Environmental Earth Sciences (2020) 79:521

https://doi.org/10.1007/s12665-020-09266-y

ORIGINAL ARTICLE

Umbrella arch method performance, structural behavior and design

elements utilizing in collapsing zones
Parisa Bagherzadeh1   · Kamran Goshtasbi1 · Mojtaba Kashef2

Received: 10 May 2020 / Accepted: 30 October 2020 / Published online: 16 November 2020
© Springer-Verlag GmbH Germany, part of Springer Nature 2020

Abstract
The umbrella arch method is a pre-reinforcement technique that ensures tunnel stability in poor ground conditions or shallow
overburden. Quantitative assessment of this method is still challenging due to its complexity, there is no accurate method
to model the behavior of reinforced arch forming with steel pipes and grout. This paper uses finite difference numerical
simulation to model and analyze the effectiveness of the umbrella arch method in Qazvin–Rasht railway tunnel as well as
in-situ measurement. Umbrella arch has modeled considering pipe, grout and soilcrete as pipe umbrella components; each
pipe simulated individually. The structural behaviour of pipes in the excavation sequence is analyzed as well as the impor-
tance of geometrical and physical design elements, on tunnel stability using FLAC 3D software. Results indicate that the
umbrella arch method has reduced tunnel convergence by 49%. By continuing the excavation step, residual pipe length ahead
of the face and tunnel arch, the location of maximum pipe displacement, force, and moment are changed. Furthermore, by
considering executive limitation, an increase in pipe angle, diameter and thickness, length and overlapping length, as well
as pipes distance decrease, center to center grout penetration radius, and water-cement ratio 0.5 have had a positive effect in
providing more strength arch and tunnel stability.

Keywords  Umbrella arch · Numerical modeling · Instrumentation · Design element · Structural behavior

Introduction This method is well described in the literature by, Pelizza

and Peila (1993),Yoo and Shin (2003), Carrieri et al. (1991),
Tunnel construction in the weak ground would arise several Carrieri et al. (2002), Lunardi (2000), Kim et al. (2005),
problems including tunnel instability, damage to existing Aksoy (2010), Ocak (2008), Song et al. (2013), Bagherza-
structures, and ground settlement. To overcome these prob- deh et al. (2013), Bagherzadeh and Goshtasbi (2014),Schu-
lems, some auxiliary techniques have been adopted in tun- macher et al. (2016), Gao et al. (2015), Yahya (2016), Salmi
neling. Grouting with pipes is one of the most popular pre- et al. (2017), Taromi and Eftekhari (2018) and Klotoé and
reinforcement methods of tunnel heading that is called the Bourgeois (2019).
umbrella arch method or UAM. In this technique, by creating Volkmann (2006) has subdivided this method into five
a longitudinal arching parallel to the tunnel axis, safe exca- categories mainly according to the material used and ways of
vation would be allowed even in poor ground conditions. reinforcing the ground, including forepoling, pipe roof, sub-
horizontal jet-grouting, pipe jacking, and freezing. There
* Parisa Bagherzadeh are substantial differences among these methods in terms of
[email protected] execution procedure, equipment, and reinforcing material.
Kamran Goshtasbi However, the basic concept and design philosophy are the
[email protected] same. The pipe roof method is commonly used as Umbrella
Mojtaba Kashef arch support that consists of pipes, installed from the actual
[email protected] face to the front. The pipes are 12.0 m or 15.0 m long and
arranged at the outer shape of the tunnel. The outer diam-
1
Department of Mining Engineering, Faculty of Engineering, eter of the pipes ranges from 70 to 200 mm. The steel pipes
Tarbiat Modares University, Tehran, Iran
are usually installed with a dip of 5–10 degrees, to form an
2
Department of Mining Engineering, Faculty of Engineering,
University of Zanjan, Zanjan, Iran

