Tunnel Support

Use of Lattice Girders in Sedimentary Rock

Einar Hrafn Hjlmarsson

Faculty of Civil and Enviromental

Engineering
University of Iceland

Tunnel support

Einar Hrafn Hjlmarsson

30 ECTS thesis submitted in partial fulfilment of a

Magister Scientiarum degree in Civil Engineering

Supervisor
Sigurur Erlingsson, University of Iceland
Advisor
Haukur Eirksson, Hnit engineering
Faculty Representative
Gsli Eirksson

Faculty of Civil and Environmental Engineering

School of Engineering and Natural Sciences
University of Iceland
Reykjavk, October 2011

Tunnel Support, use of lattice girders in sedimentary rock.

Tunnel Support.
30 ECTS thesis submitted in partial fulfilment of a
Magister Scientiarum degree in Civil Engineering.
Copyright 2011 Einar Hrafn Hjlmarsson
All rights reserved
Faculty of Civil and Environmental Engineering
School of Engineering and natural Sciences
University of Iceland
VR-II, Hjararhaga 2-6
107, Reykjavk
Iceland
Telephone: +354 525 4000
Bibliographic information::
Hjlmarsson, E.H., 2011, Tunnel support, use of lattice girders in sedimentary rock,
Masters thesis, Faculty of Civil and Environmental Engineering, University of Iceland, 78
pages.
Printing: Hsklaprent ehf.
Reykjavk, Iceland, October 2011

Abctract
Icelandic geology is highly influenced by its location on the Mid Atlantic ridge. Changes
in geological conditions are therefore frequent in Icelandic tunnels and flexible rock
support methods are required to deal with constant variation of rock mass properties.
Main focus of this thesis is tunnel excavation and rock support in sedimentary rock mass.
Instability problems in thick sedimentary rock layers in the shl tunnel are discussed.
Installed support is evaluated using finite element modelling of the rock mass and bearing
capacity calculations of the tunnel lining. Special attention is given to the usage of lattice
girders and how it was used as both temporary support to secure safer working
environment at the tunnel face and as a part of the final rock support for the tunnel.
Good estimation of rock mass properties to construct a reliable finite element model is
important. Limited laboratory test data of the sedimentary rock mass turned out to be a
problematic factor. A number of uni-axial compression tests were made on core samples
from sedimentary layers in the tunnel but no tri-axial tests were executed. That led to very
conservative approach in the estimation of the rock mass parameters. Result from the
modelling indicates that the tunnel lining used in the shl tunnel is sufficient if average
material parameters are used for the sedimentary rock mass. The rock support although
seems to be slightly insufficient if lower limit (90% of test data above) is used for sediment
strength. These results emphasise the importance of estimation of rock mass properties in
preparation of tunnel constructions.

vi

Index
Figures ................................................................................................................................. ix
Tables .................................................................................................................................. xii
Acknowledgement............................................................................................................. xiii
1 Introduction ..................................................................................................................... 1
2 Rock mass properties...................................................................................................... 2
2.1 Icelandic rock mass ................................................................................................. 3
2.2 Characteristics of rock mass .................................................................................... 5
2.2.1 Generalized Hoek-Brown criterion ................................................................ 5
2.2.2 Mohr-Coulomb failure criterion .................................................................... 6
2.3 Stresses in rock mass ............................................................................................... 8
2.4 Discontinuties in the rockmass .............................................................................. 10
3 Types of rock support ................................................................................................... 11
3.1 Rock bolts .............................................................................................................. 11
3.1.1 General ......................................................................................................... 11
3.1.2 Types of rock bolts....................................................................................... 12
3.2 Shotcrete ................................................................................................................ 14
3.2.1 General ......................................................................................................... 14
3.2.2 Shotcrete methods ........................................................................................ 16
3.2.3 Shotcrete mix ............................................................................................... 16
3.2.4 Steel fiber ..................................................................................................... 17
3.2.5 Curing of shotcrete ....................................................................................... 17
3.3 Wire mesh.............................................................................................................. 18
3.4 Shotcrete ribs ......................................................................................................... 19
3.4.1 General ......................................................................................................... 19
3.4.2 Types of shotcrete rips ................................................................................. 20
3.4.3 Installation procedure (Norwegian standard)............................................... 22
3.4.4 Usage............................................................................................................ 22
3.5 Lattice girders ........................................................................................................ 23
3.5.1 General ......................................................................................................... 23
3.5.2 Types of lattice girders................................................................................. 24
3.5.3 Installation procedure................................................................................... 25
3.6 Steel arches ............................................................................................................ 29
4 Estimation of required rock support .......................................................................... 30
4.1 Rock quality designation index (RQD) ................................................................. 30
4.2 The RMR-system .................................................................................................. 31
4.3 Rock tunneling quality Index, Q-system ............................................................... 33
4.3.1 Determination of the Q-value ...................................................................... 33
4.3.2 Required rock support for estimated Q-value .............................................. 34
4.4 Numerical analysis ................................................................................................ 36

vii

5 Deformation of rock mass ............................................................................................. 37

5.1 Theory .................................................................................................................... 37
5.2 Modeling deformation curve ..................................................................................39
5.2.1 Axisymmetrical FEM model ....................................................................... 40
6 Case study: Use of lattice girders in shl tunnel. .................................................... 43
6.1 Introduction ............................................................................................................ 43
6.2 Rock mass properties in shl tunnel .................................................................. 45
6.3 Estimation of rock mass parameters ...................................................................... 46
6.4 Calculated capacity of installed support ................................................................ 51
6.5 FEM-modelling ...................................................................................................... 54
6.5.1 Modelling method........................................................................................ 55
6.6 Result ..................................................................................................................... 57
7 Conclusions .................................................................................................................... 62
Bibliography ....................................................................................................................... 63
Appendix ............................................................................................................................. 65
A.1 Rock classification systems. ..................................................................................... 65
A.2 GSI and D value in Hoek-Brown criterion ...............................................................70
A.3 shl tunnel: Geological mapping for st 16.300-16.450 ........................................ 72
A.4 Bearing capacity of shotcrete lining. ........................................................................ 73

viii

Figures
Figure 2.1 Simplified geological map of Iceland (Weisenberger, 2010). ............................. 3
Figure 2.2 Mixed face. Basalt layers with approximately 2m sediment interbed. ................ 4
Figure 2.3 Mohr-Coulomb and Hoek-Brown failure criterions (Hoek, 2000) ...................... 7
Figure 2.4 H vs depth at various locations in Iceland........................................................... 9
Figure 2.5 Ideology of most stress criterias......................................................................... 10
Figure 3.1 Fixing of a single block and systematic bolting (Palmstm et al, 2000) ........... 11
Figure 3.2 Most used rock bolts in Icelandic tunnels. Figures from (Statens
Vegvesen, 2000) ............................................................................................... 13
Figure 3.3 Application of spiling bolts in a weakness zone. ............................................... 14
Figure 3.4 Shotcreting ......................................................................................................... 15
Figure 3.5 Deflection curves for concrete slabs with and without steel fibers
(Palmstm & Nilsen, 2000) .............................................................................. 17
Figure 3.6 Normal and measured curing of C30 shotcrete (Einarsson, 2010) .................... 18
Figure 3.7 Wire mesh to support local unstable area. ......................................................... 19
Figure 3.8 Radial rock bolting of a shotcrete rib (Vegagerin, 2008)................................ 20
Figure 3.9 Single layered shotcrete rib (Statens vegvesen, 2009) ....................................... 21
Figure 3.10 Double layered shotcrete rib (Statens vegvesen, 2009) ................................... 21
Figure 3.11 Installation of single layer shotcrete ribs in Stjrdal tunnel, Norway
(Gumundsson, 2011). ..................................................................................... 23
Figure 3.12 Thee and four-chord lattice girders (Jorimann, 2010) ..................................... 24
Figure 3.13 Scaling of under breaks and loose rock blocks ................................................ 25
Figure 3.14 Initial layer of shotcrete ................................................................................... 26
Figure 3.15 Installation of lattice girder .............................................................................. 26
Figure 3.16 Girder half embedded in shotcrete ................................................................... 27
Figure 3.17 Installation of spiling bolts ............................................................................... 27

ix

Figure 3.18 Lattice girder fully embedded with shotcrete .................................................. 28

Figure 3.19 Steel arch with sliding gap (Hoek et al.,2008) ................................................. 29
Figure 4.1 Example of RQD-value estimation (Hoek, Practical Rock Engineering,
2000) ................................................................................................................. 30
Figure 4.2 Rock support according to the Q-system (Hoek, 2000)..................................... 34
Figure 4.3 Generated element mesh of finite element model. ............................................ 36
Figure 5.1 Deformation vector around advancing tunnel (Hoek, Practical Rock
Engineering, 2000) ........................................................................................... 37
Figure 5.2 Longitudial deformation profile based on elastic and empirical models for
Rt =4,6m and Rp =7,45m. ................................................................................ 39
Figure 5.3 Axisymmetric model of the tunnel .................................................................... 40
Figure 5.4 Deformation vs decreasing inclusive e-modulus. .............................................. 41
Figure 5.5 Deformation curve for walls .............................................................................. 41
Figure 6.1 Overbreak due to weak sedimentary layer in shl tunnel. ............................. 43
Figure 6.2 Use of lattice girders and spiling bolts in shl tunnel. ................................... 44
Figure 6.3 Typical layered rock mass strata in the Vestfirir peninsula
(Gumundsson et al., 2007). ............................................................................ 45
Figure 6.4 Normal distribution of UCS data ....................................................................... 47
Figure 6.5 UCS vs Ei-module.............................................................................................. 47
Figure 6.6 Hoek and Brown failure envelop for lower limit strength parameters of
sediment............................................................................................................ 48
Figure 6.7 Hoek and Brown failure envelop for mean strength parameters of
sediment............................................................................................................ 48
Figure 6.8 Hoek and Brown failure envelope for basalt ..................................................... 50
Figure 6.9 Installed tunnel support at station 16.400 and 16.410. ...................................... 51
Figure 6.10 Cross section of lattice girder used in shl tunnel. Units are in mm. .......... 51
Figure 6.11 On left: Assembled lattice girders (Efla, 2011). On right: Final shotcrete
lining with lattice girders (Pedersen, Kompen, & Kveen, 2010). .................... 51
Figure 6.12 UCS result of shotcrete in shl tunnel. ........................................................ 53
Figure 6.13 Case 1 and 2. Station 16.400............................................................................ 54

Figure 6.14 Case 3 and 4. Station 16.410. ........................................................................... 55

Figure 6.15 Maximum wall deformation and plastic radius. X indicates a shear
failure in the rock mass and o indicates tension failure. ................................... 55
Figure 6.16 Deformation curve for case 1 ........................................................................... 56
Figure 6.17 Inclusion modulus vs deformation ................................................................... 56
Figure 6.18 Result for case 1. Lower limit strength of sediment 12m in thickness
from tunnel invert. ............................................................................................ 57
Figure 6.19 Results for case 2. Mean strength of sediment 12m in thickness from
tunnel invert. ..................................................................................................... 58
Figure 6.20 Results for case 3. Lower limit strength of sediment 12m in thickness
from mid section. .............................................................................................. 58
Figure 6.21 Results for case 4. Mean strength of sediment 12m in thickness from mid
section. .............................................................................................................. 59
Figure 6.22 Highest moments on the boarder of sediment and basalt in case 3.................. 60
Figure 6.23 Result for case 4 with variable k value from 0,5 to 1,5. .................................. 61

xi

Tables
Table 3.1 Properties of commonly used rock bolts (based on Palmstm & Nilsen,
2000) ................................................................................................................. 14
Table 3.2 Typical mix for C40 wet Shotcrete ..................................................................... 16
Table 4.1 Rock mass Rating system-RMR (enlarged in appendix A1) (Hoek,
Practical Rock Engineering, 2000) ................................................................... 31
Table 4.2 Guidelines for rock support and excavation of 10 span tunnel according to
RMR system. (Hoek, Practical Rock Engineering, 2000) ................................ 32
Table 4.3 ESR values (Hoek, Practical Rock Engineering, 2000) ...................................... 34
Table 4.4 Typical parameters for estimation of Q value in Iceland (Loftsson, 2009). ....... 35
Table 5.1 Rock mass parameters for axisymmetric FEM-model ........................................ 40
Table 6.1 UCS values of intact sedimentary rock samples in shl tunnel ...................... 46
Table 6.2 Rock mass properties of lower limit strength of sediment rock mass................. 49
Table 6.3 Rock mass properties of mean strength sediment rock mass. ............................. 49
Table 6.4 Min requirements for shotcrete in shl tunnel ................................................ 52
Table 6.5 Main model parameters ....................................................................................... 54

xii

Acknowledgement
I would like to thank the following people for their advice and assistance during my work
on this thesis.
Dr. Sigurur Erlingsson for his guidance, helpful discussions and high interest in the
project.
Haukur Eirksson for his guidance and helpful advice.
Bjrn Hararson for providing data and literature.
Hlynur Gumundsson for providing data and pictures.
Gsli Eirksson for providing test data from shl tunnel.
Special thanks to my parents for all their support.

