Tunnelling and Underground Space Technology 82 (2018) 39–49

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology

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

A simple analytical approach to simulate the arch umbrella supporting T

system in deep tunnels based on convergence confinement method
⁎
Masoud Ranjbarniaa, , Nima Rahimpoura, Pierpaolo Oresteb
a
Department of Geotechnical Engineering, Faculty of Civil Engineering, University of Tabriz, Tabriz, Iran
b
Department of Environmental, Land and Infrastructure Engineering, Politecnico di Torino, Turin, Italy

A R T I C LE I N FO A B S T R A C T

Keywords: In this study, a simple analytical method is presented to investigate the behavior of arch umbrella supporting
Tunnel system (the forepolling system) in deep tunnels. For this purpose, a qualitative distribution of the load on the
Forepoling arch umbrella elements (the beams) is proposed by using longitudinal deformation profile of unsupported
Weak rock mass tunnel. The flexural stiffness of an element is then obtained and finally, the deformation of supported tunnel are
CCM
calculated by convergence-confinement method. That is, further advancement of tunnel leads the more length of
umbrella elements to be activated; and therefore, the corresponding stiffness is reduced. Hence, the non-linear
support characteristic curve is drown to intersect the ground response curve.
Sensitive analyses are performed, and on the basis of the proposed analytical method, some graphs are
prepared to give the ratio of total displacement of supported tunnel to that of unsupported for the various arch
umbrella systems installed in the tunnels with different potential of squeezing. Some examples are solved, and
the results are compared with those obtained by FLAC3D software.

1. Introduction et al. 2014; Gao et al. 2015).

Regarding the issue of tunnel face stability and ground surface
In the underground excavations, according to the geological con- settlements control, Peila (1994) by a tridimensional numerical simu-
ditions, various excavation techniques and stabilization methods are lation demonstrated that the displacement of tunnel face was reduced
used. For good rock mass quality in a low depth, a light support system around 40% when the longitudinal bolts in tunnel face were also in-
is sufficient while for a weak rock mass at a great depth (i.e. the very stalled (in addition to foreploes). The ground settlements can be con-
squeezing condition), a combination of advanced excavation methods trolled efficiently using arch umbrella (Ocak, 2008), and it is sub-
and pre-support approaches are necessary prior to installation of very stantially reduced if the longitudinal bolts are accompanied with the
stiff support system. In fact, the arch umbrella supporting system ac- pre-supporting (Aksoy and Onargan, 2010). Song and Wang (2011)
companied by the sequential excavation methods must be used for the studied the face stability of excavation using theory of limit analysis.
safe excavation (Pelizza and Peila, 1993; Hoek and Marinos, 2000; Peila Considering structural performance of the arch umbrella elements
and Pelizza, 2003). in the shallow tunnels, on the other hand, a number of studies have
However, the arch umbrella supporting system is utilized in the been carried out. Peila and Oreste (1998) proposed the distribution of
shallow tunnels, e.g. in urban metro tunnels, too. In these tunnels, in load such as that shown in Fig. 1 where the value of load was assumed
addition to stabilize tunnel face, the control of ground surface settle- to be equal to the overburden (i.e. the soil weight from ground surface
ments is the other important aim. Numerous researches have been to tunnel crown). In this analytical study, the system was simulated by
conducted to explore different aspects of this pre-support system per- cantilever beam element, which the buried side leaned on the tunnel
formance (Peila, 1994, 2013; Peila and Oreste, 1998; Volkmann and face and the other side on the main support system. The Winkler elastic
Schubert, 2006, 2010; Ocak, 2008; Wang and Jia, 2008, 2009; Aksoy springs was used to model the reactions of soil (as the elastic founda-
and Onargan, 2010; Song and Wang, 2011; Song et al. 2013, Zhang tion). Volkmann and Schubert (2006, 2007) and Volkmann et al. (2006)

⁎
Corresponding author.
E-mail address: [email protected] (M. Ranjbarnia).

https://doi.org/10.1016/j.tust.2018.07.033
Received 29 June 2017; Received in revised form 7 March 2018; Accepted 28 July 2018
0886-7798/ © 2018 Elsevier Ltd. All rights reserved.
M. Ranjbarnia et al. Tunnelling and Underground Space Technology 82 (2018) 39–49

for preliminary design through the prepared graphs. In this method, the
load distribution is obtained by employing convergence confinement
method (CCM). As a result, the longitudinal deformation profile (LDP)
of supported tunnels is found.
However, some assumptions are made to simplify developing of the
simulation as the follows:

