Centrifuge investigation into the effect of new shield tunnelling on an existing underlying large-diameter tunnel

Tunnelling and Underground Space Technology 42 (2014) 59–66

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology journal homepage: www.elsevier.com/locate/tust

Centrifuge investigation into the effect of new shield tunnelling on an existing underlying large-diameter tunnel Peng Li a, Shou-Ji Du b,⇑, Xian-Feng Ma c, Zhen-Yu Yin d, Shui-Long Shen e,1 a

Department of Civil Engineering, School of Naval Architecture, Ocean and Civil Engineering, Shanghai Jiao Tong University, Shanghai 200240, China Department of Civil Engineering, School of Naval Architecture, Ocean and Civil Engineering, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai 200240, China c Department of Geotechnical Engineering, Key Laboratory of Geotechnical and Underground Engineering of Ministry of Education, Tongji University, Shanghai 200092, China d Department of Civil Engineering, Shanghai Jiao Tong University, 800 Dongchuan Road, Minhang District, Shanghai 200240, China e Department of Civil Engineering, Shanghai Jiao Tong University and State Key Laboratory of Ocean Engineering, 800 Dong Chuan Road, Minhang District, Shanghai 200240, China b

a r t i c l e

i n f o

Article history: Received 6 July 2013 Received in revised form 11 January 2014 Accepted 2 February 2014 Available online 25 February 2014 Keywords: Shield tunnelling River-crossing tunnel Centrifuge model test

a b s t r a c t A centrifuge model test was carried out to investigate the effect of new shield tunnelling on an existing underlying large-diameter tunnel. Three construction steps of a new shield tunnel were simulated in the test. Soil excavation with ground loss and grouting in each step was simulated by discharge and injection of a dense solution. The vertical displacement and the longitudinal stress of the existing tunnel, as well as the pore water pressure at its spring line, were measured. The vertical displacement and the longitudinal stress of the existing tunnel increase as a result of the excavation of the new shield tunnel. However, the vertical displacement and the longitudinal stress decrease when grouting is injected for the new tunnel. The vertical displacement and the longitudinal stress exhibit an approximately linear change with increases in ground loss ratio and grouting ratio of the new shield tunnel. In the test, the heave zone of the existing tunnel is within 39 m of the axis of the new shield tunnel. The variation in pore water pressure at the spring line of the existing tunnel is low during construction of the new shield tunnel. Grouting is an effective measure to mitigate the responses of the existing tunnel. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction In congested urban areas of Shanghai, more and more subway tunnels and river-crossing tunnels have been constructed to meet the needs of increasing transportation (Shen et al., 2009, 2010, 2014; Wu et al., 2014). Consequently, new tunnels are inevitably constructed across existing tunnels (Klar et al., 2008; Li and Yuan, 2012; Liao et al., 2009; Marshall et al., 2010). In order to maintain services and to ensure the safety of existing tunnels in very soft clays (Chang et al., 2001a, b; Yin et al., 2009, 2010, 2011; Shen et al., 2013a, b), it is necessary to investigate the influence of new tunnelling on adjacent existing tunnels, particularly aging tunnels in a complicated geological environment. There exist many aging river-crossing tunnels in Shanghai. As a typical case, this paper considers a scenario, in which a new subway tunnel is constructed perpendicularly over an existing river-crossing tunnel.

⇑ Corresponding author. Tel.: +86 21 3420 6334; fax: +86 21 3420 6197. E-mail addresses: [email protected] (P. Li), [email protected] (S.-J. Du), [email protected] (X.-F. Ma), [email protected] (Z.-Y. Yin), slshen@sjtu. edu.cn (S.-L. Shen). 1 Tel.: +86 21 3420 4301; fax: +86 21 6419 1030. http://dx.doi.org/10.1016/j.tust.2014.02.004 0886-7798/Ó 2014 Elsevier Ltd. All rights reserved.

