1 INTRODUCTION
Large deformations of the tunnel lining may occur in tunnels excavated in squeezing rock under high overburden pressure. These deformations may continue with time and damage may be observed after a long period of tunnel construction (Wu et al., 2020). In fact, the deformation rate of squeezing mainly depends on geological conditions and the ratio of rock mass strength to in situ stress (Arora et al., 2021; Bian et al., 2019). Table 1 shows a list of squeezing cases reported in the literature and the related fundamental parameters of each tunnel.
Table 1. Example of squeezing cases and related fundamental parameters.
| Tunnel |
Depth, H (m) |
Equivalent diameter, D (m) |
σcm/p0 |
ε (%) |
References |
| Saint Martin La Porte tunnel, France |
310 |
5.5 |
0.23 |
18.0 |
Barla et al. (2008) |
| Tawarazaka tunnel, Japan |
260 |
5.2 |
1.30–2.20 |
3.0 |
Wang et al. (2021) |
| New Tienlun Headrace, China |
400 |
6.5 |
0.07–0.38 |
14.0 |
Chern et al. (1998) |
| Maan project, China |
200 |
6.0 |
0.14–0.33 |
3.7 |
| Fréjus road tunnel between France and Italy |
1067 |
8.6 |
0.35–2.50 |
0.9–1.8 |
De La Fuente et al. (2020) |
| Pinglin Tunnel, China |
150–200 |
4.8 |
0.19–0.32 |
3.3–4.2 |
Barla and Borgna (1999) |
| Yacambu-Quibor tunnel, western Venezuela |
600 |
5.5 |
0.06 |
>30.0 |
Arora et al., 2021 |
| Loktak tunnel, India |
300 |
4.8 |
0.10 |
7.0 |
Singh et al. (1992) |
| Maneri Bhali, India |
350 |
4.8 |
0.10 |
7.9 |
| Nathpa Jhakri tunnel, India |
300 |
10.0 |
0.25 |
20.0 |
Hoek (2000) |
| Chameliya hydroelectric headrace tunnel 3 + 296 m, Nepal |
252 |
5.4 |
0.43 |
12.5 |
Basnet (2013) |
| Chameliya hydroelectric headrace tunnel 3 + 404 m, Nepal |
284 |
5.4 |
0.35 |
36.7 |
| Maneri-Ultarkashi power tunnel, India |
900 |
4.8 |
0.10 |
8.9 |
Goel et al. (1995) |
Abbreviations: p0, in situ stress; ε, ratio of tunnel closure to tunnel diameter; σcm, global rock mass strength.
The problem of large deformations in a squeezing ground must be analyzed by considering the factors of rock strength reduction and creep properties in the long term (Agan, 2016; Wu et al., 2017). Fraldi and Guarracino (2010) used the Hoek–Brown criterion to propose a comprehensive analytical solution for the tunnel collapse mechanism, with special attention paid to the progressive failure evolution of circular tunnels. In addition, field investigations and measurements are implemented as common approaches to assess different modes of squeezing failure, such as invert heaving, wall convergence, and cracking of support (De La Fuente, 2018; Iasiello et al., 2021). Furthermore, Zhang et al. (2020) used seven machine learning classifiers with weighted voting methods to suggest a robust classifier that can predict the squeezing phenomenon in rock tunnels. Under such risk of squeezing-induced failure, understanding the squeezing potential and designing the appropriate support system can be challenging in tunnel engineering (Aksoy et al., 2012; Bo et al., 2023; Liu et al., 2022). Many methodologies and solutions have been proposed to improve the convergence–confinement method (CCM) and consequently calculate the ground response and the required support stiffness so as to ensure tunnel stability in a squeezing ground (Kitagawa et al., 1991; Oke et al., 2018; Sakai & Schubert, 2019). Through a series of load tests on full-scale segmental linings, it is shown that the support of steel ribs or steel plates could increase the ultimate loading capacity by approximately 30%–40% (Qiu et al., 2018; Zheng et al., 2022). Zhang et al. (2017) used CCM to suggest four quick remedial measures of backfilling, grouting, reinforcement, and drainage. However, due to the adverse geological conditions and regional differences, the reasons for and scale of tunnel failure are clearly different from one tunnel to another. Therefore, it is impractical to suggest countermeasures systematically. Squeezing failure in operational tunnels necessitates further study to analyze the deformation mechanism and provide valuable engineering experience in tunnel collapse and countermeasure evaluation. Accordingly, Tishreen tunnel in the west of Syria was selected as a case study to analyze the squeezing potential based on empirical approaches. The efficiency of numerical modeling, based on the Burger–creep constitutive model, was then discussed and verified. On this basis, appropriate support measures were proposed and implemented in the field, and their effectiveness in maintaining the operationality of the tunnel under study was proven.