- The complex interaction between the single umbrella element and

the surrounding ground is not studied. That is, the shear stress
which might be developed between the ground and element is ig-
nored, and the connection between the element and ground is as-
sumed to be rigid. Therefore, the stress principle direction does not
rotate.
Fig. 1. Simulation of arch umbrella elements by Peila and Oreste (1998). - The effect of the umbrella elements deflection (or displacements) on
the load redistribution of elements is ignored.
- The arching phenomenon of ground is not developed between the
through calibration of the 3D numerical model (FLAC3D) with in-situ umbrella elements i.e. the loose rock mass between the elements in
monitoring of settlements a head of tunnel face clarified the influence the circumferential direction does not fallen into tunnel. As a result,
of the design parameters to optimize the used support system. Based on the umbrella elements contribution is considered in the form of
back-calculations of in-situ measurements, they defined the structural uniform pressure to the zone influence of each element.
properties for the arch umbrella elements and the changes in load re-
distribution (Volkmann and Schubert, 2010). In the other study, Galetto 2. Analytical simulation
et al. (2013) through physical modeling demonstrated that the dis-
placement of umbrella elements in sandy soil starts ahead of the face The convergence confinement method is used to predict deforma-
with a distance of about a tunnel radius. The solutions presented by tion of tunnel wall through considering the interaction of support
Wang and Jia (2009), Marchino et al. (2010), Song et al. (2013), Zhang system and rock surrounding tunnel. In the case of arch umbrella ele-
et al. (2014); Gao et al. (2015) are the other attempts to obtain the ments as the support system, it is necessary to find the corresponding
displacement and (or) load of the arch umbrella system in the shallow stiffness which itself depends on load distribution.
tunnels.
The loading mechanism of the arch umbrella supporting system in 2.1. The load distribution on a single umbrella element
the deep tunnels is different. That is, unlike shallow tunnels where the
load on the umbrella elements can be quickly found through con- The loading process of a single umbrella element is explained in
sidering overburden, the ground arching effect influences the loading in Fig. 2. In each excavation step, tunnel wall displacement causes that the
deep tunnels so that it is not easy task to obtain it in very squeezing point (A) is much loaded, and at the same time it further moves toward
conditions. the equilibrium point in ground response curves. Since much length of
Oke et al. (2012) conducted a parametric study by a numerical si- the umbrella element is loaded in foregoing excavations, the stiffness is
mulation to demonstrate the influence of arch umbrella elements and reduced, and support characteristic curve (SCC) might be non-linear.
rock mass interaction. Then, they also investigated the influence of In this study, the behavior of umbrella elements in the entrance
elements number (say density) and their inclination, and concluded section of tunnels to the weak ground are simulated. This section is the
that the number of the elements as well as size of the elements are the critical, because, as shown in Fig. 3, the umbrella elements are sup-
most important factors in controlling the stability (Oke et al., 2014a). In ported from tunnel face meanwhile their other side are free. The fol-
the other effort, from numerous case studies, they presented a metho- lowing advancement of the excavation and then installation of the main
dology for selection of umbrella arch system based on the geological support leads the free side of next groups of umbrella elements to be
conditions, stress and required stiffness of support (Oke et al., 2013; lean on.
2014b). Recently, they also proposed a second order equation for dis- As the arch umbrella elements cause to reduce the ultimate total
tributed load through a semi-analytical solution in which, it is assumed, displacement, longitudinal deformation profile (LDP) of the supported
the beam lays on the elastic foundation (Oke et al., 2016). Heidari and tunnel can be considered as that for un-supported tunnel (Heidari and
Tonon (2015) considered the hardening effect of jet grouting umbrella Tonon, 2015). Vlachopoulos and Diederichs (2009) proposed the LDP
elements in controlling of deep tunnel convergence. of a circular tunnel can be calculated by
In fact, design of the arch umbrella supporting system (or any other
complicated structure) is often carried out by 3D numerical simulation ⎡ 1 r x' ⎤
u (x ) = Uul ⎢1−⎛1− exp ⎛−0.15 e ⎞ ⎞ exp ⎛⎜−1.5 ⎞⎟ ⎥ for x' ≥ 0
to take its complex interaction with ground, into account. However, the 3 ⎝ r ⎠⎠ re ⎠
⎣ ⎝ ⎝ ⎦
analytical solutions provide formula to high speed and preliminary
(1-a)
design which are sometimes sufficient for practical purposes, and de-
termine the influence of each parameter explicitly. On the other hand, 1 ⎡ x '−0.15re ⎞ ⎤
u (x ) = U exp ⎜⎛ ⎟ for x' ≥ 0
there is no comprehensive analytical solution to explore the unknown 3 ul ⎢ ⎝ r ⎠⎦
⎥
(1-b)
aspects of arch umbrella supporting system in deep tunnels i.e. the load ⎣
distribution as well as the load magnitude on this support system is not where the parameters Uul and re respectively are ultimate unsupported
still clear. tunnel convergence and elastic-plastic boundary radius which can be
To bridge this gap, this paper presents a new analytical method to calculated by the standard analytical approaches having been presented
simply explain the behavior of arch umbrella system and provides a tool till now. The variables r and x ' are tunnel radius, and distance of a

