Design of pressure tunnels using a

finite element model

ID.Y.F. Simanjuntak and M. Morence, UNESCO-IHE Institute for Water Education, The Netherlands
A.J. Schleiss, Ecole Palytechnique Federole de lausanne, Switzerland
A.E. Mynell, UNESCO-IHE and Delft University of Technology, The Netherlands
In this study, a two-dimensional plane strain finite element model is used for the prediction of mechanical and hydraulic behaviour of
pre-stressed concrete-lined pressure tunnels. Results obtained from numerical models are presented and compared with the one calculated
using analytical solutions.

his paper is part of the study on the coupled another reason for the popularity of this technique is

T stress-seepage numerical design of concrete-

lined pressure tunnels and is particularly con-
centrated on the predicted behaviour of uncracked
that a tight contact between the lining and the rock
mass can be achieved, and thus the load from the lin-
ing to the rock mass can be transferred continuously.
Principally, the level of grout pressure injected
concrete-lined pressure tunnels based on finite ele-
ment model (FEM) analysis. The study covers the through radial boreholes depends on the rock mass
modelling of tunnel excavation, installation of sup- strength, which should not exceed the smallest main
port systems, pre-stressing of concrete lining and the axial stress in the rock mass so as to avoid the hydro-
loading of internal water pressure. A circular con- jacking and/or hydro-fracturing of surrounding rock
crete-lined tunnel is modelled. This design was mass. The tangential compressive strains induced by
selected as being the most suitable cross-section for pre-stressing are estimated using the load-line dia-
pressure tunnels in a deep homogeneous isotropic gram, which was developed based on the linear elas-
rock mass subject to a constant in-situ compressive tic relationship between the rock mass and concrete
stress. The jointed rock mass properties are defined lining. However, rock masses do not have a linear
based on the Hoek-Brown failure criterion. behaviour. In reality, the rock mass failure is con-
The analysis begins with the definition of normal trolled by numerous joint surfaces and therefore, the
in-situ principal stresses along the longitudinal axis non-linear Hoek-Brown criterion is applied in this
of excavation to predict excavation-induced stresses study to estimate the performance of the structural
and deformations. As a result of the excavation behaviour of pressure tunnels.
works, the stress levels surrounding the site are Because of the applicability of the FEM in dealing
changed. In this case, the behaviour of rock mass with rock mass problems, among others, non-linear
may no longer remain elastic and therefore the zone deformability, material inhomogeneity and complex
where the plasticity occurs has to be predicted. boundary conditions, this method has been widely
The mechanical performance of the rock mass in applied. In this paper, the FEM is used and an
the elastic-plastic condition was studied using the attempt has been made to describe the mechanical
FEM. An adequate support system, taking into and hydraulic behaviour of concrete-lined pressure
account the stress release in front of the excavation tunnels.
face, was designed to anticipate the initiation of The objective of the study is to provide an
plastic failure and instability problems of the tunnel. overview of a practical application of FEM for the
Once the equilibrium condition around the supported design of pre-stressed concrete-lined pressure tun-
tunnel is reached, a final lining can be installed on the nels which are divided in three parts consecutively,
support element and a high-pressure cement grout dedicated to tunnel excavation and support installa-
injected through the radial boreholes. This passive pre- tion, pre-stressing of the final lining, and activation
stressing technique produces enough compressive of internal water pressure. Only a few examples of
stress in the final lining to suppress crack openings. the implementation of FEM for the design of pre-
stressed pressure tunnels are given in the literature,
among others are found in Stematiu et ai. [1982 3 ]
1. Background and Marence [1996 4 ]. While Stematiu et ai. assumed
Over the last 25 years, the design of pressure tunnels the rock mass to behave as elastic material, Marence
at hydropower schemes has developed considerably. considered the plastic behaviour of the rock mass
The need for a more economical design has conse- defined by Mohr-Coulomb parameters.
quences for the choice of improving the bearing It has been recognized that the Mohr-Coulomb law is
capacity of concrete-lined pressure tunnels. One of not suitable for a rock mass. The predicted behaviour
the techniques available to improve their applicabil- of rock mass as a result of excavation is therefore
ity is by pre-stressing the final lining with a high- based on the non-linear yield function given by Hoek-
pressure cement grout. Brown rock plasticity. Using the convergence confine-
Since it first appeared in Seeber [1985 1], the load- ment method, the appropriate location of the installa-
line diagram has been widely used to determine the tion of the support lining, taking into account the
bearing capacity of pre-stressed concrete-lined pres- three-dimensional effect of excavation, can be deter-
sure tunnels. This diagram remains widely referred mined. The redistribution of stresses and deformations
to and the most recent edition is found in Vigl and after the installation of the support lining is predicted
Gerstner [2009 2]. As well as economic benefits, using a FEM. Then, the modelling of pressure tunnels

