LULEL

2001 :64
UNIVERSITY
OF TECHNOLOGY

Shotcrete Rock Support Exposed

to Varying Load Condi�ions

LARS MALMGREN

Department of Civil and Mining Engineering

Division of Rock Mechanics

2001 :64 • ISSN: 1402 - 1757 • ISRN: LTU - LIC - - 01/64 - - SE

PREFACE

The research work presented in this thesis was carried out at the Department of Civil and Mining
Engineering, Division of Rock Mechanics, Luleå University of Technology, Sweden. The field
work was done in the Kiirunavaara mine, LKAB, Sweden and the laboratory tests at Test Lab at
the Department of Civil and Mining Engineering, Luleå University of Technology.

This work is also a part of the research consortium "Väg/Bro/tunnel", the research consortium is
supported by Vinnova, Cementa, Elforsk, LKAB, NCC, Skanska, PEAB and SBUF. The
financial support from LKAB and the research consortium made this research work possible.

Professor Erling Nordlund at the Division of Rock Mechanics, Luleå University of Technology,
Sweden has been my supervisor, I am gratefully acknowledged for his support and trust.

I am thankful to Dr. Chunlin Li and Mr. Tomas Svensson, for the help with the laboratory tests
and field tests respectively, and Professor Bill Hustrulid for important help during the initiation
of the project.

I will also thank all my colleagues at the Department of Mining research at LKAB for help with
field tests, technical and scientific guidance.

Finally, special thanks to my beloved wife Marit and our children Therese, Andreas and
Ingeborg who have made this work possible with their patience and support.

Kiruna, November 2001

Lars Malmgren
 11

ABSTRACT

Field and laboratory tests and analyses

This Licentiate thesis deals with the function of shotcrete as rock support. An extensive failure
mapping of shotcrete in the Kiirunavaara mine has been performed to improve the understanding
of the performance of shotcrete. Furthermore, the relationship between adhesion and
compressive strength and the adhesion strength obtained for different types of scaling/cleaning
methods were investigated in field tests in the mine.

The failure mapping showed that most of the observed failures of shotcrete are in areas with a
thin shotcrete layer (<2 cm) together with a low adhesion strength. Because of that, it is very
important to have a well-cleaned surface and to get sufficient thickness all over the shotcreted
area i.e. to avoid areas with thin shotcrete. Furthermore, the growth of the adhesion strength of
shotcrete on a sandblasted concrete wall and the growth of the compressive strength of shotcrete
showed a clear correlation. The results from the field tests showed that the adhesion strength was
significantly higher on rock surfaces that had been water-jet scaled (with 22 MPa water pressure)
than those treated by mechanical scaling followed by cleaning of the rock surface (water pressure
0.7 MPa).

Higher rock stresses in the Kiirunavaara mine will probably require a more ductile surface
support, which will increase the use of reinforced shotcrete. Panels and beams of reinforced
shotcrete have been tested. The following issues were addressed, identification of failure modes
of the shotcrete panels, comparison of the bearing capacity of shotcrete panels reinforced with
two types of steel fibres and two dosages of steel fibres. Furthermore, comparison of the bearing
capacity of steel fibre reinforced shotcrete with shotcrete reinforced with welded steel mesh.

The test results indicate that the typical failure mechanism was a bending failure followed by
punching failure. Furthermore, the tests showed that panels reinforced with steel fibres were
more ductile and had higher punching capacity than panels reinforced with steel mesh. The effect
of dome action was apparent in the tests and the measured bearing capacity was much higher
than that calculated using yield line theory.

Vibration measurements and dynamic analyses

Rock support is mostly designed for static loads. In many mines, however, the openings are
subjected to dynamic loads caused by blasting. The behaviour of the dynamic load from the
production blasting has been investigated by vibration measurements. To study the behaviour of
shotcrete exposed to the load induced by the stress wave generated by the production blasting a
single degree of freedom model (SDOF) was developed. The model consisted of a shotcrete
layer and a symmetric rock prism. The recorded vibration was used as the external disturbance in
the analysis. Observations in the field (failure mapping) have been done to identify failure
mechanisms and the validity of the model.

The vibration measurements showed a wide scatter of magnitudes of the particle acceleration,
which was expected. The maximum particle velocity was 1.1 m/s at a horizontal distance of 4.5
m away from the blast holes.

The results from the linear elastic analyses, presented as a displacement response spectra,
showed that a prism can be ejected from the roof as a result of production blasting, since the
 iii

displacement was larger than the displacement at failure of the joints of the prism. This was also
observed by the failure mapping of the roof of the cross cuts close to the drawpoint.

Furthermore, the non-linear analyses showed that the reinforced shotcrete had the necessary
bearing capacity to support actual prisms for load histories measured at a horizontal distance of
4.5 m away from the blasted fan. The failure mapping showed that failures in the reinforced
shotcrete were observed at a horizontal distance of 1- 4 m away from the drawpoint. The analysis
also showed that plain shotcrete was too brittle. Field observations confirmed that the area or
zone of failed shotcrete was much larger in areas with plain shotcrete compared to areas with
reinforced shotcrete.

Key-word's:
Rock mechanics, tunnels, shotcrete, failure mapping, adhesion strength, water-jet scaling,
vibrations, blasting, dynamic analysis, beam test, panel tests, dome effect.
 iv

SAMMANFATTNING

Fält och laboratorieförsök med tillhörande analyser

I denna licentiatavhandling har sprutbetongens funktionen som bergförstärkning studerats. För

att undersöka hur sprutbetongförstärkningen fungerar idag har en omfattande skadekartering
utförts i LKAB:s Kiirunavaaragruva. Vidare har sambandet mellan sprutbetongens
tryckhållfasthet och vidhäftningshållfasthet studerats i fähprov. Även skrotningens och
rengöringens betydelse för vidhäftningen har undersökts i fältprov.

Skadekarteringen visade att de vanligaste utfallen var tunn sprutbetong (<2 cm) på bergytor med
dålig vidhäftning. Detta visar hur viktigt det är att vidhäftningen är god och sprutbetongen har
tillräcklig tjocklek. Vidare visade resultaten att det var en klar korrelation mellan sprutbetongens
tryckhållfasthet och sprutbetongens vidhäftningshållfasthet som funktion av tiden, vidhäftningen
provades mot en betongyta. Resultaten från fältprovningarna visade att vattenskrotning (22 MPa
vattentryck) gav signifikant bättre vidhäftning mellan sprutbetongen och berg än normal
mekanisk skrotning där bergytan rengjordes med vatten med normalt vattentryck (0.7 MPa).

Högre bergspänningar i Kiirunavaaragruvan kommer antagligen att kräva en ytförstärkning med

högre seghet, vilket kommer att öka användningen av armerad sprutbetong. För att undersöka
sprutbetongens brottformer främst kring bergbulten och att jämföra olika armeringar har
omfattande platt- och balkförsök utförts. Sprutbetong armerad med stålfiber med varierande
längd och kvantitet har provats. Vidare har sprutbetong armerad med avspänningsglödgad
nätarmering jämförts med sprutbetongen armerad med stålfiber.

Provresultatet visar att den typiska brottformen för plattorna var böjbrott följt av
genomstansning. Plattor armerade med stålfiber hade högre bärförmåga än plattor armerade med
nätarmering. Effekten av kupoleffekten var tydlig i plattproven, bärförmågan överskred den
beräknade bärförmågan enligt brottlinjeteori.

Vibrationsmätningar och dynamisk analys

Bergförstärkning är vanligtvis dimensionerad för statiska laster, men i många gruvor utsätts
förstärkningen för dynamiska laster till exempel i form av vibrationer från
produktionssprängning. Vibrationsmätningar har utförts på ett horisontellt aystånd av 4.5 till 8 m
från borrkransen för att undersöka vibrationernas amplituder och frekvensinnehåll. För att
undersöka den dynamiska bärförmågan för sprutbetong belastad av vibrationer från
produktionssprängningarna utvecklades en dynamisk modell med en frihetsgrad. Modellen
bestod av ett sprutbetonglager och en symmetrisk bergkil med de uppmätta vibrationssignaler
som extern störning. I tillägg till vibrationsmätningarna och analysen gjordes också
skadekartering av området nära produktionssalvorna för att bestämma brottformer i berg och
förstärkning samt för verifiering av den numeriska modellen.

Resultatet från vibrationsmätningarna hade som väntat stor spridning, den största
partikelhastigheten var 1.1 m/s på ett horisontellt aystånd av 4.5 m från den borrkransen.
Resultatet från dorn linjär-elastiska beräkningarna, presenterade som förskjutningsspektrum,
visade att bergkilar kan stötas ut på grund av vibrationerna vilket också verifierades med
skadekarteringen. Vidare visade den olinjära numeriska analysen att armerad sprutbetong hade
tillräcklig bärförmåga att bära bergkilen utsatt för den dynamiska belastningen på ett horisontellt
aystånd av 4.5 m från borrkransen. Skadekarteringen visade att skador på armerad sprutbetong
uppstod på ett horisontellt aystånd av ca 1 - 4 m ifrån brynet. Analysen visade också att oarmerad
sprutbetong hade otillräcklig förmåga att bära en bergkil utsatt för vibrationer nära
produktionssalvan, oarmerad sprutbetong är för spröd. Detta verifierades också i
skadekarteringen.
 vi

LIST OF SYMBOLS

a,a„ Normal stress

al = Major principal stress
az = Intermediate principal stress
a, = Minor principal stress
(3-'1 = Major principal effective stress
cr '3 = Minor principal effective stress
Co = Tangential stress
ac = Uniaxial compressive strength of the intact rock
0-cm = Uniaxial compressive strength of the rock mass
ah = Horizontal stress parallel to the ore body
ax = Horizontal stress perpendicular to the ore body
(i. = Effective normal stress
as, = Vertical stress
4 = Distance to the drawpoint, damping ratio
1" = Shear stress
•rf = Shear stress at failure
zn, = Mean bond strength at failure between fibre and matrix
(I) = Friction angle
Or = Residual friction angle
(5 = Beam or panel deflection
åcr = First crack deflection of beam or panel
&max Beam deflection caused by the load Fmax
A = Slenderness of a panel or slab
e = Apical angle of crack pattern of a panel
4 = Damping ratio
co = Natural frequency (rad/s)
a = Semi-apical angle of wedge
P = Density
0 = Angular deformation
er = Angular deformation, radial cracks
Ot = Angular deformation, tangential cracks
A = Cross section
a.J = Average aperture thickness
b = Width of wedge
bb = Width of test beam
c = Cohesion
C = Damping coefficient
Ccrit = Critical damping coefficient
d = Diameter of fibre
E = Young's modulus for the intact rock
E, = Young's modulus for shotcrete or concrete
Em = Young's modulus for the rock mass
F = Force
f = Natural frequency = clY2ir (Hz)
fs,ro, fio,2o• • • = Residual strength
fa = Adhesion strength between shotcrete and rock
 vii

f = Compressive strength of shotcrete or concrete

F„ = First crack load
Fdamp, Fdamp(t) Damping force
D FID = Damping force
Fo
Ff = Load at failure
hie, = First crack strength in shotcrete/concrete beam tests
fibs = Ultimate flexural strength shotcrete/concrete beam tests
Fol ,F,' Inertial force
Fi = Spring force of the joints
Fif = Bearing capacity of the joints
F„,,,, = Maximum load
F, = Spring force of the support
s , Fis = Spring force
Fo
fsh = Shear strength or cohesion between new shotcrete or concrete and old
concrete
Fspring, Fspring(t) = Total spring force
F, = Yield load capacity of the support
fü' = Tensile strength of the fibre
fr. = Tensile strength of the matrix
Fi0,(0) -= Total load applied on the test beam as a function of angular deformation
Fy = Yield load capacity
H = Horizontal force
h = Height of a wedge
hb = Height of beam
hp = Height of panel
Ho = Horizontal force caused by rock stresses
1 = Area moment of inertia, impulse
15, 110, 120 • • • = Toughness indices for beam tests, defined in the ASTM C1018 standard
JCS = Joint Wall Compressive Strength.
JRC = Joint Roughness Coefficient
K, K(u) = Total stiffness = KJ+K,
Ki = Stiffness of the joints
Kin = Normal stiffness of the joints
Kis = Shear stiffness of the joints
kni = Initial normal stiffness (per unit area)
kii = Normal stiffness of the joint (per unit area)
K, = Stiffness of the support
k, = Shear stiffness of the joint (per unit area)
L = Anchor length of a fibre
L = Span of beam or panel
le = Critical anchor length of a fibre
icy! = Length of cylindrical shaped specimen
Li = Joint length
Lf = Failure zone of shotcrete close to the drawpoint
Lin = Retreated distance
m = Mean value
M = Mass of the wedge
m(0) = Moment capacity, per unit length, of the panels as function angular
deformation
M(0) = Moment capacity of the test beam as a function of angular deformation
 viii

mr = Moment capacity, per unit length, of the panels, along radial cracks
m(0) = Moment capacity, per unit length, of the panels as function angular
deformation, along radial cracks
mt = Moment capacity of the panels, per unit length, along tangential cracks
inded = Moment capacity, per unit length, of the panels as function angular
deformation, along tangential cracks
n = Number of samples
No = Normal joint force
AIV = Normal joint force caused by a forced displacement of the wedge
PPV = Peak Particle Velocity
r = Radius of crack pattern of panels
R5,10, R10,20 • • • = Residual strength factor, defined in the ASTM C1018 standard
RQD = Rock Quality Designation
s = Material constant for the rock mass, standard deviation
So = Shear joint force
AS = Shear joint force caused by a forced displacement of the wedge
t = Student's distribution
At = Time increment, numerical analysis
it, u(t) -= Displacement
Ug = Ground motion
Ilif = Displacement at failure of the joints
Un = Normal displacement in the joint
us = Shear displacement in the joint
usf = Displacement at failure of the support
u, = Yield displacement of the support
u = Velocity
ü = Acceleration
W = Energy
Welastic = Elastic energy
We = External energy
WI = Internal energy
Vf = Fibre content
vf cru = Critical fibre content
V, = Maximum closure of the joint
VOD = Velocity of Detonation
x, y = Co-ordinate
Z, z = Co-ordinate ,depth
 ix

TABLE OF CONTENTS

PREFACE

ABSTRACT ii

SAMMANFATTNING iv

LIST OF SYMBOLS vi

TABLE OF CONTENTS ix

1 INTRODUCTION 1
1.1 Background 1
1.2 Objectives and scope of the work 3

2 SHOTCRETE AS ROCK SUPPORT - A LITERATURE REVIEW 5

2.1 General 5
2.2 Shotcrete as rock support 6
2.2.1 General 6
2.2.2 Shotcrete subjected to vibrations 11
2.3 Mix design 12
2.3.1 General . 12
2.3.2 Cement, aggregates and additives 13
2.3.3 Reinforcement 16
2.3.4 Experience with steel fibre reinforced shotcrete and mesh
reinforced shotcrete at LKAB:s underground mine in Malmberget 18
2.4 Shotcrete technique 19
2.4.1 Cleaning of rock surfaces 19
2.4.2 Spraying 19
2.4.3 Curing 20
2.4.4 Training 20
2.5 Mechanical properties .20
2.5.1 Strength between rock and shotcrete 20
2.5.2 Beam test 24
2.6 Design criterion for shotcrete 31
2.6.1 The New Austrian Tunnelling Method (NATM) 31
2.6.2 Rock mass classification 32
 x

2.7 Experimental studies, panel tests 33

2.7.1 Shotcrete as rock support for falling stones . 33
2.7.2 Shotcrete as rock support in weak or very jointed rock 36
2.7.3 EFNARC panel tests and the Round Determinate panel 41
2.7.4 Dynamic testing of retaining elements .42

3 THE KBRUNAVAARA MINE 46

3.1 General 46
3.2 Regional geology .47
3.3 Mechanical properties of intact rock 49
3.4 Rock mass structures 49
3.5 Mechanical properties of discontinuities 50
3.5.1 Typical values at the Kiirunavaara mine 50
3.5.2 Normal stiffness 51
3.5.3 Shear stiffness 51
3.5.4 Shear strength 52
3.6 In-situ stresses 52
3.7 Mining method — sublevel caving 53
3.8 Rock support 55
3.9 Stress changes caused by the sublevel caving . 56

4 SHOTCRETE — FIELD AND LABORATORY TEST 60

4.1 General 60
4.2 Failure mapping of shotcrete 60
4.3 Adhesion strength and compressive strength of shotcrete as a function of time 63
4.3.1 General 63
4.3.2 Compressive strength tests 63
4.3.3 Adhesion strength tests 64
4.3.4 Comparison between adhesion strength and compressive strength 66
4.4 Adhesion strength of shotcrete when using different types of
treatment methods 67
4.4.1 Scaling and cleaning methods 67
4.4.2 Results 68
 xi