The river-crossing tunnel has a diameter of 14.5 m. Fig. 1 illustrates this situation. In recent decades, field observations and theoretical analyses have been performed to investigate the effects of overlying excavation on existing tunnels (Hu et al., 2003; Huang et al., 2006; Sharma et al., 2001; Zhang et al., 2013). Numerical analyses have also been conducted to estimate the effects of overlying excavation on existing tunnels (Dolezalova, 2001; Jiang and Yin, 2012; Li and Du, 2012a, b; Liu et al., 2009, 2011). However most of these studies did not investigate the effect of overlying tunnelling on existing large-diameter tunnels in detail. In a geotechnical centrifuge test, the stress state of a model is the same as that in the prototype. Moreover, a geotechnical centrifuge test is highly reliable, controllable and stable. Thus, centrifuge model testing in shield tunnel modelling has developed in recent decades (Meguid et al., 2008; Ng et al., 2013; Nomoto et al., 1999). Centrifuge model testing is a suitable approach, which can not only simulate the interaction between tunnels and soil, but can also be used to carry out quantitative research on relevant parameters, to simulate tunnelling over an existing large-diameter tunnel. However, there is very limited case study into centrifuge

P. Li et al. / Tunnelling and Underground Space Technology 42 (2014) 59–66

700 145 50 62 138 50

60

2.1. Model configuration and materials The geotechnical centrifuge used in the test is located at Tongji University in China, and has an effective radius of 3 m, a maximum payload of 2 tonnes (at 75 g), and a maximum centrifugal acceleration of 200 g. The geotechnical centrifuge is composed of a power system, a rotational system, a data acquisition system, a control system, and a video surveillance system. The centrifuge can keep running for 24 h. The strong box designed for the centrifuge has an internal plan area of 0.7 m by 0.9 m and an internal height of 0.7 m. The test reported in this paper was carried out at an acceleration level of 100 g. Consequently, the dimensions of the prototype were scaled down by a factor of 0.01, while the stress and strain were scaled by a factor of 1. The prototype of the existing tunnel referred to the Jungong Road River-crossing Tunnel in Shanghai, which has an external diameter of 14.5 m and an internal diameter of 13.3 m. The maximum longitudinal slope of this tunnel is 4.5%. It passes through about 6 soil layers in the longitudinal direction. The soils around it are mainly soft silty clay. This type of soil has a high water content, low shear strength, high compressibility, and high sensitivity. Therefore, it is important to reduce the disturbance to the surrounding soil of this tunnel when a new tunnel is constructed. In order to avoid leakage of water, low deformation and small displacement of this tunnel are strict requirements, as this tunnel is beneath the Huangpu River and the surrounding soil is in a saturated state. The structure of the new shield tunnel was based on the Shanghai metro tunnels, of which the external and internal diameters are 6.2 m and 5.5 m, respectively. The layout and the dimensions of the model are presented in Fig. 2, in which the model unit is used. The model represents a prototype where a new shield tunnel is constructed perpendicular to an existing tunnel with a buried depth of 25 m (from the ground surface to the existing tunnel crown). The vertical clearance from the invert of the new shield tunnel to the crown of the existing tunnel is set at 5 m. The basic characteristic of geotechnical centrifuge testing is that the model can be made using prototype material. The stress state of the model in a centrifuge test is the same as that of the prototype, as long as the centrifugal acceleration is based on the ratio

205

Grey silty clay

50 50

Existing tunnel

Sand layer 50

900

50

Strong box New shield tunnel

175

200

175

Step 1

Step 2

Step 3

Existing tunnel Clap board Space for data wires and pipes

205

700

145 50 62 138 50

(a) Front view

Grey silty clay

50 50

2. Centrifuge model test

Strong box

(Unit: mm)

Fig. 1. Schematic illustration of the problem.

testing performed to investigate the responses of existing tunnels due to overlying tunnelling. In this paper, a centrifuge model test was carried out to investigate the effect of new shield tunnelling on an existing underlying large-diameter tunnel. The construction process of a new shield tunnel, including soil excavation, ground loss, and grouting, was simulated in the test. The vertical displacement and the longitudinal strain of the existing tunnel were measured during the construction process of the new shield tunnel, as well as the pore water pressure at the spring line of the existing tunnel.

New shield tunnel Grey silty clay

Sand layer 50

550

(Unit: mm) 5

145 50

(b) Side view Fig. 2. Configuration of centrifuge model.

