See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/329923313

Tunnel face stability analysis using Kinematical Element Method

Article  in  Tunnelling and Underground Space Technology · December 2018

DOI: 10.1016/j.tust.2018.11.024

CITATIONS READS

0 58

5 authors, including:

Mahmoud Qarmout
Ruhr-Universität Bochum
2 PUBLICATIONS   11 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Local transient face support within hydro-shield tunneling View project

THM considerations View project

All content following this page was uploaded by Mahmoud Qarmout on 07 January 2019.

The user has requested enhancement of the downloaded file.

 Tunnelling and Underground Space Technology 85 (2019) 354–367

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology

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

Tunnel face stability analysis using Kinematical Element Method T

a,⁎ a b c a,1
Mahmoud Qarmout , Diethard König , Peter Gussmann , Markus Thewes , Tom Schanz
a
Ruhr-Universität Bochum, Chair of Foundation Engineering, Soil and Rock Mechanics, Universitätsstr. 150, 44780 Bochum, Germany
b
Am Bächle 3, D 74629 Pfedelbach, Untersteinbach, Germany
c
Ruhr-Universität Bochum, Chair of Tunneling, Utility Engineering and Construction Management, Universitätsstr. 150, 44780 Bochum, Germany

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

Keywords: This paper introduces a new approach to investigate the tunnel face stability. In the proposed approach a 3D
Kinematical Element Method KEM (Kinematical Element Method) model is developed which consists of two rigid blocks, tetrahedron wedge
Support pressure block (lower part) and triangular prism block (upper part). In order to incorporate the influence of the 3D soil
Face stability arching in silo for predicting the minimum support pressure, it is presumed a 3D active earth pressure acting on
Shallow tunnel
the vertical prism’s slip surfaces from the adjoining soil. The results given by the KEM model are compared with
the results of numerical simulations, analytical approaches, and physical model tests available in literature.
Similar to Terzaghi’s superposition method commonly used in bearing capacity analysis, the minimum support
pressure is expressed by summation of cohesion (c), surcharge load (q) and unit weight of the soil (γ ) multiplied
by non-dimensional bearing capacity coefficients Nc , Nq and Nγ . Simple formulas have been developed to
compute the non-dimensional coefficients. The Nγ , Nc , and Nq formulas are presented as function of internal
friction angle (φ ), and cover depth to tunnel diameter ratio (C/D). The proposed KEM model for the tunnel face
stability gives good results in term of minimum support pressure compared to physical model test results, various
analytical and numerical solutions for the tunnel face stability analysis of shallow tunnels.

1. Introduction
One of the major aspects for the mechanized tunneling process is to
adequately support the soil at the tunnel face during tunnel construc-
tion. To prevent the tunnel face from collapse a minimum pressure at
the tunnel face is required. In practice, several analytical methods to
estimate the minimum face support pressure are available. These ana-
lytical methods can be subdivided into two main approaches, Limit
Analysis Method (LAM) and the Limit Equilibrium Method (LEM). Horn
(1961) was the first to present a 3D wedge-silo model. Horn failure
mechanism consists of prismatic wedge and vertical silo above the
wedge. Using Limit Equilibrium Method, Anagnostou and Kovari
(1994), Jancsecz and Steiner (1994), Broere (2001), Anagnostou
(2012), and Chen et al. (2015) proposed 3D tunnel face stability model
based on Horn failure mechanism, including the effect of soil arching in
the silo by using Janssen’s silo theory (1895). In Janssen’s analysis of
soil arching, it is required to have explicit value for the lateral earth
pressure coefficient for the silo (Ksilo ). Furthermore, in order to calcu-
late the shear stress acting on each triangle surface of the wedge, the
lateral earth pressure coefficient (Kwedge ) for the wedge must be as-
sumed in advance. Anagnostou and Kovari (1994) assumed the lateral
earth pressure coefficient for the silo Ksilo = 0.8 and for the wedge
Kwedge = 0.4 . Jancsecz and Steiner (1994) suggested the lateral earth
pressure coefficient for the wedge as Kwedge = (K a + K 0 )/2 with K a
(K a = tan2 (45 − φ /2) ) and K 0 (K 0 = 1 − sin φ ; Jaky, 1944). Broere
(2001) and Anagnostou (2012) studied the layered soil, using the in-
finitesimally slices method in the wedge, described by Walz and Prager
(1978) for slurry-filled trenches. Broere (2001) proposed
Ksilo = Kwedge = K 0 . Anagnostou (2012) assumed Ksilo = 1 and
Kwedge = 0.5. Kirsch and Kolymbas (2005) include the effect of soil
arching in the silo by assuming a quadratic parabolic distribution of
vertical stress above the tunnel crown. Chen et al. (2015) considered
the height of the silo on the basis of physical model tests results, their
results indicated that the average value of Ksilo ≈ 1. Krause (1987) es-
tablished a 3D shell failure mechanism on front of the tunnel face.
The wedge-silo models have also been extended by several re-
searchers considering seepage flow (Perazzelli et al., 2014; Zingg and
Anagnostou, 2016), tunnel face reinforcement (Anagnostou and
Perazzelli, 2015; Anagnostou and Serafeimidis, 2007), excess pore
pressure at the tunnel face (Broere, 2001; Dias and Bezuijen, 2016) and
the influence of slurry infiltration process (Broere and Tol, 2000; Dias
and Bezuijen, 2016).

⁎
Corresponding author.
E-mail address: Mahmoud.Qarmout@rub.de (M. Qarmout).
1
The authors would like to dedicate this work to the memory of Professor Tom Schanz, who passed away during the preparation of this paper.

