ARMA 08-146

Design of Ductile Tunnel Linings

Schubert, W.
Geotechnical Group Graz, Graz University of Technology, Austria

Copyright 2008, ARMA, American Rock Mechanics Association

This paper was prepared for presentation at San Francisco 2008, the 42nd US Rock Mechanics Symposium and 2nd U.S.-Canada Rock Mechanics Symposium, held in San Francisco, June 29July 2, 2008.
This paper was selected for presentation by an ARMA Technical Program Committee following review of information contained in an abstract submitted earlier by the author(s). Contents of the
paper, as presented, have not been reviewed by ARMA and are subject to correction by the author(s). The material, as presented, does not necessarily reflect any position of ARMA, their
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper for commercial purposes without the written consent of ARMA is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgement of where and by whom the
paper was presented.

ABSTRACT: In weak rock or under high overburden, considerable displacements occur during excavation of tunnels and
galleries. The strains developing in many cases exceed the deformability of standard linings, frequently leading to severe damages
and the necessity of costly repairs. To allow for a safe and economical tunnel construction, strategies have to be used, which
guarantee support characteristics compatible with the strains, and at the same time utilize the supports as much as possible.
After a review of traditional methods, mainly used in mining in the past recent developments to deal with high displacements in
combination with modern standard supports, such as shotcrete and rock bolts are shown. The different systems currently available
are critically reviewed. For the design of such supports the development of the expected displacements must be predicted and the
time dependent properties of shotcrete considered. Special tools have been developed to predict displacements.
A relatively simple analysis method to design shotcrete linings with integrated steel elements, based on predicted displacements
and the transient lining properties is used to demonstrate the effectiveness and practical applicability of the various systems
available.

1. INTRODUCTION
Large displacements during excavation of tunnels
due to poor rock and high stresses are a challenge
for designers and contractors. Displacements can
reach several tens of centimeters, in some cases
displacements of one meter and more have been
Associated with those large
reported [1,2].
displacements are difficulties in predicting their
magnitude and development, as well as problems of
the limited deformability of standard supports.
Tunnel supports on the one hand should provide as
much resistance against deformation as possible, on
the other hand should be able to sustain the large
imposed strains. Various methods have been
developed over the decades to cope with the
difficulties. This paper addresses some aspects of
consequences of large displacements in relation to
the lining design. This includes a review of support
techniques used in the past and recent developments
of yielding supports, and the experience made with
their application on site.

2. DEVELOPMENT OF SUPPORTS FOR

LARGE DISPLACEMENTS
A traditional method in mining when experiencing
large displacements was to use U-shaped steel sets
with sliding couplings in combination with wire
mesh or lagging.

Figure 1. Destroyed steel set support with sliding couplings.

Photo: DMT

Figure 1 illustrates the deficiencies of such

supports. In particular in cases of anisotropic
deformation the steel sets buckle, and costly and
dangerous repairs are required. Timber supports had
to be replaced many times until stabilization was
reached.
2.1. Timber elements
With the introduction of concrete and shotcrete
linings in the late nineteen fifties, the previous
problem of excessive loosening diminished, but the
comparatively low deformability of concrete lead to
destruction of the lining in case of larger
displacements. Rabcewicz proposed timber
elements integrated into a concrete lining as early as
1950 to provide sufficient deformation capacity of
the system [3]. Depending on the required ductility
and resistance different types of timber can be used.

worked quite well. The same concept was later used

at the Arlbergtunnel [5], the Karawankentunnel [6]
and Inntaltunnel [7,8] in western Austria with quite
some success. The relatively homogeneous,
although extremely poor rock mass conditions in
those projects contributed to an execution without
major technical problems.
A serious collapse occurred during the excavation
of the Galgenbergtunnel in a very heterogeneous
fault zone, where also open slots in the lining in
combination with heavy rock bolting had been used.
The investigation into the accident showed that
among other factors contributing to the failure, a
certain resistance of the lining would at least have
slowed down the evolution of the failure [9].
3. YIELDING ELEMENTS INTEGRATED IN
LINING
The need to better utilize the linings in combination
with the required ductility has led to the
development of a number of different yielding
elements.

