Introduction

Any underground construction process and the use of underground utilities will cause interactions with surrounding ground and may cause hazards to existing facilities. These hazards include injection of hazardous materials (Du et al. 2012, 2014) and ground movements (Xu et al. 2012a; Wu et al. 2014, 2015a; Shen et al. 2015a, b). The underground construction process generally involves temporary works, e.g., deep excavation (Tan and Wang 2013a, b; Cui et al. 2015a, 2016), tunnelling excavation (Liao et al. 2009; Shen et al. 2009, 2010; Cui et al. 2015b), installation of deep mixing columns (Shen et al. 2003) and jet-grouting columns (Wang et al. 2013; Shen et al. 2013a, b), and pile installation. The use of underground utilities may cause long-term deformation due to both loading and geological movement, e.g., the extraction or leakage of groundwater (Xu et al. 2012a, b, 2013a, b, 2014, 2015; Shen et al. 2006, 2014, 2015a, b; Ma et al. 2014; Wu et al. 2014, 2015b, c, d).

Recently in China, with the process of rapid urbanisation, large numbers of underground structures have been constructed in various types of ground conditions. Railway and metro tunnels make up the largest proportion of underground construction currently in China. As metro railways are usually constructed among crowded urban buildings and structures, it is necessary to investigate the tunnelling excavation effects on adjacent buildings, and hazards prevention methods should be proposed (Liao et al. 2009; Du et al. 2012, 2014).

Many literatures on the physical and mechanical properties have been reported (Dearman 1974, 1978; BSI 1981; Norbury et al. 1995; Ehlen 2002). The granite rock is of high uniaxial compressive strength, which caused the granite or weathered granite to be difficult to break. When constructing the metro tunnels by shield tunnelling method, this kind of weathered granite will appear in front of the excavation face. A number of mixed-ground tunnelling excavations have been reported (Zhao et al. 1994; Shirlaw et al. 2000; Hassanpour et al. 2009; Delisio et al. 2013; Fargnoli et al. 2013). Zhao et al. (2007) described the mixed ground phenomenon and the difficulties encountered when tunnelling through mixed ground in Singapore. Tóth et al. (2013) presented a classification system categorising mixed-face ground and proposed a method to maintain safety in construction and to protect the surrounding ground. Numerical simulations have been used to investigate the interaction between tunnelling excavations and adjacent structures (Addenbrooke and Potts 2001). A new strategy was proposed to predict ground movements and potential damage to the adjacent structures (Shin et al. 2006). Blindheim et al. (2002) studied the effect of mixed face conditions on hard rock Tunnel Boring Machine performance and proposed a protection method for face stability. However, there are very few published field investigation cases recording the impact on adjacent buildings in an “upper-soft and lower-hard” ground, which consists of the soft residual soils formed in Quaternary in the upper layer and the hard weathered granite in the lower layer.

The objective of this paper is to investigate hazards on adjacent buildings during tunnel excavation in completely weathered granite overburdened by residual sandy soil. The interaction between the seven adjacent buildings and the construction procedure is also analysed via field monitoring both of the behaviour of the buildings and the construction parameters.

Project and geology

The Sui-Guan-Shen interurban railway (SGSIR) was a high speed railway linking Guangzhou, Dongguan, and Shenzhen located in Guangdong province, China (Fig. 1a, b). The SGSIR was 74.885 km long and included 14 stations. Taiping Tunnel (14.490 km) and Airport Tunnel (7.816 km) were two of the tunnels of the SGSIR, and were constructed by the shield tunnelling method. Three earth pressure balance shield boring machines (EPBM) were used to excavate Taiping tunnel, which included three stations and two tunnel sections. The test site was located at the south side of Humen Bridge, which was one section of the Taiping Tunnel running from Humen station to Changanxiabian station, measuring 2.893 km in length. There are two parallel tunnels. One is excavated towards the north (northbound tunnel) and the other one is excavated towards the north (southbound tunnel) (see Fig. 1c). Their centre-to-centre distance is 15.5 m. The subject in this paper is the southbound tunnel which was excavated first. The centreline of the tunnel was 17 m below the ground surface. Lining rings had an inner and outer diameter of 7.7 and 8.5 m, respectively, and were 0.4 m thick and 1.6 m long.

