Excavation Sequence and Surrounding Rock Mass Stability of Large-Scale Underground Engineering with 8 Tunnels

In some large-scale hydroelectric power projects, there are more than 3 tunnels that are too close to each other to eliminate the mutual influence during the excavation period, especially for large-scale tunnel groups. In this paper, aimed at analyzing the Bukun hydropower station consisting of 8 tunnels in Malaysia, the displacement, stress, and plastic zone of the surrounding rock mass are analyzed to study the effect of the excavation sequence on the stability of the surrounding rockmass for large-scale tunnel groups. On the one hand, the in situ monitoring of the surface displacement of the rock mass surrounding the tunnel using extensometers is performed to obtain the deformation characteristics on the excavation limit under the typical excavation sequence. On the other hand, a series of elastic-plastic 3D numerical experiments are carried out to explore the displacement characteristics, stability of the large-scale tunnel groups, and safety of the initial supporting system. *e results show that the tendencies of the displacement increase corresponding to the tunnel face movement are similar for the three excavation sequences. *e displacement under initial excavation sequence 2 (IES2) is the smallest among the three sequences; the area of the plastic zone under IES2 is the smallest among them; and the stresses in the shotcrete layer and axial forces in the rock bolt under the three excavation sequences are within the safety limitation. Initial excavation sequence 2 is an optimized excavation sequence, in which tunnels #1 and #5 are excavated first; after an advance of 3 times the diameter of the tunnel, tunnels #3 and #7 are excavated; tunnels #2 and #6 are excavated after an advance of 3 times diameters; and tunnels #4 and #8 are excavated after an advance of 3 times diameters.


Introduction
Tunnels are one of the main structures that allow water to flow from reservoirs to power plants in hydroelectric projects. ere are always two or more power tunnels due to the electric generating unit in a large-scale project [1,2]. In some projects, there are so many power tunnels, their construction, such as excavation, often takes years to complete, and these tunnel excavations are the main time-consuming subproject in the entire project [3]. Many attempts have been made to minimize the influence of excavation on the stability of the surrounding rock mass in scheduled construction times [4,5]. In tunnel excavation applications, engineers and scholars are eager to find a good solution that can reduce the tunnel construction time to reduce project cost and improve the stability of tunnels, especially for large-scale tunnel groups. e excavation sequence is very important to the scheduled construction time and stability of large-scale tunnel groups. In this paper, a hydroelectric project with 8 power tunnels, the Bukun hydropower station in Malaysia, is selected to investigate the effect of excavation sequences on the stability of the rock mass surrounding the tunnel. e optimized excavation sequence is proposed based on the stability of the rock mass and the tunnel and the supporting system safety of largescale tunnel groups.
Regarding the mutual influence of tunnels during excavation, there are several reports on the influence of tunneling on the existing tunnels in the soil rock stratum using numerical or in situ monitoring methods, even by means of theoretical tools. A parametric study is performed based on numerical methods for a three-dimensional model to explore the complex interaction of the different parameters, which may affect the tunnel response on the open-pit mining sequence of existing tunnels [6]. In another example, the deformation of an existing tunnel is analyzed by 3D finite element methods (FEM), and the settlement of the overburden is obviously influenced by the structure of the existing tunnels through the plastic development of the soil layer [7,8]. An empirical method is developed to estimate the settlement trough caused by the second of twin tunnels, and it can provide a preliminary predictive tool for shield tunnels based on the Heathrow Express project in London, UK, as well as a database of case histories for twin tunnels prepared from reported information worldwide [9]. In addition, a field monitoring application is presented to evaluate the close proximity tunneling influences of a new tunnel on an existing railroad tunnel. e principal component analysis-based monitoring framework is developed using the sensor datum collected from the existing tunnel while the new tunnel is under excavation [10]. A case of closely spaced twin tunnels excavated beneath other closely spaced existing twin tunnels is presented, and the settlements of the existing tunnels and the ground surfaces associated with the new tunnel construction were systematically monitored [11].
ese results show that a close existing tunnel has a substantial influence on the newly constructed tunnel, and the new tunnel excavation also has a remarkable influence on the deformation and stress of the old tunnel in the surrounding rock mass. e failure of the support structure is the main problem for old tunnels. e stability of the surrounding rock mass is the main problem for newly constructed tunnels, especially for large-scale tunnel groups.
Physical modeling tests play an important role in studies related to the excavation of tunnels in soft ground. A variety of modeling techniques have been developed by researchers to study the ground response to tunneling worldwide [12]. Among the physical modeling test methods for geotechnical engineering, the centrifuge test is an important method in many studies. Some experts carried out research work by means of centrifuge model tests, and many results were achieved. A centrifuge model test was carried out to investigate the effect of new shield tunneling on the existing underlying large-diameter tunnel, and three construction steps of a new shield tunnel were simulated in this centrifuge model test [13]. e deformation and stress of the existing tunnel that was affected by the construction of new tunnels were systematically monitored. e settlement development of existing tunnels has four stages [14]. A series of threedimensional centrifuge model tests were conducted to investigate the effect of twin tunnel construction on the existing single pile in dry sand. e influence of the depth of each tunnel relative to the pile was investigated by constructing twin tunnels either close to the mid-depth of the pile shaft or near the pile toe [15]. Although centrifuge model testing is a suitable and available approach, there are very limited cases studying the effect of the large-scale tunnel group during the construction period.
In this paper, with the object of studying the influence of the excavation sequence on the stability of the rock mass surrounding a tunnel, a series of in situ monitoring tests of the rock mass surface displacement surrounding a tunnel using tape extensometers are carried out to obtain the deformation or displacement on the excavation limit under typical excavation sequences. In addition, to obtain more information on the surrounding rock mass and supporting system, a systemic series of elastic-plastic 3D numerical experiments are carried out by FEM to explore the deformation characteristics and stability of the rock mass surrounding the tunnel under different excavation sequences based on the Bukun hydropower station in eastern Malaysia. Finally, an optimized excavation sequence is suggested for the Bukun hydropower station.