Table 1 Soil properties. Modulus of compressibility (MPa) Void ratio Water content (%) Unit weight (kN/m3) Cohesion (kPa) Angle of friction (degrees) Permeability coefficient (cm/s)

3.34 1.2 44 17.8 14 19.2 9  10

8

of prototype dimension and model dimension. Therefore, the grey silty clay on site, in which the existing tunnel is mainly situated, was taken to simulate the ground in the centrifuge test. The mechanical and physical properties of the grey silty clay are given in Table 1. The thickness of the existing tunnel lining is 6 mm after scaling down by a factor of 0.01, so the model tunnel cannot be made from concrete. A hollow plexiglass cylinder, with a thickness of 7.5 mm and a Young’s modulus of 3 GPa, was used to simulate the existing tunnel. The use of this cylinder ensured that the longitudinal bending stiffness of the existing tunnel model was equivalent to that of the existing tunnel prototype.

2.2. Data acquisition The vertical displacement and the longitudinal internal force of the existing tunnel were the main focus of the centrifuge test, since these indexes can perfectly reflect the responses of the existing tunnel due to the overlying shield tunnelling. Eight displacement transducers, arranged in a line at the crown of the existing tunnel model, were used to measure the vertical displacements of the existing tunnel, and the intervals between

P. Li et al. / Tunnelling and Underground Space Technology 42 (2014) 59–66

61

Fig. 3. Layout of measuring instruments.

these transducers are presented in Fig. 3. Strain gauges were pasted at the crown and the invert of the existing tunnel model in a longitudinal direction with an identical interval of 90 mm, as illustrated in Fig. 3. Simultaneously, a pore-water pressure gauge was located at the spring line of the existing tunnel model. 2.3. Experimental procedure The centrifuge model test consisted of four stages, as illustrated in Fig. 4. In the preparation stage, the soil sample, the new shield tunnel model with the simulation system of tunnelling, and the existing tunnel model with the data acquisition system were prepared. In the second stage, the sand, the soil sample, and the model tunnels were installed in the strong box step by step, simulating the stress loading created by the mass of soil. Fig. 5 presents the prepared strong box before it was put into the centrifuge. The third stage was the simulation of the new shield tunnelling, which

is the key stage in the whole centrifuge test. In this stage, the shield tunnelling consisted of three construction steps. Soil excavation, ground loss, and grouting were simulated in each construction step. Details of this stage are given in the following section. In the last stage, the centrifuge was kept running for 9.5 h after the completion of the shield tunnelling. The new shield tunnel model consisted of a hollow aluminium alloy cylinder with four enlarged bands, which was divided into three segments in order to simulate the three construction steps. Each segment was sealed within a latex membrane, and was put into a container before the test. Simultaneously, the containers and the gaps between the latex membrane and the outside of the aluminium alloy cylinder were filled with saturated CaCl2 solution which was used to simulate soil excavation, ground loss and grouting. The ground loss and the grouting were simulated by control the volume of injected or discharged solution. Six solenoid valves were used to control the discharge and the injection of the

Fig. 4. Centrifuge test procedure.

62

P. Li et al. / Tunnelling and Underground Space Technology 42 (2014) 59–66

Displacement transducer

New shield tunnel model Strong box Fig. 5. Prepared strong box.

solution. The magnitude of tunnel volume loss and grouting volume was controlled using a hydraulic jack connected to the piston of a fluid-filled sealed cylinder, which was connected to the new shield tunnel model and had an internal diameter of 30 mm. A given upward unit displacement (e.g. 1 mm) of the hydraulic jack corresponded to a specific volume loss in the new shield tunnel model, and therefore the relationship between hydraulic jack displacement and tunnel volume loss, as well as grouting volume, was established. For example, a ground loss ratio of the new shield tunnel is given, and then the amount of solution volume discharged from the new shield tunnel model can be calculated. The discharged solution will enter the hydraulic cylinder. The internal area of the hydraulic cylinder is a determined value, and then the required upward displacement of the hydraulic jack can be calculated. This method is illustrated in Fig. 6. Based on the statistical analyses of the data of ground loss ratio induced by shield tunnelling, Wei (2010) found that ground loss ratio ranges from 0.35% to 2.02% for cohesive soils in Shanghai. In view of this, the amount of ground loss ratio and grouting ratio were specified to be 1% and 150%, respectively. The construction of the model shield tunnel consisted of 3 steps. Excavation lengths were 175 mm in step 1, 200 mm in step 2, and 175 mm in step 3, which is presented in Fig. 2(b). The simulation procedure is given in detail as follows: (a) Excavation in step 1. Turn on solenoid valve 4 and let the solution in sac 1 discharge. Simultaneously, turn on solenoid valve 1. Give the hydraulic jack upward displacements of 1.6 mm, 3.2 mm, 4.8 mm, 6.4 mm, and 8 mm to simulate ground loss ratios of 0.2%, 0.4%, 0.6%, 0.8%, and 1%, respectively.