Fig. 1
figure 1

Location of test site: a map of China, b location of SGS interurban railway and the test site; c plan view of the tunnel and seven buildings

Full size image

There were seven buildings at the western side of the tunnel, as shown in Fig. 1c. Table 1 lists the general profile of these buildings. Their distance from the tunnel ranged from 2.1 to 7.2 m. A total of 428 jet-grouted columns were installed between the tunnel and the buildings to prevent large displacement of the buildings and high deformation of the adjacent soil. The total length of the seven buildings was about 80 m, and the jet grouting zone is 100 m, along the direction in which the tunnel was advanced. Figure 2 depicts the layout of the jet-grouted columns. Cement and sodium silicate mixed grout (CSG) was grouted into the ground as the secondary grouting material. Table 2 lists the parameters of the CSG.

Table 1 General profile of the seven buildings
Full size table
Fig. 2
figure 2

a Plan and b sectional view (A–A) of tunnel and building foundations

Full size image
Table 2 Parameters of cement and sodium silicate mixed grout
Full size table

The stratigraphy of the site is shown in Fig. 3a. The subsoil profile here consisted of backfill (from 0 to 10.1 m), quaternary soft deposit (GVI, from 3.4 to 20.8 m), completely weathered granite (GV, from 14 to 34 m), and moderately weathered granite (GIV, from 26.9 m). The excavation face mainly consisted of sand and completely weathered granite. The typical subsurface profiles and soil properties of the test site are depicted in Fig. 3b. The cohesion of these soils ranged from 13.2 to 36 kPa, with internal frictional angle varying between 10° and 25°. A standard penetration test (SPT) is conducted to check the strength of the soil. The mean SPT values of the completely weathered granite and moderately weathered granite ranged from 43 to 52 and from 61 to 75. The groundwater level was 3.5 m below the ground surface.

Fig. 3
figure 3

a Cross-sectional view and b soil properties of the construction site

Full size image

The weathering classification system is divided into five grades: fresh, slightly weathered, moderately weathered, highly weathered and completely weathered (MOHURD 2014), which is consistent with weathering classification proposed by Ehlen (2002). The structure of completely weathered granite is like hard soil, which collapses easily when pressed by the hands or washed with water. Highly weathered granite is like broken pieces of granite in soil or soil sand. It collapses easily when soaked in water. Moderately weathered granite, which is block-shaped with coarse grain in fragment to massive, has abundant weathered fractures. Slightly weathered granite has few weathered fractures and has high strength, with a high uniaxial compressive strength (UCS) of more than 100 MPa.

The property of granite in Humen is the same to that of granite all over Guangdong Province, including Shenzhen (Cui et al. 2016). The granite was generally formed in the Cretaceous and Jurassic periods. The mineral composition of the granite consists of feldspar and quartz. In southeast China, there are four genetic types of granite, remelting type, syntectic type, differentiation type and metamorphic–metasomatic type. Most of the granites in Humen are the syntectic type, and the rest are the re-melting type. The syntectic granite, which distributed adjacent to the fault zone, is formed by the synthetic effect of intrusive rock and continental crust material. The lithology mainly includes adamellite, granodiorite and biotite granite, and the occurrence is stock or dike.

The granite in Humen, whose tectonic is granular, porphyritic or porphyroid, is medium to coarse grain and exists in massive structures. The granite mainly includes minerals like silica and alkali feldspar. It has mainly been weathered by chemical weathering, which changes the minerals and chemicals present. During the process of weathering, the plagioclase decomposes first, and then the orthoclase and the biotite. The main products of weathering are clay minerals and silica, which have large pores that keep the granite residual soil within the structure of the parent rock.

The EPBM employed in this project was an earth pressure balance (EPB) shield machine with a diameter of 8810 mm to excavate the tunnel in the layers composed of completely to slightly weathered granite, which was the upper-soft lower-hard ground. Figure 4 shows the cutter head of both rippers and disc-cutters of the TBM, which can cut through hard rock having an unconfined compressive strength (UCS) of more than 200 MPa. The strength of the rock environment shown in this manuscript was only up to 50 MPa. However, full-face rock granite with unconfined compressive strength up to 180 MPa has been encountered in some other part of the rock along the tunnel alignment. Therefore, we used the earth pressure balance shield boring machine (EPBM) in this project.