40
M. Ranjbarnia et al. Tunnelling and Underground Space Technology 82 (2018) 39–49

Fig. 2. The loading process of arch umbrella elements in excavation stages of tunnels.

desired section from tunnel face. 1 ⎡ x '−0.15(C2 × re) ⎞ ⎤

Therefore; the modified LDP of the tunnel can be found by Eq. (2) u(x) = (C1 × U ) ⎢exp ⎛⎜ ⎟
⎥ for x≥0
3 ul ⎝ r ⎠⎦
⎣
(2-b)
⎡ 1 (C × re) ⎞ ⎞ x' ⎞⎟ ⎤
u(x) = (C1 × Uul) ⎢1−⎛1− exp ⎛−0.15 2
⎜ exp ⎛⎜−1.5
⎟
⎥
3 ⎝ r ⎠⎠ (C 2 × re) ⎠ ⎦
⎣ ⎝ ⎝ where the parameter C1 and C2 are the reduction coefficients which are
for x ' ≥ 0 (2-a) unknown.
The load distribution on a single umbrella element can be obtained

41
M. Ranjbarnia et al. Tunnelling and Underground Space Technology 82 (2018) 39–49

the umbrella element located in the excavated length, and Lt-buried is the
part of buried length of the element which the load acts on and might
be about half of the tunnel diameter (Hoek, 1999).
As the displacement, shear force, and moment are negligible at the
end of the buried length, the corresponding boundary condition can be
assumed such as shown in Fig. 6. In this condition, from static equili-
brium point of view, the ground reaction shown by the springs beneath
the buried length equals to the distributed load on total length of the
element.
As the magnitude of the load is still unknown, neither the ground
reaction nor force of each spring, the following assumptions are made
to simplify the model:

- The load distribution on the buried length is considered triangular

but on a smaller length which its value is chosen so that total
magnitude of load approximately remains constant; The springs are
removed; instead, the cantilever support is replaced (Fig. 7).
Fig. 3. Arch umbrella elements in the entrance section of tunnel to very weak
rock and in the following section.
Note that a little difference in the value of moment at cantilever
support due to assumption of smaller length is neglected.
Peq
2.2. The stiffness of a single umbrella element

The maximum displacement of an umbrella element i.e. at end of

free length, is found by super position principle. That is, an umbrella
Uul-re
LDP element can be considered as two beams (Fig. 8)
The load in the buried length is
r
L' buried−x
q(x) = Pi
L' buried (6)

Therefore, the rotation and displacement can be obtained by Eqs.

(7) and (8), respectively
x
3
⎡ x3 x4 L' ⎤ 1
Fig. 4. Schematic distribution of load on a single umbrella element. θ (x) = ⎢ Pi− '
Pi− buried Pi ⎥
6 24L buried 8 EI
⎣ ⎦ (7)

by 3 4
x ⎡ 4x 5 xL'buried 11L'buried ⎤ Pi
q(x) =
d4 (w(x ')) w(x) = ∫ θ (x)dx = ⎢ 24 −
5 × 24L'
−
8
+
5 × 24 EI
⎥
4 buried (8)
⎣ ⎦
dx ' (3)
or where EI is the umbrella element flexural rigidity, L' buried is length of
buried element, Pi is the load of elements, and x is the excavation length
'
x of elements.
q(x) = C3exp ⎜⎛−1.5 ⎞⎟
⎝ re ⎠ (4) The displacement at the end of the buried length due to triangular
distributed load is
where the parameter C3 is the coefficient of the final load, which is
4
induced by tunnel crown displacement. 11L'buried Pi
The schematic distribution of load on a single umbrella element as δ1 =
5 × 24EI (9)
well as longitudinal deformation profile are shown in Fig. 4a
For simulation, the load distribution is considered on a single um- Similarly, the displacements due to concentrated moment and shear
brella element as shown in Fig. 5, in which Lfree indicates the length of force are found by Eqs. (10) and (11), respectively

Fig. 5. Load distribution acting on a single umbrella

element.

42
M. Ranjbarnia et al. Tunnelling and Underground Space Technology 82 (2018) 39–49

Fig. 6. The boundary condition of a single umbrella

element.

Fig. 7. Simplified modeling of a single umbrella element.

Fig. 8. Replacement of the umbrella element by two equivalent beams.

