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

net/publication/315783464

Analysis of rock-lining interaction for circular tunnels using the finite element
simulation.

Conference Paper · May 2005

CITATIONS READS

0 165

3 authors, including:

H. A. Hasan Fathalla M. El-Nahhas

3 PUBLICATIONS   1 CITATION   
Ain Shams University
69 PUBLICATIONS   265 CITATIONS   
SEE PROFILE
SEE PROFILE

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

Rehabilitation of Ahmed Hamdy Tunnel View project

Tunnelling and Underground Structures in Soft Ground View project

All content following this page was uploaded by Fathalla M. El-Nahhas on 05 April 2017.

The user has requested enhancement of the downloaded file.

 11th ICSGE Ain Shams University
17-19 May 2005 Faculty of Engineering
Cairo - Egypt Department of Structural Engineering

Eleventh International Colloquium on Structural and Geotechnical Engineering

ANALYSIS OF ROCK-LINING INTERACTION FOR CIRCULAR

TUNNELS USING THE FINITE ELEMENT SIMULATION
(1)
Capt. Eng. Hazem Ahmed Hasan
(2)
Prof. Dr. Fathalla Mohamed El-Nahhas
(3)
Col. Dr. Adel Mahmoud Belal
ABSTRACT
The finite element technique is used to model the two phases of tunneling process
namely; excavation and rock-lining interaction. The excavation phase is responsible for
determining the pre-lining rock mass deformations and the reduced in-situ stresses. The
interaction phase models the compatibility of the rock-lining system. The deformations and
stresses of the rock-lining system and the final rock mass pressure acting on the lining are
determined. The finite element results are compared with the result of the Convergence-
Confinement method for the case study (Whitemud-Creek Tunnel) that was guided by field
measurements.
A parametric study proved that for circular tunnels with diameter up to12m excavated
through very poor rock full lining with shotcrete and steel ribs is needed.

Keywords: Convergence-Confinement approach, Rock Tunnels, Ground Reaction

Curve, Rock bolts, Shotcrete, Structural interaction, Support Characteristics Curve.
1. INTRODUCTION
Realistic analysis of the tunnel behavior and lining interaction requires an understanding
of the tunneling procedure to simulate its effect in the analytical model. The ground movements
due to tunneling, the lining installation and the rock-liner behavior are different elements of the
mechanism which must be idealized in the analytical model. Simulation of these elements
requires adequate representation of the strain-strength characteristics of the ground and some
major details of the tunneling construction process.
The interaction between the rock and the lining requires input data concerning the rock and
the liner materials properties and the structural idealization of the liner in the overall system. The
interface properties between the rock and the liner must be simulated in the analytical model to
ensure the compatibility between the liner and the surrounding rock. Therefore, the simulation
procedure of the problem goes through the following two stages:
 Excavation Phase: this phase of the analysis starts with the initial state of the
stresses present in the ground up to the liner activation including ground
deformation associated with tunneling operation.
 Interaction Phase: this phase of the analysis defines the interaction behavior
between the rock and the liner.
(1)
Graduate student, Engineering Department of Armed Forces.
(2)
Prof. of Geotechnical Engineering, Ain Shams University.
(3)
Assistant Prof. of Geotechnical Engineering, Military Technical College.

1/9
2. FINITE ELEMENT SIMULATION
Idealized ground profile at the Whitemud-Creek Tunnel is shown in Fig. 1. Analysis of
displacements and stresses around the tunnel was carried out using a 2-D plane strain finite
element package ANSYS version 5.6 taking into consideration the linear-elastic behavior of the
lining and the ground materials. The finite element mesh used to carry out the simulation for the
excavation phase is shown in Fig. 2 with 2836 nodes and 5270 2-D elements representing ground
layers. The mesh used for the interaction simulation phase is shown in Fig. 3 with 1057 nodes
and 1739 2-D elements representing the ground and the lining material.