https://doi.org/10.1016/j.tust.2018.11.024
Received 5 January 2018; Received in revised form 10 October 2018; Accepted 22 November 2018
0886-7798/ © 2018 Elsevier Ltd. All rights reserved.
M. Qarmout et al.
Nomenclature
Notations
C cover depth of the tunnel
c soil cohesion
D diameter of the tunnel
C/D cover depth to diameter ratio
K3D 3D active earth pressure coefficient
K2D 2D active earth pressure coefficient
Based on the review of wedge-silo models cited above, the effect of
the soil arching in the silo is interpreted by adopting Janssen’s silo
theory (1895). For computing the pressure/force acting at the bottom of
the silo using Janssen’s silo theory, three main steps are adopted (1) The
lateral earth pressure coefficient value has to be assumed in advance.
Meanwhile, the lateral earth pressure coefficient has a significant effect
on the minimum support pressure. (2) The lateral earth pressure coef-
ficient has the same value over the silo depth. (3) The shear force be-
tween the silo and the wedge is omitted in equilibrium of the forces.
In the past few years, upper bound theorem of limit analysis has
been gradually applied to assess the stability of tunnel face. Leca and
Dormieux (1990) proposed two mechanisms for the failure zone at the
front of the tunnel face; one is consisted of a single conical block, the
other one is composed of two solid conical wedges with elliptic cross
section. Mollon et al. (2010) modified the 3D failure mechanism de-
veloped by Leca and Dormieux (1990) using several rigid conical blocks
with circular cross-section. Ibrahim et al. (2015) improved the failure
mechanism of Mollon et al. (2010) to compute the minimum support
pressure in dry multilayered purely frictional soil. Senent and Jimenez
(2015) amended the solution of Mollon et al. (2010) to study the pos-
sibility of partial collapse in layered soils. Senent and Jimenez (2015)
model examined the influence of soil properties of the crossed soil and
the cover soil on the minimum support pressure. Park et al. (2007)
included the effect of seepage forces emerging from the groundwater
flow in the upper bound solution.
Generally, the upper bound solution employed one or more trun-
cated conical sliding wedge as the failure mechanism leading to reduce
the self weight and thereby to include 3D soil arching effect in eva-
luation the support pressure.
Vermeer et al. (2002) presented finite element calculations to study
the influence of friction angle and cohesion on the minimum support
pressure. The soil was modeled with elastic, perfectly plastic Mohr-
Coulomb constitutive model. A numerical study with FEM was used by
Kirsch (2009) to model the stability of the tunnel face in sand. Two
different material models were used for his simulation: hypoplasticity
and Mohr-Coulomb soil models. Ohta and Kiya (2001) used Finite
Difference Method (FDM) to examine the influence of the soil properties
and groundwater level on the stability of the tunnel face. They illu-
strated the relationship between support pressure at the tunnel face and
the initial water levels for different values of Youngs modulus of the
soil. Chen et al. (2011) constructed a 3D DEM model to analyze the face
stability of shallow tunnels. The minimum support pressure and geo-
metry of the failure zone were discussed in terms of the process of
tunnel face failure.
In addition to the aforementioned numerical studies, several nu-
merical simulations were conducted by other researchers, through FEM
(e.g., Peila, 1994; Ng and Lee, 2002; Mayer et al., 2003; Sterpi and
Cividini, 2004; Kim and Tonon, 2010, etc.), FDM (e.g., Li et al., 2009;
Dias, 2011; Senent and Jimenez, 2015, etc.) and DEM (e.g., Maynar and
Rodriguez, 2005; Funatsu et al., 2008; Zhang et al., 2010, etc.).
Various physical model tests were conducted to investigate tunnel
face stability and failure mechanism. Large and small scale model test
at 1g, and increased g-level (ng) using a geotechnical centrifuge were
355
Tunnelling and Underground Space Technology 85 (2019) 354–367
i the number of each block
n the number of slip surface
m the vertexes number of the failure surface
pu minimum support pressure
ΔV the relative tangential displacement
x, y, z coordinates of vertex of the failure surface
φ angle of internal friction of the soil
γ dry unit weight of the soil
σv the vertical stress a above the tunnel crown
σh the horizontal stress a above the tunnel crown
performed. Chambon and Corte (1994) performed a series of ng-tests to
determine the minimum support pressure in dry sand. Their results
indicated that the same failure geometry is found with different soil
densities and tunnel depths. Kirsch (2009) studied the evolution of
support pressure in dense and loose sands by using 1g-tests. The results
indicated that the necessary support pressure is independent of initial
density of the soil. Idinger et al. (2011) performed ng-tests using dry
sand. The influence of the overburden was examined for three cover-to-
diameter ratios (e.g. C/D = 0.5, 1.0, and 1.5). They observed that the
failure of sand soil took the form of a narrow chimney extending from
the tunnel face to the ground. The slip surfaces arising from the bottom
the tunnel face propagated at an angle of about 45 + φ /2 to the hor-
izontal until it turned more or less vertical reaching the ground surface.
Chen et al. (2013) conducted a set of 1g-tests with a tunnel of 1 m
diameter for various cover to diameter ratios to investigate the evolu-
tion of soil arching during face failure in dry sand. Their results in-
dicated that the failure zone at the front of the tunnel face was a
chimney-like failure mechanism, the minimum support pressure in-
creases with the increase of the C/D ratio and then tends to be constant
for larger C/D ratio.
Other physical model tests are reported by Kimura and Mair (1981),
Sterpi et al. (1996), Plekkenpol et al. (2006), Vardoulakis et al. (2009),
Juneja et al. (2010), Konig et al. (2012), Berthoz et al. (2012), Hirata
et al. (2013). The results of these tests are used to validate and calibrate
the numerical and the analytical models, as well as provided the re-
levant factors influencing ground movements induced by tunneling.
In this paper, Kinematical element method (KEM) is applied to the
analysis of the tunnel face stability introducing a new three-dimen-
sional (3D) failure mechanism for dry frictional and frictional-cohesive
soil. KEM provides a computer implemented solution for transitional
rigid body failure mechanism including optimization tools to find the
most relevant failure mechanism. It can be applied to the analysis of
bearing capacity, earth pressure as well as slope stability in 2D and 3D.
KEM can be easily adapted to complex soil conditions (soil with pore
water pressure either static or in seepage flow and drained or undrained
soil condition), different geometrical soil profiles (plane, or inclined
ground surface, and horizontal or sloped layers, etc.) and surcharge
loads.
For quantifying the contribution of 3D silo arching in estimating the
support pressure, a feasible approach in calculating the three-dimen-
sional lateral earth pressure coefficient (K3D ) is proposed. Within this
approach, it is assumed a 3D active earth pressure acting on silo slip
surfaces. The K3D value between the silo and the adjoining soil is ac-
curately obtained as a function of friction angle and cover depth ratio.
The horizontal friction force and cohesion between the silo and the
wedge are adopted in predicting the minimum support pressure. To
consider the 3D wedge arching at the front of the tunnel face, the failure
mechanism of the soil is simulated by 3D tetrahedron wedge-shaped
block, sliding along two intersecting faces. The tunnel face is assumed
to be a triangle that has the same cross sectional area as the circular
tunnel. The effect of groundwater is not taken into account in the
present study.
This paper is organized as follows: The fundamental assumptions of
M. Qarmout et al. Tunnelling and Underground Space Technology 85 (2019) 354–367

KEM and the failure mechanism for the stability analysis of the tunnel
face are introduced in Sections 2 and 3, respectively. The calculation
procedure for estimating the 3D lateral earth pressure coefficient is
proposed in Section 4. The results of KEM calculations are displayed
and discussed in Section 5. The minimum support pressure obtained
from the present failure mechanism and the physical model tests in
literature are compared in Section 5.1. A comparison between the re-
sults of the present study and existing approaches is performed in
Section 5.2. Finally, stability design equations are developed for cal-
culating the minimum support pressure and given in Section 6.

2. Kinematical Element Method (KEM)

2.1. Assumptions

Kinematical Element Method (KEM) was developed by Gussmann

(1982). The soil is idealized to follow Mohr-Coulomb failure criterion
(τ = c + σ · tan φ ) with shear strength properties, cohesion (c) and in-
ternal friction angle (φ ). In KEM, the failure mechanism consists of a
limited number of rigid blocks, each block is called kinematic element.
The limit load inducing the collapse is calculated by considering the soil
as rigid perfectly plastic material.
The calculation procedure of the Kinematical Element Method
(KEM) consists of three parts, namely kinematic analysis, static analysis
and optimization process.
Fig. 2. Geometry of the failure mechanism.