4.5 Panel and beam tests 70

4.5.1 Test method and program 70
4.5.2 Steel fibre rebound 73
4.5.3 Compressive strength 74
4.5.4 Beam tests 74
4.5.5 Panel tests 77
4.5.6 Bearing capacity of the panels determined by the yield line theory 79
4.6 Discussion 83
4.6.1 Field and laboratory tests in the Kiirunavaara mine .83
4.6.2 Panel and beam tests 84

5 VIBRATION MEASUREMENTS IN THE KIIRUNAVAARA MINE WITH

BELONGING FAILURE MAPPING 86
5.1 General 86
5.2 Production blasting 88
5.3 Explosives used in the production blasting . 89
5.4 Acceleration recordings in the cross cut 89
5.5 Results 91
5.6 Failure mapping of the drawpoint 96
5.7 Discussion 97

6 THE RESPONSE OF SHOTCRETE DUE TO VIBRATIONS CAUSED

BY BLASTING
99
6.1 Introduction .99
6.2 The dynamic behaviour of a symmetric roof prism caused by blasting 100
6.3 The behaviour of the roof prism 102
6.4 The behaviour of the shotcrete lining 108
6.5 Dynamic load 109
6.6 Results 111
6.6.1 Static bearing capacity for the roof prism 111
6.6.2 Natural frequencies for the roof prism 113
6.6.3 Dynamic analyses 114
6.6.4 Shotcrete as rock support after failure has occurred
in the rock joints 121
6.6.5 Ductility of the shotcrete lining 122
6.7 Discussion 123
 xii

7 DISCUSSIONS WITH CONCLUSIONS AND PROPOSAL FOR FURTHER

RESEARCH AND DEVELOPMENT 124
7.1 General 124
7.2 Shotcrete — field and laboratory tests .124
7.2.1 Field and laboratory tests performed in the Kiirunavaara mine .124
7.2.2 Panel and beam tests 125
7.3 Shotcrete subjected to dynamic loads from production blasting 126
7.4 Proposal for further research and development . 127
7.4.1 Research 127
7.4.2 Development 128

REFERENCES .129

APPENDIX A4.1 Results — panel tests

APPENDIX A4.2 Yield line theory and the moment capacity of test beams
expressed as a function of angular deformation, 9

APPENDIX A6.1 Numerical integration

APPENDIX A6.2 Shorter time increment for the numerical integration

APPENDIX A6.3 The Duhamel integral equation

 1

I INTRODUCTION

1.1 Background

Shotcrete is widely used as rock support in mines and in civil engineering projects. It is applied
through a process by which concrete or mortar is sprayed onto a surface to produce a compacted
self-supporting and load bearing layer. Depending on the addition of water to the mix, distinction
is made between the dry and the wet process. The quality of the final shotcrete layer is closely
related to the application procedures used. Some important procedures are surface preparation,
shotcrete mix, spraying technique, curing and crew training.

During the last decades the introduction of silica and fibre reinforcement are one of the most
important factors for the successful use of shotcrete as rock support (Hoek et.al. 1995). Silica is
an extremely fine pozzolan and is a by-product of the ferrosilicon metal industry. Silica makes it
possible to increase the thickness of the shotcrete layer, increase the strength and improve the
resistance to alkali aggregate reaction and improve the sulphate resistance (Melbu, 2001).
Reinforcement by fibres, (steel fibres are the most used fibres today), has advantages compared
with traditional mesh reinforcement. The production fits better into the tunnelling as it eliminates
the heavy and time consuming manual application of mesh (Franzén, 1992). It is also easier to
adapt fibre-reinforced shotcrete to an irregular surface compared to the steel mesh in a shotcrete
lining.

The design of a shotcrete lining is an imprecise process, and a large number of questions
concerning the behaviour of shotcrete in interaction with rock and rock bolts remain. This mainly
depends on the heterogeneous nature of the rock. Therefore, the support design is often based on
common practice or classification systems. However, in addition to the experience many load
conditions have been analysed by experimental simulations and been used as guidelines for the
design. Loose rock, which requires immediate support, makes the use of shotcrete in short-term
stabilisation clearly inappropriate because of the low strength of young shotcrete.

Generally one can say that the development of the shotcrete and the application of shotcrete have
gone faster than the understanding of the behaviour of shotcrete as rock support. One reason is as
mentioned above the complex interaction between shotcrete and rock. Another and a more
important reason is the fact that shotcrete has worked very well to stabilise failing rock and this
experience has driven the application of shotcrete faster than the development of the
understanding of the mechanism of shotcrete as rock support.

At the Kiirunavaara and the Malmberget underground mines, which are owned and operated by
the Luossavaara Kiirunavaara Ltd, (LKAB), shotcrete has been used as rock support for more
than 20 years. Increasing depths in the underground mining will set new demands on the rock
support, both technical requirements how to reach the desired stability of the openings as well as
cost efficient solutions.

The mining method is large-scale sublevel caving, see Figure 1-1, which has been the
dominating mining method in the mines. The mining induced stress can both increase or
decrease the rock stress and can, in addition change the direction of the principal rock stresses.
One consequence of this is that the temporary support for drifting is not the ideal support for the
following mining operations, the support must be designed for the new load conditions which
 2

occur with the mining. This can appear obvious but because ofinsufficient lack ofknowledge of
how the shotcrete behaves in interaction with the rock and rock bolts and because of
uncertainties in the rock stress state and the properties ofthe rock mass this is a difficult design
process.

Waste dilution and ore losses are the drawbacks for sub-level caving. Extensive research and
development have been done to determine the ore flow in the cave, to identify ways to reduce ore
losses and minimise waste dilution. One important factor to get a proper mass flow is to keep the
profile ofthe cross cut intact. Close to the drawpoint, the dynamic influence from the production
blasting can be significant. Therefore, ifthe rock around the cross cuts must be supported it must
also be supported to withstand the dynamic loads from production blasting

With respect to the relatively imprecise design process, the low strength ofyoung shotcrete and
the problems ofinsuring the right quality ofthe shotcrete, there are still many questions to be
solved by the researcher, designer and the practician. Because ofthe challenges and problems
discussed above and the fäet that the knowledge ofhow shotcrete acts as rock support is limited
a project was initiated to (i) investigate the function ofshotcrete as rock support and (ii) develop
ofshotcrete support for varying load conditions. The second part ofthe project (ii) dealt among
other things with the behaviour ofshotcrete when subjected to dynamic load from the production
blasting. The project is ongoing and this thesis report is part ofthe project.

a) b)

___
�/
-.... ---�.,--:..:..

Figure 1-1 Large-scale sublevel caving in the Kiirunavaara mine

a) Large-scale sublevel caving, principles, after Hamrin (1986)
b) Fans for production blasting
 3

1.2 Objective and scope of the work

The objective of the thesis is to

increase the understanding of shotcrete as rock support
determine the load conditions caused by production blasting
investigate how shotcrete acts as a rock support for rock subjected to dynamic loading
induced by the production blasting

To fulfil the objective a number of activities have been performed. Following this introduction, a
literature review of shotcrete is presented in Section 2.

The geological model for the Kiirunavaara mine is briefly described in Section 3. Assessments of
representative rock properties from the Kiirunavaara mine for the further analyses are presented
in the same Section. The use of shotcrete is not only depending on the rock conditions but also
on the mining method. Therefore, the large-scale sublevel caving used in the Kiirunavaara mine
is presented.

Section 4 covers field and laboratory tests of shotcrete. An extensive failure mapping was done
to identify typical failures of shotcrete. About 7 km of drifts were mapped. The damage to the
shotcrete was classified and important parameters, such as thickness, dimension of fallouts and
geology, were noted.

Many tests have been done on the growth of the compressive strength of shotcrete with respect to
time, but the author has not found very much information about the growth of the adhesion
strength of shotcrete with respect to time. Therefore, tests have been performed which deal with
the compressive and adhesion strength of shotcrete as function of curing time. The tests were
performed underground in the Kiirunavaara mine.

In the Kiirunavaara mine, scaling is performed in all drifts and cross cuts, normally with a
hydraulic scaling hammer. Before shotcreting, the rock surface is cleaned by spraying water on it
with a pressure of 0.7 MPa (7 bar). During the past few years, water-jet scaling has been tested
as an alternative method to mechanical scaling. The adhesion strength of shotcrete on water-jet
scaled rock surfaces is compared with shotcrete on rock with normal treatment.

A rock mass is not a well-defined material. The actual loads on a rock support system are often
unknown. However, knowledge of the supporting effect can be obtained by studying models of
shotcrete linings in laboratory tests. Depending on the rock quality and the rock stress situation,
various test models can be used. Many of the various test methods are discussed in the literature
review. To supplement the literature review the bearing capacity of the reinforced shotcrete has
been tested by panel and beam tests at Luleå University of Technology.

Section 5 presents a field measurement programme which was carried out to estimate to what
extent the rock support in the openings were subjected to destructive dynamic load from
production blasting. In addition to the vibration measurements, failure mapping close to the
drawpoint was carried out to investigate the influence of the dynamic loading induced by the
production blasting on the stability of the roof.

To study the dynamic response of shotcrete support, a number of dynamic analyses have been
done. A single degree of freedom (SDOF) model of a symmetric roof prism subjected to stress
 4

waves induced by blasting has been developed and the results from the analyses are discussed
and compared with mapped failures. The dynamic analysis is presented in Section 6.

Finally, conclusions from the conducted research and recommendations for future work are
presented in Section 7.
 5

2 SHOTCRETE AS ROCK SUPPORT - A LITERATURE REVIEW

2.1 General

The original process of spraying dry-mix mortar was given the name "Gunite" (Vanderwalle,
1994). In the early 1930's the American Railway Association introduced the term "shotcrete".
In 1966, the American Concrete Institute (ACT) applied the term "shotcrete" to all
pneumatically applied mortar and concrete, both the dry-mix and the wet-mix processes. In
some European guidelines and standards the term "sprayed concrete" is used instead of
"shotcrete "for example in the Norwegian guideline NCA No 7 (1999) and EFNARC (1999).
In the ongoing harmonisation of a European standard for shotcrete the term "Sprayed
concrete" has been agreed as a general term for the technique. However, in this thesis, the
term "shotcrete" will be used.

Shotcrete has two primary application techniques - wet-mix and dry-mix, see Figure 2-1. Wet-
mix shotcrete is mixed as a low slump concrete, which is pumped to the nozzle. Accelerator
and compressed air is added in the nozzle. Dry-mix shotcrete is mixed dry, water and
accelerator is added at the nozzle. Table 2-1 shows the advantages/disadvantages for the two
types of shotcrete application techniques.

Shotcrete has been used as rock support in underground excavations for many years especially
in civil works but also in underground mines. At LKAB' s underground mines in Malmberget
and Kiruna (Sweden) shotcrete has been used as rock support for more than 20 years.

a) b)

Dry mix
Wet mix
Compressed air
Air - —

Accelerator Water + liquid

accelerator/additives

Figure 2-1 A schematic representation of the two primary application techniques for
shotcrete
a) Wet-mix method
b) Dry-mix method
 6

Table 2-1 Wet-mix or dry-mix after Brady and Brown (1999) and Vandewalle (1994)
Properties Wet-mix Dry-mix

Rebound, Lower rebound when Rebound can be 15-40 % on walls and 20-50% on
shotcrete spraying, less than 10% roof
Rebound, 12 - 18 % ' 35 - 78 % u)
steel fibres
Dust 1/6 of the amount of dust -
from dry-mix
Quality Control of water-cement Less homogeneous quality as the water addition is
ratio regulated by the operator and discontinuous
material supply

Higher strength, due to lower W/C-ratio

More difficult to obtain
high strength due to
higher W/C-ratio b)
Equipment Dry-mix equipment is less expensive

Capacity Higher spraying capacity Rarely exceeds 5 m3/h in place

than similar dry-mix
equipment, 2-10 m3/h
with handheld nozzle up
to 20 m3/h with
manipulator
Impact General adequate for Higher impact velocity, gives better adhesion
velocity tunnel and mining works
a) Banthia etal.(1994)
b) W/C-ratio is water/cement —ratio

2.2 Shotcrete as rock support

2.2.1 General

The main design principle for rock support is to help the rock carry itself. Shotcrete can be
applied with and without reinforcement. In the case of plain shotcrete, see Figure 2-2, a loose
block is held by the shotcrete. For this case, the adhesion between the shotcrete and the rock
mass is one of the most important properties. Also the effect of sealing the joints by shotcrete
helps to maintain the integrity of the rock mass, (Morgan & Mowat, 1984). Failure normally
develops as an adhesion failure and at a deformation of only a few millimetres, Stille (1992).
According to Swan et.al. (1996) shotcrete in contrast to screen can prevent loosening because
of higher stiffness and because it isolates joints from air and water.
 7

/1 \ Wedge
/ \
1
I

Perimeter of the wedge

Drift
Shotcrete

Figure 2-2 Support of a loose or key block by shotcrete

If ductility is required the shotcrete must be reinforced. The reinforced shotcrete is very often
used together with rock bolts. In this case the function of the shotcrete is to give stability to
the rock between the rock bolts. The interaction between surface support, rock and rock bolts
is complex. Fredriksson and Stille (1992) have described this situation as shown in Figure
2-3. In the first case, Figure 2-3 a), the rock is arching between the rock bolts and the loose
rock below the arch is carried by the surface support. The arch is obtained by adhesion
between the rock, the grouting and the rock bolt and support from the washer of the rock bolt.
The surface support must have necessary stiffness to avoid disintegration of the rock between
the bolts. If not, large deformations can arise in the surface support, this may lead to a
situation where the rock bolts cannot support the rock as planned. The load will be transferred
to the surface support, which is not designed for this load and unstable conditions will occur.
In the second case, Figure 2-3 b), the rock bolt is designed to carry the load from the surface
support into stable rock and is therefore not dependent of the stiffness of the surface support.

a) b)

ARCHING SUSPENSION

Stable condition Unstable condition Stable condition

surface support failure Stable condition
roof failure

Figure 2-3 Support functions (Fredriksson & Stille, 1992)

The example above shows the importance of the interaction between the support and the
ground. The rock does not only act as a load but also as a bearing element. Therefore, the
ground condition is very important for the design of the support. Figure 2-4 gives a simplified
 8

description of the various types of failure which are observed in underground structures. In
the following there is a presentation of how shotcrete can work as a rock support for some of
the ground conditions shown in Figure 2-4.

Low stress levels High stress levets

.ll:
u
E
�
'iii
CII
IIS
:E
@]
Massivc rock subjected to low in situ
@]
Massive rock subjected to high in situ
stress lcvels. Spe.lling, slabbing and
crushing initiates at high stress
stress levels. Lincar clastic rcsponse concentration points on the boundary
with little or no rock failure. and propegates into tlte surrounding
rockmass.

.ll:
u
E �

Massive rock, witlt relatively few dis-

[j
Massive rock, with relatively few dis-
·o continuities, subjected to low in situ
stress conditions. Blocks or wedges,
continuitics, subjected to high in situ
stress conditions. Failure OCCIIIS as a
result of sliding on discontinuity
released by inter.lecting discontinuities, surfaces and also by crusbing and
fall or slide due to gravity loading. splitting of rock blocks.

a Il
.ll:
u
,,E
·o
.....
C

>, Heavily jointed rock subjectcd to low Heavily jointed rock subjected to high
�
:
in situ stress conditions. The opening in situ stress conditions. The rock mass
swfacc fails as a result of umavelling surrounding lhe opening fails by slid-
% of small interlocking blocks and ing on discontinuitics and crusbing of
wedges. Failure can propagate a 1"111! rock pieccs. Floor heave and sidewall
way into the rock mass if it is not closwe are typical rcsults of !bis typc
controlled. of failure.

Figure 2-4 Types of failure modes, which occur in different rock masses (after Hoek,
Kaiser and Bawden, 1995)
 9

Jointed rock with low stresses

For low stress magnitude, the main function for a rock support, such as shotcrete, is to prevent
a key block or loose blocks falling or sliding due to gravity. If the loose blocks are not
supported this failure process will continue until natural arching in the rock mass prevents
further unravelling or until the opening is full of caved material. The ideal application of
shotcrete is for (more) closely jointed rock masses where wedge failure would occur as a
progressive process, starting with smaller wedges exposed at the excavation surface and
gradually working its way back into the rock mass. In this situation, rock bolts alone can not
stabilise the rock mass, a surface support or lining must be added. The low strength of young
shotcrete is of course a problem when supporting a wedge. The wedge requires immediate
support, and the use of shotcrete in short term stabilisation is clearly inappropriate.

Rock with few joints and low stresses

Shotcrete should mainly be used as a surface rock support (Holmgren, 1992). In case of larger
blocks other types of support should be used, for example rock bolts. However, shotcrete can
be used to seal the joints and in that way help to maintain the integrity of the rock mass.