Location and Main Building.
e Bukun hydropower station is located on the Balui River in eastern Malaysia, approximately 37 km from Belaga town and 180 km from port town Bintulu. is hydrojunction project consists of four main buildings: concrete-faced rockfill dams, tunnels and a powerhouse on the left bank, a discharge structure, and a diversion structure on the right bank. e elevation of the earth dam is 235.00 m, the maximum height of the dam is 205.0 meters, and the total reservoir capacity is 44.0 Gm3. e layout of the power tunnels is shown in Figure 1. ere are 8 tunnels that can introduce water from the reservoir to the powerhouse; the inner diameter of the tunnel is approximately 9.0 meters, and the tunnel cross-section is also present in Figure 1. e length of the tunnel is approximately 600 meters. e orientation of the tunnel axis is NW 341°, the overburden of the tunnel is in the range of 50-235 m, and the diameter of the tunnel excavation limit is approximately 10 meters.

General Geological Information of the Tunnel.
According to the geology report by Lahmeyer Limited Company, there are mainly three types of stratum in the area of concern, such as greywacke, shale, and interbedding of greywacke/shale. e main rock strata along the tunnel are shown in Figure 2. e surrounding rock mass at the power tunnel is predominantly slightly weathered～fresh greywacke, shale (mudstone), and greywacke/shale interbeds, of which greywacke accounts for 50.0～62.6% (with an average of 54.8%), shale (mudstone) accounts for 30.9～43.0% (with an average of 38.4%), and greywacke/shale interbeds account for approximately 4.5～9.2% (with an average of 6.8%). e greywacke is thick-bedded with sparse joints, and bedding is poorly developed with a high strength, and it can be regarded as an isotropic rock mass. Shale (mudstone) and graywacke/shale interbeds are anisotropic due to obvious interlayer joints. e surrounding rock mass is classified into 4 grades (from grades I to IV), and the geomechanical property parameters of each surrounding rock mass grade are shown in Table 1.

