APCOM & ISCM
th
11-14 December, 2013, Singapore
Finite element modeling of a tunnel affected by
dislocation of faults
*Xingwen Luo1,2, Zhenjun Yang2,3
1
State Key Laboratory of Geomechanics and Geotechnical Engineering, Institute of Rock and Soil
Mechanics, Chinese Academy of Sciences, Wuhan, Hubei 430071,China
2
School of Mechanical, Aerospace and Civil Engineering, The University of Manchester, Manchester,
M13 9PL, UK
3
College of Civil Engineering and Architecture, Zhejiang University, Hangzhou 310058, China
*Corresponding author: luoxw2005@hotmail.com
Abstract
Many tunnels built in mountainous areas go through active faults whose dislocation
can severely affect the stability, strength and serviceability of the tunnels. In this
study, finite element method was used to simulate the dislocation of the faults, and we
investigated the effects of main factors such as fault dip angle and tunnel lining
segment length on the internal stress and deformation of lining structures of the
tunnels. Failure mechanism of the lining structures was studied, and the sensitivity of
the main factors was analyzed. It is found that the optimal lining segment length of
the tunnel was 5m.
Keywords: Finite element method, Tunnels through faults, Lining structures,
Dislocation of faults, Internal stress, Deformation.
Introduction
Faults are often encountered in highway tunnel constructions. It is a normal way to
avoid faults in highway tunnel constructions in existing design standards[1]. But with
the rapid development of Chinese national economy and the exploitation of land
resources, a number of tunnels constructions projects inevitably encounter faults
problems, Especially, in strong earthquake areas. In these areas the faults which the
tunnels go through are often active faults. The tunnels go through active faults are
hugely impacted by the dislocation of the faults, which have direct impacts on the
constructions and operation safety of the tunnels.
The impacts of the faults fracture zone on the tunnel structures are mainly manifested
in below two aspects. The first is that the faults fracture zone take place dislocation
along the faults surfaces which are often called dislocation problems. The second is
that dynamic response characteristics of tunnels in the faults fracture zone under the
earthquake which are often called vibration problems. After investigating the designs
and studies on tunnels engineering which go through active faults, a lot of studies are
concentrated on the anti-seismic of tunnels. Appropriate engineering measures of
anti-seismic are relatively mature, but the measure of anti-relatively dislocation is
very little.
John Caulfield, et al (2004) designed the tunnel in southern California Claremont
water pressure tunnel by expanding its section size and reducing the length of the
lining segments. A lot of shear seams in the lining of the tunnel were set in southern
California Claremont water pressure tunnel projects[2,3].
1
Russo M, et al (2002) proposed to build relatively flexible connections in Turkey
Bolu highway tunnel projects.[4-6].
In China in the construction of WuXiaoLing tunnel (2004) also met the same problem.
The tunnel go through an activities fault. The design of the tunnel taken a way of
expanding section size and set aside shear displacement seams in the lining of the
tunnel[7].
A R Shahidi, et al (2005), in Greece Koohrang-Ⅲ water pressure tunnels project also
adopted a relatively flexible connection design. Numerical calculations was used to
certificate that the tunnel was safe when the fault take place dislocation[8].
Jiang Shuping, et al (2008), respectively discussed both of the methods of
overexcavation design and hinged design of the tunnels[9].
Hinge design is a way which minimize the length of lining segments. A certain range
of the lining segments structures besides the faults become relatively independent
after the tunnels which adopted hinge design. The rigid linings of the tunnels are
connected by flexible connectors. When the fault take place dislocation the
mechanical behavior are concentrated on the flexible connectors or part of the
structures, which will cause localized structural damage and not lead to an overall
destruction of the whole tunnel structures.
In order to fully reveal the mechanics, deformation and failure characteristics of the
lining structures of the tunnels under the dislocation of faults. Numerical simulation
of finite element method was adopted to analyze the mechanical and deformation
behavior of the tunnel structures under the activity of the dislocation of the faults.