Highly stressed rock

If the rock is heavily jointed, the failure is more ductile compared with failure in the case of
rock with few joints. The heavily jointed rock gives freedom for individual rock pieces to
translate and rotate within the rock mass. This gives a more plastic behaviour of such rock
masses. In this case the shotcrete (if designed in the right way) will not only act as support for
a single loose wedge (a kinematic constraint for a wedge) but also act as a global support for
the whole tunnel section, see Figure 2-5. The load on the shotcrete ring can be determined by
the ground reaction curve concept (Hoek & Brown, 1980) which takes into consideration the
interaction between the shotcrete ring and the rock.

Radial
displacement
A = Po) excavated
profile

tunnel profile
Radia l support p

support pressure pi
7--support reaction
required support line
Zfor tunnel roof H —
s
required support line
,/for tunnel side wall
G
radial displacement, 8,

Shotcrete ring (Hoek & Brown, 1980) The radial support pressure-displacement curves for
the rock mass and the support system (Brady &
Brown, 1999)

Figure 2-5 A circular tunnel section supported by shotcrete

 10

Figure 2-6 shows how effective a closed ring is compared to an open ring. If the ring is open
the support capacity decreases dramatically. However, when the rock stresses increase the
closed shotcrete ring can be over stressed and not strong enough to stabilise the imposed
deflection (Kaiser and Tannant, 1997). In this situation, shotcrete must be used as a retaining
element rather than as a support arch. (In many applications such as mining a closed ring does
not apply for practical reasons.) Therefore, to avoid failures in the shotcrete in highly stressed
rock it is necessary to ensure that the shotcrete is able to deform relatively freely and therefore
the shotcrete must not be applied as a stiff ring or arch, but as a retaining element, see Figure
2-7.

10
1."' • high :
L support capaclty - Thickness Properties for shotcrete:
1mm)
E,=25 GPa (Young's modulus)
e.a f,=35 MPa (Compressive strength)
_ .. ,,.....................
_ ;

..
.....
0.01
. -.......
1 2 3 4 5 e 7
Tunnel radius [m]

Figure 2-6 Support capacity of closed and open circular support of different thicicness,
(Kaiser and Tannant, 1997)

Stiff ring Stiff arch

Flexible shotcrete
panels

Figure 2-7 Shotcrete as a support ring, support arch or flexible panels (Kaiser and Tannant,
1997)

When the shotcrete layer acts as a retaining element Kaiser and Tannant (1997) proposed a
simplified method to estimate the load on the shotcrete: The load acting on the shotcrete layer
can be estimated as the weight from the loose rock. Martin et.al. (1997) have estimated the
weight of the loose rock (Figure 2-8). They also showed that the depth of failure increases
proportionally to the maximum stress near the excavation. The dynamic effect of spalling and
slabbing is not taken into consideration in this design approach, but experiences from South
Africa (Durrheim et.al, 1998) and Norway (Grimstad, 1997) have shown that reinforced
shotcrete together with flexible rock bolts give good results to control spalling and slabbing in
highly stressed rock.
 11

The shotcrete lining follows the contour of the tunnel, which means, the irregularities of the
tunnel shape will influence on the structural behaviour of the shotcrete lining. Chang (1994)
has done studies on this phenomenon, but further studies have to be done to clarify the impact
of irregularities on the structural behaviour of the shotcrete lining.

2 +090
• 1851e7 & Koko, 1979
0 .Aayou en al, 1991
• Martin re al, 1994
Winin. 1989 is the uniaxial compressive strength
CI 0,110PP 8. Gay. 1984
Pelletal. 1091
is the maximal tangential stress at the
S../ I de JOflgh. 1977 boundary of the opening
‘
92
1.4

02 0.4 e6 ee
ammick

Figure 2-8 Depth of failure based on case studies (Martin et.al., 1997)

2.2.2 Shotcrete subjected to vibrations

Shotcrete in tunnelling and mining is often subjected to vibrations from blasting. In tunnelling
the shotcrete is subjected to vibrations from blasting during drifting but in mining the
shotcrete may also be subjected to vibrations from production blasting.

Tannant (1997) performed a full-scale test at the Bousquet mine in Sudbury, Canada, to
determine shotcrete performance near blasts. Reliable measurements of peak particle velocity
(PPV) are difficult to obtain close to a blast-hole. The problems are the response
characteristics of available instrumentation as well as the natural variability in the response of
the rock-support system, thus there is considerable variability in the measured PPV values at
different locations along the test sections.
Several tentative conclusions were made by Tannant (1997):
The area of ejected shotcrete and the volume of ejected rock for the fibre-reinforced
section were about twice that for the mesh-reinforced section.
Mesh or steel fibre reinforced shotcrete can maintain its supporting function for peak
particle velocities of up to 1.5 to 2 m/s at the drift surface when the rockmass itself is not
severely damaged.
When straining creates fractures wider than about 3 to 5 mm, the steel-fibres (30 mm)
either pull out or break. Under these conditions, the more robust wire mesh has a
significant advantage because it continues to reinforce the shotcrete even when the gap
between shotcrete panels exceeds 5 to 10 mm. Mesh also strengthens the linkage between
the shotcrete and the rockbolt and provides toughness to the shotcrete.
 12

Young shotcrete

Anse11 (1999) performed a number of tests in LKAB:s mine in Kiruna with young shotcrete
subjected to vibrations from blasting. He concluded, among other things, that 0-24 hours old
shotcrete was undamaged for vibration levels below 0.5 m/s.

2.3 Mix design

2.3.1 General

Shotcrete mix design follows the same principles that govern normal concrete mix design.
The prime factors controlling strength and durability are the W/C-ratio (water/cement-ratio),
air content and compaction. Figure 2-9 shows the relationship between W/C-ratio and the
compressive strength. A typical mix for the wet-mix method is shown in Table 2-2.

eo

4°
A 30

12°
10
0
0,3 04 0,5 06 07 08 0,9 10 11 12
W/C-ratio

Figure 2-9 Compressive strength of concrete as a function of water/cement ratio,

Betonghandboken, (1997)

Table 2-2 Mix design of shotcrete, wet mix method (Malmgren, 1999)

Cement 500 kg/m3

Silica 25 kg/m3
Aggregate 1520 kg/m3 (dry weight)
Accelerator (waterglass) 4-6 % of the weight of the
cement
Water-(cement+silica) ratio 0.42
Slump 150 mm
 13

2.3.2 Cement, aggregates and additives

Cement

For most applications, cements for shotcrete should fulfil demands for normal standard or
rapid hardening Portland cement (Malmberg, 1993a). The reaction of cement and water is
referred to as hydration. It leads to the formation of the hydrated cement paste.

The most important clinker phases are:

Tricalcium silicate 3Ca0. S i02 (C3S)
Dicalcium silicate 2CaO•Si02 (C2S)
Tricalcium aluminate 3Ca0-A1203 (C3A)
Tetracalcium aluminoferrite 4CaO.A1202•Fe203 (C4AF)
Calcium sulfate dihydrate (Gypsum) CaSO4-2H20 (CSH2)

The formulas given in parenthesis correspond to the shorthand notation of the clinker phases
commonly used in cement chemistry.

The various clinker phases differ in their rate of reaction as well as in their contribution to the
development of strength. C3A and C3S have the highest reaction rates whereas C2S reacts
much more slowly. The early rate of hydration of C3A is slowed down by the presence of
gypsum. Whereas the clinker phase C3S is responsible for the early strength development, C2S
contributes to the strength development at higher ages. The amount of heat liberated during
hydration is the highest for C3A, it is moderate for C3S and C4AF and low for C2S.
Consequently, high early strength cements have high contents of the clinker phases C3S and
C3A, whereas low heat cements have low contents of C3A and C3S but a high content of C2S.

Table 2-3 The content of clinker phases etc. in common Portland cement
The content of clinker phases etc. in common Portland cement
Material According to the According to Grasser et.al.
Betonghandboken (1997) (1994)
C3S 50-70% 25-60%
C2S 15-20% 15-50%
C3A 5-10% 3-12%
C4AF 5-15% 8-12%
Gypsum 5% -
CaO (Calcium oxide) 1-2% -
 14

The hydration of the calcium silicates results in the following:

2C3S+6H20—>C3S2-3H+3Ca(OH)2

and

2C2S+4H20—>C3S2•3H+Ca(OH)2

where

C3S2.311 is a shorthand formula of 3CaO2Si02.3 H20 (Calcium silicate hydrate or cement gel)

and

Ca(OH)2 is Calcium hydroxide

The reactions between C3A and sulfates and the associated volume increase are of particular
significance for the sulfate resistance of concrete when exposed to an external source of
sulfates, therefore the amount of C3A shall be limited in such cases.

Aggregates

The maximum aggregate size for wet-mix shotcrete is normally 8 mm and for dry-mix
shotcrete 10-16 mm (Malmberg, 1993a). The natural moisture content in the aggregate must
be as constant as possible and should not exceed 6% for dry-mix shotcrete (Malmberg,
1993a). Natural gravels are preferred to crushed stone because of the better pumping
characteristics of the rounded natural aggregate particles (Hoek & Brown, 1980).

Silica

Condensed silica fume and ferrosilicon dust may be added to increase the cohesion and
stability of the shotcrete for example to decrease the rebound on the tunnel roof. When silica
fume is added, a superplasticizer must be added to achieve the required workability. The
normal dosage of silica is 3-15 % of the weight of the cement (Malmberg, 1993a). The
addition of Silica results in more cement gel and less Calcium hydroxide.

Accelerators

Accelerator admixtures are used in shotcrete to (Vanderwalle, 1994):

help the shotcrete to gain more quickly so that it is more competent to support the tunnel
at an early stage;
reduce the level of rebound of aggregate from shotcrete;
- enable thicker layers of shotcrete to be sprayed in one pass;
- enable shotcrete to be applied to a wet surface subjected to slight infiltrations;
- enable wet-mix shotcrete to be used for overhead works.

The main types of accelerators in use are:

- alkali silicates as Na2Si02 and alkali aluminates
- alkali free accelerator, pH 2-6
 15

Alkali silicates and alkali aluminates are still the most common types of accelerators used for
shotcrete. Sodium silicate also known as waterglass, is the classic accelerator used for
shotcrete. Alkali silicates and alkali aluminates will decrease the final strength of shotcrete.
Figure 2-10 shows how the compressive strength decreases with increasing amount of
waterglass. Furthermore, Lukas & Kusterle (1990) have shown that higher amount of
waterglass also decreases the adhesion strength of shotcrete to the substrate, see Figure 2-11.
The normal dosage of most accelerators is 2-8 To of the weight of the cement.

During the last years the use of alkali free accelerators have been increased because of better
working safety, by reduced aggressiveness of the accelerator and faster early strength
development, and less or no reduction of the final strength of shotcrete as shown in Figure
2-12.

50 100%
100%
3c= 7r=1 28 ow rage _ 90
I 40
70% 72% 67% _ 2
›;
-x
- 50
1;1 20 se - ig

10 - 20
- 10
0
0% 5% 10% 15%
Dosierung

Figure 2-10 Compressive strength as a function of the amount of waterglass (Stiglat et.al,
1985)

SUBSTRATE: GNEISS SUBSTRATE: LIME-STONE

3 3
LIQUID ALUMINATE LI QUID ALUMI NATE
›-
j ACCELERATOR ACCELERATOR
2 2
cn ••• "'"••••...
1 WATER GLASS WATER GLASS
• /
•— • —
0 5 10 15 0 5 10 15
ACCELERATOR DOSAGE 1% 1 ACCELERATOR DOSAGE I%

SUBSTRATE: CONCRETE
3
LIQUID ALUMINATE
/ACCELERATOR
2 —
•
•
—
WATER GLASS

0 5 10 15
ACCELERATOR DOSAGE I %

Figure 2-11 Adhesion strength of shotcrete as a function of the amount of accelerator (Lukas
& Kusterle,1990)
 16

Sodium silicaMMIkalifroe

100

0,1
0,1 1 10 100 1000
Hours after spraying

I .0—Alkalitres —in— Saditrn sdele I

Figure 2-12 Early and final strength development alkalifree accelerator vs Sodium silicate
(waterglass), after Erlien (1999)

Plasticisers or water reducing agents and superplasticisers

The name comes from the ability of the product to make concrete wetter without increasing
water or to reduce water in the concrete without loosing the original workability.
As a very rough comparison a super-plasticiser can reduce water in a given mix up to three
times the amount that can be removed using standard water reducing agents (Reynolds, 1998).
The workability time for the shotcrete with superplasticisers is between 0.25 —1.75 hours,
depending on the type of superplasticiser and concrete temperature (Betonghandboken, 1997)

Air entraining admixtures

Air entraining admixtures are used in concrete exposed to freeze/thaw. This technique is not
possible to use for dry mix shotcrete, but the freeze/thaw resistance is normally fulfilled by
the low porosity governed by the low W/C-ratio (Malmberg, 2001). For the wet-mix shotcrete
many attempts have been made to add air-entrancing admixtures but the problem is that the
high compressive force during shotcreting will destroy the air pores. Some good results have
been reported when an overdose of admixtures was used which ended in a sufficient air
content after shotcreting. Freeze/thaw resistance has also been enhanced by use of low W/C-
ratio, but finding a proper method to ensure necessary freeze/thaw resistance of the wet-mix
shotcrete is still to be solved (Malmberg, 2001).

2.3.3 Reinforcement

The main function for reinforcement is to increase the tensile strength and the ductility of
shotcrete. Another important function is to control the size and position of cracks.

Chain-link/woven mesh/diamond mesh or welded mesh are used to reinforce shotcrete.

Welded mesh is sometimes preferred to chain-link mesh because of the difficulty of applying
shotcrete satisfactorily through the smaller openings in the chain-link mesh (Hoek et.al.
1995). But on the other hand welded mesh is stiffer than chain-link/woven mesh, which
owing to its stiffness cannot accommodate large deflections beyond peak load to an extent
comparable to chain-link mesh (Kirsten, 1993).
 17

In the early 70's the first experimental work was undertaken with steel fibre reinforced
shotcrete (Vandewalle, 1994). The function of fibre reinforced shotcrete is the same as for
shotcrete reinforced with mesh and bars. Shotcrete may be reinforced with steel, glass,
polymer or carbon fibres, but the most used fibres are steel fibres.

Important factors for the strength are (Ay, 1999):

bond strength between fibres and matrix
fibre length
fibre aspect ratio
tensile strength of fibres
amount of fibres
numbers of fibres
shape of fibres
dispersion and orientation of fibres

To properly estimate the mechanical properties of fibre reinforced shotcrete, some brief
theoretical considerations given in the following, are useful. The mechanical behaviour of
fibre reinforced composites, particularly for the case of a ductile fibre in a brittle matrix, may
be described with the classical theory of composite materials. The stresses acting in the fibres
and in the matrix respectively, may be estimated from the volume concentration of the fibres
vf and the stiffness ratio of both components n = ElE,,, where Ef and Em are the Young's
moduli of the fibres and of the matrix respectively.

The tensile strength of the composite material depends on the fibre content, on the fibre
orientation and on the bond strength between fibre and matrix. A substantial increase of the
tensile strength occurs only, if the fibre content is larger than a critical value vf„" principally
shown in Figure 2-13.

v, -0

Elongation Al Elongation Al
Concentric Tension Concentric Compression

Figure 2-13 Load deformation relations for fibre reinforced shotcrete, (Grasser et.a1.1994)
 18

In the simplest case, the critical fibre content can be expressed as

V I, frit =
(2.1)
11071 fif

where

= tensile strength of the matrix

ftf = tensile strength of the fibre
17b = takes into account the bond properties between fibres and the matrix. If nb < 1,0 the
bond properties are not sufficient to transmit a tensile stress from the matrix into the
fibres which is equal to the tensile strength of the fibres. The bond properties are not
only influenced by the bond strength but also by the fibre length as expressed by (2.2).
710 -= the coefficient is < 1. It takes into account, that not all fibres are oriented in the direction
of the applied stress

Various values are given in the literature for the fibre orientation coefficient no used in (2.1).
Grasser et.al.(1994) gives values from theoretical considerations, no = 0.45 —0.7 for two-
dimensional fibre orientation in concrete and no = 0.33 — 0.66 for random, three- dimensional
orientation in concrete. When applying shotcrete the fibres have almost a two-dimensional
orientation, according to Betongrapport nr 4 (1995).

The tensile strength of the steel fibre F,= f /4 shall be equal to or greater than the bond
strength of the steel fibre Fb= r„,d1 to get a ductile failure. rn, is the mean bond strength at
failure and l is the anchor length of the steel fibre. The reason is that the fibre should be pulled
out from the matrix instead of break.