xiii

xiv

1 Introduction
First road tunnel in Iceland was excavated through Arnadalshamar in 1948. Since then 9
other road tunnels have been constructed along with numerous tunnels associated with
construction of hydro power plants. Each tunnel has contributed to the knowledge of the
Icelandic rock mass characteristics and usage of classification systems or other methods to
estimate required rock support in the tunnels.
Tunnel support methods in Iceland have mainly been derived from experience and
conventions in the Norwegian tunnelling industry and the Norwegian standard of
tunnelling has been the foundation of Icelandic tunnel design. Q-value is used in the
Norwegian standard to classify rock mass into rock support categories. Norwegian rock
mass is however completely different from the Icelandic rock mass and adjustments has
therefore been needed to adopt the Q-system to Icelandic tunnelling conditions.
One of the newest challenges in Icelandic tunnelling was excavation through thick
sedimentary layers in shl tunnel (also referred to as Bolungarvk tunnel). Rock support
method based on the usage of lattice girders was used for the first time in Icelandic
tunnelling to deal with low strength sedimentary rock layers.
Detailed description of the usage of lattice girders in sedimentary rock will be carried out
in this report along with short overview of other main support methods used in Icelandic
tunnelling. Pros and cons of all methods will be evaluated.
Numerical analysis will be used to estimate deformation and stresses acting on such lining
for real cases from shl tunnel. Bearing capacity of reinforced shotcrete lining with
lattice girders will be calculated according to the Eurocode 2 standard.
Foundation of high quality numerical analysis is the input parameters used to describe the
rock mass behaviour around the underground opening. Main properties of rock mass
characteristics will therefore be discussed and methods used to estimate rock mass
properties of sedimentary layer in shl tunnel.
Support method based on lattice girders has been used around the world for decades
because of their flexibility and easy assembling ability. Since the knowledge of the usage
of lattice girders in Iceland is very limited a detailed description of their application in
shl tunnel is considered contributing to the knowledge base of the Icelandic tunnelling
industry.

2 Rock mass properties

Reliable design of an underground excavation requires reliable estimation of the rock mass
surrounding the underground opening (Hoek, 2000).
Rock mass is a complex matrix of intact rock and weakness zones and number of
parameters influence its strength, deformability, permeability and stability behaviour.
According to Kirkaldie (1988) at least 28 parameters could be used to describe rock mass
behaviour (Kirkaldie, 1988).

10 rock material properties

10 discontinuities properties
8 hydro geological properties

To include all these parameters in estimation of rock mass is hard or even impossible.
Limited number of the most representative parameters is therefore often used to describe
the rock mass behaviour. Number of design and classification systems has been derived for
this purpose and some of them are described in later chapter (Palmstm & Nilsen, 2000).
Compression and shear strength of rock along with properties of the discontinuities in the
rock mass is usually the foundation of classification and design system in geotechnical
engineering. Knowledge of the virgin stress field around the underground opening is also
necessary to estimate deformations and stress concentrations around the opening
The estimation of rock mass properties is usually the most challenging factor in
geotechnical design. Quality of the structure design is never higher than the quality of the
data used to estimate the rock mass parameters. Pre-geological investigation plays
therefore an important role in the design process of every underground structure and
should be comprehensive enough to make the design believable and reliable.
Rock mass properties are highly dependent on its origin. Properties of igneous rock which
have solidified from magma differ slightly dependent of the geological circumstance its
formed in but rock mass of sedimentary type has completely different characteristics. Basic
knowledge of geology is there for needed to choose a suitable design or classification
systems.

2.1 Icelandic rock mass

Iceland is located on the boundary between the Eurasian and North American tectonic
plates called the Mid Atlantic ridge. The plates drift apart at the speed of 1-2 cm/yr and the
void between the plates is constantly filled with igneous rock.
The North Atlantic opened around 60 million years ago and has from that time formed the
North Atlantic sea floor generated from the modern day plate boundary the Mid Atlantic
ridge (Sigmundsson, 2006).

Figure 2.1 Simplified geological map of Iceland (Weisenberger, 2010).

The youngest rock is located near the volcanic active zones of Iceland and the oldest rock,
14-16 million years old, is located in the north west and east coast. Icelandic bedrock
consists 80-90% of sub aerial basalt, 5-10% of acidic and igneous rocks and 5-10% of
sedimentary interbeds. Typical Icelandic bedrock consists of relatively thin basalt layers
with thin scoria layers on top and bottom of each layer and thin sedimentary interbeds. The
strata usually dip slightly towards the volcanically active zone. Since the basalt layer are
usually thin <10m a mixed face condition is very common in Icelandic tunnels and
sometimes the dominating condition in tunnel projects (Hararsson, 1991). Mix faced
condition can be seen on Figure 2.2.

Figure 2.2 Mixed face. Basalt layers with approximately 2m sediment interbed.

Other rock formations like dykes and breccias are also frequently crossed during
excavation of tunnel. Sedimentary intebeds is also frequent in Icelandic rock mass but are
usually thin layered, from few centimetres to 1 or 2 meters. Occasionally thicker
sedimentary layers occur.
Main focus of this report is tunnel excavation in sedimentary rock and properties of
sedimentary rock will therefore be most discussed.
Sedimentary rock can be divided in two main categories by its formation:
1. Where material particles has been transported to the place of deposition, known as
clastic or detrital rock. Conglomerates, sandstones, siltstones, mudstones are of that
kind.
2. By an aggregation of organic matter or chemically/biochemically. Limestone, chalk
and coal are of that kind.
Sedimentary rock is not interlocked like basalt rock but cemented together with an
intergranual matrix by diagenesis (i.e. hardening of loose materials to rock) (Palmstm &
Nilsen, 2000).
Sedimentary rocks in the Tertiary bedrock are mostly fine grained tuffaceous interbeds and
some thicker conglomerates. The Pleistocene bedrock is usually more abundant of

sedimentary rock, mostly sandstones and conglomerates of glacial origin (Hararsson,

1991).
Thick sedimentary interbeds can cause some serious instability problems specially if joined
with low strength, swelling potentials or sensitivity to water exposure.

2.2 Characteristics of rock mass

Estimation of rock mass characteristics is required to design an underground excavation.
Methods such as the generalized Hoek-Brown criterion and Mohr-Coulomb failure
criterion can be used to describe the characteristic behaviour of rock mass like strength and
deformations. Measured data from core samples are often used to estimate the properties of
intact rock (no weakness planes) and from that point through empirical approach to
estimate the behaviour of the overall characteristics of the rock mass surrounding an
underground opening. Strength of intact rock sample is usually higher than the overall
strength of the rock mass and method are therefore needed to convert measured data from
core samples to the rock mass(Hoek, 2000).
2.2.1 Generalized Hoek-Brown criterion
Hoek and Brown proposed in the 1980s a method to estimate the strength and properties
of a jointed rock mass called Hoek-Brown failure criterion. The method is based on
estimation of interlocking between rock blocks and shear conditions in the joints. This
method was derived to be used to estimate strength of jointed rock mass where rock blocks
are small relative to the excavation considered. The method has been modified over the
years but the version introduced in this chapter is a modified version from 2002 (Hoek,
Carranza, & al, 2002). The following series of equations represents the criteria:

Eq. 2-1

Eq. 2-2
Eq. 2-3
1
2

1
6

Eq. 2-4

where
and
are the maximum and minimum effective principal stresses at failure,
mb is the Hoek-Brown constant for the rock mass and mi is the Hoek-Brown constant for
the intact rock samples, s and a are constants related to the rock mass characteristics and
is the uniaxial compressive strength of intact rock sample.
GSI is the geological strength index introduced by Hoek in 1994 to simplify the conversion
between the intact rock strength and the rock mass strength.

A disturbance factor D is used to take into account the disturbance from blasting and stress
relaxations in the rock mass.

To estimate the value of mi the equation 2-1 is used with S=1 and a=0,5 and becomes:

Eq. 2-5

A series of triaxial test on core samples can therefore be used to determine the value of mi
and . Authors of the method recommends that series of at least five triaxial tests should
be used (Hoek, Carranza, & al, 2002).
Authors also recommend that the range of
should be equally distributed between zero
and 0,5 time the intact compression strength.
can be set to zero in equation 2-5 to reveal the uniaxial compression strength of the
rock mass and becomes:

Eq. 2-6

The tensile strength of the rock mass can be found in a similar way by setting
zero and the tensile strength becomes:

equal to

Eq. 2-7

Figure that can be used to estimate the value of GSI and D can be found in appendix A2.
2.2.2 Mohr-Coulomb failure criterion
The Hoek-Brown failure criteria is well suited for jointed or heavily jointed igneous rock
types like basalt. But for other rock types like sedimentary or metamorphic rock the MohrCoulomb failure criterion can be a better choice (Hoek, 2000).
Estimation of shear strength can be made by the Mohr-Coulumb eqution:

tan

Eq. 2-8

where is the internal friction angle of the intact rock sample, C is the cohesion and
the normal stress acting on the plain of failure (Erlingsson, 2009).

is

The Mohr-Coulomb equation can also be written as:

1
1

sin
sin

2
1 sin

tan

Eq. 2-9

Uniaxial compression strength and tensile strength can be derived from equation 2-9 by
putting and to zero respectively, thus
2
1 sin
2
1 sin

Eq. 2-10
Eq. 2-11

Figure 2.3 Mohr-Coulomb and Hoek-Brown failure criterions (Hoek, 2000)

2.3 Stresses in rock mass

Stresses within the rock mass are of great interest when designing an underground
excavation.
Compressive stresses are by convention positive and tensile stresses have negative prefix.
Both high and low stresses can cause instability problems in underground excavation. If
stresses exceed the compression strength of the rock mass it yields and low stress condition
can cause rock blocks to slide due to low normal stress in the joints.
Stresses surrounding an underground opening are influenced mainly by the stress condition
prior to the excavation (virgin stresses) and the geometry of the opening. Number of
components influences the direction and magnitude of the virgin stress field.
Main components influencing magnitude and direction of the stress field are (Palmstm &
Nilsen, 2000):
Gravitational stresses are caused by gravitational force that pulls all material towards the
centre of the earth. Gravitational stresses increase with depth and 0,027 MPa/m is a
common number to calculate the vertical components of the gravitational stress field.
Poisson ratio is often used to describe the relationship between the vertical and horizontal
component of the gravitation stresses in elastic materials. Poisson ration does not however
correspond very well with stress conditions in jointed rock mass.
Topographic stresses are caused by uneven ground surface. Uneven surface causes stress
concentrations and uneven stress field in the rock mass near the ground surface like in the
bottom of deep valleys and mountain slopes. The horizontal component is the overall stress
field can become the dominant factor near the surface in such cases.
Tectonic stresses are mainly caused by plate tectonic or the continental drift. Major faults
and folds are caused by tectonic stresses. Variation in tectonic stresses is high and that is
one on the reasons why the horizontal component of virgin stresses are much more difficult
to estimate than the vertical one.
Residual stresses. During the earlier geological stages of the rock mass it locks in some
internal stresses.
Vetical component can generally be calculated by equation 2-12.
0,027

Eq. 2-12

The letter k is usually used to describe the ratio between horizontal and vertical stresses
and the horizontal stress field becomes:
Eq. 2-13

Origin and behaviours of virgin stress field can be very complex and hard to estimate. The
vertical component of the stress field can easily be calculated with certain accuracy but the
8

horizontal components will always be a rough estimation unless some rock stress
measurement has been performed. Stress measurements are time consuming and expensive
but also very necessary to perform if a design of an underground excavation need relatively
accurate estimation of the horizontal stress field.
Number of stress measurements has been done in Iceland, mostly related to construction of
hydroelectric power stations. Figure 2.4 shows overview of various stress measurements in
Iceland.
0
100
200

Reydarfjordur
Teigsbjarg

depth[m]

300

Sandfell
400

Blanda
Frskrsfjrurtunnel

500

v(0,027xd)
600

H/v=0,5
H/v=1,5

700
800
0

10

15

Hmax[Mpa]

20

25

30

Figure 2.4 H vs depth at various locations in Iceland.