Fig. 4
figure 4

Cutter head of the TBM

Full size image

Figure 5a depicts the tunnelling progress curve. It took 28 days for the EPBM to pass through the seven buildings from R368 to R435. The average advance speed was 2.72 m (1.7 rings) per day. The position of the EPBM cutter head is also shown in Fig. 5b. CSG was grouted four times into the soil. It took almost 7 days to pass through building 3 and 6 days to pass through building 6, which were relatively longer than the other five buildings.

Fig. 5
figure 5

Tunnelling progress curve of passing below the seven buildings

Full size image

Field test programme

During the tunnelling excavation under the seven buildings, lateral displacements of the subsoil, and horizontal displacements and settlements of the buildings were monitored. The monitoring instruments included one total station, 28 settlement gauges (C1–C28), and four soil inclinometers (I1–I4). Figure 6 shows the layout of the monitoring points. The settlement gauges were installed at the four corners of each building. The horizontal displacements were monitored by the total station. The horizontal displacement measuring points of the buildings were positioned at two corners close to the tunnel. The inclinometers were installed at 25 m depth. The four soil inclinometers (I1–I4) were 2.5, 2.1, 3.7 and 3.8 m away from the tunnel. The test monitoring frequency for the three items was twice a day (12 h interval).

Fig. 6
figure 6

Field instrumentation: a plan and b sectional view

Full size image

In case of cutter wear during passing under the seven buildings and the risk to replace cutters, when the EPBM was 15 m from the buildings, it was stopped for the examination of the cutter disc, which might be damaged and required replacement in the past tunnelling. Test monitoring started when the EPBM was stopped. It took 12 days to replace the cutters. After this, the EPBM was restarted to continue the excavation, and reached the buildings in 3 days. The total duration for the monitoring of tunnelling excavation effects on the adjacent buildings was 70 days, including monitoring at 15 days before, 28 days during and 21 days after passing under the seven buildings.

Observed results

Vertical displacements of buildings

Figure 7 shows the vertical displacements of the seven buildings induced by the tunnelling excavation. Positive vertical displacements indicate the heave of the building, whereas negative values denote the settlement of the building. The pattern of vertical displacements of one side of each building is very similar. During the 12 days of cutter replacement, it can be observed that 1 mm heave occurred during the first 7 days when the EPBM was stopped, and then almost 2 mm settlement occurred in the following 5 days before the EPBM was restarted (Fig. 7a, b), which indicates that the field monitoring made later is reliable and the error is acceptable, and the former tunnelling excavation had little influence on the test monitoring of the seven buildings.

Fig. 7
figure 7

Settlement of buildings: a building 7; b building 5 and 6; c building 3 and 4; d building 1 and 2

Full size image

As shown in Fig. 7a, there is a 2 mm heave before the EPBM arrived at building 1. After the EPBM left building 1, the building settlements increased. But there is a heave of about 1 mm due to the second grouting of CSG after which the building settlements increased. Another obvious cause of heave of 0.5–2 mm was found due to the fourth grouting of CSG (Fig. 5). The settlements of C3 and C4 are larger than those of C1 and C2 because C3 and C4 are closer to the tunnel (Fig. 6). The deviation between them is increasing. The building settlements of C3 and C4 are within 10 mm, which means building 1 was little affected by the tunnelling excavation.

The behaviour of the building settlements of the other six buildings is similar to that of building 1. For each building, the maximum settlements occurred near the tunnel (C3, C4, etc., as shown in Fig. 7). The maximum settlements observed for each building were 9.5 mm (C4), 12.3 mm (C7), 18.9 mm (C10), 17.8 mm (C14), 10.1 mm (C19), 8.7 mm (C22), and 5.7 mm (C26), when the EPBM was 80 m away from building 7. As can be seen, the maximum settlements of C10 (building 3) and C14 (building 4) were higher in comparison with the other five buildings. It can be concluded that before the EPBM head arrived at the building, there was 0–2 mm heave (Phase 1). Then the building settlements began to increase (Phase 2). The building settlements would reduce by 1–3 mm with the CSG grouting.