According to the expressions for Fb and F, the fibre length must be less than 2/, to get a
ductile failure

d (2.2)

2.3.4 Experience with steel fibre reinforced shotcrete and mesh reinforced
shotcrete at LKAB:s underground mine in Malmberget

At the Malmberget mine steel-fibre reinforced shotcrete has been tested since 1980 but
widespread implementation of steel fibre reinforced shotcrete was delayed until 1994 because
of lack of proper equipment. Savilahti (1995) has summarised the experiences with steel-fibre
reinforced shotcrete in Malmberget's mine. Table 2-4 shows that steel-fibre reinforced
shotcrete is very competitive when considering costs and capacity compared to shotcrete
reinforced with steel mesh.

Failure mapping of crosscuts in poor rock conditions showed that both reinforcement
techniques have been working very well. However, fixing the mesh to the back is difficult and
sometimes hazardous because the miners are exposed a longer time in the unsafe area.
 19

Table 2-4 Comparison of costs and capacity for steel fibre reinforced shotcrete and
shotcrete reinforced with steel mesh (Savilahti, 1995)
Rock support Capacity (1994) Costs (1994)
Shotcrete reinforced with 9 m drift/shift 3550 SEK/m drift
steel fibres
Shotcrete reinforced with 3 m drift/shift 4900 SEK/m drift
steel mesh

2.4 Shotcrete technique

2.4.1 Cleaning of rock surfaces

Because the adhesion strength is one of the most important parameters for shotcrete as rock
support, the rock surface must be clean when applying the shotcrete. Therefore, in addition to
scaling, all surfaces, which require to be bonded, should be cleaned by water or an air/water
combination.

According to NCA No 7 (1999), a minimum of 7 bars air pressure and a minimum water flow
of 0.3 m3/min should be used.

2.4.2 Spraying

The critical parameters to control when spraying on a surface are according to Spearings
(1998):
- spray angle (the angle between the nozzle and the surface being sprayed)
- spray distance (the distance between the nozzle and the surface being sprayed)

The spray angle is critical as it influences the rebound and the compaction of the sprayed
material. The optimum angle is perpendicular (90°) to the surface and the angle should never
be less than 45°.

The optimum distance between the nozzle and the rock face being sprayed, depends on
(Spearing, 1998):
the maximum aggregate size
- the grading curve
- the capacity of the shotcrete machine
- whether the nozzle is hand held or applied by a remote controlled machine. As a general
rule, a distance of about 0.5 to 1.6 m gives the best results.

In addition, the nozzle motion is important to reduce rebound, increase compaction and
eliminate sand pockets. To achieve this the nozzle must be rotated in slow circular movements
across the surface.
 20

2.4.3 Curing

Shotcrete is just pneumatically applied concrete and hence needs the same curing
considerations as concrete. This aspect is frequently overlooked in shotcrete operations.
Premature drying out of the shotcrete will reduce its strength and durability. Depending on the
site conditions, curing can be obtained by:
- repeated moist spraying
- curtain-type moistened overlays or by curing membranes
- using an internal curing admixture

In mining applications, although the humidity is generally high and would tend to keep the
shotcrete surface moistened, the air velocity is generally high and tends to counteract this.
Internal curing is therefore usually the most efficient method according to Spearings, (1998).

2.4.4 Training

All experience points out how important the training of personnel associated with the
placement of shotcrete is. Especially the nozzleman plays a vital role in the application.

2.5 Mechanical properties

As mentioned earlier the prime factors controlling strength and durability are the W/C-ratio
(water/cement-ratio), air content and compaction. The most important mechanical properties
for shotcrete are strength in the rock—shotcrete interface, compressive strength, tensile
strength, as well as Young's modulus of the shotcrete.

2.5.1 Strength between rock and shotcrete

The adhesion strength between shotcrete and the rock mass is one of the most important
parameters for the function, especially for plain shotcrete or shotcrete without rock bolts.

To describe the strength of the interface between the shotcrete and the rock mass a failure
criterion can be used. The criterion can be expressed as

✓ z, No failure
(2.3)
z> I Shear failure

and

o- fr, No failure
(2.4)
• > f, Tensile failure
 21

where rf is the shear stress at failure according to Mohr-Columb

rf = c+0;" tan 0 (2.5)

where c is the cohesion between shotcrete and rock mass, cr'„ is the effective normal stress in
the layer, 0 is the friction angle between shotcrete and rock mass andfa is the adhesion
strength between shotcrete and rock.

The most important parameters for adhesion strength are:

Cleaning of substrata
Geology
Shotcrete (from mixing to spraying the shotcrete)

Hahn (1983) tested the adhesion strength for shotcrete sprayed on different rocks (intact
rock), Figure 2-14 summaries the results. The adhesion strength between the shotcrete and the
rock mass can be determined using the test equipment shown in Figure 2-15.

Gabbro, fine to middle grained

Granite, fine to middle grained 11/212a1

n8t2M2arig

Granite, middle grained

Granite, coarse porphyroblastic

Sandstone, porous
13 Smooth surface M Rough surface

,ttittiett
Sandstone, crystalline

Marble, middle grained

Reef lime stone, fine grained

Lime stone-marlestone

eteiwhiatte
Gneiss, parallell cut

Gneiss, perpendicular cut

ittetUtiaffletettikif
Mica-schist, middle grained

Clay shale, very fine grained F92

0.0 0.5 1.0 1.5 2.0

f, (MPa)
Figure 2-14 Adhesion strength (Hahn, 1983)
 22

a) b) c)

Figure 2-15 Adhesion strength test equipment (from Swedish Standard, SS 13 72 43, 1987)
a) Drilling two circular slots with double diamond drill
b) The inner circular slot is drilled through the shotcrete
c) The test machine, attached to the shotcrete drill core for the pull out test

Table 2-5 shows the adhesion strength determined in two civil engineering projects in Sweden
and for LKAB:s underground mine in Malmberget, Sweden. As one can see there is a great
difference between the adhesion strength determined in the civil engineering projects and in
the mine. The adhesion strength is calculated according to the Swedish standard SS 13 72 43
(1987), see Figure 2-15.

Table 2-5 Adhesion strength in tunnels and in LKAB:s underground mine

Location Adhesion Adhesion No of References
strength strength tests
Mean value Std deviation
(MPa) (MPa)
Södra länken, Stockholm, 1.37 0.71 78 Ellison (2000)
Sweden (Wet-mix method)
Grödingebanan, Sweden 0.86 - 78 Malmberg
(Wet-mix method) (1993b)
Grödingebanan, Sweden 0.97 - 48 Malmberg
(Dry-mix method) (1993b)
LKAB:s underground mine in 0.40 0.38 23 LKAB (1999)
Malmberget, Sweden (Wet-mix
method)

In adhesion strength tests, failures do not only occur between shotcrete and rock but also in
the rock or in the shotcrete or combinations of contact surface/shotcrete/rock. Karlsson (1980)
carried out more than 230 adhesion strength tests to find out where the failure surfaces were
located, the results are presented in Figure 2-16.
 23

Localisation of the failure surface

Other
Rock, shotcrete and contact area 2
Rock and shotcrete

Contact area and rock

Contact area and shotcrete TJ

Rock -
Shotcrete '
Contact area

0 5 10 15 20 25 30 35
Part of the failures (%)

Figure 2-16 Location of failure surface in adhesion strength tests between shotcrete and
rock, after Karlsson (1980)

The author has not found results from the literature on the shear strength between shotcrete
and rock. However, Silfwerbrandt (1987, 1988) has measured both the adhesion strength and
shear strength, fib between new and old concrete and between shotcrete and old concrete. The
shear strength was determined by rotating a core of the new concrete relative to the old. Since
the normal stress was equal to zero when the tests were performed the cohesion is equal to the
shear strength. The results from the tests are shown in Table 2-6. Results from field tests are
also shown in the same table (Magnusson, 1987).

Table 2-6 Relationship between shear strength or cohesion (fsh) and adhesion strength (f,)
between new concrete/shotcrete and concrete
Description Adhesion Shear Ratio fs ja References
strength, fa strength, f'sh
(mean values) (mean values)
MPa (MPa)
The strength between Silfwerbrand
new and old concrete 1.70 3.67 2.16 (1987)
(Laboratory tests)
The strength between
new shotcrete and old 1.74 3.22 1.85 Silfwerbrand
concrete. (1988)
(Laboratory tests)
The strength between
new and old concrete 1.46 3.50 2.40 Magnusson (1987)
(Field tests)

Because the shear strength,fsh, between rock and shotcrete is unknown for the author as
mentioned above, the ratio fsbf, between rock and shotcrete is assumed to be the same as
those presented in Table 2-6. In the further analysis the ratio 2 will be used. This means that
 24

the cohesion c in (2.5) can be estimated as two times the actual adhesion strength fa. Finally,
the friction angle q5in (2.5) is assumed to be 30 degrees, which is a common value between
two layers of concrete (Betonghandboken (1997).

2.5.2 Beam test

Bending flexural strength and flexural toughness of fibre reinforced concrete/shotcrete is

normally determined using beam tests. Five of the most used beam test methods are briefly
discussed in the following, the methods are:

- ASTM C1018 (1994) (USA)

- JSCE-SF4 (Japan)
- NCA No.7 (1999) (Norway)
- Betongrapport nr 4 (1995) (Sweden)
- CEN/TC 104/WG 10 N 62 (1999) proposal for European standard

The descriptions of the two first methods are based on Groth's (2000) review of beam test
methods. The various standards use different principles to determine the flexural toughness.
The methods are based on recorded load vs. deflection in the-mid span for a simply supported
beam loaded with two point loads, see Figure 2-17.

F/2 F/2

\fr \fr

150 150 150

/ / / /
450

Figure 2-17 A four point test beam

ASTM C1018

This standard, (which is one of the most used standards) evaluates the flexural performance of
toughness by defining flexural toughness values. These relative flexural values are called
toughness indices, .4=5,10,203o ... Furthermore, residual strength factors R510, R10,20 • • •, first
crack strengthffl„ and ultimate flexural strength ffh, are calculated. All these quantities are
defined in Figure 2-18.
 25

etc

g. - too

M idspan-deflection
D
538

Figure 2-18 Flexural toughness values according to ASTM C1018 standard

For linear elastic perfectly plastic behaviour the toughness indices I will be the same as the
subscript of the indices, L=5 = 5 and the residual strength factors R will always be equal to
100.

If the deformation is measured without excluding the extraneous deformations from support
settings the evaluated indices will be affected (Groth, 2000). This is because all indices are
dependent on the deflection for the first crack According to Groth (2000) 8cr can differ up
to an order of magnitude if the deformations are not correctly determined.

The toughness parameter is not only strongly dependent on the location of the first crack but
also how this location is determined. ASTM C1018 defines this location in a way which can
cause subjectivity (Groth, 2000).

Betongrapport nr 4 (1995), Sweden

This guideline propose an alternative way to determine the first crack in a more objective way
compared with ASTM. Normally, the beam size is (75.125450 mm, depth • width • span).
Betongrapport nr 4 (1995) use the same quantities as ASTM. The first crack load F„ and
deflection 8 is calculated as follows, (Groth, 2000):

a) Locate the first crack point on the curve. This point is defined as "the point at which the
curvature first increases sharply and the slope of the curve exhibits a definite change".This
first guess is considered as an approximate value and the corresponding load is denoted
see Figure 2-19.
b) Determine the slope a of the curve, between the points 0.25F: and 0.75F:, . This line is
denoted L1 in Figure 2-19.
c) Draw a line parallel to L I, but at a distance of W2000 to the right in the diagram (h = beam
height), this line is denoted L2.
d) The point P is the point where L2 intersects the load-deflection curve.
e) Determine the new value, F„ at point Q, F„ shall be the maximum value of the load along
the load-deflection curve until the point P.
 26

f) Check that F„ does not deviate more than 10 % from F. If it does, repeat step a) to f)
by putting F:, = F„
g) The deflection corresponding to the first crack, point Q, is denoted 80.
h) Draw a line, L3, parallel to L1 that go through point Q.
i) Determine the point R as the point where L3 intersects the abscissa. This deflection is
denoted JR.
j) Calculate the first crack deflection as 8„ = 80 - 8R. Move the curve in the diagram so that
8= 0 becomes point R.

Fr
Fcr

3
4 'cr

Figure 2-19 Determination of first crack load F„ according to step a) to j) above,

(Betongrapport Nr 4, 1995)

JSCE-SF4

In the Japanese standard JSCE-SF4 the flexural toughness is an absolute value, as defined in
Figure 2-20, in contrast to the ASTM standard, which use relative toughness parameters. The
toughness factor fjscE can be considered as a mean flexural strength for deflections up to
L/150. The beam size is the same as for ASTM.
 27

Figure 2-20 Flexural toughness values according to the Japanese standard JSCE-SF4.

This method is compared to the ASTM test, independent of the calculation of the first crack
(Groth, 2000). The disadvantage with an absolute toughness value is that the value is size
dependent.

Nazi No.7 (1999)

NCA No.7 (1999) defines the toughness as residual flexural strength as shown in Table 2-7.
The beam size is (75.125.450 mm). This method is (as JSCE-SF4) independent of the
calculation of the first crack, a disadvantage with absolute toughness values is that they are
size dependent.

Table 2-7 Toughness classes, requirements for residual strength, NCA No.7 (1999)
Toughness class Residual flexural strength at deflection
0.5 mm 1 mm 2 mm 4 mm
0 Shotcrete without reinforcement
1 2.5 MPa 2.3 MPa 2.0 MPa 1.5 MPa
2 4.5 MPa 4.3 MPa 4.0 MPa 3.5 MPa

CEN/TC104/WG 10 N62 (1999)

This standard specifies as the other described standards a method for determination of the
flexural and residual strength of beam specimens. To determine the flexural toughness this
standard uses an absolute value as NCA No.7. The residual flexural strength shall be
calculated from the minimum loads on the flexural stress/deflection curve between 0.5, 1, 2
and 4 mm.
 28

Deformation class R1 (Low deformation):

The residual strength (f,j) shall be calculated from the minimum load (Pr]) recorded between
mid-span deflection of 0.5 to 1.0 mm.

Deformation class R2 (Normal deformation):

The residual strength (f.2) shall be calculated from the minimum load (1',2) recorded between
mid-span deflection of 0.5 to 2.0 mm.

Deformation class R4 (High deformation):

The residual strength (f4) shall be calculated from the minimum load (/),.4) recorded between
mid-span deflection of 0.5 to 4.0 mm.

Typical values for bending capacity for test beams

In mining applications the ductility of the reinforced shotcrete is important, see for example
the discussion in section 2.2. Nordström (1996) collected test results from 273 (91*3) tested
shotcrete beams. The method in these tests was an earlier version of the Swedish Guidelines
for shotcrete which was based on ASTM C1018 but with different beam size (75*125*450
mm) and different rate of loading. The compressive strength of the shotcrete was 46-50 MPa
and 56-60 MPa for grade K40 and K45 respectively. He found that the first crack strengthfficr
approximately could be described as

ffic,..0.355f ° 72 (MPa). (2.6)

The number of test beams for reinforcement is shown in Table 2-8, Figure 2-21 and
Figure 2-22 show the first crack strength and the residual strength for shotcrete beams
reinforced with Dramix 30/0.5 and Dramix 40/0.5 respectively. The residual strength is
presented as /5,io = R5,løfflcr andfio,30 = Rio,3offl, . The test results for beams reinforced with 60
and 65 kg/m1Dramix 40/0.5 are based on relatively few beams.

Table 2-8 Number of test beams for the test results (Nordström, 1996)
Amount of steel fibre (kg/m3) Number of beams reinforced Number of beams reinforced
with Dramix 30/0.5 with Dramix 40/0.5
45 - 12
50 6 33
55 6 -
60 15 3
65 24 3
70 93 -
 29

Bending strength (MPa)

8
4)-
6
--0— First crack
4 —a--f 5,10

2 —e— f 10,30

40 45 50 55 60 65 70
Amount of steel fibre (kg/in')

Figure 2-21 First crack strength and residual strength for shotcrete beams reinforced with
Dramix 30/0.5 (Nordström, 1996)

—4-- First crack

.
Ce 6
- f 5,10
r„. 4
- f 10,30
2 2-

40 45 50 55 60 65 70
Amount of steel fibre (kg/m3)

Figure 2-22 First crack strength and residual strength for shotcrete beams reinforced with
Dramix 40/0.5 (Nordström, 1996)

Comparison of normal concrete (NPC) and high performance concrete (HPC) and various
length of steel fibres

Will the bearing capacity and the toughness of shotcrete increase with higher strength?
Groth (2000) did an interesting investigation where the behaviour of HPC was compared with
the behaviour of NPC, both reinforced by steel fibres. Two concrete qualities with the
compressive strength fc = 122 MPa (HPC) and f. =25 MPa (NPC) were tested. The steel fibres
were delivered from Hörle, Sweden and the amount of fibres varied from 30 to 50 kg/m3. The
steel fibres were either of 35 or 60 nun length. The load — deflection diagrams are shown in
Figure 2-23 and Figure 2-24. The beam tests were performed according to Betongrapport nr 4
(1995).