Teigbjarg, Sandfell and Blanda (Haimson, 1981)

Reyarfjrur (Haimson & Rummel, 1981)
Frskrsfjrur tunnel (Dahle, 2005)
Figure 2.4 shows that horizontal stresses at relatively shallow depth gives in most cases
higher k value than measurements at greater depth. This is not surprising since both
topographic and tectonical stresses are responsible for higher portion of horizontal stresses
at shallow depth.
Measurements of horizontal stress around the world also show that the value of k tends to
be high at shallow depth and decreases at depth (Hoek, 2000).

2.4 Discontinuties in the rock mass

Discontinuities divide the rock mass into blocks of different scales. The blocks can be
categorised by their scale as (Palmstm & Nilsen, 2000):
1.
2.
3.
4.
5.

First order fault block defined by larger weakness zones or faults.

Second order blocks formed by small weakness zones or seams.
Third order blocks formed by normal joints
Small discontinuities such ad bedding or schistosity partings.
Small fragments or grains in the rock.

Discontinuities that form blocks that fall into category 5 are usually considered as part of
the rock property and generally included in the strength characterisation of the rock.
Shear strength of discontinuities is an important factor in estimation of characteristic of
jointed rock mass. Numerous criterias and test methods have been derived to estimate
shear strength of discontinuities (Hoek, 2000), but since the main focus in this report is
cemented sedimentary rock, a detailed description of those criterias is outside the scope of
this report.
Figure 2.7 shows the main ideology of shear stress criterias in discontinuities.

Figure 2.5 Ideology of most stress criterias

10

3 Types of rock support

Rock support is added to improve stability of underground opening. Pre-geological
investigation are used to reveal the main characteristics of the rock mass surrounding the
opening and a design is made to deal with various geological conditions in the tunnel.
Geological conditions often change very rapidly on the tunnel route and that require
flexible support methods that can be quickly adjusted to the current circumstances
(Palmstm & Nilsen, 2000).
Rock bolts and shotcrete are dominant support methods in Icelandic and Scandinavian
tunnelling industry and has been for a long time. Shotcrete rips were first prescribed as an
optional support method in Hvalfjrur subsea tunnel in 1996 but were not used until in
Frskrsfjrur tunnel in 2003 (Hararson, 2011). Steel arches are not commonly used
but have been used for example in Oddskars tunnel and hydro tunnel in the Krahnjkar
project (Loftson, 2011)
Main rock support methods will be described in this chapter and there pros and cons
discussed.

3.1 Rock bolts

3.1.1 General
Development of rock bolts began in the 1920s and has since then become the most
dominant support method in underground construction (Luo, 1999)
Rock bolting is a flexible method very commonly used for rock support. Rock bolts are
frequently used as initial support at the tunnel face to obtain safe working conditions for
the crew and they also form part of the final rock support (Palmstm & Nilsen, 2000).
After securing loose block at the excavation face by spot-bolting, a systematic bolting is
often used. Then a previously defined pattern of rock bolts are installed based on the
geological conditions. Geological mapping, Q-value or other methods are mainly used for
the design of the required pattern to secure the rock mass. A systematic bolting can be
installed at the end of excavation or during excavation of the tunnel. Figure 3.1 shows
spot-bolting and systematic bolting.

Figure 3.1 Fixing of a single block and systematic bolting (Palmstm et al, 2000)

11

The following equation for bolts lengths has been suggested by Palmstm (2000) for
bolting of single loose block:
1,4
1,4

where

0,08

0,16

1
0,5

0,1

Eq. 3-1
0,1

Eq. 3-2

Db = the block diameter (in meters)

Wt = the tunnel wall height (in meters)
Dt = the diameter or span of the tunnel (in meters)

3.1.2 Types of rock bolts

Rock bolts are produced of various types and lengths. Rock bolts can be categorized as
follows:

End anchored bolts.

Fully grouted bolts.
Combination bolts.
Swelling bolt
Self drilling bolts

End anchored bolts can be used when fast activation of the bolts is needed. The anchor can
be expansion shell, polyester grout or mortal (cement grout). End anchored bolts are often
used for spot bolting to secure unstable blocks at the tunnel face.
Fully grouted bolts are passive bolts and are not activated in the installation phase.
Deformations in the rock mass will eventually activate the bolt after the grout has cured.
Fully grouted rock bolts are often used for systematic rock bolting. Cement based grout is
mostly used to fasten the bolt but polyester is also possible to use, especially for short
bolts.
Grout gives the bolt protection from corrosion but the bolts are also often galvanized or
epoxy coded.
Combination rock bolts are end anchored bolts that can be fully grouted afterwards. The
most common use combination bolts in modern Scandinavian tunnels are CT-bolts.
Polyethylene sleeve gives the CT-bolt additional protection from corrosion. CT-bolt are
relatively stiff since it is both end anchored and fully grouted and are therefore very
convenient for use in weak rock or heavily jointed rock. CT-bolts are however not

12

applicable in rock mass where high deformation can be expected since its stiffness can lead
to failure in the rock bolt (Statens Vegvesen, 2000).

Figure 3.2 Most used rock bolts in Icelandic tunnels. Figures from (Statens Vegvesen,
2000)

Swelling rock bolts can be used as initial rock support but are not qualified as permanent
rock support by the Icelandic road authorities. Swelling bolts have direct contact with the
rock mass and therefore in great threat of corrosion. Installation procedure is however very
fast and therefore favourable under special circumstances.
Self drilling rock bolts are sometimes used in very weak or heavily jointed rock mass.
Sometimes its very hard to keep the bore hole unblocked after drilling and therefore the
installation of normal rock bolts are hard or impossible. That problem can be avoid with
the self drilling rock bolts since it doesnt require stable bore hole since it drills itself in.
Self drilling rock bolts are however expensive because the drill bit at the end of the bolt
cannot be retained (Statens Vegvesen, 2000).
Rock bolts are sometimes installed ahead of excavation to support assumed unstable rock
mass or rock mass with insufficient rock cover. Those rock bolts are called spiling bolts
and are installed in a fan shaped pattern oriented 10-25relativly to the tunnel axis. Spiling
bolts are usually fully grouted with spacing of 30-80cm (Palmstm & Nilsen, 2000).
Figure 3.3 shows typical application of spiling bolts.

13

Figure 3.3 Application of spiling bolts in a weakness zone.

Table 3.1 Properties of commonly used rock bolts (based on Palmstm & Nilsen, 2000)

Typeofrock Diameter Yieldstrength Failureload Elongationat Elongationfor3

bolt
[mm]
[kN]
[kN]
failure
mbolt[mm]
Roundsteelbar
20
6070
100
8%
240
Deformedbar
20
120
150
3%
90

25
220
250
1%
30
CTbolt
20
120
150
3%
90

22
200
250
2%
60
Hollowbolt
27
100
130
8%
240

Standard
boltlength
[m]
0.86.0
0.86.0
0.86.0
1.56.0
1.56.0
2.06.0

3.2 Shotcrete
3.2.1 General
Shotcrete is a widely used method for tunneling support. This type of rock support is
obtained by spraying concrete on the rock surface. Shotcrete for rock support has been
used for several decades and has become increasingly popular because of its favorable
properties together with high capacity and flexibility (Palmstm & Nilsen, 2000).

14

Figure 3.4 Shotcreting

The main advantages of shotcrete as rock support are (based on Palmstm & Nilsen,
2000):

Ready to use on short notice (batching plant on site)

Framework is not needed
Applicable in uneven excavation profile
Easy to combine with other support methods
Flexible deformation properties.

The main disadvantages are:

Low tension strength (can be improved by mixing fibers or reinforce with casting
steel)
Can collapse when applied on swelling rock types
Hard to apply on rock types with low cohesion (some types of sedimentary rock)
Hard to apply on rock with flowing water.

Shotcrete can be used in multiple variations as rock support:

15

1. Ordinary shotcrete spayed in layers of up to about 60-100 mm thickness.

2. Mesh reinforced shotcrete. This is produced by first spraying a layer of concrete
before installing a mesh with typically 5-6 mm diameter steel bars. Then a second
and sometimes more layers are applied to cover the mesh entirely.
3. Fiber reinforced shotcrete. Steel or plastic fibers are mixed with the wet concrete.
4. In combination with steel ribs, lattice girders or steel beams.
Al these mentioned methods are typically used with combination of rock bolt to fasten
loose block or to fasten the wire mesh or steel sets (Palmstm & Nilsen, 2000).
3.2.2 Shotcrete methods
Wet mix shotcreting is now the dominant shotcrete method in modern construction of
tunnels and has almost replaced the previous dry mix method. The difference between
these two methods are that the wet-mix method uses fully mixed concrete and blows it out
of the nozzle of the shotcrete robot using compressed air.
The mixing water is however not added to the dry-mix until in the nozzle in the dry-mix
method. That gives the operator the change to adjust the dosage of water during
shotcreting. The wet mix method is however more convenient for ordinary tunnel
shotcreting since it has less rebound, higher capacity and gives better working conditions.
3.2.3 Shotcrete mix
When producing concrete one uses a various types of recipes based on what properties of
the concrete one wishes to get. Table 3.1 shows a typical mix for concrete used for
shotcrete support in tunnelling in Norway (similar in Iceland).

Table 3.2 Typical mix for C40 wet Shotcrete

Portland cement
Micro-silica
Aggregates 0-8mm
Superplasticizer (BNS)
Plasticizer (lignosulphonate)
Steel fibers
Accelerator (modified silicate)

470 kg
8%
1670 kg
5 kg
3,5 kg
50 kg
5%

The amount of micro-silica and accelerator are calculated as percentage of cement. Microsilica in the mix improves the strength properties of the shotcrete and makes it easier to
distribute the steel fibers in the fresh shotcrete. It also reduces permeability and improves
frost resistance (Palmstm & Nilsen, 2000).

16

3.2.4 Steel fiber

Unreinforced shotcrete is a brittle material that experiences cracks and displacement under
relatively low tensile stress compared to its compression strength. Steel fibers have been
used since the late 1950s to improve the concretes tensile properties.
By adding 1 vol % of steel fibers into a shotcrete mix can increase the load capacity of a
C50 shotcrete slab by 85% (Palmstm & Nilsen, 2000). The usage of steel fibers has made
it possible to use a combination of rock bolts and fiber reinforced shotcrete instead of full
cast lining. A full cast lining is a very time consuming process and expensive so replacing
that with rock bolts and fiber shotcrete is very economically favourable.

