Tunnelling and Underground Space Technology 71 (2018) 329–339

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology

journal homepage: www.elsevier.com/locate/tust

Comparative study of construction methods for deep excavations above MARK

shield tunnels
⁎
Ming-Guang Lia, Jin-Jian Chena, Jian-Hua Wanga, , Yan-Fei Zhub
a
State Key Laboratory of Ocean Engineering, Department of Civil Engineering, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai, China
b
Shanghai Tunnel Engineering Co., Ltd., 118 Dalian Rd., Shanghai 200082, China

A R T I C L E I N F O A B S T R A C T

Keywords: Although an appropriate construction method contributes to minimizing the excavation-induced impact on
Deep excavation existing facilities, few studies have focused on the comparison and optimization of the construction methods for
Tunnels engineering projects in practice. In this paper, Three-dimensional (3D) numerical simulations are carried out to
Numerical simulation compare the construction methods for an excavation above shield tunnels through a case history, which has been
Construction method
constructed by a novel excavation method, referred to as divided alternate excavation method (DAEM). A 3D
Comparative analysis
numerical model is developed in the finite element (FE) software, ABAQUS 6.10, to simulate the practical
construction process of the DAEM, and the numerical results are analyzed in combination with the field data.
Comparative studies are performed by simulating the hypothetical construction schemes of two conventional
excavation methods that were proposed for the investigated case. Numerical results confirm the effectiveness of
the DAEM on controlling the excavation-induced underlying tunnel deformation. Extensive parametric studies of
the DAEM are performed for further investigation. Through the comparative studies and parametric studies,
suggestions are proposed for optimization of construction methods for deep excavations above shield tunnels.

1. Introduction remedial measures or countermeasures are adopted, including soil re-

Shanghai has experienced unprecedented infrastructure construc-
tion activities in the past decades. To relive the traffic pressure in this
congested urban area, many metro tunnels have been excavated in this
period using earth pressure balanced (EPB) shield machines.
Meanwhile, more and more highway tunnels are being constructed,
which calls for deep excavations supported by retaining structures.
Deep excavations will cause potential large soil deformations inside and
outside of the excavation pit, and will result in serious influence on
adjacent existing shield tunnels (Chai et al., 2014; Son and Cording,
2005). If excessive stress and deformation are induced in the shield
tunnels, cracks will be developed and these tunnels may be damaged
(Chang et al., 2001). Therefore, to minimize the adverse impact, it is of
particular importance to compare and optimize the schemes before
construction.
Tunnel deformations resulted from an adjacent deep excavation are
a common problem encountered by engineers (Cooper et al., 2002; Lu
and Hanyga, 2005a; Tan et al., 2015; Chen et al., 2016; Li et al.,
2017a,b; Zhang et al., 2017). Chang et al. (2001) reported a case history
in which a section of a tunnel was damaged as a result of an adjacent
excavation. To keep the existing tunnels from damaging, several
inforcement within excavation area (Hu et al., 2003; Huang et al., 2006;
Chen et al., 2013) and stiffness enhancement of retaining structures
(Huang et al., 2013). Moreover, a rational excavation sequence also
helps to control the deformation of the existing tunnels efficiently by
following the time-space effect (Hu et al., 2003; Lu and Hanyga, 2005b;
Liu et al., 2011; Li et al., 2014). However, the effectiveness of these
remedial measures or countermeasures on the tunnel deformation
needs to be evaluated.
Numerical method is a feasible tool and has been widely used to
evaluate the response of tunnels to adjacent deep excavations (Zhang
et al., 2013; Ng et al., 2013, 2015a, 2015b). Two dimensional (2D)
numerical simulation is preferred because of its high computational
efficiency (Chen et al., 2011; Sharma et al., 2001). Using the 2D FEM,
Hu et al. (2003) predicted the vertical and horizontal displacements of
two adjacent tunnels induced by the excavation. Dolezalova (2001)
predicted the deformation and stress change in a tunnel induced by an
open excavation, and the computed results agreed well with the mon-
itoring data. Zheng and Wei (2008) studied the response of existing
tunnels at three typical locations to the excavation, namely at the
central line under the excavation bottom, directly under the base of the
diaphragm wall and outside of the diaphragm wall. Considering the

⁎
Corresponding author.
E-mail addresses: Lmg20066028@sjtu.edu.cn (M.-G. Li), chenjj29@sjtu.edu.cn (J.-J. Chen), wjh417@sjtu.edu.cn (J.-H. Wang), zyflisp@163.com (Y.-F. Zhu).

