Transportation Geotechnics 44 (2024) 101171

Contents lists available at ScienceDirect

Transportation Geotechnics
journal homepage: www.elsevier.com/locate/trgeo

A mechanical analysis model for tunnels under strike-slip faulting

considering the fault zone width and nonlinear tunnel-stratum interaction
Xiao Zhang a, b, c, Li Yu a, b, Mingnian Wang a, b, *, Henghong Yang a, b
a
School of Civil Engineering, Southwest Jiaotong University, Chengdu 610036, China
b
Key Laboratory of Traffic Tunnel Engineering, Ministry of Education, Southwest Jiaotong University, Chengdu 610036, China
c
School of Transportation and Logistics, Southwest Jiaotong University, Chengdu 610036, China

A R T I C L E I N F O A B S T R A C T

Keywords: Tunnels crossing active fault zones are seriously damaged under the action of faulting. However, the influence of
Tunnel engineering the fault zone width or nonlinear tunnel-stratum interaction is neglected in existing analytical models, resulting
Strike-slip fault in incorrect estimation of the tunnel mechanical response. To aid in the seismic design of tunnels, a mechanical
Fault zone width
analysis model for tunnels under strike-slip faulting considering the effect of fault zone width and nonlinear
Nonlinear tunnel-stratum interaction
Mechanical response
tunnel-stratum interaction is first proposed in this work. The calculation results of the proposed model agree well
with those obtained from the model test and numerical simulation, and the accuracy is obviously higher than
that of the existing analytical methods. Compared with the FEM numerical model, the average errors of the peak
bending moment, shear force and axial force calculated by the proposed model are 6.37 %, 11.00 % and 9.91 %,
respectively. Neglecting the effect of fault zone width or nonlinear tunnel-stratum interaction, the tunnel
response will be misestimated, e.g., when the fault displacement is 3 m and the crossing angle β is 90◦ , the
overestimated value is approximately 40–110 %. Finally, the results of parameter analysis show that under the
condition of 0◦ < β < 90◦ , the tunnel failure scope caused by the left-lateral strike-slip faulting is obviously larger
than that caused by the right-lateral strike-slip faulting, while the conclusion is opposite under 90◦ < β < 180◦ ;
the maximum strain within the tunnel correlates directly with fault displacement and the strength of the fault
zone, exhibiting an inverse correlation with the width of the fault zone; meanwhile, the scope for tunnel failure is
directly influenced by fault displacement, fault zone width, and the crossing angle (where 90◦ < β < 180◦ ), while
inversely affected by the crossing angle (where 0◦ < β < 90◦ ) and the strength of the fault zone.

l Element length
t Thickness N Axial force
D Equivalent outer diameter M Bending moment
C Buried depth W Fault zone width
L Length WF Fault zone width in the footwall
n Number of elements WH Fault zone width in the hanging wall
LF Length in the footwall Δf Strike-slip displacement
LH Length in the hanging wall Δfx Axial displacement
β Crossing angle Δfy Transverse displacement
V Shear force δx Axial relative tunnel-stratum displacement
ux Axial displacement δy Transverse relative tunnel-stratum displacement
uy Transverse displacement K0 Lateral pressure coefficient
A Cross-section area Δx Axial yield displacement
E Young’s modulus Δy Transverse yield displacement
I Inertia moment Pxu
χ
Axial bearing capacity

* Corresponding author at: School of Civil Engineering, Southwest Jiaotong University, Chengdu 610036, China.
E-mail addresses: [email protected] (X. Zhang), [email protected] (L. Yu), [email protected] (M. Wang).

https://doi.org/10.1016/j.trgeo.2023.101171
Received 18 July 2023; Received in revised form 9 November 2023; Accepted 4 December 2023
Available online 10 December 2023
2214-3912/© 2023 Elsevier Ltd. All rights reserved.
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

Pχyu Transverse bearing capacity and cracking may occur. Additionally, Zhong et al. [21] pointed out that
kχy Transverse stiffness the increase in stratum stiffness in the fault zone reduces the extension
Eχ Young’s modulus but increases the severity of structural damage to the tunnel lining,
Px Axial stratum stress which indicates that the fault zone width should be considered in tunnel
Py Transverse stratum stress response analysis. This point has also been widely recognized by many
f1 Coating dependent factor authors [17,22].
γχ Unit weight However, the numerical simulation of a tunnel crossing an active
cχ Cohesion fault zone takes several hours even in a high-performance computer,
φχ Friction angle which is not applicable to engineering design. Therefore, a series of
kxx Axial stiffness analytical models have been proposed for the mechanical response
νχ Poisson’s ratio analysis of a tunnel or pipeline under faulting. Newmark and Hall [23]
Nch , Nqh
χ χ
Bearing capacity factor first proposed a cable-like model for a pipeline under a strike-slip fault
without considering the contribution of lateral stratum resistance.
Note: Considering the stratum parameters of the fault zone and host
Kennedy [24] extended this model by considering stratum resistance,
rock, the parameter variable χ is introduced. The parameters of the fault
but the pipeline’s flexural stiffness is still not considered. Later, Wang
zone for χ = f and host rock for χ = h.
and Yeh [25] partitioned the pipeline into four segments to compensate
for the limitation of the cable-like models in neglecting the contribution
Introduction
of the pipeline’s flexural stiffness. In this model, the two high-curvature
segments proximal to the fault are treated as circular arcs, while the two
Active fault zones are widely distributed around the world, espe­
segments at the far ends are treated as elastic beams on an elastic
cially in China, the USA, Turkey and Japan. Generally, a fault zone
foundation. After that, this model was improved and developed by many
consists of a fault core and a damage zone, and its width varies from tens
authors [26–31], and some shortcomings have been overcome. How­
to hundreds of meters [1,2], as shown in Fig. 1.
ever, in these analytical models, the fault zone is simplified to a thin line,
With the development of infrastructure construction, it is inevitable
which is applicable to shallow buried pipelines (buried a few meters in
that some tunnels will cross active fault zones, especially in seismically
the soft stratum), as shown in Fig. 3(a). The highway, railway and hy­
active regions. However, the displacement of the active fault zone would
draulic tunnels crossing active fault zones are usually buried tens to
result in severe damage to tunnels, i.e., the 1906 San Francisco earth­
hundreds of meters, and the fault zone width has a significant effect on
quake caused severe damage to two tunnels crossing the San Andreas
the tunnel response [21], as shown in Fig. 3(b). Therefore, these models
fault [3]; the 1999 Chi-Chi earthquake in Taiwan caused significant
are not applicable to tunnels crossing active fault zones due to the
damage to many nearby tunnels due to the displacement of the Che­
neglect of the influence of fault zone width.
longpu fault [4]; the 2008 Wenchuan earthquake in China resulted in
Additionally, to assess the tunnel response under the action of fault
severe damage to several tunnels along the Duwen expressway [5]; and
zone displacement, a series of analytical models have been developed
the 2016 Kumamoto earthquake in Japan severely damaged the Toyama
[32–37], and the effect of fault zone width is considered based on the
tunnel near the fault [6]. Most notably, during the 2022 Menyuan
elastic foundation beam theory. In these models, a tunnel crossing an
earthquake in China, the Daliang tunnel suffered severe damage due to a
active fault zone is simplified as an elastic beam acting on an elastic
3 m strike-slip fault displacement [7,8], as plotted in Fig. 2. Therefore,
foundation, and the tunnel-stratum interaction is assumed to be linear.
the mechanical response of tunnels under faulting has been widely
However, the relevant model tests [38] and numerical simulations [39]
investigated.
showed that the tunnel-stratum interaction stress increases nonlinearly
A series of model tests [9–14] and numerical simulations [15–20]
with increasing fault displacement, indicating that the tunnel-stratum
have been conducted to investigate the tunnel response under faulting.
interaction has obvious nonlinearity. Therefore, this assumption is un­
The results show that the tunnel is subjected to a combination of tensile
realistic and can lead to incorrect estimation of the internal force of the
(compressive), bending and shearing and experiences an elongated ‘S’-
tunnel under faulting.
shaped deformation along the tunnel axis; the tunnel internal forces
To this end, a mechanical analysis model for tunnels under strike-slip
increase with increasing fault displacement, stratum stiffness and tunnel
faulting considering the fault zone width and nonlinear tunnel-stratum
stiffness; under the action of fault zone displacement, the tunnel will be
interaction is proposed in this work, and the above weaknesses have
severely damaged, especially near the fault zone, where tunnel collapse
been fixed.

