Numerical solutions for tunnels excavated in strain-softening rock masses

considering a combined support system

Fei Song1, Alfonso Rodriguez-Dono1,2*

1 Department of Civil and Environmental Engineering,

Technical University of Catalonia (UPC), Spain, 08034

2 Institute of Environmental Assessment and Water Research (IDAEA),

Spanish National Research Council (CSIC), Spain, 08034

* Corresponding author: Alfonso Rodriguez-Dono

E-mail address: [email protected]

1
Abstract

In this article, an alternative numerical procedure to calculate displacements and stresses of supported

circular tunnels is proposed, considering the whole process of tunnel advancement, and sequential

installation of the primary and secondary support systems. In the derivation, the plastic area of the rock

mass is divided into a large number of annulus around the tunnel, and then the Finite Difference Method

is employed. First, the strain-softening behaviour model is taken to simulate the post-failure behaviour of

the rock mass. Furthermore, the Mohr-Coulomb or the Hoek-Brown failure criteria can be chosen, a non-

associated plastic flow rule is assumed and the dilatancy of the rock mass is considered. After that, the

fictitious support forces concept is used to simulate the process of tunnel advancement, and thus, the

three-dimensional effect of the tunnel face is considered. Finally, the solutions of displacements and

stresses for the rock mass and the supports can be obtained, by using the compatibility conditions of

stresses and displacements at both rock-support and support-support interfaces. The results obtained from

these solutions agree well with those of the self-similar solutions for circular openings, and the

compatibility conditions of supported tunnels were verified. The proposed method has been compared

with the convergence-confinement method. Parametric analyses are then carried out to investigate the

sensitivity of support forces and displacements to the rock mass behaviour model selection. Then, the

application of the proposed solutions in the design of tunnels is presented. The proposed method provides

a convenient alternative method for the preliminary design of tunnels.

Keywords:

Strain-softening, Supported tunnels, Tunnel advancement, Numerical solutions, Ground reaction curves,

Combined support system

2
1 Introduction

The tunnels supported with a system comprising several supports are in high demand due to the

increasing demands of transportation, geological disposal, hydraulic and mining engineering. The primary

support system is usually put in place to seal the rock and withstand the loads that may arise during the

excavation, while the secondary support system is installed to ensure the long-term stability of tunnels [1-

3]. Thus, a reliable analysis of the secondary support system constitutes a critical topic to the long-term

stability of tunnels. At the same time, the acting support forces at the support-support interface directly

affect the stability of the secondary support system.

In addition, many accidents are related to fractures when dealing with mining or civil engineering

underground projects. Therefore, the complexity of the behaviour of the different geomaterials is a critical

topic in the mitigation of the accidents that may occur in works such as tunnel excavation [4, 5]. Thus,

both rock mass behaviour model selection and the design method of the support system are crucial topics

in the design of tunnels. A proper simulation of the whole process of tunnelling and support installation is

critical to achieve an optimal design of tunnels.

The design of tunnels can be tackled by various different methods, such as analytical solutions, numerical

simulations or the convergence-confinement method (CCM) [2-4, 6-51]. Numerical simulations have

been widely used in the analysis of underground projects, considering more complex geological

conditions. Although numerical simulations can provide some useful results, usually they require long

runtimes and sometimes meet numerical difficulties, especially when complete parametric analyses need

to be performed [3]. On the other hand, analytical solutions provide an efficient and quick approach to

gain insights into the nature of the problem [11]. However, solutions for plastic problems are not easy to

develop. The CCM provides an efficient way for determining support forces by considering the rock-liner

interactions [6, 8]. Although the CCM has been widely used in the design of tunnels, the application of

the CCM including two different support systems is limited [3]. In addition, it is not possible to use the

3
CCM in the design of a tunnel support system if ground reaction curves, support characteristic curves or

longitudinal deformation profiles are not available [30]. As an alternative, a simple stepwise approach has

been adopted to obtain solutions for tunnels excavated in strain-softening rock masses in this study,

considering sequential installation of two different support systems, as explained below.

Moreover, revising the literature so far on existing solutions for supported tunnels, some researchers

presented elastic solutions for their stresses and displacements [13, 44, 52]. Some other researchers

presented solutions for supported tunnels constructed in time-dependent viscoelastic rock masses [3, 31,

35, 38]. However, no plastic behaviour is considered in these references, ergo the solutions obtained may

lead to an unsafe design of tunnels, since plastic behaviour is common in many real engineering projects.

On the other hand, many researchers developed so far various approaches for the plastic problems of

tunnelling in rock masses, but mostly for elastic-perfectly-plastic (EPP) or elastic-brittle (EB) behaviour

of rock masses [48, 49, 53-55]. Nonetheless, the average quality rock masses behave in a strain-softening

(SS) manner and thus, it is important to study the response of tunnels in rock masses exhibiting SS

behaviour [6-8, 10]. Note that both EPP and EB models are special cases of the SS model.

Regarding rock masses exhibiting strain-softening behaviour, through defining a fictitious ‘time’ variable

and re-scaling some variables, Carranza-Torres [56] and Alonso et al. [10] presented self-similarity

numerical solutions of circular openings excavated in strain-softening geomaterials. Later, based on those

solutions in Alonso et al. [10], Alejano et al. [7] presented ground reaction curves considering EPP, EB

and SS behaviour models and increasingly realistic parameters.

Furthermore, by using a simple stepwise procedure that successively determines the stresses and strains

on the boundaries of a number of annulus into which the plastic zone is divided, some researchers [49, 51,

57-62] presented numerical solutions for tunnelling in rock masses that exhibit strain-softening behaviour.

Combining them with the CCM, the above plastic solutions can be used for designing supported tunnels.

However, the application of the CCM is limited to tunnels with a single support [3].

4
On the other hand, Oreste [2] proposed support characteristic curves for a combined support system,

which extended the CCM in the application of tunnels with several different supports [2, 7, 9, 63].

However, in order to make a correct design of the secondary support system, the modified ground reaction

curve due to the influence of the primary support system is needed; a research that was missing until now.

To do that, the compatibility conditions at the interface between supports should be considered.

Therefore, at this point the CCM meets one of its limitations.

In summary, the solutions of supported tunnels were mostly concerned with elastic or viscoelastic

problems, and rarely geomaterial-support/support-support coupled interaction is considered in the plastic

solutions. Instead, in this research, we have proposed solutions for tunnels with two different support

systems constructed in rock masses showing strain-softening behaviour. In these solutions, we consider:

(1) the tunnel advancement, (2) the sequential installation of primary and secondary support systems, (3)

different rock mass post-failure behaviour models: elastic-perfectly-plastic (EPP), strain-softening (SS)

and elastic-brittle (EB), and (4) both displacements and stress compatibility conditions on the rock-

support and support-support interfaces.

In our approach for unsupported circular tunnels, the numerical solutions for the stresses and

displacements are first derived for the strain-softening rock masses. Then, based on the coupled model at

both rock-support and support-support interfaces, the solutions for the supported circular tunnels are

presented, considering the tunnel advancement and different installation time of the primary and

secondary support systems. Finally, the sensitivity of support forces and deformation of tunnels to the

rock mass behaviour model selection are analysed, and an application example is presented to illustrate

the application of the proposed method in the design of tunnels.

2 Statement of the problem

2.1 Assumptions and description of the problem

5
This research considers the design of tunnels with two different support systems constructed in rock

masses showing strain-softening (SS) post-failure behaviour. The process of tunnel advancement and the

sequential installation of primary and secondary support systems have been considered. Throughout the

analyses, the following assumptions have been made:

(1) The surrounding rock is homogeneous and isotropic, and exhibits an elastic-plastic mechanical

behaviour. The strain-softening (SS), the elastic-perfectly-plastic (EPP) as well as the elastic-

brittle (EB) models are adopted to simulate the post-failure behaviour of rock masses.

(2) The initial isotropic stress field around the tunnel is idealized as an axisymmetric condition, and

the stresses in the far field can be assumed as P0.

(3) Both the primary and secondary support systems are assumed to behave linearly elastic. The

primary and secondary supports are instantaneously installed at a distance from the tunnel face xd

= x1 and xd = x2, respectively. The outer and inner radii of the primary (secondary) support system

are R1 (R2) and R2 (R3), respectively, and thus, the thickness of the primary (secondary) support is

d1 = R1 - R2, d2 = R2 - R3, as shown in Figures 1 and 2.

(4) The excavation is slow enough that it may be assumed that it does not induce any dynamic stress.

Small strain theory is considered.

Considering the tunnel advancement, the analysis of the distribution of the stresses and displacements

near the periphery of the tunnel is a three-dimensional (3D) problem. As shown in Figure 1, the support

forces from the tunnel face (named fictitious support forces in this article), Pf, progressively decrease

along the increased distance to the tunnel face, from the initial stress P0 (ahead of point A), to zero

(behind point D) when the tunnel face is at a distance where it has no influence on the considered section.

Moreover, after the installation of the primary support system (behind point B), both the fictitious support

forces Pf and the interaction support forces between the rock mass and the primary support system PS1 j (j

6
= 1, 2 represents the primary or secondary support stage, respectively) act on the periphery of the tunnel.

At last, with the advancement of the tunnel face, Pf will eventually decrease to zero while PS1 j will

increase from zero to a constant value. Before the installation of the secondary support (between point B

and C), only the support forces PS11 act on the outer boundary of the primary support; after the installation

of the secondary support (behind point C), PS12 and PS22 act on the outer and inner boundaries of the

primary support, while PS22 acts on the outer boundary of the secondary support (Figures 1 and 2).

By introducing the fictitious support forces Pf, to account for the effect of the tunnel face, the 3D problem

can be treated as an equivalent plane-strain one [13, 64], as shown in Figure 2. This infinite plane,

subjected to uniform isotropic far field stresses and made of an elastic-plastic medium, contains a circular

tunnel with primary and secondary support systems sequentially installed. Figure 2 presents the

mechanical models of the tunnel and the support system. Rp and Rs represent the peak and residual plastic

radii of the rock mass, respectively. Polar coordinates (r, ) will be employed in the derivation of the

solutions. The sign convention is defined as positive for compression and negative for tension.

2.2 Rock mass behaviour models

Some researchers [7, 65] provided post-peak strength guidelines. These guidelines are based on the

geotechnical quality of the rock by means of the geotechnical strength index (GSI). According to these

guidelines, (1) for high quality rock masses (GSI > 60), the rock mass behaves in an EB manner; (2) for

average quality rock masses (40 < GSI < 60), the rock mass presents a SS behaviour, which can also be

referred to as strength-weakening behaviour; (3) for weak rock masses (GSI < 40), EPP behaviour is

assumed (Figure 3). Note that these ranges are just a reference and they may vary in some cases.

SS behaviour can accommodate EB behaviour (the SS model with a drop modulus, M, equal to infinite)

and EPP behaviour (the SS model with a drop modulus, M, equal to zero). Thus, EPP and EB behaviours

7
are just two particular cases of SS behaviour. In this article, the SS model will be adopted to simulate the

post-failure behaviour of rock masses.

Figure 1. Three-dimensional effect of the tunnel advancement, including the boundary conditions in
the rock mass and in the primary and secondary support systems.

8
Figure 2. Geometry and boundary conditions for: (a) rock mass and support systems; (b) rock mass;
(c) the primary support system; (d) the secondary support system.

Figure 3. Different post-failure behaviour of rock masses with different geological strength indexes
(GSI). Based on the work of Alejano et al. [6, 7].

One of the main features of the SS behaviour model is that the yield surfaces depend not only on the

stress tensor , but also on the plastic or softening parameter  [6, 7], as shown in Figure 4. The failure

criterion is defined as shown in Eq. (1).