13
Vol.:(0123456789)
521 
Page 2 of 15 Environmental Earth Sciences (2020) 79:521

umbrella with a truncated cone shape to make overlapping Zarei et al (2019) have studied UAM importance in weak
of two adjacent umbrellas possible (Peila 2005). ground conditions using a finite difference program. Earth
Various experimental, numerical and analytical studies pressure could effectively be transferred to the primary sup-
have been implemented to discover umbrella arch method port systems, improving tunnel stability. Results show that
performance in tunnel construction. Ocak (2008) had stud- the installation angle and pipe distance decrease and pipe
ied the UAM method in second excavation stage of Istan- diameter increase can reduce tunnel crown displacement.
bul Metro. The results indicated that surface deformation Wang (2012) accomplished an analytical approach analyzing
has been controlled effectively in clay-bearing formations. the effect of UAM design parameters such as pipe diameter,
Yoo and Shin (2003) conducted experimental and numeri- pipe length, and overlap length on the mechanical behavior
cal investigations on the deformation behavior of tunnel of pipe roof reinforcement. The study showed that each rein-
face reinforced by longitudinal pipes. They concluded that forcement parameter has a critical value to reach maximum
face reinforcement technique can reduce face deformation tunnel stability.
and surface settlements significantly. Klotoé and Bourgeois In this study, the influence of the umbrella arch applica-
(2019) conducted 3D finite element simulations to analyze tion on tunnel stability and decreasing displacement was
the influence of UAM on settlement induced by shallow examined at Qazvin–Rasht railway tunnel (km: 110 + 345
tunneling. to 111 + 08). In this section, the rock formation is classified
Hisatake and Ohno (2008) conducted centrifugal model as poor and very poor, and the water level makes tunneling
test to clarify the effect of pipe roof supports and excava- conditions harsh. In this condition, tunnel construction is
tion methods on tunnel stability. A computerized excava- not possible with conventional supports such as rock bolts,
tion robot, controlled in three-dimension, was designed to shotcrete, and steel sets; thus, utilizing other special pre-
model full face excavation and ring-cut excavation method. supports like the umbrella method is inevitable.
The results represent that employing UAM could decrease Although this method has been widely used in weak
ground displacement by 75%. Aksoy and Onargan (2010) ground and shallow depth excavation since the 1980s, there
have studied the effects of UAM and face bolts on ground is no simple and definite method to model the actual behav-
settlement using numerical assessment and in-situ measure- ior of UAM. Some studies used simplified assumptions such
ments in the Izmir metro construction. Results indicate that as primary support system elements. Where, Shotcrete, wire
UAM and bolt application has reduced settlement and con- mesh, and steel frame are modeled as equivalent sections. To
vergence by 69% and 57%, respectively. Heidari and Tonon develop a suitable model, the umbrella arch has simulated
(2015) have considered jet grouting progressive hardening using FLAC 3D finite difference method. The pipes have
in controlling tunnel convergence using the ground reac- modeled individually considering grout and soilcrete influ-
tion curve. Unsafe tunnel design is resulted by ignoring jet ence. Laboratory experiments on grout and the reinforcing
grout hardening. Qian et al. (2019) developed an analytical area around pipes have been performed to obtain accurate
approach using the Winkler-spring model to simulate UAM values of numeric code. Tunnel convergence, measured
pipes’ behavior. Oke et al. (2014) have studied influential through in situ instrumentation and monitoring during con-
design parameters (spacing and overlap) of UAM method struction, has confirmed numerical modeling results.
using 2D and 3D models. They later developed a semi-ana- The supporting mechanism of the umbrella method is to
lytical model for distributed load where beams lay on an stabilize the excavation face in both the transverse and lon-
elastic foundation in squeezing ground conditions (Oke et al. gitudinal direction by an arch-like reinforced zone. In the
2016). Song et al. (2013) developed a beam-spring structural transverse direction, an arch-like reinforced zone carries the
analysis model to optimize the design parameters of UAM ground load, and in the longitudinal direction, the ground
method using a finite element programming code. load of the freshly excavated portion is supported by a beam
Some studies have been carried out designing parameters with one end supported by the tunnel support and the other
of the UAM method. Shin et.al (2008) had studied the effect end supported by the ground.
of reinforcing pattern and pipe length considering ground By supporting ground load in this way, the following
displacement and stress changes. The results have shown that effects on tunnel excavation occur:
longitudinal arching along the tunnel axis has reduced set-
tlement and improved face stability. Ranjbarnia et al. (2018) • Restraining ground settlement ahead of the face
has developed an analytical method to study the behavior of • Increasing face stability
UAM in deep tunnels. Displacement of the supported span • Reducing dimensions of tunnel supports (shotcrete, steel
was calculated using the convergence-confinement method. arch support.)
The results showed that pipe diameter is an important factor • Enlarging the cutting face, this makes it possible to use
to control settlement in deep tunnels. large machinery, which in turn, leads to rapid excavation.

13
Environmental Earth Sciences (2020) 79:521 Page 3 of 15  521

Stabilizing the excavation face in both transverse and lon- longitudinal direction elements (pipe length and overlap-
gitudinal directions by an arch-like reinforced zone is the ping length), and the water-cement ratio at grout mix design.
supporting mechanism of the umbrella arch method (Muraki
1997). The structural behavior of the pipe umbrella also
would be discussed in this paper by considering maximum Geology of research field
pipe deformation and applied forces on pipes.
Forepoling design elements had discussed in some previ- The case study is the 4th section of Qazvin–Rasht railway
ous researches. For example, research by Shin et al. (2008) tunnel, located in the north of Iran, shown in Fig. 1. This
has shown that within the pipe length of 1.5D, where D tunnel section has 663 m length and is placed in 110 + 345
is tunnel diameter, an increase in pipe length causes more to 111 + 08 km. The tunnel is drilled in the mountainous
stability. Nevertheless, when the pipe length is longer than area in Shemshak formation (Fig. 2). It can be perceived that
1.5D at the crown, the structural advantage on stability is by moving from point 1 to point 5 and 6 (exit portal of the
not remarkable. Moreover, investigation using the centri- tunnel), sedimentary layers of conglomerate and sandstone
fuge model by Juneja et al. (2010) has proved that fore- are more dominant. Considering longitudinal profile of the
poles reduce the length of settlement ahead of the tunnel tunnel and laboratory experiments of project samples, the
face. However, the width of the settlement does not change. geotechnical profile of the tunnel route can be classified into
Besides, tunnel stability depends on both unsupported length three different zones as below:
and forepoles length. Wang (2012) found that when L = 1.5 h
(h = excavation height) reinforcement would extend to the 1. Sand and clay with relatively weak geotechnical proper-
theoretical Rankin active line, and any further increase ties that are located in the entrance part of the tunnel.
would yield to no extra benefit. Forepoling designing ele- 2. Clay with very weak mechanical properties in the middle
ments have not been comprehensively investigated in previ- part of the tunnel
ous studies. In this paper, finite difference numerical models 3. Sandstone layers at the end part of the tunnel.
have been created to investigate the effects of geometrical
and physical elements through sensitivity analysis. These
parameters are divided into four categories: pipes inclina- The soil and rock properties around the tunnel are rela-
tion, transverse direction parameters (pipes lateral distance, tively weak since most portion of the tunnel is surrounded
pipe diameter and thickness, grout penetration radius), by sandy and clayey materials.

Fig. 1  Location of the study

area (Qazvin–Rasht railway
tunnel)

13
521 
Page 4 of 15 Environmental Earth Sciences (2020) 79:521

conducted to define geotechnical characteristics of the for-

mations. The results are given in Table 1.
In this section of the tunnel, the features of the mate-
rial don’t have the required coherence and carry a very low
plasticity index, making the tunnel condition very sensitive
to disturbance. So to make tunnel construction more stable,
special pre-support like umbrella method is vital.