Establishment of the proposed model

Outline

The proposed model is plotted in Fig. 4, where the strike-slip

displacement is Δf, the crossing angle is β (0◦ < β < 180◦ ), the tunnel
equivalent diameter is D (i.e., the diameter of a circular tunnel with the
same area as the prototype tunnel), the lining thickness is t, the burial
depth is C, the fault zone width is W (WF in the footwall and WH in the
hanging wall) and the tunnel length is L (LF in the footwall and LH in the
hanging wall). Note that the fault core is simplified to a fault plane, as
the width of the fault core is much smaller than that of the damage zone.
This simplification has been widely employed in previous works [17].
The left- and right-lateral strike-slip faults and crossing angles β are
considered in this work, as shown in Fig. 4(a) and (b).
As plotted in Fig. 4, the axial and transverse fault displacements Δfx
and Δfy acting on the tunnel can be calculated by Eq. (1) and Eq. (2). The
Fig. 1. Conceptual model of a fault zone illustrating the main architectural value of Δf is positive (+) for left-lateral strike-slip faults and negative (-)
elements: damage zone and fault core. for right-lateral strike-slip faults.

2
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

Fig. 2. Daliang tunnel damaged due to strike-slip faulting in the 2022 Menyuan earthquake [7,8]: (a) circumferential crack; (b) collapsed lining; (c) bulging invert;
(d) spalling lining.

Fig. 3. Schematic diagram of (a) a shallow pipeline and (b) a deep tunnel crossing an active fault zone.

Δfx = Δf cosβ (1) where N is the axial force caused by the combined action of axial and
transverse displacements, which will be discussed later.
Δfy = Δf sinβ (2) Based on Eq. (3), the following equation of moment equilibrium is
obtained:
Derivation of the governing equation
dM duy
Based on previous studies [31,40], a tunnel crossing an active fault − =N +V (4)
dx dx
zone is simplified as a large-deformed beam acting on a nonlinear
foundation, as shown in Fig. 5. The tunnel-stratum interaction is where uy is the transverse displacement of the tunnel.
described by a series of nonlinear springs, and the fault displacements The bending moment of the tunnel is [41]:
are applied at the ends of the springs. The different mechanical char­ d2 uy
acteristics of the fault zone and host rock are reflected by different M = − EI (5)
dx2
stratum stiffnesses. Considering the different stratum parameters of the
fault zone and host rock, the parameter variable χ is introduced. The where I and E are the inertia moment and Young’s modulus of the
parameters of the fault zone for χ = f and host rock for χ = h. For tunnel, respectively.
f Based on Eq. (4) and Eq. (5), yields:
instance, the nonlinear transverse stiffness of the fault zone is ky , and the
nonlinear stiffness of the host rock is khy . Under the action of fault dis­ d3 uy duy
EI =N +V (6)
placements (Δfx and Δfy) and stratum stresses (Px and Py), the tunnel is dx3 dx
subjected to axial force N, shear force V and bending moment M, as By taking the derivative from both sides of Eq. (6), yields:
plotted in Fig. 5.
According to Fig. 5, the microelement equilibrium of moment yields: d4 uy dN duy d2 uy dV
EI = +N 2 + (7)
dx4 dx dx dx dx
Py
dM + M + Vdx + Ntanθdx = M + (dx)2 (3) From the microelement equilibrium of forces in the y-direction,
2

3
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

Fig. 4. (a) A tunnel crossing an active fault zone, (b) left-lateral strike-slip fault with 0◦ < β < 90◦ , (c) left-lateral strike-slip fault with 90◦ < β < 180◦ , (d) right-
lateral strike-slip fault with 0◦ < β < 90◦ and (e) right-lateral strike-slip fault with 90◦ < β < 180◦ .

yields:
δy = Δfy − uy (10)
dV
= Py (8) Based on Eq. (4) ~ Eq. (10) yields the governing equation of trans­
dx
verse tunnel displacement uy:
The nonlinear transverse stratum stress Py is proportional to the
value of the relative transverse tunnel-stratum displacements δy and EI d4 uy N d2 uy 1 dN duy
− − + uy = Δfy (11)
nonlinear stratum stiffnesses kχy : kyχ dx4 kxχ dx2 kxχ dx dx

Py = kyχ δy (9) Derivation of the axial force equation

The relative transverse tunnel-stratum displacements δy can be Based on previous studies [31,40], the axial force N of a large
calculated by deformed tunnel consists of a frictional axial force Na (caused by fric­
tional axial tunnel-stratum interaction) and a membrane axial force Nb

4
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

Fig. 5. Schematic diagram of a tunnel in bending, tension and shearing.

(caused by large transverse deflection). Neglecting the term Nb leads to The membrane axial strain εb of the tunnel owing to this large
erroneous results of tunnel response under faulting [27,31,40]. transverse deflection is derived as [31,40]:
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
N = Na + Nb = EA(εa + εb ) (12) du
εb = 1 + ( y )2 − 1 (18)
where A is the lining cross-sectional area; εa and εb are the axial dx
strains caused by frictional axial tunnel-stratum interaction and trans­ Using the following series expansion yields:
verse deflection, respectively. √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
du 1 du 1 du
εb = 1 + ( y )2 = 1 + ( y )2 − ( y )4 … (19)
(1) The frictional axial force Na dx 2 dx 8 dx
The higher-order terms in Eq. (19) are deleted under the premise of
In the axial direction, the tunnel is subjected to the axial fault ensuring the calculation accuracy. Therefore, the membrane axial strain
displacement Δfx and frictional stress Px, resulting in axial tunnel εb and axial force Nb can be calculated by
displacement ux and axial force Na, as shown in Fig. 6.
The axial force Na of the tunnel is [41]: 1 duy 2
εb = ( ) (20)
2 dx
dux
Na = EAεa = EA (13) ( )2
dx EA duy
Nb = EAεb = (21)
According to Eq. (13), the axial tunnel displacement ux must be ob­ 2 dx
tained to calculate the axial force Na. Therefore, the governing equation Note that the key problem in analysing the tunnel response under
of the axial tunnel displacement ux is derived herein. From the free body faulting is to solve the governing equations of the axial and transverse
equilibrium of forces in the x-direction in Fig. 6, yields tunnel displacements ux and uy, i.e., Eq. (11) and Eq. (17).
Na = Na + Px dx + dNa (14)
The nonlinear axial stratum stress Px is a function of the relative axial Solution of the governing equations
tunnel-stratum displacement δx and nonlinear stratum stiffnesses kχx :
Due to the nonlinear stratum stiffnesses kχx and kχy , it is difficult to
Px = kxχ δx (15)
obtain the analytical solution of tunnel displacements ux and uy.
The relative axial tunnel-stratum displacements δx can be calculated Therefore, the finite difference method is used to solve the governing
by equations Eq. (11) and Eq. (17). As shown in Fig. 7, a tunnel with a
length of L is discretized along the axial direction into n + 5 node ele­
δx = Δfx − ux (16)
ments (including four virtual node elements), and the element length is l
Based on Eq. (14) ~ Eq. (16), the differential equation of axial tunnel = L/n.
displacement ux is:

EA d2 ux
− + ux = Δfx (17)
kxχ dx2

(2) Membrane axial force Nb

Fig. 7. Discretization of the tunnel.

