- Email: [email protected]
Case studies of groundwater flow into tunnels and an innovative water-gathering system for water drainage
Tunnelling and Underground Space Technology 24 (2009) 260–268
Contents lists available at ScienceDirect
Tunnelling and Underground Space Technology journal homepage: www.elsevier.com/locate/tust
Case studies of groundwater flow into tunnels and an innovative water-gathering system for water drainage Diyuan Li a,b,*, Xibing Li a, Charlie C. Li b, Bingren Huang a, Fengqiang Gong a, Wei Zhang a a b
School of Resources and Safety Engineering, Central South University, Changsha, Hunan 410083, China Department of Geology and Mineral Resources Engineering, Norwegian University of Science and Technology, Trondheim, Norway
a r t i c l e
i n f o
Article history: Received 20 May 2008 Received in revised form 20 August 2008 Accepted 27 August 2008 Available online 11 October 2008 Keywords: Case study Numerical modeling Groundwater flow Tunnel Leakage Water-gathering system
a b s t r a c t Groundwater inflow into tunnels can constitute a potential hazard and also is an important factor influencing the speed of tunnel excavation. In this paper the results of numerical modelling are presented to investigate the groundwater flow and the distribution of the pore pressure around tunnels. Two types of tunnels, double-arch tunnel and twin-tube tunnel, were studied. Potential leakage places are identified for the two types of tunnels. The most permeable place in the double-arch tunnel is at the contact interface between the middle wall and the overlying rock. The results of numerical modelling are compared with field observations in the case studies. Based on the results of numerical modelling and the field investigations, an innovative water-gathering system for reducing water leakage was proposed and applied in some tunnels on ChangJi Expressway in China. The water-gathering system can be quickly glued to the rock surface and easily installed for tunnelling. It can be applied in tunnels where waterbearing fractures are well-developed in the rock mass. Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
2. Background
Groundwater inflow into tunnels can constitute a potential hazard and also is an important factor influencing the speed of excavation advancement. Nowadays more and more road tunnels of complex opening shapes are constructed in mountainous areas all over the world. The groundwater flows in such special opening shapes are complicated and different from those of simple opening shapes. Groundwater inflow and leakage problems are difficult to manage in tunnels. Based on site investigations and field water pressure tests, two cases of numerical study about groundwater flow were carried out using code FLAC3D in this paper. The distributions of seepage field around double-arch tunnel and twin-tube tunnel were obtained. Potential permeable places were identified based on the results of the numerical modellings. The modelling results from case studies were compared with field observations at Bimaxi tunnel and Qingshangang tunnel of the ChangJi Expressway in Hunan Province, China. According to the modelling results and site investigations, an innovative watergathering system for water drainage was put forward and applied in fields. The proposed method for reducing water leakage in tunnels is easy to carry out and effective in reducing groundwater flow into tunnels.
Groundwater inflow is an old but common and difficult problem in tunnelling project. It may retard the speed of excavation under tunnel construction, while on the other hand it may lead a potential hazard to the stability of the tunnel in the long-term run. One of the practical difficulties in tunnel construction and maintenance is related to groundwater inflow and leakage. On one hand, some of the most disastrous events in tunnelling are associated with groundwater inflow of large volumes in highly fractured water-saturated rocks. For instance, in the Pinglin tunnels of the Taipei-Ilan Expressway Project, the sudden groundwater inflow was up to 750 l/s (Tseng et al., 2001). It leaded to a collapse at the tunnel heading and the TBM in the pilot tunnel was trapped and damaged. Therefore, it is essential to have a good understanding on the water inflow rule in the rock mass under tunnel construction, so that appropriate changes can be taken in tunnel routes and drainage systems. On the other hand, water leakage exists in most of the finished tunnels. It is very difficult to realize a totally watertight tunnel. Yuan et al. (2000) gave a review of the practices in tunnel waterproofing in China. It is pointed out that it is hard to achieve a completely watertight tunnel using one waterproofing method alone. Consequently, multiple waterproofing measures must be integrated in underground works. For instance, membranes and watertight concrete were used together to prevent water from entering the tunnel space for immersed tunnels (Grantz et al., 1997).
* Corresponding author. Tel.: +86 731 8877276. E-mail address: [email protected] (D. Li). 0886-7798/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.tust.2008.08.006
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
Groundwater inflow depends on a number of factors, such as the permeability of the rock mass, the groundwater table, the aperture of rock fractures, and the size of the excavation. Although it is very difficult to accurately predict the water inflow during tunnel construction, a large number of researches have been done to tackle the problem, particularly through numerical simulations in aiming at rational and visible solutions. There are several analytical expressions in the literature to predict discharges into tunnels. In a homogeneous infinite water table aquifer, Goodman (1965) proposed an expression for the flow rate per unit tunnel length, q,
q¼
2pKH0 lnð2H0 =rÞ
ð1Þ
where K is the hydraulic conductivity, r is the radius of the tunnel and H0 is the depth of the tunnel below the water table. Eq. (1) is widely used in tunnelling owing to its simplicity and good accuracy of prediction. Zhang and Franklin (1993) proposed another analytical solution to predict the water inflow with a gradient hydraulic conductivity on the basis of the solution by Goodman. Hwang and Lu (2007) proposed a semi-analytical approach for water inflow. For the constant flow and variable drawdown problem of sources and sinks, the governing equation is written as:
Sðr; tÞ ¼ Q FðP; r; tÞ
ð2Þ
where S is the drawdown of water level, Q is the inflow, F is a function of the drawdown, P is a parameter for all relevant hydrogeological factors (including permeability, coefficient of storage, thickness of aquifer and distance to the boundary), r is the distance to the location of water inflow and t is the time since the beginning of water inflow. The formulae were developed for a tunnel intersecting an inclined aquifer, which is the most common case of tunnel inflow. Analytical solutions rely on given hydro-geological assumptions with simple circular or rectangular openings. In complex hydrogeological conditions such as fractured rocks, numerical models can be adopted to study groundwater flow around tunnels. A numerical flow model is a simplified representation of an aquifer which aims at capturing the most relevant features of groundwater flow. Construction of a numerical model requires formulating first a conceptual model, which is a qualitative description of the main properties of the flow system. On water flow through fractured media, several conceptual models have been proposed so far, including (1) equivalent porous media (EPM) models, (2) discrete fracture network (DFN) models and (3) hybrid models. Berkowitz (2002) gave a detailed review on these conceptual models and Samardzioska and Popov (2005) compared three different models for modelling of flow and solute transport in fractured porous media, in terms of their predictions of the flow and solute transport field variables. The three models can be applied to different conditions. It is claimed that the EPM models are efficient and faster to calculate than other types of models to simulate the fluid flow in the fractured media. Some numerical modelling work for groundwater inflow can be found in literatures (for instance, Meiri, 1985; Molinero et al., 2002; Shin et al., 2002; Or et al., 2005). For example, Meiri (1985) used the finite element method for nonsteady groundwater flows with a free surface and compared the numerical results with those obtained from analytical approximations. Shin et al. (2002) considered the effect of groundwater movement in the long-term. The lining permeability was varied from fully permeable to impermeable, and an approach to model finite lining permeability was presented and assessed. Or et al. (2005) presented a simple and physically based model for estimation of the seepage into tunnels excavated in unsaturated fractured rocks. Most of these studies, however, concentrate on the prediction of groundwater inflow into tunnels and the tunnel shapes are always quite simple.
