- Email: [email protected]
Model test of immersed tube tunnel foundation treated by sand-flow method
Tunnelling and Underground Space Technology 40 (2014) 102–108
Contents lists available at ScienceDirect
Tunnelling and Underground Space Technology journal homepage: www.elsevier.com/locate/tust
Model test of immersed tube tunnel foundation treated by sand-flow method Wei Li a,b, Yingguang Fang a, Haihong Mo a, Renguo Gu a, Junsheng Chen a,c,⇑, Yizhao Wang a, Deluan Feng a a
School of Civil Engineering & Transportation, South China University of Technology, Guangzhou 510640, Guangdong, PR China Guangzhou Electric Power Design Institute, Guangzhou 510075, Guangdong, PR China c State Key Laboratory of Subtropical Building Science, Guangzhou 510641, Guangdong, PR China b
a r t i c l e
i n f o
Article history: Received 30 July 2012 Received in revised form 7 August 2013 Accepted 28 September 2013 Available online 20 October 2013 Keywords: Immersed tube tunnel Foundation treatment Sand-flow method Model test Sand deposit Water pressure
a b s t r a c t In order to explore the immersed tunnel foundation treated by sand-flow method, modeling principles for the full-scale model test of sand flow were put forward. In addition, a sand-flow test model was built, which consisted of model system, equipment system and measurement system. The situation of sanddeposit expanding and the water pressure in the foundation trench were evaluated through the model test. The results show that a semi-closed space was formed between the model board and the expanding sand deposit, which made the water pressure in it rising with little range of volatility. The sand deposit gradually became non-circular truncated cone which shaped with its expanding radius, and the difference of the water pressure increased at each direction. The water pressure in crater had a linear increase with the sand-deposit radius, with a maximal value of 0.015 2 MPa. The volatility of the water pressure under the tunnel board and the water pressure value in the crater could be used as bases of construction control. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction The main underwater tunnel methods for crossing rivers and straits are mining methods, shield methods and immersed tube methods. Since 1910, the advent of the first true sense immersed tube tunnel-the Detroit River immersed tube tunnel built, immersed tube tunnels have been developed in USA (Grantz, 1997), Europe (Glerum, 1995; Rasmussen, 1997) and Asia (Kiyomiya, 1995; Janssen et al., 2006; Yang et al., 2008) vigorously. Its advantages are obviously evident. There are several factors affecting the settlement of the immersed tube tunnel, such as sub-soil conditions, foundation, siltation, surcharge, and trench dredging methods (Grantz, 2001a), but it is difficult to accurately predict and control the final settlement. Grantz (2001b) shared his immersed tunnel experience with several typical tunnel types and actual settlement records, which shows that the immersed tube tunnel foundation and the treated methods of it have a significant influence on the final settlement. Considered as one of the treatment methods of immersed tunnel foundation used nowadays, sand-flow method (Glerum, 1995;
⇑ Corresponding author at: School of Civil Engineering & Transportation, South China University of Technology, Guangzhou 510640, Guangdong, PR China. Tel.: +86 020 22236638. E-mail address: [email protected] (J. Chen). 0886-7798/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tust.2013.09.015
Chen et al., 2002), with a number of advantages, is one of the mainstream construction methods. Sand-flow method was first used in Vlake Tunnel of Holland (Tongeren et al., 1978), corresponding model test was carried out and the mechanism of the method and construction parameter were explored before construction. Water pressure near the injection opening and sand content in the sand-flow mixture were recorded in the construction. Though immersed tunnel construction started relatively late in China, a large number of sand-flow model tests were taken. By the Guangzhou Pearl River Tunnel, Li (2001) built a sand-flow test model and pointed out that the pressure near the injection opening and buoyancy force increased with the expansion of sand deposit, providing a technical basis for the construction. Before the construction of the Shengwudao–Daxuecheng Tunnel, Wang et al. (2009) studied the sand-deposit radius, void ratio and pressure by the 1:5 reduced scale model test, verified the original design and construction program. The authors take the Guagnzhou Zhoutouzui Tunnel (GMED&RI and MCAL, 2008) in China as the background, and put forward modeling principles for the full-scale model test of sand flow. Sand-flow test model, consisting of model system, equipment system and measurement system, was built. The relationship of sand deposit and time, as well as the distribution law of water pressure in the foundation trench was gained. Furthermore, forecasting methods of the sand-deposit radius and construction-control index
W. Li et al. / Tunnelling and Underground Space Technology 40 (2014) 102–108
for engineering practice were proposed. The authors hope that this paper could provide basis for the sand flow method and the test model could provide a reference for further studies. 2. Model system 2.1. Zhoutouzui immersed tunnel project The length of Guangzhou Zhoutouzui immersed tunnel (GMED&RI and MCAL, 2008) is 340 m. The designed pipe segments of the tunnel in Pearl River are E1 (85 m), E2 (85 m), E3 (79.5 m), E4 (90.5 m), of which the E1 and E4 are variable cross-section segments. The tunnel foundation will be treated by sand-flow method. The typical pipe segment and the designed sand deposit are shown in Fig. 1. 2.2. Modeling principles for immersed tunnel In order to guarantee the rationality and validity of the model test, the test model was designed in accordance with the following principles: (1) The plane shape of the model board was equivalent to the floor of the tunnel in a full-scale sand-deposit area, which simulated the top boundary condition of the fluid in construction. (2) The space between the model board edge and the tank wall was similar to the one between the lateral wall of immersed tunnel and the trench slope, which was equivalent to the lateral boundary condition of the fluid. (3) By plugging the side of the trench, the various ‘‘artificial boundary plugging’’ caused by the sand deposit already existed at the one side or both sides of the constructing sand deposit could be simulated. (4) The model board was immersed in water to simulate the underwater construction condition. The only difference happened between the depth of tested water and that of the
Fig. 1. Plane and cross-section of Guangzhou Zhoutouzui immersed tube tunnel/ cm.