9
 F (σ, )  0 (1)

The SS behaviour model is characterized by a gradual transition from the peak failure surface to the

residual failure surface, which is governed by the plastic or softening parameter . In a SS model, a

softening regime occurs whenever 0 <  < *, and the residual regime takes place when  ≥ *, as shown

in Figure 4. * is the value of the softening parameter whereon the softening phase ends, beginning the
p p
residual phase. The softening parameter  is defined as shown in Eq. (2), where  ( ) represents the

accumulated plastic strain [7, 10].

  p   rp (2)

(a) (b)

Figure 4. (a) Theoretical strain-softening model for a triaxial test performed on a rock sample.
(b) Yield surface evolution for a rock mass with strain-softening behaviour. Based on the
work of Alejano et al. [6, 7] and Song et al. [4].

In this study, a linear decrease function of the plastic parameters k() is adopted to represent the strain-

softening behaviour [7, 10], as shown in Eq. (3), where kp and ks are the peak and residual values of k,

respectively. If a Mohr-Coulomb strain-softening model is adopted, k represents cohesion c, friction angle

10
 and dilatancy angle , while k represents the mHB, sHB, aHB and  when considering the Hoek-Brown

strain-softening model.

 kp , for   0

  ks  kp 
k     kp    , for 0     * (3)
 *
  

ks , for    *

The Mohr-Coulomb strain-softening model can be shown in Eq. (4) [10].

FMC     N    r  S   (4)

where N   = 1  sin     1  sin     , S   2 c   N   . The peak and residual

Mohr-Coulomb failure criteria can be expressed as in Eqs. (5) and (6), respectively.

FMC,p     N p  r  Sp (5)

FMC,s     N s  r  Ss (6)

where N i = 1  sin  i  1  sin  i  , Si  2ci N i , and i = p (or s) represent the peak (or residual)

parameters in the Mohr-Coulomb strain-softening model.

The Hoek-Brown strain-softening model is represented by Eq. (7), where m HB(), sHB() and a HB() are

the plastic strain-dependent parameters in the Hoek-Brown strain-softening model [66].

a HB ( )
  
FHB      r   ci  mHB   r  sHB    (7)
  ci 

GSI-100 GSI-100
1 1   GSI 
20

where mHB  mi e 28-14D , sHB  e 9-3D
and aHB    e 15  e 3  ; D is disturbance factor for rock
2 6 

masses. From an engineering point of view, the tunnel deformations obtained using the Hoek-Brown

model and the Mohr-Coulomb model are similar [4, 8] and thus, we think it is reasonable to use the

equivalent Mohr-coulomb model to represent the tunnel deformation of the Hoek-Brown model. Hence,

11
in this article, if a Hoek-Brown strain-softening model is chosen, the equivalent cohesion and friction

angle for each rock mass and stress range will be determined through fitting process. Then, the equivalent

Mohr-Coulomb parameters will be used in the process of derivation. Based on the work of Hoek et al.

[66], the equivalent cohesion and friction angle can be obtained from Eqs. (8) and (9).

  
 '  
6aHB mHB  sHB  mHB 3max 
   ci  
  sin 1  aHB 1 
(8)
   3max
'
 
 
2 1  a HB  2  aHB   6 a m s
HB HB  HB  mHB
 ci 
 
  

aHB 1
  3max
'
  3max
'

 ci 1  2aHB  sHB  1  aHB  mHB   sHB  mHB 
  ci    ci 
c (9)
   '

aHB 1 

1  a  2  a 
HB HB 1  6aHB mHB  sHB  mHB
 
3max
 

1  aHB  2  aHB  
 
  ci 


where  3max
'
is the upper limit of the confining stress over which the relationship between the Hoek-

Brown and the Mohr-Coulomb criteria is considered. The equivalent method has been adapted from Hoek

et al. [66]. The reader is referred to the original source for a detailed description of the equivalent method

between the Hoek-Brown and Mohr-Coulomb failure criteria.

The plastic potential in the form of Eq. (10) has been adopted, and () is the plastic strain-dependent

dilatancy angle.

G     K    r (10)

where K   = 1  sin     1  sin     .

3 Numerical solutions for supported tunnels

In this work, the process of construction of the supported circular tunnel is divided into 3 stages:

12
(1) The unsupported circular tunnel stage: which extends from xd = 0 to xd = x1, with x1 being the distance

to the tunnel face where the primary support system is installed;

(2) The primary support stage: which extends from xd = x1 to xd = x2, where x2 represents the distance to

the tunnel face where the secondary support system is installed;

(3) The secondary support stage: which extends from xd = x2 onwards, where the tunnel is supported by

the primary and secondary support systems.

Both displacements and stress compatibility conditions on the interface between the rock mass and the

primary support system (rock-support interface), and the interface between the primary and secondary

support systems (support-support interface) should be satisfied. Section 3.1 introduces the basic theory of

the Finite Difference Method (FDM) used in the current research for solving the strain-softening

problems. Section 3.2 presents the method of determination of the fictitious support forces Pf. Sections

3.3 and 3.4 present the derivation of the solutions for the unsupported tunnel stage and the

primary/secondary support stage, respectively.

3.1 The Finite Difference Method (FDM) for strain-softening problems

The Finite Difference Method (FDM) [58] has been used for deriving the solutions of strain-softening

problems. In the FDM, the plastic zone of the rock mass (including softening and residual area) has been

divided into a set of annuli, and the total number of annuli is assumed as n, as shown in Figure 5. The

outer radius of the first annulus is the radius of plastic zone, i.e. r(0) = Rp, where Rp represents the radius of

the plastic zone; the inner radius of the n-th annulus is the radius of the tunnel wall, i.e. r(n) = R1, where R1

represents the radius of tunnels. The outer and inner radii of the i-th annulus are r(i-1) and r(i), respectively.

In each annulus, the material properties are assumed as isotropic, uniform and unchangeable: the material

13
parameters in the i-th annulus are equal to the one at the outer boundary of the i-th annulus, i.e. ki = k(i-1),

where k represent the material parameters.

Figure 5. The radii of the different n annuli used in the numerical procedure.

3.2 The determination of the fictitious support forces Pf

In this study, the fictitious support forces Pf are adopted to simulate the effect of the tunnel face. The

expression of Pf can be determined by combining the ground reaction curve (GRC) and the longitudinal

deformation profile (LDP), and the method to determine Pf has been introduced in detail in Cui et al. [18].

1j
For unsupported tunnels, no interaction support forces act on the tunnels ( PS equal to 0), so only the

14
fictitious support forces Pf act on the tunnel wall ( = R1). As e.g. shown in Figure 6, at a specific distance

to the tunnel face, xd = xk(i) (i = 1, 2), the radial displacements ur = ur  (i = 1, 2) for unsupported tunnels
i

can be determined by using the LDP; then, combining it with the GRC, the corresponding fictitious

(i)
support forces Pf (i = 1, 2) can be determined.

To determine the fictitious support forces Pf, the LDP and the GRC should be presented firstly. A

numerical procedure to calculate the GRC, considering the strain-softening behaviour model of rock

masses, is presented in the sub-section 3.3 (solutions for unsupported tunnels). The solutions for LDP

found in Vlachopoulos and Diederichs [67] have been adopted in this study. Vlachopoulos and Diederichs

[67] presented the LDP for elastic-plastic rock mass shown in Eq. (11), with respect to the tunnel

geometry, the distance to the tunnel face xd and the maximum plastic radius Rpmax . Initially, the empirical

formula by Vlachopoulos and Diederichs [67] has been developed for rock masses exhibiting EPP

behaviour. Furthermore, by comparison with numerical simulations in FLAC [68, 69], Alejano et al. [8]

and Rodriguez-Dono [70] have proved that the empirical formula of Eq. (11) by Vlachopoulos and

Diederichs [67] can be also used for strain-softening rock masses, although the plastic radius used would

be the strain-softening one. Therefore, it is reasonable to choose the empirical formula of Eq. (11) by

Vlachopoulos and Diederichs [67] in this paper.

 xd 0.15 Rp
max

 1 e R1 , for xd＜0
ur  R1 , xd  
Rock  3
 (11)
Rock
ur,max   0.15 Rpmax  3 xd

 1 

 max
2 Rp
1  1  e
R1
r  R1
e , for xd  0
  3 


Rock
Rock
where ur,max represents the maximum radial displacements and ur represents the radial displacements

of unsupported tunnels at a distance of xd. Figure 7 presents the flow chart for determining the fictitious

support forces.
15
Figure 6. Determining of the support forces from tunnel face Pf, by combining the ground reaction
curves (GRC) and the longitudinal deformation profiles (LDP). Based on the work of Cui
et al. [18].

Figure 7. Flow chart in determining of the support forces Pf, by combining the ground reaction
curves (GRC) and the longitudinal deformation profiles (LDP).

16
3.3 Solutions for the unsupported circular tunnels

3.3.1 Elastic solutions

Based on the theory of elasticity, the stresses and strains in the elastic area can be expressed as in Eqs. (12)

and (13), respectively.



 p   Rp  p0
 re   0  Rp2 
r2 

   (12)
    p  p  
  Rp2 
e

 0 0 

Rp
r2 

 re  1     R 2   R p  p0 
 
p
(13)
E  r 2   
     p0   R 
e

 p 

where r (or  ) and r (or  ) represent elastic radial (or hoop) stresses, and incremental radial (or
e e e e

circumferential) strains, respectively;  R represents radial stresses at the elastic-plastic boundary (r = Rp).
p

Note that the strains are the incremental strains occurred after the excavation of the tunnel and the stresses

are the total stresses.

It should be noted that if the acting force Pinner along the tunnel wall is bigger than the critical support

force Pcritical, the whole region of the circular opening will exhibit elastic behaviour, i.e. no plastic zone is

formed. For the Mohr-Coulomb model, Pcritical can be expressed in Eq. (14) [58].

2 P0  S p
pcritical  (14)
1  Np

3.3.2 Plastic solutions for circular openings

17
The outer boundary of the 1-st annulus is the elastic-plastic interface (see Figure 5). By substituting r = Rp

into Eqs. (12) and (13), the stresses, strains at the outer boundary of the 1-st annulus can be expressed as

in Eqs. (15) and (16).

 r  0    Rp 
   
   (15)
   0  2 p0   R 
   p 

 1   
  Rp 
 p0 

 r  0    
   E
   (16)
   0    1   
    p

p0   R 

 E 

where  R represents the radial stresses along the elastic-plastic interface (r = Rp).
p

Based on the FDM described in the section 3.1, in the 1-st annulus, the material parameters should be

equal to the material parameters at the outer boundary of the 1-st annulus (plastic zone), i.e. the peak

values of plastic parameters. And thus, the properties of the 1-st annulus should satisfy the peak failure

criterion, which can be expressed as in Eq. (17).

F1M C     N  0     r  S  0    (17)

where N  0    = N p  1  sin  p  1  sin  p  , S  0     2 c p N p . Combining the expressions of

stresses (15) and the peak failure criterion (17), the radial stresses in the elastic-plastic interface  R can
p

be expressed as in Eq. (18).

2 P0  S 0   
R = (18)
p
1  N  0  

No plastic strain occurs in the outer boundary of the 1-st annulus, i.e.  rp 0    p 0   0 . The failure

criterion in the i-th annulus, can be expressed as in Eq. (19).

18
 Fi MC   i  Ni 1 ri  Si 1 (19)

where N  i 1    1  sin   i 1    1  sin   i 1    , S i 1    2ci 1   N  i 1   ; c i 1   and

i 1   represents the cohesion and friction angle at the outer boundary of the i-th annulus.