98 HVdropower & Dams Issue Five, 2012

progresses to the installation of the final lining, pre- of plastic dilation and thus plastic dissipation [Wan,
stressing of the final lining with high-pressure 1992 15].
cement grout and loading of the internal water pres-
sure. The pre-stressing is modelled by applying a 3. Excavation-induced stress and deformation
uniform load at the support and final lining interface.
For model validation, results of excavation- As a result of the tunnel excavation, the equilibrium
induced stresses and deformations as well as their state of initial stress is disturbed and the first deforma-
redistribution after the installation of support tions occur. The rock mass surrounding the excavation
obtained from FEM are compared with those calcu- may not remain elastic anymore and may deform non-
lated using the elastic-plastic solution as described elastically. To define the stress and deformation asso-
in Carranza-Torres [2004 5 ]. The seepage pressure ciated plasticity, the Hoek-Brown criterion and the
distribution is compared with the analytical solution non-associated flow rule material behaviour are used.
proposed by Schleiss [1986 6] and the prediction of Studies on the elastic-ideal plastic analysis of stresses
integrated bearing capacity of the system based on and displacements around a circular excavation in the
the stress-seepage coupling mechanism, is compared Hoek-Brown media are available in the literature and
with the one proposed by Simanjuntak et al. [2012 7 ]. in this paper. The results obtained from numerical
models for the calculation of excavation-induced
stress and deformation in the elastic region are com-
2. The Hoek-Brown rock plasticity model pared with the one proposed by Sharan [2005 9]. For
In cases of rock masses that exhibit non-significant the plastic region, the series of formulae described in
anisotropy in strength and deformability, the Carranza-Torres [2004 5 ] and Carranza-Torres and
assumptions of isotropic behaviour are reasonable. Fairhurst [2000 16] are precise.
However, non-linear criteria have to be used, since
the strength of the rock mass does not increase lin-
early with the level of stresses. Since being first 4. Convergence confineme t method
introduced in 1980, the failure of a jointed rock mass Since the state of stresses and displacements around
in response to induced stresses is usually described the tunnel excavation is actually three dimensional,
by using the Hoek-Brown failure criterion. three-dimensional methods should be used for the
Consequently, the definition of rock mass parame- proper analysis of distribution stresses and displace-
ters used in the numerical model is determined based ments during the tunnelling processes. If one of the
on the Hoek-Brown criterion. Some relevant exam- principal components of the in-situ stress is acting
ples can be found in Carranza-Torres and Fairhurst parallel to the longitudinal axis of the excavation,
[19998], Carranza-Torres [2004 5], Sharan [2005 9] this complicated analysis is usually replaced by
and 'Clausen and Damkilde [2008 10 ]. cross-sectional plane-strain analysis [Unlu and
Since the full mathematical treatise of the Hoek- Gercek,2003 17].
Brown criterion is already presented in Hoek et al Methods based on the two-dimensional plane strain
[2002 11 ], this paper concentrates on the plastic analysis have been developed in the past on estab-
potential that can be distinguished for non-associat- lishing the relative position of the tunnel face and
ed and associated material behaviour. The plastic the sections under consideration using certain
potential, g, is defined as: approximations which takes into account the stress
release in front of the face of excavation, among oth-
ers, convergence-confinement method [Panet and
Guenot, 1982 18 ], progressive softening method
[Swoboda et al., 1993 19] and hypothetical modulus
... (1) of elasticity soft lining method [Karakus and Fowell,
2003 2°]. These approximations reflect the deforma-
The plastic potential is needed to control volume tions that occur between the excavated area and the
change through plastic dilation. The rate of dilation application of the support system.
is controlled by the parameter lng, in which with rela- Nevertheless, a method to predict the support loads
tion to the dilation angle 'ljJ, this parameter is calcu- must consider not only the stress release occurring
lated using: before the installation of the support system, but also
the plastic behaviour of the ground, as well as that of
l+m '" 1+ sin II' the elastic ground [Kim and Eisenstein, 2006 21 ]. The
g 1- sin II' elastic-plastic behaviour of the rock mass is incorpo-
... (2) rated in the convergence confinement method and
therefore this method is selected in this study to cal-
Practical examples of the implementation of the culate the load imposed to the support lining.
Hoek-Brown criterion considering the associated Several studies have been done in the past on the cor-
and non-associated flow rule conditions are avail- rect approximation of longitudinal deformation pro-
able in Carranza-Torres and Fairhurst [1999 8 ] and file. When the measured data are available, an empiri-
Clausen and Damkilde [2008 10]. However, in cases cal best-fit relationship as suggested by Carranza-
of plane strain in isotropic rocks with failure criteria Torres and Fairhurst [2000 16] is adequate. In cases
independent of the main intermediate stress, the use where measured data cannot be realised, three-dimen-
of plastic non-associated flow rule is more appropri- sional numerical models can be used. Formulae devel-
ate [Hoek and Brown, 1997 12 ; Serrano et al., 2011 13 : oped by Vlachopoulos and Diederichs [2009 22 ] is rec-
Wang, 1996 14], meaning that the rock mass under- ommended in this paper so as to validate the results
goes no change in volume during plastic deforma- from numerical modelling given that the influence of
tion. Another reason to disregard the associated flow the development of a plastic zone is taken into account
rule is that its application overestimates the degree in the prediction of deformation after shotcrete instal-