In the free length, the displacement due to the uniform distributed

load is

Pi L 4free
δ6 =
8EI (14)

Finally, the total displacement at the end of the free length can be
calculated with summing all displacement

δt
4 3 2
Pi (11L'buried + 55L'buried L free + 90L 2freeL'buried + 60L3freeL' buried + 15L 4free)
Fig. 9. Schematic of each elements load area. =
(24 × 5)EI
3 (15)
(Pi L free)L'buried
δ2 = Therefore, total stiffness of the element is
3EI (10)
2 ⎡ (24 × 5)EI ⎤
Pi L 2freeL'buried k=⎢ ⎥
δ3 = '4 '3 2L '
2
3L ' 4
4EI (11) ⎣ 11L buried + 55L buried x + 90x buried + 60x buried + 15x ⎦ (16)
On the other hand, the rotation of buried length causes the dis- As L free is not constant at each step of excavation, it is replaced by
placement of free length. In this case, for the concentrated moment and variable x in Eq. (16).
shear forces, the corresponding displacements at the free length re- To include the width of load area the stiffness of each element (see
spectively are Fig. 9), Eq. (16) can be rewritten as
2
Pi L 2freeL'buried 1⎡ (24 × 5)EI ⎤
δ4 = θ3 × L f = k= ⎢ ⎥
2EI (12) dt 11L'4 '3 2 '2 3 ' 4
⎣ buried + 55L buried x + 90x L buried + 60x L buried + 15x ⎦ (17)
Pi L3freeL' buried
δ5 = θ4 × L f = In fact, it is assumed that each element contributes its zone influ-
2EI (13)
ence uniformly.

43
M. Ranjbarnia et al. Tunnelling and Underground Space Technology 82 (2018) 39–49

Fig. 10. LDP of supported tunnel by (a) the first row of arch umbrella elements and (b) the following rows of umbrella elements.

Fig. 12. Installation angle of the arch umbrella elements on tunnel crown.

Table 1
Approximate criterion to predict tunnel squeezing problem in the very weak
Fig. 11. Equivalent linear stiffness of arch umbrella elements. rock mass (Hoek and Marinos, 2000).
Class D (very squeezing) E (extreme squeezing)
2.3. The displacement of supported tunnel σcm 0.15 ≪ 0.2 < 0.15
p0

Due to the fact that excavation cause the more length of the um- tunnel closure
∗100 5 ≪ 10 > 10
tunnel diameter
brella elements to be activated (Fig. 1); here, the total excavation length
of the tunnel is divided to n steps where n is arbitrary value

44
M. Ranjbarnia et al. Tunnelling and Underground Space Technology 82 (2018) 39–49

Table 2 dFi = ki (dx ) i = 1, 2, ⋯, n (19)

The results of the proposed approach and FLAC3D simulation for tunnels sup-
where ki is the stiffness of a single umbrella element in each step, and
ported by arch umbrella elements.
can be calculated by substituting of total length of excavation till to ith
step in Eq. (17).
The total load is obtained by summing all loads obtained from the
excavation steps.
n
σcm Class Umbrella elements
δ / r (× 10−2) Ψ
F= ∑ dFi
p0
properties* (20)
i=1
3D 3D
proposed FLAC D (cm) t (cm) proposed FLAC The ultimate displacement of supported tunnel i.e. Uul − sup is calcu-
method method lated by summing all steps’ displacements. The reduction factor (ψ) is
0.2 D 1.26 1.268 30 2 0.89 0.87
defined as
40 2.5 0.67 0.66 Uul − sup
50 3 0.49 0.51 ψ=
Uul (21)
0.17 D 1.76 1.824 30 2 0.85 0.82
40 2.5 0.61 0.60 where parameter ψ define as a reduction factor, and it shows the ratio
50 3 0.44 0.45 of the total displacement of supported tunnel to unsupported dis-
0.15 D 2.272 2.048 30 2 0.83 0.78 placement, which is obtained by the GRC diagrams. This parameter
40 2.5 0.58 0.53 gives LDP of supported tunnel through multiplying it by that of un-
50 3 0.40 0.46 supported tunnel.
0.13 E 3.114 2.7 30 2 0.80 0.75 Since some displacements has been occurred before installation of
40 2.5 0.55 0.54 first row of arch umbrella elements; thus in this case, the parameter ψ
50 3 0.39 0.43
should be applied to the LDP corresponding to front of the face
0.11 E 4.554 5.12 30 2 0.74 0.69 (Fig. 10a). However, for the following rows of arch umbrella elements,
40 2.5 0.50 0.46
the parameter ψ can be applied to total LDP (Fig. 10b) provided that the
50 3 0.34 0.42
main support of tunnel have not been installed.
0.10 E 5.686 6.528 30 2 0.70 0.685
40 2.5 0.47 0.45
50 3 0.31 0.365 2.4. Equivalent linear stiffness of arch umbrella system
0.05 E 36.8 – 30 2 0.37 –
40 2.5 0.22 –
In this section, nonlinear stiffness of arch umbrella system is ap-
50 3 0.15 – proximated by an equivalent constant stiffness in order to provide quick
calculation rather than more complex approach in practical purposes.
* dt = 0.5 m. From Fig. 11, it can be seen that tangential linear Kinitial and Kfinal are
the upper and lower limits of stiffness. Thus, the average value of these
two is assumed as the equivalent stiffness of system i.e.
kinitial + kfinal
keq =
2 (23a)
or