Fig. 6. Simulation system of new shield tunnelling.

(b) Grouting in step 1. Give the hydraulic jack downward displacements of 2.4 mm, 4.8 mm, 7.2 mm, 9.6 mm, and 12 mm to simulate grouting ratios of 30%, 60%, 90%, 120%, and 150%, respectively. When the jack achieves a specified displacement of 12 mm, turn off solenoid valves 1 and 4. (c) Excavation in step 2. Turn on solenoid valves 2 and 5. Give the jack upward displacements of 1.8 mm, 3.6 mm, 5.4 mm, 7.2 mm, and 9 mm to simulate ground loss ratios of 0.2%, 0.4%, 0.6%, 0.8%, and 1%, respectively. (d) Grouting in step 2. Give the hydraulic jack downward displacements of 2.7 mm, 5.4 mm, 8.1 mm, 10.8 mm, and 13.5 mm to simulate grouting ratios of 30%, 60%, 90%, 120%, and 150%, respectively. When the jack achieves a specified displacement of 13.5 mm, turn off solenoid valves 2 and 5. (e) Construction in step 3. Repeat (a and b) to simulate construction in step 3. It must be pointed out that solenoid valve 3 was used to control ground loss and grouting, and solenoid valve 6 was used to control soil excavation in this step, although other operations were identical to step 1. Note that the geotechnical centrifuge was operational during the above stages.

3. Results and analysis 3.1. Vertical displacement of existing tunnel The measuring locations of vertical displacement were labelled as D1–D8 from left to right, and the cross section, which was just beneath the axis of the new shield tunnel, was defined as a symmetric plane, as illustrated in Fig. 3. During the test process, due to malfunction of the fifth displacement transducer, displacement data at D5 was not obtained. Because of symmetry of configuration, structure, material, and boundary condition, it was assumed that the displacement at D5 was identical to that at D4. The test results are presented and discussed in terms of prototype units in the following section. Fig. 7 shows the vertical displacement pattern at the crown of the existing tunnel in a longitudinal direction at different stages. Fig. 8 shows the change in vertical displacements at D1–D4 during the construction process of the new shield tunnel. The upward vertical displacement is assumed to be positive, and the downward vertical displacement is assumed to be negative. From the two figures, it can be observed that the maximum vertical displacement and the minimum curvature radius of the vertical displacement curve occur in the symmetric plane during the construction process of the new shield tunnel. With the increase of offset from

Fig. 7. Vertical displacement of existing tunnel.

P. Li et al. / Tunnelling and Underground Space Technology 42 (2014) 59–66

63

of grouting ratio, the vertical displacements decrease. The vertical displacement at D3 has a slight increase at the start of grouting. It is assumed that this slight increase is the result of test error. The vertical displacement at D4 declines from 10.6 mm to 4.4 mm as the grouting ratio increases from 0% to 150%. It is also noted that the change rate of vertical displacement at D4 is significantly faster than those at the other measuring locations as the ground loss ratio and grouting ratio increase. Fig. 10 shows the post construction displacements of the existing tunnel at D1–D8. It is indicated that the vertical displacements of the existing tunnel undergo a comparatively rapid declining process in the first year, and an insignificant decreasing process to reach stable values within the range from 2 mm to 1 mm in the following period. 3.2. Longitudinal internal force of existing tunnel Fig. 8. Vertical displacements at D1-D4 versus construction process of new tunnel.

the symmetric plane, the vertical displacements gradually decline and finally reach stable values. The cumulative vertical displacement can reach about 12 mm in the symmetrical plane. The heave range of the existing tunnel extends at least to the position which has an offset of 39 m from the symmetrical plane. On the other hand, after the completion of excavation at each step, an apparent increment of vertical displacement of the existing tunnel can be observed. Conversely, grouting at each step causes a decline in vertical displacement. This observation implies that grouting can effectively reduce the displacement of the existing tunnel, and sufficient backfilling for the stabilization of the existing tunnel is important. It is also noted that the vertical displacement of the existing tunnel experiences a sequence of slight uplift and settlement, apparent uplift and settlement, and finally slight uplift and settlement in the three construction steps. Furthermore, excavation in step 2 causes the maximum vertical displacement of the existing tunnel, which confirms that construction just over the existing tunnel will lead to the greatest impact on the displacement of the existing tunnel. Attention should be paid to this period in order to maintain the structural safety and normal operation of the existing tunnel. Fig. 9 shows the change in vertical displacements at D1–D4, along with the increases of the ground loss ratio and the grouting ratio in construction step 2. From this figure, the vertical displacements exhibit approximately linear growth as the ground loss ratio increases. Taking as an example the vertical displacement at D4, the value increases from 2.9 mm to 10.6 mm as the ground loss ratio increases from 0% to 1%. Generally speaking, with the increase