Hydropower & Dams Issue Five, 2012 99

 lation. The best-fit relationship between radial dis- 6. Integrated bearing capacity of pre-stressed
placement and the points located ahead and behind the
face are calculated consecutively using: concrete-lined pressure tunnels
The bearing capacity of pressure tunnels can be
e-O.15(Rpil R)
_---,-_ _ .e(xIR)
improved by creating a certain pre-stress in the lining,
3 through the injection of high pressure cement grout
... (3) into the circumferential gap between the shotcrete and
final lining. The calculation of pre-stress-induced
strain at the intrados of concrete lining assuming the
tunnel is impervious can be found in Seeber [1985 1] •
... (4) In reality, a concrete lining cannot be made totally
impervious. The bearing capacity of pressure tunnels
with Rand Rpl as the radius of excavation and the elas- therefore has to take into account the effect of seepage
tic-plastic interface respectively. pressure. Using the double thick-walled porous cylin-
Once the longitudinal deformation profile has been der theory, the seepage-induced stress at the intrados
established, another question of maximum tolerable of concrete lining can be predicted. A full mathemati-
convergence arises, since it is impossible, except when cal treatise of the method has been presented in
very special techniques are used, to install the shot- Schleiss [1986 6].
crete directly at the tunnel face. In some cases, the crit- Recent studies on the mechanical and hydraulic
ical convergence of up to 5 per cent may be used; nev- effects of grouting towards the improvement of bear-
ertheless, as recommended by Hoek [2000 23 ], a critical ing capacity of pre-stressed concrete-lined pressure
value of I per cent should not be exceeded since it tunnels has been carried out by Simanjuntak et al.
would cause distress. [2012 7 ]. The integrated bearing capacity of pre-
stressed concrete-lined pressure tunnels is calculated
s. Grouted zone using the coupled stress-seepage analysis expressed in
terms of the residual tangential compressive strain at
If the rock mass in the vicinity of pressure tunnels has
a relatively high permeability, the rate of leakage flow- the intrados of concrete lining, and this should not fall
ing out of the tunnel has to be minimized to avoid pos- beyond zero for the most unfavourable loading combi-
sible hydro-jacking of the surrounding rock mass, nations.
washing out the joint fillings and associated hazards,
such as landslide, environmental impacts, flooding of
... (5)
the adjacent powerhouse or even the collapse of the
tunnel. If the safety of the tunnel is not put at risk, the
E'c,pp and E'c,Pi are respectively pre-stress-induced strain
rate of water leakage, q, in the order of 2 Us/km/bar is
still tolerable [Marence, 2008 24 ]. The excessive leak- and seepage-induced strain at the intrados of concrete
age imposed by the internal water pressure can be pre- lining.
vented by application of consolidation grouting. The
necessary radius of rock mass which needs to be grout- 7. Pressure tunnels modelling
ed, Tg , can be calculated using Eq. (6) given in Schleiss This section is intended to present an overview of the
[1986 25 ]. modelling of pre-stressed concrete-lined pressure tun-
nels simulated in three consecutive phases: tunnel
excavation and installation of support lining; installa-
Table IA: Tunnel geometry
Parameters
,:Symbol Value Unit
tion of final lining and pre-stressing; and, the activa-
tion of the loading of internal water pressure inside the
tunnel.
Internal radius of final lining f, 1.6 m The calculation of stresses and deformations from
each phase is predicted using commercial finite ele-
External radius of final lining fa 1.9 m
ment software DIANA [2012 26]. It is a deep tunnel and
Shotcrete thickness t, 0.1 m assumed to be located above the groundwater level in
Radius of excavation R 2.0 m an isotropic homogenous rock mass subjected to a
Radius of grouted zone f
g 3.0 m constant hydrostatic in-situ stress. In DIANA, two
types of analysis are coupled, namely the structural
Table IB: Rock mass properties non-linear and steady-state groundwater flow analysis.
Parameters Symbol Value Unit The general description of the modelling procedure
Uniaxial compressive strength (intact) ad 30 MPa and assumptions used for each phase is presented in
the following sub-sections.
Hoek-Brown constant mb 1.677 -

Hoek-Brown constant s 0.00387 - 7.1 Model set-up

Dilation angle l/J 0 0
For computation, an opening with a 2 m radius of
Modulus of deformation E~ 5.5 GPa excavation in an infinite elastic-plastic rock mass was
Poisson's ratio v 0.25 - used in this study. The data for intact rock properties
were adopted from the one given in Carranza-Torres
Table IC: Concrete properties
and Fairhurst [1999 8]. For the determination of rock
Parameters Symbol Value Unit mass parameters satisfying the Hoek-Brown criterion,
Unconfined compressive strength ace 35 MPa formulae described in Hoek et al [2002"] were
Young's modulus Ec 30 GPa applied. The thickness of support and final lining are
taken consecutively as 10 and 30 cm. The thickness of
Poisson's ratio Vc 0.25 - the grouted rock mass is I m.