Figure 2. Concrete support with integrated timber element, as

proposed by Rabcewicz [3]

2.2. Open slots

The improvement of the tunneling technique during
the nineteen sixties, in particular the increased use
of shotcrete and rock bolts, considerably reduced
the problems in poor ground. Severe problems with
considerable displacements and destruction of the
shotcrete lining were experienced in a fault zone at
the Tauern tunnel in the Austrian Alps in the early
nineteen seventies [4]. To prevent the failure of the
shotcrete, open deformation slots were left in the
lining, which closed with deformation. In
combination with a dense bolting, this concept

3.1. Requirements
Support for tunnels in fault zones nowadays
commonly consists of reinforced shotcrete, steel
sets and rock bolts. To cope with the large
displacements, a certain strain tolerance of the
support elements is required. Unfortunately the
highest strain rates are imposed on the lining when
the strength of the shotcrete is lowest. To avoid
overloading of the shotcrete in its early age, the
response of the yielding elements to strains
immediately after installation has to be rather
soft. Besides the final magnitude of the
displacements, the displacement characteristics and
the advance rate influence the evolution of the
displacements over time.
Additional factors to be considered for a design of
ductile linings are the strength and stiffness
development, as well as creeping and shrinking of
shotcrete.
3.2. Yielding steel elements
Following the accident on the Galgenberg tunnel
ductile steel elements, which are integrated in the
shotcrete lining were developed at the Institute for
Rock Mechanics and Tunnelling at the Graz
University of Technology [9]. The system consisted
of a set of steel pipes with a foot and head plate.
The onset of the buckling should start before the
capacity of the lining was reached.

Figure 4. Left: section through improved yielding element

with concentrically arranged steel tubes [10]; right: group of 4
elements assembled for the installation in the lining

Figure 3. Buckling of steel tube in the laboratory test.

Foto: Schubert

The required capacity can be controlled by the

number of tubes per element. Experiments showed
that the load to initiate the initial buckling is
relatively high, and after the first peak a strong drop
in the resistance was found. To ease the initiation of
buckling, the tubes were weakened by drill holes or
reduction of the section by slots. The system was
then successfully used at the Galgenbergtunnel, as
well as at the Strenger tunnel in western Austria.
Development of Lining Stress Controllers (LSC)
The resistance of the yielding steel elements shown
above varies in a wide range, depending on the state
of buckling. Moritz conducted research with the aim
to decrease the variation of the load with shortening
of the tubes, and thus increase the energy absorbed
by the elements [10].

The system
and known
Controller)
projects in
Greece.

One of the critical issues, when applying such

systems with shotcrete linings is the time dependent
hardening of the shotcrete in combination with the
highest displacement rates immediately after
installation of the lining. To provide a soft response
of the lining system, not all elements in a group
have the same length. Figure 5 shows the
development of the load over shortening for a 4element LSC Type II. It can be seen, that the peak
load is reached after about 80 mm of shortening.
The peak load for this type is approximately
2.500 kN, and the load variation during further
shortening due to the buckling is between 200 to
400 kN.
3000

2500

Following requirements had to be met:

Relatively low initial resistance to account for
the strength development of shotcrete
Small variation in resistance after reaching the
peak load
Use of standard, off the shelf elements
After a series of tests and a number of numerical
simulations, a system of concentrically arranged
tubes emerged as the most feasible solution (see
Figure 4). The guiding tubes influence the buckling
of the load tube, in this way smoothen the load line.
Different dimensions of tubes were chosen, with a
capacity of up to 700 kN per tube.

2000

load (kN)

shown above is produced by ALWAG

in the market as LSC (Lining Stress
and has been used on a number of
Austria, Switzerland, Slovenia, and

1500

1000

500

0
0

20

40

60

80

100

120

140

160

shortening (mm)

Figure 5. Load versus shortening of a 4-element LSC

(length 90 cm)

180

200

BE yielding elements
With the construction of a number of tunnels in the
Alps, the demand in ductile linings increased. This
demand triggered the development of a number of
yielding elements in different layouts.

One of those is the so called HiDCon element,

which is distributed by Solexperts, Switzerland.

One system, which also uses steel tubes, is

produced by Bochumer Eisenhtte. The steel tubes
are loaded perpendicular to their axis, which leads
to an oval shape, when loaded. Several tubes are
arranged parallel and in layers (see Figure 6).

Figure 8. HiDCon element installed in shotcrete lining (from

(12)
3000

2500

Inserts of smaller diameter tubes can be used to

increase the resistance. The response of the element
to loading is rather smooth. However, due to the
direction of loading the resistance developing on the
first 200 mm of deformation is rather low (Figure
7), when compared to the LSC elements. At the
same time, the steel consumption for one element is
higher than for the LSCs.

1500
with Styrofoam layers
1000

500

0
0

20

40

60

80

100

120

140

160

180

200

shortening (mm)

Figure 9. Load versus shortening for HiDCon element with a

length of 90 cm, simplified from [11]

The elements are composed of a mixture of cement,

steel fibers and hollow glass particles and a
transverse reinforcement [11, 12]. The characteristic
of those elements is a very stiff response, reaching a
relatively high load already at approximately 2% of
strain.

3000

2500

2000

Load (kN)

without Styrofoam layers

2000

load (kN)

Figure 6. Yielding steel element, as produced by Bochumer

Eisenhtte (BE)

1500

1000

500

0
0

20

40

60

80

100

120

140

160

180

200

shortening (mm)

Figure 7. Shortening versus load for the BE elements; lower

line without inserts, upper line with inserts (length 100 cm)

Cement based elements

At least two producers attempted to produce
yielding elements based on deformable concrete.