2 INFORMATION ON THE DISRUPTED TUNNEL SITE
2.1 General context and engineering status of Tishreen tunnel
Tishreen tunnel is a water conveyance tunnel, designed to provide 6 m3 discharge under free water surface conditions and pass the water from the reservoir of the hydro-engineering complex construction on El-Kebir River to irrigate lands of the coastal plain in the west of Syria. With a total length of 7295 m, the tunnel has been in use since April 1988. The tunnel passes underground at depths ranging from 14 m to more than 250 m. The thickness of the soil layers above the currently deformed area increases with the distance from the tunnel shaft and reaches approximately 150 m, as shown in Figure 1.
Tishreen tunnel: Schematic representation of the longitudinal and vertical profile with squeezing zone formation.
Two principal types of linings are adopted during construction. Lining I consists of a precast concrete ring, including six reinforced blocks; this lining is made by the shielding method of tunneling. Lining II is a mass-reinforced concrete with a horseshoe shape with an inner size of 300 cm × 310 cm and 30 cm thickness; it is made using the mining method and corresponds well to the deformed part of the tunnel. The original dimensions and deformed shape of the lining are shown in Figure 2.
Dimension of lining II where tunnel failure occurs. (a) Original dimensions and (b) deformed shape.
2.2 Geological profile and geotechnical properties
The region of the tunnel has a complicated geological profile characterized by intensive folded and cracked rock masses. Four distinguished geological profiles can be observed, as shown in Figure 3.
Geological profile of the tunnel alignment.
The first profile spreads up to St.115, consisting of stable rocks represented by diabase and porphyrite, where cracks and shattering zones are often welded by quartz, calcite, and feldspar. Further on (up to St.800), the second profile is composed of clays, rich in fragments of cataclasite, melonite, and other rocks. Spreading between St.800 and St.4800, the third profile includes a layer of clay shale and thinly interbedded radiolarite, among which there are bands of argillites and marls. The designer mentioned that the rocks here are susceptible to squeeze and bulge in the presence of water. Between St.4800 and the outlet portal, the rocks are gently dipped and consist of intercalations of marls, clayey, and cracked limestone. The rocks here are rather stable but easily yield to weathering. The failure problem occurred in the third geological profile (at a distance of 450 m from the tunnel shaft), where the tunnel is surrounded by marly clayey shale characterized by high squeezing ability and weak resistance to overburden load. A geotechnical survey of the surrounding rock was carried out to extract the required properties, as summarized in Table 2.
Table 2. Geotechnical properties of rock-surrounded tunnel in a deformed zone according to the specifications of (ASTM-D2487).
| Property |
Source |
Value |
| Density (g/cm3) |
|
2.13–2.30 |
| Uniaxial compressive strength of the intact rock, σci (kN/m2) |
|
850–940 |
| Elastic modulus of rock (MN/m2) |
|
125–175 |
| Angle of internal friction (°) |
|
13.9–17.8 |
| Cohesion of rock (kN/m2) |
|
127–152 |
| Rock mass rating, RMR (%) |
Bieniawski (1989) |
14–20 |
| Rock class number and classification |
V—very poor rock |
| Estimated average stand up time (min/m) |
10 h for 2.5 m span |
| Geological strength index
|
Hoek and Brown (1997) |
10 |
| Rock mass quality, Q |
Barton (2002) |
0.021–0.037 |
2.3 Distortion of the tunnel
During the irrigation season of the summer of 2021, it was found that the outflow from the tunnel was much less than the inflow. Careful inspection of the tunnel and measurement of the resulting deformation showed that there were different types of distortions represented by the following aspects (Figure
4):
1.
Tunnel floor heave, reaching the state of slab crashing. It constituted a barrier of up to 75 cm height in front of inrushing water. This observation was accompanied by the rising and buckling of the rail fixed at the top surface of the tunnel floor.