Horizontal displacements of seven buildings

Figure 8 shows the horizontal displacements caused by the tunnelling excavation. That all the horizontal displacements are positive suggests that the buildings were displaced away from the tunnel. About 1.5 mm horizontal displacement (Fig. 8a, b) was observed during the 12 days of cutter replacement, which suggests that the field monitoring made later is reliable and the error is acceptable, and the former tunnelling excavation barely affected the test monitoring of the seven buildings.

Fig. 8
figure 8

Horizontal displacements of building: a building 7; b building 4, 5 and 6; c building 1, 2 and Building 3

Full size image

As shown in Fig. 8a, there is a 2 mm horizontal displacement before the EPBM arrived at building 1. But the horizontal displacements did not increase while the EPBM passed building 7. After the EPBM left building 1, the building horizontal displacements increased by 1 mm due to tail skin grouting, and then they continued to increase. However, there is a decline of about 1 mm, possibly due to the unloading of the surrounding soil. Then the building horizontal displacements quickly increased after this due to the second grouting of CSG.

The behaviour of the building settlements of the other six buildings is similar to that of building 1. The horizontal displacements of each building increased with the advance of the EPBM. The maximum horizontal displacements observed were 6.5 mm (L2), 6.9 mm (L3), 8.8 mm (L6), 8.3 mm (C7), 7.8 mm (L9), 7 mm (L12), and 6.1 mm (C13). As can be seen, the maximum horizontal displacements of L6 (building 3) and C7 (building 4) were larger in comparison with the other five buildings.

Lateral ground displacements of ground

Figure 9 shows the lateral displacements caused by the tunnelling excavation as observed by inclinometer I1, I2, I3 and I4 after the EPBM passed the seven buildings 21 days later, which was 80 m away. The pattern of the lateral displacements is very similar. For every inclinometer, the maximum lateral displacements were 6.4 mm (I1), 6.7 mm (I2), 6.4 mm (I3) and 5.6 mm (I4), approximately at the level of the tunnelling zone, with significant mass movements of the ground. The lateral displacements of I3 and I4 were 2 mm larger than those of I1 and I2 up to a depth of 8 m, which is in accordance with the building horizontal displacements.

Fig. 9
figure 9

Lateral displacements of subsoil at inclinometer I1, I2, I3 and I4

Full size image

Analysis on construction procedure

Figure 10 presents the field observed total thrust curve during excavation. The total thrust ranged between 20,000 and 40,000 kN. There was an upward shift of the total thrust around R410 as the stratum was changing into one with a higher percentage of residual soil, which may have caused larger settlements and horizontal displacements of building 3 and building 4 in comparison with the other buildings. Before and after this stage, the disturbance induced by the tunnelling excavation is relatively small. Actually, the mean total thrusts before and after ring number R410 were 27,000 and 34,000 kN, an increase of 31 %. The reason for the variation of these parameters was the stratum change from completely weathered granite to the composite of residual soil and completely weathered granite (see Fig. 2), which was the upper-soft lower-hard ground. This kind of ground caused many difficulties in the tunnelling such as high cutter wear and flat cutters and tunnel face instability (Zhao et al. 2007). And this is why the EPBM was stopped to examine of the cutter disc when it was 15 m away from the seven buildings.

Fig. 10
figure 10

Variation of total thrust during tunnelling excavation

Full size image

The soil chamber pressure, affected by the thrust of the EPBM, the rotation rate of the screw, and the opening at the end of the screw, was adjusted to balance the face pressure at the cutter head. As can be seen in Fig. 11, the soil chamber pressure decreased before R410 as the percentage residual soil became higher with the excavation face. And after the R410, the soil chamber pressure increased rapidly and then stabilised as the percentage of residual soil became stabilised. After R410, there were no large displacements of building 5, 6 and 7 which mean the adjustment of the soil chamber pressure is effective.