2.2. Kinematic analysis

The failure mechanism of the soil is simulated by a mechanism of 3D
rigid blocks (see Fig. 1), separated by contact surfaces. Each block has
its own faces, edges and vertices. These vertices are called kinematic
nodes which are defined by Cartesian coordinates in x, y and z direc-
tions, chosen to be either fixed or free nodes. The rigid blocks can only
slide tangentially along inner and outer contact surfaces. Outer contact
surfaces are located between the blocks and the surrounding soil. Inner
contact surfaces are located between the blocks. In all contact surfaces,
no dilatancy is taken into account and rotation is not allowed. The
movement of the rigid blocks is initiated by a pre-known displacement
(virtual displacement) on a flexible boundary. The flexible boundary is
chosen based on the direction of the external (unknown) force which
causes the failure.
By taking the compatibility conditions for the movable rigid blocks,
the kinematic equation for the system can be written as the follows:
[Kv]j × n ·[V ]̂ n × 1 + [V ]j × 1 = 0 (1)
where n denotes the number of contact surfaces, j denotes the number
of degree of freedoms for displacements of rigid blocks, in which
j = 3 · k (k denotes the number of blocks), [Kv ] is the kinematics matrix
which contains the direction cosine of unit normal vector of each sur-
face with respect to the global coordinates system, [V ]̂ is the vector of
unknown (absolute) displacements values of each block and [V] is the
vector of the pre-known displacement on the flexible boundary which
causes the displacements for the kinematic system.
After solving Eq. (1), the displacements of each block are obtained.
Then, the relative tangential displacement between any adjacent block
or boundary can be evaluated.

Fig. 1. KEM model for 3D passive earth pressure (Gussmann, 1986).

356
M. Qarmout et al. Tunnelling and Underground Space Technology 85 (2019) 354–367

Fig. 3. Geometry of isosceles triangle.

Fig. 6. The free-body diagram for KEM model.

Fig. 4. Kinematic of the blocks.

Fig. 5. Displacement hodograph with respect to the soil at rest O.

2.3. Static analysis

The forces acting in every contact surface (see Fig. 1) are the vector
of normal force Nn , the cohesion force Cn , and Rn = Nn· tan φn for the Fig. 7. Half of KEM model.
friction of the soil. Wi (i denotes the number of each block) is the weight
of each the element.
The direction of shear forces is opposite to the relative tangential
movement of the blocks. Therefore, the direction of the shear force can
be obtained in every contact surface for each kinematic block.

357
M. Qarmout et al. Tunnelling and Underground Space Technology 85 (2019) 354–367

Fig. 8. Shape of the failure mechanism observed in physical model tests and numerical simulation.

The force equilibrium equations of each rigid block along the x, y where [Ks ] is the static coefficient matrix, [N] is the normal vector of
and z-axis, and the resultant (Si ) of all forces is given by unknown forces and [F] is the vector of known forces containing inertia
forces, surface loads and cohesion, etc.
n n n
For more details about kinematic and static analysis in KEM, it is
Si = ( ∑ Six = 0, ∑ Siy = 0, ∑ Siz = 0)
i=1 i=1 i=1 (2) refereed to Gussmann (1986) and Gussmann (2000).

By assembling the force equilibrium equations over all rigid blocks
and representing in matrix form, taking the normal forces acting at the
contact surfaces and the virtual force at the flexible boundary as un-
knowns, we can obtain the following equation:
[Ks]j × n ·[N ]n × 1 + [F ]j × 1 = 0
2.4. Optimization process
The aim of optimization process in KEM is to find the proper geo-
metry of the failure mechanism with respect to specific boundary
conditions. In the case of tunnel face stability analysis (active failure),
(3) maximizing the normal force acting on the tunnel face is used in the