3. MATERIAL MODELS AND PROPERTIES

3.1 Ground
The linear elastic model is used for characterizing the ground material, which consists of
two different layers with the following average mechanical properties as recommended by
Thomson and El-Nahhas (1980) as follows:
 Clay till
E : Young’s modulus = 138 MPa.
υ : Poisson’s ratio = 0.4
γ : unit weight density = 19.6 kN/m3
 Clay-Shale bedrock
E : Young’s modulus = 352 MPa.
υ : Poisson’s ratio = 0.4
γ : unit weight density = 21 kN/m3
To represent the gradual increase of clay-shale stiffness with depths, the following
properties were used:
 Shale (1)
E : Young’s modulus = 250 MPa
υ : Poisson’s ratio = 0.4
γ : unit weight density = 21 kN/m3
 Shale (2)
E : Young’s modulus = 352 MPa
υ : Poisson’s ratio = 0.4
γ : unit weight density = 21 kN/m3
 Shale (3)
E : Young’s modulus = 500 MPa
υ : Poisson’s ratio = 0.4
γ : unit weight density = 21 kN/m3
 Shale (4)
E : Young’s modulus = 700 MPa
υ : Poisson’s ratio = 0.4
γ : unit weight density = 21 kN/m3
3.2 Lining
 Steel ribs
E : Young’s modulus = 21E 4MPa
υ : Poisson’s ratio = 0.25
γ : unit weight density = 79 kN/m3
 Shotcrete and steel ribs
E : Young’s modulus = 21E 6 kPa
2/9
υ : Poisson’s ratio = 0.25
γ : unit weight density = 25 kN/m3

4. STRUCTURAL STATIC ANALYSIS

The solution of the problem depends on a static analysis, which was used to determine the
displacements, stresses, strains and forces in the ground and different lining systems caused by
acting loads. These loads depend mainly on the depth of the overburden above the tunnel under
the effect of initial in-situ stresses which can be divided into vertical stresses and horizontal
stresses and taking into consideration the effect of coefficient of earth pressure (k◦ =0.67). The
effect of any expected dynamic loads is neglected.

5. LOADING AND BOUNDARY CONDITIONS

5.1 Stress Reversal Method
The simulation of this problem by stress reversal method depends mainly on applying
tension radial forces on each node of the tunnel boundary in the direction of the free surface of
the tunnel which resulted from the excavation process. The loads acting on tunnel perimeter is
simulated as two components of tension forces in x-direction and y-direction at each node. The
total value of each force is divided into two ratios depending on the excavation and interaction
phases as follows:
 70% of each force will be applied in the excavation phase to simulate the stress
releases after excavation.
 30% of each force will be applied in the interaction phase to simulate the ground-
lining interaction to achieve the state of equilibrium between ground and steel
ribs lining.
In case of shotcrete with steel ribs lining, the loads will be divided to 75% for excavation
phase and 25% for interaction phase because the stress releasing will be greater due to the using
of combination between two different lining types.
5.2 Boundary Conditions
While applying loading, the boundary conditions are as follows :
 The displacement in the x-direction at the two vertical sides of the model is equal
to zero ( ux =0).
 The displacement in the y-direction at the bottom horizontal boundary of the
model is equal to zero (uy =0).

6. ANALYSIS OF WHITEMUD-CREEK TUNNEL

6.1 Excavation Phase
The effect of tunnel excavation process on the vertical ground displacements values for the
whole profile of the excavated tunnel are shown in Fig. 4. It is noticed that the vertical
displacement (settlement) above tunnel is increased from the ground surface down to the tunnel
crown. The settlement at the tunnel crown is downward due to the downward unloading. The
vertical displacement below the tunnel (heave) decreased gradually for points located at lower
levels below the tunnel invert. The displacement at the tunnel invert is upward and it is always
smaller than the displacement at the tunnel crown due to the increasing of the ground stiffness.
The horizontal ground displacements adjacent to the tunnel is shown in Fig. 5. The
maximum values of final horizontal displacement was about 4mm. It can be neglected at the
crown and the invert due to the boundary of the ground mass at these locations. The horizontal
displacement decreased from the tunnel surface going through ground mass.