100 Hydropower & Dams Issue Five, 2012

 Since the load and geometry of the pressure tunnels Fig. 1. Model
are symmetrical, only half of the model geometry is geometry and mesh.
considered to represent the whole tunnel. The model
domain is made free to move in the radial direction
but not in the longitudinal direction. A two-dimension~
al plane strain condition is used meaning that only the
tangential and radial strain components are consid-
ered. To avoid using an excessively large model, the
minimum required model boundary suggested by Ruse
[2003 27 ] was applied. A highly refined mesh based
quadrilateral elements was set (Fig. Ib).

7.2 Loading steps

7.2.' Phase 0 - Stress initialisation
This step is aimed to simulate the primary state of
stress in the rock mass before the excavation. The
Hoek-Brown rock mass is assumed to be in elastic-
ideal plastic condition. The load was set by introduc-
ing a constant in-situ compression stress of 30 MPa. (al (b)
The horizontal-to-vertical stress coefficient was taken
as 1.
tensile stresses in the shotcrete-fmallining interface. The
7.2.2 Phase , - Tunnel excavation and support lining installation pre-stressing of final lining was modelled by setting a
Two cases are considered in this phase: modelling of constant uniform pre-stress load at the shotcrete-fmal
the tunnel excavation without and with a shotcrete lin- lining interface. The pre-stress grout of 10 bar was used
ing. The compulsory partitions for these two cases are as an example and again, the combination of yield crite-
summarized as follows. ria of Rankine and Von Mises was applied.
Principally, for modelling of tunnel excavation, two
7.2.4 Phase 3 - Loading of internal water pressure
boundary conditions are applied. The first boundary
condition, which is located at the outside model For thi~ phase, two permeable boundaries are set up, i.e.
domain gives an upper value of the radial stresses at the mtrados of concrete lining and at the outside of
while the other, located at the tunnel wall, represent~ model domain. Some quite moderate improvements by
the support pressure. For the modelling of tunnel exca- consolidation grouting are assumed to occur to the
vatio~. wi~h a shotcrete lining, the second boundary
hydraulic parameter of the rock mass. As an example,
conditIOn IS Illtroduced to the model by activating the the permeability of the concrete lining and shotcrete,
element representing the shotcrete lining. For the rock grouted zone, and ungrouted rock mass, were taken as
mass, the non-linear analysis according to the Hoek-
10-8 , 5 X 10-8 and 10-6 mis, respectively. The internal
water pressure of 14 bar is applied as the maximum load-
Brown yield criterion was performed and the yield
ing of the internal water pressure.
value as given in Table IB [Carranza-Torres and
While the permeable boundary at the outer model
Fairhurst, 19998] was applied.
domain characterizes the hydrostatic head imposed by
The material model for shotcrete was designed
~e groundwater level, the permeable boundary at the
according to the concrete model, that means, it is capa-
mtrados of concrete lining represents the hydrostatic
ble of handling combined tension and compression.
head imposed by the internal water pressure. Since the
Here, the biaxial stress state in the concrete can be
groundwater level is not present, the hydrostatic head at
modelled by a combination of the yield criteria of
the outer permeable boundary was set to zero. The other
Rankine and Von Mises. The relevant concrete proper-
hydrostatic head representing the internal water pressure
ties, based on Eurocode 2 and presented in Table 1C,
at the intrados of concrete lining was set to 14 bar.
were used as an example.
The distribution of stresses and deformations induced
7.2.3 Phase 2 - Instaffation affinal Dning and pre-stressing by the internal water pressure was simulated based on
non-linear structural analysis. 14 bar-internal-water-
While the shotcrete lining absorbs some of the stresses
pressure was introduced to the model by setting the uni-
and deformations resulting from tunnel excavation, the
form pressure load of 14 bar at the intrados of the fmal
fmal lining is designed to take the loading imposed by
lining for the simulation of seepage-induced stress, the
the mternal water pressure. The final lining is concreted
steady state groundwater flow analysis was applied.
directly on the shotcrete lining after the equilibrium con-
dition has been reached. As a result of the concreting
processes, self-weight of lining, creep and shrinkage 8. Results and discussion
effects, the fmal lining detaches from the shotcrete While for Phase 1 and Phase 2 only the structural non-
developing the circumferential gap between them. This linear analysis was performed, the structural non-linear
occurs especially at the tunnel crown. In practice, this analysis of Phase 3 was coupled with the groundwater
gap is closed with contact grouting, using low pressure flow stress analysis. For each phase, successive itera-
(of up to 5 bar), to achieve full contact in the system. tions were performed until the equilibrium in forces and
The modelling of pressure tunnels proposed in this deformation was achieved.
paper is oriented towards the concept of the compatibil-
ity condition meaning that full contact between shotcrete 8.1 Modelling results
and fm~1 lining is achieved after the contact grouting B.'.' Phase , • Tunnel excavatian and support lining instaffation
applicatIOn. Therefore, there are no tensile strains and While Figs. 2(a) and 3(a) represent the distribution of