2
Advances in Civil Engineering

Initial Excavation Sequence (IES).
ere are 8 power tunnels in the Bukun hydropower station, and a reasonable excavation sequence is important to shorten the construction time and improve the safety of the excavation. ree initial excavation sequences (IES1, IES2, and IES3) are initially suggested and are described further in Table 2.
Initial Excavation Sequences 1 (IES1): tunnels #1 and #3 are excavated first at the same time. Tunnels #2 and #4 are     excavated after the faces of tunnels #1 and #3 move approximately 3 times the diameter; tunnels #5 and #7 are excavated after tunnel #2, and the #4 tunnel face moves approximately 3 times the diameter; tunnels #6 and #8 are excavated after tunnel #5, and the face of tunnel #7 moves approximately 3 times the diameter. IES1 is shown in Figure 3(a). Initial Excavation Sequences 2 (IES2): tunnels #1 and #5 are excavated first at the same time, and then tunnels #3 and #7 are excavated after tunnel #1 and the #5 tunnel face moves approximately 3 times the diameter; then, the tunnel faces of tunnels #3 and #7 move approximately 3 times their diameters, and tunnels #2 and #6 are excavated. e next step is for tunnels #4 and #8 to be excavated. IES2 is shown in Figure 3 Initial Excavation Sequences 3 (IES3): tunnels #1, #3, #5, and #7 are excavated first at the same time, and tunnels #2, #4, #6, and #8 are excavated after the #3, #5, and #7 tunnel faces move approximately 3 times their diameter. IES3 is shown in Figure 3 e distance between the adjacent tunnels is 3 times the tunnel diameter. at is, the tunnel face of the next step is at a location of 3 times the tunnel diameter after the previous step, as shown in Figure 3.

Monitoring Scheme.
On the construction site, basic displacement monitoring is conducted to obtain the deformation information of the rock mass surrounding the tunnel and evaluate the validation of the three IESs. Displacement monitoring is conducted using the tape extensometer tool. ree sections are selected to perform the monitoring in tunnels #1-#4. Each section will be excavated by lengths of 50 m during monitoring, as shown in Figure 4.
In preparing the monitoring point, a measuring crosssection consists of three displacement rings, and a measuring set consists of five measuring cross-sections. Each measuring cross-section has an interval distance of 3 meters, as shown in Figure 5. All 3 displacement rings in the same crosssection are located in the same plane perpendicular to the tunnel axis. ere are three measuring lines in a measuring cross-section. e length of the measuring line can be obtained by a tape extensometer. e vertical displacement of the top point and the convergence deformation of the two sidewalls can be calculated by three lengths of the measuring line based on the constant area of the triangle formed by three measuring lines. e supporting systems are the same in the three tunnel sections, and the shotcrete layer is type C20, with a thickness of 15 cm, a rock bolt diameter of 22 mm, a length of 3.5 m, and a row and column spacing of 1.0 m × 1.5 m.
e displacement measuring rings are installed 0.5 m behind the tunnel face. e first measurement is taken approximately 2 hours after installation. After the first measurement, the subsequent measurement is performed at each 3 meter distance ahead of the tunnel face.

Displacement Datum and Discussion.
Based on the monitoring datum on tunnels #1 to #4, the effect of the IES on the deformation characteristics of the rock mass surrounding the tunnel is discussed. In this section, only the datum from tunnels #1 and #2 is chosen for discussion.
ese displacement curves at both the top point and sidewall in tunnels #1 and #2 under different IESs are shown in Figure 6. e full displacement curves from the surface monitoring can be roughly divided into two stages: the initial rapid growth stage when the tunnel face is close to the monitoring cross-section and the stable convergence stage when the monitoring cross-section is far from the tunnel face.
e displacements of both the top and sidewall points are dramatically affected by the IES in tunnel #1, but the influence is not obvious in tunnel #2. In tunnel #1, the total displacements of the top points are 7.82, 6.24, and 8.51 mm from IES1 to IES3, respectively. e total displacements of the sidewall points are 6.33, 5.15, and 6.85 mm. For tunnel #2, the top point displacements are 5.46, 5.31, and 5.48 mm, respectively. e sidewall point displacements are 4.57, 4.45, and 4.59 mm from IES1 to IES3, respectively. From the displacement curves of the top and sidewall points in tunnel #1, after installation of the displacement ring, the displacement increase is triggered by the advance of the tunnel and close tunnel face.
Comparing the displacements under the three IESs, the smallest displacement is produced under IES2, and the largest displacement is produced under IES3. us, IES2 is the best initial excavation sequence for the surrounding rock mass stability of the tunnel among the three IESs.