358
M. Qarmout et al.
Fig. 9. Flow chart for the procedure of calculating the support pressure in KEM model.
Fig. 10. KEM model for active earth pressure problem.
objective function. Because the geometry of the block system is ex-
pressed by contact surfaces which are defined by the nodes, the opti-
mization process for the geometry is able to be done in terms of the
coordinates of the nodes. The coordinates of the free kinematic nodes
will be varied until the critical failure mechanism is found. To do so, the
authors have used the algorithm of Particle Swarm Optimization (PSO)
(Kennedy and Eberhart, 1995).
3. KEM model for tunnel face stability
3.1. Geometry of the failure mechanism
In the case of 3D KEM model for the tunnel face stability, the
method of vector analysis provides relatively simple formulations to
359
Tunnelling and Underground Space Technology 85 (2019) 354–367
determine all the relevant features of each block including the volume
of block, the area of the faces and the positions of its vertices. The
Cartesian coordinate system x, y and z is selected as global coordinate
system and defined by orthonormal unit vector {e1, e2, e3} , with e1 = [1,
0, 0], e2 = [0, 1, 0] and e3 = [0, 0, 1]. The failure surfaces are re-
presented by m vertices [P1, P2…Pm], and n surfaces, [F1, F2…Fn], with the
coordinates of nodes [(x1, y1 , z1), (x2 , y2 , z2), …, (x m , ym , z m)].
The sliding (translational) failure mechanism consists of two parts
(see Fig. 2). The soil wedge block (lower part) is enclosed by four
surfaces: The tunnel face, two outer contact surfaces, and the inner
contact surface with the upper silo block. The silo block raised from the
crown of the tunnel to the ground surface (upper part) enclosed by five
surfaces: the ground surface, the inner contact surface with the wedge
block and three vertical outer contact surfaces to the adjacent soil. The
M. Qarmout et al.
Fig. 11. Forces upon to tunnel face stability model.
positions of the nodes (P1, P2, P3, P6, P7) in the Cartesian coordinate
system are chosen depending on the cover (C) and diameter (D) of the
tunnel. The degree of freedom for each node during the optimization
process to find the relevant geometry of the failure mechanism is
chosen as follows:
1. The nodes P1, P2, P3, P6 and P7 are fixed in x, y and z direction.
2. The node P4 is fixed in y, and free in x and z direction.
3. The node P5 is fixed in y and z direction and is bounded to the node
P4 in x direction.
3.2. Equivalent area of the tunnel face
The tunnel face is approximated by an isosceles triangle (see Fig. 3),
with the same area like the tunnel face. The sides of triangle are cal-
culated as follows:
a = 0.25·π·D
b= D 2 + a2
where D is the diameter of tunnel, b is the side of equilateral triangle.
3.3. Kinematic analysis
The kinematic process will start by initiating a unit displacement υ
(virtual displacement) on the face of the tunnel V = −υe1 (see Fig. 4).
Each block i (i is the number of blocks) moves with a global displace-
ment Vi = [Vix , Viy, Viz ], with respect to the soil at rest(O). The soil at rest
is considered as an element with a zero displacement VO = [0, 0, 0]. The
relative displacement of each block is determined by a hodograph, as
shown in Fig. 5.
The directions of the relative tangential displacements between any
adjacent blocks and between the blocks with surrounding soil are cal-
culated as follow:
ΔVn (2 − 1) = V(2) − V(1)
ΔVn (i − O) = V(i) − V(O)
It must be noted that υ is a virtual displacement and not a real
displacement, the value of the relative displacements do not influence
on the results and only necessary to get the shear forces directions.
Tunnelling and Underground Space Technology 85 (2019) 354–367
3.4. Acting forces
Fig. 6 presents the internal and the acting forces on each block for
KEM model. These forces are defined as follow:
1. Active forces
(a) Wi : The weight of each block (i is the number of blocks).
(b) Q: The surcharge load acting on the ground surface.
2. Resistance forces
(a) Nn : The normal reactions force on each of the slip surfaces.
(b) Rn = Nn· tan φn : The shear force for each of the slip surfaces (n is
the number of slip surface).
(c) Cn : The cohesion force on each of the contact surfaces.
3. P: The support force on the tunnel face.
3.5. Problem of indeterminacy
(4) In simplified method, only half of KEM model is modeled due to the
symmetry of the failure mechanism (see Fig. 7) according to the fol-
(5) lowing assumption:
1. No displacement in Y direction over the symmetric plane.
2. No shear forces along the symmetric plane.
3. The normal forces on the symmetric plane are taken into count.
the static system-matrix for the KEM model is formulated as follows:
[Ks]6 × 7 ·[N ]7 × 1 + [F ]6 × 1 = 0 (8)
From Eq. (8), the system is statically indeterminate, the number of
unknowns is greater than the number of available equations. Hence,
equilibrium equations alone are insufficient for obtaining a solution. In
order to manage the possible solution, using the physical model tests
and the numerical results, one can predict the lateral pressures imposed
on the silo leading to estimate the value of one unknown force of the
system, which is discussed in the following section.
(6)
3.6. Setting up the static system
(7)
According to the observations from physical model tests (Chambon
and Corte, 1994; Messerli et al., 2010; Chen et al., 2013), for lower
cover to diameter ratio (e.g., C/D = 0.5) a silo mechanism forms di-
rectly above the tunnel face and a sinkhole develops at the ground
360
M. Qarmout et al.
Fig. 12. Comparison of K3D value obtained from the other methods with KEM model.
surface (global failure), see Fig. 8a. However at a more deeper cover
depth (e.g., C/D = 2) a bulb-shaped failure zone is observed above the
tunnel face (local failure), where there is no obvious settlement on
ground surface. Moreover, the outcomes of the physical model tests
demonstrate that the soil mass bounding the silo mechanism moves
slightly in the direction to the silo, allowing the soil mass to expand
horizontally. With this freedom movement, the small movement of the
surrounding soil allows a reduction in the horizontal stress acting on
the silo. This reduction of horizontal stress in the soil surrounding the
silo is observed in the physical model tests (Chen et al., 2013) and
numerical results (Chen et al., 2011), as shown in Figs. 8b, c and d.
Clearly, the moving soil mass inside the silo must support the sur-
rounding soil mass by a minimum pressure to avoid extending the
collapse. With the relaxation in the surrounding soil mass, the hor-
izontal stress around the silo is less than at rest (Chen et al., 2011,
2013). Besides, the horizontal stresses around the silo cannot drop
below their active values. Based on the measurements of earth pressures
and failure patterns of tunnel face revealed by physical model tests
361
Tunnelling and Underground Space Technology 85 (2019) 354–367
Chen et al. (2013), it could be observed that the surrounding soil mass
is brought into a state of limiting active “plastic”. Due to this, it is as-
sumed that a 3D active earth pressure acts on the outer contact surface
F6 (P1, P2, P6, P7) (see Fig. 6) of the silo which is located above the
tunnel face. This assumption solves the problem of indeterminacy. In
addition, the influence of this assumption for the forces acting on the
two other out contact surfaces F5 (P1, P4, P5, P7) and F7 (P4, P5, P8, P9)
will be discussed later.
3.7. Optimization calculation procedure
Searching for critical failure mechanism is a problem of max-
imization. Fig. 9 presents the flow chart for optimization algorithm of
calculating minimum support pressure. The iterative optimization al-
gorithm is started by specifying input data for soil parameters (φ , γ , and
c), diameter of the tunnel (D) and the cover depth (C), also by defining
surcharge load acting on ground surface (Q), see Fig. 6.
After setting up initial input data, the cover depth (C) and the half
M. Qarmout et al.
Fig. 13. Comparison of experimental and numerical results with KEM model,
φ = 31.7°.
Fig. 14. 3D lateral earth pressure coefficient (K3D ) for different friction angles
(φ ).
length between nodes (P1, P2) (0.5B) are switched from the subroutine
“tunnel face stability” to the subroutine “3D active earth pressure”, as
shown in Figs. 10 and 11. Thereafter, randomly selected values for the
x-coordinate of the nodes P10, and P11 as well as for y-coordinate of
node P10 are initiated. An iterative procedure using PSO is im-
plemented to the subroutine “3D active earth pressure” to find the
maximum normal force (0.5N6 ) for 3D active earth pressure failure.
The value of maximum normal (0.5N6 ) force is transformed to the
subroutine “tunnel face stability” acting at the slip surface between the
nodes (P1, P7, P8, P9). Subsequently, in the subroutine “tunnel face
stability”, the value of the shear force at the slip surface between the
nodes (P1, P7, P8, P9) is computed as 0.5N6·tanφ. Then, from the static
equilibrium equations, the support pressure is calculated using Eq. (3).
To continue searching for the maximum support force in the subroutine
“tunnel face stability”, new coordinates for free nodes (nodes P4, and
P5) are chosen based on their current position finding the most ade-
quate failure mechanism described by the x-coordinate of the node P4
and P5 and by z-coordinate of the node P4. The optimization procedure
is terminated, when PSO in the subroutine “tunnel face stability”
reaches termination criteria, that can be either when a predefined
maximum number of iterations is reached or achieving a constant value
362
Tunnelling and Underground Space Technology 85 (2019) 354–367
of support pressure for a certain number of iterations. Finally, the
minimum support pressure is obtained.
4. 3D lateral earth pressure coefficient
During excavation of the tunnel, the stress history of soil mass
surrounding excavation area is disturbed. Thereafter, a new stress re-
distribution is induced in the vicinity of tunnel face due to deformation
of the soil mass. The mechanism that leads to this stress redistribution is
so-called the arching phenomena (Terzaghi, 1943). This mechanism
yield a reduction of vertical stress. The generation of soil arching effect
relies on the shear strength of the soil. Therefore, with certain shearing
resistance, the soil mass can always generate stable arching action. The
arching soil effect has been reported by Terzaghi (1936) in trap door
experiments. Also, by Chambon and Corte (1994) and Chen et al.
(2011) conducting a series of physical model tests to investigate the
stability of tunnel face. As well as by Chee et al. (2011) based on nu-
merical simulation (FDM).
One of the main aspects that has to be implemented in estimating
the minimum support pressure is the arching effect in the silo. The
evaluation of arching effect is described by the lateral stress ratio (the
so-called lateral earth pressure coefficient) between the horizontal
stresses (σh ) and the vertical stresses (σv ) (Terzaghi, 1943).
Within this paper a new calculation procedure for estimating 3D
lateral earth pressure coefficient (K3D ) with the consideration of soil
arching is proposed. In the following part, the results of calculations are
presented and discussed. The load acting from the adjoining soil of the
silo is considered as a 3D active earth pressure exerting on slip surface
of the silo (see Fig. 11). The 3D lateral earth pressure coefficient be-
tween the silo and the adjoining soil is adopted equal to the calculated
3D active earth pressure coefficient.
The horizontal normal force of 3D active earth problem is defined
by:
N6 = 0.5·K3D·γ ·C 2·B (9)
Therefore, the calculated 3D active earth pressure coefficient (K3D ) is
N6
K3D =
0.5·γ ·C 2·B (10)
Fig. 12(a)–(c) shows the comparison of the present solution with the
methods of Huder (1972) derived from the silo theory, Walz and Prager
(1978) based on the theory of element slices, and Washbourne (1984)
modified the shape of the sliding wedge, assuming a combined planar
ends with the wedge to approximate the curved geometry observed for
trench failure mechanism. Walz and Prager (1978) suggested the value
of earth pressure coefficient (Ky ) between the range K c ⩽ Ky ⩽ K o ,
where K 0 = 1 − sinφ and K c = cos 2 φ . Fig. 12(a)–(c) represents the
value of the K3D /K2D ratio as a function of C/B ratio (see Fig. 10).
Where the 2D active earth pressure coefficient (K2D ) is calculated as
follows:
1 − sinφ
K2D =
1 + sinφ (11)
The results indicate that K3D /K2D ratio decreases with increasing C/B
ratio. It can also be observed that K3D /K2D ratio for a specific C/B ratio
decreases with increasing friction angle of the soil. Also, the KEM model
gives almost identical or slightly greater results than the results ob-
tained using the theory of Walz and Prager (1978) (Ky = cos 2 φ ), which
are less than the value obtained by Huder (1972), Walz and Prager
(1978) (Ky = 1 − sinφ), the results from Washbourne (1984) are quite
different compared to those of the present paper.
In Fig. 13, the small scale test results and the numerical model
performed by Worden and Achmus (2013) are used for the validation of
KEM model. The best agreement with the numerical simulation and
experimental results is obtained with KEM model and Walz and Prager
(1978) (Ky = cos 2 φ ) results.
M. Qarmout et al. Tunnelling and Underground Space Technology 85 (2019) 354–367

