applied

sciences
Article
Combined Application of Pipe Roof Pre-SUPPORT
and Curtain Grouting Pre-Reinforcement in Closely
Spaced Large Span Triple Tunnels
Ran Li 1 , Dingli Zhang 1, *, Peng Wu 2 , Qian Fang 1 , Ao Li 1 and Liqiang Cao 1
1 Key Laboratory for Urban Underground Engineering of Ministry of Education, Beijing Jiaotong University,
Beijing 100044, China; 17115300@bjtu.edu.cn (R.L.); qfang@bjtu.edu.cn (Q.F.); 15115279@bjtu.edu.cn (A.L.);
14115304@bjtu.edu.cn (L.C.)
2 China Highway Engineering Consulting Group Company LTD, Beijing 100088, China; 16121127@bjtu.edu.cn
* Correspondence: dlzhang@bjtu.edu.cn; Tel.: +86-10-5168-8111




Received: 13 April 2020; Accepted: 30 April 2020; Published: 2 May 2020


Abstract: The platform of the Badaling Great Wall Underground Station consisted of large span
triple tunnels. The triple tunnels passed through several major faults. Based on preliminary in situ
monitoring, the interaction behavior among triple tunnels was significantly severe, and the original
supports failed to meet the safety requirement. Therefore, pipe roof pre-support (PRPS) and curtain
grouting pre-reinforcement (CGPR) were used as the pre-construction techniques. By numerical
investigation, the pipe diameter, the arrangement angle of PRPS, and the grouting thickness of
CGPR were optimized as 108 mm, 150◦ , and 3 m, respectively. According to numerical results,
PRPS predominantly bore loose rock and reduced mean crown settlement (MCS), while CGPR
primarily improved the mechanical properties of rock and decreased the pillar plastic ratio (PPR).
PRPS and CGPR had complementary advantages in time, space, and mechanical properties.
PRPS could timely form a strong arch structure far ahead of the tunnel face, CGPR could effectively
restrict the development of plastic zone. They mutually and actively formed a strong permanent
ring in front of the tunnel face. After the field application of PRPS and CGPR, the mean surrounding
rock pressure reduced by 33.4%, the MCS reduced by 58.7%, and no support damage was observed.
The excavation safety was guaranteed.

Keywords: triple tunnels; large span; pipe roof; curtain grouting; field monitoring;
numerical simulation

1. Introduction
For large span tunnels excavated in the weak and broken surrounding ground,
excessive deformation and potential damage may take place ahead of the excavation face.
Pre-construction techniques have been extensively used to minimize pre-deformation and pre-failure
of the rock mass. Pre-construction techniques can be divided into pre-support techniques and
pre-reinforcement techniques. The core of pre-support techniques is transferring and bearing the
rock load to prevent tunnel collapse, while the core of pre-reinforcement techniques is improving the
strength and rigidity of surrounding rock to maintain tunnel stability. In tunnel engineering, the pipe
roof and curtain grouting are commonly used as the representative techniques of pre-support and
pre-reinforcement, respectively.
Pipe roof pre-support (PRPS) consists of grouted steel pipes with large moment inertia. These pipes
are circumferentially inserted from the tunnel face with the pipe diameter of 50–200 mm and the
advanced length of 20–50 m. With the great stiffness of the steel pipe, the PRPS system could redistribute
the stress state ahead of the tunnel face and eventually decrease the ground settlement [1]. Since its first

Appl. Sci. 2020, 10, 3186; doi:10.3390/app10093186 www.mdpi.com/journal/applsci

Appl. Sci. 2020, 10, 3186 2 of 18

application in Japan in 1971, the PRPS method has been successfully applied to minimize excavation
interference and prevent ground collapse in the weak stratum or crossing projects. Ocak [2] found
that PRPS could effectively control the ground settlements caused by metro tunnel excavation in
the clayey ground. The PRPS method was approximately three times more efficient than the new
Austrian tunneling method (NATM) in controlling surface settlement in the same geological ground.
Volkmann et al. [3] compared two pipe roof installation methods, including the cased-drilling method
and the predrilling method. By on-site measurement with the horizontal inclinometer, they summarized
that the cased-drilling method was preferred in strain sensitive ground owing to its better performance
in preventing water saturation and flushing erosion. Rostami et al. [4] concluded that PRPS could
significantly improve roof conditions by bearing and transferring the vertical pressure. They also
noticed that with the increase of the diameter and amount of PRPS, the ground subsidence was
considerably reduced. Schumacher et al. [5] numerically modeled the pipe roof with beam elements,
and they found that the bending moments of the pipe roof varied with different pipe diameter and pipe
spacing. A design criterion for PRPS was put forward based on the numerical bending moment and
the analytical bending stress derived by the flexure formula. Then, they suggested a broken rock mass
that the pipe roof should hold up ranged from 1.5 to 3 m. Zhang et al. [1] proposed a double-parameter
elastic beam model to analytically investigate the working mechanism of PRPS used in the tunnel
entrance. The results showed that the load carried by PRPS greatly increased with the deterioration
of the surrounding ground, and the load transferring capacity improved with the stiffness increase
of primary lining. Niedbalski et al. [6] underlined that the PRPS was necessarily demanded in the
excavation of the Laliki tunnel in the Carpathian flysch. Through field monitoring, they verified the
important role of PRPS in preventing rock falling and reducing surface deformation.
Curtain grouting pre-reinforcement (CGPR) is injecting cement-based slurry ahead of the tunnel
face with the grouting thickness of 1–10 m and the diffusion distance of 10–25 m. Owing to the filling,
permeating, compaction, and splitting effect of the CGPR, the broken rock mass could be reinforced,
and the ground permeability could be reduced. The CGPR method has already been widely adopted
in dam construction [7,8] and slope stabilization [9,10]. In complex tunneling engineering, the grouting
method has also shown a significant effect in ground consolidation [11], settlement compensation [12],
shear strength enhancement, and water tightness increment. Ganerød et al. [13] reported the
pre-grouting application in two subsea tunnels with fault zones, and they observed that pre-grouting
volume considerably varied in different fault-related sub-zones. The injected cement ratio among
the fault core, inner damage zone, and outer damage zone was given as 1:2:1. Stille et al. [14]
reviewed the grouting project of the Namntall Tunnel, and they found the geologic and hydraulic
conditions of rock mass greatly influenced the grouting performance. Then, they suggested a novel
method to predict hydrogeological conditions and to optimize the grouting process. Zhang et al. [15]
presented a practical curtain grouting case in the highly weathered rock mass of Xiang’an subsea
tunnel, the process of double compaction grouting and double fracture grouting were elaborated.
And the subsea tunnel safely passed through the fault zones with the ground treatment of CGPR.
Moreover, Fang et al. [16] proposed a convergence-confinement model [17] to theoretically study the
reinforcing effect of CGPR in a circular tunnel, they found the plastic zone may initiate from the outer
boundary of the reinforced zone, the inner boundary of the reinforced zone, or both the outer and the
inner boundaries simultaneously. Six different configurations were deduced based on the distribution
and extent of the plastic zone(s) and the grouting thickness. Seven critical unconfinement factors were
solved to define the possible transition between two consecutive configurations.
Previous studies primarily focused on the independent application of PRPS or CGPR in the
single tunnel. However, for the closely spaced tunnels excavated in the fault zone or jointed zone,
an independent pre-construction technique may fail to meet the high safety requirements. In some
cases, the PRPS and the CGPR should be adopted jointly. Besides, due to the inevitably complicated
interaction behavior between neighboring tunnels [18], the effects of pre-construction techniques are
more challenging to evaluate.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 3 of 19

Appl. Sci. 2020, 10, 3186 3 of 18

interaction behavior between neighboring tunnels [18], the effects of pre-construction techniques are
more challenging to evaluate.
This paper
paper reports
reportsthe
thecombined
combinedapplication
application of of
PRPSPRPSandand
CGPR
CGPRin the
in excavation of large
the excavation span
of large
closely spacedspaced
span closely triple tunnels of the Badaling
triple tunnels Great Wall
of the Badaling Underground
Great Station. In
Wall Underground situ data,
Station. In including
situ data,
surrounding rock pressure
including surrounding rockand crown and
pressure settlement, were initially
crown settlement, monitored
were initiallytomonitored
show the urgent
to showneed
the
for pre-constructions.
urgent ReasonableReasonable
need for pre-constructions. design parameters of PRPSofand
design parameters PRPSCGPR werewere
and CGPR determined
determinedby
numerical
by numericaloptimization.
optimization.TheThe
working mechanism
working mechanism of the PRPS
of the andand
PRPS the the
CGPRCGPRwere investigated
were investigatedby
comparative
by comparativeanalysis. TheThe
analysis. mutual
mutualeffects of of
effects PRPS
PRPSand
andCGPR
CGPRwere
wereevaluated
evaluatedandandverified
verified byby the
numerical analysis and the engineering application. This study may provide a practical reference for
other similar tunneling projects.

