Abstract
To learn the airproof capacity of filter cakes as opening chambers under air pressure, a series of tests were carried out. The variations of discharged water with air pressure and time were observed, and the relationship between airproof capacity of filter cakes and surrounding air pressure was analysed. The test results indicated that there were three stages as compressed air acting on filter cakes: completely not infiltration, a very small amount of infiltration, and penetration leakage. The certain air pressure between the first and second stages was called the airproofing value of filter cake. And a capillary bundle model was used to explain the mechanism of air tightness of filter cakes. In Nanjing Yangtze River Tunnel, a 5 cm thickness filter cake was formed in gravel sand, and its airproofing value was a little lower than 0.12 MPa. The air pressure used as opening chamber should be equal to the summation of water pressure in sand and airproofing value of filter cake. While the air pressure is larger than the summation, the filter cake would be gas permeable. The slurry formulation and airproofing value of filter cakes obtained in the tests were applied successfully in Nanjing Yangtze River Tunnel.
1. Introduction
Nowadays, shield tunneling method has been widely used in the construction of subway and river-crossing tunnels in China [1]. However, shield-tunneling machine malfunctions and highly worn cutting wheel or cutters were encountered frequently and nagged the construction enterprises seriously in the difficult geological conditions [2–4]. To restore the excavation, technicists have to open chambers to maintain equipment or exchange the cutters. Meanwhile maintaining the stability of tunnel face is the most vital problem as opening chambers in shield tunneling projects [5]. Generally, compressed air is used to support the tunnel face during opening chamber operations in some subjects. However, in some soils with high permeability, the tunnel face cannot be supported effectively without any accessory measures as the massive air leakage. This problem has attracted the attention of construction enterprises in the world and some accessory measures, such as filter cakes formed on the surface, have been taken to reduce the permeability of soils as opening chamber under air pressure [2, 5].
In slurry-shield-driven project, the method of opening the working chamber under compressed air has been widely used, such as in Nanjing Yangtze River Tunnel [6–8]. A filter cake plays an important role on the working face stability in the normally tunneling [9, 10]. As opening an excavation chamber under compressed air, the air pressure is in direct action on the filter cake, and it is the key to the excavation face stability. However, few theoretical researches on the mechanism of air tightness of the filter cake are reported.
In shield tunnel construction, opening an excavation chamber under compressed air is an engineering challenge. Cases of excavation chamber opening have been reported in some countries, such as the Elbe 4 Tunnel in Germany, which was excavated in the slurry shield mode of the TBM composed of sand, marlite, and boulder in the chamber replacement center cutters [6]. The Red Line Metro in the St. Petersburg subway in Russia is under a pressure of 0.55 MPa, with a clay and fine sand formation opening an excavation chamber under compressed air [1]. The Holland Westerschelde Tunnel has a gas pressure of MPa in a depth of 60 m of a clay-layer open chamber [7, 8]. These aforementioned cases of opening an excavation chamber belong to engineering accident treatment. Thus, this study focused on describing the processes and points in chamber opening, such as symptoms of decompression sickness. However, theoretical and practical analyses on the mechanism of air tightness of filter cakes and an excavation face during chamber opening are limited.
A filter cake is a special porous medium whose airproof capability mainly varies with the movement of water and gas in a porous medium. Richards and Ogata advanced the concept of the intake value of a ceramic plate. The intake value is the air pressure that the saturated porous ceramic plate begins to drain; the air enters but not through this plate [11]. Fredlund and Rahardjo considered that air entry into unsaturated soils causes air pressure to break through the surface tension of the bound water on the surface of soil particles [12]. Ye et al. indicated that the threshold pressure of gas infiltrating in unsaturated soils causes the bound water to fill the pore channel of the clay [13].
This paper presents an experimental and a case study on the Nanjing Yangtze River Tunnel accessing the working chamber under compressed air to study the breathability of a filter cake under the air pressure. The concept of moderate air pressure and the filter cake air tightness mechanism were discussed.