103
actual construction water, which would generate no significant differences in the test and had been verified by the preliminary experiment (Fang et al., 2012). (5) The performance parameters of the test equipment system were fully consistent with those of the construction equipment, which could provide stable sand–water flow as those in construction. (6) Measuring system worked accurately and continuously without influencing the test. (7) Test sand had the same category and the similar particle size distribution with the construction sand material. 2.3. Model system of immersed tunnel According to the modeling principles above, a variable crosssection pipe segment in a sand-deposit range was modeled and a full-scale test model for sand-flow method was built. The simulated area is shown by coarse dotted line in Fig. 1(a). The model system of immersed tunnel consists of a model board, a experimental tank, a circulation tank and a underground reservoir, as shown in Fig. 2. The model board is a reinforced concrete, with a dimension of 8.2 m 15.2 m (length width), which was used to full-scale simulate the floor of the immersed tunnel in a sand-deposit range. The clearance between the model board and the ground is 1.0 m, which simulated the clearance of foundation trench. An injection opening was simulated by a 6-in. steel nozzle, which was installed on the board centre, connecting the sand pump by steel-wired hose. Opening–closing holes for observation and sampling were installed at the centerline of the board.
Fig. 2. Test model system. Note: The E, S, W, N in the paper expressed the East, South, West and North direction of the model board.
104
W. Li et al. / Tunnelling and Underground Space Technology 40 (2014) 102–108
The dimension of the experimental tank is 10.5 m 17.2 m 2.5 m (width length height). The model board was placed in the tank, with a clearance of 0.8–1.3 m between it and the tank wall. In the test, the tank was filled with water and the model board was under the water surface, which simulated the underwater condition. Weep holes and overflow holes were installed on the N side of the tank wall, while water-resistant steel door on the W side. The dimension of circulation tank is 4.0 m 3.0 m 2.0 m (length width height), one of its sides was connected to experimental tank, while the other side was linked to water pump through steel-wired hose. As a result, buffering fluid and re-circulating water were achieved. The 400 m3 – capacity underground reservoir was located at the bottom of the test site, which was used to provide the testing water to the tank. 3. Equipment system 3.1. Test equipment system The equipment system of sand-flow test consists of four components: water conveyance part, sand conveyance part, pressure injection part and controller part. All the equipment were selfdeveloped and the performance parameters could be used directly in construction for equipment selection. The water conveyance part composed of water pump, butterfly valve, return valve and 6-in. steel-wired hoses. It was used to sup-
ply quantitative water to sand–water mixer. The sand conveyance part comprised shovel car, first conveyor belt, Sand storage device, sand distributor and second conveyor belt. It was used to supply quantitative sand to sand–water mixer. The pressure injection part was made of sand–water mixer, sand pump, return valve and 6-in. steel-wired hoses. It was used to inject the sand–water mixture to the bottom of model board. The controller part, also known as control electric cabinet, was used to control all the equipment. Consequently, the unified control to equipment and the monitor to electric current and voltage of the system were achieved. The test equipment system refers to Fig. 3 and the photos of main test equipment are shown in Fig. 4. 3.2. Sand–water ratio control The sand and water conveyance rates were determined before the test. The sand distributor was installed under the sand storage device. The sand flow rate was controlled by adjusting the distributor and the water flow rate was controlled by adjusting the return valve and butterfly valve of the water pump. The flow of water pump and sand distributor were adjusted to the pre-set sand–water ratio and to the appropriate rate (i.e., the total sand–water input of the equipment system matched the output of the sand pump per unit time, and the mixer did not overflow or adsorbed dry) in the test. Therefore, steady sand–water flow was guaranteed. 4. Measurement system 4.1. Water pressure measurement Hydro-dynamic gauges (0–0.2 Mpa) were installed in the trench and listed in Fig. 5 and Table 1. The measurement time interval was 5 min and the maximal value was recorded whenever it occurred. 4.2. Sand-deposit measurement
Fig. 3. Layout of equipment system. Note: 1 – First conveyor belt; 2 – sand storage device; 3 – sand distributor; 4 – second conveyor belt; 5 – mixer; 6 – sand pump; 7 – water pump; 8 – control cabinet.
The process of sand-deposit expansion was measured by sanddeposit detectors (Fang et al., 2012), which were installed on the
Fig. 4. Photos of main equipment.
W. Li et al. / Tunnelling and Underground Space Technology 40 (2014) 102–108
105
unchanged with time under a certain measuring point, it could be considered that the sand-deposit radius expanded to that measuring point at that time, and the time–history curves of sand-deposit radius on each trend could be drawn by connecting the time data of the sand deposit expanded to different radius. 5. Test scheme and procedures
Fig. 5. Layout of hydro-dynamic gauges in trench. Note: The serial number indicates the hydro-dynamic gauges located at the underside of model board (the top surface of sand-deposit).
Table 1 Serial number and position of hydro-dynamic gauges. Radius Direction
0.5 m
2.0 m
3.5 m
6.5 m
E S W N
1056# — 250# —
1041# 244# 1064# 1048#
1055# 247# 1036# 1029#
256# — 1058# —
According to the design of Guangzhou Zhoutouzui immersed tube tunnel (GMED&RI and MCAL, 2008), the sand–water ratio, namely mass ratio, was set as 1:8 and the trench clearance was 1.0 m. The model boundary was widely open. The model board was made by fair-faced concrete and placed at a depth of 0.9 m under water. In the test, sand was drawn into sand–water mixer at the preset flow through first conveyor belt, sand storage device, sand distributor and second conveyor belt. Water was pumped into sand– water mixer from circulation tank by water pump at the pre-set flow. After being fully mixed in the mixer, the sand–water mixture was drawn and pumped into the bottom of model board by sand pump. The sediment of sand particles formed sand deposit and the water entered into circulation tank. 6. Test results and analysis 6.1. Sand-deposit expansion The typical time–history curves of sand-deposit height on H & M axes are illustrated in Fig. 7, the law on other axes is similar to these figures. Typical sand-deposit profiles on H–M axis are obtained by connecting the data point of sand-deposit height of H & M axes, which measured at the same time, shown in Fig. 8. Fig. 7 shows that, at the beginning of the test, sand particles of the mixture deposit in the trench where it is close to the injection opening and expand gradually in radial direction. Clearance is likely to exist between the model board and the sand-deposit sur-
Fig. 6. Layout of sand-deposit detectors.