NPC showed better residual flexural strength for longer fibres (60 mm) compared to short
fibres (35 mm). The reason is according to Groth (2000) that the critical length le is close to
30 mm (=1/2•60 mm), see Equation (2.2), for the NPC (few fibre breakage's). Since there was
a low risk of fibre breaking for fibre length shorter than 60 mm, the longer fibres are much
more effective with respect to toughness compared with the shorter fibres.
 30

If no fibre breaking had occurred for the HPC, the higher matrix strength had given a more
ductile response, but since the stronger matrix had higher bond strength capacity than the
tension capacity for the fibres, fibre breaking occurred for the HPC. Therefore, the strong
matrix in the HPC gives higher peak strength but lower toughness for larger deflections
compared with the NPC. As mentioned earlier, ductility is an important property for rock
support in mining and because of that shotcrete with normal strength better fulfils the
requirements.

12
BOO: Plain concrete

B07: 0.38% 35mm

10
BOB: 0.64% 35mm

- B09: 0.38% 60mm

- BIO: 0.64% 60mm

1 2 3 4
Mid-span deflection, (mm)

Figure 2-23 Load — deflection curve for normal concrete (NPC),f, =25 MPa, after Groth
(2000)

BOO: Plain concrete

B2: 0.38% 35mm

B3: 0.64% 35mm

804: 0.38% 60mm

B05: 0.64% 60mm

1 2 3 4
Mid-span deflection, (mm)

Figure 2-24 Load — deflection curve for normal concrete (HPC),f, =122 MPa, after Groth
(2000)
 31

Young's modulus

Young's modulus, E„ can be calculated from beam tests, assuming linear elastic behaviour.
The slope for the load-deflection curve of an uncracked beam is used. Young's modulus for a
four point beam with dimensions according to Figure 2-17 is

23 F„.1?
= (2.7)
1296 801

where F. and 4) are the crack load and corresponding deflection calculated according to
Betongrapport nr 4 (1995), /is the area moment of inertia and Lis the span of the beam. In
Table 2-9 Young's modulus, Ec, is presented from various tests.

Table 2-9 Young's modulus, E,

f E, E, according E, Test method Source

to BBK 94 p for E,
-BBK
BBK 94
(MPa) (GPa) (GPa)

Shotcrete 34.0 13.2 31.0 0.43 - Myrvang & Davik,

38. 7 11.5 32.2 0.36 (1997)
Shotcrete 52.0 16.1 " 35.3 0.46 Beam test Malmberg (1993b)
Shotcrete 49.5 20.5 ' 34.9 0.59 Beam test Nordström (1996)
Shotcrete 53.6 22.9' 35.5 0.64 Beam test Nordström (1996)
Concrete 25.0 32.0 28.5 1.12 Beam test Groth (2000)
122.0 49.7 42.5 b) 1.17 1')
Concrete 49.5 24.3 34.9 0.70 Beam test Nilsson (2000)
55.1 24.0 35.7 0.67
a) Calculated by the author according to Equation (2.7)
b) High Performance Concrete Structures, Design Handbook (2000)

2.6 Design criterion for shotcrete

The support design can be based on common practice, empirical systems with and without
measurements, rigorous calculations and comparative experimental simulations. Empirical
systems are briefly discussed in this section. Experimental simulations are discussed in
section 2.7.

2.6.1 The New Austrian Tunnelling Method (NATM)

The basic principles for the NATM were developed in the 1940's by professor Rabcewicz. In
1963 "The New Austrian Tunnelling Method" was introduced at the Geomechanics
Colloquy in Salzburg. This is an empirical system based on the concept of ground response.
 32

The main principles are according to Golser (1995):

With a flexible support system a new equilibrium shall be reached. The support system
shall be flexible in order to minimise bending moments and to facilitate the stress
rearrangement process without exposing the lining to unfavourable sectional forces.
Additional support requirements shall not be added by increasing lining thickness but
bolting. The lining shall be in full contact with the exposed rock. Shotcrete fulfils this
requirement.
The Equilibrium State shall be "controlled" or be under observation by in-situ
deformation measurements. With the monitoring and interpretation of deformations, strain
and stresses it is possible to optimise working procedures and support requirements.
- After the Equilibrium State is obtained the permanent support system (an inner arch) shall
be built. In specific cases this inner arch can be omitted.

One of the most important objectives with the NATM concept is to maintain triaxial loading
conditions for all construction phases and to minimise situations with uniaxial or biaxial stress
states in the rock mass.

To reach this goal it is important to avoid detrimental loosening by careful excavations and by
immediate application of support and strengthening means. Furthermore, rounded tunnel
shapes should be used to avoid stress concentration in corners where progressive failure starts.

The main advantages and disadvantages of the NATM are according to Jodl (1995):

Advantages:

- Applicable in a wide range of ground conditions

Simple and flexible adaptation to different cross sections
- High economy by optimising necessary support measures
- Economic application for short contract sections
- Easy combination with TBM
Relatively small investment with quick amortization

Disadvantages:

application in ground water only with additional measurements

- rate of advance is relatively small and cannot be increased decisively
- very high requirements with regard to education, training and practice of personnel
high requirements with regard to the quality of execution and material
- difficult formulation and distribution of risks for client and tender
limited possibility of automatization
support for tunnels in hard rock that are designed with the NATM-concept tend to be
expensive (Kirsten,1998)

2.6.2 Rock mass classification

The main purpose of rock mass classification systems is to help the engineers to judge the
stability of a rock mass and to serve as an aid while setting contractual issues regarding rock
support (Brantmark, 1997). A great variety of rock mass classification systems have been
derived. The RMR-system, developed by Bieniawski (1976) together with the Q-system,
developed by Barton et al. (1974), are the most used systems in Sweden and the world. It is
 33

important to remember that the classification systems give reliable results only for the similar
rock masses and circumstances for which the systems were developed. Both of these systems
are widely used in civil engineering practice but not so much in mining practice.
From the RMR and the Q-system, modified classification systems have been developed by for
example Laubscher & Taylor (1976) Laubscher (1984), Stille et.al. (1982) and Hoek et al.
(1995, 1997).

The classification systems are described and advantages and disadvantages are discussed by
many engineers /scientists such as Hoek & Brown (1980), Hoek et al. (1995), Lundman
(1998), Brady and Brown (1999), Nordlund et al. (1997).

2.7 Experimental studies, panel tests

During the last decades a large number of tests have been carried out to improve the
understanding of shotcrete as rock support. To find a test method to model the interaction
between shotcrete and rock is very difficult. This is mainly because of the heterogeneous
properties of the rock mass but also difficulties to include the effect of rock stresses. This
means that experimental studies can only be done for relatively simple ground conditions.

The test methods can mainly be divided into two different groups:
testing the function of the shotcrete as rock support
compare different types or amounts of reinforcement or different types of shotcrete mixes
with a specified testing method, for example the beam tests discussed in Section 2.5.2

These two groups will not be treated separately in this review.

2.7.1 Shotcrete as rock support for falling block

If the rock stresses are low compared to the strength of the rock, the main function for the
rock support, is to prevent fallouts of loose rock material. A typical load situation is shown in
Figure 2-25 a).

a) b)
_$HOTCRETE
./
OUTER MIDDLE OUTER
ROCK SLAB ROCK SLAB ROCK SLAB

APPLIED
LOAD P/2
IP

CRACKS BEETWEEN
ROCK SLABS

Width= 1200 mm
Thickness, shotcrete = 80 mm

Figure 2-25 Rock support for a key block

a) Load situation (Holmgren, 1985) b) Test rig (after Holmgren,1979)
 34

Tests, which have been designed to simulate this kind of loading condition, have been
performed by for example Holmgren (1979, 1985), Torsteinsen & Kompen (1986) and
Mörsch (1993). Figure 2-25 b) shows a typical test rig. The response depends on the adhesion
strength between the rock and the shotcrete, shear, bending moment and membrane capacity
of the shotcrete. The adhesive failure shows a brittle behaviour as shown in Figure 2-26 and
the ductility of the shotcrete depends on the type and amount of reinforcement.

P
PRIMARY FAILURE WAD

MIDDLE
SLAB 'i rSOUTE
LABR

ADHESION CRACK
PROPAGATION LOAD

1-117-1
tr, P/2 ?re P/2 lp Pa
UNCRACKED "ADHESION CRACKING" "FLEXURAL
STATE CRACKING"

Figure 2-26 Typical response for shotcrete (Holmgren, 1979)

Holmgren (1985) studied the influence of various types of washers on the performance of
steel-fibre reinforced shotcrete (SFRS). The SFRS had a high amount of steel-fibres (1.5% by
volume). The tested washers are shown in Figure 2-27. Figure 2-28 shows that there is no
noticeable difference between stiff and flexible washers. In fact the flexible washers worked
better together with the SFRS than the stiff washers. According to Holmgren (1985) SFRS is
not as sensitive to stress concentration as shotcrete reinforced with welded steel mesh. The
interaction between shotcrete and rock bolts without washers was not very good, one reason
was the difficulty to spray around the rock bolts.
STEEL FIBRE
REINFORCEMENT:
Rock bolt o20 Rock bolt o 20

80 .
.,,,
e Outer . Middle rock 4
300 ,TOCk slab , F2
,e_ rev err rr PI
1 550 I, 4 50 1, 800
k

Plate 0160 t.5

Sohericol slate 160 t=12ltest 5:e 601 1.

T
,
e 1 F3 - F9
er ,.• I. r 'r Y Y Y Yi

Spherical plate 0160, t•12 E511 REINFORCEMENT:

F13 -14, F22

,

oh rein oratmeM F13-14: Swedish Nps 50, d 5,o 100.

FU: Mild steel o 6, c 150 deformed bars

Figure 2-27 Different types of washers tested (after Holmgren, 1985)

 35

Type of washers and steelfibres

(The shotcrete was reinforced with 118 kg/m3 steelfibre)
P/2 (kN)
77160 mm stiff or flexible washers, Dramix ZL35/0.35
60 mm stiff washers, Dramix ZL35/0.35

_____Bended rock bolt without washer, Dramix ZL45/0.35

50 /fly
---- Straight rock bolt without washer, Dramix ZL45/0.35

******
4 tea w
_ ....
\. ...

0 50 100 (mm)

Figure 2-28 Test results with different type of washers (after Holmgren, 1985)

Holmgren (1985) also tested SFRS with a steel fibre content of 0.5-0.6% by volume and
compared the capacity with shotcrete reinforced with cold-formed or stress-relieved anneal
welded mesh. The SFRS panels were fixed with rock bolts with flexible washers and the
panels reinforced with welded mesh were fixed with bolts with stiff washers. Figure 2-29
shows that the SFRS panels had higher load capacity than the shotcrete panels reinforced with
welded mesh. The ductility of SFRS panels was better than for shotcrete reinforced with cold-
formed welded steel mesh but slightly lower compared with shotcrete reinforced with stress-
relieved anneal steel mesh.
P/2 (kN) Reinforcement:

0.5-0.6% Dramix ZL 45x0.35 (39-47 kg/m3)

Welded steel mesh 05c100 f=510-560 N/mm2

Welded mild steel 06c150 fi =220 N/mm2

50
(stress-relieved anneal steel mesh)

.••••••••

0 50 100 (mm)

Figure 2-29 Test results with different reinforcement (after Holmgren, 1985)
 36

2.7.2 Shotcrete as rock support in weak or very jointed rock

In this case the rock applies more or less a uniform load on the shotcrete lining, i.e. the
shotcrete lining acts as a retaining element for the loosened rock, see Figure 2-30. To
determine the capacity of shotcrete in this loading situation, many types of tests have been
performed. Some of the more well-known test methods are shown in Figure 2-31 and Figure
2-32 and shortly described in Table 2-10.

Rock
stresses
Rock loads

Hold

Reinforce
(strengthen rock
and control bulking)

Figure 2-30 Support functions (Esplay et.a1.1995)

a)

b)

a) Simply supported circular panel with a point load in the centre of the panel
or clamped circular panel loaded with a point load in the centre of the panel,
Torsteinsen et.al. (1986), Mörsch (1993) and Nilsson (2000)
b) Simply supported square panel loaded with a point load in the centre of the
panel EFNARC , Bernard (1997) or clamped square panel loaded with a
point load in the centre of the panel, Bernard (1997)

Figure 2-31 Testing methods

 37

a)

\ EDGE OE TEST PANEL

POIEL

A-A

b) c)

a) South African water bed test. A square panel supported by four bolts and
loaded with a uniform load, Kirsten (1993)
b) Panel supported on a continuous media and loaded by a point load, (Becket
& Humphreys, 1989)
c) A square panel supported by four bolts or four columns, one in each corner
and loaded by a point load in the centre of the panel, Morgan and Mowat
(1984), Tannant and Kaiser (1997)

Figure 2-32 Testing methods

 38

Table 2-10 Cormnents on some of the testing methods presented in Figure 2-31 and
Figure 2-32
Type of test Function of the test method Authors comments
method

Figure 2-31 Shows the stress situation around Advantage:

Simply a rock bolt (the point load is Easy to carry out.
supported simulating the rock bot)l
Disadvantage:
EFNARC (1999) use the test rig
Do not give the response from the
in Figure 2-31 b) as a standard
distributed load (pressure from
test method
loosing rock), but indirectly by the
point load (load from rock bolt).
Figure 2-31 Shows the stress situation around Disadvantage:
Clamped a rock bolt (the point load is Do not give the response from the
boundary simulating the rock bot).
l distributed load (pressure from
condition The clamped boundary simulates loosing rock), but indirectly by the
the conditions for an inner panel. point load (load from rock bolt).
Difficult to reach clamped conditions.
Figure 2-32 a) Simulates a uniform pressure Advantage:
from loosening rock on the This test method gives the response
surface of the shotcrete panel, from a uniform load (pressure from
The panel is supported with rock loosening rock)
bolts and the steel frame.
Disadvantage:
Relatively complicated testing method
compared with the other described
methods.
Figure 2-32 b) The supporting media describes Advantages:
the loosened rock, which is Easy to carry out.
supported by the shotcrete panel.
Disadvantages:
The point load on the test panel
simulates the impact from the The capacity of the shotcrete depends
rock bolt. on the stiffness of the supporting media
— difficult to make repeatable tests.

The South African waterbed test, see Figure 2-32 a) describes the behaviour of shotcrete very
well when the shotcrete acts as a retaining element. Therefore, only results from this test
method will be presented. The centre distances between rock bolts, one metre is according to
Kirsten et.al. (1993) realistic for mines in South Africa. Kirsten (1993) tested two types of
steel fibres, Dramix and Malt extract. In this review only the results with Dramix fibres are
reported. Panels reinforced with steel fibres were compared with panels reinforced with
diamond mesh (0 3.2c100 mm). The tests were performed in two series and shotcreted with
the dry mix method.
 39

The compressive strength varied from 47 MPa to 80 MPa in series 1 and from 62 MPa to 64
MPa in series 2. Figure 2-33 shows that the load bearing capacity of the panels reinforced
with on average 63 kg/m3 steel fibres and the panels reinforced with diamond mesh had the
same magnitude. When the amount of steel fibres was decreased to 34 kg/m3 the load bearing
capacity of the diamond mesh reinforced panels was superior to the steel fibre reinforced
panels.

a) b)
P (kN)
A
Dramix steel-fibre
t= 50 mm
/ t=100 mm
400 fr t=150 mm

300 Diamond mesh

t= 50 mm P (kN)
t=100 mm
200 200
t=150 mm

100 100

50 100
50 100
Deflection
Deflection

Figure 2-33 Load-deformation behaviour of reinforced shotcrete panels, after Kirsten (1993)
a) Panels reinforced with on average 63 kg/m3 (30 mm long) Dramix steel fibre
b) Panels reinforced with on average 34 kg/m3 (30 mm long) Dramix steel fibre

A new series of tests with the same method was presented by Kirsten, Ortlepp and Stacey
(1997). The panels were reinforced with Dramix steel-fibre or Polypropylene fibres or
diamond mesh. The test panels had a nominal thickness of 75 mm. Data for the reinforcement
is presented in Table 2-11. The test result was evaluated with the following deformation
criterion: when 150 mm central panel deflection was reached, the panel should retain a load
bearing capacity equal to or greater than 50% of the peak load capacity.