http://dx.doi.org/10.1016/j.tust.2017.09.014
Received 17 January 2017; Received in revised form 7 September 2017; Accepted 10 September 2017
Available online 15 September 2017
0886-7798/ © 2017 Elsevier Ltd. All rights reserved.
M.-G. Li et al.
asymmetric nature between the excavation and the adjacent tunnels,
three dimensional (3D) numerical simulations are needed to achieve
more accurate estimations. Through 3D simulations, Liu et al. (2011)
analyzed various influence factors, such as the ground reinforcement
depth, the excavation sequence and the skew angle between the new
tunnel and the existing tunnels, on the deformation of underlying
tunnels. More recently, Huang et al. (2013) studied parameters that
may affect the tunnel response, including the relative position of the
tunnel with respect to the excavation, tunnel diameter, excavation di-
mensions, and tunnel protection measures. Until now, many numerical
simulations have been carried out to simulate the excavation process
and the structural response of the adjacent tunnels. However, only few
focused on quantifying the effect of construction methods on the de-
formation of tunnels, although an appropriate construction method
plays an important role in practice (Ferrada and Serpell, 2014; Zhang
et al., 2016).
In this paper, 3D numerical simulations are performed to compare
three construction methods through a case history of a deep excavation
located above metro tunnels. The effect of the construction method on
the excavation-induced underlying metro tunnel deformation is dis-
cussed and quantified. Moreover, extensive parametric studies are
conducted to optimize the novel construction method of DAEM.
2. Overview of the case
The investigated case was an excavation of a highway tunnel lo-
cated above an existing metro tunnel in Shanghai. The highway tunnel
was constructed by the cut-and-cover method. The excavation of the
highway tunnel was 27.1 m in width and 11.5 m in depth, and the
crossing length over the metro tunnel is 150 m. The underlying metro
lines were composed of two shield tunnels, i.e. the inbound and the
outbound, which were excavated using earth pressure balanced (EPB)
shield machines. The internal diameter and the thickness of the lining
structure of the tunnels were 6.2 m and 0.35 m, respectively. More
detailed information of the project was reported by Chen et al. (2015).
Metro tunnels are the lifelines of a city and are sensitive to adjacent
constructions. To safeguard the safety of these lifelines, strict guidelines
have been made for deep excavations adjacent to metro tunnels, e.g. the
construction-induced maximum vertical displacement of the tunnels
must be controlled within 20 mm (SZ-08-2000 Shanghai Municipal
Engineering Authority, 2000). In this project, the clearance between the
new highway tunnel and the existing tunnels’ crown was only 4 m, so
appropriate countermeasures were needed before and during con-
struction. Soil improvement is an efficient way to reduce the rebound of
soil below the excavation surface (Wang et al., 2010). The deep soil
mixing (DSM) method was used in this project to improve the sur-
rounding soil (Chen et al., 2013) and the tunnels were uplifted with a
maximum value of 3.7 mm during the soil improvement process.
Therefore, the heave of the underlying metro tunnels caused by the
excavation must be controlled within 16.3 mm.
To safeguard the metro tunnel, Chen et al. (2015) proposed a novel
construction method of a divided alternate excavation method (DAEM)
for the project by following the time-space effect. Fig. 1 illustrates the
principle and concept of the DAEM by comparing to the slope excava-
tion and the divided successive excavation. In the slope excavation as
shown in Fig. 1(a), the influence of unloading and the unloading width
is increased compared with a vertical-sided excavation method. This
will increase the uplift of the underlying tunnels during a single ex-
cavation. In the divided successive excavation, the soil in the excava-
tion is removed from one side to the other side in turn, so the con-
struction duration is long and the superimposed effect is significant as
shown in Fig. 1(b). To overcome the disadvantages of conventional
excavation methods and to better utilize the principle of the time-space
effect and stiffness of tunnel, the DAEM was proposed to decrease de-
formation of the metro tunnel and to shorten the construction time as
shown in Fig. 1(c). In this construction method, isolated piles and
330
Tunnelling and Underground Space Technology 71 (2018) 329–339
resistance piles are installed on the both sides of the tunnels. Dividing
walls are also constructed to divide the big excavation into small pits so
that a single unloading volume and lateral earth pressure can be re-
duced (Wang et al., 2003; Chen et al., 2017; Li et al., 2017a,b). 24
dividing walls are constructed before excavation with the spacing from
6 to 8 m, as shown in Fig. 2. These small pits are grouped and the pits in
a group are excavated simultaneously in each construction stages.
Three level struts are designed to support the retaining walls in each
small pit, and soil above the third level strut is excavated in layers,
leaving the last layer of soil with a thickness of 3–4 m. The corre-
sponding struts should be installed timely after each layer of soil is
excavated. In order to minimize the tunnel deformations, the last layer
of soil is excavated in blocks. After the construction of base slabs is
completed, the next construction stage can be carried out. The novel
construction method of DAEM was successfully applied in the project.
Although the deformation of the underlying metro tunnels was con-
trolled within the allowable value, the efficiency of this construction
method needs to be evaluated.
3. Numerical simulation
Numerical simulation is a feasible method to model the construction
process and capture the structural response of the metro tunnels to
adjacent excavations. In this section, the finite element (FE) software,
ABAQUS 6.10, is used to evaluate the deformation of the underlying
metro tunnel due to soil unloading in the highway tunnel excavation
with the construction method of DAEM.
3.1. Numerical model
Considering the skew angle between the highway tunnel and ex-
isting tunnels, it is necessary to model all existing tunnels in three-di-
mensional (3D) conditions. To eliminate the influence of boundary
conditions on numerical results, the dimensions of the model are set to
328 m in length, 150 m in width and 60 m in depth, as shown in Fig. 3.
As for boundary conditions, the horizontal movements of nodes on the
four side boundaries are restrained with vertical movement allowed. All
nodes at the bottom of the model are fixed.
The model consists of subsoil layers, retaining structures, struts,
slabs and tunnel linings. The soil mass is modeled by a total of 133,040
solid elements. The tunnel linings, slabs and retaining structures are
modeled as shell elements, and the struts are simulated using beam
elements. The structural model of the deep excavation and the under-
lying metro tunnels is presented in Fig. 4. 16,127 shell elements are
used for modeling the tunnel linings, slabs and retaining structures. The
struts are modeled by 1700 beam structural elements.
3.2. Input parameters
The concrete structure units of both the shell elements and beam
elements were modeled as an isotropic linear elastic material with a
Poisson’s ratio of 0.15. Considering stiffness reduction due to cracks in
the concrete, the stiffness of retaining walls, protecting walls, dividing
walls and struts is reduced by 20% (Lim et al., 2010) from the nominal
value with an input Young’s modulus of 24 GPa. To consider the in-
fluence of joint on the stiffness, the effective rigidity ratios of the lining
for circumferential and longitudinal direction are set to be 0.7 and 0.17
as suggested by Huang et al. (2013) and Chen et al. (2016). Thus the
input Young’s moduli in circumferential direction and longitudinal di-
rection for lining are 21 GPa and 5.1 GPa, respectively. For modeling
soil behavior in excavation, Hardening-Soil Model (HS) is now be-
coming prevalent among users of FEM analysis for its competence in
considering stress history of soils. Nevertheless, it still has difficulties in
defining reasonable model parameters due to lack of laboratory and
field tests. A fairly simple model with credible physical parameters is
preferred (Liao et al., 2013; Ou et al., 2013; Hsieh et al., 2013). The
M.-G. Li et al. Tunnelling and Underground Space Technology 71 (2018) 329–339