2. Site
Site Description
Description and Experimental Section Monitoring

2.1. Project
2.1. Project Overview
Overview
The Badaling
The Badaling Great
Great Wall
Wall Underground
Underground Station Station isis the
the largest
largest underground
underground high-speed
high-speed railway
railway
station in the world. This station will serve for the 2022 Beijing
station in the world. This station will serve for the 2022 Beijing Winter Olympics. MoreWinter Olympics. More information
information
associated with
associated with the
theexcavation
excavationdesign
designand andthe thestructural
structural response
response of of
thisthis
station could
station couldbe found
be found in the
in
studies by Zhang’s research group [19–22]. This station consists of the concourse
the studies by Zhang’s research group [19–22]. This station consists of the concourse and the platform, and the platform,
and the
and the platform
platform waswas constructed
constructed earlier
earlier than
than the
the concourse.
concourse. The The platform,
platform, withwith aa length
length ofof 398
398 m m
(from DK67 + 851 to DK68 + 249), is composed of triple tunnels with
(from DK67 + 851 to DK68 + 249), is composed of triple tunnels with narrow pillars. The triple tunnelsnarrow pillars. The triple
tunnels
are deeplyareburied
deeplywith
buriedthewith the overburden
overburden depthfrom
depth ranges ranges 70 from
to 10270mto(Figure
102 m (Figure 1). The
1). The triple triple
tunnels
tunnels
were were excavated
mainly mainly excavated in moderately–completely
in moderately–completely weathered weathered
monzonite monzonite
granite. granite.
There are There are
several
several faults (F1, F2, and F3) and a sizeable jointed zone in the tunnel
faults (F1, F2, and F3) and a sizeable jointed zone in the tunnel site. The faults are torsion- site. The faults are torsion-
compression fractures,
compression fractures,and andthethe
fault zones
fault primarily
zones comprise
primarily comprisecrushed weak weak
crushed rock. The
rock. sizeable jointed
The sizeable
zone intersects
jointed the tunnel
zone intersects thebound
tunnelranging
bound fromrangingDK67from+ 900DK67 + 900 +
to DK68 to000,
DK68and+ the
000,dip andandthestrike
dip and are
86 ◦ and 335◦ , respectively. The surrounding rock mass in the fault zone and the jointed zone is generally
strike are 86°and 335°, respectively. The surrounding rock mass in the fault zone and the jointed zone
broken
is withbroken
generally 8–10 setswithof considerably developed joints,
8–10 sets of considerably most of
developed themmost
joints, are closed
of them joints without
are closed any
joints
fillings. Referring to the basic quality classification system (BQ system) in
without any fillings. Referring to the basic quality classification system (BQ system) in China, the China, the surrounding rock
in the fault zone
surrounding wasinclassified
rock the faultaszone
class was
V. Table 1 [23] shows
classified as class the V.
simplified
Table 1relationship
[23] showsofthe thesimplified
rock mass
qualities
relationshipbetween therock
of the BQ system and the widely
mass qualities between used
thequality
BQ system(Q) system.
and the Besides,
widelytheused
triplequality
tunnels (Q) are
located in the middle–strong water-rich area, and the normal yield of groundwater
system. Besides, the triple tunnels are located in the middle–strong water-rich area, and the normal is approximately
of /d.
2400 m 3
yield groundwater is approximately 2400 m3/d.

Zhang Jiakou Excavation Direction Beijing

DK68+249 DK67+851
Triple tunnels
Legend
700
50
Granite 58 67
670 85
Monzonitic Granite 86
Elevation (m)

86
Joint 640

Fault 610
Tunnel bound
580
Test section
550
Monitoring section F1 76 58 F2 236 80 F3 161 35 Jointed zone 335 86
Rock mass quality V (poor) III (fair) V (poor) III (fair)
Chainage DK68+200 DK68+000 DK67+900
Test section Applied section
Monitoring section
(DK68+220) (DK67+950)

Geological profile of triple tunnels.

Figure 1. Geological
Appl. Sci. 2020, 10, x FOR PEER REVIEW 4 of 19
Appl. Sci. 2020, 10, 3186 4 of 18

Table 1. Relationship between the basic quality classification system (BQ system) and the quality
Table 1. Relationship between the basicsystem
quality(Qclassification
system). system (BQ system) and the quality
system (Q system).
CLASS I (Very CLASS II CLASS III CLASS IV CLASS V (Very
Value
Value Good) (Good)
CLASS I (Very Good)
CLASS II (Fair)CLASS III (Poor)
CLASS IV Poor)
CLASS V
(Good) (Fair) (Poor) (Very Poor)
BQ >550 451–550 351–450 251–350 <250
BQ >550 451–550 351–450 251–350 <250
Q Q >40 >40 10–40 10–40 4–10 4–10 1–41–4 <1<1

The triple tunnels are symmetrically distributed (Figure 2), with the Left–Right–Middle
The triple tunnels are symmetrically distributed (Figure 2), with the Left–Right–Middle excavation
excavation sequence. The left and the right tunnels, having the same horseshoe cross-sections, are
sequence. The left and the right tunnels, having the same horseshoe cross-sections, are arrival–departure
arrival–departure single-track tunnels, and the excavation span and height are 15.94 and 12.15 m,
single-track tunnels, and the excavation span and height are 15.94 and 12.15 m, respectively. The middle
respectively. The middle tunnel is a double-track tunnel with a 14.38 m excavation span and 12.24 m
tunnel is a double-track tunnel with a 14.38 m excavation span and 12.24 m height. The pillar width is
height. The pillar width is narrowly 6 m (i.e., 0.38 D, D is the excavation span of both left and right
narrowly 6 m (i.e., 0.38 D, D is the excavation span of both left and right tunnels). The triple tunnels
tunnels). The triple tunnels were excavated using the drilling and blasting method. Each tunnel was
were excavated using the drilling and blasting method. Each tunnel was excavated in three parts,
excavated in three parts, including the top heading, bench, and the invert. Thus, there were a total of
including the top heading, bench, and the invert. Thus, there were a total of nine excavation stages.
nine excavation stages.

Figure 2.
Figure 2. Excavation sequence and
Excavation sequence and geometry
geometry layout
layout of
of triple
triple tunnels
tunnels (Unit:
(Unit: m).
m).

2.2. Field Monitoring

2.2. Field Monitoring of
of Experimental
Experimental Section
Section
In
In the
the initial
initial excavation
excavation period,
period, thethe section
section from DK68 ++ 249
from DK68 249 to DK68 +
to DK68 200 was
+ 200 was selected
selected asas the
the
preliminary
preliminary experimental
experimental section
section toto reveal
reveal the
the interaction
interaction characteristics
characteristics of of the
the triple
triple tunnels.
tunnels. InIn the
the
experimental section, the regular supports, which had been successfully applied in a previous single
tunnel excavation, were
were adopted
adopted for for the triple tunnels.
tunnels. The primary lining consisted of C30 shotcrete
with
with 300
300mmmmthickness, steel
thickness, frame
steel framewithwith mm2 mm
4213 4213 cross-sectional area, and
2 cross-sectional area,rockandbolt (radially
rock inserted
bolt (radially
around tunnel contour) with 4.5 m length. The secondary lining was 400 mm thick
inserted around tunnel contour) with 4.5 m length. The secondary lining was 400 mm thick reinforced reinforced concrete
with a 28-day
concrete with compressive strength ofstrength
a 28-day compressive 35 MPa.ofThe 35 12
MPa.mmThe thick
12 composite
mm thick waterproofing membrane
composite waterproofing
was sandwiched
membrane between thebetween
was sandwiched primarythe andprimary
secondary andlinings.
secondary linings.
In the
theexperimental
experimentalsection,
section,thethe monitoring
monitoring layout layout is described
is described in Figurein 3. Figure 3. The
The crown crown
settlement
settlement
was monitored wasvia monitored via thewith
the total station total station with
a distance interval a distance interval
of 5 m (Figure 4a). of
The5 DK68
m (Figure 4a).
+ 220 was
The DK68 +selected
particularly 220 wasfor particularly
surrounding selected
rock for surrounding
pressure monitoring.rock For
pressure
each monitoring.
tunnel, sevenFor each
double-
tunnel,
membrane seven double-membrane
pressure pressure
cells were fixed cellssteel
at the wereframe
fixed and
at thedirectly
steel frame andthe
against directly
rock against the rock
to monitor the
to monitor therock
surrounding surrounding
pressure rock pressure
(Figure (Figure
4b). The 4b). The
pressure pressure
cells were cells wereimmediately
installed installed immediately
after the
after the excavation
excavation of each
of each part parttunnel.
of the of the tunnel. Therefore,
Therefore, the pressure
the pressure cells atcells
the at the knees
knees were were installed
installed later
later
than than others.
others. The measured
The measured pressure
pressure data were
data were manually
manually collected
collected by thebyacquisition
the acquisition instrument
instrument with
with the frequency
the frequency onceonce
per per
day.day. Noted
Noted thatthat
allall monitoring
monitoring instrumentswere
instruments wereimplemented
implemented behind
behind the
tunnel face, which meant the pre-deformation and the pre-load were
pre-deformation and the pre-load were not observed. not observed.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 5 of 19
Appl. Sci. 2020, 10, 3186 5 of 18
Appl. Sci. 2020, 10, x FOR PEER REVIEW Crown 5 of 19