Table 2-11 Reinforcement for test panels

Reinforcement Dimensions Amount of fibres

Diamond wire mesh 0 3.2c100 -
Dramix, steel-fibre L=40 mm 2.5% by mass
(55 kg/m3)
Polypropylene L=40-50 mm 0.35% by mass
(-.L...8 kg/m3)
 40

Figure 2-34 shows that the mesh-reinforced panels have a good yielding capacity. The panels
reinforced with fibres also met the deformation criterion. Both types of fibres straddled the
cracks as deformation took place, and tended to pull out rather than to fail in tension. The
peak load capacity of the SFRS panels was significantly higher than the panels reinforced
with polypropylene fibres as well as the panels reinforced with diamond mesh.
110-

140

ISO

141411« MKS

I 1
P013146011110C

03

40

20

0
20 40 00 00 IIO 1i0 140 110
Ma» (nee)

Figure 2-34 Load deformation behaviour of reinforced shotcrete panels (Ortlepp et.a1.1999, a
summary of results by Kirsten, Ortlepp and Stacey (1997))

Morgan et.al. (1999) tested shotcrete with the same type (and size) of test rig as Kirsten
(1993) and Kirsten, Ortlepp and Stacey (1997). Two types of welded steel mesh and four
types of different synthetic fibres were tested, see Figure 2-35. The panels were shotcreted
with the wet method

Abe y
1.03 1100 e
Si I
48 J 19
.. 4 x 4 8/8 1.41
4.0
1.03 14-41
S2 1 100
i., 48 0.30

i 1Z3, 0.50
S3
50 i 10 (111)
6 x 6 6/6 111150
4.I3

S4
..---,1
F. 35 150
Dimensions of synthetic fibres (mm) Dimensions of welded wire mesh fabric (mm

Figure 2-35 Description of welded steel meshes and synthetic fibres, Morgan et. al(1999)

The compressive strength of the shotcrete varied from 66 to 75 MPa. The amount of fibres is
percent by volume. As shown in Figure 2-36 panels reinforced with synthetic fibres Si and S2
displayed superior residual load carrying capacity to panels reinforced with the welded steel
mesh.
 41

a) b)
120 120

100 100

,..-- 6 036/6 mesh

80
ll •
40 8/8 mesh
[60
3 40
Plain
20 20

0
0 20 40 60 80 100 20 40 60
Central Point Deflection (mm]
60 100
Central Point Deflection (mm]

Figure 2-36 Test results from the panel tests Morgan et.a1(1999)
a) Fibre reinforcement
b) Mesh reinforcement

2.7.3 EFNARC panel test and the Round Determinate panel

The purpose with the EFNARC test is mainly to compare shotcrete panels with different
fibres and/or amount fibres. The test method is shown in Figure 2-37, and is described in the
EFNARC standard (1999). The absorbed energy corresponding to the area under the load-
deflection curve between 0 and 25 mm deflection shall be determined.
100x100 mm
Load point 100x600x600 mm
/ square panel

500x500 mm
simple supports Rigid supporting
fixture

Figure 2-37 EFNARC panel test (Clements, 1999)

One disadvantage of the method is the scatter of the results when the test is repeated,
Clements (1999). Another disadvantage is that the ratio Ä = span/thickness (= 500/100 = 5)
makes the panel act more like a dome than a slab according to Nordström, (2000).

The Round Determinate panel test developed by Dr Stephen Bernard at the University of
Western Sidney (Australia), consists of a circular plate with a diameter of 800 mm and a
 42

thickness of 75 mm, see Figure 2-38. The aspect ratio is 10, which ensures a bending failure.
Because the panel is supported in three points the panel is statically determinate. Furthermore,
no cutting is required.

100 mm diameter
hemispherical load point
75x800 mm diameter
round panel

Pivoted support I Support points

points with steel 4, on 750 mm
bearing plates / diameter
11 .
11.1111PZ
( el Oa I LW'

Figure 2-38 Round Determinate panel test (Clements, 1999)

2.7.4 Dynamic testing of retaining element

A number of tests have been performed in South Africa (Ortlepp and Stacey, 1997) and
Canada (Tannant and Kaiser, 1997) to determine the behaviour of shotcrete exposed to
dynamic loads. In the test rig used by Ortlepp and Stacey (1997) the panels were supported by
four bolts, see Figure 2-39. Tannant and Kaiser (1997) used a test rig where the test panel was
supported by columns as shown schematically in Figure 2-40. In both tests an impact hammer
was dropped on the panels. Data for the impact hammer and kinetic energy is presented in
Table 2-12. The South African tests were done with higher impact energy than the Canadian
tests but on the other hand the impact was distributed through concrete blocks in direct
contact with the retaining support, which damped the response, see Figure 2-39. The packed
concrete block should simulate the rock mass and distribute the load to the whole support
area.
 43

muffle WM»
""17-all »Maul»

ie Mu"
•»62*#
i ::71
ri11141111111111ffl 1111111111
2nemndmä

E EMI Kea323 'M

Figure 2-39 Test rig, Ortlepp and Stacey (1997)

Figure 2-40 A schematic representation of the Canadian test rig, Tannant and Kaiser (1997)

Table 2-12 A summary of the two testing methods

Test method Weight of Drop height Maximum Maximum
drop weight velocity kinetic energy
(kg) (m/s) kJ
(m)
South African test 1048 8.5 38
Ortlepp and Stacey, 2706 - 7.3 70.6
1997
Canadian test 565 3 7.7 16.6
Tannant and Kaiser,
1997
 44

In the South African tests various types of retaining support were tested such as welded mesh,
diamond mesh, wire rope lacing and plain/reinforced shotcrete. A summary of the test results
is shown in Figure 2-41 and Table 2-13. The figure comes from Ortlepp et.al.(1999) which is
a summary of the results from Ortlepp and Stacey (1997). The tensile strength"; of the welded
mesh and the diamond mesh was 485 MPa and 285 MPa respectively. With addition of wire
rope lacing the capability of the support system to absorb energy was considerably increased
as Figure 2-41 shows.

Table 2-13 Summaries of results from the South African tests, after Ortlepp and Stacey
(1997)
Support Reinforcement Energy Range Dimensions No of Tests
(kJ)
Welded mesh 9
03.5c100 and - 1.3 to 8.9
04c100 2m x 2m of mesh
Diamond mesh supported by four rock 9
03.2c100 and - 6.4 to 15.9 bolts spaced 1 m apart.
03.2c75
Shotcrete, Plain shotcrete 6.1 1
nominal thickness
=100 mm Welded mesh 15.4 1
04c100 1.6mx1.6m shotcrete
panel supported by
Dramix 30 mm four rock bolts spaced 4
(2.75% by mass 10.3 to 20.6 1 m apart.
(.-._ 60 kg/m3)
50 mm 3
monofilament 10.3 to 15.4
fibres
0.5% by mass
(11 kg/m3)

500
A
e
.—.
11 400 •
• • •
ss-z' 300 •• •
40

t 200
• • A.
•
4 •
a> •
fl 100 *4 • • •
••
0
0 10 20 30 40 50 60
Kinetic Energy (kJ)

Figure 2-41 Results from the South African tests, Ortlepp et.al. (1999)
 45

The test results presented by Tannant and Kaiser (1997) were focused on shotcrete panels
with two different thicknesses, reinforced by welded steel mesh. The result was presented as
minor (A), moderate ( ) and severe (0) damage. Minor damage was associated with
significant cracking in the shotcrete but no broken mesh wires. Moderate damage was
characterised by extensive shotcrete fracturing and spalling with a few broken mesh wires.
Severe damage occurred when the panel was torn open, extensively fractured, and sustained
more than ten broken mesh wires. When severe damage occurred the mesh reinforcement lost
its retaining function. Figure 2-42 compares the kinetic energy contained in the drop weight at
impact with the final deflection of the panel at the central point. Tannant and Kaiser (1997)
observed that the severity of damage is largely controlled by the panel's deflection resulting
from the impact. Different thicknesses of shotcrete sustain similar damage at the same state of
deflection, although the amount of energy required causing a given deflection or damage
increases with the thickness of the panel.

25
•
total estruction
20

Du
>, 15 damage
- minor moderate severe levels
C

0.1 0.2 0.3 0.4 05

Centre Deflection (m)

Figure 2-42 Impact energy versus total panel deflection (shaded symbols for 75 to 90 mm
thick panels, solid symbols for 60 70 mm thick panels), Tannant and Kaiser
(1997)
 46

3 THE KIIRUNAVAARA MINE

3.1 General
The mining company Luossavaara Kiirunavaara Ltd, (LKAB), has been mining iron ore for
more than 100 years in the mines of Malmberget and Kiruna; both towns are situated in the
north of Sweden, above the Arctic Circle.

The Kiirunavaara mine (in Kiruna) is one of the largest underground mines in the world, with
an annual production of 20 to 21 million tonnes of iron ore. The ore body strikes nearly north-
south and dips 60° to the east, is 4000 m long of which 3200 m is mined, 80 meters wide in
average and extends to an estimated depth of 2000 m, see Figure 3-1. A horizontal section of
the ore body is shown in Figure 3-2.

Metres from Above sea level,

mountain top Metres

142
230 Luossai.vi lake
224
320
420

Figure 3-1 The ore body

Since the start, the mining level has moved successively deeper. The new main level situated
1045 m below the former top of the mountain will ensure production in Kiirunavaara for
another 18 years. Up to the present, nearly 900 millions tonnes of iron ore have been mined.
 47

Horisontalsn It I.
Ski va 792
410410,22H

11200 11000 11200 /1400 00 11200 12000 12400 12.0 rem°. 13000 13200 Y 13200 11000 SOO 12400 14210L
2000:000

25100 North 2f10

25400 20100

211200 20200

250
B-ore, see Section 3.2
D-ore, see Section 3.2
12000 22000

1.00 11300 11 CO V/ GO 11200 11000 12200 12400 12000 Y21100 12000 CO 12 GO 13800 12000 14200 14 00 14330'
0''

Figure 3-2 A horizontal section of the ore body at level 792

3.2 Regional geology

Many authors, among others Hansson (1999), have interpreted the regional geology in
Kiirunavaara, therefore, only a short summary will be given. The ore body is formed as an
intrusive sill, now tilted as shown in Figure 3-1. The footwall mainly consists of trachyte,
internally designated as syenitporphyry, see Figure 3-3 . A distance to the west a slightly older
rhyolite with trachytic intercalations occurs. The hanging wall consists of rhyolite, internally
designated as quartzporphyry. Lenses of trachyte are observed. Dikes of diabase and porphyry
occur and at depth some bodies of intermediate intrusives appear. The main iron ore consists
of magnetite. Internally it is divided into two qualities, apatite rich ore (D-ore) with a phospor
content of 0.4 to 4 % and an apatite poor iron ore (B-ore) with phospor content of less than
0.1%. The main ore has the average in-situ grades with 40% cut-off presented in Table 3-1
and is fine-grained, mm. The calculated ore-reserve down to 1500 meters level is 1100
million tonnes.
 48

Geology, Kiruna Area

After Paul Forsell 1985
Simplified end modified
1997 1(•/1. Hague.

Legend

MI Granite

77 Schist, Quartzite, Conglomerate

111. Syenite
NM Gabbro
Phorpyry dike
MI Iron Ore
71Kiruna Rhyolite
▪ Trachyte
11111 Syenite
Rhyolite
Trachytic intercalations
$ II Kurravaara Conglomerate
LA Andesite
•• • Copper Ore
▪ Tuff, Schist, Chert, Marble
▪ Greenstone
1111111 Basement Granite

Figure 3-3 Geological map of the Kiruna area, after Hansson (1999)

Table 3-1 In-situ grade of the Kiirunavaara ore (Hansson 1999)

Fe% P% Na20% K20%
64 0.41 0.06 0.18
 49

3.3 Mechanical properties of intact rock

The uniaxial compressive strength, cre, density, p, Young's modulus, E, and Poisson's ratio,
v, for intact rock at the Kiirunavaara mine are shown in Table 3-2 and Table 3-3.

Table 3-2 A summary of the uniaxial compressive strength for intact rock cr, in the
Kiirunavaara mine
Uniaxial compressive
strength, a,
Footvvall 90 - 430 4
Hangingwall 100 — 280 b)
Iron ore 135 -l85 4
a) Holmstedt (1995)
b) Sjöberg et.al. (2001)
c) Holmstedt (1994)

Table 3-3 Density, p, Young's modulus, E and Poisson's ratio, v for the intact rock, in
the Kiirunavaara mine
Density, Young's modulus for Poisson's ratio
p (kg/m3) intact rock E, (MPa) v
Footwall 2800 45 — 70 a) 0.24— 0.27 4
Hangingwall 2700 37 — 81 4 0.20— 0.26 4
Iron ore 4600 — 4800 d) 40 — 110 b) 0.25 '
a) Sjöberg et.al. (2001)
b) Holmstedt (1994)
c) Paganus and Stephans son (1990)
d) p = 4500 kg/m3 used in the further analyses

3.4 Rock mass structures

To quantify the behaviour of rock is difficult. It depends on the heterogeneous nature of the
fractured rock mass. The nature and distribution of structural features are known as rock
structures (Brady and Brown, 1999). These rock structures mainly cause the heterogeneous
nature of rock. In this thesis the terms discontinuity or joint are used as term for all fractures
and features in a rock mass that have zero or relatively low tensile strength.

The discontinuities in the mine have been mapped by among others Holmstedt (1994, 1995)
and Magnor and Mattson (1999). The orientation of the main discontinuities in the mine is
shown in Figure 3-4 a) and the frequency of the discontinuities is shown in Figure 3-4 b). As
one can see the discontinuities are mainly oriented along the main ore body and perpendicular
to the ore body.
 50

b)

Strui.rer 10.4.1
104 0.104 00 I 04041041040 I I

Figure 3-4 The main discontinuities in the Kiirunavaara mine, (Magnor and Mattson, 1999)
a) Schematic presentation of the discontinuities
b) Contours of pole concentrations

3.5 Mechanical properties of discontinuities

3.5.1 Typical values at the Kiirunavaara mine

In this section the mechanical properties of the discontinuities at the Kiirunavaara mine are
presented.

The strength and stiffness of a discontinuity can be characterised by the friction angle, Ø , the
residual frictional angle, Ø,. , the Joint Roughness Coefficient, JRC, and the Joint Wall
Compressive Strength, JCS. Typical values for the Kirunavaara mine are summarised in Table
3-4. The influence from ground water in the production area is small and fully drained
conditions are therefore assumed. The cohesive strength c is assumed to be equal to zero.

Table 3-4 Mechanical properties for the discontinuities in the iron ore
Part of Dip of the Spacing of the Friction Residual JRC a) JCS a)
the mine discontinuities discontinuities angle friction angle
(°) (m) (0) (0) (MPa) (MPa)
Y18-24 6065 b 0206 b 3135b 28° 5-8 80-112
Y24-28 6070b 0. 6-( 1.0) b 3135b 28° 5-8 80-112
a) Barton (1988)
b) Holmstedt (1999)
 51

3.5.2 Normal stiffness

The normal joint stiffness is according to Bandis et.al (1983) defined as

1
k„ = k (MN I m2I film) (3.1)
o;, )2
(1
Vk +

where
k = —7.15+1.75. JRC + 0.02- (JCS I a j ) (3.2)
is the initial normal stiffness

V„, = A+ 13 • (JRC)+C • (JCS I a j )D (3.3)

is the maximum closure of the discontinuity and

a j = (JRC I 5) • (0.2 • (o-, /JCS)— 0.1) (mm) (3.4)

is the average aperture thickness. o-„ is the normal stress, ac is the uniaxial compressive
strength, JRC is the Joint Roughness Coefficient and JCS is the Joint Wall Compressive
Strength. The empirical parameters A, B, C, and D in (3.3) are listed in Table 3-5.

Table 3-5 The empirical parameters used in Equation (3.3), after Bandis (1983)
Parameter Cycle 1 Cycle 2 Cycle 3
A -0.2960 -0.1005 -0.1032
B -0.0056 -0.0073 -0.0074
C 2.240 1.0082 1.350
D -0.2450 -0.2301 -0.2510

In this thesis the values corresponding to cycle 2 are used.

3.5.3 Shear stiffness

The shear stiffness is according to Barton and Choubey (1977) defined as

Ics = (100/Li ) • a„ • tan[JRC • logia [JCSIo-J + (3.5)

where k, is the peak shear stiffness (MN/m2/m), L3 is the joint length (m) and Or is the residual
frictions angle.
 52

3,5.4 Shear strength

In this thesis the relationship between the shear strength and the normal stress according to
Mohr-Coulomb will be used,

r f No failure
(3.6a)
> z f Shear failure
where is the shear stress at failure according to Mohr-Columb

1".1- =C+0" tan 0 (3.6b)

where c is the cohesion, ci- is the effective normal stress, 0 is the friction angle.

3.6 In-situ stresses

In the late 50's Prof. Nils Hast measured rock stresses in the Kiirunavaara mine. His results
were, however, very controversial, due to the fact that his results showed horizontal stresses
which were higher than the vertical stresses. At that time the general opinion was that the
horizontal stresses were caused only by gravity, see for example Hanzagi (1965).

During the last 20 to 30 years rock stresses have been measured in the Kiirunavaara mine, the
most used method has been the overcoring rock stress measurements using soft inclusion
cells.