The pasted strain gauges were labelled as C1–C9 from left to right at the crown of the existing tunnel model, and labelled as I1–I9 from left to right at the invert of the existing tunnel model, as illustrated in Fig. 3. According to the relationship between stress and strain, the longitudinal stress of the existing tunnel can be obtained. Fig. 11 shows the longitudinal stresses at both the crown and the invert of the existing tunnel at each step. It should be pointed out that the stresses are additional bending stresses regardless of initial stresses. Moreover, tensile stress is assumed to be positive, while compressive stress is assumed to be negative. From the three figures, it can be observed that the overlying shield tunnelling will subject the crown of the existing tunnel to tensile stresses, and the invert of the existing tunnel to compressive stresses. This can also be deduced from the vertical displacement curves in Fig. 7. Moreover, the maximum tensile stress and the maximum compressive stress (absolute value) always occur in the symmetric plane during the construction process of the new shield tunnel. Furthermore, with the increase in offset from the symmetric plane, the stresses at both the crown and the invert of the existing tunnel decrease rapidly within the offset of 18 m from the symmetric plane, then decline slightly, and finally tend to be constant. Fig. 12 shows the change of stresses at C1–C5 and I1–I5 during the construction process of the new shield tunnel. It can be observed that the longitudinal stresses (absolute value) both at the crown and the invert of the existing tunnel increase as excavation occurs, and decrease as grouting occurs. This variation trend is in accordance with that of vertical displacement in Fig. 8. In addition, as excavation occurs in step 2, the cumulative tensile stress at C5 attains its maximum value of about 500 kPa, and the cumulative compressive stress at I5 attains its maximum value of about

Fig. 9. Vertical displacements at D1–D4 versus ground loss ratio and grouting ratio in step 2.

Fig. 10. Post construction displacements of existing tunnel.

64

P. Li et al. / Tunnelling and Underground Space Technology 42 (2014) 59–66

500

500 Excavation Grouting

300

At crown

200 100 0 -100 -200

E-Excavation G-Grouting 1-In step 1 2-In step 2 3-In step 3

400

Longitudinal stress (kPa)

Longitudinal stress (kPa)

400

300

C1 C2 C3 C4 C5

200

100

At invert 0

-300

-36

-27

-18

-9

0

9

18

27

36

(a) In step 1

E2

G2

E3

G3

Construction process E1

G1

E2

G2

E3

G3

0

At crown

200

Longitudinal stress (kPa)

Longitudinal stress (kPa)

Initial state

Excavation Grouting

400

100 0 -100 -200 -300

At invert -36

-27

-18

-9

0

9

18

27

Excavation Grouting

400

I1 I2 I3 I4 I5

Fig. 12. Longitudinal stress of existing tunnel versus construction process of new tunnel.

At crown

200 100 0 -100

-300

E-Excavation G-Grouting 1-In step 1 2-In step 2 3-In step 3

(b) At the invert of existing tunnel

500

-200

-200

-300

(b) In step 2

300

-100

36

Offset from symmetric plane (m)

Longitudinal stress (kPa)

G1

Construction process

(a) At the crown of existing tunnel

500

300

E1

Initial state

Offset from symmetric plane (m)

At invert -36

-27

-18

-9

0

9

18

27

36

Offset from symmetric plane (m)

(c) In step 3 Fig. 11. Longitudinal stress of existing tunnel.

300 kPa, implying that this construction stage has a significant influence on the longitudinal stress of the existing tunnel. Furthermore, the stresses (absolute value) at the crown of the existing tunnel are greater than those at the invert, which confirms that axis tensile force exists in the existing tunnel. From the stresses at the crown and the invert of the nine cross sections of the existing tunnel (Fig. 3), the longitudinal bending moment of the existing tunnel can be obtained using material mechanics formulae.