Fig. 11
figure 11

Variation of soil chamber pressure during tunnelling excavation

Full size image

Figure 12 shows the grouting pressure at the tail-skin grouting which aims at filling the annular void between the ground and the lining. The grouting pressure shows the same tendency as the total thrust curve. The range of the mean grout pressures per ring varies from 0.1 to 0.28 MPa. Figure 13 shows the daily grouting volume. The mean grouting volumes before and after R410 are 9 and 10.8 m3/day. The larger grouting pressure and grouting volume after R410 are to prevent the building 5, 6 and 7 from large displacements.

Fig. 12
figure 12

Variation of grouting pressure during tunnelling excavation

Full size image
Fig. 13
figure 13

Variation of grouting volume during tunnelling excavation

Full size image

As shown in Fig. 5, it took 28 days for the EPBM to pass the seven buildings. The mean excavation speed is 1.7 rings per day. Regarding building 3 and building 5, the mean excavation speeds are 0.95 rings and 3.33 rings per day, respectively. Despite the lower excavation speed of building 3, its displacement is larger than that of building 5. Therefore, it is possible that within a certain range, the excavation speed is not the main factor for the displacements of the building.

Table 3 lists a comparison of the EPBM operation parameters in normal tunnelling excavation and the tunnelling excavation passing through the seven buildings. It can be concluded that a low penetration rate and a low excavation speed should be adopted and the thrust and the soil chamber pressure should be adjusted according to the stratum for the EPBM to safely pass through the seven buildings.

Table 3 TBM operation parameters
Full size table

After the analysis of the tunnelling excavation effects on the seven adjacent buildings, requirements of tunnelling excavation passing near buildings in upper-soft lower-hard ground are proposed:

  1. 1.

    Conduct a comprehensive geotechnical exploration of the construction site.

  2. 2.

    Fully overhaul the EPBM ahead of tunnelling construction, especially the cutter head, foam-filled tube, grouting tube, and the tunnel tail sealing, in case of machine faults during the “passing”.

  3. 3.

    Install jet-grouting columns or other separation structures between the tunnel and the buildings.

  4. 4.

    Predict and adjust the tail-skin grouting and secondary grouting by adjustment of the grouting material, grouting proportion, grouting pressure and grouting volume.

  5. 5.

    Modify the spoil at the working face if using the EPB method.

  6. 6.

    Monitor the soil deformation and the building displacements and feed back to the construction parameters in time.

  7. 7.

    Adjust the thrust force, excavation speed, screw conveyor mass per ring, penetration speed and posture of the EPBM.

Conclusions

Field monitoring has been conducted to investigate hazards on surrounding facilities within a very close distance of a shield tunnelling machine passing under them. The results of this field investigation are summarised as follows:

  1. 1.

    The hazards on seven buildings during shield tunnelling were within the controllable range. The maximum building settlement was 18.9 mm and the maximum building lateral displacement was 8.8 mm. The maximum lateral ground displacement was 6.7 mm. All of these displacements were within the allowable range of 20 mm for hazards prevention. The construction parameters optimised according to the gradually changing strata through the seven buildings rather than adopted in normal tunnelling could effectively reduce the hazards of surrounding environment during tunnel construction.

  2. 2.

    Different vertical and horizontal displacements have been observed for the seven buildings. The reason was attributed to the different tunnelling operation parameters that varied with the ground stratum. When tunnelling in the upper-soft lower-hard ground as the stratum was changing into one with a higher percentage of residual soil, there was an upward shift of total thrust as well as the soil chamber pressure and grouting pressure, leading to larger settlements and horizontal displacements of the nearby buildings. Within a certain range from 1 to 3.5 rings per day, excavation speed is not the main factor for the displacements of the building.

  3. 3.

    Tail grouting cannot mitigate excessive settlements of the surrounding soil and the adjacent buildings. However, secondary grouting can sometimes restrain the displacements of buildings, but can increase the buildings’ horizontal displacements.

  4. 4.

    Requirements to mitigate hazards of environments during tunnelling excavation passing below buildings in upper-soft lower-hard ground are proposed based on this case. This successful case provides an example for future environmental protection construction of shield tunnel in similar geological strata.