3/9
 The values of radial displacements on the tunnel perimeter are shown in Fig. 6. The average
radial displacement ahead the tunnel face is estimated as 6.4mm along the perimeter of excavated
tunnel which resulted from the releasing of stresses after excavation process. This value on the
Ground Reaction Curve is estimated as 6mm. The values of vertical, horizontal and radial
displacements at the tunnel crown, spring line and invert of the tunnel are listed in Table 1.
Table 1: Values of displacements due to excavation
Vertical Horizontal Radial
Location
displacement displacement displacement
Crown 10mm 0 10mm
Spring line 1mm 4mm 4.1mm
Invert 6.8mm 0 6.8mm

The vertical in-situ stresses around the tunnel are changed due to excavation process. These
changes are shown in Fig. 7. It is obvious that the changes of stresses at the crown and the invert
are tension due to the inward movement of the ground in the vertical direction. It is downward at
the crown and upward at the invert. But, at spring line the increase of stresses is compression
because of stress arching.
The increase of horizontal stresses at the crown and the invert is compression due to stress
arching and at the spring line, the change of horizontal stresses is tension due to the inward
ground movement in the horizontal direction. The values of change in horizontal stresses are
shown in Fig. 8 and the values of vertical, horizontal and radial stresses at the crown, spring line
and invert of the tunnel are summarized in Table 2.
Table 2 Values of change in stresses due to excavation
Vertical Horizontal Radial
Location
stresses stresses stresses
Crown 586.5kPa -121kPa 598.9kPa

Spring line -710.1kPa 558kPa 903.1kPa

Invert 1086kPa -204.8kPa 1105.1kPa

6.2 Rock-Lining Interaction

6.2.1 Evaluation of the Ground Movements and Stresses Associated with Steel Ribs Lining
Fig. 9 and Fig.10 illustrate the additional vertical and horizontal displacements of the ground
profile after steel ribs lining activation respectively. It is clear that these values are less than the
corresponding values at the excavation process because of the activation loads are smaller than
those associated with the excavation process. The values of vertical, horizontal and radial
displacements in the rock mass at the tunnel crown, spring line and invert are summarized in
Table 3.

Fig.11 and Fig.12 illustrate the change in vertical and horizontal stresses in rock mass after
steel ribs lining activation. The values of the change in vertical, horizontal and radial stresses at
the tunnel crown, spring line and invert are summarized in Table 4.

4/9
 Table 3 Values of displacements in rock mass due to lining activation
Vertical Horizontal Radial
Location
displacement displacement displacement
Crown 0.7mm 0 0.7mm

Spring line 0.25mm 0.5mm 0.56 mm

Invert 1mm 0 1mm

Table 4 Values of change in stresses in rock mass due to lining activation

Vertical Horizontal Radial
Location
stresses stresses stresses
Crown 84kPa 35.7kPa 91.3kPa

Spring line -86.6kPa -71.6kPa -112.4kPa

Invert 118.5kPa 50kPa 128.6kPa

The final equilibrium stresses of the ground profile are estimated as the sum of the in-situ
stresses as compression and the induced stresses at the excavation and lining phases.
σ final = -σ º + σ excavation phase + σ lining phase (1)
Fig.13 and Fig.14 illustrate the final vertical and horizontal stresses of the ground profile at
equilibrium position. The values of the final vertical, horizontal and radial stresses in rock mass
at the tunnel crown, spring line and invert are listed in Table 5.
For the tunnel perimeter the radial displacement after steel ribs lining activation is shown in
Fig.15.
Table (5) Values of Final Stresses in Rock Mass
Vertical Horizontal Radial
Location
stresses stresses stresses
Crown -187.4kPa -660.1kPa -686.2kPa
Spring line -1718.7kPa -131.34kPa -1723.7kPa
Invert -219kPa -815.1kPa -844kPa

The total average radial displacement is equal to the sum of the average radial displacement
of the excavation and interaction phases, which are equal to 6.4 mm and 0.6 mm respectively.
The resulted value is equal 7 mm while the value estimated using the Convergence-Confinement
approach varies between 7.5 to 9 mm as shown in Fig.16. It is clear that the resulted values of the
two analysis procedures are compatible. The final acting pressure on the steel ribs lining at the
tunnel crown, spring line and invert are summarized in Table 6 and Fig.17. The average acting
pressure on steel ribs lining is equal to 215 kPa while the average estimated value from the
Convergence-Confinement approach is equal to 228 kPa as shown in Fig.16. It is clear that the
resulted values of the two analysis procedures are compatible.