Fig. 1. Principle and concept of the DAEM.

Fig. 2. Layout of the supporting system of the DAEM.

331
M.-G. Li et al.
subsoil in the construction field is mainly soft soil comprising Qua-
ternary alluvial and marine deposits. In general, the soil surrounding
the metro tunnels and near the excavation surface is very soft silty clay.
In this paper, Modified Cam-Clay Model is adopted as the constitutive
relationship for subsoil in the numerical analyses (Roscoe and Burland,
1968).
The related parameters for Modified Cam-Clay Model are listed in
Table 1, in which λ, κ and M can be obtained using Eqs. (1)–(3), re-
spectively.
Cc
λ=
ln10
Cs
κ=
ln10
sinφ′
M=
3−sinφ′
where Cc and Cs are the compression index and rebound index respec-
tively. φ′ is effective friction angle of triaxial test. Poisson’s ratio (ν) can
be set equal to 0.3.
To minimize the unloading effect, soil in the pit and around the
excavation area is improved by the triple-shaft deep soil mixing (DSM)
method. Fig. 5 shows the range of the improved soil. The maximum
depth of the improved soil above the tunnel is 13.8 m. The improved
soil in the excavation range from 9.0 m below the ground surface (BGS)
to 26.0 m BGS. The maximum depth of the improved soil outside the
excavation is 26.0 m BGS. The improved soil is assumed as elastic
material, and its elasticity modulus is 240 MPa according to the em-
pirical laboratory test results.
Tunnelling and Underground Space Technology 71 (2018) 329–339
Fig. 3. Geometry and mesh of whole model.
3.3. Simulation process
According to the implementation of the DAEM in the project, the
excavation procedure is divided into four stages. By following principle
of the time-space effect and taking advantages of the tunnel stiffness
(Chen et al., 2015), the pits numbered 5, 11, 18, 20, 22, 24 and 25 are
excavated in the first stage and followed by the excavation of pits
numbered 1, 2, 3, 7, 13, 15 and 23. In the third stage, pits 6, 8, 10, 17
and 19 are excavated and the others are excavated in the final stage.
The numerical simulation procedure is consistent with the practical
excavation procedure to ensure the comparability of the numerical re-
(1) sults and the field data.
3.4. Analysis of the numerical results
(2)
The main concern of the project was the deformation of the un-
derlying metro tunnels. As mentioned before, the maximum heave of
the tunnels during excavation must be controlled within 16.3 mm
(3)
owing to the heave during soil improvement prior to excavation. To
ensure the safety of the running metro tunnels, a comprehensive long-
term monitoring program was carried out throughout the excavation
process. Fig. 6 presents a comparison of the vertical displacements of
the metro tunnels between the numerical results and the field data at
the completion of excavation. On the whole, the heave profiles of the
tunnels recorded from the FE model follow similar patterns to those
recorded from the field. The FE model slightly underestimates the
movements of the tunnels outside the excavation area. This might be
due to the following reasons: (i) it is difficult to simulate the excavation
process realistically since the management is easily influenced by many
factors in practice; (ii) the soil properties are significantly complex and
at present no constitutive model can model the soil behavior realisti-
cally. Despite this, the predicted maximum heaves and the locations of
Fig. 4. Structural model of DAEM.
332
M.-G. Li et al. Tunnelling and Underground Space Technology 71 (2018) 329–339

Table 1
Engineering properties of the soil layers.