Figure 3.5 Deflection curves for concrete slabs with and without steel fibers (Palmstm
& Nilsen, 2000)

Plastic (polypropylene) fibers can also be used in increase the tensile strength of concrete
and are often used for surface shotcrete for example as fire-protection for PE-water
membrane.
3.2.5 Curing of shotcrete
To increase the curing rate of shotcrete it is necessary to use so called accelerator. Its
mixed with the shotcrete in the nozzle and reacts with cement in the shotcrete. The dosage
is usually 4-8% of cement weight in the mixture.
Usage of accelerator gives shotcrete initial curing to some degree so it doesnt fall of the
rock surface. The influence of the accelerator on curing and measurements of compression
strength of core specimens of shotcrete from shl tunnel can be seen on figure 3.6. After
15 hours the compression strength has reach 20 MPa. Shotcrete with no accelerator will
reach this value after 4-5 days (Einarsson, 2010).

17

100

UCS[MPa]

10

0,1
0,1

10

100

Curingtime[hours]
6%accelerator
Noaccelerator
Measuredshltunnel5%

4%accelerator
Measuredshltunnel6%

Figure 3.6 Normal and measured curing of C30 shotcrete (Einarsson, 2010)

When estimating the bearing capacity of shotcrete its necessary to have information on the
curing rate of the shotcrete. Bearing capacity of a shotcrete lining is time dependant but the
stresses form the tunnel wall is mainly dependent on the distance from the tunnel face.
Right timing to activate a tunnel lining is therefore critical and is discussed further in
relations with ground reaction curves in Chapter 5.

3.3 Wire mesh

Wire mesh is commonly used with combination of rock bolts and shotcrete. Use of fiber
reinforced shotcrete has however reduced the need for wire mesh since installation of wire
mesh is time consuming and is therefore only used where fiber reinforced shotcrete is not
considered to be sufficient (Palmstm & Nilsen, 2000).
Wire mesh is also commonly used in combination with shotcrete ribs and lattice girders to
distribute forces from the rock mass to the bearing elements.

18

Figure 3.7 Wire mesh to support local unstable area.

Wire mesh is also used where low cohesion between shotcrete and rock surface is causing
fallouts of shotcrete. These circumstances often occur during shotcreting of very wet,
heavily jointed or sedimentary rock surfaces. Local shear failures due to pull weight of the
shotcrete is also common in rock types with low cohesion strength and are often supported
with wire mesh before shotcreting.
Main advantages:

Increases shear strength and ductility of a shotcrete lining.

Reduced fallout of shotcrete during shotcreting.

Main disadvantages:

Installation is time consuming and therefore replaced by steel fiber if possible.

3.4 Shotcrete ribs

3.4.1 General
Then the normal support methods like shotcrete, rock bolt and wire mesh are not
considered sufficient to secure long term stability in the rock mass, one must consider
alternative heavier rock support methods.
A full concrete lining in weak tunnel rock is a time consuming and expensive support
method. Therefore a reinforced shotcrete ribs are a very favourable alternative to be used
where its applicable (Palmstm & Nilsen, 2000).

19

Shotcrete rips are built by fastening rebar steel along the tunnel profile with rock bolts and
then covered with shotcrete. Diameter of the rebars is usually 20mm (Norwegian standard)
and number of rebars and spacing between rips is adjusted to the local condition and
required bearing capacity.

Figure 3.8 Radial rock bolting of a shotcrete rib (Vegagerin, 2008)

3.4.2 Types of shotcrete rips

Shotcrete rips can be either single or double layered as shown on Figure 3.7 and 3.8. Single
layered shotcrete ribs are constructed by tying the rebars to a fastening rod fixed at the end
of the rock bolts. Main disadvantage with single layered rips are their low moment
capacity. Rebars on top and bottom of a double layered rip however provide the rip with
much larger moment capacity.

20

Figure 3.9 Single layered shotcrete rib (Statens vegvesen, 2009)

Figure 3.10 Double layered shotcrete rib (Statens vegvesen, 2009)

21

3.4.3 Installation procedure (Norwegian standard)

Before installing the first (or only) layer of rebars, the rock surface must be smoothened
with shotcrete. This layer is often 100-200mm thick fiber reinforced shotcrete (Palmstm
& Nilsen, 2000).
After smothering of rock surface, the following procedure takes place:

Rock bolts installed with 1-1,5 m separation based on rock class (Norwegian
standard)
Fixing bars are fastened on all rock bolts.
Reinforcement bars (20mm) are fastened (welded or tied) to the fixing bars.
The rib is shotcreted so that minimal shotcrete cover is 50mm (75mm for subsea
tunnels)

40mm spacing from reinforcement bars to the smoothening shotcrete is required before
shotcreting the rip. Min 110mm spacing between parallel rebars is required. Point 2-4 in
the procedure is repeated to build a double layer shotcrete rip (Pedersen et al, 2010).

3.4.4 Usage
Shotcrete rip can be considered as a beam that experience load from the deforming rock
mass. Because of the circular formation of the rip the load from the rock mass is
transferred to axial stress in the rib. The rib can however also experience moment where
the point load or uneven loads from the rock mass occur. Steel reinforcement is supposed
to grant the rip with enough moment capacity to withstand those moments.
Shotcrete rips are also commonly used where overburden of the tunnel is low. Where
stresses and deformation are low can the ribs be installed relatively far from the face if the
short time stability is sufficient to provide the workers with safe working conditions at the
tunnel face. Therefore they can be installing in greater number to gain productivity. Figure
3.11 shows installed shotcrete ribs in the Stjrdal tunnel where overburden is only 2-4m
(Gumundsson, 2011).

22

Figure 3.11 Installation of single layer shotcrete ribs in Stjrdal tunnel, Norway
(Gumundsson, 2011).

Main advantages of shotcrete rips are:

Material is on site or available with short notice.

Low preparation time.
Adapt to irregular profile.
High moment capacity if double layered

Main disadvantages are:

Needs to be rock bolted in the whole contour.

Difficult to use as a support for spiling bolt if profile is irregular.
Low moment capacity if single layered.

3.5 Lattice girders

3.5.1 General
Lattice girders have been used for tunnel support since the late 1970s. In recent decade a
move from heavy rolled steel arches to lighter, more manageable lattice girders has taken
place (Komselis et al, 2005).
Steel arches and lattice girders have basically the same function. They can be used as an
element of temporary support lining or a part of the permanent lining. Lattice girders were
for example used as part of the permanent lining in the underground transport system in
Bavaria, Munich in Germany with great success (Baumann & Betzle, 1984).

23

Lattice girders are very similar to normal shotcrete ribs, the main thing distinguishing these
two support methods are the installation method. The final product is a bar reinforced
shotcrete lining.
Main advantages of lattice girders are (based on Komselis et al, 2005):

Simple and fast installation

Solid support for spiling bolts
Temporary support for shotcrete until in gains sufficient strength to support itself.
Immediate support in the area of the tunnel face.
High moment capacity

Main disadvantages are:

Needs to be ordered in time or constructed on site (min 1-2 weeks in Iceland).

Leads frequently to increased usage of shotcrete

3.5.2 Types of lattice girders

Two types are most commonly used, three-chord and four-chord. Three-chord lattice has
triangular section with a larger bar (25-40mm) at the apex and two smaller diameter bars at
the base corners. Sinusoidal bars (10-12mm) separates the apex and the main bars. Fourchord lattice has four equally sized bars (20-40mm) at the corners of rectangular section.
Sinusoidal side bars and cross bars (16mm) separates the main bars.

Figure 3.12 Thee and four-chord lattice girders (Jorimann, 2010)

Design diameters stated in this chapter are the most common used bar diameters in lattice
girder production. Other bar sizes is also possible (Komselis et al, 2005).
24

3.5.3 Installation procedure

Installation procedure of lattice girders is highly related to the geological circumstances
and the purpose on their installation. It can be used as a temporary rock support to secure a
safer working environment for the workers at the face or considered as part of the final
rock support lining.
Following description of installation procedure is considered applicable for installation of
lattice girders where sedimentary rock is partly or the only rock type in the profile and
where the rock mass stresses are not preventing that final support lining is at least partly
installed very close to the tunnel face.
The installation procedure is in ten steps and every step will be described.

1) Proper scaling of loose rock blocks is always the first step in every sequence of
installation of rock support. It minimizes the risk of fall down of shotcrete and
makes the working environment safer at the tunnel face. Scaling of under breaks in
the profile must also take place at this stage before shotcreting.
Scaling in sedimentary rock with very low cohesion strength can though be risky.
Few cubic meter of lost shotcrete is then better than risking large overbreaks by
scaling to harshly with an excavator.

Figure 3.13 Scaling of under breaks and loose rock blocks

25

2) After scaling the tunnel walls and roof must be secured by installation of shotcrete.

Installation of the initial layer of shotcrete in sedimentary rock can be very time and
material consuming if the rock is weak. The weight of the shotcrete is sometimes
causing too much shear stress in the rock and causing both shotcrete and rock to
fall down. The shotcrete operator should though be able to build up the shotcrete
shell by starting at the floor or harder rock and slowly cover the unsupported area.

Figure 3.14 Initial layer of shotcrete

3) In this case(as usually) the lattice girders is the first structural member to be

installed since the initial shotcrete layer is not considered a structural member but a
safety measure and is not involved in bearing capacity calculations since its
thickness varies.
The lattice girder is assembled on the tunnel floor and raised by an excavator or by
the drill rig. To make sure that the girder is at the right position it must be check by
a laser guidance system or a surveyor.

Figure 3.15 Installation of lattice girder

4) The girders must be fastened into place. All girders must at least be fastened as
close to the floor as possible. This is done to prevent that the end of the girder
slides from the wall when forces from the deforming rock mass starts to act on the

26

girder. Extra rock bolt can be added if hard stable rock is somewhere in the profile.
5) Wire mesh is usually installed behind the lattice girder. It both adds tensional
strength to the shotcrete between the girders and also help with transferring the load
to the main bearing units which in this case is the bar reinforced concrete beam
build up by lattice girder and shotcrete. Rebars must be installed to overlap the
junctions of the girder to unify the moment and shear capacity of the girder.
6) Shotcrete is now used to cover the wire mesh and lattice girder. In this round the

girder is only half embedded in shotcrete to prevent that the girder is overloaded
with uncured shotcrete that is not able to support itself.

Figure 3.16 Girder half embedded in shotcrete

7) The girder is now used as a cantilever for spiling bolts used to support the rock for

next blasting round. The holes for spiling bolts are drilled through the lattice girder
and the bolts installed. By installing a tight series of spiling bolt the rock ahead has
been supported and chances of overbreaks reduced.

Figure 3.17 Installation of spiling bolts

27

8) The next tunnel round is now excavated.

9) After repeating stage 1-5 above in order to install the next lattice girder the

previously installed girder is connected to the new one with side bars and covered
with shotcrete. A continuous supporting lining is achieved by connecting all girders
with sidebars.

Figure 3.18 Lattice girder fully embedded with shotcrete

10) Installation of the previous lattice girder is now finished and stage 6-10 can now be
repeated for the current lattice girder.
As previously stated is this working procedure of lattice girders adjusted for use in
conditions where sedimentary rock layers are causing instability at the tunnel face. In other
geological condition the work procedure could be slightly different and must always be
adapted to the local conditions. Similar work procedure was successfully used in the
construction of shl tunnel.