2.
Major disaster of large deformation of the side wall and floor heave. The deformation of the side wall was up to 35 cm in some cases, and the maximum floor heave was nearly 75 cm, as shown in Figure 2. The horizontal convergence of the side wall was far greater than that of the spring line and arc spandrel. The floor heave had a large amount of vertical movement, which was more prominent than the displacement of the crown.
3.
Fissured lining. Field investigation and core tests revealed that the lining had only 30–35 cm thickness, which was actually less than the designed minimum thickness (40 cm).
Different failure modes observed in Tishreen tunnel: (a) Floor heave; (b) wall spalling; (c) crashing of slab; and (d) lining convergence reaching failure limit.
3 ASSESSMENT OF SQUEEZING PHENOMENA
It is noticed that the lining of Tishreen tunnel (excavated in clay shale in the damaged area) undergoes large plastic deformation, and the tunnel keeps on substantially converging over 25 years, which meets the definition of squeezing ground condition proposed by Barla and Borgna (1999). Various approaches have been established to predict the potential of squeezing in tunnels. These approaches can be classified into three categories: (1) empirical approaches; (2) semi-analytical approaches; and (3) numerical modeling approaches.
3.1 Empirical approaches
Based on a large number of observations, empirical approaches define squeezing according to predominant parameters and simplified engineering judgment:
3.1.1 Singh et al. (1992)
This empirical approach is established according to 39 case studies. The rock mass quality index
Q-System is used to identify the potential of squeezing as follows:
(1)
where
H denotes the overburden thickness in meters. The
Q index is given by the Barton relationship (Barton et al.,
1974):
(2)
If the overburden thickness (H) is greater than 350 × Q1/3, the tunnel is subjected to a squeezing condition, given the properties and the classification of surrounding rock (RQD = 20; Jr = 1 [small and planar joint roughness]; Jn = 15 [heavily jointed]; Ja = 4 [softening or low-friction mineral coatings, swelling clay]; Jw = 0.66 [medium inflow or pressure]; Srf = 10 [very loose rock, weakness zones containing clay]). By substituting the resulting value, Q = 0.022 is obtained.
3.1.2 Goel et al. (1995)
Goel et al. (
1995) used the data collected from 99 tunnel sections to propose a simple empirical equation, which is mainly related to the rock mass number (
N) by using the stress reduction factor (
Srf = 1) in Equation (
2). They proposed different degrees of squeezing according to three parameters (
H,
N, and
B), where
B is the tunnel width. However, for the squeezing phenomenon to occur, Equation (
3) must be satisfied:
(3)
3.2 Semi-analytical approaches
These methods provide indicators for prediction of squeezing conditions by using closed-form analytical solutions for deep tunnels in a hydrostatic stress field. The competency factor is generally used as a common starting point to evaluate the squeezing potential of rock. This factor represents the ratio of the uniaxial compressive strength of rock mass to overburden stress (Azizi et al., 2018). The most popular semi-analytical approaches are proposed by (Hoek & Marinos, 2000; Jethwa et al., 1984).
3.2.1 Jethwa et al. (1984)
Jethwa et al. (
1984) used the competency factor to determine the degree of squeezing as defined by Equation (
4):
(4)
where
γ denotes the rock density. If
N
c < 2, the squeezing condition may affect the tunnel with different degrees of deformation.
3.2.2 Hoek and Marinos method
Moreover, using finite element analysis (FEA) results and various numerical case studies, Hoek and Marinos (
2000) and Hoek and Guevara (
2009) proposed a criterion based on the tunnel strain (
ε
t) and described it in Equation (
5) as follows:
(5)
Using Equation (4), the curve given in Figure 5 can be used to conduct a preliminary estimate of tunnel squeezing problems and provide a set of approximate guidelines on the degree of difficulty that can be encountered for different levels of strain.
Tunneling problems associated with different levels of strain without support in weak ground according to the results obtained by Hoek and Marinos (
2000).