Fig. 13 shows the longitudinal bending moment curves of the existing tunnel. The amplitude of the longitudinal bending moment always occurs in the symmetric plane, and the longitudinal bending moment in the symmetric plane attains its maximum value of about 35,000 kN m in the excavation of step 2. With the increase of offset from the symmetric plane, the longitudinal bending moments decrease rapidly within the offset of 18 m from the symmetric plane, and then decline slightly beyond this region. This change pattern coincides with that of longitudinal stress in Fig. 11. Fig. 14 shows the change of the longitudinal bending moments in five cross sections of the existing tunnel along with the construction process of the new shield tunnel. The change pattern of the longitudinal bending moment is similar to that of longitudinal stress. It is noted that the longitudinal bending moments increase as excavation occurs, while they decrease as grouting occurs. This implies that grouting is a powerful approach to mitigate the influence of new shield tunnelling on the internal forces of the existing tunnel. Moreover, the longitudinal bending moments in these cross sections attain their maximum values in the excavation of step 2, which indicates that construction just over the existing tunnel induces a comparatively significant effect on the longitudinal bending moment of the existing tunnel. Furthermore, the variation in longitudinal bending moment is more apparent in the symmetric plane than those in the other four cross sections, which indicates that safety in the symmetric plane of the existing tunnel should

P. Li et al. / Tunnelling and Underground Space Technology 42 (2014) 59–66

65

as the grouting ratio increases from 0% to 30%, which is assumed to be the result of test error. The change rate of the longitudinal bending moment in the symmetric plane is significantly more rapid than that in the other cross sections as the ground loss ratio and the grouting ratio increase. The above change patterns are similar to those of vertical displacement shown in Fig. 9.

3.3. Pore water pressure at the spring line of existing tunnel

Fig. 13. Longitudinal bending moment of existing tunnel.

Fig. 16 shows the change in pore water pressure at the spring line of the existing tunnel during the construction process of the new shield tunnel. The pore water pressure decreases as excavation occurs, and increases as grouting occurs. These variation trends are contrary to the changes in vertical displacement and longitudinal internal force of the existing tunnel with the construction process of the new shield tunnel. In addition, the pore water pressure varies within the range from 269.8 kPa to 279.0 kPa, and the amplitude of variation is less than 10 kPa. Furthermore, the value of the pore water pressure during the construction process of the new shield tunnel is less than the initial value. Consequently, the influence of variation of pore water pressure on the existing tunnel is slight. Taking as an example the change of the pore water pressure in step 2, Fig. 17 presents the correlation between the pore water pressure at the spring line of the existing tunnel with the ground loss ratio, and with the grouting ratio. It is noted that the pore water pressure is almost constant as the ground loss ratio increases

Fig. 14. Longitudinal bending moment of existing tunnel versus construction process of new tunnel.

be given more attention during the construction process of a new shield tunnel. Fig. 15 shows the change in longitudinal bending moment in five cross sections of the existing tunnel with the increases of ground loss ratio and grouting ratio in construction step 2 of the new shield tunnel. The longitudinal bending moments exhibit approximately linear growth with the increase of ground loss ratio. It decreases with the increase of grouting ratio. The longitudinal bending moment in the symmetric plane shows a slight increase

Fig. 16. Pore water pressure around existing tunnel versus construction process of new tunnel.

Fig. 15. Longitudinal bending moment of existing tunnel versus ground loss ratio and grouting ratio in step 2.

Fig. 17. Pore water pressure around existing tunnel versus ground loss ratio and grouting ratio in step 2.

66

P. Li et al. / Tunnelling and Underground Space Technology 42 (2014) 59–66

from 0% to 0.4%. With the follow-up increase of the ground loss ratio, the pore water pressure begins to decrease rapidly. The same phenomenon can be observed in the relationship between the pore water pressure and the grouting ratio. The pore water pressure increases slightly as the grouting ratio increases from 0% to 90%, but it then increases rapidly as the grouting ratio continues to increase further. It is concluded that pore water pressure is slightly affected by a small ground loss ratio and a small grouting ratio. The above variation trends are contrary to those of the vertical displacement and the longitudinal bending moment of the existing tunnel.