2. Experimental Methods and Materials
2.1. Experimental Apparatus
A self-developed filter cake air tightness device was used as the test apparatus in this paper. Figure 1 shows the schematic of the test apparatus. The apparatus mainly comprised a Perspex cylinder for pressure stabilization and filtering acquisition. The Perspex cylinder had an inner diameter of 8.4 cm and a height of 120 cm. It can withstand a maximum pressure of about 0.7 MPa. The control precision of the pressure stabilization system was 0.01 MPa. The accuracy of the filtering acquisition system was 0.1 g. Simulation of the shield working chamber-working face-stratum system was achieved by loading strata and slurry into the Perspex cylinder, whose end was sealed by a flange. An air compressor was used to provide the permeation pressure for slurry and air pressure for the air tightness of the filter cake. At the same time, the air pressure was controlled by the stabilization system. The pressure indicator monitored the pressure variation inside the apparatus. As placed at 1.2 m above the ground and filled with water in it, the pressure drainage apparatus provided a stable head as water or gas passed through the strata. And then the seepage discharge of water or gas was measured by electronic scales. The variation in seepage discharge with time was automatically recorded by the data acquisition apparatus.
2.2. Experimental Materials
2.2.1. Slurry Materials
Two slurries were used in this experiment, namely, slurry 1 and slurry 2. Their properties are shown in Table 1. The major components of the slurry included bentonite, silt, and an appropriate amount of slurry additives. The bentonite was sodium bentonite from Nanjing Tangshan and the silt was obtained from Baima Lake in Jiangsu Province. The grain size distribution curve is shown in Figure 2.
2.2.2. Stratum Materials
In this experiment, a uniform sand with particle sizes ranging from 0.25 mm to 0.5 mm was used as stratum materials. It was obtained by screening between the nearest two sizes of standard sieves, the 0.25 mm and 0.5 mm. The stratum was prepared by three times with a depth of 9 cm per layer, and the load height was 27 cm. The soil specimen was compacted to a dry density () of 1.5 g/cm3 and the corresponding porosity () was 0.43. The hydraulic conductivity () of soils was measured by the constant head permeability tests, and its hydraulic conductivity was about 6.5 × 10−2 cm/s. To prevent sand draining, a 5 cm sand filter layer with particle sizes ranging from 2 mm to 5 mm and hydraulic conductivity of 4.2 cm/s was placed at the bottom.
2.3. Experimental Methods
Preparation of the Filter Cake Test. The filter layer was first loaded, hit, and compacted. The strata were then loaded, the layer was hit and compacted, and the surface was screed. Water was slowly flooded from the bottom to the top until the strata were completely saturated [14]. The whole saturation process required around 30 min. The slurry was slowly injected along the diversion rod into the strata and the experimental apparatus was sealed. The air compressor was turned on, and the air pressure of 0.1 MPa was acting on the slurry in the apparatus. The drain valve was then opened and the slurry permeated into the strata under the action of pressure. When the increment of filtering quantity was lower than 1 cm3/min and stable, the filter cake was deemed to be formed. Finally, the intake and drain valves were closed, the slurry valve was opened, and the slurry was drained. The shape and thickness of the filter cake were then recorded.
Filter Cake Air Tightness Test. The slurry valve was closed, the intake valve was opened to exert air pressure to the filter cake, and then the drain valve was opened. The air pressure and variation in seepage discharge with time were then recorded. At the end of the experiment, the thickness of the filter cake was recorded. The thickness of every filter cake and infiltrated zone formed in the tests were measured three times by a ruler with a precision of 1 mm. The average value was recorded as the final thickness of filter cakes and infiltrated zone.
Two Sets of Experiment. (1) The filter cake formation test was performed four times using slurry 1 under 0.1 MPa pressure. Then, air pressure was exerted to each filter cake stepwise, with an initial pressure of 0.02 MPa and an increase of 0.02 MPa every 2 min until the variation in the air permeation quantity became very obvious. The seepage discharge of the filter cake under each stepwise increase in air pressure and the deformation of the filter cake were recorded. (2) The filter cake formation test was performed five times using slurry 2 under 0.1 MPa pressure. Different air pressures (0.1, 0.14, 0.15, 0.16, and 0.2 MPa) were then exerted to each filter cake continuously for 24 h. The variation in seepage discharge with time until air leakage appeared in the filter cake was recorded.
3. Experimental Results
3.1. Form of the Filter Cakes
The forms of the filter cake of the two groups of nine experiments were recorded. The thickness of the filter cake and the infiltrated zone for each experiment are summarized in Table 2.
The thickness of filter cake 1 and the infiltrated zone were approximately 4.9 and 29.6 mm, respectively. The thickness of the filter cake 2 and the infiltrated zone were approximately 6.0 and 26.9 mm, respectively. The thickness of the formed filter cake and infiltrated zone were essentially identical for the same slurry under the same conditions, which indicated that the experiment was of reusable operability and the experimental method was feasible.