Hydropower & Dams Issue Five, 2012 101

 radial and tangential stress for unsupported tunnels, tion at the tunnel wall were transferred continuously to
Figs. 2(b) and 3(b) represent the stresses when the tun- the shotcrete once it was installed. The radial stress at
nel is suppOlted. When the tunnel is not supported, the the intrados of shotcrete lining was predicted to be
plastic zone develops significantly around the opening 0.58 MPa with the corresponding radial displacement
and the total deformation at the tunnel wall was found of 4.4 cm.
to be 8.01 cm (Fig. 4a) which falls beyond the critical
convergence. The development of a plastic zone needs 8.1.2 Phase 2 . Installation of linallining and in;ection of high
pressure grout
to be reduced with the installation of support lining.
With the aid of the convergence confinement method In this phase, the 30 cm-final-lining was installed on
(Fig. 5), the optimal location of the support installation the shotcrete lining and radial stress and deformation
was determined by considering the arching effect. The acting at the intrados of shotcrete lining was trans-
chosen location was at a distance of 2.1 m behind the ferred continuously to the extrados of the final lining.
tunnel face. FEM was used to predict the distribution of Since the equilibrium condition was reached and there
stresses and deformations at that location. The results of was no additional load from the rock mass, the radial
radial and tangential stresses as well as radial deforma- stress and displacement imposed by the shotcrete
tions are shown in Figs. 2(b), 3(b) and 4(b) respectively. diminished to zero on the intrados of the final lining.
In the case of unsupported tunnels, the radial stress at After that, a uniform pre-stressing load of 1 MPa was
the tunnel wall was zero (Fig. 2a). Once the 10 cm- applied at the shotcrete-final lining interface resulting
shotcrete-lining had been installed, as much as 1.95 III as much as 1.15 MPa of compressive stress in a

MPa of stress, in a radial direction supporting the rock radial direction at the extrados of final lining (Fig. 7a).
mass, acted at the boundary between the rock mass and
8.1.3 Phase 3 . Loading of internal water pressure
shotcrete lining. Accordingly, the radial displacement
at the tunnel wall decreased to 4.17 cm. The radial dis- As shown in Fig. 8a, as little as 0.14 Pa of tensile stress
placement in equilibrium is within the acceptable limit in a radial direction was found at the intrados of con-
of critical convergence. The classic jump for both tan- crete lining as a result of the loading of 14 bar-inter-
gential and radial stresses occurring at the elastic-plas- nal-water-pressure. However, this tensile stress will
tic boundary was predicted to take place at a 4.6 m not develop cracks in the final lining, since the tan-
radius. Fig. 6 shows a comparison of results for both gential stress at the intrados of concrete lining still
unsupported and supported tunnels. These figures remained in a compressive state. During the operation
show agreement between the results of numerical and
analytical models.
According to the elasto-plasticity and compatibility
condition, the excavation-induced stress and deforma-

Fig. 2. Distribution .....

.......",..........
~

ofradial stress.
"
.........
......
,".... ,.....
___ ··lhoOO!

".
,...
_ ·Ut<OOl'
'".,,,,,,,,

...".
.1.&..aJl'
.,,.,,,,
'l.l'"

·le.--OOl

0 ....'"
Jf;"tt-

.'"
~, ...
""0.0((:(1
'" 22.<0)1
(al (b)
'" ,;>J.<iXP'

'", .....
W""
.
,,, 1'8o'-rot
.,....",
Fig. 4. Distribution of radial deformation.

(a) (b)
-10

,... ~ ..... Jf'

-8
.
--0:
U
c
·3Coo(B -6
'"
,... ""......... -4
i
0 ~
,,'
",
.,.
,
·1~¢J1

". .......
.. - Jf' 0.2
--e: 0.0
-2
0
"
I-

'"x
.,.,q.... ~ l.o(Il1 -0.2
.,. ~1.qJ1
-0.4
".,....", -0.6 6
"" .... .~

'1'"
·)604JF
-0.8 8

.....".....
111'4
.." -1.9 10
~

Fig. 3. '" .......

Distribution of (a) (b)
tangential stress. Fig. 5. Predicted location ofshotcrete installation.