Fig. 6. The frictional axial tunnel-stratum interaction.

5
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

Solution of the axial tunnel displacement ux

uy,− = 2uy,0 − uy,1 (36)
According to the finite difference method, the difference expression 1

of Eq. (17) is:

uy,− 2 = 4uy,0 − 4uy,1 + uy,2 (37)
EA ( )
− χ 2 ux,i+1 − 2ux,i + ux,i− 1 + ux,i = Δfx,i (22)
kx,i l uy,n+1 = 2uy,n − uy,n− 1 (38)

where kχx,i , ux,i and Δfx,i are the axial stratum stiffness, tunnel uy,n+2 = 4uy,n − 4uy,n− 1 + uy,n− 2 (39)
displacement and fault displacement of node i, respectively.
According to Eq. (31) ~ Eq. (39), the transverse tunnel displacement
The frictional axial force Na at both ends of the tunnel is zero, i.e.,
uy can be calculated by
Na,0 = Na,n = 0. According to the finite difference method and Eq. (13),
[ ] [ ]
the difference expressions of the boundary conditions are: [B] uy = Δfy (40)
EA ( )
where
Na,0 = ux,1 − ux,− 1 = 0 (23)
2l
[B](n+1)×(n+1) = [B1 ](n+1)×(n+1) − [B2 ](n+1)×(n+1) − [B3 ](n+1)×(n+1) + [B4 ](n+1)×(n+1)
EA ( )
Na,n = ux,n+1 − ux,n− 1 = 0 (24) (41)
2l
[ ]
Therefore, the axial tunnel displacement ux of nodes − 1 and n + 1 uy = [ uy,0 uy,1 uy,2 ... uy,i ... uy,n ]Tn+1 (42)
can be obtained as follows:
[ ] [ ]T
ux,− = ux,1 (25) Δfy = Δfy,0 Δfy,1 Δfy,2 ... Δfy,i ... Δfy,n n+1
(43)
1

The calculation of matrix [B] is detailed in Appendix B.

ux,n+1 = ux,n− 1 (26)

According to Eq. (22) ~ Eq. (26), the axial tunnel displacement ux Nonlinear tunnel-stratum interaction
can be calculated by To solve the equations of tunnel displacements, Eq. (27) and Eq. (40),
the key point is to calculate the nonlinear stratum stiffnesses kx and ky .
χ χ
[A][ux ] = [Δfx ] (27)
The widely recognized nonlinear structure-stratum interaction curves
where recommended by design guideline [42] are employed in this work as an
example, see Fig. 8. In addition, to assess the tunnel response under both
[A](n+1)×(n+1) = [A2 ](n+1)×(n+1) − [A1 ](n+1)×(n+1) (28)
left- and right- lateral strike-slip faulting (see Fig. 4) with uniform
governing equations, it is assumed that these springs can withstand
[ux ] = [ ux,0 ux,1 ux,2 ... ux,i ... ux,n ]Tn+1 (29) compression and tension. Otherwise, it is necessary to establish different
[ ]T calculation models for different situations. The force vector applied to
[Δfx ] = Δfx,0 Δfx,1 Δfx,2 ... Δfx,i ... Δfx,n n+1
(30) the tunnel by these springs is shown in Fig. 8.
The calculation of matrix [A] is detailed in Appendix A. According to Fig. 8, the axial and transverse stratum stiffnesses kχx
and kχy can be calculated by
Solution of the transverse tunnel displacement uy ⎧P
x
⎪
According to the finite difference method, the difference expression ⎪
⎨ |δ | Δx⩽|δx |
(44)
x
of Eq. (11) is: kxχ =
⎪ χ
⎩ Pxu
⎪
0⩽|δx | < Δx
EI ( ) Ni ( ) Δx
χ 4 uy,i+2 − 4uy,i+1 + 6uy,i − 4uy,i− 1 + uy,i− 2 − χ 2 uy,i+1 − 2uy,i + uy,i− 1
ky,i l ky,i l ⎧
Py ⃒ ⃒
⎪ Δy⩽⃒δy ⃒
1 ( ) ⎪ ⃒⃒δ ⃒⃒
⎪
⎨
− χ 2 (Ni+1 − Ni− 1 ) uy,i+1 − uy,i− 1 + uy,i = Δfy,i y
4ky,i l kyχ = (45)
⎪
⎪ Pχ ⃒ ⃒
(31) ⎪
⎩ yu 0⩽⃒δy ⃒ < Δy
Δy
where Ni, kχy,i , uy,i and Δfy,i are the axial force, transverse stratum
where Pχxu and Pχyu are the bearing capacities of the axial and trans­
stiffness, transverse tunnel displacement and transverse fault displace­
ment of node i, respectively. verse stratum, respectively; Δx and Δy are the corresponding yield
The axial force N, bending moment M and shear force V at both ends displacements of the tunnel-stratum interaction; and the relative axial
of the tunnel are zero. According to the finite difference method and the and transverse tunnel-stratum displacements δx and δy can be calculated
elastic beam theory, the difference expressions of the boundary condi­ by Eq. (16) and Eq. (10), respectively.
The axial and transverse stratum bearing capacity Pxu and Pyu can be
χ χ
tions are:
calculated by [42]:
EI ( )
M0 = − uy,1 − 2uy,0 + uy,− 1 = 0 (32) 1 + K0
l2 Pχxu = παχ cχ D + πCDγχ tan(f1 φχ ) (46)
2
EI ( )
Mn = − 2 uy,n+1 − 2uy,n + uy,n− 1 = 0 (33)
l
χ χ
Pχyu = Nch χ χ
c D + Nqh γ CD (47)

EI ( ) where c is the stratum cohesion, γ is the stratum unit weight, K0 is

χ χ
V0 = − uy,2 − 2uy,1 + 2uy,− 1 − uy,− 2 = 0 (34) the lateral pressure coefficient, αχ is the adhesion factor, φχ is the stra­
2l3
tum friction angle, f1 is the coating dependent factor, and the bearing
Vn = −
EI ( )
uy,n+2 − 2uy,n+1 + 2uy,n− 1 − uy,n− 2 = 0 (35) capacity factors Nch
χ
and Nqh χ
can be interpreted from design charts
2l3 [42].
Therefore, the transverse tunnel displacement uy of nodes − 1, − 2, n
+ 1 and n + 2 can be obtained as follows:

6
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

Fig. 8. Nonlinear tunnel-stratum interaction curves [42] in the (a) axial and (b) transverse directions.