No. Soil layer name H (m) w (%) γ (kN/m3) e c (kPa) φ λ κ M

1 Filled soil 1.8 18.0 2.0

3 Silty clay 7.2 34.7 18.2 1.13 12 22.0 0.081 0.007 0.86
4 Very soft silty clay 8.0 48.6 16.9 1.37 14 11.5 0.142 0.012 0.43
5-1 Clay 7.0 34.2 18.2 0.98 17 19.0 0.060 0.005 0.73
6 Hard clay 6.0 22.5 19.9 0.68 49 17.0 0.035 0.003 0.65
7-1 Sandy clay 5.0 29.9 18.7 0.84 3 33.0 0.024 0.002 1.33
7-2 Fine sand > 10 27 19 0.76 0 34.5 0.019 0.002 1.40

Note: H = Average thickness, w = water content, γ = unit weight, e = void ratio, c, φ = cohesion and friction angle of direct shear test, λ and κ are the slopes of the compression line
and the swelling-recompression line, respectively, M is the stress ratio at the critical state.

Fig. 5. Range of the improved soil.

Fig. 6. Comparison between numerical prediction and mea-

sured results.

Fig. 7. Conventional excavation methods for pit above

shield tunnels.

the maximum heave are in good agreement with the field data, which are reasonable and can be used for comparative studies of the con-
provide an evidence for further study. It should be noted that the aim of struction methods for this deep excavation above shield tunnels.
this paper is to perform the comparison and optimization of construc-
tion methods for deep excavation above shield tunnels rather than to
4. Comparative studies of construction methods
evaluate the numerical method. Therefore, it can be tentatively con-
cluded that the input soil parameters used in the numerical simulation
The time-space effect in deep excavations indicates that the

333
M.-G. Li et al. Tunnelling and Underground Space Technology 71 (2018) 329–339

Table 2 the PIEM, PPEM and DAEM, respectively. Moreover, the maximum
Comparison between previous projects and this project (Chen et al., 2015). tunnel heave in each project and related key factors that influence the
response of the tunnel are also listed in this table. It can be seen from
Projects He (m) Hf (m) L (m) δ (mm) Excavation
method the table that the values of the excavation depth, clearance and the span
length vary in each case and the effectiveness of construction methods
Yanggao Rd. 7.4 7 33 13 PIEM on controlling the tunnel deformation is unclear. In order to carry out
Undercrossing
the comparative study of these construction methods, this section pre-
Subsided square at 5.0 3 38 8
Nanjing Rd. sents two hypothetical construction schemes for this deep excavation
Metro air shaft at Huaihai 9.1 7.3 50 6.5 with PIEM and PPEM, respectively. Numerical models are developed to
Rd. simulate the hypothetical construction schemes, and the tunnels’ re-
Dongfang Rd. 6.5 3 18 17 PPEM sponses caused by the three construction methods are comparatively
Undercrossing
analyzed.
Agile square 4.9 3.6 120 8.5
Xin Jinqiao square 5.0 4 38 9.8
EWRTC (this project) 11.5 4 150 14.8 DAEM
4.1. Hypothetical schemes of conventional construction methods
Note: He: excavation depth, Hf = clearance between excavation base and the tunnel
crown, L = span length, δ = tunnel heave. In the hypothetical construction scheme of PIEM, soil around the
metro tunnels is improved to reduce the deformation of both the sup-
different excavation methods will result in different responses of the porting structures in the excavation and the metro tunnels. Dividing
underlying tunnels. Other than the DAEM, there are two conventional piles and resistance piles are installed at both sides of each tunnel to
construction methods that consider the time-space effect and are also divide the whole excavation into five small pits, as shown in Fig. 8. Soil
available for deep excavation above metro tunnels. The conventional around the tunnels, i.e. in zones D1, D3 and D5, is first removed in
construction methods are the parallel isolation excavation method layers and in blocks, leaving soil in zone D2 and D4 to resist the tunnel
(PIEM) and perpendicular partitioned excavation method (PPEM), deformation during the construction. After the construction of the base
which were proposed by Lu (2002) and Li et al. (2006), respectively. slab in these three zones is completed, soil is backfilled on the base slab
Fig. 7 delineates features of the PIEM and the PPEM. In the PIEM, di- to reduce the unloading. The remaining pits above the tunnels, i.e.
viding walls are installed on both sides of each tunnel and divide the zones D2 and D4, are excavated simultaneously. Since soil unloading
whole excavation into several small pits, as shown in Fig. 7(a). Soil in above metro tunnels has a serious impact, a layer of soil in zones D2 and
zone A1 and zone A3 are excavated simultaneously in layer and in D4 is removed first. Then the remaining soil is divided into seven
block, leaving zone A2 to resist the tunnel heave. After construction in segments based on the width of the zones D2 and D4 and these strips
zone A1 and zone A3 is finished, zone A2 is excavated in strips (Lu, are excavated alternately from the middle to the sides. After the com-
2002). In the PPEM, dividing walls should not be constructed since the pletion of each excavated block, the shaped steels inside the protecting
longitudinal retaining walls are close to the tunnel structure. To reduce walls at both sides of the tunnel are pulled out and then the base slab is
the tunnel heave, the overlying excavation is constructed in strips, and constructed immediately. To prevent further uplift of the tunnels, the
each strip is perpendicular to the axis of the tunnel, as shown in base slabs should be backfilled when their strength reaches a certain
Fig. 7(b). These small strips are so small that the principles of “time- value. Then the excavation of the next block can be carried out. After
space effect” and “tunnel stiffness” are better utilized (Li et al., 2006). the construction of all the base slabs is completed, the inner wall and
Table 2 lists several projects in practice, which were constructed by the roof are constructed in turn.
In the hypothetical construction scheme of PPEM, resistance piles

Fig. 8. Parallel Isolation Excavation Method (PIEM) for the

case history.