5/9
 Table (6) Final acting pressure on steel ribs lining
Location Final acting pressure on lining
(kPa)
Crown 144.3
Spring line 339.2
Invert 162

6.3 Evaluation of the Ground Movements and Stresses Associated with Shotcrete and Steel
Ribs Lining
The radial displacement along tunnel perimeter after shotcrete with steel ribs lining
activation is shown in Fig.18. The total average radial displacement is equal to the sum of the
average radial displacement of the excavation and interaction phases, which are equal to 6.6 and
0.4 mm respectively. The resulted value is equal 7 mm, while the average estimated value from
the Convergence-Confinement approach is equal to 11.4 mm as shown in Fig.19. The values of
final radial displacements and radial stresses in rock mass at tunnel crown, spring line and invert
are summarized in Table 7. The final acting pressure on the shotcrete with steel ribs lining at the
tunnel crown, spring line and invert are summarized in Table 8.
Table 7 Values of final radial displacement and final radial stresses in rock mass
Final radial Final radial
Location
displacements stresses
Crown 11mm 709.7kPa

Spring line 5mm 1823.1kPa

Invert 7.4mm 1131.2kPa

Table 8 Final acting pressure on shotcrete with steel ribs lining

Location Final acting pressure on lining
(kPa)
Crown 109.3
Spring line 82
invert 91.2

The average acting pressure on shotcrete with steel ribs lining is equal to 94 kPa while the
average estimated value from the Convergence-Confinement approach is equal to 115 kPa as
shown in Fig.19. It is clear that the resulted values of the two analysis procedure are similar.
The final acting pressure on shotcrete with steel ribs lining are shown in Fig.20.
It should be noted that the value of released stresses in excavation phase is equal to 75% of the
total in-situ stresses because the activation of the two lining systems (steel ribs and shotcrete)
take more time than the activation of steel ribs only. Therefore, the value of average radial
displacement is greater than the corresponding value in case of steel ribs lining system but the
efficiency of this system is better because its load capacity is higher.
6/9
 Finally, after calibration of Whitemud-Creek Tunnel by the finite element analysis with the
Convergence-Confinement approach which was guided by field measurements, the results proved
that the utilized finite element procedure is a good procedure for the ground-lining interaction.
Hence, the finite element analysis was used for a parametric study to evaluate the effect of
different parameters on rock-lining interaction.

7. PARAMETRIC STUDY
One of the main objectives of this study is to investigate the effect of different parameters
on the behavior of excavated tunnel before and after lining activation. The analysis followed the
same procedure which had been applied in the analysis of the case of Whitemud-Creek Tunnel
taking into consideration the different values of the tunnel radius and the depth of excavated
tunnel through different qualities of rock ranging between very poor rock and poor rock.
The parametric study has been conducted for circular tunnel as follows:
 Circular tunnel
R1 = 3m R2 = 4.5m R3 = 6m
d1 = 30m d2 = 40m d3 = 50m d
R

The depth of the tunnel (d) is measured from ground surface to the center line of tunnel.
Table 9 contains the geotechnical properties of different rock mass qualities which are used for
the parametric study.The purpose of this study is to display the effect of the different parameters (
such as, the dimensions of the tunnel Young's modulus (E), depth of the overburden above the
tunnel center line (d), shear modulus (G) and unconfined compressive strength (σci)) on the
displacement of the ground and the lining and the final load of the lining.
Table 9 Geotechnical properties of different rock mass qualities
σci Erm G γ
Type of rock RMR GSI υ
MPa MPa MPa kN/m3
Very poor rock 5 15 10 0.3 500 192.3 22.1
Poor rock 10 30 25 0.25 1000 400 24

Where
σci unconfined compressive strength.
RMR Rock Mass Rating.
GSI Geotechnical Strength Index.
υ Poisson’s ratio.
Erm Young’s modulus of rock mass.
G Shear modulus.
γ Unit weight.
Fig.21 shows the finite element mesh used to carry out the simulation of the circular tunnel
shape for the parametric study using the material properties given in Table 9. Due to the quality
of rock mass, the total value of each force is divided into two ratios depending on the excavation
and interaction phases as follows:
 80% of each force will be applied in the excavation phase to simulate the stress
releases after excavation.