Calculation process EI ( )
To obtain the values of the nonlinear stratum stiffnesses kχx and kχy , a Mi = − uy,i+1 − 2uy,i + uy,i− 1 (49)
l2
novel iterative method is proposed in this section. In short, the stratum
stiffnesses kχx and kχy are continuously decreased until the tunnel dis­ Vi = −
EI ( ) Ni (
uy,i+2 − 2uy,i+1 + 2uy,i− 1 − uy,i− 2 − uy,n+1 − uy,n− 1
)
(50)
placements and axial force N converge. The specific calculation process 2l3 2l
is as follows (see Fig. 9): Under strike-slip faulting, the maximum stress occurs at the tunnel
hances, as shown in Fig. 4. The stress σi and strain εi at the left and right
(1) Pretreatment tunnel hances of node i are respectively:
• Input parameters.
Ni Mi (D − t)
• Calculate Δfx,i and Δfy,i by Eq. (1) and Eq. (2). σ i,left = + (51)
• Assume the membrane axial force Nb,i,0=0.. A 2I
• Assume the the stratum stiffnesses khx,i,0 = Phxu /Δx, khy,i,0 = Phyu /Δy, Ni Mi (D − t)
f f f f σ i,right = − (52)
kx,i,0 = Pxu /Δx and ky,i,0 = Pyu /Δy. A 2I
(2) The 1st iteration calculation. σ i,left
• Calculate the initial tunnel displacements ux,i,1 and uy,i,1 by Eq. (27) εi,left = (53)
E
and Eq. (40).
• Calculate the initial relative tunnel-stratum displacements δx,i,1 and εi,right =
σ i,right
(54)
δy,i,1by Eq. (16) and Eq. (10). E
f
• Calculate the initial stratum stiffnesses khx,i,1 , khy,i,1 , kx,i,1 and ky,i,1 by
f
Note that the positive strain indicates the tensile state, and the
Eq. (44) and Eq. (45). negative strain represents the compressive state.
• Calculate the initial membrane axial force Nb,i,1 by Eq. (21).
(3) The j (j > 1)th iteration calculation. Comparison with the model test
• Calculate the jth tunnel displacements ux,i,j and uy,i,j by substituting
f f Parameters
the j-1th stratum stiffnesses khx,i,j− 1 , khy,i,j− 1 , kx,i,j− 1 and ky,i,j− 1 into Eq.
(27) and Eq. (40).
Wang et al. [43] conducted 1 g model tests to investigate the me­
• Calculate the jth relative tunnel-stratum displacements δx,i,j and δy,i,j
chanical response of a tunnel subjected to right-lateral strike-slip fault­
by Eq. (16) and Eq. (10).
ing, as shown in Fig. 10.
f f
• Calculate the jth stratum stiffnesses khx,i,j , khy,i,j , kx,i,j and ky,i,j by Eq. The detailed testing parameters are shown in Table 1.
(44) and Eq. (45). In the theoretical model analysis, the tunnel-stratum interaction
• Calculate the jth membrane axial forceNb,i,j by Eq. (21). properties are calculated by the guideline [42], as shown in Table 2.
Note that the coating-dependent factor f1 of the tunnel-stratum interface
Equations of the tunnel forces and stresses is 0.7, as smooth gypsum is used to simulate the tunnel in the model
tests.
The tunnel axial force Ni, bending moment Mi and shear force Vi of
node i can be calculated according to the finite difference method in
Results
difference form:
EA ( ) EA( )2 The longitudinal strain and transverse displacement at the right
Ni = ux,i+1 − ux,i− 1 + 2 uy,i+1 − uy,i− 1 (48)
2l 8l hance obtained by the proposed model and model test are plotted in
Fig. 11. Note that, in the condition of only considering the nonlincar

7
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

tunnel-stratum interaction (i.e., analytical model II), the parameters of

the host rock are employed in the entire model; in the condition of only
considering the fault zone width (i.e., analytical model III), the linear
stratum stiffnesses are calculated by the equations listed in Appendix C
[44]. Specifically, the strain is symmetric about the fault plane, first
increases and then decreases as the fault distance increases and finally
reaches zero at both ends of the tunnel. The value of longitudinal strain
at the fault plane is zero, and the maximum value is located near the
fault. Under strike-slip faulting, the transverse displacement of the
tunnel shows an ‘S’ shape. In addition, the results of the proposed model
agree well with the 1 g model test, but ignoring the influence of fault
zone width or nonlinear tunnel-stratum interaction will result in obvi­
ously larger results.

Comparison with the 3D FEM numerical model

To comprehensively verify the proposed model, the calculation re­

sults of tunnel displacements and forces obtained by the proposed model
are compared with the results from the 3D FEM numerical model.

Establishment of the FEM numerical model

As shown in Fig. 12, an FEM numerical model with a geometry of

Table 1
Model testing parameters [43].
Parameter Value

Tunnel D Width 0.24 m

t Thickness 0.015 m
C Buried depth 0.72 m
L Length 2.8 m
LF Length in the footwall 1.4 m
LH Length in the hanging wall 1.4 m
β Crossing angle 90◦
E Young’s modulus 0.5 GPa
ν Poisson’s ratio 0.25
Fault W Width 0.72 m
WF Width in the footwall 0.36 m
WH Width in the hanging wall 0.36 m
Δf Fault displacement − 0.013 m
Stratum γf Unit weight of the fault zone 16.3 kN/m3
cf Cohesion of the fault zone 3.2 kPa
φf Friction angle of the fault zone 26◦
Ef Young’s modulus of the fault zone 17 MPa
νf Poisson’s ratio of the fault zone 0.45
γh Unit weight of the host rock 16.8 kN/m3
Fig. 9. Calculation flow chart. ch Cohesion of the host rock 11.4 kPa
φh Friction angle of the host rock 38◦
Eh Young’s modulus of the host rock 108 MPa
νh Poisson’s ratio of the host rock 0.35

Fig. 10. Model testing equipment [43].

8
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

Table 2
Tunnel-stratum interaction parameters[42].
Parameter Value

Phxu Axial bearing capacity of the host rock ch 0.274 0.695 11.4 0.274 0.695
αh = 0.608 − 0.123 − +( = 0.608 − 0.123 × − ( ) +( )
100 ( ch )2 ch
)3 100 11.4 2 11.4 3
+1 +1 +1 +1
100 100 100 100
= 1.018
h1 + K0 ( )
Phxu h h
= πα c D + πDCγ tan f1 φh
2
1+1
= 3.14 × 1.018 × 11.4 × 0.24 + 3.14 × 0.24 × 0.72 × 16.8 × × tan(0.7 × 38◦ )
2
= 13.306kN/m
f
Pxu Axial bearing capacity of the fault zone f cf 0.274 0.695 3.2 0.274 0.695
α = 0.608 − 0.123 − +( = 0.608 − 0.123 × − ( ) +( )
100 ( cf )2 cf
)3 100 3.2 2 3.2 3
+1 +1 + 1 +1
100 100 100 100
= 1.025
1 + K0 ( )
Pfxu = παf cf D + πDCγf tan f1 φf
2
1+1
= 3.14 × 1.025 × 3.2 × 0.24 + 3.14 × 0.24 × 0.72 × 16.3 × × tan(0.7 × 26◦ )
2
= 5.380kN/m
[ ]
Phyu Transverse bearing capacity of the host rock 0.72 11.063 7.119
Nhch = min 6.752 + 0.065 × − ( )2 + ( )3 , 9
0.24 0.72 0.72
+1 +1
0.24 0.24
= 6.367
( ) ( ) ( )
0.72 0.72 2 0.72 3 0.72 4
Nhqh = 6.816 + 2.019 × − 0.146 × + 7.651 × 10− 3
× − 1.683 × 10− 4
×
0.24 0.24 0.24 0.24
= 11.752
Phyu = Nhch ch D + Nhqh γh CD = 6.367 × 11.4 × 0.24 + 5.180 × 16.8 × 0.72 × 0.24
= 51.536kN/m
[ ]
f
Pyu Transverse bearing capacity of the fault zone 0.72 11.063 7.119
Nfch = min 6.752 + 0.065 × − ( )2 + ( )3 , 9
0.24 0.72 0.72
+1 +1
0.24 0.24
= 6.367
( ) ( ) ( )
0.72 0.72 2 0.72 3 0.72 4
Nfqh = 3.332 + 0.839 × − 0.090 × + 5.606 × 10− 3
× − 1.319 × 10− 4
×
0.24 0.24 0.24 0.24
= 5.180
Pfyu = Nfch cf D + Nfqh γf CD = 6.367 × 3.2 × 0.24 + 5.180 × 16.3 × 0.72 × 0.24
= 17.479kN/m
Δx Axial yield displacement 0.01 m
Δy Transverse yield displacement 0.015 m
K0 Lateral pressure coefficient 1
f1 Coating dependent factor 0.7

Fig. 11. Tunnel response comparison with the 1 g model test [43]: (a) longitudinal strain and (b) vertical displacement.