334
M.-G. Li et al.
are installed at both sides of each tunnel and the soil around the tunnels
is also improved to resist the tunnel heaves during excavation. The
excavation process is illustrated in Fig. 9. The first layer of soil is re-
moved as a whole. Then the excavation proceeds longitudinally from
one side to the other with slope excavation. During excavation, the
struts should be installed timely. Since the crossing length over metro
tunnels is 150 m, the pit is excavated in 41 small strips with the width
about 3 m, as shown in Fig. 9. Since the width of the excavation is 27 m,
each small strip should be divided into several blocks along the
335
Tunnelling and Underground Space Technology 71 (2018) 329–339
Fig. 9. Perpendicular Partitioned Excavation Method
(PPEM) for the case history.
Fig. 10. Structural model of PIEM.
Fig. 11. Structural model of PPEM.
transverse direction of the pit. After construction of each strip is com-
pleted and the base slab is reloaded, excavation of the next strip can be
carried out. When the construction of all the base slabs is completed,
the inner walls and the roofs are constructed in sequence.
4.2. Numerical simulation of conventional construction methods
Based on the hypothetical construction scheme of PIEM, a related
numerical model is built in the FE software. The geometry of the soil
M.-G. Li et al.
Fig. 12. Comparison of base heaves using different construction schemes: (a) DAEM, (b) PIEM, (c) PPEM.
model of PIEM is the same as that of DAEM, while their supporting
systems are different, as shown in Fig. 10. In the model of PIEM, a total
of 142,186 solid elements are used for the subsoil mass. 11,300 shell
elements are used for tunnel linings, slabs and retaining walls, dividing
piles and resistance piles. The struts are modeled by 1299 beam ele-
ments. Both the constitutive model and input parameters of soil and
structure element used in this model are the same as those in the model
with construction scheme of the DAEM.
The numerical model for the construction scheme of PPEM is built
in ABAQUS 6.10 with same geometry as those in previous models. The
layout of the retaining structures and the tunnels is shown in Fig. 11. A
total of 162,601 solid elements are used for the soil mass, and 13,190
shell elements are used for tunnel linings, slabs, retaining walls and
resistance piles. The struts are modeled by 992 beam elements. Both the
constitutive model and parameters are the same as those in the nu-
merical models with the construction scheme of DAEM and PIEM.
4.3. Results and analysis
Fig. 12 shows a comparison of the vertical displacement contours at
the completion of the excavation between different construction
336
Tunnelling and Underground Space Technology 71 (2018) 329–339
schemes. It can be seen that the maximum base heave of all the schemes
are located right above the existing tunnels. In the PIEM construction
scheme, the base heave is noticeable larger than the other construction
schemes with a maximum heave of 27.0 mm. The base heaves of the
DAEM and PPEM are well controlled with the maximum heave values
of 16.9 mm and 21.1 mm respectively. In DAEM and PPEM construction
schemes, the whole excavation is divided into dozens of small parts.
After the excavation of each part is finished, the bottom slab is con-
structed and the base of this part is reloaded. Thus the soil rebound can
be restrained and the heaves of the underlying tunnels are reduced
consequently.
Fig. 13 illustrates a comparison of the structural responses of the
metro tunnels to the deep excavation constructed by DAEM, PIEM, and
PPEM. Although the uplifted metro tunnels have similar profiles, the
tunnel heaves caused by excavation using different construction
methods are different in magnitude. The maximum heave caused by the
excavation with DAEM is noticeably smaller than that caused by the
excavation with PIEM and PPEM, which confirms the ability of the
DAEM in controlling the deformation of the metro tunnels located
under excavations.
To further understand the process of excavation-induced uplift of
M.-G. Li et al. Tunnelling and Underground Space Technology 71 (2018) 329–339

Fig. 13. Comparison of tunnel heaves using different con-

struction schemes.

