification. Tunnelling and Underground Space
Technology, 93, 103083.

Zhang, L., Feng, K., Li, M. R., He, C., Xiao, M. Q., & Zhang, Z. (2020). Characteristic analysis on the compression-
bending capacity of segmental joint based on the investigated joint parameters. Tunnelling and Underground
Space Technology, 102, 103444.

Zhang, Y., Zhu, P., & Shi, J. Q. (2020). Flexural behavior of precast UHPC beam with prestressed bolted hybrid joint.
Engineering Structures, 206, 110100.

28
Zhao, H. L., Liu, X., Bao, Y. H., Yuan, Y., & Bai, Y. (2016). Simplified nonlinear simulation of shield tunnel lining
reinforced by epoxy bonded steel plates. Tunnelling and Underground Space Technology, 51, 362–371.

Zhu, H. H., Huang, B. Q., Li, X. J., & Qiao, B. Z. (2014). Unified model for internal force and deformation of shield
segment joints and experimental analysis. Chinese Journal of Geotechnical Engineering, 36(12), 2153–2160 (in
Chinese).