100 m × 200 m × 85 m is established using ABAQUS to simulate the tangential friction coefficients are tan(φf) and tan(0.6φχ) [42], respec­
tunnel’s response under strike-slip faulting. Both the stratum and the tively, because there is usually a plastic waterproof plate between the
tunnel are modelled by 3D solid continuum elements. The stratum is an tunnel and the stratum.
ideal elastic-plastic material conforming to the Mohr–Coulomb strength
criterion, and the elastic model is adopted for the tunnel. The normal
Parameters
contact behavior between the hanging wall and footwall, as well as
between the stratum and tunnel, is defined as hard contact, while the
Note that a large strike-slip displacement of 3 m is considered, and
tangential behavior is set as penalty contact. The corresponding
the calculation includes three simplified steps:

9
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

Fig. 12. Geometry and mesh of the FEM numerical model.

(1) The initial ground stress balance.

Table 3
(2) The installation of tunnel lining.
Calculation parameters.
(3) The strike-slip displacement Δf is applied to the model bound­
aries to simulate fault dislocation, as shown in Fig. 13. Parameter Value
The detailed calculation parameters are listed in Table 3. Note that Tunnel W Width 10 m
the host rock has a larger unit weight than the fault zone, and this paper t Thickness 0.35 m
H Height 10 m
takes 21 kN/m3 and 19 kN/m3 as an example.
C Buried depth 30 m
The tunnel-stratum interaction parameters employed in the pro­ L Length 200 m
posed analysis model are obtained by Eq. (46) and Eq. (47), as listed in LF Length in the footwall 100 m
Table 4. Note that since the normal and tangential contact behaviors LH Length in the hanging wall 100 m
between the stratum and the tunnel are set as hard and penalty contact, β Crossing angle 90◦
E Young’s modulus 31.5 GPa
respectively, the value of αχ equals 0.
ν Poisson’s ratio 0.2
Fault W Width 50
WF Width in the footwall 25
Results WH Width in the hanging wall 25
Stratum Ef Young’s modulus of the fault zone 0.5 GPa
As plotted in Fig. 14, the transverse tunnel displacement and forces νf Poisson’s ratio of the fault zone 0.4
γf Unit weight of the fault zone 19 kN/m3
(bending moment, axial force and shear force) are symmetric about the
cf Cohesion of the fault zone 100 kPa
fault. The bending moment is almost zero at the tunnel-stratum crossing φf Friction angle of the fault zone 18◦
point (at x = 0 m) and increases with increasing fault distance. The axial Eh Young’s modulus of the host rock 1.2 GPa
and shear forces reach their peak value at the tunnel-stratum crossing νh Poisson’s ratio of the host rock 0.33
point and decrease with increasing fault distance. The maximum value γh Unit weight of the host rock 21 kN/m3
ch Cohesion of the host rock 400 kPa
of the bending moment is reached approximately 25 m away from the φh Friction angle of the host rock 28◦

fault plane and finally converges to zero at approximately 50 m away

from the fault plane. In addition, the comparisons also show that the
calculation results of the theoretical model (i.e., analytical model I)
considering the influence of fault zone width and nonlinear tunnel-
stratum interaction are in good agreement with the numerical simula­
tion results. Otherwise, the tunnel response will be highly enhanced by
approximately 40 %~110 % (i.e., analytical models II and III) under this
condition.
The peak axial force, shear force and bending moment calculated by
the proposed model and the 3D FEM model are plotted in Fig. 15.
Compared with the numerical model, the relative errors of the peak
bending moment, shear force and axial force calculated by analytical
model I (considering the fault zone width and nonlinear tunnel-stratum
Fig. 13. Numerical model boundary conditions.

10
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

Table 4
Tunnel-stratum interaction parameters.
Parameter value

Phxu Axial bearing capacity of the host rock αh = 0

1 + K0 ( ) 1+1
Phxu = παh ch D + πDCγh tan f1 φh = 0 + 3.14 × 10 × 30 × 21 × × tan(0.6 × 28◦ )
2 2
= 5972.538kN/m
f
Pxu Axial bearing capacity of the fault zone αf = 0
1 + K0 ( ) 1+1
Pfxu = παf cf D + πDCγf tan f1 φf = 0 + 3.14 × 10 × 30 × 19 × × tan(0.6 × 18◦ )
2 2
= 3414.226kN/m
[ ]
Phyu Transverse bearing capacity of the host rock 30 11.063 7.119
Nhch = min 6.752 + 0.065 × − ( )2 + ( )3 , 9
10 30 30
+1 +1
10 10
= 6.340
( )2 ( )3 ( )4
30 30 30 30
Nhqh = 4.565 + 1.234 × − 0.089 × + 4.275 × 10− 3
× − 9.159 × 10− 5
×
10 10 10 10
= 7.574
Phyu = Nhch ch D + Nhqh γh CD = 6.340 × 400 × 10 + 7.574 × 21 × 30 × 10
= 73075.427kN/m
[ ]
f
Pyu Transverse bearing capacity of the fault zone 30 11.063 7.119
Nfch = min 6.752 + 0.065 × − ( )2 + ( )3 , 9
10 30 30
+1 +1
10 10
= 6.340
( )2 ( )3 ( )4
30 30 30 30
Nfqh = 2.399 + 0.439 × − 0.030 × + 1.059 × 10− 3
× − 1.754 × 10− 5
×
10 10 10 10
= 3.473
Pfyu = Nfch cf D + Nfqh γf CD = 6.340 × 100 × 10 + 3.473 × 19 × 30 × 10
= 26136.879kN/m
Δx Axial yield displacement 0.01 m
Δy Transverse yield displacement 0.1 m
K0 Lateral pressure coefficient 1.0
f1 Coating dependent factor 0.6

Fig. 14. Tunnel response comparison with the 3D FEM numerical model.

11
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

16000 2500

2000
12000
1500
8000
1000
4000
500

0 0

Fig. 15. Comparison of the tunnel peak forces with different analytical models.

interaction) are 6.37 %, 11.00 % and 9.91 %, respectively; the relative and analytical model III are 9.09 %, 41.02 % and 113.43 %, respectively.
errors of the peak bending moment, shear force and axial force calcu­
lated by analytical model II (only considering the nonlinear tunnel- Parameter analysis
stratum interaction) are 31.67 %, 60.01 % and 31.32 %, respectively;
and the relative errors of the peak bending moment, shear force and In previous tunnel and pipeline analytical models [26–36], the ulti­
axial force calculated by analytical model III (only considering the fault mate strain of the corresponding materials is usually employed as the
zone width) are 63.00 %, 199.92 % and 77.36 %, respectively. In criterion of structural safety. Therefore, to assess the safety of tunnels
addition, the average errors of analytical model I, analytical model II crossing active strike-slip fault zones, the strains at tunnel hances and

Fig. 16. Strain distribution of tunnel hances under left- and right- lateral strike-slip faulting.

12
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

failure scope with different parameters are analysed in this section. The strike-slip faulting both linearly increase with increasing fault
failure scope is determined by the ultimate tensile and compressive displacement. The maximum strains are 6220, 8158, 9985, 11,730 and
strain of the concrete. Additionally, the ultimate compressive strain εcu 13,356 με for 1.0, 1.5, 2.0, 2.5 and 3.0 m of fault displacement,
of C35 concrete is 3300 με, and the ultimate tensile strain εtu is 100 με respectively. Additionally, the tunnel failure scope also linearly in­
[45]. The tunnel length L is 1000 m, the fault displacement Δf is 2 m, the creases as the fault displacement increases. Specifically, when the fault
crossing angle β is 80◦ or 100◦ , and the other parameters are the same as displacement is 1.0–3.0 m, the tunnel failure scope is 208–352 m and
in Table 3. 79–97 m under left- and right- lateral strike-slip faulting, respectively.