261
It is a big issue on how to deal with the groundwater when a tunnel is constructed below the groundwater table. To drain or to seal? Kolymbas (2005) pointed out that sealed tunnels do not influence the groundwater but their lining has to support the full water pressure. Drainage can influence the surrounding groundwater considerably but can relieve the lining from hydrostatic pressure. Actually tunnels below the groundwater table can be either sealed or drained. Some useful methods have been put forward by Caputo and Huez (1987), Nilsen (2001), Kleven (2004), Kolymbas (2005) and Aksoy (2008). Geosynthetics are applied for drainage (geospacers and geocomposite drains) and for waterproofing (geomembranes). Grouting and chemical injection are used for improving ground conditions and waterproofing. However, water leakage phenomena do exist in tunnels no matter they are drained or sealed by traditional ways. This problem has not been well-resolved. 3. Description of the problem A great number of expressways have been constructed for the last two decades in China. One of them is the expressway from Changde to Jishou, called ChangJi Expressway, in southwest of China. The construction of this expressway began in 2004. Many tunnels have to be constructed in the course of the expressway. The total length of the ChangJi Expressway is 224 km, more than 40 km of which are tunnels. Most of the short tunnels (shorter than 500 m) are constructed as double-arch tunnels, while long tunnels (longer than 1000 m) are constructed as largely-spaced twin-tubes. A double-arch tunnel is characterized by its concrete wall constructed in the middle of the tunnel. Largely-spaced twin-tube tunnels are excavated separately so that they are not affected by each other in mechanics. The excavation method is the so-called NATM – the new Austrian tunnel method. Fig. 1 shows examples of these two different types of tunnels in the ChangJi Expressway. Water flows into tunnels can cause a lot of problems not only under construction but also in the long-term operation. The tunnels of the ChangJi Expressway are located in subtropical mountainous areas in Hunan Province. The climate there is characterized by the moist air and abundant rainfalls. The invasion of groundwater and surface water into tunnels would cause serious unstable problems in tunnels. Some previous work (Li et al., 2007, 2008) about the coupling effect can be proof to this conclusion. After a thorough investigation programme and field observations, it was found that there were four types of water inflows in tunnels of ChangJi Expressway: (1) Groundwater inrush: This type of water inflow would cause rock collapse in the tunnel face. This kind of collapse is a big problem for tunnel construction but it does not happen very often. It occurs only when abnormal geological structures like aquifer faults are exposed on the tunnel face. This type of incident happened twice in a twin-tube long tunnel– Qingshangang tunnel where rock collapses led to a big hole in the roof close to the tunnel face. (2) Groundwater inflow and water drops occurred at the shoulder of the tunnel. Most of the moderately long and short tunnels in ChangJi Expressway were designed to the double-arched structure. The cross-section of a double-arch tunnel is shown in Fig. 2. The total span (D) of the tunnel is about 25 m. The potential permeable place is marked in the figure and the water leakage photos are shown in Fig. 3a. (3) Groundwater inflow occurred in the position of ‘I’ steel sets, which were the major support elements of the tunnel. Because the groundwater was not well drained by the drainage pipes, the high water pressure resulted in seepage problem along ‘I’ steel. It can be seen in Fig. 3b.
262
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
Fig. 1. Two different types of tunnels on ChangJi Expressway: (a) Bimaxi double-arch tunnel; and (b) Qingshangang twin-tube tunnel.
Fig. 2. Cross-section of the double-arch tunnel (unit, cm).
Fig. 3. Water leakage phenomena in tunnels by site investigation: (a) water leakage at the contact interface between the middle wall and overlying rock in a double-arch tunnel; and (b) water leakage occurred at the ‘‘I” steel sets in a twin-tube tunnel.
263
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
(4) Groundwater inflow from the tunnel floor. The floor water worsened the work environment and may affect the construction of the tunnel floor. (5) The groundwater flow would build up water pressures at the back of the watertight lining in case of an unsatisfactory drainage. A good understanding on the distribution of the underground seepage field would be helpful for the design of water sealing and drainage.
9.81 m/s2 and; qw is the density of water, which equals to 1000 kg/ m3. For water at 20 °C, qwg = 9.81 103 Pa/m, the following conversion can be derived to calculate k in SI units for water in FLAC3D:
k ðin SI unitsÞ kh ðin cm=sÞ 1:02 106
ð5Þ
The hydraulic parameters obtained from the in situ water pressure tests are shown in Table 1. 4.2. Case study 1: groundwater flow in a double-arch tunnel
4. Case studies of groundwater flow into tunnels 4.1. Simplifications and parameters of numerical modelling The numerical modelling of groundwater flow into tunnels has to adopt some assumptions and simplifications. On one hand, it is difficult to simulate the groundwater flow around tunnels because of the heterogeneity of the rock mass and the complexity of groundwater flow in the rock joints. On the other hand, empirical methods and waterproof guidelines for underground engineering will be always employed when one designs a tunnel below the groundwater table. However, they are usually conservative. In this paper, the code FLAC3D is adopted to model the groundwater flow into tunnels. FLAC3D models the fluid flow in a permeable medium. The permeability of a rock mass is mainly determined by the fractures in the mass rather than by the porosity of the intact rock. The fractures are so ubiquitous and random that it is hard to simulate them numerically. However, as mentioned above, from the macroscopic point of view, a fractured rock mass can be treated as an EPM. The assumption is valid as long as it is possible to define a representative elementary volume (REV) (Bear, 1993). This requires that the model domain should be much greater than the average fracture spacing. In the formation concerned in this study, there is not any fault in the near field of the tunnels. The rock mass has been simplified to an equivalent porous medium with groundwater flows obeying the Darcy law. To gain the permeability of the rock mass, small-scale in situ tests, called water pressure tests, were conducted in Bimaxi tunnel. It is assumed that the permeability is isotropic in the rock mass. The parameter of permeability can be calculated by (Zhang, 2005)
Q al kh ¼ 0:528 log h0 l r0
ð3Þ
where the constant 0.528 is dimensionless, Q denotes the measured volume flow rate, in m3/s; h0 is the piezometer head at the center of the test section, in m; l is the length of the test section, in m; a is a coefficient, which is set to 0.66 when the test section is far from the impermeable layer, otherwise 1.32; r0 is the radius of the borehole for water pressure test, in m; and kh is the permeability of the rock mass, in m/s. Through the water pressure tests, the permeability of the surrounding rock mass was obtained in the range of 103– 105 cm/s. In FLAC3D the conversion between permeability coefficient k and kh is (FLAC3D, 2005):
kh ¼ kg q
ð4Þ
w
where k is the isotropic permeability coefficient used in FLAC3D, m2/(Pa s) in SI units; g is the gravitational acceleration, which is
The Bimaxi double-arch tunnel is selected as the case for numerical modelling to analyze the groundwater seepage field in the rock mass, and is located in a hill near the city of Jishou. The tunnel is about 500 m long. In the stretch of the tunnel there are a few gullies which primarily strike north with some striking northeast. The tunnel axis is nearly perpendicular to the strike of the hill in the section of Changde part, but becomes oblique or parallel to the hill strike in the section of Jishou part. The hill slopes are steep and form a V-shape gully. The overall strike of the hill is about 340° and the slope angle is about 15–35°. The vegetation is mainly small bamboo bushes and herbaceous plants. Rock outcrop is visible in some places. In the tunnel stretch there is no large fracture structures and faults. Three sets of rock joints exist: set 1 with a dip direction of 148° and a dip angle of 89°; set 2 with a dip direction of 350° and a dip angle of 56°; set 3 with a dip direction of 225° and a dip angle of 77°. The joint spacing varies from 5 to 20 cm and the connectivity of the joint aperture is fairly good. The meshes of 3D numerical model of the double-arch tunnel are shown in Fig. 4. The X-axis and the Y-axis are in the horizontal plane and the Z-axis is vertical. The model size is 125 m 15 m 80 m (X Y Z). From the top to the bottom of the model, it is applied different material properties simulating the sandstones in different weathering degrees and it is also applied different hydraulic parameters. In FLAC3D, boundaries are impermeable by default and saturation cannot be imposed as a boundary condition. A leaky boundary condition has the form of (FLAC3D, 2005):
q
n
¼ hðp p
e
Þ
ð6Þ
where qn is the component of the specific discharge normal to the boundary in the direction of the exterior normal, h is the leakage coefficient in (m3/N s), p is the pore pressure at the boundary surface, and pe is the pore pressure in the region to or from which leakage is assumed to occur. Before tunnel excavation the pore pressure below the groundwater table is hydrostatic. After excavation, the boundary of the tunnel is modelled by a free water seepage boundary where the adjacent underground water infiltrates into the tunnel. The concrete wall in the middle is modelled as a fluid null group in the code as it is a relatively impermeable. The seepage flow field in the rock mass has been changed after excavation. The mechanical response of the excavation is not considered in the numerical analysis. Fig. 5a shows the contour diagram of the pore pressure distributions at a vertical section (y = 6 m) after excavation but before lining. Fig. 5b shows the groundwater flow vectors in the near field of the double-arch tunnel.
Table 1 Hydraulic parameters of the surrounding rock mass in numerical modelling by FLAC3D Tunnel Bimaxi tunnel
Qingshangang tunnel
Weathering Heavily weathered Moderately weathered Slightly weathered Moderately weathered
Permeability kh (cm/s) 3
1.21 10 4.61 104 3.00 105 5.02 104
Porosity ratio n
Permeability coefficient in FLAC3D k (m2/Pa s)
0.45 0.40 0.25 0.30
1.23 109 4.70 1010 3.06 1011 5.12 1010
264
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
Fig. 4. A numerical model of the double-arch tunnel in FLAC3D.