Left Right
shoulder Crown shoulder
Left Crown settlement Right
shoulder Primary lining shoulder
Left Crown settlement Right
Second lining
waist Primary lining waist
Left Right
Pressure cell lining
Second
waist
Left waist
Right
knee Pressure cell knee
Left Right
knee knee
Figure 3. Layout of monitoring points for each tunnel (monitoring instruments at the knees were
Figure 3. Layout of monitoring points for each tunnel (monitoring instruments at the knees were
installed later).
installed
Figure 3.later).
Layout of monitoring points for each tunnel (monitoring instruments at the knees were
installed later).

(a) (b)

Figure 4. Installation of
(a)monitoring instrument: (a) Total station; (b) Double-membrane
(b) pressure cell.

2.3. Monitoring Results of

of of
Figure 4. Installation
2.3. Monitoring Results Experimental Section
monitoring instrument:
Experimental Section (a) Total station; (b) Double-membrane pressure cell.

During theconstruction
During the constructionofofthe
the experimental
experimental section,
section, steelsteel
frame frame distortion,
distortion, shotcrete
shotcrete crack,crack,
small
2.3. Monitoring Results of Experimental Section
small scale tunnel collapse, and primary lining leakage were observed several times
scale tunnel collapse, and primary lining leakage were observed several times (Figure 5). (Figure 5).
During
The the construction
development of the experimental
of the surrounding section,
rock pressure at steel + 220distortion,
DK68frame is plotted shotcrete crack,
in Figure 6. The small
final
scale tunnel collapse, and primary lining leakage were observed several times (Figure
distribution of the surrounding rock pressure is shown in Figure 7. The positive value indicates 5).
a compressive force between the primary lining and rock mass. The stable distribution of the crown
settlement of triple tunnels in the experimental section is shown in Figure 8.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 6 of 19
Appl. Sci. 2020, 10, 3186 6 of 18

(a) (b)

(c) (d)
Figure
Figure 5.
5. Tunnel damages in
Tunnel damages the experimental
in the experimental section:
section: (a)
(a) Steel
Steel frame
frame distortion;
distortion; (b)
(b) Shotcrete
Shotcrete crack;
crack;
(c) Small scale tunnel collapse; (d) Primary lining leakage.

The development of the surrounding rock pressure at DK68 + 220 is plotted in Figure 6. The final
distribution of the surrounding rock pressure is shown in Figure 7. The positive value indicates a
compressive force between the primary lining and rock mass. The stable distribution of the crown
settlement of triple tunnels in the experimental section is shown in Figure 8.
Appl.
Appl. Sci. 2020, 10,
Sci. 2020, 10, 3186
x FOR PEER REVIEW 77 of 18
19

Stage 4 Stage 5&6 Stage 7- 9

Stage 1 Stage 2&3 Stage 4~6 Stage 7~9
Left knee Left waist Left shoulder
350 Crown Right shoulder Right waist
Left knee Left waist Left shoulder 450 Right knee
300 Crown Right shoulder Right waist
400
Right knee
250 350
300
200
Pressure (kPa)

Pressure (kPa)
250
150 200
100 150
100
50
50
0 0
① Pillar ⑦ Pillar ④ ① Pillar ⑦ Pillar ④
-50 -50
② ⑧ ⑤ ② ⑧ ⑤
③ ⑨ ⑥ -100 ③ ⑨ ⑥
-100
0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 60 80 100 120 140 160 180 200 220 240 260 280
Time (d) Time (d)

(a) (b)

Stage 7 Stage 8&9

Left knee Left waist Left shoulder

600
Crown Right shoulder Right waist
550 Right knee
500
450
400
350
Pressure (kPa)

300
250
200
150
100
50
① Pillar ⑦ Pillar ④
0
-50 ② ⑧ ⑤
③ ⑨ ⑥
-100
160 180 200 220 240 260
Time (d)

(c)
Figure 6.
Figure 6. Development
Development of
of surrounding
surrounding rock
rock pressure
pressureat
atsection DK68++ 220:
sectionDK68 220: (a) Left
Left tunnel;
tunnel; (b)
(b) Right
Right
tunnel; (c) Middle tunnel.
tunnel; (c) Middle tunnel.

547
291 351
Damaged
220 383 453 312 240

① ⑦ ④
49 72 95 115 102 56
② ⑧ ⑤
84 ③ 122 130 ⑨ 208 128 ⑥ 91
Left tunnel Middle tunnel Right tunnel
Figure7.
Figure 7. Distribution
Distribution of
of surrounding
surrounding rock
rockpressure
pressureat
atsection DK68++ 220
sectionDK68 220 (Unit: kPa).
(Unit: kPa).
Appl. Sci. 2020, 10, 3186 8 of 18
Appl. Sci. 2020, 10, x FOR PEER REVIEW 8 of 19

80 DK68+200
DK68+210
70 DK68+220
DK68+230
DK68+240

Crown settlement (mm)

60

50

40

30

20

10
1 2 3
Left tunnel Middle tunnel Right tunnel
Triple tunnels

Figure
Figure 8. Crown settlements
8. Crown settlements of
of triple
triple tunnels
tunnels in
in the
the experimental
experimental section.
section.

According to the monitoring data, we noted the following trends:

1. For each tunnel, with the stress release of the surrounding ground, the surrounding rock pressure
jumped up sharply in the earlier periods. The pressure increments during the top heading
excavations primarily accounted for 80% of the final pressure value. It could be inferred that the
steel frame and the shotcrete immediately bear rock load after installation.
2. The crown and shoulders were the least favorable parts with considerably larger surrounding
rock pressures ranging from 220 to 550 kPa. This trend was also validated by more structural
failures, and more small scale tunnel collapses observed at the crown and shoulders.
3. The middle tunnel was found to be the most unfavorable tunnel among the triple tunnels.
The mean surrounding rock pressure and the mean crown settlement of the middle tunnel were
275.9 kPa and 60.9 mm, considerably higher than that of the left tunnel (139.7 kPa and 38.8 mm)
and the right tunnel (182.9 kPa and 33.6 mm). This behavior might be expressed by that the
middle tunnel was excavated in the weak surrounding ground, which had already been severely
disturbed by the leading tunnel excavations.
4. Pillar width and excavation sequence greatly influenced the multiple interactions among triple
tunnels. Due to the pillar width of 1.65D, the left tunnel was hardly affected by the right tunnel
excavation. Conversely, the left tunnel is heavily influenced by the middle tunnel excavation
because of the rock pillar of 0.38D. This finding also showed that, with the decrease of pillar width,
the rock pillar was more severely worsened and weakened by multiple excavation disturbances.

In conclusion, due to the complex and intense multiple interactions of triple tunnels, the roof
conditions of triple tunnels were considerably deteriorated; the rock mass, especially the rock pillar,
might be already weakened and damaged ahead of the tunnel face. The original supports could not
meet the safety requirement of triple tunnel constructions. Due to the extreme importance of the station
and the high durability of station operation, proper pre-construction techniques must be used.