Horseshoe-shaped section tunnels are the most widely used in tunnels engineering, so
horseshoe-shaped section tunnels were chosen in numerical simulation of the tunnels
which go through the faults. In order to reveal better the deformation and failure
mechanisms of the dislocation of the faults on the lining structures of the tunnels, we
carried out study on the mechanical behavior of the faults.
Engineering geology and tectonic
According to the regional geological data, geological survey and geophysical drilling
results, the engineering geology of tunnel zones were below successively: Quaternary
Holocene strata colluvium (Q4c+dl), Quaternary Holocene alluvial (Q4al+pl), Fourth
Department of Holocene debris accumulation layer (Q4sef) and the Triassic
Feixianguan group (Tf), Permian Yangxin group (Py), Liangshan group (Pl) and
Carboniferous Zongchanggou (Cz)[10].
Folds traces were not obvious in this tunnel area. The occurrence of the rock in the
entrance of the tunnel was 121°∠49°. The occurrence of the rock in the exit of the
tunnel was 321°∠71°. The inclination and incidence were subject to effects of the
tectonic, and they change quickly in the tunnel area. There were two kinds of faults
respectively named F22 and F22-1 going through the tunnel area. Fault F22 go
through the tunnel near the entrance of the tunnel, and Fault F22-1 go through the
tunnel at the middle of tunnel area. Fault F22-1 and Fault F22 were secondary faults
in Longmen Central Fault in China. The two faults active in the 5·12 Earthquake in
China and formed cracks in the body of the mountain. A 3-8m width fracture can be
seen in Fault F22, and this fracture formed a groove ground subsidence. The large
sinking height was approximately 2m. The fracture extends continuously.
2
Finite element model establishment and parameters selection
Modeling and model size
The study on the stress and deformation of tunnels structures were strict and complex
three dimensional problems in condition of faults taking place dislocation. The
longitudinal of tunnels was the largest impacted by the activity of the dislocation of
the fault. Therefore, the fault and the tunnel structures could be set orthogonal
relation in space when establishing a 3d finite element model. In order to fully reveal
the longitudinal deformation and mechanical behavior of the tunnel lining structures
under the dislocation of the tunnel and reduce the impact of the boundary of the
tunnel, the finite element model size was below: Length×Width ×Height = 120 m ×80
m × 80 m, namely vertical length of the tunnel was 200 m, the lateral width of the
tunnel was 60 m, the depth of the tunnel was 40 m and the depth of the top of tunnel
was 100 m.
Boundary conditions and fault activity simulation implementation
On the left and the right sides of the finite element calculation model, X direction and
Y direction horizontal displacement were restricted respectively. Z direction
displacement of the lower plate of the fault was restricted, while its upper plate could
active. The fault was a thrust fault. The active pattern was the upper plate of the fault
move up while its lower plate kept stable. In the model a displacement was added on
the upper plate to simulate the vertical dislocation of the fault. The average
dislocation on the fault was 0.5m. The basic conditions of the model were below:
A. The model was composed of bedrock, concrete lining of the tunnel and fault
fractured zones.
B. The tunnel went through the faults fractured zones in the model. The left part of
the fault was called the lower plate, and the right part of the faults was called the
upper plate. The two parts could dislocate along the fault surface. The dislocation
pattern was reverse faulting.
Material constitutive and calculations parameters
As geotechnical materials have heterogeneity, anisotropy, strength-difference effect
and other characters. Any one model can not fully accurately represent these
characteristics of geological materials. There are three categories of constitutive
models which are respectively elastic model, plastic model and elastic-plastic model
currently in rock and soil. In this analysis model soil and tunnel lining were assumed
to be ideal elastic-plastic material. D-P yield criterion and the associated flow rule
were adopted because D-P yield criterion could take into account the effects of
principal stress and hydrostatic pressure on yield and destruction of the tunnel. Its
yield surface smooth and non-angular. It help to determine the direction of the
increment of plastic strain and numerical calculations. The material parameters are
less and easy to be test or conversed by the Mohr-Coulomb criterion and its material
constant. D-P parameters can be conversed as follow formulas.
9c cos ϕ sin ϕ
σy = α=
9 + 3 sin 2 ϕ 9 + 3 sin 2 ϕ
Contact surface was set between the upper plate and the lower plate of the fault.