Tishreen tunnel passes at an average depth of 150 m, which means that the in situ stress level will be p0 = 3.24 MPa. The global rock mass strength (σcm) can be estimated as 0.05σci according to the findings of Hoek and Marinos (2000), and thus, the ratio σcm/p0 = 0.1, corresponding to the radial convergence of
for our case without support (Figure 5). This means that the tunnel designer should therefore anticipate having to deal with very severe squeezing problems in this section. The results of squeezing potentials are summarized in Table 3, according to different approaches.
Table 3. Evaluation of the squeezing potential according to different criteria for the case of Tishreen Tunnel.
| Equation no. and reference |
Criterion |
Related value Tishreen tunnel |
Behavior |
| (1) Singh et al. (1992) |
|
|
Squeezing condition |
| (3) Goel et al. (1995) |
|
|
Very mild squeezing |
| (4) Jethwa et al. (1984) |
|
Nc = 0.1 |
High squeezing |
| (5) Hoek and Marinos (2000) |
|
|
Extreme severe squeezing problem |
3.3 Numerical modeling approaches
Besides the traditional empirical and semi-analytical approaches in predicting the squeezing potential, many numerical simulations have been conducted to explore the stability of tunnel linings subjected to squeezing conditions. The determination of lining stability can be successfully achieved by using three-dimensional stress conditions and a time-dependent constitutive model. The choice of failure criteria plays a crucial role in squeezing evaluation or tunnel support design (Do et al., 2014; Kabwe et al., 2020; Sun et al., 2018).
4 NUMERICAL MODEL
4.1 Geometry and boundary condition
The numerical model is developed using the finite difference scheme (FLAC3D program). Half of the domain is considered due to the symmetry of both the loading and the geometry. The vertical plane of symmetry through the tunnel center is assumed as shown in Figure 6. The crown of the tunnel is located at 150 m below the ground surface. The origin of coordinate axes is defined at the center of the tunnel; the z-axis points upward and the y-axis points along the axis of the tunnel. The size of the model is 50 m × 100 m × 200 m (length/width/height).
Three-dimensional numerical mesh of the tunnel lining and surrounding rock (80 733 elements and 88 244 nodes).
The mesh is refined in the vicinity of the tunnel lining and increases gradually when approaching boundaries along the x- and z-axis. The mechanical boundary corresponds to roller boundaries along the symmetry line and the far boundary of the model (x-direction), roller boundaries on both sides of the plane of analysis (y-direction), and fixed displacements in three directions at the model base. For the initial condition, the in situ stress state is initially applied in the total domain (De La Fuente, 2018). The three principal stresses are considered to be equal and the stress ratio K0 is set equal to 1.
4.2 Constitutive model
A time-dependent constitutive model needs to be used to obtain an accurate description of tunnel response associated with squeezing conditions (Do et al., 2021; Paraskevopoulou & Diederichs, 2018). The significant viscoelasticity under creep conditions should not be neglected. Both the attenuated and steady creep stages are critical to the rheological characterization of such weak surrounding rock. Several rheological models have been proposed to simulate the time-dependent characteristics of rock mass. It is clearly noted that weak surrounding rock shows viscoelastic deformation characteristics, which enables the use of the Kelvin, Kelvin–Voigt, Maxwell, and Burger's models as alternatives to represent the viscoelastic rheological characteristics of the weak surrounding rock (Liu et al., 2022; Sun & Pan, 2012; Zaheri et al., 2023). Therefore, in the numerical simulations of Tishreen tunnel, the assumed constitutive behavior for the ground is a creep viscoplastic model (CVISC). This model simulates the visco-elasto-plastic deviatoric behavior by using a Burger's element, which is comprised of a Maxwell model and a Kelvin model connected in series, as shown in Figure 7.
Schematic of Burger's model, time-related strains (Liu et al.,
2022).