Considering the relatively low strength of young

shotcrete, this stiff response easily leads to damages
of the lining. Reportedly, Styrofoam sheets of a
thickness of 40mm have been used on the top and
bottom of the elements at the Ltschberg Basis
tunnel to prevent failure of the lining at an early
stage. This Styrofoam inserts prevent failure of the
lining, but on the other hand lead to a practically
zero resistance of the elements (and thus the lining)
until the inserts are totally compressed.

stress intensity

A similar approach was chosen by Schretter,

Austria. Styrofoam bubbles are mixed with cement;
in addition a light wire mesh is used to control the
inevitable disintegration of the elements.

in all cases. The model used for the calculation of

the lining utilization assumes radial symmetric
displacements, time dependent development of
shotcrete properties and deformation behavior of the
yielding elements. The response of the shotcrete is
modeled with the rate of flow method [14]. The
approach of Guenot et al. is used for modeling the
evolution of displacements [15].

Figure 10. Compex element in laboratory test [13]

1,00

0,50

0,80

0,40

0,60

0,30

0,40

0,20

0,20

0,10

pi_act [MPa]

Compex element

0,00

0,00
0

10

time (days)

3000

Figure 12. Development of stress/strength ratio (stress

intensity) and radial support pressure (pi_act) for a HiDCon
element without Styrofoam inserts

2500

1500

1000

500

0
0

20

40

60

80

100

120

140

160

180

200

shortening (mm)

As can be seen from Figure 12, the capacity of the

lining is exceeded within a few hours after
installation. This has shown also on site. The
situation with 2x4cm Styrofoam layers on each side
is shown in Figure 13. Although the stress in the
lining is slightly below the strength, a strong peak
can still be observed due to the stiff reaction of the
element after the Styrofoam layers are compressed.

Figure 11 shows a very stiff initial response, an then

a constant load level over about 100 mm of
shortening. Site tests with this system on the
Tauerntunnel due to the complete disintegration of
the elements were not successful.
3.3. Comparison of yielding elements
The comparison of the element types LSC,
HiDCon, and BE is done first in terms of lining
utilization for a given imposed displacement, and
activated support resistance. Then the accumulated
energy in the single yielding elements is shown over
a shortening of 150mm.
A case was chosen with a tunnel of 10 m diameter,
a final radial displacement of about 30 cm, an
advance rate of 3 m/day, a shotcrete final strength
of 25 MPa, and a lining thickness of 35 cm. Six
elements with a length of 50 cm each are installed

stress intensity

Figure 11. Load versus shortening for Compex element with a

length of 90 cm, derived from [13]

1,00

0,50

0,80

0,40

0,60

0,30

0,40

0,20

0,20

0,10

pi_act [MPa]

Load (kN)

2000

0,00

0,00
0

10

time (days)

Figure 13. Development of stress/strength ratio (stress

intensity) and radial support pressure (pi_act) for a HiDCon
element with Styrofoam inserts

The support resistance developing with this system

is approximately 0,45 MPa, when neglecting shear
bond between lining and ground. Due to the
increase in the shotcrete strength and creeping, the
lining utilization rate drops with time to about 30%.

1,00

0,50

0,80

0,40

0,60

0,30

0,40

0,20

0,20

0,10

Table 1 shows the accumulated energy of above

element types for a shortening of 150mm.
It shows that the HiDCon element without
Styrofoam layers accumulates the same energy as
the LSC, while with Styrofoam it yields only 63%.
The BE elements, although a lot of steel is used
one element weighs about 100kg accumulates less
than half the energy of a LSC element.
Table 1. Accumulated energy after 150 mm of shortening
Element

pi_act [MPa]

stress intensity

Figure 14 shows the poor utilization of the steel

with the BE elements. The relatively stiff response
in the beginning of loading yields stress a intensity
peak of some 35% within the first day. The
utilization rate then drops to around 10%. The
activated lining resistance is around 0,15 MPa.

0,00

0,00
0

10

Energy (kNm)

% (LSC = 100%)

HiDCon 1)
307
100%
HiDCon+Styro
195
63%
BE w/o inserts
109
35%
BE w inserts
147
48%
LSC 4-tube
307
100%
1) stiff reaction without Styrofoam inserts leads to damage of
the shotcrete

time (days)

Figure 14. Development of stress/strength ratio (stress

intensity) and radial support pressure (pi_act) for a BE
element with additional steel inserts

4. CONCLUSION

1,00

0,50

0,80

0,40

0,60

0,30

0,40

0,20

0,20

0,10

pi_act [MPa]

stress intensity

As last example the development of a 4 tube LSC

element is shown in Figure 15.