Excavation and installing support system

NATM is used in the studied area of Qazvin–Rasht railway

tunnel and has been constructed in two phases. The top sec-
tion, a semi-circular with a radius of 3.4 m and a cross-
sectional area of 18.15 m2, is excavated and supported at
first. When this section is drilled about 15–30 m, then the
lower portion of the tunnel having a cross-sectional area of
39.78 m2 will start. Excavators and hydraulic hammers are
Fig. 2  Geological map of Shemshak formation around the tunnel site used in all tunnel faces. As soon as the lower bench excava-
tion is finished, the rock support must be completed and all
the cross-sections must be transformed into a horse-shoe
The lithology of the first and second zones strata is more shape.
close to the studied area. Dominant lithology is conglomer- Figure 3, has shown the excavation and installing the sup-
ate, sandstone, claystone, and mudstone. port system in the top section of the tunnel. The characteris-
In the project site, a borehole of 35 m deep was drilled tics of the primary support system are presented in Table 2.
to obtain required geotechnical data. Laboratory and in situ It should be noticed that the success of excavation and sup-
tests based on ASTM and ISRM standards have been porting method is related to the results of deformation meas-
urements during tunneling. Thus, convergence measurement

Table 1  Geotechnical properties of rock mass in the studied area

Strata ASTM 7263 ASTM D2166, (ASTM D2850) Direct Shear test (ASTM
ASTMD7012 D3080)
Unit weight (KN/m3) UCS (MPa) Modulus of elas- Poisson’s ratio Angle of Cohesion (KPa)
ticity (MPa) friction

Claystone 20 4 50 0.3 35 40
Mudstone 20 3 30 0.3 40 30
Conglomerate 24 15 500 0.24 35 150

Fig. 3  Excavation and install-

ing the support system in top
section

13
Environmental Earth Sciences (2020) 79:521 Page 5 of 15  521

Table 2  Characteristics of used support (Passillo consulting engineer- It is perceived that different approximates makes different
ing 2011) strength and deformation results. In this study, a method is
Support system Properties Value proposed to model the umbrella arch method closer to real
operational conditions.
Shotcrete 𝜎C(MPa) 23 25 cm thickness
E (GPa) 20
γ (kg/m3) 2200
ν 0.15
Steel sets 𝜎C(MPa) 240 ST 37 by 1 m spacing
𝜎t(MPa) 240
E (GPa) 200
ν 0.3
Wire mesh Ø 6 @ 10 mm* 10 mm

stations are located at every 15 m in the tunnel. Any changes

in measurements are recorded periodically during the tunnel
excavation. Numerical modeling is used to examine the role Fig. 4  Umbrella arch components
of the umbrella arch in tunnel stability.

Numerical simulation

Tunneling in weak rock masses usually generates local fail-

ure in unsupported span as well as unstable face conditions.
Only a three dimensional model with an adequate mesh may
relieve these problems and allow detailed investigations on
the supporting systems. Numerical analyses had been carried
out to some of the field cases to model umbrella method.
Nishimaki et al. (1995) conducted a three-dimensional finite
element analysis to model the steel pipe umbrella. The three-
dimensional solutions were used to conduct a back analysis
whereby the umbrella arch is simulated by improving the
material properties of the ground in two-dimensional finite
element analyses. Similarly, using the improvement of the Fig. 5  Soilcrete and grout laboratory samples
material property method, Ohtsu et al. (1995) also studied
the behavior of the ground due to tunnel excavation using a
three-dimensional finite element analysis. Ohtsu et al. (1995)
assumed that the pre-lining arch (forepoles and shotcrete)
acted as a combined shell element. Sato and Ito (1993) used
a three-dimensional finite element method to analyze the
effect of the umbrella arch method on tunnel face stabil-
ity. Most numerical modelers of the umbrella arch method
had crudely simulated the steel pipes and grout material as
a composite material. Hefny et al. (2004) had compared
the ground behavior predicted by two simple and practical
numerical solutions; the first method is based on improving
geotechnical material properties and the second one is mod-
eling steel pipes individually. While beam elements were
used to represent steel pipes. Hoek (2000) showed that the
weighted averages method could be used to estimate the
strength and deformation of the reinforced zone surround-
ing the tunnel periphery. Fig. 6  Uniaxial compressive test

13
521 
Page 6 of 15 Environmental Earth Sciences (2020) 79:521

Table 3  Umbrella arch Umbrella component E (GPa) UCS (MPa) E (GPa) UCS (MPa)
components characteristic
Grout 7 days 28 days
w/c = 0.5 0.9 8.26 2.5 11.42
w/c = 0.7 0.6 4.75 2.1 9.37
Soilcrete 7 days 28 days
w/c = 0.5 g = 70% 2.5 15.82 3 18.10
w/c = 0.7 g = 70% 2.3 13 2.7 14.82
w/c = 0.7 g = 50% 2.3 11.5 2.7 14.48
Pipe 200 240

Table 4  Umbrella arch Pipe diam- Pipe length Overlapping Pipe thick- Pipe inclina- Pipes dis- Water/cement
properties eter (mm) (m) length (m) ness (mm) tion (degree) tance (mm) ratio