Table 2 Number and position of sand-deposit detectors/m. Measuring point Axis
1
2
3
4
5
6
7
8
F H L K M
0.9 2.0 2.0 1.0 1.0
2.0 3.1 3.0 2.0 2.0
3.0 4.1 4.0 3.0 3.0
3.8 5.1 5.0 3.8 4.0
4.8 6.1 — — 5.0
5.7 7.1 — — 6.0
6.5 8.0 — — 7.1
7.5 — — — 8.0
model board, shown in Fig. 6 and Table 2. The measurement time interval was 3 min. The real-time sand-deposit height data of the place of each measuring point (shown in Fig. 6) were obtained by detectors directly. The time–history curves of sand-deposit height on each axis could be got by connecting the above mentioned data in the whole experiment process. When the sand-deposit height remained
Fig. 7. Time–history curves of sand-deposit height on H & M axes.
106
W. Li et al. / Tunnelling and Underground Space Technology 40 (2014) 102–108
Fig. 8. Sketch map of sand-deposit profiles.
face in the center area of the board, near the injection opening, while the sand deposit could fully filled upon the trench in the surrounding area. Based on Fig. 8, it is clear that annular sand dam and circular plash are gradually formed around the injection opening at the early stage of test. According to Fig. 7, it is obvious that the height of the sand dam (as shown in Fig. 8) will no longer change when it expands to about 95 cm height. Till then, the sand dam has expanded to a sand deposit (also known as ‘‘pancake’’) with a relatively stable height, and the plash has turned into a crater with a constant depth. Till now, a semi-closed space under injection opening has been formed. The clearance on the top of sand deposit shown in Figs. 7 and 8 is a gap, which was formed in the process of test and existed in the central area of sand deposit top, around the injection opening. Besides, chutes also formed in the outer area of the gap area, with limited number and irregular channel-like distribution, which are the reasons why the chutes cannot be detected by detectors. Typical sand-deposit profiles are shown in Fig. 9. The real-time data of sand-deposit radius at each direction and the fitting
average expansion trend, described by the dotted line are shown in Fig. 10. It is shown from the features of sand-deposit profiles in Fig. 9 that the sand–water flow which injected into the crater emitted out through the gap and chutes at high pressure and speed. Then, the pressure and the speed lowered down and the sand particles settled, forming oblique layered texture structure. The sand deposit expands in this manner. The dip angle of the texture is 27–31°, which is similar to the repose angle of the test sand in water. Fig. 10 shows that the data points of radius expanding in all directions almost overlap when sand-deposit radius is relatively small. In this stage, sand–water flow jetting out through gap and the sand particles has a continuous deposit on the outer margin surface of the sand deposit, with the sand deposit expanding at the same speed in all directions and forming truncated cone. This is verified by the smooth and continuous curves measured by Point 1 to 3 shown in Fig. 7. Fig. 10 also shows that the discreteness of the radius data points increased when the sand deposit expanded to a large radius. Sand– water flow was conveyed to the outer margin by chute in this stage. The uneven distribute of chute, and the properties of blocking and constant detouring make sand particles convey to each direction discontinuously and unequally. Thus, expanding speeds of sand deposit are significantly different at each direction and form an irregular shape. Furthermore, the larger the sand-deposit radius is, the more uneven the chute distributes and the more unbalanced the sand deposit expands. This is verified by the step shape curves measured by Point 4 to 8 shown in Fig. 7. 6.2. Water pressure distribution
Fig. 9. Photo of typical sand-deposit profile.
Fig. 10. Time–history curves of sand-deposit radius on each trend.
The time–history curves of water pressure in the trench are illustrated in Fig. 11. The times of sand deposit expands to different radius and corresponding times of water pressure changes are shown in Table 3. (1) Fig. 11 shows that at the initial stage of test, curve slope and increment in Fig. 11(a) are significantly larger than the one in Fig. 11(b–d). Combined with the observation of the test water level, it is obvious that the curve slopes and increments in Fig. 11(b–d) almost conformed, which is just caused by the slight rose of the static water level. These show that at the initial stage of test, aside from the influence of the static water lever, the impact of sand–water flow on the water pressure is great, in the central area of the model bottom but little on the one which in the peripheral area. Combined with the analysis of segment 6.1, the fluid boundary condition under injection opening changed with the forming process of crater and sand-deposit, which means that the trench clearance decreases gradually and the sand–water flow pumped in the crater ran over more and more difficultly. Thus, the water pressure increased in the central area of model board bottom. On the contrary, sand particles have not deposited in the peripheral area of model board bottom, and the trench clearance (i.e., the boundary
W. Li et al. / Tunnelling and Underground Space Technology 40 (2014) 102–108
107
Fig. 11. Time–history curves of water pressure in the trench.
Table 3 Timetable for sand-deposit reaches different radius and water pressure changes/min. Radius
2.0 m
3.5 m
6.5 m
Times of sand-deposit arriving Times of water-pressure increasing
77 85
180 195
422 435
condition) has not changed yet. Therefore, water pressure in this place is not affected. (2) Table 3 shows that the times that sand deposit expanded to different radius are generally correspond to those when water pressure started rising. It is shown in Figs. 8 and 11 that, where the sand deposit expanded, the water pressure at corresponding position started rising with little range volatility, otherwise, there were no significant changes on it. The sand deposit and crater formed a semi-closed space with the model board, gap and chutes that became the passageway of sand–water flow, causing the increase of flow resistance and the water pressure within the range of the formed sand deposit. Water pressure in the range of sand deposit, which rose instantly with the gap and chutes, were blocked by sand particles in the process. Then the weak places of sand deposit were washed out and the chutes rechanneled, leading to a direct result of leakage of the sand–water flow and reducing the water pressure. Again and again, water pressure presented increasing in little range volatility. (3) Fig. 11(a) shows that, in the middle and latter stage of the test, the average water pressure in the central area of model bottom can be linear fitted as
P ¼ 1:33 105 t þ P0
ð1Þ
where P is the water pressure (MPa), t is the test time (min), P0 is the water head of model board bottom (MPa), or the hydrostatic
pressure of the immersed tunnel board bottom in engineering practice. Eq. (1) indicate that the tendency of the water pressure in the central area of model bottom (crater) presents linear grow with volatility. However, the one shown in Fig. 11(b–d) are different in each direction, and could not be fitted. According to Fig. 11, the peaks of water pressure at different radiuses all appear at the end of the test, among which the maximum value 0.015 2 MPa appears in the central area of model board bottom. After the formation of sand deposit and semi-closed space, the depth of crater is essentially invariant, namely the positive fluid boundary condition of the flow under the injection opening keeps constant. At present, the gap and chutes, as the passageway, turned into the side boundary condition of flow, which determined frictional head loss of fluid directly. According to the observation, the length of chutes increased and the plane distributed sparsely with the expansion of sand deposit. Therefore, the fluid resistance increased continuously, so does the water pressure. The total fluid resistance under injection opening was less than output pressure of sand pump in the test, that is to say, the limit state of sand pump did not reached. (4) A comparison with each group of curves in Fig. 11 indicates that water pressures in the central area of model bottom at all directions are the same while different at the same radius as the radius increases. Combined with the sand-deposit expansion process, it shows that when sand deposit was relatively small, the sand–water flow jetted out evenly through the large-area gap, influenced little by model boundary condition and sand-deposit size. Therefore, the fluid pressure is small and consistent. While, when the sand deposit is relatively large, the chutes, as the main channel of the flow, distributed unevenly and randomly and continuously redirected. Accordingly, water pressure rose and distributed unevenly at each direction: The pressure was relatively high in chute areas while relatively low without chutes.