⎡
120EI ⎢ 1
k eq = ⎢
dt ⎢ '4 '3 Lfree Lfree 2 '2 Lfree 3 ' Lfree 4
⎢ 11L buried + 55L buried ⎛ n ⎞ + 90 ⎛ n ⎞ L buried + 60 ⎛ n ⎞ L buried + 15 ⎛ n ⎞
⎣ ⎝ ⎠ ⎝ ⎠ ⎝ ⎠ ⎝ ⎠
⎤
1 ⎥
+ ⎥
'4 '3 '2
11L buried + 55L buried Lfree + 90Lfree L buried + 60Lfree L buried + 15Lfree ⎥
2 3 ' 4
⎥
⎦ (23b)

2.5. Influence of installation angle

Fig. 13. The View of the 3D numerical model for arch umbrella elements sti-
mulations. Model width: 50 m; model height: 50 m; model thickness: 80 m. The In practice, this system is used with inclination from tunnel axis (see
tunnel zone is subject to 30 elements in a radial direction and 30 elements in a Fig. 12).
circumferential direction, 50 elements are foreseen in the axial direction of the It can be rewritten that
tunnel).
L1' = L1cos(θ), L2' = L2cos(θ) (24)
Lfree Replacing Eq. (24) in Eq. (17) gives:
dx = (n: number of steps )
n (18)
1⎡ (24 × 5)EI ⎤ 1
K' = 4 3 2 2 4⎥
where dx is the length of excavation in each step. dt ⎢ 3 4
⎣ 11L 2 + 55L 2 x + 90x L 2 + 60x L 2 + 15L1 ⎦ cos (θ) (25)
The load in each step is obtained by
where θ is the installation angle of umbrella elements.

45
M. Ranjbarnia et al. Tunnelling and Underground Space Technology 82 (2018) 39–49

Class D (ıcm/p0=0.17) , dt = 0.4 m Class E (ıcm/p0=0.11) , dt = 0.4 m

Class D (ıcm/p0=0.17) , dt = 0.5 m Class E (ıcm/p0=0.11) , dt = 0.5 m
Class D (ıcm/p0=0.17) , dt = 0.6 m Class E (ıcm/p0=0.11) , dt = 0.6 m

1
0.9
0.8
0.7
0.6
Ȍ

0.5
0.4
0.3
0.2
0.1
0
0 20000 40000 60000 80000 100000 120000 140000 160000
Moment of Inertia, I (cm^4)
Fig. 14. Influence of each umbrella element’s stiffness on controlling of tunnel convergence.

3. Discussion As observed in Fig. 14, by increasing the umbrella element’s stiff-

ness, the convergence of squeezing tunnels are significantly reduced.
According to Hoek and Marinos (2000) studies, the stability of the However, the rate of this reduction dwindles at great stiffness.
tunnels in weak rock is controlled by the ratio of uniaxial compressive As well, umbrella elements can best represent their efficiency in
strength of rock mass to the maximum in situ stress i.e. σcm . Based on very severe squeezing tunnels where the unstable rock mass displace-
p0
considerable experience, they predicted very severe squeezing problem ment needs to be much controlled. That is, the performance of this
when σcm is less than 0.2, and proposed arch umbrella elements to be supporting system is much more highlighted as encountered to very bad
p0
ground conditions. However, it seems that the performance of different
installed prior to main support. Two types of classes for this case is
umbrella elements in various weak ground conditions follows the same
shown in Table 1.
GSI − 100 trends. It can be due to the simplifying assumptions made in developing
Note that σcm = s σci where s = exp 9 − 3D and σci is the uniaxial
the theoretical model, and discussed in the introduction part.
strength of intact rock mass, and GSI is the Geological Strength Index of
In the following, Fig. 15 show the corresponding LDP of some for-
rock mass, and D is the disturbance factor (Hoek et al., 2002).
going analyses.
In Table 2, the foregoing classes of weak rock mass are discussed in
detail when accompanied with different combinations of arch umbrella
elements. For the circular tunnels excavated in different weak rock 4.2. The influence of umbrella elements density on ψ ratio
mass (i. e . GSI < 30 ) at several depths, the calculations were performed
by the proposed approach for three types of umbrella elements to ob- The number of umbrella elements in unit width of circumferential
tain the tunnel wall displacements; and as a result, the coefficient ψ was direction of tunnels has the inverse relation to the stiffness of this
computed to obtain LDP of supported tunnels. The results of simulation system (see Eq. (16)). To explore and to quantify its effect in confining
by FLAC3D (see Fig. 13) was also reported which relatively are in good displacements, it was changed within practical limits to evaluate the
agreement with those of theoretical model. result.
For the cases in very weak rock masses in great depths i.e. As observed in Fig. 16, unlike a single umbrella element properties,
σcm
p
< 0.10 , the unsupported tunnel is unstable and thus, it is not pos- this parameter is not considerably effective. That is, reducing the width
0
sible to find displacement nor ψ by FLAC3D. of loading area from 0.6 m to 0.4 m leads the tunnel displacement to be
confined in the maximum value of 25%. Furthermore, the number of
elements’ effect is magnified when the stiffer elements are employed at
4. Case study and design charts
extreme squeezing condition. However, it should be reminded that the
umbrella elements density has the most effective role in prevention of
4.1. Effect of a single umbrella element’s stiffness on ψ ratio
ground arching between the elements which is not the case of discus-
sion here.
The stiffness of each umbrella element is controlled by its diameter
and thickness which are involved in moment of inertia in Eq. (17)
π
(I = 8 D3t ). To gain insight into influence of stiffness on controlling of 4.3. The influence umbrella elements’ inclination angle on ψ ratio
tunnel displacement, some calculations are performed in two classes of
squeezing tunnels. Three practical values are assumed for the element installation