28

3.6 Steel arches

Steel arches and lattice girders are very similar construction elements in tunnels. They can
work as a temporary support or as a part of final support lining. Arches are often used as a
temporary support in multiple drift tunnelling since it easy to disassemble (U.S.
Department of transportation, 2011).
Steel arched are rarely used in Iceland but has been used in Krahnjkar hydro project and
Oddskar tunnel.
Based on (U.S. Department of transportation, 2011) and (Hoek et al., 2008):
Main advantages of steel arches are:

Immediate support after installation

Excellent support for spiling bolts
Easy to disassemble and for forming temporary shotcrete wall in multiple drift
excavation
High deformability if granted with sliding joint. See Figure 3.19

main disadvantages are:

Heavy and expensive

Poor bonding with shotcrete

Figure 3.19 Steel arch with sliding gap (Hoek et al.,2008)

29

4 Estimation of required rock support

Estimations of required rock support is usually based on observation, experience and
personal judgment of those involved. Engineers generally uses support guidelines or
methods to back up their estimation of required rock support. Three types of methods are
mostly used for this purpose (Palmstm & Nilsen, 2000):

Analytical methods, involving analysis of stress distributions and deformations

using methods like numerical analysis, analogue simulation or physical modeling.
Observational methods, like the New Australian Tunnelling Method which uses
measurements of movements in the rock mass during excavation. Observations are
off course also used to check if the chosen installed rock support was the right way
to go or not and adjustment made if required.
Empirical methods, often illustrated in table or graphs that connect classification of
rock mass to curtain rock support. Number of empirical methods has been derived
such as the RMR system end the Q-system.

General overview of the most commonly used empirical methods will be given in this
chapter and the use of the Final Element as an analytical method will be used in case
studies later on in this report.

4.1 Rock quality designation index (RQD)

Deere et al. developed the RQD system in 1967 to estimate the quality of rock mass from
drill cores. The RQD value is defined as the summarized length of all core pieces longer
than 100mm divided by the total length of drilled core (Hoek, 2000).

Figure 4.1 Example of RQD-value estimation (Hoek, 2000)

Nowadays the RQD system is mostly used as a sub-system in the RMR and Q-system.

30

4.2 The RMR-system

The RMR-system was introduced by Bieniawski in 1973. The tunnel route is divided into
section where the geological properties are the same or similar and the tunnel support is
assigned to those sections according to Table 4.1 (Palmstm & Nilsen, 2000).
Six geotechnical parameters are used to classify the rock mass (Erlingsson, 2009):
1.
2.
3.
4.
5.
6.

Uniaxial compression strength

RQD-value
Spacing of discontinuities
Conditions of discontinuities
Ground water conditions
Orientation of discontinuities

Table 4.1 Rock mass Rating system-RMR (enlarged in appendix A1) (Hoek, 2000)

31

Results from the RMR classification are given in number from 0-100. Table 4.2 shows
how guidelines have been derived for rock support in 10m span tunnel for estimated RMR
value.
Table 4.2 Guidelines for rock support and excavation of 10 span tunnel according to
RMR system. (Hoek, 2000)

32

4.3 Rock tunneling quality Index, Q-system

4.3.1 Determination of the Q-value
The Q-system has been the dominated rock mass classification system since 1980 in
Iceland. The system was developed by Barton et al. in 1974 based on tunnelling experience
mainly in Norway and Sweden (Palmstm & Nilsen, 2000). In this classification system
are the joints and other discontinuities measured or estimated from core samples or visible
rock mass surface and thereafter the Q-value estimated for the rock mass. (Loftsson, 2009)
The Q-value is a number that varies on logarithmic scale from 0,001-1000 and is defined
as:
Eq. 4-1

where:
RQD

is the Rock Quality Designation

is the joint set number
is the joint roughness number
is the joint alteration
is the joint water reduction

The Q-value can be considered as a function of tree parameters; where each parameter
represents different aspects of general rock mass strength (Erlingsson, 2009):
is and estimation of block sizes.
is an estimation of shear strength in joints.
is an estimation of the active stresses in the rock mass.

Table to determine the value of individual parameter of the Q-system can be found in the
Appendix A1.
The Q-system has been used in Iceland to estimate required rock support in tunnelling
since 1980. Determination of individual parameters has been adjusted to Icelandic rock
mass based on experience in various tunnel projects in Iceland (Loftsson, 2009). The
Icelandic road Administration published a report in 2009 (Loftsson, 2009), where this
experience is used to establish some guidelines to estimation of individual parameters of
the Q-values for different types of Icelandic rock mass. An overview of the guidelines for
determination of different rock types for Icelandic rock mass is given in Table 4.4

33

4.3.2 Required rock support for estimated Q-value

Barton et al. introduced in 1974 an addition parameter, Excavation Support Ration (ESR)
to take into account a factor of safety correlated to the importance of tunnel construction.
Table 4.3 gives Bartons suggested ESR values for different types of tunnel constructions.
Table 4.3 ESR values (Hoek, 2000)

Figure 4.2 shows estimated support needs for given Q-value and span of tunnel (Hoek,
2000). Support giuedlines from the Norwaigan standard for estimated Q-value is in the
Appendix A1.

Figure 4.2 Rock support according to the Q-system (Hoek, 2000).

34

Table 4.4 Typical parameters for estimation of Q value in Iceland (Loftsson, 2009).

Rocktype

RQDvalue

Jnvalue

Basalt

Usually 60-85 for low or medium jointed. 30-60

for heavily jointed rock mass

Usually 9 but 12 for flaky olivine basalt.

Scoria

For hard well cemented, 50-80 but 30-50 for low

cemented rock mass

Usually no joint systems. 9-12 is used for well cemented and 1520 for weak cemented.

Values of 40-60 are common but 10-30 if layered.

9 in most cases but 6 if not jointed. 15-20 if very weak UCS< 3

Mpa

Breccia

Usually immeasurable so lowest value of 10 is

typically used.

15-20 is normal value. But use 20 if an excavator is able to tear

it down.

Dykes

Usually 50-70 for for low or medium jointed. 2040 for heavily jointed rock mass

Usually 12-15. Use 15 if very jointed.

Sediment If UCS <3Mpa then 10.

Jr value

Ja value

Basalt

Usually 2-3.

Usually 2-3. 2 if fillings are very thin and 3 if filling involves

clay particles. Up to 4-8 if depending on thickness and amount
of clay.

Scoria

Usually 2-3. But 2 or lower if the rock mass is

shifted.

Usually 2-3 and similar to the surrounding basalt.

Sediment

Usually 3-4 since joints are usually smooth. Use 3

for sandstone and conglomerate rock.

Usually 3-4. 2 can be appropriated for hard sandstone.

Usually 0,5-1 since joint are usually smooth.

Usually > 4. Up to 8-12 if filling are preventing interlocking of

rock blocks.

Usually 2-3 if joints are irregular.

Usually 2-3. If filling include clay then 4.

Breccia
Dykes

SRF

Jw
Basalt

1 if the tunnel is dry < 5l/min. 0,66 if water

pressure is < 2,5 bar and 0,33 if water pressure is > Usually 60-85 for low or medium jointed. 30-60 for heavily
2,5 bars.
jointed rock mass

Scoria

1 if the tunnel is dry < 5l/min. 0,66 if water

pressure is < 2,5 bar and 0,33 if water pressure is > For hard well cemented, 50-80 but 30-50 for low cemented rock
2,5 bars.
mass

Sediment

1 if the tunnel is dry < 5l/min. 0,66 if water

pressure is < 2,5 bar and 0,33 if water pressure is > Values of 40-60 are common but 10-30 if layered. If UCS
2,5 bars.
<3Mpa then 10.

Breccia

Usually < 1 since water flow is often connected

with breccia. 0,33 if water pressure is 2,5-10 bars.

Usually immeasurable so lowest value of 10 is typically used.

Dykes

Usually 0,66 if dry or 1 if low water flow. 0,33 if

water pressure is higher than 2,5 bars.

Usually 50-70 for low or medium jointed. 20-40 for heavily

jointed rock mass

35

4.4 Numerical analysis

Numerical modelling of rock mass is done by dividing the rock mass into large number of
individual elements and use computer to calculate how they interact for given stress and
stiffness conditions. The finite elements method used in this report is in the category of
continuous models which means that the rock mass is considered to be continuous rock
mass and only a limited number of discontinuities (joints, faults, etc.) may be used. To
construct a finite element model of rock mass one must define geological properties of the
rock mass and the virgin stress field. Element mesh of the rock mass is then generated to
divide the rock mass into individual elements and boundary conditions determined. Finally
the magnitude of stresses and deformations are calculated for every nodal point in the mesh
for given rock mass and boundary conditions (Palmstm & Nilsen, 2000).
The computer program Phase2 (version 7) is used in this report to construct finite element
model of rock mass surrounding tunnel in sedimentary rock mass. The FEM is used to
calculate the deformations occurring during excavation and stresses that reacts on installed
rock support. Phase2 is able to automatically generate the element mesh. An example of
generated mash can be seen on Figure 4.3. Notice how the elements get smaller as it closes
on the excavation profile.

Figure 4.3 Generated element mesh of finite element model.

36

5 Deformation of rock mass

Stability in advancing tunnel in weak rock is strongly related to the deformation occurring
after excavation. Deformations of the rock mass starts about one and half diameter of the
tunnel in front of the tunnel face and has reached 25-33% of the final deformation already
at the tunnel face. In weak rock mass these deformations might already result in instability
at the tunnel face depending on the rock mass strength and the in situ stress conditions
(Hoek, 2000).

Figure 5.1 Deformation vector around advancing tunnel (Hoek, Practical Rock
Engineering, 2000)

In order to determine the appropriated timing of the installation of specific tunnel support
one must have knowledge of the rock mass deformation behaviour and displacement
capacity of the support (Hoek et al., 2008).
Creation of longitudinal deformation profiles and ground reaction curves are convenient
tools to gain knowledge of the rock mass behaviour. Longitudinal deformation profiles can
be measured in situ or created by using analytical or numerical methods such as the finite
element method (FEM). Use of analytical approach is however only possible for very
simple shape of excavation (circular) and isotropic elastic or elastic-plastic rock mass. To
model more complex shaped excavation profile and rock mass with plastic behaviour one
must use axisymmetric or three dimensional finite element models. Empirical best fit
analysis can though be used to create deformation profiles if maximum displacement and
plastic radius has been measured or modelled.

5.1 Theory
Panet derived in 1995 a relationship base on elastic analysis for short term longitudinal
deformation profile for known maximum deformation umax (Hoek et al., 2008):

37

1
4

3
1
4

3
3 4

Eq. 5-1

where dt=X/Rt, X is the distance from face and Rt is the runnel radius.
Alternative expressions have been derived for elastic analysis for deformation profiles
suggesting that the deformation doesnt form a continuous curve in front and behind the
advancing tunnel face. Unlu and Gercek suggested in 2003:
1

Eq. 5-2

Eq. 5-3

where u0 is the deformation at the tunnel face and Aa, Ab, Ba and Bb are functions of the
Poisson ratio:
0,22
0,22
0,22

0,19

0,19
0,81

Eq. 5-4

0,73

0,81

Eq. 5-5

0,39

0,65

Eq. 5-6

where is the Poissons ratio.

Numerous empirical best fit solutions based on plastic modelling and empirical best fit to
actual measured closure date have been suggested.
Based on measured data, Chern in 1998 suggested the following expression (Hoek, er al,
2008):
,

Eq. 5-7

The shape of the deformation curve is however directly connected to the radius of the
plastic zone Rp. Therefore Hoek et al. (2008) have suggested the following relationship for
short term longitudinal deformation curve:
1
3

Eq. 5-8

where Pr is given as:

Eq. 5-9

38

Eq. 5-10

Eq. 5-11

Relationship for short term longitudinal deformation curve derived by Hoek et al.(2008)
will be used as part of the case study in Chapter 6. Figure 5.2 shows comparison of elastic
and empirical models mentioned above.

1
0,9
0,8

Ur/Umax

0,7
0,6
0,5

Hoek

0,4

Rt=4,6m

0,3

Chern
Elasticmodel

Rp=7,45m

0,2
0,1
0
20

10

10

20

30

40

Distancefromface[m]
Figure 5.2 Longitudial deformation profile based on elastic and empirical models for
Rt =4,6m and Rp =7,45m.