90 12 3 7.2 7 400 0.5

Injected grout penetrates from pipe holes to the annular

gap between pipes, open pores, and joints around the bore-
hole, forming the reinforcing zone called soilcrete. It means
that the method is composed of three components: grout,
pipe, and soilcrete, shown in Fig. 4.
To define grout and soilcrete mechanical parameters,
samples are prepared in different mixtures. Grout mixed in
water/cement ratios of 0.5 and 0.7 that are widely used in
umbrella arch operation and soilcrete combined in three dif-
ferent mix designs as follow:
70% grout (w/c 0.5) and 30% field materials,
50% grout (w/c 0.7) and 50% field materials,
70% grout (w/c 0.7) and 30% field materials.
Soilcrete section (left) and grout laboratory sample (right) Fig. 7  Three rows of pipe umbrella
are shown in Fig. 5. Strength and modulus of elasticity of
samples derived at seventh and 28th days of curing, using the 28th day. It should be considered that the strength
the UCS apparatus (Fig. 6). The strength of each forepoling of soilcrete and grout in 7th day of curing is 70% of its
component is shown in Table 3. ultimate value, thus 14th-day parameters are assumed to
In this study, Flac3D was used for numerical mod- be the same as 28th). As top heading excavation is con-
eling  (Itasca 2013). The assumed failure criterion is tinuing and initial linings are installed, the second row of
Mohr–Coulomb. By utilizing the pile element, homogeni- umbrella pipes would be installed to maintain 3-m over-
zation of each pipe, and grout inside with their reinforcing lapping length. After top heading excavation, the bench
area simulated individually. Pile element is selected, as it would be excavated and supported similarly (shotcrete
can simulate shear behavior, bending moments, and forces. and steel sets). Three rows of modeled pipe umbrellas are
The weaker bond is pipe-soilcrete boundary, so normal and shown in Fig. 7.
shears coupling springs parameters are devoted to this joint.
Umbrella method characteristics in Qazvin–Rasht
railway tunnel are given in Table 4. Pre-supports pipes Umbrella arch influence on tunnel stability
are installed, hence after one week of pipes installation;
excavation would be started under a reinforced area. One The necessity of the umbrella arch method in Qaz-
meter of top heading is excavated and would be supported vin–Rasht railway tunnel has been investigated by elimi-
immediately by shotcrete and steel sets (250 mm thick nating its application in numerical modeling. Figure 8
shotcrete and steel ribs IPE 18 spacing 1 m). The exca- shows the plastic zone and displacement magnitude con-
vation round is 1 m per day, so after the excavation of tours with and without the umbrella arch method. It can
7th round, umbrella method parameters are replaced by be seen that the role of forepoles are crucial as they have a

13
Environmental Earth Sciences (2020) 79:521 Page 7 of 15  521

Fig. 8  Numerical comparison of umbrella arch performance in tunnel stability

positive effect on lowering plastic zone and tunnel defor- numerical modeling and periodical reading of instru-
mation and increasing the tunnel stability. mentation are closely matched. The results indicated
Tunnel behavior has been monitored by using five-point that using the umbrella method has reduced convergence
convergence pins during tunnel construction, every 20 m. magnitudes by 49% and has increased tunnel stability
Five pins were installed around the tunnel. After each significantly.
face advance, the distance between the pairs of pins is
measured by 0.01 mm accuracy, Fig. 9. In 110 + 600 km,
there is a monitoring station where the umbrella arch is Investigating pipes structural behavior
installed. Convergence magnitudes are predicted with and
without forepoling method using numerical simulation According to pipes axial force, bending moment, and dis-
and compared with in situ measurement using the UAM placement at different stages of excavation, the overall
method (Fig. 10). It can be perceived that the results of behavior of pipes is similar to beams with the free end at

13
521 
Page 8 of 15 Environmental Earth Sciences (2020) 79:521

pipes could move with ground displacement. The mechani-

cal model of pipe umbrella reinforcement during tunnel
excavation is shown in Fig. 11. Point B has shown primary
lining which can be substituted by simply supported end
with w0 as initial displacement and θ0 as the angle, consid-
ering the delay effect of initial lining, BC is an unsupported
section, CD disturbed section and DE undisturbed section.
According to the tunnel size and 400  mm distance
between pipes in the transverse direction, there should be
29 pipes in the tunnel periphery, one at the center of the
tunnel crown, 14 pipes at the right side, and the others
at the left. Figure 12, illustrates shear force in the central
pipe and other 14 pipes at one side of the tunnel periphery
in a 9-m excavation round (just before the next pipes row
have prepared for installation). It should be noticed that
because of the axisymmetric condition, the other 14 pipes
have the same behavior. The maximum load is applied at
Fig. 9  Convergence meter arrangement around the tunnel the top pipe. The maximum vertical load for first to seventh
pipe (upper half of tunnel arch) occurs at one-third of pipe
length and from eighth to 14th pipe (lower half of tunnel
arch) happens at one-fourth of pipe length. It means that as
the tunnel arch changes, the location of structural parame-
ters would be changed. Though maximum load-bearing has
happened at the central pipe in the tunnel crown, the impor-
tance of other pipes should not be considered negligible.
It should be noticed that the location of maximum pipe
moment, force, and displacement would change as the exca-
vation continues as well as residual pipe length ahead of
the face.
Vertical displacement of the central pipe is shown in
Fig. 13. Maximum deformation has happened near the tun-
nel face, for example in the 4th excavation round, the maxi-
mum displacement of the pipe is located in about 4 m. While
Fig. 10  Tunnel convergence obtained from numerical simulation and pipe length ahead of face reaches to 0.5L or even shorter
in situ monitoring (L is the length of the pipe), maximum deformation would
be at 0.5L to 0.75L. By continuing the drilling process and
approaching the next pipe series, maximum displacement
in overlapping lengths (here 10, 11, and 12 m) would be
located at 0.5L. This is emphasizing that the next pipe series
would tolerate extra load and deformation.
Shear force variation is presented in Fig. 14. It is per-
ceived that shear force increases slightly as the face
advances, the maximum amount is located at one meter away
from the face, which indicates that load transfer takes place
to an un-excavated part in tunnel direction. By decreasing
the undisturbed zone ahead of the tunnel face and before
installing the next pipes series, the maximum shear force
Fig. 11  Mechanical model of pipe umbrella reinforcement
would be at 0.75L and behind the tunnel face. It means that
the next pipe row should be installed before the effectiveness
of the pipe umbrella foundation ahead of the face decreases.
the face side and supported by primary support at the other The inflection point in the momentum diagram is follow-
side. Although each pipe is founded in the ground (ahead ing the same pattern as the load diagram. Positive amounts
of the face) in the longitudinal and transversal directions, represent the pressure, and negative values stand for tension.