Fig. 15. Iteration steps in the optimization process.

Table 1
Selected physical model tests for verification of KEM model.
Author Model Tested material φ [°] c [kPa] C/D [–]

Chen et al. (2013) 1g test Yangtze River sand 37 0.5 0.5, 1, 2

Chambon and Corte (1994) ng test Fontainebleau sand 38–42 0–5 0.5, 1, 2, 4
Kirsch (2009) 1g test Ottendorf-Okrilla sand 32.5 0 0.5, 0.75, 1, 1.5, 2

The values of the 3D lateral earth pressure coefficient as a function surface F5 (P1, P4, P5, P7) of the silo part is about 0.38, which is close to
of C/B ratio with different friction angles are presented in Fig. 14. the 3D active earth pressure value calculated by KEM for the surface F6
Suggested equation for calculating K3D as a function of C/B ratio and φ (P1, P7, P8, P9). For all geometrical parameters as well as for the
has been developed as follows: support pressure the values have been fastened after about 50 itera-
tions, starting from specific initial geometries of course shape of the
2·e−0.054·φ
K3D = curve and number of iteration steps to reach final values depends on the
1.75 + C / B − (0.025·φ) (12)
initial geometry.

5. Support pressure on the tunnel face 5.2. Verification by physical model tests

5.1. Development of failure mechanism by optimization process
The development of the failure mechanism, the minimum support
pressure, and the lateral earth pressure coefficient at surface F5 (P1, P4,
P5, P7) during the iteration steps process is shown in Fig. 15. For one
example, it can be concluded that the inclination of the lower edge (β )
of the wedge is about β = 55°, which is close to the active slip surface
angle for 2D active state Rankine’s theory (Rankine, 1857)
(θa = 45 + φ /2 ). The lateral earth pressure coefficient acting on the
Three series of physical model tests from literature have been
chosen to verify the accuracy of KEM model for predicting the
minimum support pressure. The specifications of the physical model
tests are summarized in Table 1.
Fig. 16 presents the comparison of normalized support pressure
( pu / γD ) calculated by the KEM model with the results from Chambon
and Corte (1994), Kirsch (2009) and Chen et al. (2013).
From Fig. 16, we can see the normalized support pressure obtained
from ng tests Chambon and Corte (1994), and 1g tests Kirsch (2009)