3.2. Airproof Phenomena of the Filter Cake under Air Pressure
The variation in seepage flow of the filter cake with air pressure is shown in Figure 3. Below a specific pressure (≤0.12 MPa), the seepage flow was low and the cake was airtight. On the other hand, when the air pressure was higher than a specific pressure (≥0.14 MPa), the amount of discharged water became very large and the cake lost its air tightness. After the test, compressive deformation was generated.
Figure 4 shows the variation in the seepage flow of the filter cake with time. Under 0.1 and 0.14 Mpa air pressure, the seepage flow of the cake was very low (about 0.003 cm3/s). After continuous observation for 24 h, the seepage flow decreased to 0 cm3/s with time. During this period, the stratum was saturated and no air entered the cake. However, under 0.15, 0.16, and 0.20 MPa, the seepage flow of the filter cake obviously increased with time. A higher air pressure resulted in faster seepage flow.
The physical processes of the filter cake under air pressure can be divided into four steps as shown in Figure 5. First, the air cannot enter the cake and the cake appeared to be undergoing a nondrainage process; the air pressure was low. Second, with increased air pressure, the amount of discharged water from the cake was low. Third, the air pressure increased until the air entered the cake, and the cake appeared to undergo a transient drainage process. Fourth, the air pressure increased to the airproofing pressure of the filter cake, the amount of discharged water immediately increased considerably, and the air entered the stratum through the filter cake. The essence of process is that cake became permeable from airproofing, and the required time depended on the magnitude of the air pressure.
3.3. Airproofing Pressure of the Filter Cakes
Figure 3 shows that inflection points appeared in the curves. For filter cake 1, the inflection points appeared at 0.12 MPa air pressure for each test. Considering the air pressure at which the inflection point appeared as airproofing pressure, when the air pressure was below the airproofing pressure, the filter cake was airtight. On the other hand, when the air pressure was above the airproofing pressure, the filter cake was permeable. The test results indicated that the airproofing pressure also existed in filter cake 2. When the air pressure was below 0.14 MPa, the filter cake was airtight. The amount of discharged water was very low because of the compressive deformation of the cake under air pressure. By contrast, when the air pressure was above 0.15 MPa, the filter cake was permeable. The test results indicated that the airproofing pressure of filter cake 2 was between 0.14 and 0.15 MPa.
4. A Case Study on Opening a Slurry Shield Excavation Chamber in the Nanjing Yangtze River Tunnel
4.1. General Conditions in the Nanjing Yangtze River Tunnel
The Nanjing Yangtze River Tunnel was excavated by the two-slurry shield mode of the TBMs. The outer diameter of the segments was 14.5 m [1]. The pressure cabin of the slurry shield mode of the TBM in the Nanjing Yangtze River Tunnel was divided into two chambers: excavation and air pressure. The top of the air pressure chamber had high air pressure, and the bottom was connected with the excavation chamber by suspensions. The two-chamber design was very flexible for inspection and repair works. Until August 2008, the TBM was excavated to 650–659 rings (K4 + 900 to K4 + 918), and the torque was rapidly increased to 15 MNm and sometimes to 20 MNm. The speed was lower than 5 mm/min [5]. Thus, the TBM had to be stopped to verify the explanation of the aforementioned exceptional events. The highest tide water level throughout the project from August 2008 to December 2008 was 8.37 m, and the water pressure at the centerline of the TBM was approximately 0.53 MPa.
4.2. Method of Opening the Chamber and Study on the Filter Cake
4.2.1. Selecting the Method of Opening the Chamber
The TBM was stopped 53.25 m under the water level of the Yangtze River (Figure 6). Thus, reinforcing the soils near the excavation face was not advisable, and opening the chamber under atmospheric pressure after soil improvement would not be conducted. The 3 m mode in the upside supported by compressed air and 11.96 m in the downside still filled with suspensions were selected. The cutting wheel would be circumrotated and the other cutters would be repaired.
4.2.2. Filter Cake Formation Tests and Air Pressure Tests of the Filter Cake
Filter cake formation tests were conducted on the mixed strata under 0.3 MPa, and the slurry was the suspensions mixed with bentonite and natural clay ( g/cm3, funnel viscosity = 25 s). A 5 mm thick filter cake was rapidly formed (Figure 7).