3. Parameter Optimization of PRPS and CGPR

In this project, pipe roof pre-support (PRPS) and curtain grouting pre-reinforcement (CGPR)
were jointly used as the pre-construction techniques. A series of numerical analyses were performed
by Flac 3D to optimize the design of PRPS and CGPR. The parameter design of the PRPS and the
CGPR were carried out separately. The mean crown settlement (MCS, the average crown settlement
Appl. Sci. 2020, 10, 3186 9 of 18

of the triple tunnels) and the pillar plastic ratio (PPR, the ratio of the plastic zone area to the whole
pillar area) were selected as the evaluation parameters. Noted that PRPS and CGPR were used
mainly for minimizing the mechanical interactions among triple tunnels and preventing potential
mechanical failures. Therefore, we primarily focused on the mechanical responses of triple tunnels.
According to the previous cases [8,9,14,15,24], CGPR showed significant performance in improving
the impermeability of surrounding rock. During tunneling, although we did not observe sudden
water inrush and the seepage effect was considered in rock mass classification through decreasing
the mechanical properties of rock, it was better to take hydraulic responses into account in the
numerical simulation. We did not quantitatively investigate the hydraulic responses in the simulation,
which might be the shortage of this research.

3.1. Numerical Model

The longitudinal length, the width, and the height of this model were 100, 250, and 200 m,
respectively (Figure 9). The tunnel lining was simulated as shell elements. The pipe roof and the
surrounding rock were modeled using solid elements. The very fine mesh was used to accurately
simulate the pipe roof and the neighboring grouting rock mass [25]. This model was initially built in the
MIDAS (a finite element analysis software) program, then the mesh information data were input into
the FLAC 3D program. The bottom of the model was constrained vertically, whereas four sides were
fixed horizontally. The middle section of the model along the longitudinal direction (i.e., y = 50 m)
was Sci.
Appl. selected for
2020, 10, analysis.
x FOR PEER REVIEW 9 of 18

Surrounding rock

CGPR
PRPS Pipe roof
Tunnel

200
Surrounding rock PRPS

Left Middle Right

tunnel tunnel tunnel

CGPR Pillar

250
10
0

Figure 9. Three-dimensional finite element model of triple tunnels.

Figure 9. Three-dimensional finite element model of triple tunnels.
The ratio of horizontal to vertical in situ stress was set to 1.8 by referring to the geotechnical
The ratio of horizontal
report. According to vertical inthe
to real circumstance, situexcavation
stress wassequence
set to 1.8inbythe
referring to the
numerical geotechnical
simulation was
report.
Left–Right–Middle, each tunnel was divided into three parts, each excavation circle was 2 m, andwas
According to real circumstance, the excavation sequence in the numerical simulation the
Left–Right–Middle, each tunnel
tunnel lining was installed was divided
immediately into
after three parts,
excavation. each
The topexcavation circle
heading face was
was 102 m
m,ahead
and theof
tunnel lining was installed immediately after excavation. The top heading face was
the bench face, and the invert face was 10 m behind the bench face. The lagging distance between 10 m ahead of the
bench
the leftface, andand
tunnel the right
inverttunnel
face was
was10set
mtobehind
30 m, the bench
which wasface.
the The
samelagging
as thatdistance
betweenbetween
the rightthe left
tunnel
tunnel and right tunnel was set to 30 m, which was the same as that between
and the middle tunnel. The advanced lengths of the PRPS and the CGPR were set to 40 m and the right tunnel and the
middle tunnel. The advanced lengths of the PRPS and the CGPR were set to 40 m and 15 m,
15 m, respectively.
respectively.
The mechanical behavior of the surrounding rock was simulated with the Mohr–Coulomb
failure criterion. The reinforcement effect of CGPR was simulated by increasing the mechanical
parameters of the rock mass in the reinforced zone. The physical and mechanical properties of each
material are listed in Table 2.
Appl. Sci. 2020, 10, 3186 10 of 18

Appl. Sci. 2020, 10, x FOR PEER REVIEW 10 of 19

The mechanical behavior of the surrounding rock was simulated with the Mohr–Coulomb failure
parameters of the rock mass in the reinforced zone. The physical and mechanical properties of each
criterion. The reinforcement effect of CGPR was simulated by increasing the mechanical parameters
material are listed in Table 2.
of the rock mass in the reinforced zone. The physical and mechanical properties of each material are
listed in Table 2.
Table 2. Material properties in the numerical simulation.
Table 2. Material properties 𝝆
in the numerical
𝑬 simulation.
𝝂 𝒄 𝝋 𝑯
Material
Material ρ (kg/m3 ) E (GPa)
(kg/m3) ν (GPa)c (kPa) (kPa)
ϕ (◦ )
(°) H (m) (m)
Miscellaneous soil filling
Miscellaneous soil filling 1750 0.1
1750 0.42 0.1 20 0.42 20
20
20 5.0 5.0
Completely
Completely weathered monzonite
weathered
1860 1.0 0.40 1000.40 24 25.0
monzonite granite 1860 1.0 100 24 25.0
granite
Highly weathered
Highly weathered 2000 1.5 0.37 1600.37 25 170.0170.0
monzonite granite monzonite granite 2000 1.5 160 25
Tunnel lining
Tunnel lining 2400 26.4 2400 0.20 26.4 - 0.20 -- - 0.4 0.4

3.2.
3.2. Parameter
Parameter Optimization
Optimization of
of PRPS
PRPS

3.2.1.
3.2.1. Pipe
Pipe Diameter
Diameter Optimization
Optimization
The
The pipe
pipe roof
roof consisted
consisted ofof the
the outer
outersteel
steelpipe
pipeand
andinner
innergrouted
groutedconcrete,
concrete,the
the outer
outer diameter
diameter of
of the steel pipe was D, and the steel pipe thickness was T (Figure 10). The density
the steel pipe was D, and the steel pipe thickness was T (Figure 10). The density and the elastic and the elastic
modulus 3
modulusof ofthe
the steel
steel pipe
pipe were
were 7850
7850 kg/m
kg/m3 and
and 210
210 GPa,
GPa, respectively,
respectively,with
with those
those of
of the
the inner
inner grouting
grouting
material 3
material being 2300 kg/m3 and 25 GPa, respectively. Note that the surrounding rockreinforced
being 2300 kg/m and 25 GPa, respectively. Note that the surrounding rock was also was also
by grouting,byand
reinforced the elastic
grouting, andmodulus of modulus
the elastic the reinforced
of thezone was 12zone
reinforced GPa was
[26].12
The
GPadiameter of diameter
[26]. The the steel
pipes to be selected ranged from 89 to 186 mm.
of the steel pipes to be selected ranged from 89 to 186 mm.

Steel pipe

Reinforced rock D T
around pipe roof

Arrangement angle
Grouting
concrete

Figure 10. Arrangement

Figure 10. Arrangement sketch
sketch of
of pipe
pipe roof
roof pre-support
pre-support(PRPS).
(PRPS).

The mechanical parameters of the pipe roof could be calculated according to the equivalence
principle of bending stiffness and mass [27]:

E0 I0 = E1 I1 + kw E2 I2 , (1)

ρ0 S0 = ρ1 S1 + ρ2 S2 , (2)

where E0 , I0 , ρ0 , S0 are the equivalent elastic modulus, the moment of inertia, the equivalent density,
and the area of the pipe roof; E1 , I1 , ρ1 , S1 are the elastic modulus, the moment of inertia, the density,
and the area of the grouting concrete; E2 , I2 , ρ2 , S2 are the elastic modulus, the moment of inertia,
the density, and the area of the steel pipe. kw is the reduction factor considering the cracking of the
grouting concrete, kw = 0.6.
The pipe roof parameters of six schemes are shown in Table 3. During the pipe diameter
optimization, the arrangement angle of the pipe roof was preset to 180◦ (“180◦ ” was not the final
optimized arrangement angle). The MCS and the PPR with different pipe diameters are shown
in Figure 11.
The mechanical parameters of the pipe roof could be calculated according to the equivalence
principle of bending stiffness and mass [27]:
 optimization, the arrangement angle of the pipe roof was preset to 180° (“180°” was not the final
optimized arrangement angle). The MCS and the PPR with different pipe diameters are shown in
Figure 11.

Table 3. Cross-section parameters of the pipe roof.