363
M. Qarmout et al.
Fig. 16. Comparison of pu / γD value obtained from the KEM model with phy-
sical model tests.
Fig. 17. Comparison of Nγ value obtained from the theoretical methods using
different friction angles with KEM model.
increases slightly with the increase in C/D ratio as well as KEM model
show that C/D ratio has slight effect on normalized support pressure.
This dependency of C/D ratio was also detected by Anagnostou and
Kovari (1994, 1996), Chen et al. (2011), and Chen et al. (2013).
The values of normalized support pressure calculated by the KEM
model show a good agreement to the results obtained by Kirsch (2009).
Furthermore, when KEM model results compared with normalized
support pressure values obtained from Chen et al. (2013), the difference
between the value of pu / γD from KEM model and 1g tests of Chen et al.
(2013) varies between 6.5% and 16.4%.
The values of pu / γD obtained from ng tests (Chambon and Corte,
1994) are less than the results predicted by KEM model in cohesionless
soil (c = 0 kPa). The reason for this variance is that the soil used in ng
tests showed a little cohesion (c = 0–5 kPa). However, if the cohesion is
considered in the calculation of the support pressure (c = 5 kPa), see
Fig. 16, the KEM model predicts the minimum support pressure much
more closer to 1g test results (Chambon and Corte, 1994).
As a conclusion, it can be inferred that the minimum support
364
Tunnelling and Underground Space Technology 85 (2019) 354–367
Fig. 18. The relationship between pu / γD and C/D obtained from the different
approaches and KEM model.
Fig. 19. Comparison of Nc value obtained from the theoretical methods using
different friction angles with KEM model.
pressure predicted by KEM model approximate the physical model tests
results very well.
5.3. Comparison of the KEM results with other approaches
Analogous to the method proposed by Terzaghi (1943) for bearing
capacity analysis, the minimum support pressure ( pu ) is represented by
the following form (Vermeer et al., 2002; Mollon et al., 2010;
Anagnostou, 2012):
pu = γ ·D ·Nγ − c·Nc + q·Nq (13)
where the contribution of different loads and soil parameters including
self-weight (γ ), internal friction angle (φ ), surface surcharge (q) and
cohesion (c) are expressed by the non-dimensional bearing capacity
coefficients Nγ , Nq and Nc as a function of the friction angle of the soil.
Due to the active conditions around the tunnel face the cohesion is
reducing the necessary support pressure.
Within this paper, the above formula is adopted to express the
minimum support pressure. In the calculations with c = 0 kPa and
q = 0 kPa, the soil unit weight coefficient (Nγ ) is obtained. Also, setting
M. Qarmout et al.
Fig. 20. The value of Nγ as function of the soil friction angle.
Fig. 21. The value of Nc as function of the soil friction angle.
Fig. 22. The value of Nq as function of the soil friction angle.
365
Tunnelling and Underground Space Technology 85 (2019) 354–367
γ = 0 kN/m3 and q = 0 kPa, the cohesion coefficient (Nc ) is calculated.
In addition, with γ = 0 kN/m3 and c = 0 kPa, the surcharge load
coefficient (Nq ) can also be obtained.
To validate results of the KEM model for the stability of the tunnel
face, the non-dimensional coefficients Nγ and Nc were examined in-
dividually by KEM model and compared with the results of other ap-
proaches.
Fig. 17 compares the value of the non-dimensional coefficient Nγ
calculated by KEM model and those given by other researchers as a
function of the friction angle, for C/D ratio equal to 1.
We can see from Fig. 17, the numerical results of Vermeer et al.
(2002), as it obtained by FEM calculations and the modified wedge
model by Anagnostou (2012) which is based on the infinitesimally thin
slices method predict the value of Nγ as well as the numerical results of
Mollon et al. (2010) using the upper bound method much more closer
to KEM results than the other models. Anagnostou (2012) method and
KEM model are taking into account the contribution of horizontal arch
action over the tunnel face, which leads to similar results calculated by
both method. The results given by Krause (1987) are above the KEM
results. In Krause model the soil conditions above the tunnel face are
not taken into account in calculating the support pressure. The solution
of Anagnostou and Kovari (1994), using a limit equilibrium method
overestimates the value of Nγ . This is due to the simplified way of
considering the vertical stress distribution (linear distribution) which
disregarding the arching effect in calculating the horizontal stresses at
the slip surfaces of the wedge. The Nγ values obtained from Leca and
Dormieux (1990) using the upper bound method are clearly below the
results given by the others, this due to that the shape of the tunnel face
is considered as an elliptic cross section inscribed to the circular face
(the elliptic cross section and the circular tunnel face do not have the
same area). Also, the two-blocks mechanism given by Leca and
Dormieux (1990) is constrained by the normality condition, required by
plasticity theory, the velocity vector must make an angle φ with dis-
continuity surfaces along the sliding surfaces. However, this condition
does not allow the three-dimensional slip surfaces to develop more
freely.
Also we can see from Fig. 18, KEM results show that pu / γD slightly
increase with the increase of C/D, similar results are also obtained by
the wedge-silo models (Anagnostou and Kovari, 1994; Chen et al.,
2015) which is also confirmed by the results of DEM simulations (Chen
et al., 2011). While the results of the upper bound solution (Leca and
Dormieux, 1990; Mollon et al., 2010), FEM calculations (Vermeer et al.,
2002) and Krause (1987) model show that C/D has no effect on pu / γD .
As can be seen in Fig. 19 where C/D ratio is 2, the results of Vermeer
et al. (2002) and Anagnostou (2012) are the same. The non-dimen-
sional coefficients Nc obtained by KEM are slightly lower than the so-
lution of Vermeer et al. (2002), Anagnostou (2012), and Mollon et al.
(2010). However, Krause’s results are extremely conservative.
As shown in Figs. 17 and 19, the calculated support pressure based
on KEM model replacing the circular shape of the tunnel face by a
isosceles triangle of equivalent area, Anagnostou (2012) substituting
the circular shape of the tunnel face with a square area and Vermeer
et al. (2002) as well as Mollon et al. (2010) using a circular shape of
equivalent area are very close. Therefore, we can conclude that the
shape of equivalent area of the tunnel face has little influence on the
minimum support pressure. In addition, from Figs. 15 and 16, the va-
lues of Nγ and Nc obtained from KEM model increase obviously with the
decrease in the friction angle, which is also revealed by Krause (1987),
Vermeer et al. (2002), and Anagnostou (2012).
6. Development of design equations
Figs. 20 and 21 show the relation of the non-dimensional coeffi-
cients Nγ and Nc with different friction angles and C/D ratios. The re-
sults from the curves Nγ − φ and Nc − φ show that as the friction angle
decreases, the effect of C/D ratio becomes more prominent and the rate
M. Qarmout et al. Tunnelling and Underground Space Technology 85 (2019) 354–367

Table 2
The values of Nγ , Nc and Nq for various angles of internal friction and C/D ratios.

Nγ Nc Nq
C/D

φ (degrees) 0.5 1 1.5 2 2.5 0.5 1 1.5 2 2.5 0.5 1 1.5 2 2.5

15° 0.345 0.414 0.457 0.506 0.536 3.100 3.371 3.784 4.203 4.811 0.235 0.091 0.000 0.000 0.000
20° 0.242 0.276 0.304 0.329 0.356 2.398 2.590 2.820 3.099 3.373 0.150 0.052 0.000 0.000 0.000
25° 0.171 0.191 0.202 0.219 0.233 1.950 2.067 2.161 2.310 2.485 0.100 0.025 0.000 0.000 0.000
30° 0.126 0.139 0.147 0.155 0.163 1.601 1.689 1.770 1.814 1.940 0.060 0.000 0.000 0.000 0.000
35° 0.093 0.102 0.108 0.114 0.121 1.330 1.382 1.442 1.501 1.580 0.030 0.000 0.000 0.000 0.000
40° 0.067 0.073 0.078 0.082 0.088 1.092 1.135 1.181 1.232 1.290 0.000 0.000 0.000 0.000 0.000
45° 0.047 0.052 0.057 0.060 0.065 0.891 0.928 0.971 1.02 1.080 0.000 0.000 0.000 0.000 0.000
50° 0.037 0.041 0.045 0.048 0.052 0.741 0.771 0.802 0.841 0.891 0.000 0.000 0.000 0.000 0.000

of growing for Nγ − φ and Nc − φ curves is higher. the supporting pressure applied to the excavated tunnel face.
The results of Nq − φ in Fig. 22 and Table 2 demonstrate that the
value of Nq becomes equal to zero for C/D ratio greater than or equal to The main conclusions drawn from the proposed KEM model are:
1.5 with φ ⩾ 15°. In addition, the value of Nq becomes equal to zero
when friction angle reaches φ = 40°, for C/D = 0.5. As well as, when 1. The support pressure calculated by KEM model, published data from
φ = 30° for C/D = 1. analytical or numerical models and experiments show good agree-
An approximated equations based on fitting the results of Figs. 20, ments, indicating a proper validation of the KEM model for φ ⩾ 15°.
21 for calculating Nγ , Nc and Nq have been developed as follows: Also, in the case of lower values of the internal friction angle
Nγ ≈ a1·(tanφ)−b1 (φ ⩾ 15°) (φ < 30°), the results of KEM model indicate notable dependence of
(14)
the minimum support pressure on the C/D ratio.
a1 = 0.055 + (0.007·C / D) (15) 2. As a result of searching the minimum support pressure in optimizing
the failure mechanism, the sliding surface between the silo block
b1 = 1.50·(C / D − 0.37)0.044 (16) (upper part) and the wedge block (lower part) resembling a sloped
Nc ≈ (tanφ)−b2 (φ ⩾ 15°) (17) line instead of a horizontal line. The inclination of sloped line de-
pends on the soil strength parameters, as well as the geometry of the
b2 = 0.77 + (0.17·C / D) (18) tunnel.
3. The results indicate that for every surface of the silo part, the values
Nq ≈ e−(b3·tanφ) (For C/ D⩽ 1. 0) (19) of the lateral earth pressure coefficients are very close. A new
b3 = 1.55 + (0.22·C / D) (20) equation for calculating the lateral earth pressure coefficient be-
tween the silo and the adjoining soils which reflects the soil arching
For practical purpose Table 2 provides the values of Nγ , Nc and Nq effects is also proposed.
for different ranges of friction angle and C/D ratio. The values of Nγ , Nc 4. As a future work, a modification in the failure mechanism of KEM
and Nq can be used in Eq. (9) for a straightforward calculation of the model could be implemented by increasing the number of blocks.
minimum support pressure applied to the tunnel face. Also, the proposed KEM model can be modified to investigate the
stability of tunnel face in undrained conditions or to consider see-
7. Summary and conclusions page forces.