Between soil and the lining structure a contact surface was also set. Contact analysis
was made on these surfaces in the model. Friction coefficient of the contact surface
3
between the upper plate and the lower plate was 0.3. Friction coefficient between
strata and tunnel lining structures was 0.7. Coulomb friction model was adopted in
the numerical analysis. The inter force and deformation laws of the lining of the
tunnel were analyzed under different amount of dislocation in this model. The
average calculation dislocations amount was 50cm.
Rock, lining and fracture zones were simulated by solid element. The material of
initial support and secondary lining of the tunnel adopted elastic-plastic constitutive.
Rock adopted D-P yield criterion and incremental elastic-plastic constitutive.
Calculation parameters and calculation model are shown in Table 1 and Figure 1.
Table 1 Calculation parameters of rock, fracture zone and lining of the tunnel
Material’s Modulus Poisson's G Friction Cohesion
name /GPa ratio /kN.m-3 angle / o /MPa
Rock 0.5 0.38 22 34 0.3
Fracture
0.05 0.4 18 30 0.1
zone
Lining 28 0.20 25 58 3
Finite element numerical model of cross-fault tunnel under the dislocation of fault is
shown in Figure 1.
Figure 1 Finite element numerical model of cross-fault tunnel under the
dislocation of the fault
FEM numerical results
Four kinds of lining segment lengths of the tunnel were taken into account in FME
calculation model. They were respectively 15m, 10m, 5m and no shear seam in tunnel
which was mean the tunnel was continue. Tunnel lining segment length could be
optimized according to FEM calculation.
Response results contrast of different lining segments lengths of the tunnel are shown
in Fig.2- Fig.6.
4
0.6
vertical displacemnt /m
length of segments=5m
0.5
length of segments=10m
0.4 length of segments=15m
ling is continue
0.3
0.2
0.1
0
0 20 40 60 80 100 120
-0.1
longitudinal position /m
Figure 2 Vertical displacement of the arch top of the tunnel for different lining
segment lengths along longitudinal
0.6
vertical displacement /m
0.5 length of segments=5m
length of segments=10m
0.4 length of segments=15m
ling is continue
0.3
0.2
0.1
0
0 20 40 60 80 100 120
-0.1
longitudinal position /m
Figure 3 Vertical displacement of the arch bottom of the tunnel for different
lining segment lengths along longitudinal
2.40E+07 length of segments=5m
Mises equivalent stress /Pa
length of segments=10m
2.10E+07
length of segments=15m
1.80E+07 ling is continue
1.50E+07
1.20E+07
9.00E+06
6.00E+06
3.00E+06
0.00E+00
0 20 40 60 80 100 120
longitudinal position /m
Figure 4 Mises equivalent stress of the arch top of the tunnel for different lining
segment lengths along longitudinal
5
Figure 5 Contact stress of the tunnel which the length of segment is 10m under
the dislocation of the fault
length of segments=5m
5.0E+06
Contact stress / Pa
length of segments=10m
length of segments=15m
4.0E+06 ling is continue
3.0E+06
2.0E+06
1.0E+06
0.0E+00
0 20 40 60 80 100 120
longitudinal position /m
Figure 6 Contact stress of the arch bottom of the tunnel for different lining
segment lengths along longitudinal
Table 2 Maximum Mises stress of the tunnel for different lining segment
lengthss
Maximum Mises stress
Lining segment length
/MPa
No shear seam in tunnel 21.91
15m 20.98
10m 17.5
5m 13.12
6
Conclusions
According to dynamic response calculations and the comparison charts of the tunnel
structures for different lining segment lengths below conclusions can be get.
(1) It is better to adopt chain segmented structures than non-segmented structures in
tunnel engineering. Chain segmented structures tunnel can reduce the influence of the
dislocation of the fault on the tunnel structures.