-10 -20
z/m
-30 -40 -50 -60 -70
-40
-30
-20
-10
0
10
20
30
40
50
60
70
x/m
a. Contours of the pore pressure (unit:MPa)
b. Groundwater flow vectors
Fig. 5. Distribution of the pore pressure and groundwater flow vectors for a double-arch tunnel: (a) contours of the pore pressure (unit, MPa); and (b) groundwater flow vectors.
It can be seen that the groundwater pressure in the zones near the tunnel decreases after tunnel excavation. The groundwater begins to flow into the tunnel as soon as it is excavated. For the double-arch tunnel, the shape of the pore pressure distribution becomes a funnel around the double-arch tunnel. The relatively low pore pressures in the upper area of the tunnel indicate that the groundwater can easily flow into tunnels from the roof of the tunnel, especially from the shoulders of the double-arch tunnel. The flow vectors are large in the floors, indicating a probably large amount of groundwater flow from the tunnel floor. The large flow vectors at the shoulder of the double-arch tunnel also indicate that the contact interface between the concrete wall and the overlying rock is a potential leakage place. 4.3. Case study 2: groundwater flow in twin-tube tunnels Qingshangang tunnel is a largely-spaced twin-tube tunnel in the ChangJi Expressway. The length of the left and the right tunnels is 1245 and 1227 m, respectively. The maximum overburden of the
tunnel is about 160 m. The spacing between the two tubes varies from 30 to 45 m. The geological survey indicates that there are no large fractures near the tunnels but rock joints are welldeveloped. There are two sets of joints in the rock mass: set 1 with a dip direction of 140–190° and a dip angle of 40–55°; set 2 with a dip direction of 240–260° and a dip angle of 70–80°. Fig. 6 shows the tunnel location and the surrounding rock mass schematically. The EPM model is used in the numerical modelling and a profile is modelled in the section Zk121 + 830–ZK121 + 833. The hydraulic parameters are listed in Table 1. The boundary conditions are the same as the double-arch tunnel described above. The left tunnel is excavated first in the numerical modelling. Fig. 7 shows the groundwater flow vectors and pore pressure distributions after tunnel excavation. It is seen that the seepage field and the flow vectors, to some degree, are different from the double-arch tunnel. After the excavation of the left tunnel, the pore pressure distribution is changed similarly to what described by Hwang and Lu (2007). After both of the tunnels are excavated, the contour lines of the pore pressure become two circles around
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
265
Fig. 6. Schematic illustration of the Qingshangang tunnel and the surrounding rock mass.
Fig. 7. Distribution of the pore pressure and groundwater flow vectors for twin-tube tunnels: (a) contours of the pore pressure (unit, MPa); and (b) groundwater flow vectors.
the tunnels and are partly overlapped in the middle area. The tunnels do not much influence each other on the pore pressures. It indicates that the large spacing between the two tunnels can reduce the influence of the excavation to the seepage field. The fact that the groundwater flow vectors point toward to tunnels means the high pore pressure will lead to leakage problems in the process of excavation. 5. Discussion 5.1. Influencing factors The influencing factors, including underground water table, drainage extent, aquiferous faults and rock joints, can lead to a different flow state around a tunnel. For example, in the case of low underground water table for a tunnel, the original pore pressure around the tunnel is relatively small and the excavation operation may be conducted in a relatively dry condition. The drawdowns of the groundwater are less than those resulted from high water tables. Thus, the leakage problem may be reduced significantly when the water table is low.
In this paper, the rock mass have been treated as an EPM medium and an equivalent isotropic permeability is used in the models for the sake of simplicity. Actually, there are a number of existing joints and fissures that allow the groundwater easily to flow into tunnels as long as the excavation begins. So the flow vectors in the numerical models can just give us a qualitative indication on the direction of the groundwater flow and the potential leakage places. The drainage extent of the tunnel also has an important influence on the groundwater flow state. When a tunnel is excavated without shotcreting and lining, the tunnel face is open for drainage. However, when a watertight lining is applied, the tunnel face becomes an impermeable boundary and the lining would instead be subjected to the pore pressures. 5.2. Relationship between the numerical models and the in situ study As shown in the numerical modelling, the groundwater will flow into tunnels when the tunnel face is barely exposed. The analysis is based on some simplifications, but the results give an indication how the groundwater flow state looks like and where the
266
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
leakage most probably takes place. In the case of double-arch tunnel, a pilot tunnel was excavated first and then the middle wall was constructed. The rock mass became drained after excavation of the pilot tunnel. When the main tunnel was excavated, the groundwater would flow into the excavation space at the contact interface between the middle wall and the overlying rock. This type of leakage was indeed observed, for example, in the Bimaxi double-arch tunnel. Fig. 3a shows two pictures of leakage occurring at the shoulder of the Bimaxi tunnel. Leakage frequency is defined as the number of groundwater leakage places at the shoulder of the tunnel in every 15 m of the tunnel length, divided by the length of the counting interval, i.e., 15 m. The maximum leakage frequency 1 represents a full face inflow in that counting section. It can be seen that the water leakage problem is serious in the double-arch tunnel from Fig. 8. In average, there will be a leakage place in about every 2 m in the tunnel. Different with the numerical modelling, field investigations showed that the groundwater flow would change with the advancing direction in tunnels. The groundwater flow does not change along the trace of the tunnel in numerical modelling since the rock mass is assumed to be isotropic in permeability. The current numerical modelling is meant to give a qualitative assessment on the groundwater flow state around tunnels of different complex opening shapes. Comprehensive simulation of the groundwater flow needs more geological and hydrological input data. Even though, the flow vectors at the contact interface between the middle wall and overlying rock agree with our observation of water leakage. In twin-tub tunnels, the seepage field and the groundwater flow vectors are different from double-arch tunnels. In the QingshanLeakage at the shoulder of the left tunnel
gang twin-tube tunnels, serious groundwater drops occurred at the tunnel vaults and water gushed at the tunnel invert. These phenomena can be discovered by the indications of flow vectors in the numerical modellings. It was observed in the tunnels that the drop water came from fractures or fissures. After shotcreting or even after a secondary lining, leakage problems still existed in the Qingshangang tunnels. One of the reasons for the leakage is that the used drainage system behind the lining is not very effective. Special measures should be taken to improve the efficiency of the drainage system furthermore. 6. Innovative drainage measures 6.1. Conventional methods In many tunnelling projects in China, waterproofing and water drainage must be conducted in accordance to the Chinese national codes, for instance, ‘‘codes for design of road tunnels (JTG D702004)”. Even though a lot of waterproofing methods suggested in the codes are applied, it happens more than often that the waterproofing effect is not satisfied and water leakage always easily takes place in tunnels. In the Qingshangang tunnel, besides the designed waterproofing techniques in tunnellings, including waterproof membranes and drainage pipes, other appropriate methods are also put forward according to the site conditions. For example, probe drilling, is adopted to provide early warnings for water ingression. Grouting is also used to solidify rock formations and improve the overall water tightness in order to improve the stability of excavation heading. Moreover, improved smooth blasting method is applied to reduce the disturbance zone in the vicinity of a tunnel.
Leakage frequency (/m))
Average leakage frequency 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
6.2. New water-gathering system
0
15
30
45
60
75
90
105 120 135 150 165 180 195
Distance from Jishou port of the tunnel (m) Fig. 8. Leakage frequency accounted in every 15 m of tunnel length at the shoulder of the Bimaxi tunnel.
To further improve the drainage performance, an innovative countermeasure has been put forward. The proposed method is named as a self-gluing system for water collection in underground engineering. The aim of the method is to reduce water inflow and leakage into tunnels. It can be used in tunnels with severe water leakage from joints and fractures exposed on excavation surfaces. The system is composed of water-gathering trenches, self-gluing wings and drainage holes. The system is schematically illustrated in Figs. 9–11. The water-gathering trench is made of a plastic membrane that is reinforced by steel wires. The cross-section of the trench is approximately rectangular with a size of 50–200 mm 30 mm
A 6
1
5
1 5
2
7 4
2
A
3
A
A
4
Fig. 9. Schematic illustrations of the new water-gathering system: (1). self-gluing wings of the trench; (2) three-way adapter; (3) water-bearing fissure; (4) water-gathering trench; (5) four-way adapter; (6) water-bearing fracture; and (7) rock surface.