Appl. Sci. 2020, 10, 3186 11 of 18
Scheme 𝑫 𝑻 𝑬𝟎 𝝆𝟎
𝑺𝟎 (m ) 2 𝑰𝟎 (m ) 4 Sketch
ID (mm) (mm) (GPa) (kg/m3)
6.2 parameters
× 3.1 × of the pipe roof.
1 89 3. Cross-section
Table 4 56.7 3253
−3 −610 10
Scheme ID D (mm) T (mm) S0 (m 2) ×
9.2 (m4×)
I0 6.7 E0 (GPa) ρ0 (kg/m3 ) Sketch
2 108 6 62.9 3465
1 89 4 6.2 ×10−3
10
−3
3.1 ×1010
−6−6
56.7 3253
2 108 6 10−3×
9.2 × 1.3 1.310×−6
6.7 × 62.9 3465 T
3 127 8 67.1 3610
3 127 8 10−2
1.3 × 10 −2 1.3 ×1010−5−5 67.1 3610 D
4 159 10 10−2×
2.0 × 2.0 3.1 ×
3.110×
−5 67.0 3608
5 4 178 159 10 10 2.5 × 10−2 4.9 × 10−5−5 67.0
63.3 3608
3477
Appl. Sci. 2020, 10, x FOR PEER REVIEW 10−2 10 12 of 19
6 186 10 2.7 × 10−2 5.9 × 10−5 61.9 3429
2.5 × 4.9 ×
5 178 10 63.3 3477
10−2 10−5
Mean crown2.7
settlement
× 5.9 ×
6 186 10 61.9 3429
55 10−2 10−5plastic ratio 100%
Pillar
Mean crown settlement (mm)

PPR 98%
50

Pillar plastic ratio (%)

MCS
96%
45

94%
40

Steel pipe
92%
35
90%
D
30
88%
Grouting concrete
25
0 1 2 3 4 5 6
(0) (89) (108) (127) (159) (178) (186)
Scheme with different pipe diameter (mm)

Figure 11. Effect of pipe diameter on mean crown settlement (MCS) and pillar plastic ratio (PPR,
Figure 11. Effect of pipe diameter on mean crown settlement (MCS) and pillar plastic ratio (PPR, red
red elements represent plastic zone).
elements represent plastic zone).
Based on numerical results, we observed that the pipe roof played a more critical role in reducing
MCS Based
when oncompared
numericalwith results,
thatweinobserved
minimizingthat the
PPR. pipe
Whenroofthe
playedpipea diameter
more critical role in reducing
increased from 0 to
MCS
186 mm, the MCS decreased from the 56.2 to 26.5 mm, while the PPR slightly decreasedfrom
when compared with that in minimizing PPR. When the pipe diameter increased from0100%to 186to
mm, theItMCS
89.5%. could decreased fromthat
be expressed the the
56.2role
to 26.5 mm,
of the pipewhile the PPR
roof was mainly slightly
to formdecreased
a strongfrom
arch 100% to
structure
89.5%. It could be expressed that the role of the pipe roof was mainly to form a
above the tunnel and to bear loose rock pressure. In comparison, it had less ability in reinforcing thestrong arch structure
above the tunnel
rock pillar due to and
thetolimited
bear loose rock pressure.
arrangement area. In comparison,
Therefore, it hadshould
the MCS less ability
be theinmajor
reinforcing the
evaluation
rock pillar due
parameter to the
in the limited arrangement
optimization of PRPS. area. Therefore, the MCS should be the major evaluation
parameter in the optimization of
With the increase of pipe diameter,PRPS. the pipe roof possessed larger stiffness and showed better
With the increase
performance of pipecrown
in controlling diameter, the pipeWhen
settlement. roof possessed larger stiffness
the pipe diameter and showed
was larger than 108 better
mm,
performance in controlling crown settlement. When the pipe diameter
the downward trend of the MCS became no longer prominent. Therefore, 108 mm could be selectedwas larger than 108 mm, theas
downward trend of the
the optimized pipe diameter. MCS became no longer prominent. Therefore, 108 mm could be selected as
the optimized pipe diameter.
3.2.2. Arrangement Angle Optimization
3.2.2. Arrangement Angle Optimization
The arrangement area of PRPS is also a critical parameter in the design. Eight sets of numerical
The arrangement
simulations area ofarrangement
with different PRPS is also angle
a critical
(i.e.,parameter in◦ ,the
0◦ , 60◦ , 90 120design.
◦ , 150◦Eight
, 180◦ ,sets
210of
◦ , numerical
240◦ ) were
simulations with different arrangement angle (i.e., 0°, 60°,
performed to obtain optimal scheme, and the results are shown in Figure 12. 90°, 120°, 150°, 180°, 210°, 240°) were
performed to obtain optimal scheme, and the results are shown in Figure 12.

Mean crown settlement

55 100%
Pillar plastic ratio
Mean crown settlement (mm)

50 PPR 96%
Pillar plastic ratio

MCS
45
92%

40
88%
35 Angle

84%
30
the optimized pipe diameter.

3.2.2. Arrangement Angle Optimization

The arrangement area of PRPS is also a critical parameter in the design. Eight sets of numerical
simulations
Appl. Sci. 2020,with different arrangement angle (i.e., 0°, 60°, 90°, 120°, 150°, 180°, 210°, 240°) 12
10, 3186 were
of 18
performed to obtain optimal scheme, and the results are shown in Figure 12.

Mean crown settlement

55 100%
Pillar plastic ratio

Mean crown settlement (mm)

50 PPR 96%

Pillar plastic ratio

MCS
45
92%

40
88%
35 Angle

84%
30
0 1 2 3 4 5 6 7
(0) (60) (90) (120) (150) (180) (210) (240)
Scheme with different pipe arrangement angle (°)

Figure12.
Figure Effectofofpipe
12.Effect pipearrangement
arrangementangle
angleon
onMCS
MCSand
andPPR.
PPR.

It could be observed that the arrangement angle had a great effect on controlling crown settlement.
The MCS decreased by 46.1% when the arrangement angle increased from 0◦ to 180◦ ; in comparison,
the PPR merely reduced by 16.5% during the same period. This phenomenon showed good agreements
with pipe diameter optimization. The MCS reduced quickly until the arrangement angle reached 150◦ .
Thus, the arrangement angle could be optimized as 150◦ .
According to the above parameter optimization of PRPS, the MCS was well controlled with the
implementation of PRPS. However, the rock pillar was still broken, and the plastic zone penetrated
the rock pillar. Therefore, CGPR should also be applied to improve the mechanical properties of the
rock pillar.

3.3. Parameter Optimization of CGPR

The grouting thickness is the major parameter of CGPR design because it is directly related to
the development and the distribution of the plastic zone [28]. To obtain optimal grouting thickness,
in the design stage, there were ten grouting schemes with different grouting thickness ranging from 0
to 5 m. Due to the importance of the rock pillar, increasing the grouting volume and multiple grouting
were particularly used to improve the mechanical properties of the rock pillar [14,15]. The material
properties of the grouted rock mass and rock pillar are shown in Table 4. The grouting area was the
surrounding rock around the excavation contour. Because of the disturbance of the complex grouting
process, the primary lining was delayed with a lag distance of 4 m.

Table 4. Material properties of grouted rock mass and rock pillar.

Material ρ (kg/m3 ) E (GPa) ν c (kPa) ϕ (◦ )

Grouted rock mass 2200 2.5 0.33 300 30
Grouted rock pillar 2200 8.0 0.28 900 42

According to the numerical results (Figure 13), we found that the PPR was more sensitive to
grouting thickness than the MCS. It could be attributed to the larger reinforced zone and smaller
stiffness of CGPR. The primary role of CGPR was improving mechanical properties and self-capacity
of surrounding rock and protecting the rock pillar from excavation disturbance. Relatively, the stiffness
of the grouted rock was much smaller than the pipe roof, which led to less effect of the CGPR in
controlling crown settlement. Therefore, the PPR should be the major evaluation parameter in grouting
thickness optimization.
stiffness of CGPR. The primary role of CGPR was improving mechanical properties and self-capacity
of surrounding rock and protecting the rock pillar from excavation disturbance. Relatively, the
stiffness of the grouted rock was much smaller than the pipe roof, which led to less effect of the CGPR
in controlling crown settlement. Therefore, the PPR should be the major evaluation parameter in
grouting thickness
Appl. Sci. 2020, 10, 3186optimization. 13 of 18

Mean crown settlement

100%
55 Pillar plastic ratio
90%

Mean crown settlement (mm)

Grouting 80%
50

Pillar plastic ratio

thickness
70%
60%
45
50%
40%
40
30%
PPR
MCS 20%
35
10%
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

Grouting thickness (m)

Figure 13. Effect of grouting thickness of curtain grouting pre-reinforcement (CGPR) on MCS and PPR.
Figure 13. Effect of grouting thickness of curtain grouting pre-reinforcement (CGPR) on MCS and
The PPR reduced by 78.1% when the grouting PPR.
thickness increased from 0 to 5 m. When the grouting
thickness was larger than 3 m, this decreasing trend became less noticeable. Moreover, the high cost
and long construction period of the CGPR should also be taken into consideration. Thus, the reasonable
grouting thickness was determined to be 3 m.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 14 of 19

4. Numerical Evaluation
The and Engineering
PPR reduced by 78.1% whenApplication
the grouting thickness increased from 0 to 5 m. When the
grouting thickness was larger than 3 m, this decreasing trend became less noticeable. Moreover, the
high cost and long construction period of the CGPR should also be taken into consideration. Thus,
4.1. Numerical Evaluation of the Combined Use of PRPS and CGPR
the reasonable grouting thickness was determined to be 3 m.