A constant radial stress increment, r , is assumed for each annulus, and thus, the radial stress increment

can be expressed as in Eq. (20).

Pinner   R
 r  p
(20)
n

On the other hand, the radial stresses at the inner boundary of the i-th annulus  ri  , can be expressed by

the radial stresses at the outer boundary ri1 , as shown in Eq. (21).

 r i   ri 1   r (21)

And thus, the hoop stresses are expressed as in Eq. (22).

 i  Ni 1 ri  Si 1 (22)

In the derivation, both the radius r and displacements ur are normalized by dividing by the peak plastic

radius Rp, i.e., r  r Rp and u  u Rp , where x represent the normalized variable of x. And thus, the

normalized peak plastic radius is equal to 1, i.e., rp  r 0  1 . For the axisymmetric problem, the

equilibrium equation in the i-th annulus can be expressed as in Eq. (23).

d ri  ri    i
 0 (23)
dr r

The strain in the i-th annulus can be obtained by geometric relations, as shown in Eq. (24).

 dui
 ri  
 dr
 (24)
 ui
  i = r

19
The strain can be decomposed into an elastic and a plastic parts, as shown in Eq. (25).

 ri   rie   rip 

      (25)
  i   e i  pi 

where the elastic strain can be expressed as in Eq. (26), and the plastic strain satisfy the flow rule as

shown in Eq. (27).

 e 1  2   
 ri   ri  P0    i  P0  
 E  1  
 (26)
 e 1  2   
 ri    i  P0    ri  P0  
 E  1  

 rip  K  i-1pi  0 (27)

where K  i-1    1  sin  i-1    1  sin   i-1    .

   

The boundary conditions can be expressed as in Eq. (28).

 ri   ri-1 , for r  ri-1


 (28)
uri  uri-1 , for r  r i-1


In the i-th annulus, combining the failure criterion equation (19), the equilibrium equation (23), and the

stress boundary conditions equations (28), the stresses can be expressed as in Eq. (29).

 S  
N  i 1 1
 S i 1 
   i 1 r
   i
     
ri
1  N  r
  
r i 1
1  N  i 1  (29)
  i 1   i 1   

  i  N i 1  ri  S i 1
    

Thus, the corresponding elastic strain in the i-th annuli can be obtained by substituting Eq. (29) into Eq.

(26). Combining Eqs. (21) and (29), the inner radius of the i-th annuli can be expressed as in Eq. (30).

20
 1
 Si 1  Ni 1 1
  r  i 1   r  
 1  N  i 1 
r i  r i 1   (30)
S i 1
   
r  i 1
 1  N  i 1 
 

By combining Eqs. (24), (25) and (27), the differential equation of the normalized radial displacements

ui can be expressed as in Eq. (31).

dui u
 K  i 1 i  f i 1 (31)
dr r

where f  i 1   rie  K i 1 e i . Combining it with Eq. (31) and the compatibility condition of displacements

(28), the displacements in the i-th annuli can be expressed as shown in Eq. (32).

K  i 1
 r i 1   K  i 1 r K  i 1
ui  u i 1 
 r 
  r 
r i 1
fi 1 r dr (32)
 

Substituting Eq. (32) into Eq. (24), and combining it with Eqs. (25) and (26), the plastic strain in the i-th

annulus can be expressed as in Eq. (33).

 p ui
  i =    i
e

ri (33)

 p
 ri =  K  i 1  i
p

And thus, the softening parameter at r = ri can be expressed as in Eq. (34).

i   pi   rip (34)

Consequently, the material parameters of the rock mass can be updated by substituting Eq. (34) into Eq.

(3). For the case of i =*, the corresponding radius ri represents the normalized residual plastic radius

rs .

21
Repeating the numerical calculation procedure for n times, the stress on the last annulus r(n) reach the

value of the inner pressure acting on the surface of the circular opening with r = R1 (Pinner=Pf + PS1j ). Then,

the peak plastic radius can be expressed as in Eq. (35).

R1
Rp  (35)
r n 

The radial displacements and the residual plastic radius can be expressed as in Eqs. (36) and (37),

respectively.

ui  ui Rp (36)

Rs  rs Rp (37)

where the normalized displacements ui can be obtained from Eq. (32). Algorithm 1 (Appendix A)

presents the calculation procedure for the elastic-plastic solutions for circular openings in rock masses.

3.4 Solutions for the tunnel with supports

3.4.1 Mechanical analysis of the supports

For supports of the type of shotcrete or concrete rings, the parameters of the supports are: shear moduli

GL1 and GL2, and Poisson’s ratios L1 and L2. According to the theory of elasticity, the radial

displacements and the stresses of the primary support under the boundary conditions (see Figure 2) can be

expressed as in Eqs. (38) and (39), respectively [35].

 R12 R22  1  2 L1 R1 pS  xd 
2 11
 1
 pS  xd  
11
2 2 
 2 2
r , x1  xd  x2
  2GL1r R1  R2  2GL1 R1  R2
ur L1  r , xd    (38)
 12 R1 R2  1  2 L1 1 S  d  2 S  d 
2 12 2 22
 1 2 2 R p x  R p x
  pS  xd   pS22  xd    2 2 
 r , xd  x2
  2GL1r R1  R2  2GL1 R12  R22

22
   R22  R12
 1  2  2 2
pS11  xd  , x1  xd  x2
  r  1R  R
 rL1  r , xd   
2

  R22  R12  R12  R22

S  d  1  2  pS22  xd  , xd  x2
11
  1  2  2 2
p x  2 2
  r  R1  R2  r  R1  R2
 (39)
  R22  R12
 1  2  2 2
pS11  xd  , x1  xd  x2
  r R
 1  R
  L1  r , xd   
2

  R22  R12  R12  R22

S  d  1  2  pS22  xd  , xd  x2
11
  1  2  2 2
p x  2 2
  r  R1  R2  r  R1  R2

where pS1 j represents the support forces acting on the outer boundary of the primary support in the j

support stage, and j = 1, 2 represent the primary and secondary support stage, respectively; PS22 represents

the support forces acting on the outer boundary of the secondary support (r = R 2) in the secondary support

stage. The radial displacements and stresses of the secondary support system under the boundary

conditions in Figure 2 can be expressed as in Eqs. (40) and (41) [35].

R22 R32  1  2 L 2 R2 PS  xd 
2 22
 1
ur L 2  r , xd   P  xd  
S
22
2 2 
 2 2
r , xd  x2 (40)
 2GL 2 r R2  R3  2GL 2 R2  R3

  R32  R22
 rL2  d  1  2  2
 r , x  2 S  d
P 22 x , xd  x2
  r  R2  R3
 (41)
  R32  R22
  L 2  d  1  2  2
 r , x  2 S  d
P 22 x , xd  x2
  r  R2  R3

3.4.2 Determination of the solutions in the primary support stage, x 1 ≤ xd < x2

The primary support was installed at the distance of xd = x1. Only the fictitious support forces Pf (xd) act

on the tunnels before the installation of the primary support system, and the support force acting on the

rock mass along the tunnel face is equal to Pf (x1) at the moment of installation. After the installation of

the primary support system, both Pf (xd) and the interaction support forces PS11 (xd) act on the tunnel wall.
23
Then, with tunnel advancement, the fictitious support forces Pf (xd) decrease, while the interaction support

forces PS11 (xd) increase. Thus, the compatibility conditions of stresses at the rock-support interface (r = R1)

are satisfied. In the derivation, the additional compatibility conditions of displacements should also be

satisfied, as shown in Eq. (42).

urRock  R1 , xd   urRock  R1 , x1   urL1  R1 , xd  , for x1 ≤ xd ≤ x2 (42)

where urRock  R1 , xd  =R1  Rock

 n  , ps  xd  = r  n   pf  xd  . Then, Eq. (42) can be simplified to Eq. (43).
11 Rock

 1 R12 R22 1  2 L R13   Rock

R1  Rock
n  u Rock
r  R1 , x1   
2G r R 2
 R 2

2G R 2
 R

2   r n 
 pf  xd   =0 (43)
 L 1 2 L 1 2 

Below is shown the procedure to obtain the support forces p S11 (xd):

Rock
(1) Assume that both PS11 and Pf are 0, to obtain the maximum displacements of tunnel ur,max and the

maximum peak plastic radius Rpmax of the unsupported tunnels;

Rock
(2) Substitute xd = x1, ur,max and Rpmax into Eq. (11), to obtain the displacements at the moment of

installation of the primary support urRock  R1 , x1  ;

(3) Follow the method presented in section 3.2, to obtain the fictitious support forces Pf (xd);

(4) Find the specific stress and strain at x = xd,  rRock

 n  and  r  n  , that can satisfy Eq. (43).
Rock

The fictitious support forces Pf(xd), and the interaction support forces PS11 xd =  rRock  
 n  - Pf (xd), can be

obtained through the above steps (1)-(4). In the primary support stage, the forces acting on the rock

masses are Pf (xd) + PS11  xd  , while the forces acting on the outer and inner boundaries of the primary

support are PS11 (xd) and 0, respectively. Then, the stresses and displacements of the rock mass can be

obtained through the method in section 3.3; the stress and displacements of the primary support can be

obtained by substituting PS11 (xd) into Eqs. (38) and (39). It should be noted that the Eq. (43) cannot be

24
solved directly, but it can be easily determined by using a numerical calculation method such as the

dichotomy method [71, 72]. Algorithm 2 (Appendix A) presents the calculation procedure for the elastic-

plastic solutions of supported tunnels in the primary support stage.

3.4.3 Determination of the solutions in the secondary support stage, x d ≥ x2

The secondary support system is installed at the distance of xd = x2. At the moment of installation of the

secondary support (xd = x2), the support forces acting on the tunnel wall (r = R1) are assumed to be Pf (x2)

and PS11  x2  , while the support forces acting on the outer and inner boundary of the primary support

system are PS11  x2  and 0, respectively. After the installation of the secondary support (xd ≥ x2), both the

fictitious support forces Pf (xd) and the interaction support forces PS12 (xd) act on the tunnel wall, while

PS12 (xd) and PS22 (xd) act on the outer and inner boundaries of the primary support system, respectively.

Moreover, PS22 (xd) acts on the outer boundary of the secondary support, meanwhile at its inner boundary

the acting force is equal to 0.

With tunnel advancement, the fictitious support forces Pf (xd) decrease to 0, and the interaction support

forces PS12 (xd) and PS22 (xd) increase to the constant values. The compatibility conditions of stresses at

both the rock-support interface (r = R1), and the support-support interface (r = R2), are satisfied. In the

derivation, the additional compatibility conditions of displacements should be satisfied, as shown in Eqs.

(44) and (45).

urRock  R1 , xd   urRock  R1 , x1   urL1  R1 , xd  , xd  x2 (44)

urL1  R2 , xd   urL1  R2 , x2   urL 2  R2 , xd  , xd  x2 (45)

25
  n  , ps  xd  = r  n   pf  xd  . Moreover, ur L1  R1 , xd  , ur L1  R2 , xd  and
where urRock  R1 , xd  =R1  Rock 12 Rock

ur L2  R2 , xd  can be expressed as in Eqs. (46), (47) and (48), respectively.

ur L1  R1 , xd   A pS12  xd   B pS22  xd  (46)

ur L1  R2 , xd   C pS12  xd   D pS22  xd  (47)

ur L 2  R2 , xd   E pS22  xd  (48)

 1 R1 R22 1  2 L1 R13   1 R1 R22 1  2 L1 R1 R22 

where A=  2 2
  , B=   ,
 2GL1 R1  R2 2GL1 R12  R22  2
 2GL1 R1  R2
2
2GL1 R12  R22 

 1 R12 R2 1  2 L1 R12 R2   1 R12 R2 1  2 L1 R23 

C=  2 2
  , D=   ,
 2GL1 R1  R2 2GL1 R12  R22  2
 2GL1 R1  R2
2
2GL1 R12  R22 

 1 R2 R32  1  2 L 2 R23 
E=  2 2 
 2 2
.
  2GL 2 R2  R3  2GL 2 R2  R3 

Substituting Eqs. (46)-(48) into Eqs. (44) and (45), then Eqs. (44) and (45) can be simplified into Eqs. (49)

and (50).