108
W. Li et al. / Tunnelling and Underground Space Technology 40 (2014) 102–108
(4) Water pressure in the crater linearly increases within the sand-deposit radius, and the maximum value is 0.0152 MPa. Given the actual water depth in the river, the water pressure in the crater could be used as one of the bases for construction control.
Acknowledgments
Fig. 12. Relation curve of water pressure in crater and sand-deposit radius.
6.3. Inspiration for engineering practice So far, there are no other safe and rapid methods for determination of sand-deposit size except diver’s underwater exploration in engineering practice (Pan et al., 2004). According to the characteristics of water pressure in test, the areas where sand deposit expanded to are accompanied with the water pressure of a rise and little range volatility. Inspired by this, water pressure gauges could be installed at the bottom of tunnel for monitoring water pressure change in construction, which could be used to judge whether or not sand deposit has expanded to designed radius. The radius of sand deposit in different time could be got in Fig. 10 and the water pressure in the crater, corresponding to the times of the sand deposit expanding to different radius (Fig. 10), could be got in Fig. 11(a). The linear fitting curve between water pressure in the crater and different radius of sand deposit can be drawn based on the aforementioned data, as shown in Fig. 12. Given the actual depth condition of river or sea water, the water pressure value in the crater with the designed radius of sand deposit could be predicted concerning to this law in construction, which could be used as one of the bases for construction control. 7. Conclusion (1) Modeling principles for the full-scale model test of sand flow were put forward. A sand-flow test model, consisting of model system, equipment system and measurement system, was built, which could provide a reference for further studies. (2) A semi-closed space was formed between the model board and the expanding sand deposit, in the place where the water pressure rose with little range volatility. The water pressure at different place of trench could be monitored in construction, which could be used to judge whether or not sand deposit has expanded to the corresponding place. (3) When the sand deposit is relatively small, the water pressures at the same radius are consistent and the sand deposit expands evenly at each direction. Otherwise, the water pressures rise and distribute unevenly and the sand deposit expands unbalanced at each direction.
The authors acknowledge the financial support provided by the National Natural Science Foundation of China (51108190), Foundation of Independence Research Subject (No. 2012ZC27) from State Key Laboratory of Subtropical Building Science in China and Foundation of Research Subject (GTCC 2008-253) from Guangzhou City, China. The authors are also deeply grateful to Guangzhou Central Transportation Construction Co., Ltd., Guangzhou Municipal Engineering Design & Research Institute, CCCC Fourth Harbor Engineering Co., Ltd. and Guangzhou Salvage Bureau of the Ministry of Transport, in particular, Yuan Wei-yao, Huang Kai and Huang Wen-feng for their assistance. References Chen, S.Z., Chen, Y., Zhang, M., 2002. Immersed Tube Tunnel Design and Construction. Science Press, Beijing (in Chinese). Fang, Y.G., Li, W., Mo, H.H., et al., 2012. Experiment and analysis of law of sand deposit expansion in foundation of immersed tube tunnel treated by sand flow method. Chin. J. Rock Mech. Eng. 31 (1), 206–216 (in Chinese). Glerum, A., 1995. Developments in immersed tunnelling in Holland. Tunn. Undergr. Sp. Technol. 10 (4), 455–462. Grantz, W.C., 1997. Steel-shell immersed tunnels – forty years of experience. Tunn. Undergr. Sp. Technol. 12 (1), 23–31. Grantz, W.C., 2001a. Immersed tunnel settlements – Parts 1: Nature of settlements. Tunn. Undergr. Sp. Technol. 16 (3), 195–201. Grantz, W.C., 2001b. Immersed tunnel settlements – Parts 2: Case histories. Tunn. Undergr. Sp. Technol. 16 (3), 203–210. Guangzhou Municipal Engineering Design & Research Institute (GMED&RI), Maunsell Consultants Asia Limited (MCAL), 2008. Construction documents design of Guangzhou Zhoutouzui tunnel. Guangzhou. (in Chinese). Janssen, W., Haas, P.D., Yoon, Y.H., 2006. Busan–Geoje link: immersed tunnel opening new horizons. Tunn. Undergr. Sp. Technol. 21 (3–4), 332. Kiyomiya, O., 1995. Earthquake-resistant design features of immersed tunnels in Japan. Tunn. Undergr. Sp. Technol. 10 (4), 463–475. Li, Z.J., 2001. Study and research of sand filled foundation for underwater tunnel crossing Pearl River. Chn. Harb. Eng. (1), 18–20 (in Chinese). Pan, Y.R., Peng, J., Naotake, Saito, 2004. Technology of sand injection for the foundation of immersed tube elements on the external ring of shanghai. Mod. Tunn. Technol. 41 (1), 41–45 (in Chinese). Rasmussen, N.S., 1997. Concrete immersed tunnels – forty years experience. Tunn. Undergr. Sp. Technol. 12 (1), 33–46. Tongeren, Ir.H.Van, 1978. The foundation of immersed tunnels. In: Delta Tunnelling Symposium. Netherlands, Amsterdam, pp. 48–57. Wang, G.H., Li, Z.G., Cheng, X.M., et al., 2009. Sand flowing experiment and experiment result analysis: case study on shengwudao–daxuecheng immersed tunnel. Tunn. Constr. 29 (2), 176–180 (in Chinese). Yang, W.W., Mao, R., Zeng, C.J., 2008. Development of the Hong Kong undersea immersed tube tunnel project. Mod. Tunn. Technol. 45 (S1), 41–46.