To better investigate theEvaluation

4. Numerical combined effect ofApplication
and Engineering PRPS and CGPR, two cases, namely, without (Case 1)
and with (Case 2) the combined
4.1. Numerical implementation
Evaluation of the Combined Use of PRPS ofandPRPS
CGPR and CGPR, were numerically compared.

The ground settlementsToofbetter

theinvestigate
two cases are shown
the combined in and
effect of PRPS Figure 14.cases,
CGPR, two Thenamely,
plastic zone
without (Case contours before and
1) and with (Case 2) the combined implementation of PRPS and CGPR, were numerically compared.
after the uses of PRPS
Theand
groundCGPR
settlementsare provided
of the in Figure
two cases are shown in Figure 15.
14. The plastic zone contours before and
after the uses of PRPS and CGPR are provided in Figure 15.

(a)

(b)
Figure 14. Ground settlement of triple tunnels: (a) without PRPS or CGPR (Case 1); (b) with PRPS and
Figure 14. Ground settlement of triple tunnels: (a) without PRPS or CGPR (Case 1); (b) with PRPS and
CGPR (Case 2).

CGPR (Case 2).

Appl. Sci. 2020, 10, 3186 14 of 18
Appl. Sci. 2020, 10, x FOR PEER REVIEW 15 of 19

(a)

(b)
Figure 15. Development
Development ofof plastic
plastic zone (red elements) around triple tunnels: (a)
(a) without
without PRPS or
CGPR (Case 1); (b) with PRPS and CGPR (Case 2).

We noted the
the following
following mechanical
mechanical characteristics:
characteristics:

1. In
In Case 1, three
threesmall
smallrockrockarches
archeswere were formed
formed above
above each each tunnel.
tunnel. Besides,
Besides, a large
a large combined
combined rock
rock arch was also formed across the triple tunnels (Figure 14a).
arch was also formed across the triple tunnels (Figure 14a). This large combined rock arch was This large combined rock arch
was generally
generally entered entered
plasticplastic state (Figure
state (Figure 15a). trends
15a). Similar Similarhave trends
alsohave
been also been analytically
analytically studied by
studied by Lifinding
Li [29]. This [29]. This finding
implied thatimplied
the loosenthat themass
rock loosen rockeach
above mass above
tunnel eachbe
might tunnel mightdue
connected be
connected due to the multiple
to the multiple excavation interactions. excavation interactions.
2.
2. During
During the the left
left and
and right
right tunnel
tunnel excavations
excavations of of Case
Case 1, 1, the
the rock
rock pillar
pillar was
was severely
severely disturbed
disturbed and and
lost the ability to bear loosen rock. The stress concentration phenomenon
lost the ability to bear loosen rock. The stress concentration phenomenon of the rock pillar was of the rock pillar was
prominent,
prominent, and and the the plastic
plastic zone
zone completely
completely penetrated
penetrated the the rock
rock pillar (Figure 15a).
pillar (Figure 15a). ItIt inevitably
inevitably
worsened
worsened the the middle
middle tunnel
tunnel excavation
excavation and and caused
caused aa larger
larger plastic
plastic zone
zone around
around the the middle
middle
tunnel. The range of the final plastic zone was considerably large,
tunnel. The range of the final plastic zone was considerably large, with more than 15 m extendedwith more than 15 m extended
to
to the
the deep
deep surrounding
surrounding rock. rock.
3.
3. With
With the combined use ofPRPS
the combined use of PRPS andandCGPR
CGPR (Case
(Case2), the development
2), the development of the
of three separated
the three rock
separated
arches to the large combined rock arch was effectively restricted (Figure
rock arches to the large combined rock arch was effectively restricted (Figure 14 b), the stress 14 b), the stress state of
the
staterock pillar
of the rockwas greatly
pillar wasimproved (Figure 15b).
greatly improved (FigureThe15b).
safety Theof safety
the tunnel
of the roof was roof
tunnel ensured,
was
with a 56.3% reduction of the MCS. The plastic zone development
ensured, with a 56.3% reduction of the MCS. The plastic zone development was also significantly was also significantly
constrained.
constrained. The Theplastic
plasticzonezonewas was primarily
primarily confined
confined inside
inside the the grouting
grouting ring,
ring, andand the PPR
the PPR
sharply
sharply decreased
decreased to to 30.4%.
30.4%.
4.
4. Moreover, the combined
Moreover, the combined application applicationofof PRPS
PRPS and and
CGPRCGPR (Case(Case 2) showed
2) showed betterbetter performance
performance than
than that of the independent use of them. The MCS decreased
that of the independent use of them. The MCS decreased to 24.6 mm, considerably smaller thanto 24.6 mm, considerably smaller
than that
that of theofindependent
the independent application
application of PRPS
of PRPS (34.1(34.1
mm,mm, FigureFigure
12) or12)CGPR
or CGPR(42.0(42.0
mm,mm, Figure
Figure 13).
13). The PPR reduced to 30.4%, more effective than that of the
The PPR reduced to 30.4%, more effective than that of the independent application of PRPS independent application of PRPS
(96.8%,
(96.8%, Figure
Figure 12) 12) or
or CGPR
CGPR (35.4%,(35.4%, Figure
Figure 13).
13). ItIt could
could be be attributed
attributed to to the
the synergistic
synergistic effect
effect
between
between PRPS and CGPR. PRPS had longer advanced length, larger rigidity, faster construction
PRPS and CGPR. PRPS had longer advanced length, larger rigidity, faster construction
speed,
speed, andand smaller
smallerreinforcement
reinforcement zone,
zone, whilewhile
CGPR CGPR had shorter
had shorter advancedadvanced
length, length,
smaller smaller
rigidity,
rigidity, slower construction
slower construction speed, and speed,
larger and larger reinforcement
reinforcement zone. Owing zone.toOwing
the longer to the longer
advanced
advanced length, larger rigidity, and shorter construction period of PRPS, a strong arch structure
could be timely formed far ahead of the tunnel face. Therefore, PRPS could effectively bear
Appl. Sci. 2020, 10, 3186 15 of 18

Appl. Sci. 2020, 10, x FOR PEER REVIEW 16 of 19

length, larger rigidity, and shorter construction period of PRPS, a strong arch structure could be
vertical rock pressure
timely formed andofprotect
far ahead the face.
the tunnel tunnel crown. Owing
Therefore, to theeffectively
PRPS could broader grouting range
bear vertical of
rock
CGPR, the CGPR could greatly improve mechanical properties and the self-stability
pressure and protect the tunnel crown. Owing to the broader grouting range of CGPR, the CGPR of the
surrounding
could greatlyrock. Thus,mechanical
improve the development of plastic
properties and thezone could be of
self-stability effectively restrictedrock.
the surrounding by
CGPR.
Thus, theOverall, PRPS ofand
development CGPR
plastic zonehave
could complementary advantages
be effectively restricted by CGPR.in time,
Overall,space,
PRPS andand
mechanical
CGPR haveproperties,
complementarythey advantages
mutually and actively
in time, formed
space, a strong permanent
and mechanical properties,ring
theyinmutually
front of
the
andtunnel face.
actively formed a strong permanent ring in front of the tunnel face.