46
M. Ranjbarnia et al. Tunnelling and Underground Space Technology 82 (2018) 39–49

a) Class D (ıcm/P0 = 0.17)

Unsupported Supported (D = 30 cm, t = 2 cm) Supported (D = 50 cm, t = 3 cm)
0.07

0.06

0.05

0.04
U (m)

0.03

0.02

0.01

0
0 5 10 15 20 25 30 35 40 45 50

X (m)

b) Class E (ıcm/P0=0.11)
Unsupported Supported (D = 30 cm, t = 2 cm) Supported (D = 50 cm, t = 3 cm)
0.18
0.16
0.14
0.12
U (m)

0.1
0.08
0.06
0.04
0.02
0
0 10 20 30 40 50 60
x (m)
σcm σcm
Fig. 15. Longitudinal deformation profile of unsupported tunnels and of supported tunnels by the umbrella elements for (a) p0
= 0.17 and (b) p0
= 0.11.

angle with respect to the horizon to investigate the weight of this These graphs can be used in order to develop a fast pre-designing of
parameter. As observed in Fig. 17, the effect of this parameter is neg- umbrella elements, leaving the numerical modelling only to check the
ligible but can be considered as a safety factor if the calculations are defined geometrical configuration.
carried out for horizontal elements. As an Example from Hoek and Marinos (2000), a 4.7 km long tunnel
with cover depths of up to 220 m is considered. This tunnel passes
5. Application examples through a series of typical flysch rocks where the rock between 2.15 km
and 2.75 km is predominantly siltstone flysch and its low strength by
The proposed formulation of Eq. (17) was used in order to GSI = 20 , σci = 20MPa and mi = 8(intact rock constant) combined with
evaluate arch umbrella elements system for different mechanical the relatively high cover of maximum 220 m, results in strains of up to
and geometrical configurations that can be met. In Fig. 18, the about 40%. This value can be computed by the standard analytical
parameter ψ can be obtained for typical values of GSI and approaches assuming the elastic-perfectly plastic stress–strain behavior
oftheweakrockmassinconventionaldepths (thedisturbancefactorDwasconsideredequaltozero for the rock mass, or the simple formulation proposed by Hoek and
, typicalvalueforundisturbedrockmass ). Thesegraphscanbeusedinordertodevelopafastpre Marinos (2000) i.e.ε = 0.2(σcm/ p0 )−2 .
−designingofumbrellaelements, leavingthenumericalmodellingonlytocheckthedefinedgeome If the arch umbrella systems with properties written in Table 3 are
tricalconfiguration . employed, the coefficients ψ of each case can be obtained using Fig. 18.
of the weak rock mass in conventional depths (the disturbance factor D The supported tunnel strain can be calculated by Eq. (21).
was considered equal to zero, typical value for undisturbed rock mass). Note that for the cases indicated by (*) in Table 3, the coefficient ψ

47
M. Ranjbarnia et al. Tunnelling and Underground Space Technology 82 (2018) 39–49

Class D (ıcm/p0=0.17) , D = 30 Cm, t=2 Cm Class E (ıcm/p0=0.11) , D = 30 Cm, t=2 Cm

Class D (ıcm/p0=0.17) , D = 40 Cm, t=2.5 Cm Class E (ıcm/p0=0.11) , D = 40 Cm, t=2.5 Cm
Class D (ıcm/p0=0.17) , D = 50 Cm, t=3 Cm Class E (ıcm/p0=0.11) , D = 50 Cm, t=3 Cm
1
0.9
0.8
0.7
0.6
Ȍ

0.5
0.4
0.3
0.2
0.1
0
0.35 0.4 0.45 0.5 0.55 0.6 0.65

dt (m)
Fig. 16. Influence of umbrella elements density on controlling of tunnel convergence.