4. Conclusions A centrifuge model test was carried out to estimate the influence of new shield tunnelling on an existing underlying largediameter tunnel. The amplitude and the change pattern of vertical displacement and longitudinal internal force of the existing tunnel are obtained, as well as the change pattern of pore water pressure at the spring line of the existing tunnel. From the test results, the following conclusions can be obtained: (1) The impact of overlying shield tunnelling on the vertical displacement of the existing tunnel is significant. The cumulative vertical displacement can reach 12 mm in the cross section of the existing tunnel which is just beneath the axis of the new shield tunnel. The heave of the existing tunnel in its longitudinal direction extends to a distance of 39 m from the axis of the new shield tunnel. (2) The post construction displacement of the existing tunnel undergoes a rapid decline in the first year. Then displacement of the existing tunnel decreases gradually to reach a stable value within the range from 2 mm to 1 mm in the following years. (3) Construction of a new shield tunnel causes additional longitudinal internal force in the existing tunnel. During construction of a new shield tunnel, the change in longitudinal tensile stress of the existing tunnel is in the range of 0–500 kPa, the change in longitudinal compressive stress is in the range of 0–300 kPa, and the longitudinal bending moment can reach a maximum value of about 35,000 kN m. (4) The longitudinal bending stress and the bending moment of the existing tunnel are significantly influenced within a range of 18 m from the axis of the new shield tunnel. (5) The vertical displacements and the longitudinal bending moments of the existing tunnel exhibit approximately linear growth as the ground loss ratio increases. Conversely, they decrease with an increase in grouting ratio. (6) Grouting can effectively reduce the displacement and the longitudinal internal force of the existing tunnel, and sufficient backfilling for the stabilisation of the existing tunnel is important. (7) The amplitude of variation in pore water pressure at the spring line of the existing tunnel is less than 10 kPa. The value of the pore water pressure during construction is less than the initial value. Therefore the influence of pore water pressure on the existing tunnel is not significant. (8) The pore water pressure at the spring line of the existing tunnel is almost constant as the ground loss ratio of the new shield tunnel increases from 0% to 0.4%, and it then begins to decrease rapidly with an increase in ground loss ratio. The pore water pressure increases slightly as the grouting ratio increases from 0% to 90%, and then it increases rapidly as the grouting ratio increases further.