102 Hydropower & Dams Issue Five. 2012

of pressure tunnels, the tangential stress acting at the
intrados of final lining was predicted to be 1.59 MPa
(Fig.8b). 2.0 Plastic Elastic -\Mthoot s~pcrt (Carranza-Torres, 2004l 0,04 ,-----;Wi:;nJ1::-ou-t,--~-p-ort-,-I""Ca-"-an-'a--;-T--;o":-:-es--;. 2004;;;;;;;-]1
'---V1"lthoots~pert(FEMl ---Wrthootsl(lportlfEM)
Furthermore, seepage pressures propagate through 1.5 ~ ,,=~~::~~~~~~nza-Tmes,200U 0.03 ~~::~=~:~~~nza-TmeS,2004)
the permeable final lining, the shotcrete and the sur-
o j .........::::-.::--......_--.._-------
rounding rock mass and their magnitudes depend on ~ 1.0 ~002
the groundwater level and permeability of the sys- b
tem, that is, the concrete lining, shotcrete, grouted 0.5 0.01
zone and the rock mass. For the case where there is
no groundwater level present above the tunnel, the
0.0 ./L.-'-~ ---I
0.00 L--:::::::::~~§§3
1 2 3 5 6 8 9 10 123456 8 9 10
distribution of seepage pressures are exclusively con- rI R r/ R
trolled by the permeability properties of the system.
Then, the saturated zone is developed because water
leaks out of the tunnel and its distribution around the
tunnel is evident by the characteristic line as illus- 8.2 Integrated bearing capacity of pressure tunnels Fig. 6. Comparison
trated in Fig. 9. For the evaluation of the serviceability and reliability ofanalytical and
The distribution of seepage pressure around the of the pressure tunnels, the magnitude of residual tan- numerical results
pressure tunnel is shown in Fig. 10. From the analy- gential strain at the final loading condition is investi- for tunnel
sis, the predicted seepage pressures at the intrados excavation without
gated. Since the predicted residual tangential strain and with support.
and extrados of the grouted zone were found to be was 4.98 x 10-5 and remained compressive, it can be
4.33 and 0.82 bar, respectively. When computed concluded that the pre-designed value of internal
using the double thick-walled porous cylinder theory water pressure, which is 14 bar, is still within the safe
[Schleiss, 19866], the corresponding seepage pressure margin of tunnel design. If the coupling stress-seep-
were 4.24 and 0.7 bar. Again, results obtained using age mechanism proposed by Simanjuntak et al
numerical models were in agreement with analytical [2012 7] is used, the residual tangential compressive
solutions. strain is calculated to be 3.76 x 10-5 .

Fig. 7. Radial
stress, tangential
stress and radial
displacement after
prestressing offinal
lining.

1.31.-0001

.... .."

.... .."

(bl (e)

.. """"

·1&-0007

-llSr-oOCV

,..
,..
.l.~

.,....."

...,>So""
.. ;t$e.<07

·217.<lXP

It1.-«J1

H1~

<.
" ..
1".41/

"""",,
.."
Fig. 8. Radial
stress. tangential
stress and radial
(e) displacement after
loading ofinternal
water pressure.

Hydropower & Dams Issue Five. 2012 103

Fig. 9. Saturated References
zone. "1,,0((12 1. Seeber, G., "Power Conduits for High-Head Plants", Water
""
_.tlro012 Power & Dam Construction, Vol. 37 No.6; 1985.
"" -·1~..cm
2. Vigl, A. and Gerstner, R., "Grouting in Pressure Tunnel
,'"
_.llt.a;l2
Construction", Geomechanics and Ilmnelling, Vol. 2, No.5; 2009.
3. Stematiu, D., Lacatus, F. and Popescu, D., "A Finite Element
""
".
• •to...(XQ
Model for Excavation, Lining and Lining Prestressing of Water
Power Plant Tunnels", Proc. ISRM International Symposium
".
09.0.-0001

on Rock Mechanics: Caverns and Pressure Shafts. A.A.

,... ~'k<Ol

Balkema, Aachen, Germany; 1982.

".
.e.!ko-o(l)1
4. Marence, M., "Finite Element Modelling of Pressure Tunnel",
_ .s..7.olXIl
Proceedings of the 2nd ECCOMAS Conference. In:
'''' '04&'4)1 Computational Methods in Applied Sciences '96. J. A. Desideri,
''''
,-"lS.~l
C. Hirsch, P. Le TalJec, E. Onate, M. Pandolfi, J. Periaux, E.
""
__ .2-4•.0001
Stein. John Wiley & Sons Ltd;1996.
., .... 5. Carranza-Torres, C.,. "Elasto-Plastic Solution of Tunnel
_"I~-OO1