267
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
1
8 9
8 9
1
Fig. 10. Schematic illustrations of the water-gathering trench and the four-way adapter: (1) self-gluing wings of the trench; (8) 1-mm-thick plastic membrane; and (9) 3-mm steel wires.
(width height). The plastic membrane is 1 mm thick and flexible to adapt the fluctuation of rock surface. The steel wire is 3 mm in diameter and is laterally laid in the trench in a spacing of 70 mm. The wires are cemented to the membrane by epoxy resin. A gluing bag is fixed to the top of the trench wall. The gluing bag is made of fiber papers and a gluing agent. The agent is portland cement mixed with an accelerator. The details of the components are shown in Fig. 11. When installing the water-gathering system, the trench is pushed on to the rock surface and covers the water-bearing fractures. The water from the fractures will mix with the gluing agent so that the trench can glue to the rock surface. Nevertheless, some steel nails will be helpful to install the water-gathering trench. The groundwater gathered in the trench then is guided to the longitudinal drainage pipes, which are installed at the merges of the sides with the invert. Site experiences manifest that the new watergathering system is easy to install and very effective for groundwater drainage. 6.3. Field application The new water-gathering system was tested in the Qingshangang tunnel and the drainage effect was satisfactory. Fig. 12 is a 11
12
9
10
8
Fig. 11. Cross-section view of the new water-gathering system: (8) 1-mm-thick plastic membrane; (9) 3-mm steel wires; (10) epoxy resin; (11) fiber paper; and (12) gluing agent.
Fig. 12. The new water-gathering system installed in site.
photo of the system installed before shotcreting. Table 2 is the data collected in the tested sections of the tunnel. It is seen that the water-gathering ratio of the system reached 86.5% in average. Actually, the innovative countermeasure was applied not only in the Qingshangang tunnel but also in other two tunnels in the ChangJi Expressway. In case of double-arch tunnels, the watergathering trench can be installed at the shoulder of the tunnel and water leakage there would be dramatically reduced, creating an acceptable working environment. This method can be applied in rock tunnels with water leakage or water inflow from rock fractures or joints. 7. Conclusion Groundwater flow in the rock mass surrounding tunnels was numerically modelled. The models were constructed with the assumption of an EPM for the rock mass. Two types of tunnels, named double-arch tunnel and largely-spaced twin-tube tunnel,
Table 2 Data from field tests of the system in the Qingshangang tunnel Location ZK121 + 855–ZK121 + 858.20 ZK121 + 858.20–ZK121 + 861.40 ZK121 + 861.40–ZK121 + 864.50 ZK121 + 864.50–ZK121 + 867.45 ZK121 + 867.45–ZK121 + 870.35 ZK121 + 870.35–ZK121 + 873.30 ZK121 + 873.30–ZK121 + 876.35 ZK121 + 876.35–ZK121 + 879.25 Total
Section length (m)
Original water inflow (m3/h)
Trench length in use (m)
Water-inflow with the system (m3/h)
Water-gathering ratio (%)
3.20 3.20 3.10 2.95 2.90 2.95 3.05 2.90
7.5 12.6 10.4 9.7 1.5 1.2 14.8 11.2
33.5 53.5 42.3 37.2 0 0 38.5 31.4
1.9 0.9 1.1 0.9 1.5 1.2 1.0 0.8
74.7 92.9 89.4 90.7 – – 93.2 92.9
24.25
68.9
236.4
9.3
86.5
268
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
were modelled. The distribution of pore pressure after excavation has been obtained. The potential permeable places have been identified by the water flow vectors. In case of double-arch tunnels, attention should be paid to the water leakage at the shoulder of tunnel, where the middle wall has a contact interface with the overlying rock. Both the numerical modelling and site observations indicate that the shoulder may be a potential permeable place. Some factors like the underground water table, drainage conditions, aquifer faults and rock joints, can change the flow state as well as the pore pressure distribution. Thus, geological and hydro-geological investigations should be thoroughly achieved in order to obtain more realistic results from numerical modellings. Based on the modelling results and site investigations, an innovative countermeasure has been developed to improve the drainage performance and reduce the water leakage in tunnels. It is a new self-gluing system for groundwater gathering and drainage. The system is installed by using the groundwater to cement the plastic membrane to the rock surface before initial lining. The pore pressures on the lining would be decreased by well-draining and leakage problems could be reduced in tunnels. The drainage method can be applied in tunnels where water-bearing fractures are well-developed in the surrounding rock mass. Acknowledgement The authors would like to thank the Natural Science Foundation of China for their financial support to this study (Grants No. 50490274 and No. 50674107). References Aksoy, C.O., 2008. Chemical injection application at tunnel service shaft to prevent ground settlement induced by groundwater drainage: a case study. International Journal of Rock Mechanics and Mining Sciences 45 (3), 376–383. Bear, J., 1993. Modelling flow and contaminant transport in fractured rocks. In: Bear, J., Tsang, C.F., de Marisly, G. (Eds.), Flow and Contaminant Transport in Fractured Media. Academic Press, San Diego, CA, pp. 1–37. Berkowitz, B., 2002. Characterizing flow and transport in fractured geological media: a review. Advances in Water Resources 25 (8–12), 861–884.
Caputo, M., Huez, H.-P., 1987. Tunnel waterproofing using polymeric membranes 2 (1), 83–88. FLAC3D, 2005. Fluid-Mechanical Interaction. In: Itasca, I. (Ed.), Consulting Group, Minneapolis, Minnesota, USA. Goodman, R., 1965. Groundwater inflows during tunnel driving. Engineering Geology 2 (2), 39–56. Grantz, W., Tan, L., Sorensen, E., Burger, H., 1997. Waterproofing and maintenance. Tunnelling and Underground Space Technology 12 (2), 111–124. Hwang, J.-H., Lu, C.-C., 2007. A semi-analytical method for analyzing the tunnel water inflow. Tunnelling and Underground Space Technology 22 (1), 39–46. Kleven, O.-B., 2004. Radical approach to waterproofing using spray applied membranes. Concrete (London) 38 (9), 45–46. Kolymbas, D., 2005. Tunelling and Tunnel Mechanics: A Rational Approach to Tunnelling, vol. XV. Springer, Berlin. p. 437. Li, D., Li, X., Zhang, W., Gong, F., Huang, B., 2007. Stability analysis of surrounding rock of multi-arch tunnel based on coupled fluid-solid theorem. Yanshilixue Yu Gongcheng Xuebao/Chinese Journal of Rock Mechanics and Engineering 26 (5), 1056–1064. Li, X.-B., Zhang, W., Li, D.-Y., Wang, Q.-S., 2008. Influence of underground water seepage flow on surrounding rock deformation of multi-arch tunnel. Journal of Central South University of Technology (English Edition) 15 (1), 69–74. Meiri, D., 1985. Unconfined groundwater flow calculation into a tunnel. Journal of Hydrology 82 (1–2), 69–75. Molinero, J., Samper, J., Juanes, R., 2002. Numerical modeling of the transient hydrogeological response produced by tunnel construction in fractured bedrocks. Engineering Geology 64 (4), 369–386. Nilsen, B., 2001. Stability and Water Leakage of Hard Rock Subsea Tunnels. Swets and Zeitlinger, Lisse. pp. 497–502. Or, D., Tuller, M., Fedors, R., 2005. Seepage into drifts and tunnels in unsaturated fractured rock. Water Resources Research 41 (5), 05022. Samardzioska, T., Popov, V., 2005. Numerical comparison of the equivalent continuum, non-homogeneous and dual porosity models for flow and transport in fractured porous media. Advances in Water Resources 28 (3), 235–255. Shin, J.H., Addenbrooke, T.I., Potts, D.M., 2002. A numerical study of the effect of groundwater movement on long-term tunnel behaviour. Geotechnique 52 (6), 391–403. Tseng, D.J., Tsai, B.R., Chang, L.C., 2001. A case study on ground treatment for a rock tunnel with high groundwater ingression in Taiwan. Tunnelling and Underground Space Technology 16 (3), 175–183. Yuan, Y., Jiang, X., Lee, C.F., 2000. Tunnel waterproofing practices in China. Tunnelling and Underground Space Technology 15 (2), 227–233. Zhang, L., Franklin, J.A., 1993. Prediction of water flow into rock tunnels. An analytical solution assuming an hydraulic conductivity gradient. International Journal of Rock Mechanics and Mining Sciences and Geomechanics Abstracts 30 (1), 37–46. Zhang, Y.-T., 2005. Rock Hydraulics and Engineering. China Water Conservancy and Water Electricity Press, Beijing, China.