Paganus and Stephansson (1990) derived the following relationship for the Kiirunavaara
mine, which was based on current measurements.

o = —4 + 0.027z (MPa)
0-, = —8.1+ 0.04z (MPa) (3.7)
0-, = —8.9 + 0.036z (MPa)

where crif is the horizontal stress perpendicular to the ore body, 01 is the horizontal stress
parallel to the ore body, a„ is the vertical stress and z (m) is the vertical co-ordinate in the
mine. The vertical co-ordinate z = 0 was the old peak of the mountain Kiirunavaara.
Nowadays this peak is mined, therefore the minus sign in (3.7) and (3.8).

Sandström (1999) presented a relationship, which was based on all (reliable) measurements
from 1979 to 1995. Measurements disturbed by the mining were deducted in Sandström's
relationship

Cfv = 0.031z (MPa)

0-, = —4.3 + 0.044z (MPa) (3.8)
ah = —7.6+ 0.043z (MPa).
 53

The above relationship can be compared with Stephansson et.al (1987) relationship, which
applies to the conditions in Fennoscandia.

o-, = 0.027Z (MPa)

= 2.8 + 0.0399Z (MPa) (3.9)
ah 2.2 + 0.0240Z
= (MPa)

where Z (m) is the depth and ow > ch •

3.7 Mining method — sublevel caving

The mining method is large-scale sublevel caving with a sublevel height of 28.5 m, see Figure
3-5. In the first stage drifts along the ore body, footwall drifts, and cross cuts are mined. A
cross cut is a dead-end drift that crosses the iron ore. The cross cuts are 7 m wide, 5 m high
with a c/c distance of 25 m. Drilling is done with electric-powered hydraulic drill rigs. Rounds
of up to 60 holes, each five metres deep, are drilled. These holes are charged with explosives
and blasted.

Production drilling
Production drilling of the fans is done with remote-controlled production drill rigs, see
Figure 3-5. The operators control several rigs in the production area by remote control from
control rooms. The rig drills upward into the ore, forming fan-shaped patterns with ten to
eleven holes each. The longest holes are normally 40-45 metres long. The distance between
each fan, that is, the burden, is 3 m. Once this is completed, charging of the holes can begin.

Charging, blasting
Charging of the holes is performed by a purpose-built truck. Blasting is executed every night.
Each round produces about 10,000 tonnes of ore. In Section 5 the charging and blasting is
discussed in more detail.

Loading
Electric wheeled loaders (LHD) haul the ore to ore passes located along the ore body, see
Figure 3-5 and Figure 3-6. In the mine there are also remote-controlled electric loaders. The
operator (loader) sits in front of a monitor, in a control room, and 'drives' the machines in the
production area. The machines navigate with the help of rotating lasers and reflectors on the
walls of the drifts. Information, e.g., the position of the machine, is transmitted via a number
of wireless base stations to the control system in the control room computer. Each loader
carries a bucket payload of 17-25 tonnes, depending on the type of loaders. The ore is dumped
into ore passes, which end in bins at the main haulage level, see Figure 3-6.
 54

a) b)

28.5 m

Charging and
blasting

Production
drilling 28.5 m

25 m

Figure 3-5 Large-scale sublevel caving in the Kiirunavaara mine

a) Large-scale sublevel caving, principles, after Hamrin (1986)
b) Fans for production blasting

The main haulage level in the Kiirunavaara mine is at the 1045 m level, see Figure 3-6
The ore is tapped from the bins into a remote-controlled train. The driverless train, consisting
of an engine and 24 cars, carries the ore to one of four discharge stations. Nine locomotives
and about 185 cars are in operation at the main level. Each train carries about 500 tonnes of
ore. When the train passes the station, the bottoms of the cars are opened and the ore is
emptied into a crusher bin from which it is fed to one of four crushers. The ore is crushed into
fragments of about 100 mm in diameter.

When the ore has been crushed it is carried by a conveyor belt to the ore elevators (skips). The
ore is loaded into the skip automatically and hoisted, at a speed of 17 metres per second, up to
the processing plants at the surface. Each skip carries 40 tonnes of ore.

In the processing plants, crude ore is transformed into iron ore products such as pellets and
sinter fines. The iron ore products from Kiirunavaara are transported by train to the harbour in
Narvik, Norway, which is ice free all year around.
 55

Hoisting

KUJ 2000

Ore passes

Main transport level 1045 m

Figure 3-6 The Kiirunavaara mine, from LKAB's image archive (2001)

3.8 Rock support

The rock support in the Kiirunavaara mine has two functions: (i) to immediately stabilise the
drifts in order to secure the subsequent round, (ii) ensure the stability of the drifts during the
whole production cycle (development, production drilling, charging, blasting loading and
hauling), that is during two to three years.

The main support system in the Kiirunavaara underground mine is untensioned fully grouted
dowels and shotcrete where more than 80% of the shotcrete is plain shotcrete. In areas with
poor rock conditions or where stress induced problems are met, cable bolts and/or steel fibre
or net reinforced shotcrete may be used.

The main purpose of the shotcrete is to help the rock mass to maintain its integrity. The
adhesion strength between rock and shotcrete is therefore crucial, especially for the plain
shotcrete. Approximately 10,000 m3 of shotcrete a year is used as rock support in the mine.
 56

3.9 Stress changes caused by the sublevel caving

The changes of the stress state caused by the mining is shown in principle by stress
trajectories in Figure 3-7. Quinteiro (1998) analysed the global stress situation and how it
changes due to the mining activities with a 2-D model focusing on level 775. Input data for
the analysis is presented in Table 3-6. Furthermore, the used in-situ stress profile was

o-, =0.027z
aH =1.5av (MPa) (3.10)
ah = av

where z is the depth. This in-situ stress profile differs from the in-situ stress profiles presented
in Section 3.6.

Major principal stress, oj

Figure 3-7 Schematic section of the sub level caving (Sjöberg et.al, 2001)

Table 3-6 Input data for the numerical analysis (Quinteiro, 1998)
Young's Poisson's Cohesion Friction Density Density Ground
modulus ratio angle iron ore waste rock acceleration
(GPa) (MPa) (0) (kg/m3) (kg/m3) (m/s2)
40 0.25 15 35 4600 2700 10
_

Figure 3-8 shows the variation of the principal stresses as a function of the mining level. The
major principal stresses increase strongly close to the footwall. Furthermore, the maximum
stresses are reached when the mining level is at about 20 to 60 m above the 775 m level
depending on the distance from the footwall. The minor principal stresses are decreasing as
the major principal stresses are increasing. Figure 3-9 shows that the major principal stress
vector rotates and will have approximately the same dip as the ore body when the mining
level is close to the level 775. This variation of the direction of the principal stresses makes it
difficult to design the drift (shape and rock support) in an optimal way.
 57

Ore body

Mining level
70 -
12.5 m Distance inside footwall
60 -
775 m level
50 -
›Distance inside footwall
a_
2 40 -
cd, Major Stress
30 -
Minor Stress 12.5 m
20
100 m
10 7

0 100 200 300 400 500 600 700 800 900 1000
Mining Level

Figure 3-8 Principal stresses at the 775 m level as a function of the mining level, Quinteiro
(1998)

80 -

70 -
ore body dip
60

50 12.5
Ore body
3 40 - Mining level
a 100 m
-o
30 - Major principal stress Distance inside footwall
t71)
< 20 - 775 m level - Angle, positive clockwise

10 - Distance inside footwall

0 -

-10 -

-20
0 100 200 300 400 500 600 700 800 900 1000
Mining Level

Figure 3-9 Angle between horizontal and major principal stresses at the 775 m level as a
function the mining level, Quinteiro (1998)

Lundman (1998) studied the stress situation in the vicinity of a cross cut in the Kiirunavaara
mine. Due to the complex geometry and mining sequences, a 3D boundary element program
 58

(EXAMINE3D, 1998) was used. Both a single cross cut and multiple cross cuts were analysed
to determine the stress situation caused by mining within the cross cut. The tangential stress
cm in the roof was determined by applying a unit stress ow= 1 or crh = 1, where ow is the
horizontal stress perpendicular to the ore body and o-h is the horizontal stress parallel to the
ore body. The tangential stress (so was calculated for different stages of the retreat
characterised by the retreated distance (Lm) and the distance (4) to the drawpoint, see Figure
3-10. The influence of ow on (Ye (when o-h = 0) in the roof is local. Only a few meters in front
of the drawpoint are influenced by the ow, see Figure 3-11.

Figure 3-10 Illustration of the retreated distance (Lm) and the distance (4) to the drawpoint
(Lunclman, 1998).

0-0 0-H
0.2

-0.2

-0.4

-0.6

-0.8

-1
0 5 10 15 20 25 30

Distance to drawpoint, 4(m)

Figure 3-11 The tangential stress 0-0 in the roof vs the distance to the drawpoint (4) for the
loading case ow= 1 and o-h= 0 (Lundman, 1998).
 59

Lundman found from the numerical analyses that the tangential stress in the roof of the
drawpoint could be expressed as

= 2.20-,L„,° 12 - 0.750, (3.11)

where Lm is the retreated distance, 1 < Lm < 50 m. Figure 3-12 shows how the tangential stress
at the drawpoint, = 0, increases with increasing Lm.

Lundman (1998) also investigated how the mining in one or two cross cut(s) affected the state
of stress in an adjacent cross cut. He concluded that mining affects the adjacent cross cut
considerably with a decrease in stress level in the roof. By combining the results from both
the single cross cut and the multiple opening analysis Lundman concluded that the state of
stress at the drawpoint is governed by the position of the mining face in the actual cross cut as
well as the position of the mining face in the adjacent cross cuts.

In addition to the stress variation the drawpoint is subjected to dynamic loads from the
production blasting, which is discussed in Section 6.

4.0 -

3.0

0.0
0 10 20 30 40 50
Retreated distance, L (m)

Figure 3-12 The tangential stress S0 at the drawpoint 0) vs the retreated distance (Lm)
according to Equation (3.11), o, =r 1 and cifi = 0.
 60

4 SHOTCRETE - FIELD AND LABORATORY TESTS

4.1 General

This Section covers field and laboratory tests of shotcrete and field observations of shotcrete
linings. An extensive failure mapping was carried out to identify typical failures of shotcrete
in the Kiirunavaara mine, see Section 4.2.

One part of the laboratory tests has dealt with the compressive and adhesion strength of
shotcrete as functions of curing time (Section 4.3). As mentioned in Section 1 many tests have
been performed to investigate the development of the compressive strength of shotcrete as a
function of the ageing process. The author has found only limited information about the
development of the adhesion strength of shotcrete during the ageing process in the literature.
In the literature review in Section 2 only results from 28-day-old or older shotcrete are
presented with respect to the adhesion strength. The objective of this test was to compare the
development of adhesion and compressive strength of shotcrete during the ageing. The tests
have been performed underground at the Kiirunavaara mine.

Furthermore, the bearing capacity of the reinforced shotcrete has been tested using panel and
beam tests at the Luleå University of Technology, see Section 4.5. The test specimens were
reinforced with steel fibres or welded steel mesh, the amount of steel fibres varied from 50 to
80 kg/m3 (nominal values).

In the Kiirunavaara mine, scaling is performed in all drifts and crosscuts, normally with a
hydraulic scaling hammer Before shotcreting, the rock surface is cleaned using water. During
the past few years, water-jet scaling has been tested as an alternative method to mechanical
scaling. The adhesion strength of shotcrete on water-jet scaled rock surfaces has been
compared with shotcrete on rock with normal treatment. These field tests are presented in
Section 4.4. Finally the results are discussed in Section 4.6.

4.2 Failure mapping of shotcrete

To improve the understanding of the behaviour of shotcrete in the Kiirunavaara mine an

extensive mapping of 7 km of drifts was carried out. The damage to the shotcrete was
classified and important parameters, such as thickness, size of fallouts and geology, were
noted.

After an initial mapping the damage to the shotcrete was divided into two types; fallout of
only shotcrete and fallout of rock and shotcrete, see Figure 4-1. More than 80% of the damage
was small fallouts of plain shotcrete. This means that the adhesion strength between the
shotcrete and the rock in the area of failure was poor. Figure 4-2 shows typical cases where
fallouts of only shotcrete have occurred in a drift in the mine.

The shotcrete thickness was 2 cm or thinner for more than 90% of the fallouts of plain
shotcrete. The difficulties of shotcreting on an irregular rock surface were also confirmed by
this mapping since 60 % of the failures were in conjunction with an apex on the rock surface
together with a thin shotcrete layer.
 61

a)

Figure 4-1 Fallout ofshotcrete and rock (Malmgren and Svensson, 1999)
a) Fallout of only shotcrete
b) Fallout ofrock and shotcrete

Figure 4-2 Typical fallouts ofshotcrete only

To get an idea ofthe adhesion strength condition around a fallout, the area around the failure
was tested by knocking with a hammer to estimate the area with lack ofcontact between the
shotcrete and rock. This was a very crude method, which still gave some valuable
information. The crack length was on average 0.3 m, which can indicate low adhesion
strength in the failure area, see Figure 4-3. These crack lengths can be compared with the
results from a series oflaboratory tests carried out by Holmgren (1979). In these tests, the
crack Iength between rock and concrete prior to failure was about 5 cm with a shotcrete
thickness of2 cm, the adhesion strength was 0.9 - 2.3 MPa, which can be characterised as
high adhesion strength.
 62

Fallout of shotcrete

Area with lack of contact between shotcrete

and rock - 0.3 m on average

Figure 4-3 Typical situation around a fallout of shotcrete, after Malmgren and Svensson
(1999)

The results from the mapping are clear. The shotcrete is most likely to fail where the shotcrete
layer is thin, i.e. 2 cm or thinner, and the adhesion strength is poor to non-existent. Important
factors, which may cause the failure, are rock deformation, shrinkage and dynamic loads due
to blasting.

If plain shotcrete is to be used, the results indicate that it should be used in areas with
relatively good adhesion strength and not subjected to large deformations and dynamic loads.

In the Kiirunavaara mine the plain shotcrete is on average 4 cm thick according to results
from thickness tests performed in16 cross cuts in the mine, see Figure 4-4 a), but with a wide
scatter. The nominal thickness or ordered thickness was 3 — 5 cm. As one can see in Figure
4-4 b), 18% (5 + 13) of the tests showed a thickness less than or equal to 2 cm. The failure
mapping discussed above showed that the shotcrete thickness was less than or equal to 2 cm
for more than 90% of the fallouts of plain shotcrete. If the shotcrete lining with ordered
thickness of 30 — 50 mm had the same distribution of shotcrete thickness as shown in Figure
4-4 b), the supporting ability of as much as 18% of the shotcrete lining can be considered as
more or less negligible.

a) Average value =42 mm b)

Standard deviation = ±23 mm
Distribution of shotcrete thickness
a)
150i
e.'.
0 \,
o .--. 0 .• • .7 20% 116%
-c E 100 -. • • 13% - - 13%
`') E ••„„ .. • ..„, , • .„. 8%
1105%
% 5%fl 4%
>, 5%'fl 2 % 1% 1% 1%
u)
0% 1 n
c
.x
— h. e• 1/4* 4r* ' 0 0 0 0 0 0 0 c, o cp c,
-72t u? 9 t• 9 9 c)
a-- ..-
Cal
c. . .
I- ‘— Ol 01 xl• In CO 1,
-
• a—
a— a—

Thickness of shotcrete layer (mm)

Figure 4-4 Thickness of shotcrete layer in the Kiirunavaara mine, totally 370 tests
a) Each point represents one test (drilled holes), Svensson (1998)
b) Distribution of shotcrete thickness
 63

4.3 Adhesion strength and compressive strength of shotcrete as a function of

time

4.3.1 General

To increase the understanding of the adhesion strength of shotcrete, especially young and
curing shotcrete, a number of tests were carried out.

It was decided to use a well-cleaned concrete wall because of its uniform properties compared
with a rock surface. The wall surface was sandblasted using a dry-mix spraying machine. The
tests were carried out underground, at a temperature of approximately +12°C and at a relative
humidity of 80 %.

The shotcrete mix used in these tests is presented in Table 4-1. This is one of the standard
shotcrete mixes used at the Kiirunavaara mine. The accelerator used was waterglass (Sodium
silicate, Na20.Si02.H20). Two different concentrations of accelerator were tested; 4% and 8%
of the weight of the cement and silica.

Table 4-1 Mix design.

Cement (kg/m3) 450

Silica (kg/m3) 30
Aggregate, dry weight (kg/m3) 1565
Slump (mm) 150
W/C 0.45

4.3.2 Compressive strength tests

The compressive strength of the young shotcrete was assessed using two non-destructive
methods, for 2 to 4 hour old shotcrete a penetration method and for 18-hour-old shotcrete a
piston method (Hilti AG). The penetration method is a non-destructive method. A penetration
needle is placed against the shotcrete surface and pressed into the surface (15 mm). The
resistance or pressure is plotted on to a calibration curve where the compressive strength can
be estimated. The piston method is also a non-destructive test method. A stud is shot into the
shotcrete layer and the pullout force is measured and plotted on to a calibration curve where
the compressive strength can be estimated.