Longitudinal strain distribution along the tunnel Effect of the fault zone width W

As shown in Fig. 16, the strain distribution of the left and right tunnel The effect of the fault zone width W on the tunnel response is
hances is symmetrical about the fault plane under strike-slip faulting. investigated by changing the fault zone width as W = 0, 10, 25, 50 and
The maximum strain is located around the fault plane, and when the 100 m. In Fig. 18, the maximum tunnel compressive and tensile strains
fault plane distance is greater than 150 m, the tunnel strain tends to zero. under left- and right- lateral strike-slip faulting both decrease with
Under left-lateral strike-slip faulting with a crossing angle β = 80◦ , the increasing fault zone width until W > 50 m, and the maximum strain is
maximum tensile strain occurs at the right tunnel hance of the hanging 14224–9148 με for different fault zone widths. The tunnel failure scope
wall and left tunnel hance of the footwall, while the maximum increases with increasing fault zone width. For instance, when the fault
compressive strain occurs at the left tunnel hance of the hanging wall zone width is 0–100 m, the failure scope is 282–295 m and 79–93 m
and right tunnel hance of the footwall; under left-lateral strike-slip under left- and right- lateral strike-slip faulting, respectively.
faulting with β = 100◦ , the tunnel response is symmetric with that of left-
lateral strike-slip faulting-. In addition, under left-lateral strike-slip Effect of the fault zone cohesion cf
faulting with β = 80◦ , the failure scope is governed by the tensile strain,
which is 290 m; under left-lateral strike-slip faulting with β = 100◦ , the To investigate the influence of the fault zone cohesion cf on the
failure scope is governed by the compressive strain, which is 86 m. In tunnel responses, five cases with different fault zone cohesion cf are
addition, under right-lateral strike-slip faulting, the tunnel response is investigated, with values increasing from 50 kPa to 400 kPa. Fig. 19
symmetric with that of left-lateral strike-slip faulting. The results show shows that the maximum tension and compressive strains monotonically
that the failure scope under left-lateral strike-slip faulting is obviously increase with increasing fault zone cohesion, and the maximum strain is
larger than that under right-lateral strike-slip faulting under the condi­ 9748–11702 με for different fault zone cohesions. In contrast, the failure
tion of 0◦ < β < 90◦ because the ultimate tensile strain of concrete is scope decreases with increasing fault zone cohesion. For instance, when
much smaller than its ultimate compressive strain. However, under the the fault zone cohesion is 50–400 kPa, the failure scope is 292–281 m
condition of 90◦ < β < 180◦ , the opposite conclusion is obtained. and 89–78 m under left- and right- lateral strike-slip faulting, respec­
Then, the effect of fault displacement Δf, fault zone width W, fault tively. This conclusion is also obtained by Zhong et al. [21], which is
zone cohesion cf and crossing angle β on the maximum strain and failure because the stratum bearing capacity and stiffness increase with
scope of the tunnel is discussed under the condition of 0◦ < β < 90◦ . The increasing cohesion according to Eq. (44) ~ Eq. (47).
tunnel response under the condition of 90◦ < β < 180◦ can be obtained
based on Fig. 16. Effect of the crossing angle β

Effect of the fault displacement Δf To investigate the influence of the crossing angle β on the mechanical
responses of the tunnel, a parametric analysis is performed where the
To investigate the influence of the fault displacement Δf on the crossing angle β is 20◦ , 40◦ , 60◦ and 80◦ . As shown in Fig. 20, the
tunnel responses, a parametric analysis is performed where the fault maximum compressive strain under left- lateral strike-slip faulting and
displacement Δf is 1.0, 1.5, 2.0, 2.5 and 3.0 m. In Fig. 17, the maximum the maximum tensile strain under right- lateral strike-slip faulting in­
tunnel compressive and tensile strains under left- and right- lateral crease with increasing crossing angle, while the maximum tensile strain

400 130

120
300
110

100
200

90

100
80

70
0 0.0

Fig. 17. Tunnel response under strike-slip faulting with different fault displacements.

13
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

300 110
295
100
290
285 90
280
275 80

270
70
265
260 60
0.00 0.0

Fig. 18. Tunnel response under strike-slip faulting with different fault zone widths.

300 100
295
95
290
90
285
280 85
275
80
270
75
265
260 70
0.0 0.0

Fig. 19. Tunnel response under strike-slip faulting with different fault zone strengths.

700 350

600 300

500 250

400 200

300 150

200 100

100 50

0 0

Fig. 20. Tunnel response under strike-slip faulting with different crossing angles.

14
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

under left- lateral strike-slip faulting and the maximum compressive the maximum strain is located around the fault plane. Under left-
strain under right- lateral strike-slip faulting first increase and then lateral strike-slip faulting with 0◦ < β < 90◦ , the failure scope is
decrease with increasing crossing angle. Note that under the condition of governed by the tensile strain. Under right-lateral strike-slip
90◦ < β < 180◦ , the opposite conclusion is obtained. Because the axial faulting with 0◦ < β < 90◦ , the failure scope is governed by the
component of the fault displacement Δfx decreases with increasing compressive strain. The failure scope under left-lateral strike-slip
crossing angle in the condition of 0◦ < β < 90◦ and decreasing crossing faulting is obviously larger than that under right-lateral strike-
angle in the condition of 0◦ < β < 90◦ , resulting in the reduction of axial slip faulting because the tensile strength of concrete is much
force N. smaller than its compressive strength. However, under the con­
dition of 90◦ < β < 180◦ , the opposite conclusion is obtained.
Conclusion (4) The value of the tunnel maximum strain is directly related to the
fault displacement Δf and fault zone cohesion cf and has an in­
In this work, a mechanical analysis model for tunnels under strike- verse relationship with the fault zone width Wf; the tunnel failure
slip faulting considering the fault zone width and nonlinear tunnel- scope is directly related to the fault displacement Δf, fault zone
stratum interaction is first established. More specifically, tunnels width Wf, and crossing angle β (0◦ < β < 90◦ ) and has an inverse
crossing active fault zones are simplified as an elastic beam acting on a relationship with the crossing angle β (90◦ < β < 180◦ ) and fault
nonlinear foundation; the nonlinear tunnel-stratum interaction is zone cohesion cf.
considered by a series of nonlinear axial and transverse nonlinear
springs; the different mechanical characteristics of the fault zone and Additionally, the range of application of this proposed model and its
host rock are described by different spring stiffnesses; and the nonlinear limitations should be noted as follows:
tunnel-stratum interaction is analysed by the finite difference method
and a novel iterative method for nonlinear spring stiffness. The con­ • The proposed model applies to strike-slip faults but does not extend
clusions are summarized as follows: to dip-slip faults.
• In this model, the tunnel is assumed to possess elastic properties,
(1) The displacement, bending moment, shear force and axial force of while disregarding the elastic-plastic characteristics of concrete.
the proposed model agree well with the results obtained from • The proposed model only compares the strain and displacement of
model tests and numerical simulations. Compared with the nu­ the tunnel with the model tests. The stratum pressure and tunnel
merical model, the relative errors of the peak bending moment, failure scope should be considered in future works.
shear force and axial force calculated by the proposed model are
6.37 %, 11.00 % and 9.91 %, respectively, and the average
relative error is 9.09 %. Declaration of competing interest
(2) The proposed model significantly improves the accuracy of
existing analytical models. More specifically, neglecting the ef­ The authors declare that they have no known competing financial
fect of the fault zone width or nonlinear tunnel-stratum interac­ interests or personal relationships that could have appeared to influence
tion, the tunnel response will be misestimated, e.g., when the the work reported in this paper.
fault displacement is 3 m and the crossing angle is 90◦ , the
overestimated value is approximately 40 % to 110 %. The relative Data availability
average errors of analytical model I (considering the fault zone
width and nonlinear tunnel-stratum interaction), analytical No data was used for the research described in the article.
model II (only considering the nonlinear tunnel-stratum inter­
action) and analytical model III (only considering the fault zone Acknowledgement
width) are 9.09 %, 41.02 % and 113.43 %, respectively.
(3) The strain distribution of the left and right tunnel hances is The authors are grateful for the support from the National Natural
symmetrical about the fault plane under strike-slip faulting, and Science Foundation of China (52378411 and 52208404).