Acknowledgement This research work was funded by the Ministry of Housing and Urban-Rural Development of the People’s Republic of China (Grant No. 2011-K3-28). References Chang, C.T., Sun, C.W., Duann, S.W., Hwang, R.N., 2001a. Response of a Taipei Rapid Transit System (TRTS) tunnel to adjacent excavation. Tunn. Undergr. Sp. Technol. 16 (3), 151–158. Chang, C.T., Wang, M.J., Chang, C.T., Sun, C.W., 2001b. Repair of displaced shield tunnel of the Taipei rapid transit system. Tunn. Undergr. Sp. Technol. 16 (3), 167–173. Dolezalova, M., 2001. Tunnel complex unloaded by a deep excavation. Comput. Geotech. 28 (6–7), 469–493. Hu, Z.F., Yue, Z.Q., Zhou, J., Tham, L.G., 2003. Design and construction of a deep excavation in soft soils adjacent to the Shanghai Metro tunnels. Can. Geotech. J. 40 (5), 933–948. Huang, A.J., Wang, D.Y., Wang, Z.X., 2006. Rebound effects of running tunnels underneath an excavation. Tunn. Undergr. Sp. Technol. 21 (3–4), 399–405. Jiang, M., Yin, Z.Y., 2012. Analysis of stress redistribution in soil and earth pressure on tunnel lining by discrete element method. Tunn. Undergr. Sp. Technol. 32, 251–259. Klar, A., Marshall, A.M., Soga, K., Mair, R.J., 2008. Tunneling effects on jointed pipelines. Can. Geotech. J. 45 (1), 131–139. Li, P., Du, S.J., 2012a. Responses of cross-river tunnel due to overlying shield tunnel construction (I): influence of construction procedure. In: Proceedings of the International Conference on Pipelines and Trenchless Technology 2012. Wuhan, China, pp. 1585–1594. Li, P., Du, S.J., 2012b. Responses of cross-river tunnel due to overlying shield tunnel construction (II): influence of distance. In: Proceedings of the International Conference on Pipelines and Trenchless Technology 2012. Wuhan, China, pp. 1595–1605. Li, X.G., Yuan, D.J., 2012. Response of a double-decked metro tunnel to shield driving of twin closely under-crossing tunnels. Tunn. Undergr. Sp. Technol. 28, 18–30. Liao, S.M., Liu, J.H., Wang, R.L., Li, Z.M., 2009. Shield tunneling and environment protection in Shanghai soft ground. Tunn. Undergr. Sp. Technol. 24 (4), 454– 465. Liu, H.Y., Small, J.C., Carter, J.P., Williams, D.J., 2009. Effects of tunnelling on existing support systems of perpendicularly crossing tunnels. Comput. Geotech. 36 (5), 880–894. Liu, H.L., Li, P., Liu, J.Y., 2011. Numerical investigation of underlying tunnel heave during a new tunnel construction. Tunn. Undergr. Sp. Technol. 26 (2), 276–283. Marshall, A.M., Klar, A., Mair, R.J., 2010. Tunneling beneath buried pipes: view of soil strain and its effect on pipeline behavior. J. Geotech. Geoenviron. Eng. 136 (12), 1664–1672. Meguid, M.A., Saada, O., Nunes, M.A., Mattar, J., 2008. Physical modeling of tunnels in soft ground: a review. Tunn. Undergr. Sp. Technol. 23 (2), 185–198. Ng, C.W.W., Lu, H., Peng, S.Y., 2013. Three-dimensional centrifuge modelling of the effects of twin tunnelling on an existing pile. Tunn. Undergr. Sp. Technol. 35, 189–199. Nomoto, T., Imamura, S., Hagiwara, T., Kusakabe, O., Fujii, N., 1999. Shield tunnel construction in centrifuge. J. Geotech. Geoenviron. Eng. 125 (4), 289–300. Sharma, J.S., Hefny, A.M., Zhao, J., Chan, C.W., 2001. Effect of large excavation on deformation of adjacent MRT tunnels. Tunn. Undergr. Sp. Technol. 16 (2), 93–98. Shen, S.L., Horpibulsuk, S., Liao, S.M., Peng, F.L., 2009. Analysis of the behavior of DOT tunnel lining caused by rolling correction operation. Tunn. Undergr. Sp. Technol. 24 (1), 84–90. Shen, S.L., Du, Y.J., Luo, C.Y., 2010. Evaluation of the effect of rolling correction of doubleo-tunnel shields via one-side loading. Can. Geotech. J. 47 (10), 1060–1070. Shen, S.L., Wang, Z.F., Yang, J., Ho, E.C., 2013a. Generalized approach for prediction of jet grout column diameter. J. Geotech. Geoenviron. Eng. http://dx.doi.org/ doi:10.1061/(ASCE)GT.1943-5606.0000932. Shen, S.L., Wang, Z.F., Sun, W.J., Wang, L.B., Horpibulsuk, S., 2013b. A field trial of horizontal jet grouting using the composite-pipe method in the soft deposit of Shanghai. Tunn. Undergr. Sp. Technol. 35, 142–151. Shen, S.L., Wu, H.N., Cui, Y.J., Yin, Z.Y., 2014. Long-term settlement behavior of the metro tunnel in Shanghai. Tunn. Undergr Sp. Technol. 40, 309–323. Wei, G., 2010. Selection and distribution of ground loss ratio induced by shield tunnel construction. Chinese J. Geotech. Eng. 32 (9), 1354–1361 (in Chinese). Wu, H.N., Huang, R.Q., Sun, W.J., Shen, S.L., Xu, Y.S., Liu, Y.B., Du, S.J., 2014. Leaking behaviour of shield tunnels under the Huangpu River of Shanghai with induced hazards. Nat. Hazards 70, 1115–1132. Yin, Z.Y., Chang, C.S., Hicher, P.Y., Karstunen, M., 2009. Micromechanical analysis of kinematic hardening in natural clay. Int. J. Plasticity 25 (8), 1413–1435. Yin, Z.Y., Chang, C.S., Karstunen, M., Hicher, P.Y., 2010. An anisotropic elastic viscoplastic model for soft clays. Int. J. Solids Struct. 47 (5), 665–677. Yin, Z.Y., Karstunen, M., Chang, C.S., Koskinen, M., Lojander, M., 2011. Modeling time-dependent behavior of soft sensitive clay. J. Geotech. Geoenviron. Eng. 137 (11), 1103–1113. Zhang, J.F., Chen, J.J., Wang, J.H., Zhu, Y.F., 2013. Prediction of tunnel displacement induced by adjacent excavation in soft soil. Tunn. Undergr. Sp. Technol. 36, 24– 33.