The model describes correctly both delayed and instantaneous deviatoric strains (De La Fuente et al.,
2020). The CVISC model is incorporated into FLAC
3D and has been previously adopted in many numerical modeling studies (e.g., Abdolyousefi et al.,
2022; Barla et al.,
2010; Gao et al.,
2016; Kabwe et al.,
2020). In Burger's model, the creep processes are governed by the deviatoric stress state and the axial strain solution for a constant load
P is expressed as
(6)
where the superscripts K and M are Kelvin and Maxwell properties, respectively;
K denotes the bulk modulus;
G is the shear modulus;
η is the viscosity;
α =
G
K
/η
K; and
t represents the time (Paraskevopoulou & Diederichs,
2018). For numerical accuracy, the maximum numerical timestep can be expressed as the ratio of the ground viscosity to the shear modulus (Itasca,
2005). For the Burger–creep viscoplastic model, this could be interpreted as (Equation
7):
(7)
A numerical prediction of the average long-term response (25 years) of Tishreen tunnel has been conducted. For this purpose, the constitutive parameters of marly clay shale layers used in the creep model are listed in Table 4. These parameters are based on a back-analysis of convergence measurements proposed by Zhifa et al. (2021) and De La Fuente et al. (2020). For the lining segments, a constant Young's modulus of 12 GPa and Poisson ratio of 0.15 are chosen according to the long-term behavior of the concrete lining. The numerical analysis is carried out for different durations of squeezing ranging up to 25 years.
Table 4. Constitutive parameters for the CVISC model.
4.3 Numerical results
The results shown in Figure 8 indicate that the squeezing-induced stress behind the lining has a huge influence on the stress state of the tunnel lining. The vertical stress in the middle of floor moves upward as a result of the squeezing behavior of weak rock. The value of tension stress increases up to 6 MPa after 25 years of construction and this would lead to the development of tension cracks in the tunnel lining.
Compression and tension stress in the tunnel lining after 25 years of construction.
The presence of lining damage leads to a decrease of load-carrying capacity with an increase of the surrounding pressures. Consequently, the bearing capacity of the damaged lining would decrease again until the lining is destroyed.
Moreover, the results for creep displacement from numerical modeling are illustrated in Figure 9. It can be concluded that the upward displacement of the tunnel floor is greater than crown displacement. The plot of displacement magnitude for several periods shows clearly that the resulting displacement, at different points of the tunnel lining (arc spandrel, abutment, and spring line), is greater than the allowable strain, 1% of the tunnel diameter (30 mm). The amount of strain shows that squeezing conditions exist and lead to instability in the tunnel lining with time. Consequently, the squeezing phenomena may cause failures in the support system. It is worth noting that the deformation direction of the vault area changes from converging toward the tunnel center to squeezing outward to the surrounding rock; meanwhile, the deformation direction of the spring line and side wall behaves in the opposite way. Significant failure occurs in the tunnel floor and side wall segment, which is in agreement with the results of (Asghar et al., 2017; Wang et al., 2019).
Magnitude of displacement in the tunnel lining from a numerical model after a period of: (a) 1 month; (b) 1 year; (c) 3 years; (d) 10 years; (e) 15 years; and (f) 20 years.
Numerical analysis also reveals that side wall displacement reaches a value of (28 cm), which is more pronounced than the displacement of arc spandrel and spring line. The heave of the floor slab exceeds the value of 65 cm, while that of the crown is limited to 25 cm approximately as shown in Figure 10. The resulting values of the displacement magnitude and deformation mode are found to be in good agreement with the field observations, which are shown in Figure 2.
Curves of (a) vertical and (b) horizontal displacement in the tunnel lining.
5 REPAIRING METHODOLOGY AND SUPPORTING MEASURES
5.1 Estimation of support pressure
Empirical estimation of support pressure is generally used as a straightforward method that enables the engineer to have a general perspective on the required tunnel support pressure (Taghizadeh et al., 2020).
Hoek and Marinos (
2000), using FEA and various numerical case studies, proposed criteria to determine support pressure based on the tunnel strain (
ε
t) as follows:
(8)
where
p
i denotes the support pressure;
σ
cm/
p
0 is the strength to in situ stress. Although this equation is generally used for a circular tunnel, Chern et al. (
1998) observed that the computed deformations are acceptably accurate for application to actual tunnels.
Moreover, using the data obtained from 53 tunnel cases at various project sites, Dwivedi et al. (
2014) proposed a dimensional expression that correlates the variations of the observed ultimate support pressure as a function of
,
, and
d as
and
. This correlation is mainly based on joint factor
, which has been defined as
. Specifically,
J
n denotes the joint set number;
n is the joint orientation parameter depending upon the orientation of the joint with respect to loading direction; and
r represents the joint strength parameter, which is related to the joint thickness, condition, and alteration of joints due to weathering. The joint factor expresses the quality reduction in the intact rock, so that the higher the rock strength, the lower the joint factor. Accordingly, the values of estimated support pressure are calculated as shown in Equation (
9)
(9)
where
is the ultimate support pressure (MPa);
is the vertical in situ stress (0.027
H) (MPa);
is the uniaxial compressive strength of intact rock (MPa);
is the horizontal in situ stress (MPa), and
δ is the radial tunnel deformation (%).