B C  Rock B C PS  x2 
11

R1  Rock
 n u Rock
r  R1 , x1   A  rRock
n  pf  x 
d 
  
D+ E  r  n 
 pf  x 
d 
 
D E
(49)

C  PS12  xd   PS11  x2  
p 22
S x  
d
D E
(50)

Here is the procedure to obtain the support forces PS12 (xd) and PS22 (xd):

(1) PS11  x2  and urRock  R1 , x1  can be obtained from sub-section 3.4.2;

(2) Follow the method presented in section 3.2, to obtain the fictitious support forces Pf (xd);

26
(3) Obtain the specific stress and strain at x = xd (xd ≥ x2), finding  rRock
 n  and r n , that can satisfy Eq.
Rock

(49). Then, the support forces acting on the outer boundary of the primary support PS12  xd  can be

obtained through PS12 (xd) =  rRock

 n   Pf (xd);

(4) Obtain PS22 (xd) by using Eq. (50).

For a specific value of xd, Pf (xd), PS12 (xd) and PS22 (xd) can be obtained through the above steps (1)-(5),

and thus, we can determine the forces acting on the rock mass and on the boundaries of the supports. Then,

the stresses and displacements for the rock mass can be obtained through the method shown in section 3.3.

Moreover, the stresses and displacements of the primary and secondary support systems can be calculated

by substituting PS12 (xd) and PS22 (xd) into Eqs. (38)-(41). It should be noted that Eq. (49) cannot be solved

directly, but it can be easily determined by using a numerical calculation method such as the dichotomy

method [71, 72]. Algorithm 3 (Appendix A) presents the calculation procedure for the elastic-plastic

solutions of supported tunnels in the secondary support stage.

3.4.4 Reliability analysis of the supported tunnels

Concerning the support systems, the failure happens when the equivalent stress exceeds the allowable

stress of the constitutive material [19, 31]. The equivalent stress FL of the primary (L1) and secondary (L2)

support systems are defined as shown in Eqs. (51) and (52), respectively [19]. We assume that the support

system will fail in case FL ≤ 0.

F L1   cL1    L1   r L1 (51)

F L 2   cL 2    L 2   r L 2 (52)

27
where cL1 and cL2 represent the compressive strength of the primary and secondary support system,

respectively. Following the conclusions in Song et al. [31], the points at the inner boundary of the

concrete liners are the most critical points of the supports, i.e. the first points that would start to fail. In

this research, the safety factor of the primary and secondary support systems are defined as shown in Eqs.

(53) and (54), respectively.

L1  cL1
Fsafety  , r  R2 (53)
  L1   r L1

L2  cL 2
Fsafety  , r  R3 (54)
 L2   r L2

Consequently, the safety factor of the combined support system can be defined as shown in Eq. (55).

Fsafety  min  Fsafety

L1 L2
, Fsafety  (55)

where min [x1, x2] represents the minimum value of x1 and x2.

For a simplified case on which the support forces (PS) only act on the outer boundary of the liner (Router)

and no support forces act on the inner boundary of the liner (Rinner), e.g. the case of the secondary support

system, combining the Eqs. (39), (41) and (51), (52), the equivalent stress FL can be simplified to Eq. (56).

2
Router
F L   cL  2 PS 2 2
(56)
Router  Rinner

Thus, by assigning FL equal to 0 in Eq. (56), the maximum support forces that the support system can

accept (P Smax ) can be expressed as in Eq. (57) and its safety factor ( Fsafety
L
) can be expressed as in Eq. (58).

1  R2 
PSmax   cL  1  inner
2  (57)
2  Router 

L  cL  2
Rinner 
Fsafety   1  2  (58)
2 PS  Router 

28
 L
It should be noted that the expressions of the maximum support forces P Smax and the safety factor Fsafety

are consistent with those used in the CCM for the above simplified case [2].

However, for tunnels with a combined support system, the support forces not only act on the outer

boundary but also on the inner boundary of the primary support system, and thus, the support forces and

the safety factor should be determined by the compatibility conditions of the tunnels. Therefore, at this

point the CCM meets one of its limitations: the design of tunnels with a combined support system.

3.5 Application in the design of non-circular tunnels

Although the obtained solutions are rigorously applicable only for the supported circular tunnels, the

solutions are meaningful for a much wider range of non-circular tunnels [22, 23, 73, 74]. By using the

equivalent radius (Req), the proposed solutions may be used to approximate estimate the mechanical

response of non-circular tunnels. The equivalent radius (Req) function can be determined based on the

equal area method [22, 23, 74], as shown in Eq. (59).

A
Req  (59)


where A represents the cross-section area. The reader is referred to the original source for a detailed

description of the equivalent method and its application in the design of non-circular tunnels.

4. Verification of the proposed method

4.1 Verification of the solutions for unsupported circular tunnels

Alonso et al. [10] proposed self-similar solutions for the ground reaction curves (GRC) of unsupported

circular tunnels. In this sub-section, the results obtained with our proposed numerical solution are
29
compared with those obtained with self-similar solutions [7, 10] . In the calculation process, the number

of annuli n adopted is 500. To be consistent with the assumptions of Alonso et al. [10], no supports and no

tunnel advancement have been considered. The Mohr-Coulomb strain-softening behaviour model has

been adopted for the rock mass, and the input data is shown in Figure 8. In this figure, we can observe a

good match between the proposed numerical solutions and the self-similar solutions for the ground

reaction curves and the peak and residual plastic radii, which can serve as a verification of the proposed

strain-softening solutions for unsupported tunnels.

0.40 0.5
=p/2=p/4 =p/8 = Rp/R Rs/R
0.35 Solutions in this paper Solution in this paper
Self-similar solutions 0.4 Self-similar solution
0.30 (Alonso et al., 2003)
(Alonso et al., 2003)
0.25 p0=20 MPa, R1=3 m
p0=20 MPa, R1=3 m 0.3
pinner/p0

E=10000 MPa=0.25
pi/p0

0.20 E=10000 MPa=0.25

pdegsdeg
pdegsdeg 0.2
0.15 cpMPacsMPa
cpMPacsMPa
0.10 
psdeg
0.1 
0.05

0.00 0.0
0 10 20 30 40 50 60 70 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
ur/R1 2G/(p0-Rp) Rs/R Rp/R
(a) (b)

2.0
Solutions in this paper
1.8
Alejano et al. (2009)
1.6
p0=12 MPa, R1=7 m
1.4 E=3837 MPa=0.25
Pinner [MPa]

1.2 pdegsdeg
1.0 psdeg
0.8 cpMPacsMPa
0.6
0.4
0.2
0.0
0 200 400 600 800 1000 1200 1400
ur [mm]
(c)

30
Figure 8. Comparison of (a) normalized displacements, (b) normalized peak and residual plastic radii,
and (c) absolute displacements, for the circular openings between the proposed solutions
and the self-similar solutions.

4.2 Verification of the solutions for supported tunnels

The support forces acting on the rock mass are assumed to be the fictitious support forces Pf and the

interaction support forces PS1 j  R1 , xd  . The support forces acting on the outer and inner boundaries of the

primary support system are PS1 j  R1 , xd  and PS22  R2 , xd  , while those acting on the outer and inner

boundaries of the secondary support system are PS22  R2 , xd  and 0. The strain-softening solutions of the

ground reaction curves have been verified in the section 4.1, and thus, as long as the expressions of Pf,

PS1 j and PS22 are correct, the solutions for supported tunnels will be acceptable. The adopted method for

determining the fictitious support forces Pf has been verified in Cui et al. [18]. Thus, in this sub-section,

only the interaction support forces PS1 j and PS22 , as well as the compatibility conditions at the rock-

support interface (r = R1) and the support-support (r = R2) are checked to verify the correctness of the

derivation.

In this sub-section, an example is carried out to verify the compatibility conditions. A tunnel with radius

R1 = 7 m is excavated in a strain-softening rock mass. Table 1 presents the input parameters of the rock

mass [7]. The primary and secondary support systems present elastic moduli of EL1 = 29400 MPa and

EL2 = 33300 MPa, respectively, and the same Poisson’s ratio (vL1 = vL2 = 0.25) [12]. The thickness of the

primary and secondary supports is considered: d1 = 0.20 m and d2 = 0.25 m. The installation distance of

the supports from the tunnel face is considered as x1 = 2.0 m, x2 = 7.0 m. In the calculation process, the

number of annuli n adopted is 500.

31
Table 1. Input parameters of the rock mass.

Input parameters of the rock mass

GSI peak 55 Dilatancy p [deg] 3.1

residual 33 s [deg] 3.1

Cohesion cp [MPa] 0.744 Softening parameters * 0.017

cs [MPa] 0.397 Initial stress P0 [MPa] 12

Friction angle p [deg] 24.81 Elastic modulus E [MPa] 3837

s [deg] 15.69 Poisson’s ratio  0.25

Figure 9 shows a good agreement between the incremental displacements occurring on the rock mass after

the installation of the primary support system (xd ≥ x1; u rock

r (R1, xd) - u rock
r (R1, x1) ) and the absolute

displacements of the primary support at the rock-support interface (r = R1; urL1(R1, xd)). In addition, it can

be observed a good agreement between the incremental displacements of the primary support system

–occurring after the installation of the secondary support system (xd ≥ x2; u rock
r (R2, xd) - u rock
r (R2, x2) )–

and the absolute displacements of the secondary support system at the support-support interface (r = R2;

urL2(R2, xd) ). It can be noted that the solutions perfectly satisfy the compatibility conditions of both rock-

support interface (r = R1) and support-support interface (r = R2). Therefore, we can confirm that the

solutions for the supported circular tunnels are correct.

32
 7
urock
r
(R1,xd)-urock
r
(R1,x1) urL1(R1,xd)
6 u (R ,x )-u (R ,x ) urL2(R2,xd)

Displacements [mm]
rL1 2 d rL1 2 2

0
0 10 20 30 40 50 60
xd [m]

Figure 9. Comparison of incremental displacements on the surrounding rock mass occurring after the
installation of supports, and the displacements of supports.

5 Comparison between the proposed method and the convergence-confinement method

In this section, the results obtained from between the proposed method and the convergence-confinement

method (CCM) for designing of tunnels with the primary and two support systems are compared in

sections 5.1 and 5.2, respectively. The CCM consists of three basic graphs:

(1) The longitudinal deformation profile (LDP), which relates the radial displacements of an unsupported

tunnel section with its longitudinal distance to the tunnel face;

(2) The ground reaction curve (GRC), which establishes the relationship between the decreasing inner

pressure and the increasing radial displacements of the tunnel wall in plane-strain problems;

(3) The support characteristic curve (SCC), which represents the stress-strain relationship of the support

system.