Class D (ıcm/p0=0.17) , D = 30 Cm, t=2 Cm Class E (ıcm/p0=0.11) , D = 30 Cm, t=2 Cm

Class D (ıcm/p0=0.17) , D = 40 Cm, t=2.5 Cm Class E (ıcm/p0=0.11) , D = 40 Cm, t=2.5 Cm

Class D (ıcm/p0=0.17) , D = 50 Cm, t=3 Cm Class E (ıcm/p0=0.11) , D = 50 Cm, t=3 Cm

1
0.9
0.8
0.7
0.6
Ȍ

0.5
0.4
0.3
0.2
0.1
0
0 2 4 6 8 10 12 14 16
T (degree)
Fig. 17. Influence of umbrella elements’ inclination angle on controlling of tunnel convergence.

is found by interpolating of two adjacent known values. For instance of total displacement of supported tunnel to that of unsupported (i.e.
Uul − sup/ Uul ) in terms of ratio of uniaxial strength of rock mass to in-situ
6. Summary and conclusion stress (i.e.σcm/ p0 ) and the various arch umbrella systems. The results are
in good agreement to those obtained by FLAC 3D.
A simple analytical approach was presented for the computation of Sensitive analyses show that the flexural stiffness of umbrella ele-
tunnels displacement which had been supported by the arch umbrella ments i.e. the diameter of a single umbrella element is more effective
system, and was excavated in very and extreme squeezing ground than the adjacent distance between elements to control tunnel dis-
conditions. For this purpose, some graphs are proposed to give the ratio placement.

48
M. Ranjbarnia et al. Tunnelling and Underground Space Technology 82 (2018) 39–49

I=21200 (cm^4) dt=0.5m I=62800 (cm^4) dt=0.5m I=147200 (cm^4) dt=0.5m

I=21200 (cm^4) dt=0.4m I=62800 (cm^4) dt=0.4m I=147200 (cm^4) dt=0.4m
I=21200 (cm^4) dt=0.6m I=62800 (cm^4) dt=0.6m I=147200 (cm^4) dt=0.6m
1
0.9
0.8
0.7
0.6
Ȍ

0.5
0.4
0.3
0.2
0.1
0
0 0.05 0.1 0.15 0.2 0.25

ıcm/p0
Fig. 18. The performance of each type of umbrella elements in various squeezing conditions.