Figure 11 shows a significant difference between the value of the support pressure predicted by Hoek and Marinos (2000) and that of Dwivedi et al. (2014).
Support pressure versus radial convergence in weak ground tunnels, according to the results obtained by Hoek and Marinos (
2000) and Dwivedi et al. (
2014).
It should be noted that Equation (8) proposed by Hoek and Marinos represents lower bound conditions assuming zero dilation of the rock mass. In addition, for a given set of rock mass properties and in situ stresses, the two-dimensional closed-form solutions can predict larger displacements than the corresponding axisymmetric FEAs. Squeezing conditions can lead to much higher displacement of the tunnel support system than anticipated. In fact, it is found that the ultimate support requirement can be up to three times higher than that calculated for the ordinary tunnel convergence (Jethwa et al., 1984) and a suitable factor of safety must be adopted. Therefore, by substituting the value of different factors into Equation (9), the solution proposed by Dwivedi et al. (2014) yields a support pressure of ps = 3.4 MPa. Accordingly, a value of 6 MPa is adopted as a safe support pressure required to assure tunnel safety after squeezing.
5.2 Methodology for tunnel support selection
The effect of stiff and flexible supports on support pressure and convergence has been previously evaluated by many researchers. However, Dalgic (2002) studied the squeezing deformation and the related appropriate support of Bolu tunnel in Turkey. It was concluded that a flexible support system allows large deformations to occur and a part of the tunnel may collapse partially. This finding also aligns with the recommendations of Xu et al. (2021) that it is necessary to obtain a sufficiently large primary support pressure at the first stage of tunnel construction and that the allowable displacement rate should not be very high. Furthermore, it has been proven that stiff support (including steel ribs, rock bolts, wire mesh, and shotcrete lining) can reduce appreciably the displacement and plastic deformation surrounding the tunnel in the hydroelectric project in Sikkim, India (Panda et al., 2014). Hence, the application of the flexible support system has been abandoned later. Instead, more heavy support systems are used, and deformations are under control as a result.
The preliminary analysis of tunnel deformation and required support is conducted in the sequence presented in Table 5. The key element for designing the rigid permanent tunnel lining is to maintain the radial displacements at less than 1% of the tunnel diameter, taking into account the ratio of rock mass strength to in situ stress.
Table 5. Support estimation sequences.
| Depth below surface (m) |
Vertical in situ stress, p0 (MPa) |
Strength to stress ratio, σcm/p0 |
Strain limit, ε (%) |
Support pressure to in situ stress, pi/p0 |
Required safe support |
| 150 |
4.05 |
0.075 |
1 |
0.5 |
4 |
For the 150 m deep tunnel, the support pressures of 6 MPa for 1% strain are difficult to achieve without resorting to very heavy steel ribs embedded in shotcrete or cast-in concrete. Hoek (
1999) published equations for calculating the stiffness and capacity of different support systems, which have been used to estimate the support characteristics. The replacement of the whole section is inevitable in the case of deformation exceeding the limit, and the dimension of the existing section does not allow the desired service flow of the water tunnel. Hence, two control schemes are jointly developed:
1.
Improving the self-supporting capacity of surrounding rock and reducing the plastic radius. Reinforcement measures, such as grouting of the surrounding rock and anchor bolting in the floor slab, are adopted to improve the surrounding rock.
2.
Optimizing the support system that deals with the deformation characteristics of the surrounding rock. This can be considered from the aspects of support stiffness and allowable deformation range. As a result, the following countermeasures are adopted to ensure the stability of Tishreen tunnel in the long term: (1) radial grout injection; (2) pre-strengthening anchor bolt in the ground floor; (3) steel ribs of IPE200. The steel support is inserted into the primary shotcrete lining; (4) two layers of reinforced concrete: the primary one is sprayed as shotcrete to cover the steel ribs. In this way, the primary lining is capable of providing considerable rock support (so-called lining resistance) and the secondary lining is cast in place with adequate reinforcement. Table 6 shows more details and descriptions about the proposed countermeasures, supported by field photos of different rehabilitation works as shown in Figure 12.