" ... - +lgeoOJl

Problems Using the Generalized Form of the Hoek-Brown
Failure Criterion", International Journal of Rock Mechanics
., .... and Mining Sciences, 41, Supplement 1(0); 2004.
,'..
-90..(0]
Fig. IO. Seepage 6. Schleiss, AJ., "Bemessung von DruckstoJlen. TeiJ II: Eintluss
·2-0.<001
pressure der Sickerstromung in Betonauskleidung und Fels,
distribution. Mechanisch-Hydraulische Wechselwirkungen, Bemessungs-
kriterien", Mitteilung der Versuchsanstalt fUr Wasserbau,
- +14e+002
Hydrologie und Glaziologie, No. 86, ETH Zurich,
2 ,% Switzerland; 1986.
+13eiOOl
27% 7. Simanjuntak, T.D.Y.F., Marence, M. and Mynett, A.E.,
- +1.2e..0J2
,,% "Towards Improved Safety and Economical Design of
_+11e+002 Pressure Tunnels", ITA-AJTES World Tunnel Congress &
27%
_. +10e-+W2 38th General Assembly (WTC 2012), Bangkok, Thailand;
27% 2012.
..-9Se..aJt
..% 8. Carranza-Torres, C. and Fairhurst, C., "The Elasto-Plastic
+86eiOOl
27% Response of Underground Excavations in Rock Masses That
+7.6e..aJl Satisfy The Hoek-Brown FaiJure Criterion" International
27%
_-+67elffil Journal of Rock Mechanics and Mining Sciences, 36(6): 1999.
_ ".
+57e+OO1 9. Sharan, S.K., "Exact and Approximate Solutions for
27.. Displacements around Circular Openings in Elastic-BrittJe-
+4&..001
55.. Plastic Hoek-Brown Rock", International Journal of Rock
..
_+3.8e-+OO1
" -+2.ge..ool
55%
Mechanics and Mining Sciences, Vol. 42, No.4; 2005.
10. Clausen, J. and DamkiJde, L., "An Exact Implementation of
the Hoek-Brown Criterion for Elasto-Plastic Finite Element
" ..
_+Ige+OO1

_+9,5e..;(O)
Calculations", International Journal of Rock Mechanics and
575% Mining Sciences, Vol. 45, No.6; 2008.
",o..aD
II. Hoek, E., Carranza-Torres, C. and Corkum, B., "Hoek-
Brown Failure Criterion - 2002 Edition", In: Proceedings,
Fifth North American Rock Mechartics Symposium, Toronto,
9. Concluding remarks Canada; 2002.
12. Hoek, E. and Brown, E.T., "Practical Estimates of Rock
A review of a two-dimensional plane strain FEM for the Mass Strength", International Journal ofRock Mechanics and
predictio~ of mechanjcal-hydraulic behaviour and bear- Mining Sciences, Vol. 34, No.8; 1997.
mg capacity of pre-sltessed concrete-lined pressure tun- 13. Serrano, A., Olalla, C. and Reig, I., "Convergence of
nels has been reportJd. The distribution of stresses and Circular Tunnels in Elastoplastic Rock Masses with Non-
deformations in the lining and rock mass as a result of Linear FaiJure Criteria and Non-Associated Flow Laws",
International Journal of Rock Mechanics and Mining
tunnel excavation, installation of support and final lin-
Sciences, No.6; 201l.
ing, pre-stressing and the activation of internal water 14. Wang, Y., "Ground Response of Circular Tunnel in Poorly
pressure is obtained by a phased analysis procedure. Consolidated Rock", Journal of Geotechnical Engineering,
While the rock mass covering the pressure turmel is Vol. 122, No.9; 1996.
modelled based on the non-linear yield function of 15. Wan, R.G., "Implicit Integration Algorithm for Hoek-Brown
Hoek-Brown rock plasticity, the shotcrete and fmallining Elastic-Plastic Model", Computers and Geotechnics, 14(3):
are modelled as elastic material and that obeys a combi- 1992.
nation of the yield criteria of Rankine and Von Mises. 16. Carranza-Torres, C. and Fairhurst, C., "Application of the
Convergence-Confinement Method of Tunnel Design to Rock
The proposed modelling approach is capable of assess- Masses that Satisfy the Hoek-Brown Failure Criterion",
ing the performance of pre-stressed concrete-lined pres- Tunnelling and Underground Space Technology, No.2; 2000.
sure tunnels and overall, there is a global coherence 17. Unlu, T. and Gercek, H., "Effect of Poisson's Ratio on the
between the results obtained using two-dimensional Normalized Radial Displacements Occurring Around the Face
FEM and analytical solutions. For this reason, further of a Circular Tunnel", Tunnelling and Underground Space
studies involving FEM are encouraged to predict the Technology, Vol. 18, No.5; 2003.
behaviour of pre-stressed concrete-lined pressure tunnels 18. Panet, M. and Guenot, A., "Analysis of Convergence Behind
The Face of a Tunnel", Proceedings, International Symposium
in heterogeneous rock mass under various in-situ com- Tunnelling'82, The Institution of Mining and Metallurgy.
pressive stresses and different groundwater levels. 0 London, UK; 1982.
19. Swoboda, G., Marence, M. and Mader, I., "Finite Element
Acknowledgment Modelling of Tunnel Excavation", International Journal for
This work was financially supported by Verbund Hydro Power Engineering Modelling, Vol. 6, os. 1-4; 1993.
AG, in Austria and the authors are indebted to Support Team of 20. Karakus, M. and FoweU, RJ., "Effects of Different Tunnel
TNO-DIANA and colleagues working in the software Face Advance Excavation on the Settlement by FEM",
development. Tunnelling and Underground Space Technology, No.5; 2003.