4.2.
4.2. Engineering
Engineering Application
Application
The optimized scheme
The optimized schemeofofpre-construction
pre-construction techniques
techniques waswastimelytimely applied
applied in theinlater thetriple
later tunnel
triple
tunnel excavation.
excavation. In practical
In practical use, theuse,
pipethe pipe diameter,
diameter, the arrangement
the arrangement angle, and angle, and length
the pipe the pipe of length
the PRPS of
the
werePRPS
108 were
mm, 140 ◦
108 mm,
, and140°,
39 m,and 39 m, respectively
respectively . The pipe . The pipe amount
amount of the left,of middle,
the left, middle,
and right and right
tunnels
tunnels areand
are 45, 43, 45, 45,
43, respectively.
and 45, respectively.
Due to the Due to theof
urgency urgency of the construction
the construction period, theperiod,
designthe design
scheme of
scheme
CGPR was of notCGPR wasadopted.
entirely not entirely adopted.the
Nevertheless, Nevertheless,
grouting design thestill
grouting
provided design still provided
significant reference.
significant
The rock pillarreference.
and the Themiddle
rock pillar
tunneland the finally
were middleselected
tunnel were as thefinally selected as
key grouting zonesthe key
becausegrouting
they
zones because
were the most they were the
vulnerable areamost vulnerable
according to thearea accordingdesign.
preliminary to the preliminary design.
For
For the installation
installation ofofPRPS,
PRPS,thetheconcrete
concretearcharchwaswascastcast first
first to provide
to provide strong
strong supportsupport
for theforpipe
the
pipe inserting.
inserting. The steel
The steel pipe pipe was installed
was installed by thebycased-drilling
the cased-drillingmethod method [3].drilling
[3]. The The drillingjumbo jumbo was
was used
used to boreholes
to drill drill boreholes and insert
and insert steel steel
pipe pipe simultaneously.
simultaneously. AfterAfter the drilling
the drilling process,
process, the front
the front casing casing
pipe
pipe remained
remained in place
in place as theaspermanent
the permanent support.
support. Thepipe
The steel steelwaspipegrouted
was grouted with cement-based
with cement-based slurry,
slurry,
and theand the diffusion
slurry slurry diffusion radius
radius was notwas
lessnot
than less
0.6than
times0.6the
times
pipethe pipe spacing.
spacing.
For
For the
the execution
execution of of CGPR,
CGPR, the
the 1.0
1.0 mm thick
thick concrete
concrete wallwall was
was built
built first,
first, the
the combination
combination of of the
the
forward
forward grouting
grouting and the backward grouting was adopted. adopted. Sulfoaluminate
Sulfoaluminate cement cement singlesingle slurry
slurry
(water:cement ==0.8:1)
(water:cement 0.8:1) was
was used
used as
as the
the main
main cement-based
cement-based slurry,slurry, and the cement-sodium
cement-sodium silicate silicate
binary
binary slurry
slurry was was used
used asas the auxiliary
auxiliary slurry. The The grouting
grouting process
process should
should be be stopped
stopped when when the the
grouting
grouting pressure
pressure achieved final grouting pressure (3–5 MPa) and the grouting grouting speed was was lower
lower than
than
55 L/min.
L/min.
To practicallyevaluate
To practically evaluatethethe engineering
engineering application,
application, DK 67DK + 95067was + 950 was asselected
selected another as another
monitoring
monitoring
section. The section. The distribution
distribution of the surrounding
of the surrounding rock pressurerock pressure
is shownisin shown
Figure in 16.
Figure The16.stable
The
stable
crowncrown settlements
settlements of triple
of triple tunnels
tunnels rangerange
from from
DK67DK67 + 940+ to940DK67
to DK67+ 980 + 980
are are
shown shown in Figure
in Figure 17.
17.
TheThe comparisons
comparisons of representative
of representative measured
measured datadata
of twoof two
casescases
are are listed
listed in Table
in Table 5. 5.

403
178 Damaged 251
210 235 278 362 134

① ⑦ ④
19 ②
42 66 ⑧
83 52 40
⑤
50 ③ 106 -21 ⑨ 92 69 ⑥ 46
Left tunnel Middle tunnel Right tunnel
Figure
Figure 16.
16. Distribution
Distribution of
of surrounding
surrounding rock
rock pressure at section
pressure at section DK67
DK67 +
+ 950
950 (Unit:
(Unit: kPa).
kPa).
Appl. Sci. 2020, 10, 3186 16 of 18
Appl. Sci. 2020, 10, x FOR PEER REVIEW 17 of 19

DK67+940
30 DK67+950
DK67+960
DK67+970
25 DK67+980

Crown settlement (mm)

20

15

10

5
Left tunnel Middle tunnel Right tunnel
Triple tunnels

Figure
Figure 17.
17. Crown
Crownsettlements
settlementsof
oftriple
tripletunnels
tunnels with
with PRPS
PRPS and
and CGPR.
CGPR.

Table 5. Monitoring data comparison of two cases.

Table 5. Monitoring data comparison of two cases.
Max Surrounding Rock Mean Surrounding Rock Max Crown Mean Crown
Cases Max Surrounding Mean Pressure
Surrounding Max Crown Mean (mm)Crown
Pressure (kPa) (kPa) Settlement (mm) Settlement
CasesCase 1 Rock Pressure
547 Rock Pressure
202.5 Settlement
76.6 Settlement
44.5
Case 2
(kPa)403 134.8
(kPa) 28.7
(mm) 18.4
(mm)
Reduction rate 26.3% 33.4% 62.5% 58.7%
Case 1 547 202.5 76.6 44.5
Case 2 403 134.8 28.7 18.4
It can be observed that the surrounding rock pressure at the crown and shoulders of each tunnel at
Reduction rate 26.3% 33.4% 62.5% 58.7%
DK67 + 950 were mainly concentrated between 200–400 kPa. The mean surrounding rock pressure of
triple tunnels at DK67 + 950 was 134.8 kPa, 33.4% lower than that (202.5 kPa) at DK68 + 220. The MCS
It can be observed that the surrounding rock pressure at the crown and shoulders of each tunnel
reduced by 58.7% after pre-support and pre-reinforcement (from 44.5 to 18.4 mm). The tunnel supports
at DK67 + 950 were mainly concentrated between 200–400 kPa. The mean surrounding rock pressure
were in the good working state, no more steel frame distortion or concrete cracking was observed,
of triple tunnels at DK67 + 950 was 134.8 kPa, 33.4% lower than that (202.5 kPa) at DK68 + 220. The
and the primary lining leakage was also well controlled. It implied that the combined implementation
MCS reduced by 58.7% after pre-support and pre-reinforcement (from 44.5 to 18.4 mm). The tunnel
of PRPS and CGPR greatly improved the mechanical state of the rock–support system.
supports were in the good working state, no more steel frame distortion or concrete cracking was
observed, and the primary lining leakage was also well controlled. It implied that the combined
5. Conclusions
implementation of PRPS and CGPR greatly improved the mechanical state of the rock–support
system.Pipe roof pre-support (PRPS) and curtain grouting pre-reinforcement (CGPR) were successfully
used in the closely spaced large span triple tunnels of the Badaling Great Wall Underground Station.
The following conclusions were obtained:
5. Conclusions
1. Based
Pipe onpre-support
roof preliminary(PRPS)
field monitoring,
and curtainitgrouting
indicated that the rock mass
pre-reinforcement aheadwere
(CGPR) of thesuccessfully
tunnel face
was severely broken due to the severe multiple interactions among triple
used in the closely spaced large span triple tunnels of the Badaling Great Wall Underground tunnels, and theStation.
current
supports could not meet the
The following conclusions were obtained:safety requirements. Therefore, PRPS and CGPR should be applied.
2. A series of numerical simulations were conducted to optimize design parameters. The optimal
1. Based on preliminary
pipe diameter, field monitoring,
arrangement angle of PRPSit indicated that theto
were determined rock mass
be 108 mm ahead of the
and 150 tunnel
◦ . The face
grouting
was severely
thickness broken
of CGPR wasdue to the severe
optimized as 3 m.multiple interactions among triple tunnels, and the
3. current
According to numerical results, thesafety
supports could not meet the PRPS requirements.
showed a better Therefore,
effect in PRPS and CGPR
controlling MCS,should be
while the
applied.
CGPR showed better performance on reducing PPR. It could be attributed to their different
2. Aworking
series ofmechanism,
numerical simulations were conducted
the PRPS predominantly boretoexternal
optimize design
load parameters.
and prevented The collapse,
tunnel optimal
pipe diameter, arrangement angle of PRPS were determined to be 108 mm
while the CGPR primarily improved the mechanical properties and enhanced the self-stability and 150°. The
of
grouting thickness
the rock mass. of CGPR was optimized as 3 m.
3. According to numerical results, the PRPS showed a better effect in controlling MCS, while the
4. PRPS and CGPR had complementary advantages in time, space, and mechanical properties.
CGPR showed better performance on reducing PPR. It could be attributed to their different
PRPS could timely form a strong arch structure far ahead of the tunnel face, CGPR could
working mechanism, the PRPS predominantly bore external load and prevented tunnel collapse,
Appl. Sci. 2020, 10, 3186 17 of 18

effectively restrict the development of the plastic zone. They mutually and actively formed
a strong permanent ring in front of the tunnel face.
5. The combined implementation of PRPS and CGPR were adopted in subsequent tunnel excavation,
the mean surrounding rock pressure decreased by 33.4%, the MCS decreased by 58.7%,
respectively, and no significant tunnel damages were observed. The triple tunnel excavation was
effectively safeguarded.