Table 6. Repairing measures and parameters.
| Countermeasure |
Related parameters |
| Grout injection |
Grout mixing consists of two materials: Cement slurry (with a water-to-cement ratio of about 0.5–0.6, and sodium silicate). The mixture is injected into the surrounding area through a grouting tube with eight nozzles and a length of 5 m. The grouting pressure is controlled after many trials at a fixed value of 200 kPa |
| Anchor bolts |
Rock bolts or ground anchors are often used to protect the floor by friction resistance against squeezing |
| Anchor bolts in the tunnel slab are therefore added with 4 m depth in two parallel lines; the parameters of the anchor bolts are a bar whorl-steel bolt with a diameter of 25 mm at distances of 100 mm from each other |
| Steel ribs |
Steel rib profiles of IPE200 are adopted, with a spacing of 700 mm; longitudinal connections of steel bars are installed immediately to fix the arch frame. The steel rib is closed and fixed into the tunnel ground slab |
| Rigid concrete lining support |
Shotcrete first lining |
The first segment support uses 25-cm-thick C25 reinforced shotcrete with synthetic polypropylene fiber |
| Cast in place second lining |
The second support uses 30 cm thick C30 cast-in concrete, with a steel reinforcement mesh of 20 mm bars at 10 cm spacing in the transverse direction and 16 mm bars at 15 cm spacing in the longitudinal direction |
Rehabilitation steps of Tishreen tunnel: (a) Breaking down and removing debris; (b) installing steel ribs and anchor bolts; (c) installing steel rebars and grout pipes; and (d) shotcrete application.
After the countermeasures were adopted, the amount of maximum displacement in many sections and directions was verified and evaluated. Monitoring is ongoing at the time of writing this paper, and the initial measurement indicates stability of the tunnel lining. Specifically, the maximum horizontal convergence and vault subsidence are 1.0 and 0.1 mm, respectively, after 3 months of installation of a new rigid support lining. Consequently, the achieved results prove that the proposed countermeasures can notably decrease the displacement values as well as the rate of deformation.
6 CONCLUSION
This study aimed at exploring the causes of failure and related effective measures of rehabilitation for Tishreen tunnel. Field measurements and empirical solutions clearly demonstrated the occurrence of large time-dependent displacement. Moreover, the time-dependent behavior of the rock mass of the tunnel was studied by means of numerical modeling. The constitutive model of the rock mass used in this study is the Burger–creep viscoplastic model, which simulates the creep processes in the long term. The calibration of numerical results was verified to properly fit the field observations. The countermeasures were also proven to be valid on the basis of the in situ tunnel monitoring. According to the research, the most important results are as follows:
1.
In designing tunnels, time-dependent characteristics in weak rock masses, especially for important tunnels with horseshoe shapes, should be considered.
2.
Numerical modeling results indicate that the concrete lining can suffer from excessive convergence and floor heave (instability caused by squeezing) due to the increase in rock pressure in the long term. The concentration of deformation is easily observed in the middle of the tunnel floor and side wall.
3.
The proposed model leads to reasonable long-term predictions that are consistent with the observed field displacement.
4.
Results also show that due to squeezing conditions, a heavy support system of Tishreen tunnel is required. Therefore, the lining resistance needs to be improved, taking into account the ground response curve and the support pressure. The success of the proposed solution can provide a meaningful reference for similar projects in the future, with emphasis on the necessity to continue monitoring for a long period.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest.
Biography
Mohannad Mhanna is currently an assistant professor in the Faculty of Civil Engineering at Tishreen University in Syria. He is also the Director of the Soil Mechanic Laboratory and invited professor at Manara Private University. Prior to his recent appointment at Tishreen University, he was a Research project coordinator in IRT—Railenium in France. Dr. Mhanna received his Master's degree as well as his PhD in Civil Engineering from Lille 1 University - Science and Technology in France. Dr. Mhanna has published a number of papers in preferred journals and chapters in books and presented various academic as well as research-based papers at several national and international conferences. His research activities focus on numerical modeling in geotechnical engineering, especially traffic-induced vibrations and stability of tunnels and slopes.