In this article a numerical model is presented using KEM to estimate Acknowledgments

the minimum support pressure to stabilize the tunnel face for dry
frictional and frictional-cohesive soil. The main features and assump-
tions for the new KEM model are as follows:
1. The proposed 3D failure mechanism consists of two rigid blocks,
tetrahedron wedge block, and triangular prism block. The tunnel
face is assumed to be a triangle having the same area of the circular
tunnel.
2. The horizontal friction force and cohesion between the silo and the
wedge is adopted in predicting the minimum support pressure.
3. Based on the measurements of earth pressures above the tunnel
crown reported by physical model tests (Chen et al., 2013), a 3D
active earth pressure is assumed acting on the vertical prism’s slip
surfaces from the adjoining soil. The 3D active earth pressure re-
flects the evolution of soil arching for redistribution of horizontal
stresses in the surrounding soil of the silo as a function of friction
angle and C/D ratio. This evolution of soil arching is quite easily
applied to the failure mechanism, and does not need further as-
sumptions.
4. Design equations have been proposed to compute the non-dimen-
sional coefficients, Nc , Nq and Nγ . These equations can be used in
practice (for preliminary design) and enable a quick calculation for
This research has been supported by the German Research
Foundation (DFG) through the Collaborative Research Center (SFB
837). This support is gratefully acknowledged.
References
Anagnostou, G., 2012. The contribution of horizontal arching to tunnel face stability.
Geotechnik 35, 34–44.
Anagnostou, G., Kovari, K., 1994. The face stability of slurry-shield-driven tunnels. Tunn.
Undergr. Space Technol. 9, 165–174.
Anagnostou, G., Kovari, K., 1996. Face stability conditions with earth-pressure-balanced
shields. Tunn. Undergr. Space Technol. 11, 165–173.
Anagnostou, G., Perazzelli, P., 2015. Analysis method and design charts for bolt re-
inforcement of the tunnel face in cohesive-frictional soils. Tunn. Undergr. Space
Technol. 47, 162–181.
Anagnostou, G., Serafeimidis, K., 2007. The dimensioning of tunnel face reinforcement.
In: Underground Space the 4th Dimension of Metropolises, Prague 1, pp. 291–296.
Berthoz, N., Branque, D., Subrin, D., Wong, H., Humbert, E., 2012. Face failure in
homogeneous and stratified soft ground: theoretical and experimental approaches on
1g EPBS reduced scale mode. Tunn. Undergr. Space Technol. 30, 25–37.
Broere, W., 2001. Tunnel Face Stability and New CPT Application. PhD thesis. Delft
University Press, Netherlands.
Broere, W., Tol, V., 2000. Influence of infiltration and groundwater flow on tunnel face
stability. Geotech. Aspect. Undergr. Constr. Soft Ground, Tokyo, Japan 339–344.
Chambon, P., Corte, J., 1994. Shallow tunnels in cohesionless soil: stability of tunnel face.
J. Geotech. Eng. 120, 1148–1165.