At the distance xd = x1 to the tunnel face, i.e. point A in Figure 10, the radial displacements of the

 
tunnel u rrock R1 , x1 , can be determined by the LDP. Thus, we could obtain the inner pressure at that

distance from the GRC (point C). Moreover, the support system is installed at point B, when some

33
convergence has already occur and, thus, a part of fictitious support forces from the tunnel face Pf has

been dissipated and will not be loaded by the support. Point D represents the failure point of the support

system. The intersection between the GRC and the SCC in Figure 10 represents the equilibrium of the

rock-support system, when the load on the support matches the inner pressure in the tunnel (i.e. support

forces from the tunnel face Pf equal to 0). The detailed description of the CCM can be found in the

scientific literature [2, 6, 12].

In tunnel engineering projects, the primary and secondary support systems are put in place at different

times to ensure the temporary and long-term stability of the tunnel, respectively. Oreste [2] proposed

support characteristic curves for a compound support system, extending the application of the CCM for

the design of tunnels with a combined support system. Nonetheless, for tunnels with two different support

systems, the support forces not only act on the outer boundary but also on the inner boundary of the

primary support system, and thus, the support forces and the safety factor should be determined by the

compatibility conditions of the tunnels. However, no compatibility conditions at the interface between

supports are considered in the CCM. Therefore, at this point the CCM meets one of its limitations in the

design of secondary support systems. The reason is that the ground reaction curve is modified by the

primary support system, and thus, we should use this modified GRC for the correct application of the

CCM in the design of the secondary support system. In our proposed method, the real compatibility

conditions at both the rock-support interface and the support-support interface are considered, and, hence,

the safety factors obtained using our method are probably closer to reality than those obtained using the

CCM.

34
Figure 10. Main elements of the convergence-confinement method (CCM). Typical graphs obtained
for the longitudinal deformation profile (LDP) –above–, and for the ground reaction curve
(GRC) and the support characteristic curve (SCC) –below. Based on the work of Alejano
et al. [6-8].

5.1 Design of the primary support system

In this sub-section, the results for the design of the primary support system obtained with our proposed

method are compared with those from the CCM. In our method, the support forces acting on the rock

mass are the sum of the fictitious support forces Pf (xd), and the support forces caused by the interaction

35
between the support and the tunnel PS  R1 , xd  (hereinafter referred to as interaction support forces), while
1j

only interaction support forces PS  R1 , xd  act on the primary support system.

1j

However, in the CCM, as shown in Figure 10, at the equilibrium point E, the support forces acting on the

rock mass are equal to the support forces acting on the primary support system. In fact, when the distance

to the tunnel face xd tends to a big enough value, the fictitious support force from the tunnel face Pf (xd) is

zero and, in that case, our solutions are consistent with the solutions of the CCM. Thus, the solutions of

the CCM constitute a particular case of the solutions of this article.

To explain it better, an example is carried out herein. The geometry and the material properties are the

same as those used in section 4.2. The primary support system is installed at a distance x1 = 10 m from the

tunnel face. Based on the CCM, as shown in Figure 11, the equilibrium solutions (Point E in Figure 10)

between the support forces and the displacements on the surrounding rock mass are Pinner= PS11  R1 ,   =

0.638 MPa, and urrock = 0.434 m, respectively.

1.4 1.4

1.2 LDP 1.2

GRC
1.0 1.0
Pinner [MPa]
[m]

0.8 0.8
SCC
urock

0.6 0.6
r

0.4 0.4

0.2 0.2

0.0 0.0
0 20 40 60 80 100 120 140 160 180 200 0.2 0.4 0.6 0.8 1.0 1.2 1.4
rock
Distance to the tunnel face, xd, [m] u [m]
(a) (b) r

Figure 11. (a) The longitudinal deformation profile (LDP), (b) the ground reaction curve (GRC) and
the support characteristic curve (SCC), for a medium quality rock mass using the strain-
softening model.

36
In this article, however, we present the equilibrium solutions for the whole process of tunnel advancement,

as shown in Figure 12. In the calculation process of the proposed solutions, the number of annuli n

adopted is 500. With the advancement of the tunnel, the fictitious support force Pf (xd) decreases from the

initial stress to zero, while the support force acting on the primary support system P 11S(R1, xd) increases

from zero to a constant value at a big enough distance to the tunnel face. In CCM, only the solution for

Pf(xd)=0 –i.e. big enough distance from tunnel face– is presented. However, the equilibrium solutions of

displacements and support forces during the whole process of excavation and advancement are

significantly important, especially for the design of the secondary support system.

1.6 1.6
Unsupported tunnel
Supporting forces [MPa]
1.4 1.4 Pf (xd)+P11(R1, xd)
Supported tunnel S

1.2 1.2 P11

S
(R1, xd)
1.0 1.0
[m]

0.8 0.8
urock
r

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Distance to the tunnel face, xd, [m] Distance to the tunnel face, xd, [m]
(a) (b)

Figure 12. (a) The displacements of tunnels along r = R1 for supported and unsupported tunnels,
(b) support forces act on the rock mass (Pf (xd)+P 11S(R1, xd)) and support forces acting on the

primary support system (P 11S(R1, xd)), versus the distance to the tunnel face xd.

5.2 Design for tunnels with two support systems

Regarding the design of tunnels with a combined support system, in this sub-section, we show a

comparison between the results obtained with our proposed method and those from the CCM. Thus, an

example is carried out herein, where the geometry and the properties of the rock mass are the same as

those used in section 4.2 (Table 1). In the calculation process of the proposed solutions, the number of
37
annuli n adopted is 500. Concerning the thickness d of the support systems, we have selected d1 = 0.15 m

for the primary support and d2 = 0.20 m for the secondary support. In addition, two different cases have

been considered:

- Case (1): In this case, we consider steel fibre reinforced shotcrete as the primary support system (L1)

and a concrete liner as the secondary support (L2). Moreover, their corresponding parameters are: EL1 =

37 GPa, EL2 = 8 GPa, vL1 = vL2 = 0.2, cL1 = 45 MPa, cL2 = 16 MPa. The primary and secondary support

systems are installed at a distance x1 = 1 m and x2 = 1.2 m from the tunnel face, respectively.

- Case (2): In this case, unreinforced shotcrete have been considered for the primary (L1) support system

and a concrete liner for the secondary (L2). Their parameters are EL1 = 8000 MPa, EL2 = 10000 MPa,

vL1 = vL2 = 0.2, cL1 = 16 MPa, cL2 = 20 MPa. The primary and secondary support systems are installed at

a distance x1 = 10 m and x2 = 10.5 m from the tunnel face, respectively.

Hence, following our method, and through a number of calculations, we can obtain the maximum value of

the support forces acting on the support systems –when the fictitious support forces Pf are equal to 0 at a

certain distance from the tunnel face–, as well as the safety factor.

In this example, according to the proposed method in this paper, the maximum values of the support

forces (P 12 22 L1 L2
S,max and P S,max ) and the minimum safety factors of the primary (F safety ) and secondary (F safety )

support systems can be obtained for case (1) and case (2), as shown in Table 2. According to Eqs. (53)-

L1-L 2
(55), the safety factor of the combined support system (F safety ) for case (1) and case (2) are 0.9 and 1.4,

respectively. Regarding the maximum displacements for the tunnel wall u rock
r,max , for case (1) and case (2)

we obtain 0.18 m and 0.44 m, respectively (Table 3). In any case, it must be kept in mind that the

maximum displacements obtained –especially in case (2)– might not be tolerable, depending of the

project.

38
Furthermore, following Oreste [2, 7, 63], Figure 13 presents the results obtained by the CCM for tunnels

with a combined support system. In Figure 13, the final support forces that act on the tunnel wall reach a

value of around 1.30 MPa for case (1), and around 0.62 MPa for case (2). On the other hand, the

maximum tunnel deformations are around 0.19 m for case (1) and around 0.45 m for case (2). Based on

the CCM, the safety factors for case (1) and case (2) would be 1.1 and 1.5, respectively, which are

different from the safety factors obtained by the method proposed in this article: safety factor: 0.9 for case

(1), and 1.4 for case (2).

Table 3 shows a comparison between the results obtained with our proposed method and the CCM in

terms of maximum displacements on the tunnel wall (u rock

r,max ) and safety factors of the combined support

L1-L 2
system (F safety ). We can observe that the maximum displacements obtained with our method are similar,

but slightly lower than those obtained with the CCM. However, the safety factors obtained with our

method are significantly lower than those obtained using the CCM. The difference in the results between

both methods is due to the fact that, unlike the CCM, our method considers the real compatibility

conditions at both the rock-support interface (r = R1) and the support-support interface (r = R2). Instead,

no compatibility conditions between the primary and secondary support system are considered in the

CCM. In consequence, because of an overestimation of the safety factor, the use of the CCM can lead to

an unsafe design of tunnels with combined support systems.

39
 2.4 1.6
The GRC The GRC
2.1 The SCC for the combined support system 1.4 The SCC for the combined support system
The SCC for the primay support system The SCC for the primay support system
1.8 The SCC for the secondary support system 1.2 The SCC for the secondary support system
Pinner [MPa]

Pinner [MPa]
1.5 case (1) 1.0 case (2)

1.2 0.8

0.9 0.6

0.6 0.4

0.3 0.2

0.0 0.0
0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
rock rock
u [m] u [m]
(a) r (b) r

Figure 13. Ground reaction curves (GRC) and support characteristic curves (SCC), in the
convergence-confinement method (CCM); (a) for case (1), and (b) for case (2).

L1-L2
Table 2. Safety factors (F safety ) obtained using our proposed method.

The maximum support forces The minimum safety factors of the

Cases 12
P S,max [MPa]
22
P S,max [MPa]
L1
F safety
L2
F safety

Case (1) 1.33 0.30 0.9 (failure) 1.5

Case (2) 0.63 0.39 1.4 1.4

L1-L2
Table 3. Safety factors (F safety ) obtained using our proposed method and the convergence-

confinement method (CCM).

rock L1-L2
The maximum displacements, u r,max [m] The safety factor, F safety

Cases Our method CCM Our method CCM

Case (1) 0.18 0.19 0.9 (failure) 1.1

Case (2) 0.44 0.45 1.4 1.5

40
6 Analysis and discussion

6.1 Discussion on the selection of rock mass behaviour models

In this section, the results of our proposed numerical solutions are analysed using an example. In this

example, three different models have been used, including the elastic-perfectly-plastic (EPP), strain-

softening (SS) and elastic-brittle (EB) behaviour models. Nonetheless, note that the SS model can be

simplified to the EPP model by assigning a big enough value of * (in this example, * = 1.0), or to the

EB model by assigning a small value of * (in this example, * = 0). In addition, four different quality

rock masses have been considered in the analysis. Table 4 presents the parameters of the different rock

masses considered. In the calculation process, the number of annuli n adopted is 500.

Input parameters of four different quality rock masses

Different quality rock masses A2 B2 C2 D2
Peak value GSIpeak 75 60 50 40
GSI
Residual value GSIres 40 35 30 27

Elastic Elastic modulus E (MPa) 36500 15400 8660 4870

parameters Passion’s Ratio v 0.25 0.25 0.25 0.25
Cohesion (MPa) 3.637 2.67 2.24 1.88
Peak values
Friction angle (deg) 29.52 25.7 23.1 20.6
Cohesion (MPa) 1.878 1.71 1.54 1.43
Post-failure Residual values
behaviour Friction angle (deg) 20.64 19.4 18.2 17.5
Dilatancy angle peak=res (deg) 7.38 4.49 2.89 1.55
Critical softening
* 0.0011 0.0062 0.0288 0.119
parameter
Table 4. Input parameters of four different quality rock masses. Based on the work of Alejano et al.
[8].