Table 3 less. In construction schemes of PPEM and DAEM, the unloading vo-
Tunnel heaves caused by each construction stage. lume of soil in the first stage is almost the same as those in the second
stage. However, the magnitude of excavation-induced underlying
Construction method PIEM PPEM DAEM
tunnel heave in the first construction stage is larger than that in the
Construction stage Stage1 Stage2 Stage1 Stage2 Stage1 Stage2 second construction stage. Note that the maximum rebounds of the
underlying tunnels, with construction schemes of PIEM, PPEM and
Maximum heave (mm) 7.3 22.5 18.8 19.5 9.6 14.9
DAEM, are 22.5 mm, 19.5 mm and 14.9 mm, respectively. Since the
maximum tunnel heave must be controlled within 16.3 mm, only the
underlying metro tunnels, each construction scheme is divided into two DAEM construction scheme can be used in this project.
stages. In PIEM construction scheme, the first construction stage in-
cludes excavation of zones D1, D3 and D5 and excavation of the re- 5. Parametric studies of DAEM
maining zones is the second construction stage. In PPEM construction
scheme, excavation of the first strip to the 21th strip is the first con- 5.1. Effect of each small pit on tunnel heave
struction stage and excavation of the remaining strips is the second
construction stage. For DAEM construction scheme, excavation of the In DAEM, dividing walls are constructed to divide the excavation
odd-numbered pits is the first construction stage and excavation of the into twenty-five small pits and each pit can be excavated in-
even-numbered pits is the second construction stage. Table 3 lists the dependently. In this section, twenty-five 3D numerical simulations are
maximum tunnel heaves after each construction stage of the three conducted to study the impact of each small pit on the tunnels. Fig. 14
construction schemes. Results of PIEM construction scheme reveal that illustrates the maximum heave of the inbound and outbound caused by
the excavation right above the tunnels (D2 and D4) causes larger tunnel excavation of each independent pit. It can be seen from this figure that
heaves than the removing of soil around the tunnels (D1, D3 and D5), the responses of the inbound and outbound to excavations are similar.
even though the unload volume of soil right above the tunnels is much Despite the same unloading volume, the excavation right above the
tunnels causes the greatest influence on the tunnels and it decreases

Fig. 14. Tunnel heave caused by excavation of each in-

dependent small pit.