366
M. Qarmout et al.
Chee, C., Wen, H., Cheng, T., 2011. Stress redistribution and ground arch development
during tunneling. Tunn. Undergr. Space Technol. 26, 228–235.
Chen, R., Tang, L.J., Ling, D.S., Chen, Y., 2011. Face stability analysis of shallow shield
tunnels in dry sandy ground using the discrete element method. Comput. Geotech.
38, 187–195.
Chen, R.P., Li, J., Kong, L.G., Tang, L.J., 2013. Experimental study on face instability of
shield tunnel in sand. Tunn. Undergr. Space Technol. 33, 12–21.
Chen, R.P., Tang, L.J., Yin, X.S., Chen, Y.M., Bian, X.C., 2015. An improved 3d wedge-
prism model for the face stability analysis of the shield tunnel in cohesionless soils.
Acta Geotech. 10, 683–692.
Dias, D., 2011. Convergence-confinement approach for designing tunnel face reinforce-
ment by horizontal bolting. Tunn. Undergr. Space Technol. 26, 517–523.
Dias, T., Bezuijen, A., 2016. A different view on TBM face equilibrium in permeable
ground. In: ITA World Tunnel Congress-Uniting an Industry, at San Francisco, USA.
Funatsu, T., Hoshino, T., Sawae, H., Shimizu, N., 2008. Numerical analysis to better
understand the mechanism of the effects of ground supports and reinforcements on
the stability of tunnels using the distinct element method. Tunn. Undergr. Space
Technol. 23, 561–573.
Gussmann, P., 1982. Kinematical elements for soils and rocks. In: Proceedings of 4th
International Conference on Numerical Methods in Geomechanics. Edmonton 1, pp.
9–21.
Gussmann, P., 1986. Kinematical element method for 3d problems in geomechanics. In:
Proc. ECONMIG, Stuttgart 2..
Gussmann, P., 2000. Effective KEM solutions for the limit load and the slope stability
problem. Numer. Anal. Methods Geomech. 24, 1061–1077 (John Wiley & Sons).
Hirata, R., Yashiro, K., Haga, Y., Ueno, H., Asakura, T., 2013. The effects of long face bolts
on face stability in the squeezing ground. In: Anagnostou, G., Ehrbar, H. (Eds.),
Proceedings of ITAWorld Tunnel Congress, Underground The Way to the Future.
Taylor & Francis Group, London, pp. 77–684.
Horn, N., 1961. Horizontal earth pressure on the vertical surfaces of the tunnel tubes. In:
National Conference of the Hungarian Civil Engineering Industry, Budapest, pp.
7–16.
Huder, J., 1972. Stability of bentonite slurry trenches with some experiences in swiss
practice. In: Proceedings 5th European Conference on Soil Mechanics and Foundation
Engineering. Madrid, pp. 517–522.
Ibrahim, E., Soubra, A., Mollon, G., Raphael, W., Dias, D., Reda, A., 2015. Three-di-
mensional face stability analysis of pressurized tunnels driven in a multilayered
purely frictional medium. Tunn. Undergr. Space Technol. 49, 18–34.
Idinger, G., Aklik, P., Wu, W., Borja, R., 2011. Centrifuge model test on the face stability
of shallow tunnel. Acta Geotech. 6, 105–117.
Jaky, J., 1944. The coefficient of earth pressure at rest. J. Soc. Hung. Architects Eng. 78,
355–358.
Jancsecz, S., Steiner, W., 1994. Face support for a large mix-shield in heterogeneous
ground conditions. In: Proocreedings 7th International Symposium Tunnelling,
London: 531550, pp. 189–195.
Janssen, H., 1895. Versuche uber getreidedruck in silozellen. Zeitschrift des Vereines
deutscher Ingenieure 39, 1045–1049.
Juneja, A., Hedge, A., Lee, F., Yeo, C., 2010. Centrifuge modelling of tunnel face re-
inforcement using forepoling. Tunn. Undergr. Space Technol. 25, 526–542.
Kennedy, J., Eberhart, R., 1995. Particle swarm optimization. In: Proc of IEEE
International Conference on Neural Network, Perth, Australia, IEEE Service Center,
Piscataway N.J. pp. 1942–1948.
Kim, S.H., Tonon, F., 2010. Face stability and required support pressure for TBM driven
tunnels with ideal face membrane-drained case. Tunn. Undergr. Space Technol. 25,
526–542.
Kimura, T., Mair, R., 1981. Centrifugal testing of model tunnels in soft soil. In: Proc. 10th
Int. Conf. Soil Mech. Found. Engng., Stockholm 1, pp. 319–322.
Kirsch, A., 2009. On the face stability of shallow tunnels in sand. Advances in geo-
technical engineering and tunnelling No. 16, Logos, Berlin.
Kirsch, A., Kolymbas, D., 2005. Theoretische untersuchung zur ortsbruststabilitt.
Bautechnik 82, 449–456.
Konig, D., Detert, D., Schanz, T., 2012. Simulation of soil movement in geotechnical
367
View publication stats
Tunnelling and Underground Space Technology 85 (2019) 354–367
centrifuge testing deep excavations, tunnelling, deposit. Geotech. Eng. J. SEAGS
AGSSEA 45, 12–21.
Krause, T., 1987. Schildvortrieb mit flssigkeits – und erdgesttzter ortsbrust. ph.d. thesis.
Technischen Universitt Carolo-Wilhelmina, Braunschweig.
Leca, E., Dormieux, L., 1990. Upper and lower bound solutions for the face stability of
shallow circular tunnels in frictional material. Geotechnik 40, 581–606.
Li, Y., Emeriault, F., Kastner, R., Zhang, Z., 2009. Stability analysis of large slurry shield-
driven tunnel in soft clay. Tunn. Undergr. Space Technol. 24, 472–481.
Mayer, P., Hartwig, U., Schwab, C., 2003. Standsicherheitsuntersuchungen der ortsbrust
mittles bruchkrpermodell und fem. Bautechnik 80, 452–467.
Maynar, M., Rodriguez, L., 2005. Discrete numerical model for analysis of earth pressure
balance tunnel excavation. J. Geotech. Geoenviron. Eng. 131, 1234–1242.
Messerli, J., Pimentel, E., Anagnostou, G., 2010. Experimental study into tunnel face
collapse in sand. In: In: Springman, Laue, Seward (Eds.), Physical Modelling in
Geotechnics, vol. 1. pp. 575–580.
Mollon, G., Dias, D., Soubra, A., 2010. Face stability analysis of circular tunnels driven by
a pressurized shield. J. Geotech. Geoenviron. Eng. 136, 215–229.
Ng, W., Lee, G., 2002. A three-dimensional parametric study of the use of soil nails for
stabilising tunnel faces. Comput. Geotech. 29, 673–697.
Ohta, T., Kiya, H., 2001. Experimental study and numerical analysis on stability of tunnel
face in sandy ground. Quart. Rep. Railway Tech. Res. Inst. Jpn. 42, 156–160.
Park, J., Tanner, J., Jae, H., 2007. Upper bound solutions for tunnel face stability con-
sidering seepage and strength increase with depth. In: Underground Space - the 4th
Dimension of Metropolises, London, pp. 1217–1222.
Peila, D., 1994. A theoretical study of reinforcement influence on the stability of a tunnel
face. Geotech. Geol. Eng. 12, 145–168.
Perazzelli, P., Leone, T., Anagnostou, G., 2014. Tunnel face stability under seepage flow
conditions. Tunn. Undergr. Space Technol. 43, 459–469.
Plekkenpol, J., Schrier, J., Hergarden, H., 2006. Shield tunnelling in saturated sand – face
support pressure and soil deformations. In: Bezuijen, A., van Lottum, H. (Eds.),
Tunnelling: A Decade of Progress, GeoDelft, pp. 1995–2005.
Rankine, W.J.M., 1857. On the stability of loose earth. Philos. Trans. R. Soc. London 1,
9–27.
Senent, S., Jimenez, R., 2015. A tunnel face failure mechanism for layered ground,
considering the possibility of partial collapse. Tunn. Undergr. Space Technol. 47,
182–192.
Sterpi, D., Cividini, A., 2004. A physical and numerical investigation on the stability of
shallow tunnels in strain softening media. Rock Mech. Rock Eng. 37, 277–298.
Sterpi, D., Cividini, A., Sakurai, S., Nishitake, S., 1996. Laboratory model tests and nu-
merical analysis of shallow tunnels. In: Proceedings of the International Symposium
on Eurock 96 - ISRM, Torino, Balkema, Rotterdam 1, pp. 689–696.
Terzaghi, K., 1936. Stress distribution in dry and saturated sand above a yielding trap-
door. In: Proceedings, First Int. Conf. on Soil Mechanics and Foundation Engineering,
Cambridge, MA, pp. 307–311.
Terzaghi, K., 1943. Theoretical Soil Mechanics. Wiley and Sons, New York.
Vardoulakis, P., Stavropoulou, M., Exadaktylos, G., 2009. Sandbox modeling of the
shallow tunnel face collapse. Riv. Geotec. 1, 9–21.
Vermeer, P.A., Ruse, N.M., Marcher, T., 2002. Tunnel heading stability in drained ground.
Felsbau 20, 818.
Walz, B., Prager, J., 1978. Der nachweis der ueren standsicherheit suspensionsgesttzter
erdwnde nach der elementscheibentheorie. Verffentlichungen des Grundbauinstituts
der Technischen Universitt Berlin, Heft 4..
Washbourne, J., 1984. The three-dimensional stability analysis of diaphragm wall ex-
cavations. Ground Eng. 17, 24–29.
Worden, F., Achmus, M., 2013. Numerical modeling of three-dimensional active earth
pressure acting on rigid walls. Comput. Geotech. 5, 83–90.
Zhang, Z.X., Hu, X.Y., Scott, D., 2010. A discrete numerical approach for modeling face
stability in slurry shield tunnelling in soft soils. Comput. Geotech. 94–104.
Zingg, S., Anagnostou, G., 2016. An investigation into efficient drainage layouts for the
stabilization of tunnel faces in homogeneous ground. Tunn. Undergr. Space Technol.
58, 49–73.