Appendix A

⎡ ⎤
− 2 2
⎢ χ χ 0 0 ... 0 ⎥
⎢ kx,0 kx,0 ⎥
⎢ ⎥
⎢ 1 − 2 1 ⎥
⎢ ⎥
⎢ χ χ χ 0 ... 0 ⎥
⎢ kx,1 kx,1 kx,1 ⎥
⎢ ⎥
⎢ ⎥
EA ⎢
⎢
1 − 2 − 1
0 ⎥
⎥
[A1 ] = 2 ⎢ 0 χ
kx,2 χ
kx,2 χ
kx,2
... ⎥ (A1)
l ⎢ ⎥
⎢ ⎥
⎢ 1 − 2 ⎥
⎢ 0 0 ... 0 ⎥
⎢ χ
kx,3 χ
kx,3 ⎥
⎢ ⎥
⎢ ... ... ... ... ... ... ⎥
⎢ ⎥
⎢ ⎥
⎣ − 2⎦
0 0 0 0 ... χ
kx,n
(n+1)×(n+1)

15
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

⎡ ⎤
1 0 0 0 ... 0
⎢0 1 0 0 ... 0⎥
⎢ ⎥
⎢0 0 1 0 ... 0⎥
[A2 ] = ⎢
⎢0
⎥ (A2)
⎢ 0 0 1 ... 0⎥ ⎥
⎣ ... ... ... ... ... ... ⎦
0 0 0 0 ... 1 (n+1)×(n+1)

Appendix B

⎡ ⎤
2 − 4 2
⎢ kχ χ χ 0 ... 0 ⎥
⎢ y,0 ky,0 ky,0 ⎥
⎢ ⎥
⎢− 2 5 − 4 1 ⎥
⎢ ⎥
⎢ χ χ χ χ ... 0 ⎥
⎢ ky,1 ky,1 ky,1 ky,1 ⎥
⎢ ⎥
⎢ ⎥
EI ⎢
⎢
1 − 4 6 4
0 ⎥
⎥
[B1 ] = 4 ⎢ kχ χ
ky,2 χ
ky,2
− χ
ky,2
... ⎥ (B1)
l ⎢ y,2 ⎥
⎢ ⎥
⎢ 1 − 4 6 ⎥
⎢ 0 ... 0 ⎥
⎢ χ
ky,3 χ
ky,3 χ
ky,3 ⎥
⎢ ⎥
⎢ ... ... ... ... ... ... ⎥
⎢ ⎥
⎢ ⎥
⎣ 2 ⎦
0 0 0 0 ... χ
ky,n
(n+1)×(n+1)

⎡ ⎤
− 2N0 2N0
⎢ kχ χ 0 0 ... 0 ⎥
⎢ y,0 ky,0 ⎥
⎢ ⎥
⎢ N − 2N1 N1 ⎥
⎢ 1 ⎥
⎢ χ χ χ 0 ... 0 ⎥
⎢ ky,1 ky,1 ky,1 ⎥
⎢ ⎥
⎢ ⎥
1⎢⎢
N2 − 2N2 N2
0 ⎥
⎥
[B2 ] = 2 ⎢ 0 χ
ky,2 χ
ky,2 χ
ky,2
... ⎥ (B2)
l ⎢ ⎥
⎢ ⎥
⎢ N3 − 2N3 ⎥
⎢ 0 0 ... 0 ⎥
⎢ χ
ky,3 χ
ky,3 ⎥
⎢ ⎥
⎢ ... ... ... ... ... ... ⎥
⎢ ⎥
⎢ ⎥
⎣ − 2Nn ⎦
0 0 0 0 ... χ
ky,n
(n+1)×(n+1)

⎡ ⎤
0 0 0 0 ... 0
⎢ ⎥
⎢ − (N2 − N0 ) N2 − N0 ⎥
⎢ χ 0 χ 0 ... 0 ⎥
⎢ ky,1 ky,1 ⎥
⎢ ⎥
⎢ ⎥
⎢ − (N3 − N1 ) (N3 − N1 ) ⎥
1 ⎢
⎢ 0 0 ... 0 ⎥⎥
(B3)
χ χ
[B3 ] = 2 ⎢ ky,2 ky,2 ⎥
4l ⎢ ⎥
⎢ ⎥
⎢ − (N4 − N2 ) ⎥
⎢ 0 0 χ 0 ... 0 ⎥
⎢ ky,3 ⎥
⎢ ⎥
⎢ ... ... ... ... ... ... ⎥
⎣ ⎦
0 0 0 0 ... 0
(n+1)×(n+1)

⎡ ⎤
1 0 0 0 ... 0
⎢0 1 0 0 ... 0 ⎥
⎢ ⎥
⎢0 0 1 0 ... 0 ⎥
[B4 ] = ⎢
⎢0 0
⎥ (B4)
⎢ 0 1 ... 0 ⎥ ⎥
⎣ ... ... ... ... ... ... ⎦
0 0 0 0 ... 1 (n+1)×(n+1)

Appendix C. [44]