After filter cake formation on the surface, an air pressure test was conducted and a curve for the relationship between the air pressure and corresponding discharged water was obtained (Figure 8). Below a specific pressure, the water discharge amount was low. At an air pressure higher than the specific pressure, the discharged water immediately increased considerably. For a 5 mm thick filter cake formed in the test, no discharged water was measured because the air pressure was lower than 0.12 MPa. Thus, the compressed air cannot penetrate the filter cake. At 0.12 MPa pressure, the discharged water suddenly increased and some air penetrated the filter cake. Similar phenomena occurred when the pressure reached 0.13 and 0.14 MPa. Therefore, the critical pressure for this filter cake airproofing (the airproofing pressure) was 0.11 MPa. The threshold air pressure that can be adopted in the project was equal to the water pressure of the ground and the airproofing pressure of the filter cake formed in the test. As aforementioned, the threshold air pressure was 0.59 MPa (0.48 + 0.11 MPa). Aside from withstanding the pore water pressure, part of the air pressure (the airproofing pressure) can counteract the effective earth pressure in the tunnel face. The threshold air pressure was nearly equal to the total earth pressure at rest, that is, 0.59 MPa.
4.3. The Implementation of Opening Chamber under Air Pressure
Before opening the chamber, an air pressure test was performed in the site to verify the reliability of the filter cake formed on the surface. The level of the slurry in the working chamber decreased to 3 m after a dense filter cake was formed. The top of the chamber was filled with compressed air then. Considering the safety factor of the airproofing pressure of the filter cake, air pressure from 0.55 MPa to 0.57 MPa was set to support the tunnel face. This range of pressure was a little lower than the threshold airproofing pressure of 0.59 MPa but a little larger than the total active earth pressure of 0.547 MPa. The air pressure change in the excavation chamber was monitored to estimate the stability of the excavation surface. The monitoring data showed that the gas escape was rare and the pressure in the chamber was stable. The filter cake was sealed in this case. The technicians accessed the excavation chamber to repair the cutters, and they can work only from 2 h to 5 h in the chamber every time. After finishing the assignment, they exited the chamber and resumed the slurry level to the top. The same procedures were repeated in the next operation.
Figure 9 shows the images of filter cakes on the surface of the tunnel face. According to these photos and observations in the chamber, a dense and thick filter cake was formed on the tunnel face, with no collapse and air escape. The filter cake was also deemed stable.
5. Discussion
5.1. Physical Meaning of the Airproofing Pressure of the Filter Cake
To study the nature of the airproofing pressure and pneumatically convert it to the effective stress in the mechanism, it should be considered from the surface tension of the water in the porous medium [15]. A capillary bundle model is very commonly used to study the movement of water in the porous medium [16, 17]. Pore connectivity in the media is observed as a capillary in the model [18]. The capillary phenomenon is formed by the joint action of the wettability of the capillary wall and the surface tension of water. The negative capillary pressure is borne by the meniscus and then passes through the surface tension to the wall [19].
When air pressure is applied to the saturated filter cake, a meniscus is formed to resist the pressure through the surface tension of the meniscus. The pressure is transmitted to the particles of the filter cake and then converted to the effective stress. Within a certain pressure range, the curvature of the meniscus increases with the pressure. When the curvature threshold is reached, the meniscus is pushed down by the pressure, which is the air pressure overcoming the surface tension of water in the pores of the filter cake. This pressure is called the airproofing pressure.
The airproofing pressure overcomes the surface tension of water in the pores of the filter cake. Assuming that the meniscus is a spherical surface, the stress analysis is shown in Figure 10. is pressure, is the coefficient of surface tension of water, and is the pore radius [14, 15].
According to the balancing of forces in the vertical direction, . The pressure values are then obtained. This equation shows that the airproofing pressure is related to the pore size; smaller pores result in higher airproofing pressure. This pressure magnitude is independent of the pressure time and the pressure exerted.
Figure 3 shows that when the pressure is less than the airproofing pressure, that is, the air pressure is equal to or less than 0.12 MPa, the air pressure is insufficient to overcome the surface tension of water in the pores of the airproofing pressure. The pressure is then passed through the interaction between the water and soil particles to the soil skeleton and becomes effective stress. The cake is compressed in the role of the effective stress. Figure 11 compares the deformation and seepage discharge of filter cake when the airproofing pressure is reached, suggesting that the seepage discharge was mainly generated by the cake deformation. Figure 4 shows two curves of 0.1 and 0.14 MPa. The seepage discharge progressively decreases and finally approaches zero. The deformation of the filter cake also gradually becomes stable.
When the pressure is higher than the airproofing pressure, such as the pressure reaches 0.14 MPa in Figure 3 and 0.15 MPa in Figure 4, the air pressure overcomes the surface tension of the water in the pores of the filter cake, the gas penetrates the pores, seepage discharge significantly increases, and the filter cake is breathable. Figure 4 shows that, at different pressures (0.15, 0.16, and 0.2 MPa), the difference between the processes of air permeability of the cake indicates that a higher air pressure results in a shorter time within which the cake becomes breathable.