104 Hydropower & Dams Issue Five, 2012

21. Kim, HJ. and Eisenstein, Z., "Prediction of Tunnel Lining 23. Hoek, E., "Practical Rock Engineering", 2000, Internet
Loads Using Correction Factors", Engineering Geology, Vol. Reference: WwwJocscience.com.
85, os. 3-4; 2006. 24. Marence, M., "Numerical Modelling and Design of Pressure
22. Vlachoponlos, N. and Diederichs, M.S., "Improved Tunnels", Proceedings, HYDRO 2008 (Aqua-Media
Longitudinal Displacement Profiles for Convergence International Ltd), Ljubljana, Slovenia; 2008.
Confmement Analysis of Deep Tunnels", Rock Mechanics and 25. Schleiss, AJ., "Design of Pervious Pressure Tunnels",
Rock Engineering, Vol. 42, No.2; 2009. International Water Power & Dam Construction, 38(5): 1986.
26. DIANA, User's Manual - Release 9.4.4. TNO-DlANA.
https://support.tnodiana.com/manuals/d944/Diana.html, 2012.
27. Ruse, NM., "Raumliche Betrachtung der Standsicherheit der
T.D.Y.F. Simanjuntak graduated in Civil Engineering (cum
Ortsbrust beim Tunnelvortrieb", PhD Thesis. 1nstitut fUr
laude) from Institut Teknologi Medan, Indonesia, in 2002 and
Geotechnik, Universitat Stuttgart, Germany; 2003.
worked on projects dealing with the design of hydraulic
stlUctures. In 2007, he received his MSc (distinction) in Water
Science and Engineering, specialization Hydraulic Engineering
and River Basin Development from UNESCO-IHE Institute for
Water Education, The Netherlands. Since 2010, he has been
conducting PhD research at UNESCO-IHE entitled Coupled
stress-seepage numerical design of prestressed concrete lined
pressure tunnels.

M. Marence holds an MSc in Civil Engineering from the

University of Zagreb, Croatia (1987) and a PhD in Civil T.D.Y.F. Simanjuntak M.Marenee
Engineering from the InnsblUck Technical University, Austria
(1993). He has been working for Poyry Energy GmbH for more
than 18 years and involved in design of many hydropower
plants worldwide. In 2009, he joined UNESCO-IHE in The
Netherlands and holds position as Associate Professor in
Storage and Hydropower. He is author of numerous
publications on hydropower, numerical modelling, design
construction of pressure tunnels, rock mechanics and geo-
mechanical engineering solutions for the construction of
hydropower plants.
AJ. Sell/eiss AE.Mynett
UNESCO-IHE Institute for Water Education, Department of
Water Science and Engineering, PO Box 3015, 2601 DADelft,
The Netherlands

Prof AJ. Schleiss graduated in Civil Engineering from the

Swiss Federal Institute of Technology (ETH) in Zurich,
Switzerland, in 1978. He has a Doctorate of Technical Sciences
on the topic of pressure tunnel design (1986). He worked for 11
years at Electrowatt Engineering Ltd, Zurich, and was involved
in the design of many hydropower projects worldwide as an
expert on hydraulic engineering and underground waterways. In
1997 he was nominated full professor and became Director of
the Laboratory of Hydraulic Constructions (LCH) in the Civil
Engineering Department of the Swiss Federal Institute of
Technology Lausanne (EPFL). From 2006 to 2012, he was
Director of the Civil Engineering progranune of EPFL and
Chairman of the Swiss Committee on Dams (SwissCOD). He
obtained the ASCE Karl Emil Hilgard Hydraulic Prize and the
J. C. Stevens Award in 2006. In 2012, he was elected Vice-
President of ICOLD for zone Europe.

Ecole Polytechnique Fed€rale de Lausanne, LCH-ENAC-EPFL,

Station 18, CH-l 0 15 Lausanne, Switzerland.

A.E. Mynett has an MSc in Civil Engineering from Delft

University of Technology, The Netherlands. He has a DSc in
Hydrodynamics from the Massachusetts Institute of Technology,
USA (1980). He joined Delft Hydraulics in the departments of
Maritime Structures, Harbours Coasts and Offshore Technology,
and Strategic Research and Development. In 1984, he received
the TK. Hsieh Award from the Royal Institution of Civil
Engineers in the UK for his research on the stability of coastal
structures due to wind, wave and seismic loading. He has more
than 30 years of experience in research and development in the
fields of Civil and Maritime engineering. In 1997, he joined
UNESCO-IHE, Delft, as Professor of Environmental
Hydroinformatics. He is a member of the Science Board of
Deltares (Delft Hydraulics) and holds an appointment at the
Faculty of Civil Engineering and Geosciences, Delft University
of Technology. He serves on the Council of IAHR. Since 2011,
he has been Professor of Hydraulic Engineering and River
Basin Development at UNESCO-IHE where he is Head of
Department of Water Science and Engineering.

Delft University of Technology, Faculty of Civil Engineering Dragutina Golika 63, 10000 Zagreb, Croatia
and Geosciences, P.O. Box 5048,2600 GA, Delft, The
Netherlands. veski eski.hr I www.veski.hr

Hydropower & Dams Issue Five, 2012 105