13
Environmental Earth Sciences (2020) 79:521 Page 9 of 15  521

Fig. 12  Pipes shear force in the

vertical direction at the nine-
meter excavation

Fig. 13  Pipe deformation in the

vertical direction

Forepoling design elements Pipe inclination

Forepoling design elements are investigated in this case At the initial design layout, the inclination of the pipe is
study using sensitivity analysis through numerical mode- 7 degrees. The effect of pipe inclination on tunnel crown
ling. These elements are divided into four categories: pipes convergence is investigated at 5°, 7°, 10°, and 12°. Crown
inclination, transverse direction parameters (pipes lateral displacement is lowered by 7% by increasing the pipes angle
distance, pipe diameter and thickness, grout penetration from 5° to 12° (Fig. 16). During forepoling construction,
radius), longitudinal direction elements (pipe length and expanding angle provides the required space for the installa-
overlapping length), and water/cement ratio at grout mix tion of the next row pipes. On the other hand, over increasing
design. Figure 15, illustrates umbrella arch geometric design reduces the effective length. So, choosing the optimum angle
elements. according to the project requirements and capability of drill-
To evaluate the effectiveness of each parameter, other ele- ing equipment is important.
ments are considered as the designed layout, mentioned in
Table 4.

13
521 
Page 10 of 15 Environmental Earth Sciences (2020) 79:521

Fig. 14  Pipe shear force in the

vertical direction

Transverse direction elements (pipes lateral

distance, diameter and thickness, grout penetration
radius)

Regarding stabilization in the transverse direction, an arch-

like reinforced zone carries the ground load. In this section,
the effects of transverse direction elements, including pipes
distance and grout penetration radius, pipes thickness, and
diameter are discussed.
Pipes lateral distance is 40 cm at the initial layout. The
effect of pipe distance was examined in three cases, 30, 40,
and 50 cm. Another variable is the grout penetration radius
that is considered unchanged, 40 cm. While pipes lateral
distance is 40 and 30 cm, grout bulbs are tangent to each
other or overlapped, in 50 cm pipe distance, unreinforced
Fig. 15  Umbrella arch geometric design elements (Hoek 2004) materials remain between pipes. Figure  17 shows that
increasing pipe lateral distance and unreinforced material
between pipes have raised crown convergence by 12%.
Pipes distance depends on the rock mass or soil type
surrounding steel pipes and necessary support pressure.
The minimum distance should allow the rock mass to cre-
ate a local arch between consecutive pipes. It ensures the
designed shape of the tunnel without local failure (Volk-
mann and Schubert 2006). On the other hand, decreasing
pipe lateral distance would require more pipes in the tun-
nel periphery. It takes more excavation, which is time-
consuming and cost-intensive.
Increasing pipe thickness and diameter would make
support stiffer and control support pressure which can
reduce the number of pipes. As forepoling is a passive
support system, increasing the pipe diameter would make
support activation immediate due to the bending increase.
Fig. 16  Tunnel crown convergence in different pipe inclination

13
Environmental Earth Sciences (2020) 79:521 Page 11 of 15  521

Fig. 17  Tunnel crown conver-

gence in different pipe lateral
distance, pipe thickness, and
diameter

Hence, more stability is provided before excessive defor- longitudinal parameters (pipe length and pipe overlapping
mation takes place. length) are investigated.
As discussed in “Numerical simulation”, in the first At initial modeling, the length of the pipes is 12  m.
scheme, pipes diameter and thickness were 90 mm and Numerical analyses carried out to study the influence of pipe
7.2 mm. In Fig. 17, crown convergence is compared in length at four different sizes: 9, 12, 15, and 20 m, where pipe
D = 90 mm, t = 7.2 mm, and D = 110 mm, t = 8 mm, while overlapping length is 3 m, excavation rounds is 6 m and
other factors are the same as the initial layout. Crown dis- there is just one-row forepole at all cases. Figure 18, reveals
placement magnitudes are reduced by 3% in D = 110 mm that pipe length increase has decreased crown displacement.
and t = 8 mm. There is a significant load transfer along the tunnel
axis as the location of maximum stress has slightly moved
Longitudinal direction elements (pipes length, away from the face with an increase in pipe length (Shin
pipes overlapping length) et al. 2008). Also, considering the ground quality and tun-
nel height, the effectiveness of the forepoles foundation
To increase stability in the longitudinal direction, the ground ahead of the face decreases when there is a decline in
load of the recently excavated portion is supported through pipe length (Volkmann and Schubert 2007). In 9 m pipes,
a beam with one end supported by the tunnel support and the foundation at the deformation zone ahead of the face
the other found in the ground. In this section, the effects of has decreased and more deformation has occurred. On
the other hand, pipes with 12, 15, and 20 m length have

Fig. 18  Tunnel crown conver-

gence in different pipe length
and overlapping length (in 6 m
excavation round)