Modeling.
To obtain more information on the stability of the rock mass surrounding the tunnel during excavation, a series of 3D numerical experiments are carried out using the FEM code ABAQUS. e whole 3D model created by the ABAQUS preprocessor is shown in Figure 7(a). e element type used in the 3D model is C3D20. e model of 8 tunnels embedded in the whole 3D model is presented in Figure 7(b).      Advances in Civil Engineering e boundary of the whole 3D model includes the upper boundary, which is the original ground surface, and the bottom boundary, which is a plane with 3.0 times the tunnel diameter beneath the lowest tunnel. e two side boundaries are two planes that are located 2.5 times the tunnel diameter away from the tunnels.
In the numerical experiments, the surrounding rock mass of the tunnel is regarded as an elastic-plastic material with the Mohr-Column yield criterion. e material of the shotcrete layer and rock bolt is set as an elastic material. e physical and mechanical parameters of the rock mass, shotcrete, and bolt inputted into the numerical model are listed in Table 3.
According to a local geological report, the initial ground stress in the rock stratum around the project is a gravitydominated stress field. e vertical initial stress is expressed as σ v � cH.
e horizontal initial stress is expressed as where H is the overburden of the rock stratum, c is the unit weight of the rock mass, and k 0 is the coefficient of the confining pressure in the rock stratum, where To correctly simulate the excavation procession and response, the load releasing model is used to calculate the excavation loading on the tunnel limit, and it is applied in the numerical experiment as follows: where ΔF A is the ground stress releasing load vector, B is the element geometric matrix, α is the releasing coefficient of the stress, σ 0 is the initial ground stress vector, and S A is the area of the rock excavation. e excavation releasing load is a function including some factors, such as the excavation sequence, supporting time, and advance of the tunnel face. e stress-releasing coefficient α, a comprehensive index, can be obtained from engineering experience for a construction site. e initial excavation sequences applied in the numerical experiment are the same sequences as the in situ construction, IES1, IES2, and IES3.

Comparing the In Situ Monitoring
Results. Due to the delay in the in situ monitoring from the tape extensometer in the tunnel, the displacement from in situ monitoring is only a portion of the whole displacement during the entire excavation period. In this section, the results of the numerical experiments and in situ monitoring under the same scenario are discussed. e point at the top of the tunnel is regarded as the representative point on the excavation limit, and the datum of three IESs is discussed. All the displacement curves from tunnel #1 are plotted in Figure 8. In Figure 8, "-M" represents results obtained by monitoring, and "-N" represents results obtained by the numerical experiment.
By analyzing these displacement curves in Figure 8, it is evident that the tendency of the displacement increasing at the top point regarding to the tunnel face advance for all the curves is roughly similar by methods of both monitoring and numerical experiments. e displacement obtained by monitoring is smaller than that from the numerical experiment method. e similarities in the tendency and value of the displacement indicate that the numerical model in this paper is qualified to carry out further analysis.

Complete Displacement Curves and Total Displacement.
e results can be obtained in the numerical experiments, such as the displacements and stresses at all the points, especially the complete displacement curves from the in situ monitoring. ree curves are presented in Figure 9, and all the displacements are the deformation datum at the top point under three IES cases for tunnel #1.
Ahead of the tunnel face passing the monitoring crosssection, there is the same trend of the displacement increasing among the three IES cases. After the tunnel face passing through the monitoring cross-section, the displacements increase at different speeds at the top point under the three cases and reach different total displacements. e displacements at the top point in tunnels #1∼#8 are listed in Table 4. Under the same IES, the displacements at the top point in tunnels #4 and #5 are smaller than those in the other tunnels. By comparing the three IESs, the displacement under IES2 is the smallest.
e displacement under case IES3 is the largest displacement among them. erefore, according to the displacements of the surrounding rock mass at the top point, IES2 is a better excavation sequence for the stability of the rock mass surrounding the tunnel.

Plastic Zone of the Surrounding Rock Mass.
e plastic zone is another index of the stability of the surrounding rock mass for tunnels. e plastic zones of the rock mass surrounding the tunnel under the three IESs are presented in Figure 10. e plastic zones are distributed mainly around the tunnel excavation limit in the surrounding rock mass near the tunnel bottom, sidewall, and crown. e main differences occur in the tunnel crown and sidewall area.
By comparing the area of the plastic zone, case IES2 has the smallest plastic zone among the three cases, and case IES3 is the largest one among the three cases. erefore, case IES2 is a better choice for tunnel excavation.