Contents lists available at ScienceDirect
Tunnelling and Underground Space Technology journal homepage: www.elsevier.com/locate/tust
Case studies of groundwater flow into tunnels and an innovative water-gathering system for water drainage Diyuan Li a,b,*, Xibing Li a, Charlie C. Li b, Bingren Huang a, Fengqiang Gong a, Wei Zhang a a b
School of Resources and Safety Engineering, Central South University, Changsha, Hunan 410083, China Department of Geology and Mineral Resources Engineering, Norwegian University of Science and Technology, Trondheim, Norway
a r t i c l e
i n f o
Article history: Received 20 May 2008 Received in revised form 20 August 2008 Accepted 27 August 2008 Available online 11 October 2008 Keywords: Case study Numerical modeling Groundwater flow Tunnel Leakage Water-gathering system
a b s t r a c t Groundwater inflow into tunnels can constitute a potential hazard and also is an important factor influencing the speed of tunnel excavation. In this paper the results of numerical modelling are presented to investigate the groundwater flow and the distribution of the pore pressure around tunnels. Two types of tunnels, double-arch tunnel and twin-tube tunnel, were studied. Potential leakage places are identified for the two types of tunnels. The most permeable place in the double-arch tunnel is at the contact interface between the middle wall and the overlying rock. The results of numerical modelling are compared with field observations in the case studies. Based on the results of numerical modelling and the field investigations, an innovative water-gathering system for reducing water leakage was proposed and applied in some tunnels on ChangJi Expressway in China. The water-gathering system can be quickly glued to the rock surface and easily installed for tunnelling. It can be applied in tunnels where waterbearing fractures are well-developed in the rock mass. Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
2. Background
Groundwater inflow into tunnels can constitute a potential hazard and also is an important factor influencing the speed of excavation advancement. Nowadays more and more road tunnels of complex opening shapes are constructed in mountainous areas all over the world. The groundwater flows in such special opening shapes are complicated and different from those of simple opening shapes. Groundwater inflow and leakage problems are difficult to manage in tunnels. Based on site investigations and field water pressure tests, two cases of numerical study about groundwater flow were carried out using code FLAC3D in this paper. The distributions of seepage field around double-arch tunnel and twin-tube tunnel were obtained. Potential permeable places were identified based on the results of the numerical modellings. The modelling results from case studies were compared with field observations at Bimaxi tunnel and Qingshangang tunnel of the ChangJi Expressway in Hunan Province, China. According to the modelling results and site investigations, an innovative watergathering system for water drainage was put forward and applied in fields. The proposed method for reducing water leakage in tunnels is easy to carry out and effective in reducing groundwater flow into tunnels.
Groundwater inflow is an old but common and difficult problem in tunnelling project. It may retard the speed of excavation under tunnel construction, while on the other hand it may lead a potential hazard to the stability of the tunnel in the long-term run. One of the practical difficulties in tunnel construction and maintenance is related to groundwater inflow and leakage. On one hand, some of the most disastrous events in tunnelling are associated with groundwater inflow of large volumes in highly fractured water-saturated rocks. For instance, in the Pinglin tunnels of the Taipei-Ilan Expressway Project, the sudden groundwater inflow was up to 750 l/s (Tseng et al., 2001). It leaded to a collapse at the tunnel heading and the TBM in the pilot tunnel was trapped and damaged. Therefore, it is essential to have a good understanding on the water inflow rule in the rock mass under tunnel construction, so that appropriate changes can be taken in tunnel routes and drainage systems. On the other hand, water leakage exists in most of the finished tunnels. It is very difficult to realize a totally watertight tunnel. Yuan et al. (2000) gave a review of the practices in tunnel waterproofing in China. It is pointed out that it is hard to achieve a completely watertight tunnel using one waterproofing method alone. Consequently, multiple waterproofing measures must be integrated in underground works. For instance, membranes and watertight concrete were used together to prevent water from entering the tunnel space for immersed tunnels (Grantz et al., 1997).
* Corresponding author. Tel.: +86 731 8877276. E-mail address: [email protected] (D. Li). 0886-7798/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.tust.2008.08.006
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
Groundwater inflow depends on a number of factors, such as the permeability of the rock mass, the groundwater table, the aperture of rock fractures, and the size of the excavation. Although it is very difficult to accurately predict the water inflow during tunnel construction, a large number of researches have been done to tackle the problem, particularly through numerical simulations in aiming at rational and visible solutions. There are several analytical expressions in the literature to predict discharges into tunnels. In a homogeneous infinite water table aquifer, Goodman (1965) proposed an expression for the flow rate per unit tunnel length, q,
q¼
2pKH0 lnð2H0 =rÞ
ð1Þ
where K is the hydraulic conductivity, r is the radius of the tunnel and H0 is the depth of the tunnel below the water table. Eq. (1) is widely used in tunnelling owing to its simplicity and good accuracy of prediction. Zhang and Franklin (1993) proposed another analytical solution to predict the water inflow with a gradient hydraulic conductivity on the basis of the solution by Goodman. Hwang and Lu (2007) proposed a semi-analytical approach for water inflow. For the constant flow and variable drawdown problem of sources and sinks, the governing equation is written as:
Sðr; tÞ ¼ Q FðP; r; tÞ
ð2Þ
where S is the drawdown of water level, Q is the inflow, F is a function of the drawdown, P is a parameter for all relevant hydrogeological factors (including permeability, coefficient of storage, thickness of aquifer and distance to the boundary), r is the distance to the location of water inflow and t is the time since the beginning of water inflow. The formulae were developed for a tunnel intersecting an inclined aquifer, which is the most common case of tunnel inflow. Analytical solutions rely on given hydro-geological assumptions with simple circular or rectangular openings. In complex hydrogeological conditions such as fractured rocks, numerical models can be adopted to study groundwater flow around tunnels. A numerical flow model is a simplified representation of an aquifer which aims at capturing the most relevant features of groundwater flow. Construction of a numerical model requires formulating first a conceptual model, which is a qualitative description of the main properties of the flow system. On water flow through fractured media, several conceptual models have been proposed so far, including (1) equivalent porous media (EPM) models, (2) discrete fracture network (DFN) models and (3) hybrid models. Berkowitz (2002) gave a detailed review on these conceptual models and Samardzioska and Popov (2005) compared three different models for modelling of flow and solute transport in fractured porous media, in terms of their predictions of the flow and solute transport field variables. The three models can be applied to different conditions. It is claimed that the EPM models are efficient and faster to calculate than other types of models to simulate the fluid flow in the fractured media. Some numerical modelling work for groundwater inflow can be found in literatures (for instance, Meiri, 1985; Molinero et al., 2002; Shin et al., 2002; Or et al., 2005). For example, Meiri (1985) used the finite element method for nonsteady groundwater flows with a free surface and compared the numerical results with those obtained from analytical approximations. Shin et al. (2002) considered the effect of groundwater movement in the long-term. The lining permeability was varied from fully permeable to impermeable, and an approach to model finite lining permeability was presented and assessed. Or et al. (2005) presented a simple and physically based model for estimation of the seepage into tunnels excavated in unsaturated fractured rocks. Most of these studies, however, concentrate on the prediction of groundwater inflow into tunnels and the tunnel shapes are always quite simple.