5.2 Modeling deformation curve

An axisymmetric FEM analysis can used to model the response of an elastic-plastic rock
mass around non-circular cross section. The process can best be explained by an example.
In this chapter, a axisymmetric model and best fit empirical analysis is used to construct a
deformation curve for a given rock mass and field stress circumstances.
Cross section and rock mass properties in Table 5.1 from shl tunnel are used in this
example. Rock mass properties are analyzed data from Mannvit consulting engineering
office from core samples taken during construction of the tunnel (Mannvit, 2009).

39

Table 5.1 Rock mass parameters for axisymmetric FEM-model

Rock
type

Tensile Friction
HoekBr. HoekBr. HoekBr.
Stiffness strength angle Cohesion UCS parameter parameter parameter

Em

[MPa]
Basalt
12270
Sediment 2266

T0
[MPa]

[]

c
[MPa]

0,79

56

1,66

ci
mb
[MPa]
80
4,09

0,00855

0,503

The field stress is considered to be 16,2 MPa in vertical direction and 6,48 MPa in
horizontal direction. That corresponds to 600m over burden and k value of 0,4. This
example should be considered as hypothetical and will not be used in the case study in
Chapter 6.
5.2.1 Axisymmetrical FEM model
For this example a tunnel of 9,2 m width is considered to be crossing a 12m thick
sedimentary rock embedded in a basalt rock mass. Location of the sedimentary layer is
considered to be in the invert of the tunnel cross section.

Figure 5.3 Axisymmetric model of the tunnel

As the tunnel advances the support from the tunnel face is reduced. To simulate this effect
an elastic material is placed inside the excavation profile and its stiffness reduced in stages.
At stage 1 the E-modulus of the inclusive rock mass is the same as the surrounding rock
mass and zero at the final stage to simulate a fully excavated tunnel far from the tunnel
face. Deformations for a single point can therefore be collected for each modulus of the
inclusion material. Figure 5.4 shows the results of the deformations in the walls in the
example.

40

2500

InclusiveEmodule[MPa]

2250
2000
1750
1500
1250

Leftwall

1000

Rightwall

750

Average

500
250
0
0

0,02

0,04

0,06

0,08

0,1

Walldeformation[m]
Figure 5.4 Deformation vs decreasing inclusive e-modulus.

According to the axisymmetric model (see Figure 5.4) the maximum average mid wall
displacement is 74mm for unsupported tunnel far from the tunnel face. The maximum
plastic radius can be estimated at the final stage of the model as 7,45m.
Deformation curve as a function of distance from the tunnel face can now be created by
using the empirical relationship suggested by Hoek (see Chapert 5.1) since both maximum
deformations and plastic radius have been estimated in the model. The curve is given in
Figure 5.5.

Walldeformation[m]

0,08
0,07
0,06
0,05
0,04

Deformationcurve

0,03

Installingofsupport
Installingofsupport1,5mfromface

0,02
0,01
0
10

10

20

30

Distancefromface[m]
Figure 5.5 Deformation curve for walls

The deformation curve also shows how much of the total displacement has already taken
41

place before installation of rock support. In this case the rock support is considered to be
installed 1,5 m from the face. Figure 5.5 shows that at that time, the deformations in the
rock mass have reached 33mm or 45% of its total deformation.
Figure 5.4 shown that deformation of 33mm has occurred when the inclusion modulus has
been decreased to 100MPa. Support from the tunnel face at 1,5m distance is therefore
considered equal to inclusion modulus of 100MPa. This information could be used to make
an FEM-model that would allow undisturbed deformation up to 33mm and make the
installed support 1,5m from the tunnel face react with all deformation after that.
The method of generating relaxation to a certain inclusion modulus can be used in FEMmodelling to replace the function of load split. The load split function divides the
deformations so that the tunnel support will only react with deformations that are
considered to take place after activation of the rock support.

42

6 Case study: Use of lattice girders in

shl tunnel.
6.1 Introduction
shl tunnel is located in the Vestfirir peninsula at the north west corner of Iceland and
connects two small towns, Hnfsdalur and Bolungarvk. The tunnel is a 9,2m span road
tunnel with 53,75 m2 cross section and a length of 5,2 km. shl tunnel were constructed
in 2008-2010 and are categorized as low traffic tunnel according to the Norwegian road
tunnel standard with annual average daily traffic (AADT) around 780. The tunnel was
excavated by traditional drill and blast method.
This case study involves a FEM-design of a shotcrete lining through sedimentary rock in
the shl tunnel. Pre-geological investigation indicated number of sedimentary interbeds
along the tunnel route. Original design of the tunnel support suggested that those
sedimentary interbeds would be dealt with by conventional rock support like rock bolts,
shotcrete, wire mesh and shotcrete ribs with spiling bolts according to convention in
Icelandic tunnel construction.
During the construction of the tunnel some instability problems were experienced in the
tunnel crown when excavated through sedimentary interbeds that in some cases lead to
collapse of the tunnel crown right after blasting (Figure 6.1). An alternative working
method for tunnel support was therefore initiated to deal with these circumstances. That
support method involved installation of lattice girders and spiling bolts.

Figure 6.1 Overbreak due to weak sedimentary layer in shl tunnel.

43

Since the use of lattice girders was not expected during the design of the tunnel, the tunnel
lining involving lattice girders was designed on site with participation of engineers from
both contractor and the supervision and approved by the tunnel designer.

Figure 6.2 Use of lattice girders and spiling bolts in shl tunnel.

The FEM program, Phase 2, is used to estimate the forces that react on the tunnel lining
under the current geological circumstances. The design code Eurocode 2 is further used to
estimate the bearing capacity of the lining.
Only limited investigations of the rock mass properties were available. Number of uniaxial compression tests (UCS) were made to estimated the intact rock strength of both
sedimentary interbeds and basalt. No tri-axial tests results were available to estimate the
geological parameters in the Hoek-Brown or Mohr-Coloumb failure criterion. To
investigate how much this limitation of available data will infect the design both average
and lower limit strength parameters were used.
At station 16.360 (TM 2.095 from Hnfsdalur side) a sedimentary layer began to arise from
the bottom of the tunnel. 10m thick sedimentary interbed had been logged in core hole at
st. 16.100 and the current layer was considered to be the same one. Exploratory holes
indicated that the layer could be 12-15m thick and according to the inclination of the strata
it would be in the profile for at least 80-100m. Excavation revealed though that tectonic
fault had shifted the strata in station 16.390 and excavation in sediment layer turned out to
be around 50m. Geological circumstances around station 16.400 to 16410 is used in this
case study since UCS and cohesion test data are available for that sedimentary layer.
Geological mapping can be found in the Appendix A3 of the area discussed.
44

6.2 Rock mass properties in shl tunnel

shl tunnel are located in the Vestfirir peninsula at the north west corner of Iceland.
The rock mass of the tunnel route is part of the oldest rock mass in Iceland. The rock mass
strata is formed by highly alterated basalt layers and numerous sedimentary layers, some
with very low compression strength. The rock mass strata dips 5-7to south west and since
the tunnel route is almost parallel to the dip direction it goes through around 500m of the
rock mass strata. Most basalt layers are around 3-8m thick but goes up to 20m in
thickness. (Gumundsson et al., 2007)
Pre-geological investigation of the tunnel route indicated that the tunnel route would cross
numerous sedimentary layers with thickness from 1-5m. Much thicker sedimentary layers
was though encountered during excavation of tunnel and lead to use of heavier rock
support than previously anticipated.
Figure 6.3 shows typical layered rock mass strata in the Vestfirir peninsula.

Figure 6.3 Typical layered rock mass strata in the Vestfirir peninsula (Gumundsson
et al., 2007).

45

6.3 Estimation of rock mass parameters

The reliability of a FEM-design is never higher than the reliability of the input parameters
(Palmstm & Nilsen, 2000). Therefore an effort is made here to estimate required rock
mass properties to be able to construct a reliable FEM-model of the rock mass. Hoek and
Brown failure criterion will be used to mimic both the basaltic and sedimentary rock. The
choose of Hoek and Brown criterion instead of Mohr-Coulumb failure criterion is based on
that no tri-axial data are available and therefore hard to create a reliable Mohr-Coulumb
failure envelope.
Table 6.1 shows UCS (Uni-axial compression strength) data measured on intact rock
samples in shl tunnel at station 16.350 (Mannvit, 2009).
Table 6.1 UCS values of intact sedimentary rock samples in shl tunnel

Depth[m] UCS[MPa] E[MPa]

3,8
18,3
2630
4,85
14,6
2320
5,15
15,2
3160
5,75
10,9
2440
6,35
4,7
1015
6,45
10,9
2250
6,75
4,7
1240
7,45
17,5
3725
7,55
11,6
2850
7,65
10,0
2350
8,4
10,9
3015
8,95
5,7
940
9,2
7,0
1480
9,4
5,4
1475
10,35
14,9
2520
10,45
12,6
2840
Mean
10,9
2266
Stdev
4,5

UCS values for the sediment layer tested in this borehole are highly distributed. Value of
standard deviation of the data is 42 % of the mean value. Statistical analysis is therefore
required to justify a single value for using as a UCS in the FEM model. By using statistical
analysis it is possible to determine the percentage of values in the data set that exceeds a
certain value (see Figure 6.4). 90% of UCS measurement are exceeding 5,2 MPa and that
value will be used as input parameter in the FEM-model. The mean value of 10,9 MPa
could also be used but in that case the rock mass would in 50% of cases be less favourable
than the parameter used in the design and would be hard to justify. It will however be used
here as a comparison. Figure 6.5 shows the correlation between E-modulus and UCS.

46

0,1
0,09

Probability Density

0,08
0,07
0,06
0,05
0,04

Density

0,03
0,02
0,01

10%

0
0

90%
6

8 10 12 14 16 18 20 22 24 26 28 30

UCS [MPa]

Stiffness E [MPa]

Figure 6.4 Normal distribution of UCS data

4000
3500
3000
2500
2000
1500
1000
500
0
4,0

6,0

8,0

10,0

12,0

14,0

16,0

18,0

20,0

UCS [MPa]
Figure 6.5 UCS vs Ei-module

Measurements of cohesion strength in this same borehole indicated mean cohesion strength
of c= 0,79MPa and c10% = 0,37 MPa.
The stiffness modulus Ei10% should be around 1366 MPa according to Figure 6.5.
Other UCS measurements of sedimentary rock in shl tunnel indicates UCS values from
1,5-9,8 MPa (Plsson, 2009). That supports the conservative approach of using UCS10% .

47

Figure 6.6 Hoek and Brown failure envelop for lower limit strength parameters of
sediment.

Figure 6.6 shows the Hoek-Brown failure envelope (red) for the lower limit strength of the
sediment layer. Figure 6.7 shows the Hoek-Brown failure envelope if mean values are used
for UCS strength and corresponding Ei-modulus.

Figure 6.7 Hoek and Brown failure envelop for mean strength parameters of sediment.

48

The blue line on Figure 6.6 and 6.7 shows approximated Mohr-Coulomb failure envelope
to the Hoek and Brown envelope.
Table 6.2 Rock mass properties of lower limit strength of sediment rock mass.
Selectedrockproperties
UCS
5,2 MPa
IntactEmodulus 1366 MPa
GSI
95
mi
6
D
0,1

HoekBrownparameters
mb
s
a

MohrCoulombfit

RockMassParameters

4,972 cohesion 1,239 MPa tensilestrength 0,589 MPa

0,5629 fric.angle 28,12
UCS
3,606 MPa
0,5
Emodulus
1266 MPa

Table 6.3 Rock mass properties of mean strength sediment rock mass.
Selectedrockproperties

HoekBrownparameters

UCS
10,9 MPa mb
IntactEmodulus 2262 MPa s
GSI
95
a
mi
6

D
0,1

MohrCoulombfit

4,972 cohesion
2,11 MPa
0,5629 fric.angle 33,05
0,5

RockMassParameters
tensilestrength 1,234 MPa
UCS
8,177 MPa
Emodulus
2096,5 MPa

First column of Table 6.2 and 6.3 shows the selected parameters. GSI value is selected
relatively high since this specific type of sedimentary rock is well cemented and doesnt
have any joint systems. mi value of 6 is recommended value (Hoek, 2000) for sedimentary
rock with silty or clayey texture if no tri-axial values are available to derive a complete
Hoek-Brown failure envelope. D value of 0,1 is chosen since damage from blasting is
considered minor. Choose of higher D value does however not reduce the rock mass
strength considerably.
The vertical component of the virgin stress field can be calculated from overburden (in this
case 400m) according to Eq. 2-12.
As previously stated is the horizontal component of the stress field harder to estimate if no
stress measurements have been performed. Considering that the most problematic factor in
the excavation of the shl tunnel was instability in the tunnel crown and previous stress
measurements in Iceland, a k value of 0,5 is considered appropriate.
The virgin stress field around the excavation is therefore estimated:
10,9

5,45

Rock mass parameters of the basalt rock are shown on Figure 6.8. The rock mass
parameters for the basalt layer are chosen relatively conservative. Modelling shows
however that increasing the rock mass strength of the basalt has minimal affects on the
modelling results and the sediment rock parameters are the dominant factor due to its very
low strength.