7/9
  20% of each force will be applied in the interaction phase to simulate the ground-
lining interaction to achieve the state of equilibrium between rock and shotcrete
with steel ribs lining.
The changes of average radial displacement and radial stresses are the major aim of the
obtained results for each model analysis. The analysis consists of 18 models for circular tunnel
before lining activation for very poor rock and poor rock. Also, the analysis consists of 2 models
for circular shape with diameter D=12 m at depth equal to 50 m excavated through very poor
rock as the worst case of study and poor rock to be lined with shotcrete with steel ribs lining
system. This lining system was chosen because the very poor rock in some region of the tunnel
section may contain small fragments or huge separated blocks which would probably fall down.
Therefore, we need a fast installation and safe lining system. The results for different circular
tunnel cases before and after lining activation is shown in Tables 10, 11 and 12. Figures 22 to 25
show the average radial displacement for different circular tunnel cases due to tunnel excavation.

Table 10 The conclusion of results for different circular tunnel cases due to tunnel
excavation
Average radial
Tunnel condition R (m) d (m)
displacement (mm)
Very poor rock 3 30 3.8
Very poor rock 3 40 5
Very poor rock 3 50 6.3
Very poor rock 4.5 30 5.7
Very poor rock 4.5 40 7.5
Very poor rock 4.5 50 9.6
Very poor rock 6 30 7.6
Very poor rock 6 40 10.4
Very poor rock 6 50 13.2
Poor rock 3 30 2
Poor rock 3 40 2.7
Poor rock 3 50 3.4
Poor rock 4.5 30 2.8
Poor rock 4.5 40 3.8
Poor rock 4.5 50 5
Poor rock 6 30 4
Poor rock 6 40 5.5
Poor rock 6 50 6.7

8/9
 Table 11 The conclusion of results for circular tunnel (D=12m, d=50m) excavated through
very poor rock
Average radial pressure
Average radial
Tunnel condition due to lining activation
displacement (mm)
(kPa)
Before lining activation 13.2 Not applicable
After lining activation 1 85.8
Final 14.2 85.8

Table 12 The conclusion of results for circular tunnel (D=12m, d=50m) excavated through
poor rock
Average radial pressure
Average radial
Tunnel condition due to lining activation
displacement (mm)
(kPa)
Before lining activation 6.7 Not applicable
After lining activation 0.7 44.8
Final 7.4 44.8

The average radial displacement is shown in Fig. 26.

8. CONCLUSIONS
The stress reversal technique was used successfully to model the ground response during
the construction of the Whitemud-Creek Tunnel utilizing the finite element method. Results of
the interaction of the clay shale bedrock with different lining systems were similar with those of
the Convergence-Confinement approach that was guided with field measurements.
Results of the parametric study indicated that the greater value of the modulus of elasticity
(E) the lower value of average radial displacement, so the value of average radial displacement in
case of poor rock is nearly half its value in case of very poor rock due to the value of Young’s
modulus for poor rock is twice its value in case of very poor rock. For circular tunnel with
diameter up to 12m excavated through very poor rock, shotcrete with steel ribs lining system is
essential for supporting the tunnel.

REFRENCES
[1] Carranza-Torres , C. and Fairhurst , C. (2000), “Application of The Convergence-
Confinement Method of Tunnel Design to Rock Masses that Satisfy The Hoek-Brown
Failure Criterion.” Journal of Tunnelling and Underground Space Technology, Vol. 15,
No. 2, Published by Elsevier Science Ltd. pp. 187-213.
[2] El-Nahhas , F. (1977), “Field Measurements in two Tunnels in the City of Edmonton” ,
MSc. Thesis , University of Alberta , Edmonton , Alberta , Canada ,85 p.
[3] Bell , F. G. “Engineering Properties of Soil and Rock Mechanics”, pp 268-310.
[4] Thomson , S. and EL-Nahhas , F. (1980), “ Field Measurements in Two Tunnels in
Edmonton, Alberta” , Canadian Geotechincal Journal, Vol. 17, No. 1 , pp 20-33.

9/9

View publication stats