13
521 
Page 12 of 15 Environmental Earth Sciences (2020) 79:521

shown close results. As the tunnel face excavated, loads ahead of the tunnel face. Therefore, the overlapping length
are transferred to the area ahead of the face, and the defor- should at least cover the disturbed zone ahead of the tunnel
mation area has extended from 1D to 1.5D, where D is face.
tunnel diameter. Thus, pipes lower than 1D do not have a In projects that settlement reduction ahead of the face is a
considerable effect on tunnel stability; on the contrary in limitation factor, next row forepoles should be installed well
pipes longer than 2D, which are far from disturbing zone, before pipes support influence, and their length at failure
stability effectiveness is not remarkable. zone decrease.
Hence, reduction in pipe length, at disturbed zone ahead To investigate the effects of overlapping length, two mod-
of face, will increase plastic zones and magnifies deforma- els have been analyzed, one with 3 m overlapping length
tion. To avoid any breakdown, next row pipes should be and the other 5 m, while pipe length is 12 m in both situa-
installed timely while the efficient overlapping length is tions (Fig. 18). The long overlapping length will cover more
provided. tunnel length with double forepoles; thus, less deformation
The determination of overlapping length depends on the would be expected. This length should be set concerning
distance and length of the pipes foundation ahead of the face project requirements. According to Terzaghi (1943), the
(Volkmann and Schubert 2007). These two factors relate length of loose zone ahead of tunnel face while using hori-
to rock characteristics and tunnel geometry. Every excava- zontal pipes can be calculated as Eq. (1).
tion round would decrease for poles length and consequently ( 𝜑)
reduce pre-support efficacy. d = h tan 45 − , (1)
2
The beneficial impacts of UAM method on tunnel stabil-
ity is decreased by approaching end parts of pipes. In this where d is the length of loose zone ahead of face, h is the
condition, the UAM system cannot transfer the overburden- height of tunnel face, and φ is the friction angle of surround-
related loads to the primary support system and ground ing rock.
As minimum overlapping length should cover failure
zone ahead of tunnel face, by considering pipe inclina-
tion (Fig. 19), the following relation can be established as
Eq. (2).
� �
⎛ 𝜑 ⎞
⎜ sin 45 − 2 ⎟
Lov = h⎜ � � ⎟, (2)
⎜ sin 45 + − 𝜃L ⎟
𝜑
⎝ 2 ⎠

where L ov is pipe overlapping length and θ L is pipe

inclination.
Since in this case study, tunnel diameter is 6.8 m, the
minimum overlapping length could be 2.35 m.

Fig. 19  Pipes length and overlapping length scheme (Shin et al. 2008)

Fig. 20  Tunnel crown conver-

gence in different w/c ratio

13
Environmental Earth Sciences (2020) 79:521 Page 13 of 15  521

Water/cement ratio at grout mix horizontal in-place inclinometer chains installed in pipe
umbrella to control system behavior.
Geometrical elements of forepoling have been investigated In another study by Shin et  al. (2008), large-scale
at a water-cement ratio of 0.5. In this section, the influence UAM was simulated in the laboratory along with a three-
of water/cement ratio, regardless of other injection param- dimensional finite element simulation to investigate the
eters, has been studied at 0.5 and 0.7, which are widely reinforcement mechanism and improve the design method.
used in forepoling implementation. As shown in Fig. 20, Face stability leads to longitudinal load transfer to the
the water/cement ratio increase from 0.5 to 0.7 has soared undrilled area and reduces stress concentration that ulti-
displacement magnitudes, although these changes are not mately reduces deformation and increases stability.
significant in this modeling. In this study, the structural behavior of UAM has inves-
tigated using numerical modeling. According to pipes
axial force, bending moment, and displacement at different
Discussion stages of excavation, the overall behavior of pipes can be
predicted and optimized. Therefore, the mode of loading
The primary objective of this study is to accurately investi- and its effect on overall system behavior is analyzed dur-
gate the useful impacts of UAM on the stability of tunnels ing tunneling.
in weak ground conditions. In this research, initially, the The influential impact of design parameters of UAM
importance of using UAM method in soft ground evalu- pipes is investigated numerically and empirically by sev-
ated using numerical simulation validated by in-situ tunnel eral authors: Shin et al. (2008), Ranjbarnia et al. (2018),
deformation measurements. Different approaches of UAM and Zarei et al. (2019). These studies have only considered
numerical simulation makes different strength and defor- two or three design factors in their modeling, while in this
mation results. In most of 2D modeling, UAM is simulated study a comprehensive evaluation has been accomplished
by strengthening material properties of ground such as using sensitivity analysis on pipe characteristics such as pipe
Nishimaki et al. (1995). The homogenous model suggested inclination, transverse direction elements (pipes lateral dis-
by Hoek (2000) is a crude approach. Ohtsu et al. (1995) tance, diameter and thickness, grout penetration radius) and
have simulated forepoles and shotcrete as a combined sell longitudinal direction elements (pipes length, pipes overlap-
element using three-dimensional finite element analyses. ping length), also optimum values on tunnel stability are
Sato and Ito (1993) had simulated steel pipes and grout investigated.
material as composite materials. This modeling cannot According to the designed models of the umbrella arch
make the local arching between the structural elements as method in Qazvin–Rasht railway tunnel the final layout can
well as longitudinal stress transfer. be expressed as follows:
UAM supporting system is often designed by 3D
numerical simulation to consider the complex interaction • Pipes length: 12 m
with the ground. In this study, authors discovered that the • Pipes overlapping length: 2.5–3 m
pile element is more suitable for forepoling simulation, • Pipes diameter: 100 mm
in agreement with previous studies (Broch et al. 2006; • Pipes thickness: 8 mm
Volkman and Schubert 2006; Vlachopoulos and Sieder- • Pipes distance: 40 or 50 cm
ichs 2014; Oke et al. 2014). UAM elements comprised of • Grout penetration radius: center to center
pipes, injected grout and reinforced area between pipes i.e. • Pipe inclination: 10° or 12°
soilcrete which are simulated individually through finite • Water/cement ratio: 0.5.
difference program. Grout and soilcrete parameters are
examined in different mixtures considering cement harden- It should be noticed that the proposed design would be
ing. Forepoling strength changes are applied in continuous effective provided that center to center grout penetration
excavation steps and cement hardening procedures. radius is achieved. This factor depends on injecting char-
Monitoring structural behavior of pipe can evaluate acteristics, water/cement ratio, and rock or soil geotechni-
deformations that would be beneficial in controlling sys- cal parameters which should be precisely investigated at the
tem behavior accurately during excavation and support design stage.
installation. It would help to identify probable changes and
detect root causes and provides the opportunity to adapt
the support system at earlier stages.
Volkmann and Schubert (2005) conducted a series
of geotechnical monitoring during tunneling. They used