Contents lists available at ScienceDirect
Tunnelling and Underground Space Technology journal homepage: www.elsevier.com/locate/tust
Model test of immersed tube tunnel foundation treated by sand-flow method Wei Li a,b, Yingguang Fang a, Haihong Mo a, Renguo Gu a, Junsheng Chen a,c,⇑, Yizhao Wang a, Deluan Feng a a
School of Civil Engineering & Transportation, South China University of Technology, Guangzhou 510640, Guangdong, PR China Guangzhou Electric Power Design Institute, Guangzhou 510075, Guangdong, PR China c State Key Laboratory of Subtropical Building Science, Guangzhou 510641, Guangdong, PR China b
a r t i c l e
i n f o
Article history: Received 30 July 2012 Received in revised form 7 August 2013 Accepted 28 September 2013 Available online 20 October 2013 Keywords: Immersed tube tunnel Foundation treatment Sand-flow method Model test Sand deposit Water pressure
a b s t r a c t In order to explore the immersed tunnel foundation treated by sand-flow method, modeling principles for the full-scale model test of sand flow were put forward. In addition, a sand-flow test model was built, which consisted of model system, equipment system and measurement system. The situation of sanddeposit expanding and the water pressure in the foundation trench were evaluated through the model test. The results show that a semi-closed space was formed between the model board and the expanding sand deposit, which made the water pressure in it rising with little range of volatility. The sand deposit gradually became non-circular truncated cone which shaped with its expanding radius, and the difference of the water pressure increased at each direction. The water pressure in crater had a linear increase with the sand-deposit radius, with a maximal value of 0.015 2 MPa. The volatility of the water pressure under the tunnel board and the water pressure value in the crater could be used as bases of construction control. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction The main underwater tunnel methods for crossing rivers and straits are mining methods, shield methods and immersed tube methods. Since 1910, the advent of the first true sense immersed tube tunnel-the Detroit River immersed tube tunnel built, immersed tube tunnels have been developed in USA (Grantz, 1997), Europe (Glerum, 1995; Rasmussen, 1997) and Asia (Kiyomiya, 1995; Janssen et al., 2006; Yang et al., 2008) vigorously. Its advantages are obviously evident. There are several factors affecting the settlement of the immersed tube tunnel, such as sub-soil conditions, foundation, siltation, surcharge, and trench dredging methods (Grantz, 2001a), but it is difficult to accurately predict and control the final settlement. Grantz (2001b) shared his immersed tunnel experience with several typical tunnel types and actual settlement records, which shows that the immersed tube tunnel foundation and the treated methods of it have a significant influence on the final settlement. Considered as one of the treatment methods of immersed tunnel foundation used nowadays, sand-flow method (Glerum, 1995;
⇑ Corresponding author at: School of Civil Engineering & Transportation, South China University of Technology, Guangzhou 510640, Guangdong, PR China. Tel.: +86 020 22236638. E-mail address: [email protected] (J. Chen). 0886-7798/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tust.2013.09.015
Chen et al., 2002), with a number of advantages, is one of the mainstream construction methods. Sand-flow method was first used in Vlake Tunnel of Holland (Tongeren et al., 1978), corresponding model test was carried out and the mechanism of the method and construction parameter were explored before construction. Water pressure near the injection opening and sand content in the sand-flow mixture were recorded in the construction. Though immersed tunnel construction started relatively late in China, a large number of sand-flow model tests were taken. By the Guangzhou Pearl River Tunnel, Li (2001) built a sand-flow test model and pointed out that the pressure near the injection opening and buoyancy force increased with the expansion of sand deposit, providing a technical basis for the construction. Before the construction of the Shengwudao–Daxuecheng Tunnel, Wang et al. (2009) studied the sand-deposit radius, void ratio and pressure by the 1:5 reduced scale model test, verified the original design and construction program. The authors take the Guagnzhou Zhoutouzui Tunnel (GMED&RI and MCAL, 2008) in China as the background, and put forward modeling principles for the full-scale model test of sand flow. Sand-flow test model, consisting of model system, equipment system and measurement system, was built. The relationship of sand deposit and time, as well as the distribution law of water pressure in the foundation trench was gained. Furthermore, forecasting methods of the sand-deposit radius and construction-control index
W. Li et al. / Tunnelling and Underground Space Technology 40 (2014) 102–108
for engineering practice were proposed. The authors hope that this paper could provide basis for the sand flow method and the test model could provide a reference for further studies. 2. Model system 2.1. Zhoutouzui immersed tunnel project The length of Guangzhou Zhoutouzui immersed tunnel (GMED&RI and MCAL, 2008) is 340 m. The designed pipe segments of the tunnel in Pearl River are E1 (85 m), E2 (85 m), E3 (79.5 m), E4 (90.5 m), of which the E1 and E4 are variable cross-section segments. The tunnel foundation will be treated by sand-flow method. The typical pipe segment and the designed sand deposit are shown in Fig. 1. 2.2. Modeling principles for immersed tunnel In order to guarantee the rationality and validity of the model test, the test model was designed in accordance with the following principles: (1) The plane shape of the model board was equivalent to the floor of the tunnel in a full-scale sand-deposit area, which simulated the top boundary condition of the fluid in construction. (2) The space between the model board edge and the tank wall was similar to the one between the lateral wall of immersed tunnel and the trench slope, which was equivalent to the lateral boundary condition of the fluid. (3) By plugging the side of the trench, the various ‘‘artificial boundary plugging’’ caused by the sand deposit already existed at the one side or both sides of the constructing sand deposit could be simulated. (4) The model board was immersed in water to simulate the underwater construction condition. The only difference happened between the depth of tested water and that of the
Fig. 1. Plane and cross-section of Guangzhou Zhoutouzui immersed tube tunnel/ cm.