261
It is a big issue on how to deal with the groundwater when a tunnel is constructed below the groundwater table. To drain or to seal? Kolymbas (2005) pointed out that sealed tunnels do not influence the groundwater but their lining has to support the full water pressure. Drainage can influence the surrounding groundwater considerably but can relieve the lining from hydrostatic pressure. Actually tunnels below the groundwater table can be either sealed or drained. Some useful methods have been put forward by Caputo and Huez (1987), Nilsen (2001), Kleven (2004), Kolymbas (2005) and Aksoy (2008). Geosynthetics are applied for drainage (geospacers and geocomposite drains) and for waterproofing (geomembranes). Grouting and chemical injection are used for improving ground conditions and waterproofing. However, water leakage phenomena do exist in tunnels no matter they are drained or sealed by traditional ways. This problem has not been well-resolved. 3. Description of the problem A great number of expressways have been constructed for the last two decades in China. One of them is the expressway from Changde to Jishou, called ChangJi Expressway, in southwest of China. The construction of this expressway began in 2004. Many tunnels have to be constructed in the course of the expressway. The total length of the ChangJi Expressway is 224 km, more than 40 km of which are tunnels. Most of the short tunnels (shorter than 500 m) are constructed as double-arch tunnels, while long tunnels (longer than 1000 m) are constructed as largely-spaced twin-tubes. A double-arch tunnel is characterized by its concrete wall constructed in the middle of the tunnel. Largely-spaced twin-tube tunnels are excavated separately so that they are not affected by each other in mechanics. The excavation method is the so-called NATM – the new Austrian tunnel method. Fig. 1 shows examples of these two different types of tunnels in the ChangJi Expressway. Water flows into tunnels can cause a lot of problems not only under construction but also in the long-term operation. The tunnels of the ChangJi Expressway are located in subtropical mountainous areas in Hunan Province. The climate there is characterized by the moist air and abundant rainfalls. The invasion of groundwater and surface water into tunnels would cause serious unstable problems in tunnels. Some previous work (Li et al., 2007, 2008) about the coupling effect can be proof to this conclusion. After a thorough investigation programme and field observations, it was found that there were four types of water inflows in tunnels of ChangJi Expressway: (1) Groundwater inrush: This type of water inflow would cause rock collapse in the tunnel face. This kind of collapse is a big problem for tunnel construction but it does not happen very often. It occurs only when abnormal geological structures like aquifer faults are exposed on the tunnel face. This type of incident happened twice in a twin-tube long tunnel– Qingshangang tunnel where rock collapses led to a big hole in the roof close to the tunnel face. (2) Groundwater inflow and water drops occurred at the shoulder of the tunnel. Most of the moderately long and short tunnels in ChangJi Expressway were designed to the double-arched structure. The cross-section of a double-arch tunnel is shown in Fig. 2. The total span (D) of the tunnel is about 25 m. The potential permeable place is marked in the figure and the water leakage photos are shown in Fig. 3a. (3) Groundwater inflow occurred in the position of ‘I’ steel sets, which were the major support elements of the tunnel. Because the groundwater was not well drained by the drainage pipes, the high water pressure resulted in seepage problem along ‘I’ steel. It can be seen in Fig. 3b.
262
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
Fig. 1. Two different types of tunnels on ChangJi Expressway: (a) Bimaxi double-arch tunnel; and (b) Qingshangang twin-tube tunnel.
Fig. 2. Cross-section of the double-arch tunnel (unit, cm).
Fig. 3. Water leakage phenomena in tunnels by site investigation: (a) water leakage at the contact interface between the middle wall and overlying rock in a double-arch tunnel; and (b) water leakage occurred at the ‘‘I” steel sets in a twin-tube tunnel.
263
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
(4) Groundwater inflow from the tunnel floor. The floor water worsened the work environment and may affect the construction of the tunnel floor. (5) The groundwater flow would build up water pressures at the back of the watertight lining in case of an unsatisfactory drainage. A good understanding on the distribution of the underground seepage field would be helpful for the design of water sealing and drainage.
9.81 m/s2 and; qw is the density of water, which equals to 1000 kg/ m3. For water at 20 °C, qwg = 9.81 103 Pa/m, the following conversion can be derived to calculate k in SI units for water in FLAC3D:
k ðin SI unitsÞ kh ðin cm=sÞ 1:02 106
ð5Þ
The hydraulic parameters obtained from the in situ water pressure tests are shown in Table 1. 4.2. Case study 1: groundwater flow in a double-arch tunnel
4. Case studies of groundwater flow into tunnels 4.1. Simplifications and parameters of numerical modelling The numerical modelling of groundwater flow into tunnels has to adopt some assumptions and simplifications. On one hand, it is difficult to simulate the groundwater flow around tunnels because of the heterogeneity of the rock mass and the complexity of groundwater flow in the rock joints. On the other hand, empirical methods and waterproof guidelines for underground engineering will be always employed when one designs a tunnel below the groundwater table. However, they are usually conservative. In this paper, the code FLAC3D is adopted to model the groundwater flow into tunnels. FLAC3D models the fluid flow in a permeable medium. The permeability of a rock mass is mainly determined by the fractures in the mass rather than by the porosity of the intact rock. The fractures are so ubiquitous and random that it is hard to simulate them numerically. However, as mentioned above, from the macroscopic point of view, a fractured rock mass can be treated as an EPM. The assumption is valid as long as it is possible to define a representative elementary volume (REV) (Bear, 1993). This requires that the model domain should be much greater than the average fracture spacing. In the formation concerned in this study, there is not any fault in the near field of the tunnels. The rock mass has been simplified to an equivalent porous medium with groundwater flows obeying the Darcy law. To gain the permeability of the rock mass, small-scale in situ tests, called water pressure tests, were conducted in Bimaxi tunnel. It is assumed that the permeability is isotropic in the rock mass. The parameter of permeability can be calculated by (Zhang, 2005)
Q al kh ¼ 0:528 log h0 l r0
ð3Þ
where the constant 0.528 is dimensionless, Q denotes the measured volume flow rate, in m3/s; h0 is the piezometer head at the center of the test section, in m; l is the length of the test section, in m; a is a coefficient, which is set to 0.66 when the test section is far from the impermeable layer, otherwise 1.32; r0 is the radius of the borehole for water pressure test, in m; and kh is the permeability of the rock mass, in m/s. Through the water pressure tests, the permeability of the surrounding rock mass was obtained in the range of 103– 105 cm/s. In FLAC3D the conversion between permeability coefficient k and kh is (FLAC3D, 2005):
kh ¼ kg q
ð4Þ
w
where k is the isotropic permeability coefficient used in FLAC3D, m2/(Pa s) in SI units; g is the gravitational acceleration, which is
The Bimaxi double-arch tunnel is selected as the case for numerical modelling to analyze the groundwater seepage field in the rock mass, and is located in a hill near the city of Jishou. The tunnel is about 500 m long. In the stretch of the tunnel there are a few gullies which primarily strike north with some striking northeast. The tunnel axis is nearly perpendicular to the strike of the hill in the section of Changde part, but becomes oblique or parallel to the hill strike in the section of Jishou part. The hill slopes are steep and form a V-shape gully. The overall strike of the hill is about 340° and the slope angle is about 15–35°. The vegetation is mainly small bamboo bushes and herbaceous plants. Rock outcrop is visible in some places. In the tunnel stretch there is no large fracture structures and faults. Three sets of rock joints exist: set 1 with a dip direction of 148° and a dip angle of 89°; set 2 with a dip direction of 350° and a dip angle of 56°; set 3 with a dip direction of 225° and a dip angle of 77°. The joint spacing varies from 5 to 20 cm and the connectivity of the joint aperture is fairly good. The meshes of 3D numerical model of the double-arch tunnel are shown in Fig. 4. The X-axis and the Y-axis are in the horizontal plane and the Z-axis is vertical. The model size is 125 m 15 m 80 m (X Y Z). From the top to the bottom of the model, it is applied different material properties simulating the sandstones in different weathering degrees and it is also applied different hydraulic parameters. In FLAC3D, boundaries are impermeable by default and saturation cannot be imposed as a boundary condition. A leaky boundary condition has the form of (FLAC3D, 2005):
q
n
¼ hðp p
e
Þ
ð6Þ
where qn is the component of the specific discharge normal to the boundary in the direction of the exterior normal, h is the leakage coefficient in (m3/N s), p is the pore pressure at the boundary surface, and pe is the pore pressure in the region to or from which leakage is assumed to occur. Before tunnel excavation the pore pressure below the groundwater table is hydrostatic. After excavation, the boundary of the tunnel is modelled by a free water seepage boundary where the adjacent underground water infiltrates into the tunnel. The concrete wall in the middle is modelled as a fluid null group in the code as it is a relatively impermeable. The seepage flow field in the rock mass has been changed after excavation. The mechanical response of the excavation is not considered in the numerical analysis. Fig. 5a shows the contour diagram of the pore pressure distributions at a vertical section (y = 6 m) after excavation but before lining. Fig. 5b shows the groundwater flow vectors in the near field of the double-arch tunnel.
Table 1 Hydraulic parameters of the surrounding rock mass in numerical modelling by FLAC3D Tunnel Bimaxi tunnel
Qingshangang tunnel
Weathering Heavily weathered Moderately weathered Slightly weathered Moderately weathered
Permeability kh (cm/s) 3
1.21 10 4.61 104 3.00 105 5.02 104
Porosity ratio n
Permeability coefficient in FLAC3D k (m2/Pa s)
0.45 0.40 0.25 0.30
1.23 109 4.70 1010 3.06 1011 5.12 1010
264
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
Fig. 4. A numerical model of the double-arch tunnel in FLAC3D.