Shotcrete Layer Stress.
e maximum/minimum stresses at some specific location in the shotcrete layer are listed in Table 5. Under all IES cases, all the stresses in the shotcrete layer are compressive, which shows that all the shotcrete layers are under compressive stress conditions. For one tunnel, the maximum compressive stress appears at the middle point in both sidewalls. For all the tunnels, these stresses in number #4 or #5 are larger than those of the other tunnels at the same location in the shotcrete layer for the same IES. In detail, for case IES1, the maximum stresses are 7.33, 7.44, 12.53, 10.09, 7.30, 11.48, and 11.21 MPa for tunnels #1 to #8 in the shotcrete layer, respectively. e maximum stresses are 9.54, 9.61, 9.12, 7.12, 9.02, 10, 9.09, and 9.01 MPa for 8 tunnels under case IES2, respectively. e maximum stresses are 7.22, 7.35, 12.17, 7.61, 12.28, 7.52, 11.62, and 11.49 MPa for 8 tunnels under case IES3, respectively.

Advances in Civil Engineering
In the three IES cases, the maximum stress in the shotcrete layer is 12.53 MPa, and it is induced at the left sidewall point in tunnel #3 under case IES1. All the stresses in the shotcrete layer are less than the compressive strength of the shotcrete. Any one of three IES cases can be a choice for the excavation sequence designing only considering the stresses in shotcrete layer.     Table 6.
Under cases IES1 and IES3, both the maximum axial forces at the top point are approximately 65 kN, which occurs in tunnel #6, but under case IES2, the maximum axial force is 62.5 kN.
ere is no apparent difference in the amount of bolt axial force among the three IES cases. All the maximum axial forces in any part of the surrounding rock mass are under the bearing capacity of the rock bolt. Hence,   all IESs can be a good choice for the excavation sequence design in consideration of the axial forces in the rock bolt. For large-scale engineering with 3 or more tunnels, which tunnel is excavated first and which tunnel is excavated last are an important consideration. e excavation sequence for tunnels is an important factor in approaching a safe and economical plan. For a good excavation sequence, the most important requirement is to ensure the stability of the rock mass surrounding the tunnel and the safety of the supporting system. By comparing these results of systematic numerical experiments among cases IES1 to IES3, the safety of the supporting system can be assured by all IESs. An IES can be achieved. In our work, by considering the displacement and plastic zone, case IES2 is better optimized than IES1 or IES3. erefore, IES2 is a better excavation sequence for the Bukun project with 8 tunnels.

Conclusions
Based on in situ monitoring and systemic numerical experiments, the stability of the rock mass surrounding the tunnel and safety of the supporting system are studied in this paper. e following conclusions can be drawn: (1) By comparing the results from both monitoring and numerical experiments, although the displacement obtained by monitoring is smaller than that via numerical experiments, the similarity in the increasing tendency shows that the numerical model is qualified to carry out further analysis. (2) In the displacement and plastic zone, the tendencies of the displacement increase corresponding to tunnel face movement are similar for the three IESs, but the total amount of displacement is different. e displacement under IES2 is the smallest among them. In the plastic zone, IES2 is the smallest case, and IES3 is the largest case among the three cases. ES2 is the better excavation sequence for the surrounding rock mass stability of the tunnel. (3) e maximum stresses in the shotcrete layer are 12.53, 9.61, and 12.28 MPa for the three IESs, respectively. All the stresses in the shotcrete layer are less than the compressive strength of the shotcrete. All IESs can be a choice for the excavation sequence in consideration of the stress in the shotcrete layer. (4) e initial excavation sequence 2 (IES2) is recommended for engineering designers and owners. In IES2, #1 and #5 are excavated first; after an advance of 3 times the diameter of the tunnel, tunnels #3 and #7 are excavated; tunnels #2 and #6 are excavated after an advance of 3 times diameters; and tunnels #4 and #8 are excavated after an advance of 3 times diameters.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare no conflicts of interest.