103
actual construction water, which would generate no significant differences in the test and had been verified by the preliminary experiment (Fang et al., 2012). (5) The performance parameters of the test equipment system were fully consistent with those of the construction equipment, which could provide stable sand–water flow as those in construction. (6) Measuring system worked accurately and continuously without influencing the test. (7) Test sand had the same category and the similar particle size distribution with the construction sand material. 2.3. Model system of immersed tunnel According to the modeling principles above, a variable crosssection pipe segment in a sand-deposit range was modeled and a full-scale test model for sand-flow method was built. The simulated area is shown by coarse dotted line in Fig. 1(a). The model system of immersed tunnel consists of a model board, a experimental tank, a circulation tank and a underground reservoir, as shown in Fig. 2. The model board is a reinforced concrete, with a dimension of 8.2 m 15.2 m (length width), which was used to full-scale simulate the floor of the immersed tunnel in a sand-deposit range. The clearance between the model board and the ground is 1.0 m, which simulated the clearance of foundation trench. An injection opening was simulated by a 6-in. steel nozzle, which was installed on the board centre, connecting the sand pump by steel-wired hose. Opening–closing holes for observation and sampling were installed at the centerline of the board.
Fig. 2. Test model system. Note: The E, S, W, N in the paper expressed the East, South, West and North direction of the model board.
104
W. Li et al. / Tunnelling and Underground Space Technology 40 (2014) 102–108
The dimension of the experimental tank is 10.5 m 17.2 m 2.5 m (width length height). The model board was placed in the tank, with a clearance of 0.8–1.3 m between it and the tank wall. In the test, the tank was filled with water and the model board was under the water surface, which simulated the underwater condition. Weep holes and overflow holes were installed on the N side of the tank wall, while water-resistant steel door on the W side. The dimension of circulation tank is 4.0 m 3.0 m 2.0 m (length width height), one of its sides was connected to experimental tank, while the other side was linked to water pump through steel-wired hose. As a result, buffering fluid and re-circulating water were achieved. The 400 m3 – capacity underground reservoir was located at the bottom of the test site, which was used to provide the testing water to the tank. 3. Equipment system 3.1. Test equipment system The equipment system of sand-flow test consists of four components: water conveyance part, sand conveyance part, pressure injection part and controller part. All the equipment were selfdeveloped and the performance parameters could be used directly in construction for equipment selection. The water conveyance part composed of water pump, butterfly valve, return valve and 6-in. steel-wired hoses. It was used to sup-
ply quantitative water to sand–water mixer. The sand conveyance part comprised shovel car, first conveyor belt, Sand storage device, sand distributor and second conveyor belt. It was used to supply quantitative sand to sand–water mixer. The pressure injection part was made of sand–water mixer, sand pump, return valve and 6-in. steel-wired hoses. It was used to inject the sand–water mixture to the bottom of model board. The controller part, also known as control electric cabinet, was used to control all the equipment. Consequently, the unified control to equipment and the monitor to electric current and voltage of the system were achieved. The test equipment system refers to Fig. 3 and the photos of main test equipment are shown in Fig. 4. 3.2. Sand–water ratio control The sand and water conveyance rates were determined before the test. The sand distributor was installed under the sand storage device. The sand flow rate was controlled by adjusting the distributor and the water flow rate was controlled by adjusting the return valve and butterfly valve of the water pump. The flow of water pump and sand distributor were adjusted to the pre-set sand–water ratio and to the appropriate rate (i.e., the total sand–water input of the equipment system matched the output of the sand pump per unit time, and the mixer did not overflow or adsorbed dry) in the test. Therefore, steady sand–water flow was guaranteed. 4. Measurement system 4.1. Water pressure measurement Hydro-dynamic gauges (0–0.2 Mpa) were installed in the trench and listed in Fig. 5 and Table 1. The measurement time interval was 5 min and the maximal value was recorded whenever it occurred. 4.2. Sand-deposit measurement
Fig. 3. Layout of equipment system. Note: 1 – First conveyor belt; 2 – sand storage device; 3 – sand distributor; 4 – second conveyor belt; 5 – mixer; 6 – sand pump; 7 – water pump; 8 – control cabinet.
The process of sand-deposit expansion was measured by sanddeposit detectors (Fang et al., 2012), which were installed on the
Fig. 4. Photos of main equipment.
W. Li et al. / Tunnelling and Underground Space Technology 40 (2014) 102–108
105
unchanged with time under a certain measuring point, it could be considered that the sand-deposit radius expanded to that measuring point at that time, and the time–history curves of sand-deposit radius on each trend could be drawn by connecting the time data of the sand deposit expanded to different radius. 5. Test scheme and procedures
Fig. 5. Layout of hydro-dynamic gauges in trench. Note: The serial number indicates the hydro-dynamic gauges located at the underside of model board (the top surface of sand-deposit).
Table 1 Serial number and position of hydro-dynamic gauges. Radius Direction
0.5 m
2.0 m
3.5 m
6.5 m
E S W N
1056# — 250# —
1041# 244# 1064# 1048#
1055# 247# 1036# 1029#
256# — 1058# —
According to the design of Guangzhou Zhoutouzui immersed tube tunnel (GMED&RI and MCAL, 2008), the sand–water ratio, namely mass ratio, was set as 1:8 and the trench clearance was 1.0 m. The model boundary was widely open. The model board was made by fair-faced concrete and placed at a depth of 0.9 m under water. In the test, sand was drawn into sand–water mixer at the preset flow through first conveyor belt, sand storage device, sand distributor and second conveyor belt. Water was pumped into sand– water mixer from circulation tank by water pump at the pre-set flow. After being fully mixed in the mixer, the sand–water mixture was drawn and pumped into the bottom of model board by sand pump. The sediment of sand particles formed sand deposit and the water entered into circulation tank. 6. Test results and analysis 6.1. Sand-deposit expansion The typical time–history curves of sand-deposit height on H & M axes are illustrated in Fig. 7, the law on other axes is similar to these figures. Typical sand-deposit profiles on H–M axis are obtained by connecting the data point of sand-deposit height of H & M axes, which measured at the same time, shown in Fig. 8. Fig. 7 shows that, at the beginning of the test, sand particles of the mixture deposit in the trench where it is close to the injection opening and expand gradually in radial direction. Clearance is likely to exist between the model board and the sand-deposit sur-
Fig. 6. Layout of sand-deposit detectors.
Table 2 Number and position of sand-deposit detectors/m. Measuring point Axis
1
2
3
4
5
6
7
8
F H L K M
0.9 2.0 2.0 1.0 1.0
2.0 3.1 3.0 2.0 2.0
3.0 4.1 4.0 3.0 3.0
3.8 5.1 5.0 3.8 4.0
4.8 6.1 — — 5.0
5.7 7.1 — — 6.0
6.5 8.0 — — 7.1
7.5 — — — 8.0
model board, shown in Fig. 6 and Table 2. The measurement time interval was 3 min. The real-time sand-deposit height data of the place of each measuring point (shown in Fig. 6) were obtained by detectors directly. The time–history curves of sand-deposit height on each axis could be got by connecting the above mentioned data in the whole experiment process. When the sand-deposit height remained
Fig. 7. Time–history curves of sand-deposit height on H & M axes.