-10 -20
z/m
-30 -40 -50 -60 -70
-40
-30
-20
-10
0
10
20
30
40
50
60
70
x/m
a. Contours of the pore pressure (unit:MPa)
b. Groundwater flow vectors
Fig. 5. Distribution of the pore pressure and groundwater flow vectors for a double-arch tunnel: (a) contours of the pore pressure (unit, MPa); and (b) groundwater flow vectors.
It can be seen that the groundwater pressure in the zones near the tunnel decreases after tunnel excavation. The groundwater begins to flow into the tunnel as soon as it is excavated. For the double-arch tunnel, the shape of the pore pressure distribution becomes a funnel around the double-arch tunnel. The relatively low pore pressures in the upper area of the tunnel indicate that the groundwater can easily flow into tunnels from the roof of the tunnel, especially from the shoulders of the double-arch tunnel. The flow vectors are large in the floors, indicating a probably large amount of groundwater flow from the tunnel floor. The large flow vectors at the shoulder of the double-arch tunnel also indicate that the contact interface between the concrete wall and the overlying rock is a potential leakage place. 4.3. Case study 2: groundwater flow in twin-tube tunnels Qingshangang tunnel is a largely-spaced twin-tube tunnel in the ChangJi Expressway. The length of the left and the right tunnels is 1245 and 1227 m, respectively. The maximum overburden of the
tunnel is about 160 m. The spacing between the two tubes varies from 30 to 45 m. The geological survey indicates that there are no large fractures near the tunnels but rock joints are welldeveloped. There are two sets of joints in the rock mass: set 1 with a dip direction of 140–190° and a dip angle of 40–55°; set 2 with a dip direction of 240–260° and a dip angle of 70–80°. Fig. 6 shows the tunnel location and the surrounding rock mass schematically. The EPM model is used in the numerical modelling and a profile is modelled in the section Zk121 + 830–ZK121 + 833. The hydraulic parameters are listed in Table 1. The boundary conditions are the same as the double-arch tunnel described above. The left tunnel is excavated first in the numerical modelling. Fig. 7 shows the groundwater flow vectors and pore pressure distributions after tunnel excavation. It is seen that the seepage field and the flow vectors, to some degree, are different from the double-arch tunnel. After the excavation of the left tunnel, the pore pressure distribution is changed similarly to what described by Hwang and Lu (2007). After both of the tunnels are excavated, the contour lines of the pore pressure become two circles around
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
265
Fig. 6. Schematic illustration of the Qingshangang tunnel and the surrounding rock mass.
Fig. 7. Distribution of the pore pressure and groundwater flow vectors for twin-tube tunnels: (a) contours of the pore pressure (unit, MPa); and (b) groundwater flow vectors.
the tunnels and are partly overlapped in the middle area. The tunnels do not much influence each other on the pore pressures. It indicates that the large spacing between the two tunnels can reduce the influence of the excavation to the seepage field. The fact that the groundwater flow vectors point toward to tunnels means the high pore pressure will lead to leakage problems in the process of excavation. 5. Discussion 5.1. Influencing factors The influencing factors, including underground water table, drainage extent, aquiferous faults and rock joints, can lead to a different flow state around a tunnel. For example, in the case of low underground water table for a tunnel, the original pore pressure around the tunnel is relatively small and the excavation operation may be conducted in a relatively dry condition. The drawdowns of the groundwater are less than those resulted from high water tables. Thus, the leakage problem may be reduced significantly when the water table is low.
In this paper, the rock mass have been treated as an EPM medium and an equivalent isotropic permeability is used in the models for the sake of simplicity. Actually, there are a number of existing joints and fissures that allow the groundwater easily to flow into tunnels as long as the excavation begins. So the flow vectors in the numerical models can just give us a qualitative indication on the direction of the groundwater flow and the potential leakage places. The drainage extent of the tunnel also has an important influence on the groundwater flow state. When a tunnel is excavated without shotcreting and lining, the tunnel face is open for drainage. However, when a watertight lining is applied, the tunnel face becomes an impermeable boundary and the lining would instead be subjected to the pore pressures. 5.2. Relationship between the numerical models and the in situ study As shown in the numerical modelling, the groundwater will flow into tunnels when the tunnel face is barely exposed. The analysis is based on some simplifications, but the results give an indication how the groundwater flow state looks like and where the
266
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
leakage most probably takes place. In the case of double-arch tunnel, a pilot tunnel was excavated first and then the middle wall was constructed. The rock mass became drained after excavation of the pilot tunnel. When the main tunnel was excavated, the groundwater would flow into the excavation space at the contact interface between the middle wall and the overlying rock. This type of leakage was indeed observed, for example, in the Bimaxi double-arch tunnel. Fig. 3a shows two pictures of leakage occurring at the shoulder of the Bimaxi tunnel. Leakage frequency is defined as the number of groundwater leakage places at the shoulder of the tunnel in every 15 m of the tunnel length, divided by the length of the counting interval, i.e., 15 m. The maximum leakage frequency 1 represents a full face inflow in that counting section. It can be seen that the water leakage problem is serious in the double-arch tunnel from Fig. 8. In average, there will be a leakage place in about every 2 m in the tunnel. Different with the numerical modelling, field investigations showed that the groundwater flow would change with the advancing direction in tunnels. The groundwater flow does not change along the trace of the tunnel in numerical modelling since the rock mass is assumed to be isotropic in permeability. The current numerical modelling is meant to give a qualitative assessment on the groundwater flow state around tunnels of different complex opening shapes. Comprehensive simulation of the groundwater flow needs more geological and hydrological input data. Even though, the flow vectors at the contact interface between the middle wall and overlying rock agree with our observation of water leakage. In twin-tub tunnels, the seepage field and the groundwater flow vectors are different from double-arch tunnels. In the QingshanLeakage at the shoulder of the left tunnel
gang twin-tube tunnels, serious groundwater drops occurred at the tunnel vaults and water gushed at the tunnel invert. These phenomena can be discovered by the indications of flow vectors in the numerical modellings. It was observed in the tunnels that the drop water came from fractures or fissures. After shotcreting or even after a secondary lining, leakage problems still existed in the Qingshangang tunnels. One of the reasons for the leakage is that the used drainage system behind the lining is not very effective. Special measures should be taken to improve the efficiency of the drainage system furthermore. 6. Innovative drainage measures 6.1. Conventional methods In many tunnelling projects in China, waterproofing and water drainage must be conducted in accordance to the Chinese national codes, for instance, ‘‘codes for design of road tunnels (JTG D702004)”. Even though a lot of waterproofing methods suggested in the codes are applied, it happens more than often that the waterproofing effect is not satisfied and water leakage always easily takes place in tunnels. In the Qingshangang tunnel, besides the designed waterproofing techniques in tunnellings, including waterproof membranes and drainage pipes, other appropriate methods are also put forward according to the site conditions. For example, probe drilling, is adopted to provide early warnings for water ingression. Grouting is also used to solidify rock formations and improve the overall water tightness in order to improve the stability of excavation heading. Moreover, improved smooth blasting method is applied to reduce the disturbance zone in the vicinity of a tunnel.
Leakage frequency (/m))
Average leakage frequency 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
6.2. New water-gathering system
0
15
30
45
60
75
90
105 120 135 150 165 180 195
Distance from Jishou port of the tunnel (m) Fig. 8. Leakage frequency accounted in every 15 m of tunnel length at the shoulder of the Bimaxi tunnel.
To further improve the drainage performance, an innovative countermeasure has been put forward. The proposed method is named as a self-gluing system for water collection in underground engineering. The aim of the method is to reduce water inflow and leakage into tunnels. It can be used in tunnels with severe water leakage from joints and fractures exposed on excavation surfaces. The system is composed of water-gathering trenches, self-gluing wings and drainage holes. The system is schematically illustrated in Figs. 9–11. The water-gathering trench is made of a plastic membrane that is reinforced by steel wires. The cross-section of the trench is approximately rectangular with a size of 50–200 mm 30 mm
A 6
1
5
1 5
2
7 4
2
A
3
A
A
4
Fig. 9. Schematic illustrations of the new water-gathering system: (1). self-gluing wings of the trench; (2) three-way adapter; (3) water-bearing fissure; (4) water-gathering trench; (5) four-way adapter; (6) water-bearing fracture; and (7) rock surface.
267
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
1
8 9
8 9
1
Fig. 10. Schematic illustrations of the water-gathering trench and the four-way adapter: (1) self-gluing wings of the trench; (8) 1-mm-thick plastic membrane; and (9) 3-mm steel wires.