41
On the other hand, the primary and secondary support systems are installed at the distance x1 = 2.5 m, and

x2 = 10 m, respectively. The primary and secondary support systems are assumed to behave as elastic

materials. The elastic properties (elastic modulus and Poisson’s ratio) of the primary (L1) and secondary

(L2) supports systems are: EL1 = 8000 MPa, EL2 = 12000 MPa, and vL1 = vL2 = 0.15. The tunnel has a

diameter of 5 m and its depth of 1500 m implies a field stress of 37.5 MPa [8]. Rock mass specific weight

is 25 KN/m3 and the Poisson’s ratio is 0.25 [8]. The radii related to the primary and secondary support

systems (see Figures 1 and 2) are: R1 = 2.5 m, R2 = 2.4 m and R3 = 2.25 m.

Figure 14 shows the displacements (u rock

r ) along the tunnel wall (r = R1) versus the distance to the tunnel

face (xd) for a variety of cases, i.e. the longitudinal deformation profiles. These results meet the same

conclusion of Alejano et al. [8]: there are significant differences in the displacements of unsupported

tunnels when using different behaviour models (EPP, SS or EB), and the difference between the

displacements of tunnels calculated for EB rock masses and for SS rock masses grows as the GSI

decreases.

In addition, in this article, it can be found that the behaviour model selection also makes a significant

effect on the behaviour of supported tunnels. It is therefore important to note that SS behaviour plays a

role in the behaviour of tunnels, and hence, in the design and installation of supports. In Figure 14, it can

be observed that, initially, the displacements gradually increase with the tunnel advancement, as expected

(e.g. Alejano et al., 2012 [8]). However, after the installation of the support system, the displacement rate

decelerates suddenly and the maximum displacement achieved is significantly lower.

The support forces acting on the tunnel wall are the sum of the fictitious support forces Pf and the

interaction support forces at the rock-support interface P 1Sj (j = 1, 2 represent the primary and secondary

support stages, respectively). Moreover, the support forces acting on the outer boundary of the primary

support system (r = R1) are P 1Sj and the support forces at the support-support interface (r = R2) are P S22 .
42
 0.10 0.25
GSIpeak=75 SS EPP EB GSIpeak=60 SS EPP EB
GSIres =40 urock-supported GSIres =35 urock-supported
0.08 r 0.20 r

urock
r
-unsupported urock
r
-unsupported

0.06 0.15
[m]

[m]
urock

urock
0.04 0.10
r

r
0.02 0.05

0.00 0.00
0 5 10 15 20 25 0 5 10 15 20 25
xd [m] xd [m]
(a) (b)

0.6 1.2
GSIpeak=50 SS EPP EB
GSIpeak=40 SS EPP EB
rock GSIres =27 urock-supported
0.5 GSIres =30 u
r
-supported 1.0 r

u rock
-unsupported urock
r
-unsupported
r
0.4 0.8
[m]

[m]
0.3 0.6
urock

urock
r

0.2 0.4

0.1 0.2

0.0 0.0
0 5 10 15 20 25 0 5 10 15 20 25
xd [m] xd [m]
(c) (d)

Figure 14. Tunnel displacements along the cross-section r = R1, versus the distance to the tunnel face,
for four different quality rock masses: (a) A2 (GSIpeak = 75, GSIres = 40), (b) B2 (GSIpeak = 60,
GSIres = 35), (c) C2 (GSIpeak = 50, GSIres = 30), and (d) D2 (GSIpeak = 40, GSIres = 27).

Figure 15 and 16 respectively present the acting forces on the tunnel wall and on the support systems

versus the distance to the tunnel face. In Figure 15, we can observe that the support forces acting on the

1j
tunnel wall (Pf + P S ) decrease with the increasing distance to the tunnel face, and eventually reach a

constant value. Pf represents the fictitious support forces from the tunnel face for unsupported tunnels.

43
Figure 16 presents the variety of support forces PS1 j and P S22 versus the distance to the tunnel face xd: the

support forces ( PS1 j , P S22 ) firstly increased and then achieve a constant value when the distance to the

tunnel face is large enough.

7 7
GSIpeak=75 1j
P +Pf Pf GSIpeak=60 P1jS+Pf Pf
6 S
6
GSIres =40 SS GSIres =35 SS
5 EPP 5 EPP
EB EB
P [MPa]

P [MPa]
4 4

3 3

2 2

1 1

0 0
0 5 10 15 20 25 0 5 10 15 20 25
xd [m] xd [m]
(a) (b)

7 7
GSIpeak=50 P1jS+Pf Pf GSIpeak=40 P1jS+Pf Pf
6 6
GSIres =30 SS GSIres =27 SS
5 EPP 5 EPP
EB EB
P [MPa]

P [MPa]

4 4

3 3

2 2

1 1

0 0
0 5 10 15 20 25 0 5 10 15 20 25
xd [m] xd [m]
(c) (d)

1j
Figure 15. Support forces acting on the tunnel wall, Pf + P S , and fictitious support forces from the
tunnel face Pf (see Figures 1 and 2), versus the distance to the tunnel face xd, for four
different quality rock masses: (a) A2 (GSIpeak = 75, GSIres = 40); (b) B2 (GSIpeak = 60,
GSIres = 35); (c) C2 (GSIpeak = 50, GSIres = 30); and (d) D2 (GSIpeak = 40, GSIres = 27).

44
 2.1 3.2
GSIpeak=75 P 1j
P 22
GSIpeak=60 P1jS P22
1.8 S S 2.8 S
GSIres =40 SS GSIres =35 SS
EPP 2.4 EPP
1.5
EB EB
2.0
P [MPa]

P [MPa]
1.2
1.6
0.9
1.2
0.6
0.8
0.3 0.4

0.0 0.0
0 5 10 15 20 25 0 5 10 15 20 25
xd [m] xd [m]
(a) (b)

4.0 4.8
GSIpeak=50 P1jS P22 GSIpeak=40 P1jS P22
3.5 S 4.2 S
GSIres =30 SS GSIres =27 SS
3.0 EPP 3.6 EPP
EB EB
2.5 3.0
P [MPa]

P [MPa]
2.0 2.4

1.5 1.8

1.0 1.2

0.5 0.6

0.0 0.0
0 5 10 15 20 25 0 5 10 15 20 25
xd [m] xd [m]
(c) (d)

Figure 16. Support forces acting on the outer boundary of the primary support system, P 1Sj , and

support forces acting on the outer boundary of the secondary support system, P S22 (see
Figures 1 and 2), versus the distance to the tunnel face xd, for four different quality rock
masses: (a) A2 (GSIpeak = 75, GSIres = 40); (b) B2 (GSIpeak = 60, GSIres = 35); (c) C2
(GSIpeak = 50, GSIres = 30); and (d) D2 (GSIpeak = 40, GSIres = 27).

Moreover, in Figures 15 and 16, there are significant differences in the fictitious support forces (Pf) as

well as the support forces calculated by different behaviour models (EPP, SS or EB), and the difference

between the displacements of tunnels calculated for EPP rock masses and for SS rock masses grows as the

GSI increases. The differences of stresses and displacements calculated by SS and EPP models can be

45
negligible for lower quality rock masses, as expected, since the EPP behaviour represents well the

behaviour of low quality rock masses [8]. Furthermore, there is a significant difference between

displacements and support forces for unsupported and supported tunnels.

6.2 Application in the design of tunnels

Due to the process of tunnel excavation and construction of supports, the final support forces applied on

the outer boundary of the secondary support system are significantly affected by the installation distance

of support system, as well as the stiffness of the support system [3, 31]. Early installation of the secondary

support system leads to larger final support forces acting on the secondary support system and thus, may

result in failure of the secondary support system [3, 31].

In this sub-section, an example is carried out to present the application of our method to tunnel design in

terms of installation distance of the secondary support system from the tunnel face (x2) and thickness of

the secondary support system (d2 = R2 - R3).

The tunnel is excavated in a strain-softening rock mass with a radius of 7 m. The properties of the rock

mass are the same as those in Table 1. In all the cases, the elastic modulus of the primary (L1) and

secondary (L2) support systems are EL1 = 8000 MPa and EL2 = 12000 MPa, respectively. The Poisson’s

ratio of the support systems is vL1 = vL2 = 0.15. The installation distance of the primary support system is

x1 = 1 m, and the thickness of the primary support system is d1 = R1 - R2 = 0.1 m. In this example, the

unconfined compressive strength of the secondary support system is assumed as cL2 = 24 MPa. In the

calculation process, the number of annuli n adopted is 500.

The equivalent stress of the secondary support system FL2 can be obtained, following Eq. (52). We

assume that the secondary support system will fail in case FL2 ≤ 0, so as to approximately predict the

46
initial failure in the secondary support system. For the sake of tunnel stability, the secondary support

system is not expected to fail, and thus the equivalent stresses (FL2) should be positive in the tunnelling

design.

In the application example, Figure 17 presents the equivalent stresses (FL2) as a function of the installation

distance of the secondary support system to the tunnel face (x2), and the thickness of the secondary

support system (d2), where the dash lines represent zero equivalent stress. In Figure 17 (a), the smallest

allowable values for the installation distance from the tunnel face (x2) of the secondary support system are

obtained when the equivalent stress becomes smaller than zero. It should be noted that the bigger is the

value of the support thickness, the smaller will be the recommended values of x2. Moreover, for a constant

installation distance to the tunnel face (x2), increasing the thickness of liners benefits the stability of

tunnels although the cost of the support would increase, which is also shown in Figure 17 (b), where the

economic limit regarding the thickness of the secondary support system (d2) is obtained when the

equivalent stress is zero. Note that the aforementioned economic limit corresponds to a safety factor of 1.0.

Therefore, with the proposed method, the optimization design on both distance from the tunnel face and

thickness of the secondary support system can be conducted conveniently.

0.4 0.8
d2=0.10 m
0.3 x2=3 m
d2=0.15 m 0.6 x2=5 m
0.2 d2=0.20 m
x2=7 m
0.4
0.1
FL2
FL2

0.0 0.2

-0.1 L2
F =0 0.0
-0.2
-0.2 FL2=0
-0.3

-0.4 -0.4
2 4 6 8 10 12 14 0.0 0.1 0.2 0.3 0.4 0.5
x2 [m] d2 [m]
(a) (b)
47
Figure 17. Equivalent stresses of the secondary support system, FL2, versus: (a) installation distance
of the secondary support system, for various thickness of the support system; (b) thickness
of the secondary support system for various installation distances.

7 Conclusions

This paper provides numerical solutions for supported tunnels excavated in strain-softening rock masses,

which can be used in tunnelling for the design of the sequential installation of two different support

systems in terms of strength and installation distance. In our approach, primary and secondary support

systems are sequentially installed. The fictitious support forces have been adopted to reflect the three-

dimensional face effect. The solutions for displacements and stresses at the whole process of tunnel

excavation and installation of supports have been presented, considering the tunnel advancement and real

continuity conditions between rock-support interface and support-support interface are considered. The

Finite Difference Method (FDM) has been employed in the derivation of the solutions.

As a verification step, a good agreement between the proposed solutions and the results from the self-

similar solutions has been obtained, and the compatibility conditions of the supported tunnels have been

checked. Then, the proposed solutions have been compared with the convergence-confinement method

(CCM). Unlike the CCM, our method considers the real compatibility conditions at both the rock-support

interface and the support-support interface. Consequently, the safety factors obtained using our method

are probably closer to reality than those obtained using the CCM, which may be overestimated, leading to

an unsafe design of tunnels with combined support systems. Therefore, our proposed method may be used

with greater confidence in the design of supported tunnels.