106
W. Li et al. / Tunnelling and Underground Space Technology 40 (2014) 102–108
Fig. 8. Sketch map of sand-deposit profiles.
face in the center area of the board, near the injection opening, while the sand deposit could fully filled upon the trench in the surrounding area. Based on Fig. 8, it is clear that annular sand dam and circular plash are gradually formed around the injection opening at the early stage of test. According to Fig. 7, it is obvious that the height of the sand dam (as shown in Fig. 8) will no longer change when it expands to about 95 cm height. Till then, the sand dam has expanded to a sand deposit (also known as ‘‘pancake’’) with a relatively stable height, and the plash has turned into a crater with a constant depth. Till now, a semi-closed space under injection opening has been formed. The clearance on the top of sand deposit shown in Figs. 7 and 8 is a gap, which was formed in the process of test and existed in the central area of sand deposit top, around the injection opening. Besides, chutes also formed in the outer area of the gap area, with limited number and irregular channel-like distribution, which are the reasons why the chutes cannot be detected by detectors. Typical sand-deposit profiles are shown in Fig. 9. The real-time data of sand-deposit radius at each direction and the fitting
average expansion trend, described by the dotted line are shown in Fig. 10. It is shown from the features of sand-deposit profiles in Fig. 9 that the sand–water flow which injected into the crater emitted out through the gap and chutes at high pressure and speed. Then, the pressure and the speed lowered down and the sand particles settled, forming oblique layered texture structure. The sand deposit expands in this manner. The dip angle of the texture is 27–31°, which is similar to the repose angle of the test sand in water. Fig. 10 shows that the data points of radius expanding in all directions almost overlap when sand-deposit radius is relatively small. In this stage, sand–water flow jetting out through gap and the sand particles has a continuous deposit on the outer margin surface of the sand deposit, with the sand deposit expanding at the same speed in all directions and forming truncated cone. This is verified by the smooth and continuous curves measured by Point 1 to 3 shown in Fig. 7. Fig. 10 also shows that the discreteness of the radius data points increased when the sand deposit expanded to a large radius. Sand– water flow was conveyed to the outer margin by chute in this stage. The uneven distribute of chute, and the properties of blocking and constant detouring make sand particles convey to each direction discontinuously and unequally. Thus, expanding speeds of sand deposit are significantly different at each direction and form an irregular shape. Furthermore, the larger the sand-deposit radius is, the more uneven the chute distributes and the more unbalanced the sand deposit expands. This is verified by the step shape curves measured by Point 4 to 8 shown in Fig. 7. 6.2. Water pressure distribution
Fig. 9. Photo of typical sand-deposit profile.
Fig. 10. Time–history curves of sand-deposit radius on each trend.
The time–history curves of water pressure in the trench are illustrated in Fig. 11. The times of sand deposit expands to different radius and corresponding times of water pressure changes are shown in Table 3. (1) Fig. 11 shows that at the initial stage of test, curve slope and increment in Fig. 11(a) are significantly larger than the one in Fig. 11(b–d). Combined with the observation of the test water level, it is obvious that the curve slopes and increments in Fig. 11(b–d) almost conformed, which is just caused by the slight rose of the static water level. These show that at the initial stage of test, aside from the influence of the static water lever, the impact of sand–water flow on the water pressure is great, in the central area of the model bottom but little on the one which in the peripheral area. Combined with the analysis of segment 6.1, the fluid boundary condition under injection opening changed with the forming process of crater and sand-deposit, which means that the trench clearance decreases gradually and the sand–water flow pumped in the crater ran over more and more difficultly. Thus, the water pressure increased in the central area of model board bottom. On the contrary, sand particles have not deposited in the peripheral area of model board bottom, and the trench clearance (i.e., the boundary
W. Li et al. / Tunnelling and Underground Space Technology 40 (2014) 102–108
107
Fig. 11. Time–history curves of water pressure in the trench.
Table 3 Timetable for sand-deposit reaches different radius and water pressure changes/min. Radius
2.0 m
3.5 m
6.5 m
Times of sand-deposit arriving Times of water-pressure increasing
77 85
180 195
422 435
condition) has not changed yet. Therefore, water pressure in this place is not affected. (2) Table 3 shows that the times that sand deposit expanded to different radius are generally correspond to those when water pressure started rising. It is shown in Figs. 8 and 11 that, where the sand deposit expanded, the water pressure at corresponding position started rising with little range volatility, otherwise, there were no significant changes on it. The sand deposit and crater formed a semi-closed space with the model board, gap and chutes that became the passageway of sand–water flow, causing the increase of flow resistance and the water pressure within the range of the formed sand deposit. Water pressure in the range of sand deposit, which rose instantly with the gap and chutes, were blocked by sand particles in the process. Then the weak places of sand deposit were washed out and the chutes rechanneled, leading to a direct result of leakage of the sand–water flow and reducing the water pressure. Again and again, water pressure presented increasing in little range volatility. (3) Fig. 11(a) shows that, in the middle and latter stage of the test, the average water pressure in the central area of model bottom can be linear fitted as
P ¼ 1:33 105 t þ P0
ð1Þ
where P is the water pressure (MPa), t is the test time (min), P0 is the water head of model board bottom (MPa), or the hydrostatic
pressure of the immersed tunnel board bottom in engineering practice. Eq. (1) indicate that the tendency of the water pressure in the central area of model bottom (crater) presents linear grow with volatility. However, the one shown in Fig. 11(b–d) are different in each direction, and could not be fitted. According to Fig. 11, the peaks of water pressure at different radiuses all appear at the end of the test, among which the maximum value 0.015 2 MPa appears in the central area of model board bottom. After the formation of sand deposit and semi-closed space, the depth of crater is essentially invariant, namely the positive fluid boundary condition of the flow under the injection opening keeps constant. At present, the gap and chutes, as the passageway, turned into the side boundary condition of flow, which determined frictional head loss of fluid directly. According to the observation, the length of chutes increased and the plane distributed sparsely with the expansion of sand deposit. Therefore, the fluid resistance increased continuously, so does the water pressure. The total fluid resistance under injection opening was less than output pressure of sand pump in the test, that is to say, the limit state of sand pump did not reached. (4) A comparison with each group of curves in Fig. 11 indicates that water pressures in the central area of model bottom at all directions are the same while different at the same radius as the radius increases. Combined with the sand-deposit expansion process, it shows that when sand deposit was relatively small, the sand–water flow jetted out evenly through the large-area gap, influenced little by model boundary condition and sand-deposit size. Therefore, the fluid pressure is small and consistent. While, when the sand deposit is relatively large, the chutes, as the main channel of the flow, distributed unevenly and randomly and continuously redirected. Accordingly, water pressure rose and distributed unevenly at each direction: The pressure was relatively high in chute areas while relatively low without chutes.