49

Figure 6.8 Hoek and Brown failure envelope for basalt

50

6.4 Calculated capacity of installed support

Installed tunnel support in all cases that will be analysed (see further in Table 6.5) are
lattice girders c/c 1500mm, 150mm minimum fiber reinforced shotcrete cover between
girders and 6mm wire mesh behind the girders. Figure 6.9 shows cross section of the
installed tunnel lining and Figure 6.10 shows cross section of the lattice girder.

Figure 6.9 Installed tunnel support at station 16.400 and 16.410.

Figure 6.10 Cross section of lattice girder used in shl tunnel. Units are in mm.

Figure 6.11 On left: Assembled lattice girders (Efla, 2011). On right: Final shotcrete
lining with lattice girders (Pedersen, Kompen, & Kveen, 2010).

51

Tunnel lining in rock with non-hydrostatic stress condition will always experience both
axial stresses and moments. M-N (Moment vs Axial force) envelope is therefore useful to
estimate if the tunnel lining is able to tolerate applied loading from the rock mass. M-N
envelope can be derived by calculating the ultimate bearing capacity of the lining for given
strain condition (Eirksson, 2010).
Eight strain conditions will be used to derive the M-N envelope:
1) Pure axial force
2) Zero strain in tensional bars (positive moment)
3) Equilibrium. Max strain in shotcrete and yielding in rebars at the same time
(positive moment)
4) Pure moment (positive moment)
5) Pure tension
6) Pure moment (negative moment)
7) Equilibrium. Max strain in shotcrete and yielding in rebars at the same time
(negative moment)
8) Zero strain in tensional bars (negative moment)
Same strain conditions are used to derive M-N envelopes for structural members in
buildings (columns and beams) that experience both axial forces and moments.
Calculated bearing capacity of the lining in Figure 6.9 can be seen in Appendix A4.
Phase2 can be used to derive M-N envelopes. The method used by the program is however
purely elastic and M-N envelope is constructed for each structural member. This is not
realistic for reinforced shotcrete lining due to bonding between the shotcrete and the
reinforcement. That method is however more realistic for lining of shotcrete and steel
beams since bonding strength between shotcrete and steel beams is very low.
Characteristic strength of the shotcrete used in shl tunnel was C30/37 and other
requirements can be seen in Table 6.4.
Table 6.4 Min requirements for shotcrete in shl tunnel

UCScategory

Chloridecontendscategory
Materialcategory
Maxgrainsize

C30/37
C10,20
3
8 mm

Mincementcontent
Maxv/c

420 kg/m3
0,45

Steelfibercontent
Energyabsorption

40 kg/m3
E700

52

UCS-testing on core samples displayed in Figure 6.12 shows that 99% of core samples had
higher UCS strength than 30MPa.

10
9

Numberoftests

8
7
6
5
4
3
2
1
0
29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
UCSstrengthofshotcrete[MPa]
Figure 6.12 UCS result of shotcrete in shl tunnel.

Material partial factor for concrete c is given as 1,5 in Eurocode 2. Design strength of
shotcrete in shl tunnel should therefore be (Eurocode 2, 2002):
30
1,5

20

Eq. 6-1

Eurocode 2 however notes that the partial factor c can be lowered if reliable measurements
can be used to confirm the real strength of the concrete. Partial factor s for steel is 1,15.

53

6.5 FEM-modelling
Four cases will be modelled that represents the geological circumstances at stations 16.400
and 16.410. Geological mapping of the area can be found in Appendix 4. Estimated rock
mass parameters from Chapter 6.3 are here used to construct FEM-models that are
considered to have similar characteristic as the rock mass in those locations in shl
tunnel. Table 6.5 shows the main model parameters for each case.
Table 6.5 Main model parameters

Case1
Location

Sedimentlayer
Positionofsediment
Rockmassparameter
Virginstressfield

Case2
st16.400
Location

st16.400
12m
Sedimentlayer
12m
Frominvertup
Positionofsediment
Frominvertup
Lowstrength(Table6.2) Rockmassparameter Meanstrength(Table6.3)
400moverburden,K=0,5 Virginstressfield
400moverburden,K=0,5

Case3
Location

Sedimentlayer
Positionofsediment
Rockmassparameter
Virginstressfield

Case4
st16.410
Location

st16.410
12m
Sedimentlayer
12m
Frommiddlesectionup
Positionofsediment
Frommiddlesectionup
Lowstrength(Table6.2) Rockmassparameter Meanstrength(Table6.3)
400moverburden,K=0,5 Virginstressfield
400moverburden,K=0,5

Rock mass parameters for the basalt are given in Figure 6.8
Figure 6.13 and 6.14 shows position of the sedimentary layer (brown) in the tunnel profile
each case.

Figure 6.13 Case 1 and 2. Station 16.400.

54

Figure 6.14 Case 3 and 4. Station 16.410.

6.5.1 Modelling method

This chapter shows how axisymmetrical modelling of deformation curves are used to
construct reliable FEM-model of the geological situations of each case. All cases are
modelled in the same way and modelling process of case 1 is therefore only shown in this
chapter,
Case 1. St: 16.400, 12m Sedimentary layer from tunnel invert with lower limit strength.

Figure 6.15 Maximum wall deformation and plastic radius. X indicates a shear failure
in the rock mass and o indicates tension failure.

The analysis shows that maximum wall deformation for unsupported tunnel far from the
tunnel face is 40mm (80mm convergence) and maximum plastic radius is 7,26m. Figure
55

6.16 shows a deformation curve derived from the information according to equation 5-8 to
5-11.
0,06

deformation[m]

0,05
0,04
0,03
0,02

Supportactivated1,5m
fromtunnelface

0,01
0
10

10

15

20

25

30

Distancefromface[m]
Figure 6.16 Deformation curve for case 1

Figure 6.16 shows that 18mm deformation has already occurred 1,5m from the tunnel face.
The rest of the deformations will react with the tunnel lining.

Walldeformation[m]

0,045
0,04
0,035
0,03
0,025
0,02
0,015
0,01
0,005
0
1400

1200

1000

800

600

400

Inclusionmodulus[MPa]
Figure 6.17 Inclusion modulus vs deformation

56

200

Figure 6.17 created with same method as described in Chapter 5.2, shows that the tunnel
face is still supporting the tunnel profile 1,5m from the face with similar effect as if
material with stiffness of 500 MPa where placed inside the excavation.
Tunnel relaxation is therefore generated until the inclusion modulus is 500 MPa. It
corresponds to 1,5m from the tunnel face. All deformation after that will build up stresses
in the tunnel lining.
All four cases are modelled by the same method but different inclusion modulus according
to their corresponding axisymmetric deformation model.

6.6 Result
M-N envelopes for both characteristic strength fck and reduced design strength fcd according
to Eurocode 2 are displayed on the following figures in this chapter.
Result of axial forces and moments acting of the tunnel lining are scattered on the M-N
envelopes of the lining on Figures 6.18-6.21 for each case.
6000
5000

Axialforce[kN/m]

4000
3000
2000
1000
0
100

80

60

40

20
0
1000

20

40

60

80

100

Moment[kNm/m]
Characteristicstrength

Eurocode

Modeledstresses

Figure 6.18 Result for case 1. Lower limit strength of sediment 12m in thickness from
tunnel invert.

57

6000
5000

Axialforce[kN/m]

4000
3000
2000
1000
0

100

80

60

40

20

20

40

60

80

100

1000

Moment[kNm/m]
Characteristicstrength

Eurocode

Modeledstresses

Figure 6.19 Results for case 2. Mean strength of sediment 12m in thickness from
tunnel invert.
6000

Axialforce[kN/m]

5000
4000
3000
2000
1000
0
100

80

60

40

20

20

40

60

80

100

1000

Moment[kNm/m]
Characteristicstrength

Eurocode

Modeledstresses

Figure 6.20 Results for case 3. Lower limit strength of sediment 12m in thickness from
mid section.

58

5000

Axialforce[kN/m]

4000

3000

2000

1000

0
100

80

60

40

20

20

40

60

80

100

1000

Moment[kNm/m]
Characteristicstrength

Eurocode

Modeledstresses

Figure 6.21 Results for case 4. Mean strength of sediment 12m in thickness from mid
section.

The results clearly establish how important reliable estimations on rock mass properties are
for the finite element design of the shotcrete lining. The lack of tri-axial test data to
estimate properties of the rock mass leads to more conservative approach in selection of
rock mass parameters and might therefore lead to design of heavier support than necessary.
Case 1 and 3 shows values outside or on the M-N failure envelope when conservative
values are used for rock mass parameters. All values are however inside the failure
envelope of the lining if the mean strength of the rock mass is used.
Increased diameter of the rebars in the lattice girder would grant the lining with enough
moment capacity to include all value inside the failure envelope in cases 1 and 3.
Case 3 is the most challenging case. Figure 6.22 shows that the highest moments occur on
the border of basalt and sediment. This is probably due to squeezing effect on the boarder
since the stiffness of the sedimentary layer is one order of magnitude lower then for the
basalt.

59

Figure 6.22 Highest moments on the boarder of sediment and basalt in case 3.

To investigate how changes in k value (ratio H/ V) influence the results was case 4
repeated for k value 1,0 and 1,5. Result for case 4 with variable value of k can been seen on
Figure 6.23. Hydrostatic stress condition is constructed in the model if k value is changed
to 1,0. Figure 6.23 shows that it results in lower moments compared to non-hydrostatic
stress conditions. The tunnel lining will however fail if k value of 1,5 is selected since it
results in very high moment in the tunnel lining along with very increased axial force.
As previously stated is the k value estimated to be around 0,5. The changes that the k value
is higher than 1,0 is unlikely considering the tunnel depth and that instability problems
were in most cases isolated to the tunnel crown.

60

6000
5000

Axialforce[kN/m]

4000

100

3000
2000
1000
0
80

60

40

20
0
1000

20

40

60

80

100

2000

Moment[kNm/m]
Characteristicstrength

Eurocode

k0.5

k1

k1.5

Figure 6.23 Result for case 4 with variable k value from 0,5 to 1,5.

Q-value for area discussed was estimated 0,1. According to the Norwegian standard (see
table in Appendix 1) is the required rock support as follows:

Spiling bolts with 200-300mm spacing.

Double shotcrete ribs with 1,5 -2m spacing
150-250mm C35 fiber reinforced shotcrete between ribs.
3-6m rock bolts systematically driven through the shotcrete ribs.

The standard also notes that the shotcrete ribs can be replaced by lattice girders.
These requirements correspond quite well to the installed rock support in the shl tunnel.