Furthermore, for the sake of comparison, elastic-perfectly-plastic (EPP), strain-softening (SS) and elastic-

brittle (EB) material models have been adopted to analyse the effects of the post-failure behaviour model

selection on tunnel deformations and support forces. We have observed significant differences in the

48
displacements and the support forces considering the different behaviour models. In general, these

differences grow as the quality of the rock mass decreases.

Finally, the proposed solutions have been used for the assessment of the design and optimization of the

secondary support system, taking into account the distance from the tunnel face and the thickness of the

secondary support system. Therefore, the proposed approach constitutes an alternative method for

preliminary tunnel analysis and design, providing a greater confidence and assessment capability.

The obtained solutions are applicable only for the supported circular tunnels, although the solutions are

meaningful for a much wider range of non-circular tunnels. In the future research, numerical and

analytical solutions for the non-circular tunnels will be considered to improve the applicability of the

solutions in real engineering

Conflict of interest

None

Acknowledgements

Fei Song has been supported by the CSC Scholarship (No. 201706260240).

49
Appendix A. Algorithms for the calculation procedure to obtain different solutions

Algorithm 1. Calculation procedure to obtain the elastic-plastic solutions for circular openings in rock
masses
Input: R1, P0, n, Pinner, material parameters, xd

Calculate the fictitious support forces Pf (xd), following section 3.2 and Figures 6 and 7

For unsupported tunnels, Pinner= Pf (xd); for supported tunnels Pinner= Pf (xd) + P 1Sj (xd)

Calculate Pcritical using Eq. (14)

if Hoek-brown model, then

Calculate the equivalent Mohr-Coulomb parameters, based on Eqs. (8) and (9)

end if

if Pinner > Pcritical, then

typical elastic solutions

else

(0)= p ( represents plastic parameters)

R = (2P0-S(0))/(1+N(0)), r=(Pinner-Rp)/n
p

r(0)= Rp, (0)= 2P0-R

p

r(0)=(R -P0)/(2G), (0)=(P0-Rp)/(2G)

p

 pr 0  =0,  p  0  =0,  p0 =0

0=1, ur(0) = ur(0) R p     0   0

do i=1, n

r (i)={[r(i-1)+r-S(i-1)/(1-N(i-1))]/[r(i-1)-S(i-1)/(1-N(i-1))]}^[ 1/(N(i-1)-1)]

50
 r(i)=r(i-1)+r, (i)=N(i-1)r(i)+S(i-1)

1  2  
    p  ,     1 E     p   1     p 
2
 rei  
E 

  r  i   p0 
1 
  i 0
e
 i  i 0 r i 0

K  i 1
 ri 1  r
fi   r  r dr ; f i   r    re i   K i 1e i  for r  r i 
 K  i1 K i 1
u i   u i 1 
 r 
  r 
r i1
 

u i 
 pi  =   ei  ,  rp i  =  K  i 1p i 
i 

pi   pi    rpi 

if 0 < pi < *, then

 p  s p
i    p  i  , r s= r (i)
*

else

i   s

end if

end do

end if

Output: Rp= R1/ r (i), Rs= r s R1/ r (i) , urrock (r=R1)= u i  R1/ r (i)

51
Algorithm 2. Calculation procedure to obtain the solutions for tunnels in the primary support stage
Input: R1, R2, P0, n, Pinner, material parameters, x1, xd, f (error limit of f)

Obtain urRock  R1 , x1  , from Algorithm 1

Calculate the fictitious support forces Pf(xd), following section 3.2 and Figures 6 and 7

 1 R12 R22 1  2 L R13   Rock

f(  rRock
n )= R  Rock
1  n  u Rock
r  1 1   2G r R 2  R 2 2G R2  R 2   r n   pf  xd 
R , x  
 L 1 2 L 1 2 

Pinner_a=0, Pinner_b=P0

Pinner = Pinner_a,  rRock

 n  = Pinner

Obtain fa=f(a), from Algorithm 1

Pinner = Pinner_b,  rRock

 n  = Pinner

Obtain fb=f(b), from Algorithm 1

do while (fa – fb  f)

Pinner_c=(Pinner_a + Pinner_b)/2

Pinner = Pinner_c,  rRock

 n  = Pinner

Obtain fc=f(c), from Algorithm 1

if fafc  0 then

fa = fc, a=c

else

fb = fc, b=c

end if

end do

52
Output: Ps11 (xd)= Pinner(xd) - Pf(xd), urRock R1 , xd =R1  Rock
n  

Algorithm 3. Calculation procedure to obtain the solutions for tunnels in the secondary support stage
Input: R1, R2, R3, P0, n, Pinner, material parameters, x1, x2, xd, f (error limit of f), g (error limit of g)

From algorithms 1 and 2, obtain PS11  x2  and urRock  R1 , x1 

Calculate the fictitious support forces Pf(xd), following section 3.2 and Figures 6 and 7

 1 R1 R22 1  2 L1 R13 
A=  2 2
 
 2G L1 R1  R2 2G L1 R12  R22 

 1 R1 R22 1  2 L1 R1 R22 
B=  2 2
 
 2G L1 R1  R2 2G L1 R12  R22 

 1 R12 R2 1  2 L1 R12 R2 
C=  2 2
 
 2G L1 R1  R2 2G L1 R12  R22 

 1 R12 R2 1  2 L1 R23 
D=  2 2
 
 2G L1 R1  R2 2G L1 R12  R22 

 1 R2 R32  1 2 L2 R23 
E=  2 2 
 
 2GL 2 R2  R3  2GL2 R22  R32 

B C  Rock B C PS  x2 
11

g(  Rock
r n 
)= R1  Rock
 n u Rock
r  R , x   A 
1 1
Rock
r n 
 pf  xd   
 D+ E 
 r  n   pf  xd   
 D E

Pinner_d=0, Pinner_e=P0

Pinner = Pinner_d,  rRock

 n  = Pinner

Obtain gd=g(d), from Algorithm 1

Pinner = Pinner_e,  rRock

 n  = Pinner

Obtain ge=g(e), from Algorithm 1

53
do while (gd – ge  g)

Pinner_m=(Pinner_d + Pinner_e)/2

Pinner = Pinner_m,  rRock

 n  = Pinner

Obtain gm=g(m), from Algorithm 1

if gdgm  0

gd = gm, d=m

else

ge = gm, e=m

end if

end do

C  PS12  xd   PS11  x2  
Output: urRock  R1 , xd  =R1  Rock
 n  , PS  xd  =  r  n   Pf  xd  , pS  xd  
12 Rock 22

D E

54
References
[1] E. Leca, G.W. Clough, Preliminary design for NATM tunnel support in soil, Journal of

geotechnical engineering, 118 (1992) 558-575.

[2] P.P. Oreste, Analysis of structural interaction in tunnels using the covergence–confinement

approach, Tunnelling and Underground Space Technology, 18 (2003) 347-363.

[3] F. Song, H.N. Wang, M.J. Jiang, Analytically-based simplified formulas for circular tunnels with

two liners in viscoelastic rock under anisotropic initial stresses, Construction and Building

Materials, 175 (2018) 746-767.

[4] F. Song, A. Rodriguez-Dono, S. Olivella, Z. Zhong, Analysis and modelling of longitudinal

deformation profiles of tunnels excavated in strain-softening time-dependent rock masses,

Computers and Geotechnics, 125 (2020) 103643.

[5] K. Zhao, M. Bonini, D. Debernardi, M. Janutolo, G. Barla, G. Chen, Computational modelling of

the mechanised excavation of deep tunnels in weak rock, Computers and Geotechnics, 66 (2015)

158-171.

[6] L.R. Alejano, E. Alonso, A. Rodriguez-Dono, G. Fernandez-Manin, Application of the

convergence-confinement method to tunnels in rock masses exhibiting Hoek–Brown strain-

softening behaviour, International Journal of Rock Mechanics and Mining Sciences, 1 (2010) 150-

160.

[7] L.R. Alejano, A. Rodriguez-Dono, E. Alonso, G.F. Manín, Ground reaction curves for tunnels

excavated in different quality rock masses showing several types of post-failure behaviour,

Tunnelling and Underground Space Technology, 24 (2009) 689-705.

[8] L.R. Alejano, A. Rodríguez-Dono, M. Veiga, Plastic radii and longitudinal deformation profiles of

tunnels excavated in strain-softening rock masses, Tunnelling and Underground Space

Technology, 30 (2012) 169-182.

55
[9] E. Alonso, L. Alejano, G. Fdez-Manin, F. Garcia-Bastante, Influence of post-peak properties in

the application of the convergence-confinement method for designing underground excavations,

strain, 3 (2008) 1.

[10] E. Alonso, L.R. Alejano, F. Varas, G. Fdez‐Manin, C. Carranza‐Torres, Ground response curves

for rock masses exhibiting strain‐softening behaviour, International journal for numerical and

analytical methods in geomechanics, 27 (2003) 1153-1185.

[11] C. Carranza-Torres, C. Fairhurst, The elasto-plastic response of underground excavations in rock

masses that satisfy the Hoek–Brown failure criterion, International journal of rock mechanics and

mining sciences, 36 (1999) 777-809.

[12] C. Carranza-Torres, C. Fairhurst, Application of the convergence-confinement method of tunnel

design to rock masses that satisfy the Hoek-Brown failure criterion, Tunnelling and Underground

Space Technology, 15 (2000) 187-213.

[13] C. Carranza-Torres, B. Rysdahl, M. Kasim, On the elastic analysis of a circular lined tunnel

considering the delayed installation of the support, International Journal of Rock Mechanics and

Mining Sciences, 61 (2013) 57-85.

[14] C. Carranza-Torres, J. Zhao, Analytical and numerical study of the effect of water pressure on the

mechanical response of cylindrical lined tunnels in elastic and elasto-plastic porous media,

International Journal of Rock Mechanics and Mining Sciences, 46 (2009) 531-547.

[15] Z.F. Chu, Z.J. Wu, B.G. Liu, Q.S. Liu, Coupled analytical solutions for deep-buried circular lined

tunnels considering tunnel face advancement and soft rock rheology effects, Tunnelling and

Underground Space Technology, 94 (2019) 103111.

[16] L. Cui, Q. Sheng, Y.-k. Dong, C.-x. Miao, J.-h. Huang, A. Zhang, Two-stage analysis of

interaction between strain-softening rock mass and liner for circular tunnels considering delayed

installation of liner, European Journal of Environmental and Civil Engineering, (2020) 1-26.

56
[17] L. Cui, Q. Sheng, J.J. Zheng, Z. Cui, A. Wang, Q. Shen, Regression model for predicting tunnel

strain in strain-softening rock mass for underground openings, International Journal of Rock

Mechanics and Mining Sciences, 119 (2019) 81-97.

[18] L. Cui, J.J. Zheng, R.J. Zhang, H.J. Lai, A numerical procedure for the fictitious support pressure

in the application of the convergence–confinement method for circular tunnel design, International

Journal of Rock Mechanics and Mining Sciences, 78 (2015) 336-349.

[19] D.P. Do, N.T. Tran, V.T. Mai, D. Hoxha, M.N. Vu, Time-Dependent Reliability Analysis of Deep

Tunnel in the Viscoelastic Burger Rock with Sequential Installation of Liners, Rock Mechanics

and Rock Engineering, (2019) 1-27.

[20] A. Fahimifar, M. Ranjbarnia, Analytical approach for the design of active grouted rockbolts in

tunnel stability based on convergence-confinement method, Tunnelling and Underground Space

Technology, 24 (2009) 363-375.

[21] A. Fahimifar, F.M. Tehrani, A. Hedayat, A. Vakilzadeh, Analytical solution for the excavation of

circular tunnels in a visco-elastic Burger’s material under hydrostatic stress field, Tunnelling and

Underground Space Technology, 25 (2010) 297-304.