108
W. Li et al. / Tunnelling and Underground Space Technology 40 (2014) 102–108
(4) Water pressure in the crater linearly increases within the sand-deposit radius, and the maximum value is 0.0152 MPa. Given the actual water depth in the river, the water pressure in the crater could be used as one of the bases for construction control.
Acknowledgments
Fig. 12. Relation curve of water pressure in crater and sand-deposit radius.
6.3. Inspiration for engineering practice So far, there are no other safe and rapid methods for determination of sand-deposit size except diver’s underwater exploration in engineering practice (Pan et al., 2004). According to the characteristics of water pressure in test, the areas where sand deposit expanded to are accompanied with the water pressure of a rise and little range volatility. Inspired by this, water pressure gauges could be installed at the bottom of tunnel for monitoring water pressure change in construction, which could be used to judge whether or not sand deposit has expanded to designed radius. The radius of sand deposit in different time could be got in Fig. 10 and the water pressure in the crater, corresponding to the times of the sand deposit expanding to different radius (Fig. 10), could be got in Fig. 11(a). The linear fitting curve between water pressure in the crater and different radius of sand deposit can be drawn based on the aforementioned data, as shown in Fig. 12. Given the actual depth condition of river or sea water, the water pressure value in the crater with the designed radius of sand deposit could be predicted concerning to this law in construction, which could be used as one of the bases for construction control. 7. Conclusion (1) Modeling principles for the full-scale model test of sand flow were put forward. A sand-flow test model, consisting of model system, equipment system and measurement system, was built, which could provide a reference for further studies. (2) A semi-closed space was formed between the model board and the expanding sand deposit, in the place where the water pressure rose with little range volatility. The water pressure at different place of trench could be monitored in construction, which could be used to judge whether or not sand deposit has expanded to the corresponding place. (3) When the sand deposit is relatively small, the water pressures at the same radius are consistent and the sand deposit expands evenly at each direction. Otherwise, the water pressures rise and distribute unevenly and the sand deposit expands unbalanced at each direction.
The authors acknowledge the financial support provided by the National Natural Science Foundation of China (51108190), Foundation of Independence Research Subject (No. 2012ZC27) from State Key Laboratory of Subtropical Building Science in China and Foundation of Research Subject (GTCC 2008-253) from Guangzhou City, China. The authors are also deeply grateful to Guangzhou Central Transportation Construction Co., Ltd., Guangzhou Municipal Engineering Design & Research Institute, CCCC Fourth Harbor Engineering Co., Ltd. and Guangzhou Salvage Bureau of the Ministry of Transport, in particular, Yuan Wei-yao, Huang Kai and Huang Wen-feng for their assistance. References Chen, S.Z., Chen, Y., Zhang, M., 2002. Immersed Tube Tunnel Design and Construction. Science Press, Beijing (in Chinese). Fang, Y.G., Li, W., Mo, H.H., et al., 2012. Experiment and analysis of law of sand deposit expansion in foundation of immersed tube tunnel treated by sand flow method. Chin. J. Rock Mech. Eng. 31 (1), 206–216 (in Chinese). Glerum, A., 1995. Developments in immersed tunnelling in Holland. Tunn. Undergr. Sp. Technol. 10 (4), 455–462. Grantz, W.C., 1997. Steel-shell immersed tunnels – forty years of experience. Tunn. Undergr. Sp. Technol. 12 (1), 23–31. Grantz, W.C., 2001a. Immersed tunnel settlements – Parts 1: Nature of settlements. Tunn. Undergr. Sp. Technol. 16 (3), 195–201. Grantz, W.C., 2001b. Immersed tunnel settlements – Parts 2: Case histories. Tunn. Undergr. Sp. Technol. 16 (3), 203–210. Guangzhou Municipal Engineering Design & Research Institute (GMED&RI), Maunsell Consultants Asia Limited (MCAL), 2008. Construction documents design of Guangzhou Zhoutouzui tunnel. Guangzhou. (in Chinese). Janssen, W., Haas, P.D., Yoon, Y.H., 2006. Busan–Geoje link: immersed tunnel opening new horizons. Tunn. Undergr. Sp. Technol. 21 (3–4), 332. Kiyomiya, O., 1995. Earthquake-resistant design features of immersed tunnels in Japan. Tunn. Undergr. Sp. Technol. 10 (4), 463–475. Li, Z.J., 2001. Study and research of sand filled foundation for underwater tunnel crossing Pearl River. Chn. Harb. Eng. (1), 18–20 (in Chinese). Pan, Y.R., Peng, J., Naotake, Saito, 2004. Technology of sand injection for the foundation of immersed tube elements on the external ring of shanghai. Mod. Tunn. Technol. 41 (1), 41–45 (in Chinese). Rasmussen, N.S., 1997. Concrete immersed tunnels – forty years experience. Tunn. Undergr. Sp. Technol. 12 (1), 33–46. Tongeren, Ir.H.Van, 1978. The foundation of immersed tunnels. In: Delta Tunnelling Symposium. Netherlands, Amsterdam, pp. 48–57. Wang, G.H., Li, Z.G., Cheng, X.M., et al., 2009. Sand flowing experiment and experiment result analysis: case study on shengwudao–daxuecheng immersed tunnel. Tunn. Constr. 29 (2), 176–180 (in Chinese). Yang, W.W., Mao, R., Zeng, C.J., 2008. Development of the Hong Kong undersea immersed tube tunnel project. Mod. Tunn. Technol. 45 (S1), 41–46.