337
M.-G. Li et al.
Table 4
Tunnel heaves with different dividing wall intervals.
Wall intervals 6m 12 m 18 m 24 m
Tunnel heaves (mm) Inbound 14.9 15.5 16.0 16.2
Outbound 14.6 15.6 15.8 16.0
nonlinearly with the increase of distance between the small pits and the
tunnels. Therefore, the excavation scheme of the small pit right above
the tunnels should be improved for further optimization of the DAEM.
5.2. Effect of excavation procedure
To further reduce the excavation-induced tunnel heave, the con-
struction scheme needs to be optimized. Four excavation schemes are
designed with different grouping and excavation sequence. In this first
excavation scheme, these twenty-five small pits are excavated si-
multaneously. In the second excavation scheme, these twenty-five small
pits are divided into two groups, i.e. the odd-numbered pits in the first
group and the even-numbered pits in the second group, and the two
groups are excavated successively. In the third excavation scheme, the
small pits are divided into four groups with the pits numbered 1, 5, 9,
13, 17, 21, 25 in the group one, the pits numbered 2, 6, 10, 14, 18, 22 in
group two, the pits numbered 3, 7, 11, 15, 19, 23 in group three and the
pits numbered 4, 8, 12, 16, 20, 24 in group four, respectively. In this
excavation scheme, the small pits are excavated in turn from group one
to group four. In the fourth excavation scheme, the small pits are
grouped as in the third excavation scheme and the four groups are
excavated in order of group one, group three, group two and group
four.
The maximum tunnel heave caused by excavations using different
excavation procedures are 15.8 mm, 15.0 mm, 14.7 mm and 13.9 mm
respectively. It indicates that all the four excavation schemes can be
adopted to control the tunnel deformation with allowable value and the
construction risk is the lowest using the fourth excavation scheme, in
which the small pits are divided into four groups with these groups
excavated alternately.
5.3. Effect of dividing wall interval
In the DAEM construction scheme, dividing walls are constructed to
resist the tunnel deformation. In order to thoroughly understand the
effect of the dividing walls, four simulations are conducted with di-
viding wall intervals of 6 m, 12 m, 18 m and 24 m, respectively. Table 4
lists the numerical results of the maximum metro tunnel heaves at the
completion of excavation. As expected, the magnitude of the maximum
tunnel heaves increases with the increase of the dividing wall interval.
However, the effect of the wall interval on the tunnel heave is insig-
nificant. This might be attributed to the fact that all the investigated
intervals were within the ranges of spatial effect.
6. Suggestions for optimization of the construction method
Although the DAEM contributes to controlling the excavation-in-
duced deformation of the underlying shield tunnels, the construction
cost of the DAEM might be higher than those of the PIEM and PPEM
because of the installation of the dividing walls. To balance the cost and
effect, some recommendations for deep excavations above shield tun-
nels can be proposed based on the numerical studies and construction
experience in practice. First, regardless of the construction cost, the
DEAM has priority over the PIEM and the PPEM. Second, the interval of
dividing walls right above the tunnels should be as small as possible so
that a single unloading volume can be reduced. However, according to
previous construction experiences, the interval should not be less than
6 m considering the operability of the equipment in construction. Third,
to save construction cost, the interval of the other dividing walls can be
338
Tunnelling and Underground Space Technology 71 (2018) 329–339
increased gradually with the increase of the distance between the
tunnels and the small pits. Fourth, simultaneous excavation of alternate
small pits is suggested because it can save construction duration sig-
nificantly. As more and more metro lines have built in the cities, ex-
cavations adjacent to tunnels will be frequently encountered by en-
gineers in practice and the clearance between the excavation and the
tunnels will become smaller. To minimize the excavation-induced
tunnel deformation, additional construction measures are needed in the
DAEM.
7. Conclusions and suggestions
This paper presents a numerical study on excavation-induced de-
formation of underlying metro tunnels. 3D numerical simulations are
conducted to model three construction methods for a case history of a
deep excavation located above metro tunnels. Moreover, parametrical
studies are carried out to analyze the key factors of DAEM on the ex-
cavation-induced structural deformation of underlying metro tunnels.
Based on the numerical analyses, following conclusions can be drawn.
(1) Comparative analyses indicate that tunnel deformation due to
above unloading is closely related to the excavation methods. The
maximum vertical displacements of the metro tunnels caused by
construction schemes of PIEM, PPEM and DAEM are 22.5 mm,
19.5 mm and 14.9 mm, respectively. Comparing to the PIEM and
PPEM, the DAEM shows a better efficiency in controlling the tunnel
deformation by following the time-space effect and taking the ad-
vantage of the tunnel stiffness.
(2) Parametric studies reveal that unloading right above the tunnels
causes greatest influence on the tunnel deformations and the in-
fluence decreases nonlinearly with the increase of distance between
the pits and the tunnels. The deformation of the underlying tunnels
can be well controlled if the small pits are divided into several
groups with these groups excavated alternately. The maximum
tunnel heaves increase slightly with the increase of the dividing
wall interval.
(3) Based on the numerical analysis, suggestions can be proposed for
similar projects in the future. The DEAM has priority over the PIEM
and the PPEM regardless of the construction cost. The interval of
dividing walls right above the tunnels should be as small as possible
so that a single unloading volume can be reduced. To save con-
struction cost, the interval of the other dividing walls can be in-
creased gradually with the increase of the distance between the
tunnels and the small pits. Simultaneous excavation of alternate
small pits is suggested because it can save construction duration
significantly.
Acknowledgements
The financial support from the National Natural Science Foundation
of China (NSFC Grant No. 41602283, 41372282, 41330633) is grate-
fully acknowledged.
References
ABAQUS [Computer software]. ABAQUS user manual, ABAQUS.
Chai, J.C., Shen, S.L., Ding, W.Q., Zhu, H.H., Carter, J., 2014. Numerical investigation of