Although vertical displacement along tunnel longitudinal increase with the lining
segment length of the tunnel decrease, but the affect zone of the displacement
mutation was significantly reduced. The affected zone of dislocation of the fault on
the tunnel was 40m when there was no shear seam in the ling structures. When the
lining segment length was 5m, the affected zone was the minimum, and the affected
length along tunnel longitudinal was shorter than the lining segment length.
(2) After adopting chain segmented structures, the maximum Mises stress of the
tunnel lining reduced and the stress reduced more quickly when the lining segment
length become shorter. The maximum Mises stress was13.12MPa when the lining
segment length was 5m, and it was 59.88% of the Mises stress of the tunnel with no
shear seam . While the maximum Mises were almost the same between the tunnel
which the lining segment length was 15m and the tunnel with no shear seam.
The distribution of Mises stress was more uniform with the lining segment length
decrease. The maximum stress was mainly concentrated on the lining arch waist of
the tunnel near the fault. Contact stress in the tunnel lining reaches maximum near the
surface of the fault. With the lining segment length decreases, contact stress of the
tunnel lining decrease. Contact stress of the tunnel changed abruptly from the tunnel
with no shear seam to the tunnel which the lining segments length was 15m.
(3) The plastic zone significantly reduced when the tunnel adopt chain segmented
structures than no-chain segmented structures. When the lining segment length was
15m, the plastic zone of the lining was mainly distributed in one lining segment
length range near the fault surface. When the lining segment length was 10 m, the
plastic zone of the lining was mainly distributed in one-two lining segment length
range near the fault surface. It can be seen the plastic zone of chain segmented
structures significantly reduced with the lining segment length decreases. The
maximum equivalent plastic strain decreased with the decrease of the lining segment
length.
Based on the above analysis, when chain segment structures were adopted in tunnel
engineering, the maximum stress, the distribution of the plastic zone and the
maximum equivalent plastic strain of the lining significantly reduced. With the
decrease of section length, the stress significantly reduced, and it become more
uniform. The plastic zone and the maximum equivalent plastic strain were also
significantly reduced. When the lining segment length was 5m, the tunnel could adapt
to the dislocation of the better than that of the lining segment length was more than
5m.The tunnel which segment length was 5m give full play to the lining of carrying
capacity and had good resistance. If the segment length is too short the project costs
will increase and it will cause no necessary waste. Also it will increase the
construction process and increase the difficulty of construction control when the
lining segment length is too short. Therefore the lining segment length of chain
segment structures tunnel was recommended 5m.
References
[1] WangWei, Ren Qingwen(2006), Research on the effects of active faults of deep buried tunnel
7
outline. Earthquake engineering and engineering vibration, 26 (1), pp.175-180
[2] R.John Caulfield, D.Scott Kieffer, David F. Tsztoo, Bill Cain (2005), Seismic design measures for
the retrofit of the Claremont. RETC PROCEEDINGS, Chapter 89
[3] D. Scoa Kieffe, R. John Caulfield, Bill Cain (2001), Seismic upgrades of the Claremont tunnel.
RETC PROCEEDINGS, Chapter 68
[4] Russo M, Germani G Amberg W(2002), Design and construction of large tunnel through active
faults: a recent application, International Conference of Tunnelling &Underground Space Use,
pp.16-18
[5] Walter AMBERG, Marco RUSSO(2001), Seismic design of underground structures-the Bolu
Tunnel. AITES-ITA World Tunnel Congress, ISBN 88-555-2594-8
[6] Suleyman Dalglc7(2002), Tunneling in squeezing rock, the Bolu tunnel, Anatolian Motorway,
Turkey, Engineering Geology, 6, pp.73-96
[7] Wenhao Liang, Guoliang Li (2004), WuXiaoLing tunnel design, Modern tunnel technology, 4(2),
pp.1-7
[8] Feng Qimin, Shao Guangbiao(2004), Near-fault ground motion attenuation rule of velocity and
displacement peak research . Earthquake engineering and engineering vibration, 24 (4), pp.13-19
[9] Lipeng(2009), Active fault zone fault structure design of highway tunnel, ChongQing JiaoTong
University
[10] Sichuan Province Geological Engineering Investigation Institute(2011), MianMao road highway
engineering geological investigation report, pp.13-14