0,00

0,00
0

10

time (days)

Figure 15. Development of stress/strength ratio (stress

intensity) and lining resistance (pi_act) for a 4 tube LSC
element

Due to the smooth initial response of the system, the

lining utilization ratio never exceeds 50%, and
decreases to about 30% with the development of
shotcrete strength. The activated lining resistance is
slightly below 0,5 MPa.
This simple comparison clearly shows that the
layout of ductile lining systems has to be done with
great care.
Although the accumulated energy does not indicate
if the elements work properly under site conditions,
it can serve as an indicator for material utilization.

A number of factors have to be considered, when

designing supports for tunnels with large
displacements. Key issues are the adjustment of the
system to the evolution of the displacements with
respect to the face advance, and the development of
the shotcrete strength over time. Relatively
sophisticated material models for shotcrete have to
be used to obtain a realistic interaction between
displacements and support reaction. Most critical
for the lining system are the first one or two days,
when the displacement rate is highest and the
shotcrete strength lowest. Ductile elements
integrated into the lining have a rather soft response
to initial loading to prevent lining failure at those
early stages. The comparison shows that cement
based elements tend to build up load rather quickly,
while with steel elements the reaction to loads can
be more easily controlled. In order to prevent failure
of the lining, additional soft layers, such as
Styrofoam have to be used in combination with
cement based elements.
One of the aims is to provide as much support
resistance against displacements as possible, and at
the same time protect the lining against damages.
This is expressed as accumulated energy over
shortening. In this respect, the LSC elements due to
their controlled initial response show the best
results. The BE elements appear to be rather
uneconomical due to their poor ratio between steel
used and performance.

REFERENCES
1.

Schubert, W. 1996. Dealing with squeezing conditions

in alpine tunnels. Rock Mechanics and Rock
Engineering 29 (3): 145-153. Springer , Wien .

2.

Mahmutoglu, Y., Vardar, M., Kocak, C., Sans, G.,

2006. Tunnelling difficulties under squeezing and
flowing conditions at the Ayas tunnel. Felsbau 24, Vol
5, 44-50. VGE

3.

Rabcewicz, L.v., 1950. Die Hilfsgewlbebauweise.

Doctoral thesis, Technische Hochschule Graz

4.

Pchhacker, H. 1974. Moderner Tunnelvortrieb in sehr

stark druckhaftem Gebirge. Porr Nachrichten 57/58

5.

John, M. 1980. Construction of the Arlberg expressway

tunnel tube -3. Tunnels and Tunnelling International 12
(5), 45-50

6.

Schubert, P., Marinko, T., 1989. Vortrieb des

Karawankentunnels im tektonisch stark beanspruchten
Sdabschnitt. Felsbau 7, Nr. 2, 65-68

7.

Schubert, W. 1993. Erfahrungen bei der Durchrterung

einer Grostrung im Inntaltunnel. Felsbau 11 (6): 287290. Essen , Verlag Glckauf .

8.

Schubert, W. 1993. Importance of Shotcrete Support in

Squeezing Rock. In Kompen, R., Opsahl, O., Berg, K.
(eds), Sprayed Concrete-Modern use of wet mix and
sprayed concrete for underground support; Proc. intern.
symp., Fagernes, Norwegen, 18-21 October 1993: 277282. Oslo: Norwegian Concrete Association.

9.

Schubert, W. & Riedmller, G. 1995. Geotechnische

Nachlese eines Verbruches - Erkenntnisse und Impulse.
In Semprich, S. (ed.), Innovationen in der Geotechnik;
Proc. 10. Christian-Veder-Kolloquium, Graz, 20-21
April 1995: 59-68. Graz: Institut fuer Bodenmechanik
und Grundbau, TU-Graz.

10. Moritz, B. A., 1999. Ductile Support System for

Tunnels in Squeezing Rock. In Riedmller, Schubert,
Semprich (eds) Gruppe Geotechnik Graz, Heft 5, Graz
University of Technology
11. Solexperts, 2005. Knautschelemente
Basistunnel, Schlussbericht

Ltschberg

12. Barla, G., Bonoini, M. Debernardi, D., 2007. Modelling

of tunnels in squeezing rock. In Eberhardsteiner et al
(eds) Computational Methods in tunnelling,
EURO:TUN 2007
13. Schretter & Cie, 2008. Versuchsbericht zu Compex
Stauchelementen
14. Schubert, P., 1988. Beitrag zum rheologischen
Verhalten von Spritzbeton. Felsbau 6, Nr. 3, 150-153,
VGE

15. Guenot, A., Panet, M. and Sulem, J. 1985. A New

Aspect in Tunnel Closure Interpretation. Proc. 26th US
Symposium on Rock Mechanics, Rapid City, Vol. 1,
pp. 455 460