[22] E. Kabwe, M. Karakus, E.K. Chanda, Proposed solution for the ground reaction of non-circular

tunnels in an elastic-perfectly plastic rock mass, Computers and Geotechnics, 119 (2020) 103354.

[23] E. Kabwe, M. Karakus, E.K. Chanda, Time-dependent solution for non-circular tunnels

considering the elasto-viscoplastic rockmass, International Journal of Rock Mechanics and Mining

Sciences, 133 (2020) 104395.

[24] A.R. Kargar, An analytical solution for circular tunnels excavated in rock masses exhibiting

viscous elastic-plastic behavior, International Journal of Rock Mechanics and Mining Sciences,

124 (2019) 104128.

57
[25] A.Z. Lu, L.Q. Zhang, N. Zhang, Analytic stress solutions for a circular pressure tunnel at pressure

and great depth including support delay, International journal of rock mechanics and mining

sciences (1997), 48 (2011) 514-519.

[26] S. Maghous, D. Bernaud, E. Couto, Three-dimensional numerical simulation of rock deformation

in bolt-supported tunnels: a homogenization approach, Tunnelling and underground space

technology, 31 (2012) 68-79.

[27] P. Nomikos, R. Rahmannejad, A. Sofianos, Supported axisymmetric tunnels within linear

viscoelastic Burgers rocks, Rock mechanics and rock engineering, 44 (2011) 553-564.

[28] C. Paraskevopoulou, M. Diederichs, Analysis of time-dependent deformation in tunnels using the

Convergence-Confinement Method, Tunnelling and Underground Space Technology, 71 (2018)

62-80.

[29] S.H. Prassetyo, M. Gutierrez, Effect of transient coupled hydro-mechanical response on the

longitudinal displacement profile of deep tunnels in saturated ground, Tunnelling and

Underground Space Technology, 75 (2018) 11-20.

[30] S.H. Prassetyo, M. Gutierrez, Designing Tunnel Support Systems based on Ground Reaction

Curve and Equilibrium Strain Approach, ITA-AITES World Tunnel Congress 2018, 2018.

[31] F. Song, H.N. Wang, M.J. Jiang, Analytical solutions for lined circular tunnels in viscoelastic rock

considering various interface conditions, Applied Mathematical Modelling, 55 (2018) 109-130.

[32] H.N. Wang, X.P. Chen, M.J. Jiang, F. Song, L. Wu, The analytical predictions on displacement

and stress around shallow tunnels subjected to surcharge loadings, Tunnelling and Underground

Space Technology, 71 (2018) 403-427.

[33] H.N. Wang, X. Gao, L. Wu, M.J. Jiang, Analytical study on interaction between existing and new

tunnels parallel excavated in semi-infinite viscoelastic ground, Computers and Geotechnics, 120

(2020) 103385.

58
[34] H.N. Wang, M.J. Jiang, T. Zhao, G.S. Zeng, Viscoelastic solutions for stresses and displacements

around non-circular tunnels sequentially excavated at great depths, Acta Geotechnica, 14 (2019)

111-139.

[35] H.N. Wang, Y. Li, Q. Ni, S. Utili, M.J. Jiang, F. Liu, Analytical solutions for the construction of

deeply buried circular tunnels with two liners in rheological rock, Rock mechanics and rock

engineering, 46 (2013) 1481-1498.

[36] H.N. Wang, G.H. Nie, Analytical expressions for stress and displacement fields in viscoelastic

axisymmetric plane problem involving time-dependent boundary regions, Acta mechanica, 210

(2010) 315-330.

[37] H.N. Wang, F. Song, T. Zhao, M.J. Jiang, Solutions for lined circular tunnels sequentially

constructed in rheological rock subjected to nonhydrostatic initial stresses, European Journal of

Environmental and Civil Engineering, (2020).

[38] H.N. Wang, S. Utili, M.J. Jiang, An analytical approach for the sequential excavation of

axisymmetric lined tunnels in viscoelastic rock, International Journal of Rock Mechanics and

Mining Sciences, 68 (2014) 85-106.

[39] H.N. Wang, S. Utili, M.J. Jiang, P. He, Analytical solutions for tunnels of elliptical cross-section

in rheological rock accounting for sequential excavation, Rock mechanics and rock engineering,

48 (2015) 1997-2029.

[40] H.N. Wang, L. Wu, M.J. Jiang, Viscoelastic ground responses around shallow tunnels considering

surcharge loadings and effect of supporting, European Journal of Environmental and Civil

Engineering, (2018) 1-23.

[41] H.N. Wang, L. Wu, M.J. Jiang, F. Song, Analytical stress and displacement due to twin tunneling

in an elastic semi‐infinite ground subjected to surcharge loads, International Journal for Numerical

and Analytical Methods in Geomechanics, 42 (2018) 809-828.

59
[42] H.N. Wang, G.S. Zeng, M.J. Jiang, Analytical stress and displacement around non-circular tunnels

in semi-infinite ground, Applied Mathematical Modelling, 63 (2018) 303-328.

[43] H.N. Wang, G.S. Zeng, S. Utili, M.J. Jiang, L. Wu, Analytical solutions of stresses and

displacements for deeply buried twin tunnels in viscoelastic rock, International Journal of Rock

Mechanics and Mining Sciences, 93 (2017) 13-29.

[44] M.B. Wang, S.C. Li, A complex variable solution for stress and displacement field around a lined

circular tunnel at great depth, International journal for numerical and analytical methods in

geomechanics, 33 (2009) 939-951.

[45] S.L. Wang, X.T. Yin, H. Tang, X.R. Ge, A new approach for analyzing circular tunnel in strain-

softening rock masses, International journal of rock mechanics and mining sciences (1997), 47

(2010) 170-178.

[46] G.S. Zeng, H.N. Wang, M.J. Jiang, Analytical study of ground responses induced by the

excavation of quasirectangular tunnels at shallow depths, International Journal for Numerical and

Analytical Methods in Geomechanics, 43 (2019) 2200-2223.

[47] K. Zhang, G.M. Zhang, R.B. Hou, Y. Wu, H.Q. Zhou, Stress evolution in roadway rock bolts

during mining in a fully mechanized longwall face, and an evaluation of rock bolt support design,

Rock Mechanics and Rock Engineering, 48 (2015) 333-344.

[48] Q. Zhang, B.S. Jiang, H.J. Lv, Analytical solution for a circular opening in a rock mass obeying a

three-stage stress–strain curve, International Journal of Rock Mechanics and Mining Sciences, 100

(2016) 16-22.

[49] Q. Zhang, B.S. Jiang, S.L. Wang, X.R. Ge, H.Q. Zhang, Elasto-plastic analysis of a circular

opening in strain-softening rock mass, International Journal of Rock Mechanics and Mining

Sciences, 50 (2012) 38-46.

60
[50] Q. Zhang, H.Y. Wang, Y.J. Jiang, M.M. Lu, B.S. Jiang, A numerical large strain solution for

circular tunnels excavated in strain-softening rock masses, Computers and Geotechnics, 114 (2019)

103142.

[51] J.F. Zou, C. Li, F. Wang, A new procedure for ground response curve (GRC) in strain-softening

surrounding rock, Computers and Geotechnics, 89 (2017) 81-91.

[52] S.C. Li, M.B. Wang, An elastic stress–displacement solution for a lined tunnel at great depth,

International Journal of Rock Mechanics and Mining Sciences, 45 (2008) 486-494.

[53] Q. Shen, J.J. Zheng, L. Cui, Y. Pan, B. Cui, A procedure for interaction between rock mass and

liner for deep circular tunnel based on new solution of longitudinal displacement profile, European

Journal of Environmental and Civil Engineering, (2019) 1-19.

[54] A. Vrakas, G. Anagnostou, A finite strain closed‐form solution for the elastoplastic ground

response curve in tunnelling, International Journal for Numerical and Analytical Methods in

Geomechanics, 38 (2014) 1131-1148.

[55] H.S. Yu, Expansion of a thick cylinder of soils, Computers and Geotechnics, 14 (1992) 21-41.

[56] C. Carranza-Torres, Self-similarity analysis of the elasto-plastic response of underground

openings in rock and effects of practical variables, University of Minnesota1998.

[57] E.T. Brown, J.W. Bray, B. Ladanyi, E. Hoek, Ground response curves for rock tunnels, Journal of

Geotechnical Engineering, 109 (1983) 15-39.

[58] Y.-K. Lee, S. Pietruszczak, A new numerical procedure for elasto-plastic analysis of a circular

opening excavated in a strain-softening rock mass, Tunnelling and Underground Space

Technology, 23 (2008) 588-599.

[59] K.-H. Park, B. Tontavanich, J.-G. Lee, A simple procedure for ground response curve of circular

tunnel in elastic-strain softening rock masses, Tunnelling and Underground Space Technology, 23

(2008) 151-159.

61
[60] S.L. Wang, S.D. Yin, Z.J. Wu, Strain‐softening analysis of a spherical cavity, International journal

for numerical and analytical methods in geomechanics, 36 (2012) 182-202.

[61] S.L. Wang, X.T. Yin, H. Tang, X.R. Ge, A new approach for analyzing circular tunnel in strain-

softening rock masses, International Journal of Rock Mechanics and Mining Sciences, 1 (2010)

170-178.

[62] Q. Zhang, C.H. Zhang, B.S. Jiang, N. Li, Y.C. Wang, Elastoplastic coupling solution of circular

openings in strain-softening rock mass considering pressure-dependent effect, International

Journal of Geomechanics, 18 (2017) 04017132.

[63] P. Oreste, Analysis of the Interaction between the Lining of a TBM Tunnel and the Ground Using

the Convergence-Confinement Method, American Journal of Applied Sciences, 12 (2015) 276.

[64] J. Sulem, M. Panet, A. Guenot, An analytical solution for time-dependent displacements in a

circular tunnel, International journal of rock mechanics and mining sciences & geomechanics

abstracts, Elsevier, 1987, pp. 155-164.

[65] E. Hoek, E.T. Brown, Practical estimates of rock mass strength, International journal of rock

mechanics and mining sciences, 34 (1997) 1165-1186.

[66] E. Hoek, C. Carranza-Torres, B. Corkum, Hoek-Brown failure criterion-2002 edition, Proceedings

of NARMS-Tac, 1 (2002) 267-273.

[67] N. Vlachopoulos, M.S. Diederichs, Improved longitudinal displacement profiles for convergence

confinement analysis of deep tunnels, Rock mechanics and rock engineering, 42 (2009) 131-146.

[68] Itasca, FLAC3D Version 3. Fast Lagrangian Analysis of Continua, 3D Version, Minneapolis,

Minnesota, USA, 2007.

[69] Itasca, FLAC Version 7. Fast Lagrangian Analysis of Continua. 2D Version, Minneapolis,

Minnesota, USA, 2011.

[70] A. Rodriguez-Dono, Studies on undergrond excavations in rock masses, Universidade de Vigo,

2011.
62
[71] K.E. Atkinson, An introduction to numerical analysis, John wiley & sons2008.

[72] S.S. Sastry, Introductory methods of numerical analysis, PHI Learning Pvt, Ltd.: New Delhi, India,

(2012).

[73] M. Bonini, D. Debernardi, M. Barla, G. Barla, The mechanical behaviour of clay shales and

implications on the design of tunnels, Rock mechanics and rock engineering, 42 (2009) 361.

[74] J.Z. Zhang, X.P. Zhou, P. Yin, Visco-plastic deformation analysis of rock tunnels based on

fractional derivatives, Tunnelling and Underground Space Technology, 85 (2019) 209-219.

63