(width height). The plastic membrane is 1 mm thick and flexible to adapt the fluctuation of rock surface. The steel wire is 3 mm in diameter and is laterally laid in the trench in a spacing of 70 mm. The wires are cemented to the membrane by epoxy resin. A gluing bag is fixed to the top of the trench wall. The gluing bag is made of fiber papers and a gluing agent. The agent is portland cement mixed with an accelerator. The details of the components are shown in Fig. 11. When installing the water-gathering system, the trench is pushed on to the rock surface and covers the water-bearing fractures. The water from the fractures will mix with the gluing agent so that the trench can glue to the rock surface. Nevertheless, some steel nails will be helpful to install the water-gathering trench. The groundwater gathered in the trench then is guided to the longitudinal drainage pipes, which are installed at the merges of the sides with the invert. Site experiences manifest that the new watergathering system is easy to install and very effective for groundwater drainage. 6.3. Field application The new water-gathering system was tested in the Qingshangang tunnel and the drainage effect was satisfactory. Fig. 12 is a 11
12
9
10
8
Fig. 11. Cross-section view of the new water-gathering system: (8) 1-mm-thick plastic membrane; (9) 3-mm steel wires; (10) epoxy resin; (11) fiber paper; and (12) gluing agent.
Fig. 12. The new water-gathering system installed in site.
photo of the system installed before shotcreting. Table 2 is the data collected in the tested sections of the tunnel. It is seen that the water-gathering ratio of the system reached 86.5% in average. Actually, the innovative countermeasure was applied not only in the Qingshangang tunnel but also in other two tunnels in the ChangJi Expressway. In case of double-arch tunnels, the watergathering trench can be installed at the shoulder of the tunnel and water leakage there would be dramatically reduced, creating an acceptable working environment. This method can be applied in rock tunnels with water leakage or water inflow from rock fractures or joints. 7. Conclusion Groundwater flow in the rock mass surrounding tunnels was numerically modelled. The models were constructed with the assumption of an EPM for the rock mass. Two types of tunnels, named double-arch tunnel and largely-spaced twin-tube tunnel,
Table 2 Data from field tests of the system in the Qingshangang tunnel Location ZK121 + 855–ZK121 + 858.20 ZK121 + 858.20–ZK121 + 861.40 ZK121 + 861.40–ZK121 + 864.50 ZK121 + 864.50–ZK121 + 867.45 ZK121 + 867.45–ZK121 + 870.35 ZK121 + 870.35–ZK121 + 873.30 ZK121 + 873.30–ZK121 + 876.35 ZK121 + 876.35–ZK121 + 879.25 Total
Section length (m)
Original water inflow (m3/h)
Trench length in use (m)
Water-inflow with the system (m3/h)
Water-gathering ratio (%)
3.20 3.20 3.10 2.95 2.90 2.95 3.05 2.90
7.5 12.6 10.4 9.7 1.5 1.2 14.8 11.2
33.5 53.5 42.3 37.2 0 0 38.5 31.4
1.9 0.9 1.1 0.9 1.5 1.2 1.0 0.8
74.7 92.9 89.4 90.7 – – 93.2 92.9
24.25
68.9
236.4
9.3
86.5
268
D. Li et al. / Tunnelling and Underground Space Technology 24 (2009) 260–268
were modelled. The distribution of pore pressure after excavation has been obtained. The potential permeable places have been identified by the water flow vectors. In case of double-arch tunnels, attention should be paid to the water leakage at the shoulder of tunnel, where the middle wall has a contact interface with the overlying rock. Both the numerical modelling and site observations indicate that the shoulder may be a potential permeable place. Some factors like the underground water table, drainage conditions, aquifer faults and rock joints, can change the flow state as well as the pore pressure distribution. Thus, geological and hydro-geological investigations should be thoroughly achieved in order to obtain more realistic results from numerical modellings. Based on the modelling results and site investigations, an innovative countermeasure has been developed to improve the drainage performance and reduce the water leakage in tunnels. It is a new self-gluing system for groundwater gathering and drainage. The system is installed by using the groundwater to cement the plastic membrane to the rock surface before initial lining. The pore pressures on the lining would be decreased by well-draining and leakage problems could be reduced in tunnels. The drainage method can be applied in tunnels where water-bearing fractures are well-developed in the surrounding rock mass. Acknowledgement The authors would like to thank the Natural Science Foundation of China for their financial support to this study (Grants No. 50490274 and No. 50674107). References Aksoy, C.O., 2008. Chemical injection application at tunnel service shaft to prevent ground settlement induced by groundwater drainage: a case study. International Journal of Rock Mechanics and Mining Sciences 45 (3), 376–383. Bear, J., 1993. Modelling flow and contaminant transport in fractured rocks. In: Bear, J., Tsang, C.F., de Marisly, G. (Eds.), Flow and Contaminant Transport in Fractured Media. Academic Press, San Diego, CA, pp. 1–37. Berkowitz, B., 2002. Characterizing flow and transport in fractured geological media: a review. Advances in Water Resources 25 (8–12), 861–884.
Caputo, M., Huez, H.-P., 1987. Tunnel waterproofing using polymeric membranes 2 (1), 83–88. FLAC3D, 2005. Fluid-Mechanical Interaction. In: Itasca, I. (Ed.), Consulting Group, Minneapolis, Minnesota, USA. Goodman, R., 1965. Groundwater inflows during tunnel driving. Engineering Geology 2 (2), 39–56. Grantz, W., Tan, L., Sorensen, E., Burger, H., 1997. Waterproofing and maintenance. Tunnelling and Underground Space Technology 12 (2), 111–124. Hwang, J.-H., Lu, C.-C., 2007. A semi-analytical method for analyzing the tunnel water inflow. Tunnelling and Underground Space Technology 22 (1), 39–46. Kleven, O.-B., 2004. Radical approach to waterproofing using spray applied membranes. Concrete (London) 38 (9), 45–46. Kolymbas, D., 2005. Tunelling and Tunnel Mechanics: A Rational Approach to Tunnelling, vol. XV. Springer, Berlin. p. 437. Li, D., Li, X., Zhang, W., Gong, F., Huang, B., 2007. Stability analysis of surrounding rock of multi-arch tunnel based on coupled fluid-solid theorem. Yanshilixue Yu Gongcheng Xuebao/Chinese Journal of Rock Mechanics and Engineering 26 (5), 1056–1064. Li, X.-B., Zhang, W., Li, D.-Y., Wang, Q.-S., 2008. Influence of underground water seepage flow on surrounding rock deformation of multi-arch tunnel. Journal of Central South University of Technology (English Edition) 15 (1), 69–74. Meiri, D., 1985. Unconfined groundwater flow calculation into a tunnel. Journal of Hydrology 82 (1–2), 69–75. Molinero, J., Samper, J., Juanes, R., 2002. Numerical modeling of the transient hydrogeological response produced by tunnel construction in fractured bedrocks. Engineering Geology 64 (4), 369–386. Nilsen, B., 2001. Stability and Water Leakage of Hard Rock Subsea Tunnels. Swets and Zeitlinger, Lisse. pp. 497–502. Or, D., Tuller, M., Fedors, R., 2005. Seepage into drifts and tunnels in unsaturated fractured rock. Water Resources Research 41 (5), 05022. Samardzioska, T., Popov, V., 2005. Numerical comparison of the equivalent continuum, non-homogeneous and dual porosity models for flow and transport in fractured porous media. Advances in Water Resources 28 (3), 235–255. Shin, J.H., Addenbrooke, T.I., Potts, D.M., 2002. A numerical study of the effect of groundwater movement on long-term tunnel behaviour. Geotechnique 52 (6), 391–403. Tseng, D.J., Tsai, B.R., Chang, L.C., 2001. A case study on ground treatment for a rock tunnel with high groundwater ingression in Taiwan. Tunnelling and Underground Space Technology 16 (3), 175–183. Yuan, Y., Jiang, X., Lee, C.F., 2000. Tunnel waterproofing practices in China. Tunnelling and Underground Space Technology 15 (2), 227–233. Zhang, L., Franklin, J.A., 1993. Prediction of water flow into rock tunnels. An analytical solution assuming an hydraulic conductivity gradient. International Journal of Rock Mechanics and Mining Sciences and Geomechanics Abstracts 30 (1), 37–46. Zhang, Y.-T., 2005. Rock Hydraulics and Engineering. China Water Conservancy and Water Electricity Press, Beijing, China.