13
521 
Page 14 of 15 Environmental Earth Sciences (2020) 79:521

Conclusion on tunnel height, pipe inclination, and friction angle of

surrounding rocks.
In this study, the umbrella arch method is employed in Qaz- • The water–cement ratio of 0.5 has provided more stabil-
vin–Rasht railway tunnel and has been modeled using FLAC ity than 0.7.
3D software. Each pipe is simulated individually by consid-
ering grout and soilcrete geotechnical characteristics.

• The importance of the umbrella arch method is investi- References

gated with and without employing the forepoling tech-
nique in numerical modeling. Forepoles reduce tunnel Aksoy CO, Onargan T (2010) The role of umbrella arch and face bolt
convergence by 49%, which has been validated with in- as deformation preventing support system in preventing building
damages. Tunn Undergr Sp Technol 25:553–559
situ instrumentation results. Bagherzadeh P, Goshtasbi K (2013) Numerical modeling of
• In analyzing the structural behavior of pipes, each pipe umbrella arch method in collapsing zones. J Mine Metals Fuels
is modeled as a simple beam that is free at the face side 61(11):332–336
and supported by the primary lining of the tunnel such Bagherzadeh P, Goshtasbi K (2014) Umbrella arch modeling and its
performance at Qazvin–Rasht–Anzaly railway tunnel. Iran J Min
as lattice or steel sets on the other side. Eng IRJME 9(24):45–56
• The location of the maximum moment, forces, and dis- Broch E, Lu M, Trinh QN (2006) Three dimensional modeling of
placement in pipes would be changed with ongoing exca- tunnel cave-in and spiling bolt support. In: Leung CF, Zhou YX
vation steps and residual length of pipes in the ground. (eds) Proceedings of ISRM international symposium 2006 and
4th Asian rock mechanics symposium. World scientific Publishing
• The top pipe has tolerated the maximum deformation and Co., Pte. LTd, Singapore
shear forces occurred near the tunnel face. By changing Carrieri G, Fiorotto R, Grasso P, Pelizza S (2002) Twenty years of
the tunnel arch, the location of the maximum force has experience in the use of the umbrella-arch method of support for
shifted. Maximum forces in pipes located in the upper tunnelling. In: 4th international workshop on micropiles, Venice,
Italy
half of tunnel arch occurred in one-third of the initial Carrieri G, Grasso PG, Mahtab A, Pelizza S (1991) Ten years of expe-
length, and in the lower half of tunnel arch, is in one- rience in the use of umbrella-arch for tunnelling. In: Proceedings
fourth of initial pipe length. ofthe congress on soil and rock improvement in underground
• Pipe inclination has a positive effect on tunnel stabil- works, Milano, vol 1, pp 99–111
Gao F, Zhang K, Hou A, Guo J (2015) Deformation and internal force
ity. However, there are some limitations as expanding analysis of pipe roof support using the beam on elastic foundation
angle is applied for installing the next row pipes as well theory. Elast J Geotech Eng 20:13013–13022
as forming an umbrella in tunnel periphery, it is vital to Hisatake M, Ohno S (2008) Effect of pipe roof supports and the exca-
choose the optimum angle according to project require- vation method on the displacement above a tunnel face. Tunn
Undergr Sp Technol 23(2):120–127
ments and drilling equipment capability. Hefny A, Tan W, Pathegama R, Sharma J, Zhao J (2004) Numerical
• The more the pipes lateral distance is, the less the tun- analysis for umbrella arch method in shallow large scale excava-
nel stability is. However, due to limitations in increas- tion in weak rocks. In: Proceedings of 30th ITA-AITES world
ing the number of pipes, using thick pipes and providing tunneling congress, Singapore, 22–27 May 2004, pp 1–7
Heidari M, Tonon F (2015) Ground reaction curve for tunnel with jet
center to center grout penetration radius (which depends grouting umbrellas considering jet grouting hardening. Int J Rock
on grout, rock, and injection characteristics) is recom- Mech Min Sci 76:200–208
mended. Hoek E (2000) Numerical modelling for shallow tunnels in weak rock.
• Pipe length increase has a positive effect on tunnel sta- In: Presented during the 5th GRC Lecture, NTU, Singapore
Hoek E (2004) Numerical modeling for shallow tunneling. Discussion
bility. However, where the pipe length is more than 2D, Paper 3
its effectiveness is not significant as it is far beyond the Itasca Consulting Group Inc (2013) FLAC 3D version 5.01 User’s
deformation zone and when the length of the pipe is guide. Minnesota, USA
shorter than 1D, it does not entirely cover the disturbed Juneja A, Hegde A, Lee FH, Yeo CH (2010) Centrifuge modelling
of tunnel face reinforcement using forepoling. Tunn Undergr Sp
zone. Therefore, a desirable limit would be between 1D Technol 25:377–381
to 1.5D. Kim SH, Baek SH, Moon HK (2005) A study on the reinforcement
• As grout foundation ahead of the face and pipe length effect of umbrella arch method and prediction of tunnel crown
at disturbed zone would decrease with continuation in and surface settlement. In: Solak E (ed) Underground space use,
analysis of the past and lessons for the Future
excavation steps, it is essential to install next row pipes Klotoé CH, Bourgeois E (2019) Three dimensional finite element
at the right time. Overlapping length should at least cover analysis of the influence of the umbrella arch on the settlements
the length of the failure zone ahead of the face. An equa- induced by shallow tunneling. Comput Geotech 110:114–121
tion is proposed for estimating overlapping length based Lunardi P (2000) The design and construction of tunnels using the
approach based on the analysis of controlled deformation in rocks
and soils. Tunn Tunn Int Suppl