61

7 Conclusions
Author of this report had the privilege of working on site during the construction of the
shl tunnel and followed the installations procedure of lattice girders and other rock
support mentioned. Following conclusion are based on experience on site of shl tunnel
and investigations in this thesis:
Lattice girders are pre fabricated outside the tunnel and the time spent on building the
beam is therefore shorter at the face compared to the shotcrete ribs. Lattice girder are
therefore safer choose if the tunnel face is considered unstable. Shotcrete ribs are however
more convenient to use if the short time stability of the working area is sufficient since it
can better be adapted to irregular profile and therefore does not lead to high consumption
of shotcrete when covered.
The finite element modelling is quite helpful tool to estimate the force acting on tunnel
lining and should be reliable if required rock mass and stress data are available. The core
replacement method used here to derive a stiffness-deformation relationship worked well
and gave realistic results.
Lack of tri-axial data turned out to be a problematic factor and lead to very conservative
approach in the estimation of the rock mass parameters. Pre-geological investigations
should therefore include tri-axial testing on the rock samples at least on the rock mass that
is considered to be the most problematic during excavation.
Modelling of case 1 to 4 revealed that the highest moments occurred on the boundary of
sediment and basalt rock as expected. Weakest point for these moments is where the
squeezing effect is working perpendicular on the tunnel profile, in this case in the same
height as the radius point of the tunnel profile.
Changing the k-value from 0,5 to 1,5 tuned out to have quite dramatic influence on the
results in case 4. Stress measurements are almost inevitable if FEM-design is to be used as
a design tool for tunnels or other underground openings. Especially if the construction is
planned at low depth.
Bearing capacity calculations revealed that increased moments acting on the tunnel lining
should be dealt with by increased diameter of the lattice girder chords or decreased spacing
between girders. Tensile strength of shotcrete is low, even though mixed with steel fiber,
that the most effective way to deal with higher moments in the lining is to increase the
amount of reinforcement per tunnel metre.
Modelling of four cases was carried out. Once the M-N envelopes of the lining and rock
mass parameters had been estimated this turned out to be a relatively fast process. FEMmodelling could therefore be used on tunnel site to estimate required tunnel support in
special cases if M-N envelopes have been calculated for different types of linings in
forehand and parameters for various rock types estimated. Calculation of M-N envelopes
was however very time consuming and very sensitive for even smallest errors. M-N
envelopes should therefore not be calculated on tunnel site under time pressure and
thoroughly reviewed.

62

Bibliography
Baumann, T., & Betzle, M. (1984). Investigation of the performance of lattice girders in
Tunneling. Munich Germany: Philipp Holzmann AG.
Dahle, H. (2005). In situ rock stress in layered icelandic volcanic rocks-Fskrsfjrur
tunnel. Trondheim Norway: SINTEF.
Efla. (2011). shlargng-framkvmdaskrsla. Reykjavk: Vegagerin.
Einarsson, E. (2010). shlargng- steypuframleisla (in icelandic). Reykjavk: BM
Vall.
Eirksson, H. (2010). Lecture notes- Concrete structures 2. Reykjavk: Hskli slands.
Erlingsson, S. (2009). Lecture notes in Engineering Rock Machanics. Reykjavk, Iceland:
University of Iceland.
Eurocode 2. (2002). Eurocode 2: Design of concrete structures. Brussels: European
committee for Standardiziation.
Gumundsson, ., Kaiser, S. ., & Ward, T. (2007). shlargng, Astur til
gangagerar milli Hnfsdals og Bolungarvkur (in icelandic). Reykjavk: Jarfristofan
ehf.
Gumundsson, H. (25. 3 2011). e-mail refference. Iceland.
Haimson, B. C. (1981). Hydrofracturing Stress Measurements-Hole FV-1, Teigsbjarg.
Wisconsin U.S.A: Orkustofnun.
Haimson, B. C., & Rummel, F. (1981). Hydrofracturing stress measurements in the IRDP
drillhole at Reydarfjordur, Iceland. Wisconsin: Journal of Geophysical Research.
Hararson, B. A. (20. 9 2011). e-mail refference. Reykjavk , Iceland.
Hararsson, B. A. (1991). Tunneling Quality of Icelandic Rock. Reykjavk: JTS
Geotechnical Services Ltd.
Hoek, E. (2000). Practical Rock Engineering. Vancouver, Canada: Evert Hoek Consulting
Engineer Inc.
Hoek, E., Carranza, C., & al, E. (2002). Hoek-Brown Failure Criterion - 2002 Edition.
Vancouver.
Hoek, E., Carranza-Torres, C., & Diederichs, M. (2008). Integration of geotechnical and
structural design in tunneling. Vancouver, Canada: Evert Hoek Consulting Engineer Inc.
Jorimann. (2010). JB-Gittertrager. Walenstadt: Jorimann Stahl-Tunnelbau.
Kirkaldie, L. (1988). Rock classification systems for engineering purposes. Philadelphia:
ASTM.
63

Komselis, C., Blayney, N., & Hindle, D. (2005). The use of lattice girders in the
construction of tunnels. London, UK: London mining and Mineral Consultants LTD.
Loftson, M. (20. 9 2011). e-mail refference. Reykjavk, Iceland.
Loftsson, M. (2009). Kortlagning jarfri og styrkingum jargngum (in icelandic).
Reykjavk: Vegagerin.
Luo, J. (1999). A new rock bolt design criterion and knowlage-based expert system for
stratified roof. Virginia: Virginia faculty of the Virginia Polytechnic Institute.
Mannvit. (2009). Bolungarvkurgng - setbergslg- Samantekt um athuganir (in icelandic).
Reykjavk: Vegagerin.
Palmstm, A., & Nilsen, B. (2000). Engineering Geology and rock Engineering. Oslo,
Norwey: NBG.
Plsson, F. (2009). Enginnering geology of Oshlid tunnel, Iceland. London: Imperial
collage of London.
Pedersen, K., Kompen, R., & Kveen, A. (2010). Arbeider foran stuff og stabilitetssikring i
vegtunneler. Oslo, Norway: Statens Vegvesen.
Sigmundsson, D. F. (2006). Iceland geodynamics: crustal deformation of divergent plate
tectonics. London: Springer/Praxis.
Statens Vegvesen. (2000). Fjellbolting. Oslo, Norway: Statens vegvesen.
Statens vegvesen. (2009). Prinsipp bergsikring i tunnel- E6 Trondheim Stjrdal. Oslo:
Statens vegvesen.
U.S. Department of transportation. (2011). Technical manual for design and construction
of road tunnels. Washington DC: Federal Highway Administration.
Vegagerin. (2008). shlargng 4. hefti Byggingarvirki - Uppdrttir (in icelandic).
Reykjavk: Vegagerin.
Weisenberger, D. T. (2010). Iceland. Stt 26. 08 2011 fr DR. Tobias Weisenberger:
http://www.tobias-weisenberger.de/6Iceland.html

64

Appendix
A.1 Rock classification systems.
RMR-system

(Hoek, 2000)
65

Recommended parameter for Q-value

66

67

(Hoek, 2000)

68

Recommended rock support in Norwegian standard for certain Q-value

(Pedersen, Kompen, & Kveen, 2010)

69

A.2 GSI and D value in Hoek-Brown criterion

Estimation of GSI value in the Hoek-Brown failure criterion.

(Hoek, 2000)
70

Estimation of D- value in the Hoek-Browm failure criterion.

(Hoek, 2000)
71

A.3 shl tunnel: Geological mapping for st

16.300-16.450

(safl, 2010)
72

A.4 Bearing capacity of shotcrete lining.

Definitions of dimensions
b 1500mm

t 150mm
b1 780mm

h g 195mm

b2 330mm

b g 180mm

t2 95mm

l0 2000mm
h t t2

d1 16mm

h 0.225m

d2 20mm

dg h

30deg
co 30mm

d1

d2
2

dg 0.177m

Definitions of material properties

f cd 20MPa

f yd 435MPa

0.85

s1

s2

d1

d2

4 2
A s1 4.021 10 m

4 2
A s2 3.142 10 m

2
A c 0.176m

A c b t 0.5 t2 b2 A s1 A s2

( b b2) l 0.1 b2
b eff 2 0.2
0
2

d1
6mm
A s11 2 2 2

b eff 1.044m

b eff

4 2
A s11 5.989 10 m

150mm

t2
b eff t 0.5 t 0.5 b2 t2 t
3

h cm
b2 t2
b eff t

hcm 0.069m

73

Only in tension

Calculation of center of mass.

Pure axial force

N d f cd A c f yd A s1 A s2
3
N d 3.306 10 kN

Pure tension
N tog f yd A s11 A s2

N tog 397.187kN

Equilibrium stage (positive moment)

cu 0.0035
syd 0.00217
d h

d2
2

d 0.185m

cu
d
xbal
cu syd

0.8xbal t 1

xbal 0.114m

Fc f cd 0.8 xbal b eff

6
Fc 1.621 10 N

Fs1 f yd A s1

5
Fs1 1.749 10 N

Fs2 f yd A s2

5
Fs2 1.367 10 N

N ud Fc Fs1 Fs2

6
N ud 1.66 10 N

If 0 then adjust equation

of compression zone.

d1

Mud Fc hcm 0.4 xbal Fs1 hcm Fs2 d hcm

2

Mud 63.948m kN

74

Pure positive moment

x

f yd A s2
f 0.8 b
cd eff

x 9.625 10

Mu f yd A s2 ( d 0.4 x)

Mu 24.756m kN

Pure negative moment

d1

d 1 h
2

d1 0.217m

As11 fyd tan ()

f

cd

h c 0.094m

2 h

c
Mu2 f yd A s11 d1 3

Mu2 40.197m kN

Equilibrium stage (negative moment)

cu
d
xbal2
cu syd 1

xbal2 0.134m

beff 1.044m
hc2 0.8 xbal2

h c2 0.107m

hc2 2

Fc2 f cd tan ( )

5
Fc2 3.381 10 N

Fs11 f yd A s11

5
Fs1 1.749 10 N

Fs2 f yd A s2

5
Fs2 1.367 10 N

N ud2 Fc2 Fs2 Fs11

5
N ud2 2.143 10 N

2 h

d1
c2

Mud2 Fc2 h hcm 3 Fs2 d h cm Fs11 h cm 2

Mud2 60.387m kN

75

Zero strain in tensional bars (Positive moment)

h c3 d 0.8

h c3 0.148m

h h

c3

A c3 t beff 0.5 b2 t2 tan ( )

2
A c3 0.121m

2 h hc3
b2

tan ( )
2
t 0.5 h t h t
beff 0.5 t
c3 c3
2

h cm2

2 h h c3
b2

tan ( )
h t
beff t
2

c3
hcm2 0.06m
Fc3 f cd A c3
N ud3 Fc3 Fs1

6
Fc3 2.057 10 N
6
N ud3 2.232 10 N

d1

Mud3 Fc3 d h cm2 Fs1 d N ud3 d hcm

2

Mud3 28.218m kN

76

center mass of
compression zone

Zero strain in tensional bars (negative moment)

h c4 d1 0.8

Height of compression area.

h c4 0.174m

A c4 h c4 t2 b eff t2 b2 0.5

2
A c4 0.078m

Fc4 f cd A c4
N ud4 Fc4 Fs2
6
N ud4 1.456 10 N

h t2
t2
h t2 b h h c4
t2 b2 0.5 t
eff c4
2
3
c4

h cm3

hc4 t2 beff t2 b2 0.5

hcm3 0.099m

Center mass of compression zone.

d2

Mud4 Fc4 h cm3 0.5 d1 Fs2 d1 co N ud4 h cm 0.5 d1

2

Mud4 55.171m kN

77

ud3
M
ud
M
u
Moment
0

M
u2
M
ud2
Mud4

0
28.218
63.948

24.756
Moment
0 m kN
40.197

60.387
55.171
0

N d
N
ud3
N ud

0
Nforce N tog

0
N

ud2
N
ud4
N d

410

2.810

Nforce 1.510

2.510

3.306 103

3
2.232 10

1.66 103

Nforce 397.187 kN

214.258

3
1.456 10

3
3.306 10

110
5
110

510

510

Moment

78

110