Table 3 umbrella arch systems. J. Rock Mech. Geotech. Eng. 6 (6), 546–564.
The properties of arch umbrella systems and the output results. Oke, J., Vlachopoulos, N., Marinos, V., 2014b. Umbrella arch nomenclature and selection
methodology for temporary support systems for the design and construction of tun-
Umbrella element system properties Ψ u/ r (supported) nels. Geotech. Geol. Eng. 32 (1), 97–130.
Oke, J., Vlachopoulos, N., Diederichs, M.S., 2016. Semi-analytical model for umbrella
D (cm) t (cm) dt (m) 4
I (cm ) arch systems employed in squeezing ground conditions. Tunn. Undergr. Space
Technol. 56, 136–156.
30 2 0.6 21,200 0.54 21.6 Pelizza, S., Peila, D., 1993. Soil and rock reinforcements in tunnelling. Tunn. Undergr.
35 2 0.5 33,700 0.44* 17.6 Space Technol. 8 (3), 357–372.
Peila, D., Oreste, P.P., 1998. A new theory for steel pipe umbrella design in tunnelling. In:
50 2.5 0.4 123,000 0.22* 8.8
Tunnels and Metropolises. International Tunnelling Association, pp. 1033–1040.
Peila, D., 1994. A theoretical study of reinforcement influence on the stability of a tunnel
face. Geotech. Geol. Eng. 12 (3), 145–168.
References Peila, D., Pelizza, S., 2003. Ground reinforcing and steel pipe umbrella system in tun-
nelling. In: Kolymbas, D. (Ed.), Rational Tunneling: Advances in Geotechnical
Engineering and Tunnelling. Logos Verlag, Berlin, Innsbruck, pp. 93–132.
Aksoy, C.O., Onargan, T., 2010. The role of umbrella arch and face bolt as deformation Peila, D., 2013. Forepoling design (Lecture). Ground improvement pre support & re-
preventing support system in preventing building damages. Tunn. Undergr. Space inforcement short course. International Tunnelling and Underground Space
Technol. 25 (5), 553–559. Association, Geneva (WTC 2013).
Galetto, D., Silva, J.C.B.J., Peila, D., Assis, A., 2013. Scale Laboratory model for studying Song, H.W., Wang, H.T., 2011. Analytical approach for face stability estimate of tunnel
the behavior of pipe umbrella in sandy soil. In: SOILS & ROCKS, Brazilian Association with pipe roof reinforcement. In: Advanced Materials Research, vol. 243. Trans Tech
for Soil Mechanics and Geotechnical Engineering (ABMS) and Portuguese Publications, pp. 347–350.
Geotechnical Society (SPG), San Paolo (BR), vol. 36, no. 2, pp. 231–236. ISSN: 1980- Song, K.I., Cho, G.C., Chang, S.B., Lee, I.M., 2013. Beam–spring structural analysis for the
9743. design of a tunnel pre-reinforcement support system. Int. J. Rock Mech. Min. Sci. 59,
Gao, F., Zhang, K., Hou, A., Guo, J., 2015. Deformation and internal force analysis of pipe 139–150.
roof support using the beam on elastic foundation theory. Electron. J. Geotech. Eng. Vlachopoulos, N., Diederichs, M.S., 2009. Improved longitudinal displacement profiles
20, 13013–13022. for convergence confinement analysis of deep tunnels. Rock Mech. Rock Eng. 42 (2),
Heidari, M., Tonon, F., 2015. Ground reaction curve for tunnels with jet grouting um- 131–146.
brellas considering jet grouting hardening. Int. J. Rock Mech. Min. Sci. 76, 200–208. Volkmann, G., Schubert, W., 2006. Optimization of excavation and support in pipe roof
Hoek, E., 1999. Support for very weak rock associated with faults and shear zones. Proc. supported tunnel sections. Tunn. Undergr. Space Technol. 21 (3–4), 404.
Rock Support Reinforcement Pract. Mining 19–32. Volkmann, G., Button, E., Schubert, W., 2006. A contribution to the design of tunnels
Hoek, E., Marinos, P., 2000. Predicting tunnel squeezing problems in weak heterogeneous supported by a pipe roof. In: Golden Rocks 2006, The 41st US Symposium on Rock
rock masses. Tunn. Tunn. Int. 32 (11), 45–51. Mechanics (USRMS). American Rock Mechanics Association (January).
Hoek, E., Carranza-Torres, C., Corkum, B., 2002. Hoek-Brown failure criterion - 2002 Ed. Volkmann, G.M., Schubert, W., 2007. Geotechnical model for pipe roof supports in tun-
In: Proceedings of the NARMS-TAC, Mining Innovation and Technology, Toronto, neling. In: Proc. of the 33rd ITA-AITES World Tunneling Congress, Underground
Canada. Space–the 4th Dimension of Metropolises, vol. 1. pp. 755–760 (May).
Marchino, C., Borio, L., Peila, D., 2010. Analisi mediante modellazione numerica e Volkmann, G.M., Schubert, W., 2010. A load and load transfer model for pipe umbrella
modelli analitici di un presostegno con infilaggi in gallerie superficiali. GEAM. support. Rock Mech. Civ. Environ. Eng. 379–382.
Geoingegneria Ambientale e Mineraria 47 (3), 69–74. Wang, H., Jia, J., 2008. Analytical method for mechanical behaviors of pipe roof re-
Ocak, I., 2008. Control of surface settlements with umbrella arch method in second stage inforcement. In: International Conference on Information Management, Innovation
excavations of Istanbul Metro. Tunn. Undergr. Space Technol. 23 (6), 674–681. Management and Industrial Engineering, 2008. ICIII'08, vol. 3. IEEE, pp. 352–357
Oke, J., Vlachopoulos, N., Diederichs, M.S., 2012. Improved input parameters and nu- (December).
merical analysis techniques for temporary support of underground excavations in Wang, H., Jia, J., 2009. Face stability analysis of tunnel with pipe roof reinforcement
weak rock. In: Proceedings of the 21st Canadian Rock Mechanics Symposium based on limit analysis. Electron. J. Geotech. Eng 14.Bundle G: 15.
RockEng12 (May). Zhang, Z., Li, H., Liu, H., Li, G., Shi, X., 2014. Load transferring mechanism of pipe
Oke, J., Vlachopoulos, N., Marinos, V., 2013. The pre-support nomenclature and support umbrella support in shallow-buried tunnels. Tunn. Undergr. Space Technol. 43,
selection methodology for temporary support systems within weak rock masses. 213–221.
Geotech. Geol. Eng.
Oke, J., Vlachopoulos, N., Diederichs, M.S., 2014a. Numerical analyses in the design of

49