the failure of a building in Shanghai, China. Comput. Geotech. 55, 482–493.
Chang, C.T., Sun, C.W., Duann, S.W., Hwang, R.N., 2001. Response of a Taipei Rapid
Transit System (TRTS) tunnel to adjacent excavation. Tunn. Undergr. Space Technol.
16 (3), 151–158.
Chen, J.J., Wang, J.H., Xiang, G.W., Wen, S.L., Du, Y., 2011. Numerical study on the
movement of existing tunnel due to deep excavation in Shanghai. Geotech. Eng. J.
SEAGS & AGSSEA 40 (3), 30–40.
Chen, J.J., Zhang, L.Y., Zhang, J.F., Zhu, Y.F., Wang, J.H., 2013. Field tests, modification,
and application of deep soil mixing method in soft clay. J. Geotech. Geoenviron. Eng.
139 (1), 24–34.
Chen, J.J., Zhu, Y.F., Li, M.G., Wen, S.L., 2015. Novel excavation and construction
method of an underground highway tunnel above operating metro tunnels. J. Aerosp.
M.-G. Li et al.
Eng. 28 (6), A4014003.
Chen, R.P., Meng, F.F., Li, Z.C., Ye, Y.H., Ye, J.N., 2016. Investigation of response of
metro tunnels due to adjacent large excavation and protective measures in soft soils.
Tunn. Undergr. Space Technol. 58, 224–235.
Chen, J.J., Li, M.G., Wang, J.H., 2017. Active Earth pressure against rigid retaining walls
subjected to confined cohesionless soil. Int. J. Geomech. 17 (6).
Cooper, M.L., Chapman, D.N., Rogers, C.D.F., Chan, A.H.C., 2002. Movements in the
Piccadilly Line tunnels due to the Heathrow Express construction. Geotechnique 52
(4), 243–257.
Dolezalova, M., 2001. Tunnel complex unloaded by a deep excavation. Comput. Geotech.
28, 469–493.
Ferrada, X., Serpell, A., 2014. Selection of construction methods for construction projects:
a knowledge problem. J. Constr. Eng. Manage. 140 (4), B4014002.
Hsieh, P.G., Ou, C.Y., Lin, Y.L., 2013. Three-dimensional numerical analysis of deep ex-
cavations with cross walls. Acta Geotech. 8 (1), 33–48.
Hu, Z.F., Yue, Z.Q., Zhou, J., Tham, L.G., 2003. Design and construction of a deep ex-
cavation 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 under-
neath an excavation. Tunn. Undergr. Space Technol. 21 (3), 399.
Huang, X., Schweiger, H.F., Huang, H.W., 2013. Influence of deep excavations on nearby
existing tunnels. Int. J. Geomech. 13 (2), 170–180.
Li, Z.G., Liu, G.B., Zeng, Y., Wu, X.J., 2006. Control measures of tunnel displacement by
over-excavation unloading. Chin. J. Undergr. Space Eng. 2 (3), 430–433.
Li, M.G., Chen, J.J., Xu, A.J., Xia, X.H., Wang, J.H., 2014. Case study of innovative top-
down construction method with channel-type excavation. J. Constr. Eng. Manage.
140 (5), 05014003.
Li, M.G., Chen, J.J., Wang, J.H., 2017a. Arching effect on lateral pressure of confined
granular material: numerical and theoretical analysis. Granul. Matt. 19 (2), 20.
Li, M.G., Zhang, Z.J., Chen, J.J., Wang, J.H., Xu, A.J., 2017b. Zoned and staged con-
struction of an underground complex in Shanghai soft clay. Tunn. Undergr. Space
Technol. 67, 187–200.
Liao, S.M., Fan, Y.Y., Shi, Z.H., Shao, W., Kong, X.P., 2013. Optimization study on the
reconstruction and expansion of an underground rail transit center in Shanghai soft
ground. Tunn. Undergr. Space Technol. 38, 435–446.
Lim, A., Ou, C.Y., Hsieh, P.G., 2010. Evaluation of clay constitutive models for analysis of
deep excavation under undrained conditions. J. Geoeng. 5 (1), 9–20.
Liu, H.L., Li, P., Liu, J.Y., 2011. Numerical investigation of underlying tunnel heave
during a new tunnel construction. Tunn. Undergr. Space Technol. 26 (2), 276–283.
Lu, H.W., 2002. Design of foundation pit of protection segment of grade separation metro
under Yanggao Road in Century Avenue. Urban Roads Bridges Flood Control 2, 44–48
(in Chinese).
339
Tunnelling and Underground Space Technology 71 (2018) 329–339
Lu, J.F., Hanyga, A., 2005a. Wave field simulation for heterogeneous porous media with
singular memory drag force. J. Comput. Phys. 208 (2), 651–674.
Lu, J.F., Hanyga, A., 2005b. Fundamental solution for a layered porous half space subject
to a vertical point force or a point fluid source. Comput. Mech. 35 (5), 376–391.
Ng, C.W., Shi, J.W., Hong, Y., 2013. Three-dimensional centrifuge modelling of basement
excavation effects on an existing tunnel in dry sand. Can. Geotech. J. 50, 874–888.
Ng, C.W.W., Shi, J.W., Masin, D., Sun, H.S., Lei, G.H., 2015a. Influence of sand density
and retaining wall stiffness on three-dimensional responses of tunnel to basement
excavation. Can. Geotech. J. 52 (11), 1811–1829.
Ng, C.W.W., Sun, H.S., Lei, G.H., Shi, J.W., Masin, D., 2015b. Ability of three different soil
constitutive models to predict a tunnel's response to basement excavation. Can.
Geotech. J. 52 (11), 1685–1698.
Ou, C.Y., Hsieh, P.G., Lin, Y.L., 2013. A parametric study of wall deflections in deep
excavations with the installation of cross walls. Comput. Geotech. 50, 55–65.
Roscoe, K.H., Burland, J.B., 1968. On the generalized stress–strain behaviour of ‘wet’ clay.
In: Heyman, J., Leckie, F.A. (Eds.), Engineering Plasticity. Cambridge University
Press, pp. 535–609.
Shanghai Municipal Engineering Authority, 2000. Specification for excavation in
Shanghai metro construction. SZ-08-2000, Shanghai, China (in Chinese).
Sharma, J.S., Hefny, A.M., Zhao, J., Chan, C.W., 2001. Effect of large excavation on de-
formation of adjacent MRT tunnels. Tunn. Undergr. Space Technol. 16 (2), 93–98.
Son, M., Cording, E.J., 2005. Estimation of building damage due to excavation-induced
ground movements. J. Geotech. Geoenviron. Eng. 131 (2), 162–177.
Tan, Y., Li, X., Kang, Z.J., Liu, J.X., Zhu, Y.B., 2015. Zoned excavation of an oversized pit
close to an existing metro line in stiff clay: case study. J. Perform. Constr. Facil. 29
(6), 04014158.
Wang, J.H., Zhou, X.L., Lu, J.F., 2003. Dynamic response of pile groups embedded in a
poroelastic medium. Soil Dynam. Earthq. Eng. 23 (3), 53–60.
Wang, J.H., Xu, Z.H., Wang, W.D., 2010. Wall and ground movements due to deep ex-
cavations in shanghai soft soils. J. Geotech.Geoenviron. Eng. 136 (7), 985–994.
Zhang, Z.G., Huang, M.S., Wang, W.D., 2013. Evaluation of deformation response for
adjacent tunnels due to soil unloading in excavation engineering. Tunn. Undergr.
Space Technol. 38, 244–253.
Zhang, L.M., Wu, X.G., Liu, H.T., 2016. Strategies to reduce ground settlement from
shallow tunnel excavation: a case study in China. J. Constr. Eng. Manage. 142 (5),
04016001.
Zhang, Z.J., Li, M.G., Chen, J.J., Wang, J.H., Zeng, F.Y., 2017. Innovative construction
method for oversized excavations with bipartition walls. J. Constr. Eng. Manage. 143
(8), 04017056.
Zheng, G., Wei, S.W., 2008. Numerical analyses of influence of overlying pit excavation
on existing tunnels. J Centr. South Univ. Technol. 15, 69–75.