5.2. Airproofing Pressure of the Strata and Filter Cakes
The effects of the construction method under compressed air depend on the permeability of the strata; a lower stratum permeability results in better effects. When the permeability coefficient is higher than 10−2 cm/s, the construction method is difficult to use under compressed air. For high penetrative strata, a slurry is used to form a filter cake to decrease the permeability of the stratum and avoid the leakage of compressed air. Construction is then started under compressed air.
When forming a dense filter cake, moderate air pressure can counteract part of the effective earth pressure. Without a filter cake, compressed air can penetrate the pore and counteract the water pressure, but cannot counteract the effective pressure. The pore of the formed dense filter cake is so small that the air needs to overcome high resistance when passing through the channel. Before exceeding the resistance, the air cannot penetrate the pores but can only alter the curvature of the liquid in the channel and convert the air pressure to effective earth pressure. This mechanism explains why compressed air can counteract moderate effective pressure. In this case, the airproofing pressure of the filter cake is approximately 0.11 MPa. Considering a moderate safety factor, 0.1 MPa is used as the airproofing pressure, which is close to the effective earth pressure at rest (0.11 MPa). Therefore, filter cakes play an important role in balancing the earth pressure on the tunnel face. It also reduces air escape and enhances the stability of the tunnel face when opening the chamber.
5.3. Experience of Opening Chamber Successfully
In the present project, the air pressure less than the total earth pressure at rest and a little larger than the total active earth pressure was set to support the tunnel face when opening the chamber in the Nanjing Yangtze River Tunnel. In normal TBM excavation, the slurry pressure in excavation chamber was set 0.02 MPa larger than the total earth pressure at rest in general [18, 19]. However, the air pressure set in the present study was a little smaller than the slurry pressure. Lower cohesion of the sand could also maintain stability when supported by a pressure a little larger than the total active earth pressure because of the self-support stability of the soil caused by its own strength. Actually, the air pressure of 3 m on the top of the chamber established the slurry pressure of 12 m at the bottom. Increasing the air pressure has an important role on keeping the stability of the whole tunnel face. The slurry pressure was larger than the air pressure before it was replaced with compressed air. Whether the filter cake would fall off from the tunnel when supported by a little lower air pressure was a problem. However, the observations during physical access to the face showed that the filter cake had not fallen off and the face was stable.
6. Conclusions
(1) During chamber opening, a complete filter cake is formed after the infiltration of slurry, and a stage of airproof deformation of a filter cake occurs when compressed air acts on the surface of the filter cake. When the pressure of the compressed air is low, the air cannot pass through the filter cake and the amount of discharged water is low because of the deformation of the filter cake. When the pressure of compressed air is higher than a certain value, the air can pass through the filter cake and the discharged water rapidly increases. This pressure is defined as the airproofing pressure. This phenomenon is a draining process of pore water when its surface tension is overcome by air pressure from the exterior.
(2) The airproofing pressure of a filter cake has a certain magnitude that is independent of the pressure of compressed air and the acting time. Its magnitude only depends on the form of the filter cake. The airproofing pressure of the filter cake determines the magnitude of threshold pressure of the compressed air. If the difference between the compressed air pressure and water pressure on the working face is lower than the airproofing pressure, the stability of the working face can be guaranteed by a construction method under compressed air. However, when the buried depth of the tunnel was too deep, the needed supporting pressure was very high, and sometimes the stability of the working face cannot be guaranteed only by the construction method under compressed air; other auxiliary measures should be used.
(3) Air pressure test results show that the airproofing pressure of the filter cake is 0.11 MPa. This result suggests that, below 0.11 MPa, air cannot penetrate the pores of the filter cake and the air pressure can withstand the effective pressure. For practical application, the air pressure must be lower than the airproofing pressure of the filter cake to support the tunnel face and maintain its stability.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
Acknowledgments
The authors would like to acknowledge the financial support of the National Natural Science Foundation of China (51408191), the National Program on Key Basic Research Project of China (2015CB057803/2012CB719804), the China Postdoctoral Science Foundation (2014M560388), the Project (SKLGDUEK1306) Supported by State Key Laboratory for Geomechanics and Deep Underground Engineering, China University of Mining & Technology, and the Natural Science Foundation of Jiangsu Province (BK2011025).