13
Environmental Earth Sciences (2020) 79:521 Page 15 of 15  521

Muraki Y (1997) The umbrella method in tunnelling. M.S, Civil Engi- Taromi M, Eftekhari A (2018) Tunnel design and construction process
neering, Massachusetts Institute of Technology in difficult ground conditions with analysis of controlled defor-
Nishimaki A, Mitarashi Y, Uematsu S (1995) “Study on the effects mation (ADECO) approach; a case study. Int J Min Geo-Eng
of the AGF method” (Proposal of a simplified AGF design tech- 52(2):149–160
nique). South East Asian Symposium on Tunneling Association, Terzaghi K (1943) Theoretical soil mechanics. Wiley, New York
Bankok, pp 125–132 Torres CC, Diederichs M (2008) Mechanical analysis of circular lin-
Ocak I (2008) Control of surface settlements with umbrella arch ers with particular reference to composite supports For example,
method in second stage excavations of Istanbul Metro. Tunn liners consisting of shotcrete and steel sets. Tunn Undergr Sp
Undergr Sp Technol 23:674–681 Technol 24:506–532
Ohtsu H, Hakoishi Y, Nago M, Taki H (1995) A prediction of ground Vlachopoulos N, Diederichs MS (2014) Appropriate uses and practical
behavior due to tunnel excavation under shallow overburden with limitations of 2D numerical analysis of tunnels and tunnel support
long length forepoling. South East Asian Symposium on Tun- response. Geotech Geol Eng 32(2):469–488
neling Association, Bankok, pp 157–164 Volkmann GM, Schubert W (2005) The use of horizontal inclinometers
Oke J, Vlachopoulos N, Diederichs MS (2014) Numerical analyses in for the optimization of the rock mass –support interaction. In:
the design of umbrella arch system. J Rock Mech Geotech Eng Erdem YT, Solak AA (eds) Proceedings of the 31st ITA-AITES
6(6):546–564 world tunneling congress, underground space use: analysis of
Oke J, Vlachopoulos N, Diederichs MS (2016) Semi-analytical model the past and lessons for the future, 7th–12th May 2005, Turkey.
for umbrella arch system employed in squeezing ground condi- Balkema Publishers, London, ISBN 04 1537 452 9, pp 967–972
tions. Tunn Undergr Sp Technol 56:136–156 Volkmann GM, Schubert W (2006) A contribution to the design of
Passillo Consulting Engineers (2011) Qazvin–Rasht railway studies, tunnels Supported with pipe roof support. In: Rock mechanics in
second stage, tunnel No. 3 underground construction: ISRM international symposium 2006:
Peila D (2005) Ground reinforcing for tunnelling the example of 4th Asian rock mechanics symposium. World scientific Pub. Co.
steel pipe umbrella. In: ITA/AITES—training course tunnel Inc, Singapore
engineering Volkmann GM, Schubert W (2007) Geotechnical model for pipe roof
Pelizza S, Peila D (1993) Soil and rock reinforcements in tunneling. supports in tunneling. In: Proceeding of 33rd ITA-AITES World
Tunn Undergr Sp Technol 8(3):357–372 tunneling congress, underground space-the 4th dimension of
Qian Z, Zou J, Pan Q, Dias D (2019) Safety factor calculations of a metropolises. Taylor & Francis Group, London, pp 755–760
tunnel face reinforced with umbrella pipes: a comparison analysis. Wang H (2012) Optimization of pipe roof reinforcement applied in
Eng Struct 199:109639 tunnel construction under complex conditions. Electron J Geotech
Ranjbarnia M, Rahimpour N, Oreste PA (2018) A simple analytical Eng 17:301–310
approach to simulate the arch umbrella supporting system in deep Yoo C, Shin H (2003) Deformation behaviour of tunnel face reinforced
tunnels based on convergence confinement method. Tunn Undergr with longitudinal pipes-laboratory and numerical investigation.
Sp Technol 82:39–49 Tunn Undergr Sp Tech 18:303–319
Salmi EF, Nazem M, Giacomini A (2017) A numerical investigation Yahya S (2016) Forepoling pre-support application in the shallow
of sinkhole subsidence development over shallow excavation in tunneling through weathered granite. Dissertation submitted in
tectonised weak rocks: the Dolaei Tunnel’s excavation case. Geo- partial fulfillment of the requirements for the degree of Master of
tech Geol Eng 35(4):1685–1716 Science University of Technology, Malaysia
Sato J, Ito J (1993) Numerical analysis of the umbrella method for Zarei H, Moarefvand P, Salmi EF (2019) Numerical modeling of
tunnel excavation. In: Infrastructures Souterraines de Transports, umbrella arch technique to reduce tunnelling induced ground
Reith (ed) Balkema, Rotterdam, pp 355–360 movements. Environ Earth Sci 78(10):291
Shin JH, Choi YK, Kwon OY, Lee SD (2008) Model testing for pipe-
reinforced tunnel heading in a granular soil. Tunn Undergr Sp Publisher’s Note Springer Nature remains neutral with regard to
Technol 23:241–250 jurisdictional claims in published maps and institutional affiliations.
Schumacher FP, Kim E (2016) Modeling the pipe umbrella roof sup-
port system in a western US underground coal mine. Int J Rock
Mech Min Sci 60:114–124
Song KI, Cho GC, Chang SB, Lee IM (2013) Beam-spring structural
analysis for the design a tunnel pre-reinforcement support system.
Int J Rock Mech Min Sci 59:139–150

13