Author Contributions: Project administration: D.Z.; Conceptualization, R.L. and P.W.; Data curation, R.L. and
Q.F.; Formal analysis, R.L. and D.Z.; Investigation, R.L. and A.L.; Methodology, R.L. and D.Z.; Validation, L.C.
and P.W.; Writing—Original draft preparation, R.L.; Writing—Review & editing, D.Z. All authors have read and
agreed to the published version of the manuscript.
Funding: The authors gratefully acknowledge the financial support by the Fundamental Research Funds for
the Central Universities (Grant No. 2020YJS103), the National Natural Science Foundation of China (Grant No.
51738002) and the National Key Research and Development Program of China (Grant No. 2017YFC0805401).
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Zhang, Z.; Li, H.; Liu, H.; Li, G.; Shi, X. Load transferring mechanism of pipe umbrella support in
shallow-buried tunnels. Tunn. Undergr. Space Technol. 2014, 43, 213–221. [CrossRef]
2. Ocak, I. Control of surface settlements with umbrella arch method in second stage excavations of Istanbul
Metro. Tunn. Undergr. Space Technol. 2008, 23, 674–681. [CrossRef]
3. Volkmann, G.M.; Schubert, W. Advantages and specifications for pipe umbrella support systems.
In Proceedings of the 14th Australasian Tunnelling Conference, Auckland, New Zealand, 8–10 March 2011.
4. Rostami, A.; Dehkordi, P.K.; Ziarati, M.A.; Jahani, S.; Lotfi, K. The types of tunnels maintenance in umbrella
arch method. Open J. Civ. Eng. 2016, 06, 156–162. [CrossRef]
5. Schumacher, F.P.; Kim, E. Modeling the pipe umbrella roof support system in a Western US underground
coal mine. Int. J. Rock Mech. Min. Sci. 2013, 60, 114–124. [CrossRef]
6. Niedbalski, Z.; Małkowski, P.; Majcherczyk, T. Application of the NATM method in the road tunneling works
in difficult geological conditions—The Carpathian flysch. Tunn. Undergr. Space Technol. 2018, 74, 41–59.
[CrossRef]
7. Sadeghiyeh, S.M.; Hashemi, M.; Ajalloeian, R. Comparison of permeability and groutability of Ostur Dam
site rock mass for grout curtain design. Rock Mech. Rock Eng. 2013, 46, 341–357. [CrossRef]
8. Rostami Barani, H.R.; Khatib, M.M. Back analysis of grout treatment at Sumbar Dam using the joint hydraulic
factor. Rock Mech. Rock Eng. 2015, 48, 2485–2488. [CrossRef]
9. Davis, G.M.; Horswill, P. Groundwater control and stability in an excavation in Magnesian Limestone near
Sunderland, NE England. Eng. Geol. 2002, 66, 1–18. [CrossRef]
10. Tiwari, G.; Latha, G.M. Stability analysis and design of stabilization measures for Chenab railway bridge
rock slopes. Bull. Eng. Geol. Environ. 2020, 79, 603–627. [CrossRef]
11. Fang, Q.; Zhang, D.; Wong, L.N.Y. Shallow tunnelling method (STM) for subway station construction in soft
ground. Tunn. Undergr. Space Technol. 2012, 29, 10–30. [CrossRef]
12. Cao, L.; Zhang, D.; Fang, Q.; Yu, L. Movements of ground and existing structures induced by slurry pressure-balance
tunnel boring machine (SPB TBM) tunnelling in clay. Tunn. Undergr. Space Technol. 2020, 97, 103278. [CrossRef]
13. Ganerød, G.V.; Braathen, A.; Willemoes-Wissing, B. Predictive permeability model of extensional faults in
crystalline and metamorphic rocks; verification by pre-grouting in two sub-sea tunnels, Norway. J. Struct. Geol.
2008, 30, 993–1004. [CrossRef]
14. Stille, B.; Gustafson, G. A review of the Namntall Tunnel project with regard to grouting performance.
Tunn. Undergr. Space Technol. 2010, 25, 346–356. [CrossRef]
15. Zhang, D.; Fang, Q.; Lou, H. Grouting techniques for the unfavorable geological conditions of Xiang’an
subsea tunnel in China. J. Rock Mech. Geotech. Eng. 2014, 6, 438–446. [CrossRef]
16. Fang, Q.; Zhang, D.; Zhou, P.; Wong, L.N.Y. Ground reaction curves for deep circular tunnels considering the
effect of ground reinforcement. Int. J. Rock Mech. Min. 2013, 60, 401–412. [CrossRef]
Appl. Sci. 2020, 10, 3186 18 of 18

17. Peila, D.; Oreste, P. Axisymmetric analysis of ground reinforcing in tunnelling design. Comput. Geotech.
1995, 17, 253–274. [CrossRef]
18. Fang, Q.; Zhang, D.; Li, Q.; Wong, L.N.Y. Effects of twin tunnels construction beneath existing shield-driven
twin tunnels. Tunn. Undergr. Space Technol. 2015, 45, 128–137. [CrossRef]
19. Li, A.; Fang, Q.; Zhang, D.; Luo, J.; Hong, X. Blast vibration of a large-span high-speed railway tunnel based
on microseismic monitoring. Smart Struct. Syst. 2018, 21, 561–569.
20. Li, A.; Zhang, D.; Fang, Q.; Luo, J.; Cao, L.; Sun, Z. Safety distance of shotcrete subjected to blasting vibration
in large-span high-speed railway tunnels. Shock Vib. 2019, 2019, 2429713. [CrossRef]
21. Li, R.; Zhang, D.; Fang, Q.; Li, A.; Hong, X.; Ma, X. Geotechnical monitoring and safety assessment of
large-span triple tunnels using drilling and blasting method. J. Vibroeng. 2019, 21, 1373–1387.
22. Liu, D.; Zhang, D.; Fang, Q.; Sun, Z.; Cao, L.; Li, A. Displacement characteristics of shallow-buried
large-section loess tunnel with different types of pre-supports: A case study of New Badaling Tunnel.
Appl. Sci. 2020, 10, 195. [CrossRef]
23. Zhang, D.; Fang, Q.; Hou, Y.; Li, P.; Yuen Wong, L.N. Protection of buildings against damages as a result of
adjacent large-span tunneling in shallowly buried soft ground. J. Geotech. Geoenvironment 2013, 139, 903–913.
[CrossRef]
24. Shi, P.; Zhang, D.; Pan, J.; Liu, W. Geological Investigation and Tunnel Excavation Aspects of the Weakness
Zones of Xiang’an Subsea Tunnels in China. Rock Mech. Rock Eng. 2016, 49, 4853–4867. [CrossRef]
25. Franzius, J.N.; Potts, D.M. Influence of mesh geometry on three-dimensional finite-element analysis of tunnel
excavation. Int J. Geomech. 2005, 5, 256–266. [CrossRef]
26. Shi, Y.; Fu, J.; Yang, J.; Xu, C.; Geng, A.D. Performance evaluation of long pipe roof for tunneling below
existing highway based on field tests and numerical analysis: Case study. Int J. Geomech. 2017, 17. [CrossRef]
27. Shi, Y. Study on Mechanical and Engineering Application of Advanced Support for Lower Overburden
Tunnel in Weak Stratum. Ph.D. Thesis, Central South University, Changsha, China, 2014.
28. Sun, Z.; Zhang, D.; Fang, Q. Determination method of reasonable reinforcement parameters for subsea
tunnels considering ground reinforcement and seepage effect. Appl. Sci. 2019, 9, 3607. [CrossRef]
29. Li, P.; Wang, F.; Fan, L.; Wang, H.; Ma, G. Analytical scrutiny of loosening pressure on deep twin-tunnels in
rock formations. Tunn. Undergr. Space Technol. 2019, 83, 373–380. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).