√̅̅̅̅̅̅̅̅̅̅̅
1.3Eχ Eχ D4
χ
ky = [ ] 12
(C1)
D 1 − (νχ )2 EI

16
X. Zhang et al. Transportation Geotechnics 44 (2024) 101171

References [21] Zhong ZL, Wang Z, Zhao M, Du XL. Structural damage assessment of mountain
tunnels in fault fracture zone subjected to multiple strike-slip fault movement [J].
Tunn Undergr Space Technol 2020;104:103527.
[1] Celestino MAL, de Miranda TS, Mariano G, de Lima Alencar M, de Carvalho BRBM,
[22] Shen YS, Gao B, Yang XM, Tao SJ. Seismic damage mechanism and dynamic
da Cruz Falcão T, Topan JG, Barbosa JA, Gomes IF. Fault damage zones width:
deformation characteristic analysis of mountain tunnel after Wenchuan earthquake
Implications for the tectonic evolution of the northern border of the Araripe Basin,
[J]. Eng Geol 2014;180:85–98.
Brazil, NE Brazil[J]. J Struct Geol 2020;138:104116.
[23] Newmark NM, Hall WJ. Pipeline design to resist large faultdisplacement [C]. Ann
[2] Aygar EB, Gokceoglu C. A special support design for a large-span tunnel crossing an
Arbor: University of Michigan; 1975. p. 416–25.
active fault (T9 Tunnel, Ankara-Sivas High-Speed Railway Project, Turkey) [J].
[24] Kennedy RP, Williamson RA, Chow AM. Fault movement effects on buried oil
Environ Earth Sci 2021;80(1):37.
pipeline [J]. Transp Eng J ASCE 1977;103(5):617–33.
[3] Kenner SJ, Segall P. Postseismic Deformation Following the 1906 San Francisco
[25] Wang LRL, Yeh YH. A refined seismic analysis and design of buried pipeline for
Earthquake [J]. J Geophys Res 2000;105(B6):13195–210.
fault movement [J]. Earthq Eng Struct Dyn 1985;13(1):75–96.
[4] Wang WL, Wang TT, Su JJ, Lin CH, Huang TH. Assessment of damage in mountain
[26] Karamitros DK, Bouckovalas GD, Kouretzis GP. Stress analysis of buried steel
tunnels due to the Taiwan Chi-Chi Earthquake [J]. Tunn Undergr Space Technol
pipelines at strike-slip fault crossings [J]. Soil Dyn Earthq Eng 2007;27(3):200–11.
2001;16(3):133–50.
[27] Trifonov OV, Cherniy VP. A semi-analytical approach to a nonlinear stress–strain
[5] Cui GY, Wang MN, Yu L, Lin GJ. Study on the characteristics and mechanism of
analysis of buried steel pipelines crossing active faults [J]. Soil Dyn Earthq Eng
seismic damage for tunnel structures on fault rupture zone in Wenchuan seismic
2010;30(11):1298–308.
disastrous area [J]. China Civil Eng J 2013;46(11):122–47.
[28] Karamitros DK, Bouckovalas GD, Kouretzis GP, Gkesouli V. An analytical method
[6] Zhang XP, Jiang YJ, Sugimoto S. Seismic damage assessment of mountain tunnel: A
for strength verification of buried steel pipelines at normal fault crossings [J]. Soil
case study on the Tawarayama tunnel due to the 2016 Kumamoto Earthquake [J].
Dyn Earthq Eng 2011;31(11):1452–64.
Tunn Undergr Space Technol 2018;71:138–48.
[29] Ni PP, Mangalathu S. Simplified evaluation of pipe strains crossing a normal fault
[7] Zhang W, Li M, Ji YP, Li GL, Zhao LX, Li SG. Analysis and enlightenment of typical
through the dissipated energy method [J]. Eng Struct 2018;167:393–406.
failure characteristics of tunnels caused by the Menyuan M6.9 earthquake in
[30] Hu ZP, Ren X, Wang QY, Wang R, Pan R. Analytical method for the mechanical
Qinghai Province [J]. China Earthq Eng J 2022;44(3):661–9.
response of buried pipeline under the action of strike-slip faulting [J].
[8] Tian SM, Wu KF, Yu L, Li X, Gong JF, Wang XD. Key Technology of Anti-Seismic for
Underground Space 2022;7(2):268–77.
Railway Tunnels Crossing Active Fault Zone [J]. Tunn Construc 2022;42(08):
[31] Talebi F, Kiyono J. A refined nonlinear analytical method for buried pipelines
1351–64.
crossing strike-slip faults [J]. Earthq Eng Struct Dyn 2021;50(11):2915–38.
[9] Burridge PB, Scott RF, Hall JF. Centrifuge study of faulting effects on tunnel [J].
[32] Zhao M, Xu LH, Huang JQ, Du XL, Li HF. Analytical solutions of the tunnels under
J Geotech Eng 1989;115(7):949–67.
the fault creeping by elastic foundation beam model with considering tangential
[10] Baziar MH, Nabizadeh A, Lee CJ, Hung WY. Centrifuge modeling of interaction
interaction [J]. Soil Dyn Earthq Eng 2023;172:108047.
between reverse faulting and tunnel [J]. Soil Dyn Earthq Eng 2014;65:151–64.
[33] Liu GZ, Qiao YF, He MC, Fan Y. An analytical solution of longitudinal response of
[11] Kiani M, Akhlaghi T, Ghalandarzadeh A. Experimental modeling of segmental
tunnels under dislocation of active fault [J]. Rock Soil Mech 2020;41(3):923–32.
shallow tunnels in alluvial affected by normal faults [J]. Tunn Undergr Space
[34] Yan GM, Zhao BM. Analytical Solution for Longitudinal Seismic Responses of Fault-
Technol 2016;51:108–19.
crossing Tunnels [J]. Adv Eng Sci 2023;55(02):26–38.
[12] Cai QP, Peng JM, Ng CWW, Shi JW, Chen XX. Centrifuge and numerical modelling
[35] Li HY, Li XG, Yang Y, Liu H. Analytical solution for the longitudinal response of
of tunnel intersected by normal fault rupture in sand [J]. Comput Geotech 2019;
cross-fault shield tunnel considering plastic deformation of circumferential joints
111:137–46.
[J]. Journal of Central South University 2023;30(5):1675–94.
[13] Sabagh M, Ghalandarzadeh A. Centrifugal modeling of continuous shallow tunnels
[36] Yan GM, Zhao BM, Wang ZJ, Gao B. Simplified analytical solution for responses of
at active normal faults intersection [J]. Transp Geotech 2020;22:100325.
fault-crossing tunnels with flexible joints under fault movement [J]. Structures
[14] Liu XZ, Li XF, Sang YL, Lin LL. Experimental study on normal fault rupture
2022;45:984–98.
propagation in loose strata and its impact on mountain tunnels [J]. Tunn Undergr
[37] Yu L, Zhang X, Wang MN, Yang HH, Luo X, Qin YL, et al. A simplified analytical
Space Technol 2015;49:417–25.
method for the tunnels’ mechanical behaviour under the action of active fault
[15] Cui Z, Li JH, Fu XW,, Sheng Q, Zhou GX, Wang TT, Ma YLN. Evaluating the
zones [J]. Computers and Geotechnics 2023;164:105789.
Response of a Tunnel Subjected to Strike-Slip Fault Rupture in Conjunction with
[38] Zhang ZQ, Zhang Y, Wei RH, Yin C, Zhang H. The damage mode and forced
Model Test and Hybrid Discrete-Continuous Numerical Modeling [J]. Rock Mech
response characteristics of articulated lining structure: A case study of metro tunnel
Rock Eng 2022;55(8):4743–64.
under a reverse fault action [J]. Tunn Undergr Space Technol 2023;140:105246.
[16] Sabagh M, Ghalandarzadeh A. Numerical modelings of continuous shallow tunnels
[39] Qiao YF, Tang J, Liu GZ, He MC. Longitudinal mechanical response of tunnels
subject to reverse faulting and its verification through a centrifuge [J]. Comput
under active normal faulting [J]. Underground Space 2022;7(4):662–79.
Geotech 2020;128:103813.
[40] Talebi F, Kiyono J. Introduction of the axial force terms to governing equation for
[17] Zhao K, Chen WZ, Yang DS, Zhao WS, Wang SY, Song WP. Mechanical tests and
buried pipeline subjected to strike-slip fault movements [J]. Soil Dyn Earthq Eng
engineering applicability of fibre plastic concrete used in tunnel design in active
2020;133:106125.
fault zones [J]. Tunn Undergr Space Technol 2019;88:200–8.
[41] Wang CM, Reddy JN, Lee KH. Shear deformable beams and plates. Relationships
[18] Zeng GX, Geng P, Guo XY, Li PS, Wang Q, Ding T. An anti-fault study of basalt fiber
with classical solutions [M]. Kindlington: Elsevier; 2000.
reinforced concrete in tunnels crossing a stick-slip fault [J]. Soil Dyn Earthq Eng
[42] American Lifelines Alliance-ASCE. Guidelines for the Design of Buried Steel Pipe
2021;148:106687.
[S]. Reston, VA, USA; 2005.
[19] Cui Z, Sheng Q, Zhang GM, Zhang MC, Mei XC. Response and mechanism of a
[43] Wang HR, Zhong ZL, Zhao M, Wang Z, Zhao X, Du XL. Model Experimental Study
tunnel subjected to combined fault rupture deformation and subsequent seismic
of the Influence of Strike slip Fault Dislocation on Tunnel [J]. J Beijing Univ
excitation [J]. Transp Geotech 2022;34:100749.
Technol 2021;47(07):691–701.
[20] Zhang X, Yu L, Wang MN, Yang HH. Mechanical response and failure
[44] Attewell P.B., Yeates J., Selby, A.R. Soil movements induced by tunnelling and
characteristics of tunnels subjected to reverse faulting with nonuniform
their effects on pipelines and structures [M]. United States: N. p., 1986. Web.
displacement: Theoretical and numerical investigation [J]. Engineering Failure
[45] CABR. Code for design of concrete structures (GB 50010− 2010) [S]. China
Analysis 2023;156:107809.
Architecture & Building Press, 2015.

17