When the shotcrete was at least three days old, the compressive strength was determined by
cube testing, according to the Swedish Standard SS 13 72 20 (1984). The size of the cubes
was 100-100.100 mm. Cube tests were carried out for 3, 7, 14 and 28-day-old shotcrete. The
shotcrete was sprayed in boxes, one box for each age. Three cubes from each box were sawn
out prior to the compressive testing.

Figure 4-5 shows the compressive strength versus shotcrete age for high and low accelerator
concentrations. Each point in the diagram represents the average value for three cube tests. As
can be seen, the difference between low and high concentration of waterglass is small.
 64

a) b)

--00- 40
.c 40-
2 35 . 35-
-.E 30 30 -
g 25 71)
a) co 25 8% waterglass
to 20 cl- 20 - -4% waterglass
a)
.> 15 15-i
(. i.D.
co 10 o_ 10
E
o 5-
o
0 -
0 10 20 30 0.01 0.10 1.00 10.00 100.00
Day Day

Figure 4-5 Compressive strength, (Malmgren and Svensson, 1999)

a) linear scale
b) log scale

4.3.3 Adhesion strength tests

The adhesion strength tests were carried out in accordance with the Swedish Standard (SS 13
72 43, 1987) using the equipment shown in Figure 4-6, (CGE,1995). The adhesion strength
testing method is limited to be used on a relatively stiff shotcrete. Therefore, the adhesion
strength tests were carried out on at least two-day-old shotcrete

a) b) c)

Figure 4-6 Adhesion strength test equipment (from Swedish Standard SS 13 72 43, 1987)
a) Drilling two circular slots with double diamond drill
b) The inner circular slot is drilled through the shotcrete
c) The test machine, attached to the shotcrete drill core for the pull out test
 65

The shotcrete was sprayed in two different areas of the concrete wall, area 1 and 2. The
distance between the nozzle and the surface is normally 2-3 m. When area 2 was sprayed,
however, the distance between the nozzle and the surface was 3-4 m. The results from the
adhesion tests are shown in Figure 4-7. Figure 4-7 also shows regression curves based on a
logarithmic formula

= A • ln(Day) + B (4.1)

wherefa is the adhesion strength, Day is time (days) and A and B are constants determined in
the regression analysis.

There was a significant scatter in the adhesion strength values with respect to location on the
wall, as well as to the concentration of accelerator. Another observation was that there was a
greater difference between the adhesion strength of the different areas on the wall than
between low and high concentrations. The variation in nozzle distance was the main reason
for the variation of the adhesion strength.

When using a well-cleaned concrete wall, the adhesion strength was noticeably higher
compared with the adhesion strength of a rock surface, (compare with the results in Section
4.4).
 66

2.0
1.8
Areal - 8% acc.
8
o Areal -4% ace.
• Area 2 - 8% acc.
x Area 2 -4% ace.
C
00.8
Area 1 - 4% acc R2 =0.42
Area 1 - 8% acc = 0.45
Area 2 - 8% acc 12.2 = 0.81
0.0 Area 2 - 4% acc R2 = 0.70
10 20 30
Day

Figure 4-7 Adhesion strength against a well-cleaned concrete wall

4.3.4 Comparison between adhesion strength and compressive strength

The growth of compressive and adhesion strength as functions of time is compared in Figure
4-8, which summaries the results from Section 4.3.2 and 4.3.3. The growth of compressive
and adhesion strength show a clear correlation.

45
Compress ve strength 40
35
Area 1 - 8% ace.
30 a»
25 1-2
C Areal -4% ace.
to
20 c* • Area 2 -8% ace.
C 0.8 /C '87)
20.6 15 g
Upper and lower regression curve for >< Area 2 -4% ace.
<0.4 z 10
the adhesion strength
0.2 ri • 5 ti -•- Compressive
strength
0.0 0
0 5 10 15 20 25 30
Day

Figure 4-8 The relation between compressive strength and adhesion strength
 67

4.4 Adhesion strength of shotcrete when using different types of treatment

methods

4.4.1 Scaling and cleaning methods

In the Kiirunavaara mine, scaling is performed in all drifts and crosscuts, normally with a
hydraulic pick hammer. Before shotcreting, the rock surface is cleaned by spraying water on it
with a pressure of 0.7 MPa (7 bar). This process is hereafter referred to as the normal
treatment. During the past few years, water-jet scaling has been tested as an alternative
method to mechanical scaling. The rig used was a prototype with a water pressure of 22 MPa
(220 bar) and a water flow of 0.2 m3/minute, see Figure 4-9.

Figure 4-9 Water-jet scaling rig

Three methods of surface preparation were tested in crosscuts situated in the iron ore. Seven
different crosscuts were used for the tests, all with rock quality defined in Table 4-2 which is
a quite normal rock quality in the Kiirunavaara mine. The methods used were the normal
treatment, water-jet scaling as mentioned above and mechanical scaling with high-pressure
water washing where the water-jet machine was used only to wash the surface. The mix for
the shotcrete was the same as specified in Table 4-1. A total of 46 adhesion strength tests
were performed with the normal treatment, 14 tests with the high-pressure water washing
method, and 24 tests with water-jet scaling. The adhesion strength tests were carried out with
the equipment shown in Figure 4-6.

Table 4-2 Rock quality in the test area

RQD 40 - 60 %
Joints Unaltered to slightly
altered, clay-free
 68

4.4.2 Results
The adhesion strength for the three methods is presented in Table 4-3. Because of the large
scatter of the result a statistical analysis has been performed to investigate if there were
significant differences in adhesion strength between the three various methods. The test was
done as follows, a test variable t was defined as (Enger)

m1 — m2 (4.2a)
j s1 In1 +s2 / 2
where m1 and m2 are the mean values and ni and n2 are the number of samples for the two
measurements, SI and s2 are the standard deviation for the two measurements.

If t > t„,, (4.2b)

there is a significant difference between measurement 1 and 2 (Enger), ta,,, is the Student's
distribution (Blom, 1970). A 95% confidence interval was used, which gives

a=1— 0.95 = 0.05.

As evident from Table 4-4 there is not a significant difference between normal treatment and
high-pressure water cleaning. However, there is a significant difference between normal
treatment and water-jet scaling. There is also a significant difference between water-jet
scaling and high-pressure water cleaning.

In 36% of the tests in the area with the normal treatment there was no adhesion strength (0
MPa). Corresponding values for the high-pressure water washed areas and water-jet scaled
areas were 17% and 7% respectively.

An interesting observation is the location of the failure surface, Table 4-3. When the rock
surface was prepared with the normal treatment, most of the failure surfaces were located
within the rock. When the rock surface was prepared with water-jet scaling, a greater part of
the failures were located in the contact area between the shotcrete and the rock. For high-
pressure water cleaning no observations of the location of the failure surface were done.

One reason why more of the failures are located at the interface between shotcrete and rock
when the surface is water-jet scaled is that the water-jet scaling removes the blast damaged
layer of rock from the walls and the roof The shotcrete is therefore applied on a rock with
higher quality and strength then for normal treatment. This means that the strength of the
(remaining) rock has increased compared to the strength of the interface between shotcrete
and rock. It should also be noted that the adhesion strength (Table 4-3) was higher for the
water-jet scaled areas compared with the areas where normal treatment was used. Normal
treatment does not give this effect and therefore the failures are frequently occurring within a
layer of rock closest to the tunnel boundary
 69

Table 4-3 Adhesion strength results

Adhesion Location of the failure surface

strength
Type of No of Contact Rock Shotcrete
scaling and tests Mean Std. between rock
cleaning Dev. and shotcrete
(MPa)
method •• • • • : :: ::
(MPa) .:-:'•
.. . . :'::':: :,...:. ... .:•:.•:• ...::•:' :•''
,....ij...,.ii .•..... . . .....
..... . . . :-;:•: ,
-.••••• . • _,••......,

Normal 41 0.21 0.27 20% 75% 5%

treatment
High-pressure 14 0.33 0.22 a) a) a)
water
Water-jet 24 0.61 0.45 55 % 35 % 10 %
scaled

a) No observations

Table 4-4 Results from statistical analysis — significant differences in adhesion strength
between the various treatment methods
Type of scaling and cleaning method t according t 5% Significant difference
to Eq. 4.2b) 2 between treatment
method according to
Equation 4.2 c)
Normal treatment versus high-pressure water 1.66 2.04 No
Normal treatment versus water-jet scaling 3.96 2.03 Yes
Water-jet scaling versus high-pressure water 2.57 2.03 Yes
 70

4.5 Panel and beam tests

4.5.1 Test method and program

A rock mass is not a well-defined material. The actual loads on a rock support systems are
often unknown. However, knowledge of the supporting effect can be obtained by studying
models of shotcrete linings in laboratory tests. Depending on the rock quality and the rock
stress situation, various test models can be used. Many of the various test methods are
discussed in Section 2.

When the mining goes deeper the in-situ stresses increase. If the rock stresses are high the
shotcrete lining in some cases must be used as a retaining element rather than as a support
arch, see Section 2.2 where this support situation is discussed. The rock is assumed to apply a
lateral surface load on the shotcrete lining, i.e. the shotcrete lining acts as a retaining element
for the loosened rock, the lining is supported by rock bolts, see Figure 4-10. This support
situation is simulated in this test. The test model can be designed in many ways, various test
models are discussed in Section 2.7. Based on the discussions in Section 2.7 it was decided to
use a test model shown in Figure 4-11. The test panels were clamped square panels loaded
with a point load in the centre of the panel. The main reason for using a point load as the load
is to investigate the behaviour of shotcrete around the rock bolt, which is often a critical part
of the structure. Also, beam tests were performed which gave the moment capacity of the
shotcrete lining.

Rock
stresses
Rock loads

1 Hold

Reinforce
(strengthen rock
and control bulken)

Figure 4-10 The shotcrete lining acts as a retaining element (Esplay et.a1.1995)

Shotcrete lining

Rock bolts

Test panel - clamped square panels loaded with a point

load in the centre of the panel.

Figure 4-11 Test panel

 71

Following issues were addressed:

- Identification of failure modes of the shotcrete panels.
- Comparison of the bearing capacity of shotcrete panels reinforced with 35 mm and 50 mm
long Dramix steel fibre respectively, two dosages of steel fibres, 50 kg/m3 and 80 kg/m3
were used.
- Comparison of the bearing capacity of steel fibre reinforced shotcrete with shotcrete
reinforced with welded steel mesh.

Following tests were performed:

- A square test panel was loaded by a point load in the middle and with clamped
boundaries. The span was 1500 mm and the nominal thickness 100 mm, see Figure 4-11.
- The moment capacity was determined by beam tests.
- The compressive strength was determined by cube tests (100.100.100 mm) according to
the Swedish standard SS 13 72 20 (1984).

The number of tests, dimensions of the test specimens and reinforcement are presented in
Table 4-5, the shotcrete mix is shown in Table 4-6. The tests were performed at Test Lab at
Luleå University of Technology.
 72

Table 4-5 Test s ecimens

Reinforcement Fibre Panels Beams Cubes
dosage Item No. Item No. Item No.
Dramix 35/0.55 50 kg/m3 Panel]] Beam]] Cube]]
Pane112 Beam] 2 Cube] 2
Pane113 Beam] 3 Cube] 3
Dramix 35/0.55 80 kg/m3 Pane121 Beam21 Cube21
Pane122 Beam22 Cube22
Pane123 Beam23 Cube23
Welded steel Ø5 c 150 in Pane131 Beam31 Cube31
mesh both Pane132 Beam32 Cube32
directions Pane133 Beam33 Cube33
Dramix 50/0,62 80 kg/m3 Pane141 Beam41 Cube4I
Pane142 Beam42 Cube42
Pane143 Beam43 Cube43
Dramix 50/0,62 50 kg/m3 Pane151 Beam51 Cube51
Pane152 Beam52 Cube52
Pane153 Beam53 Cube53
SUM 15 panels 15 beams 15 cubes
Dimensions:
Panel:
Length-Width-Thickness =1900.1900-100 mm
Span = 1500 mm (in both directions)
Beam:
Length-Width.Height=500.125-75 mm
Span = 450 mm
Cube:
100-100-100 nun

Table 4-6 Shotcrete mix

Test serie Test serie Test serie Test serie Test serie
11,12,13 21,22,23 31,32,33 41,42,43 51,52,53
Cement (kg/m3) 499 499 445 498 498
Silica (kg/m3) 25 25 26 24 24
Sand (kg/m3) 1516 1516 1582 1522 1522
W/(C+silica)-ratio 0.40 0.40 0.42 0.40 0.40
Accelerator: Waterglass, the dosage was approximately 6-8 % of the weight of the cement
and silica.
 73

The specimens were shotcreted in the Kiirunavaara mine (underground) and thereafter
covered by plastic sheets to avoid dehydration. After two days underground the specimens
were stored in a storeroom on ground. The temperature was kept about 20° C and the
specimens were covered by plastic sheets.

4.5.2 Steel fibre rebound

Method

The rebound of fibres was tested for all steel fibre reinforced test beams. The weight and the
volume of the beams were measured before the beams were crushed in a crusher, the steel
fibres were collected with a magnet and the weight of the steel fibres, Wsprayed, was measured.
The rebound was defined as

(W W pry d)
Rebound = "'" 100% (4.3)
Wnom

where W,,„„, was the nominal weight of steel fibres per shotcrete test beam.

Result

The rebound of steel fibres was 7 to 14% for the 35 mm fibres and 18 to 25% for the 50 mm
fibres as shown in Table 4-7. Shotcrete reinforced with 50 mm fibres was sprayed by a nozzle
with larger diameter compared to the shotcrete reinforced with the 35 mm long fibres, because
of that the impact velocity decreased and gave higher rebound. The rebound of steel fibres is
according to Banthia et.al (1994) between 12-18% for the wet mix.

Table 4-7 Rebound of shotcrete

Steel fibres Nominal Measured Rebound Test specimens
quantities of quantities of steel according to
steel fibres fibres Eq.(4.3)
(kg/m3) Mean value (%)
(kg/n13)
Dramix 35/0.55 50 43.0 14 Beam No 11 to 13
Dramix 35/0.55 80 74.1 7 Beam No 21 to 23
Dramix 50/0,62 50 37.6 25 Beam No 51 to 53
Dramix 50/0,62 80 65.4 18 Beam No 41 to 43
 74

4.5.3 Compressive strength

The compressive strength of the 100 mm cubes was determined using the relation

F
.fc= f (4.4)
A

where fc is the compressive strength, Ff is the load at failure and A is the area of the cross-
section. The test result is shown in Table 4-8. The result is also presented as the compressive
strength of a 150 mm cube and as the compressive strength of a cylinder (length.diameter =
300.150 mm). These two dimensions of the specimens are standard sizes according to the
Swedish Standard. The conversion of the compressive strength from the 100 mm cube to the
other two dimensions was made in accordance with the Swedish Standard SS 13 72 07 (1988)
— Conversion factors.

Table 4-8 Compressive strength,f,

Reinforcement Dramix 35/0.55 Mesh Dramix 50/0,62

Amount of steel fibre 43.0 74.1 - 37.6 65.4

(kg/m3)
J . (MPa) 52.7 40.9 49.0 46.6 47.2
Cube 100.100400 mm std.dev. 2.9% std.dev. ±1% std.dev. -1-1% std.dev. ±1%
- std.dev. i-1%
fc (MPa) 42.5 33.9 40.6 38.9 39.1
Cube 150150.150 mm
fc (MPa) 31.5 25.1 30.1 28.6 29.0
Cylinder
length = 300 mm
diameter= 150 mm

4.5.4 Beam tests

Method

The beams were tested according to the Swedish guideline Betongrapport 4 (1995). The
guideline is discussed in Section 2. Figure 4-12 shows the dimensions of the test beam. The
residual strength can be expressed as

f10,20 R10,20ffdr

f10,40 = R10,40f1dr (4.5)

f10,60 = R10,60 /kit.

where R10,20, R10,40 and R10,60 are the residual strength factors andffl„ is the first crack
strength, all quantities are defined in ASTM1018 and briefly discussed in Section 2.
 75

F/2 F/2

Cross-section of the beam:

Width • Height = 125-75 mm
150 150 150f
• 450

Figure 4-12 A four-point test beam

Results

A summary of the results is shown in Figure 4-13. Each curve shows the mean value for three
beam tests. A higher amount of steel fibres gave higher toughness, expressed as the residual
strength factors in Figure 4-14. Test beams reinforced with 50 mm long fibres showed a lower
toughness than shotcrete reinforced with 35 mm fibres, mainly caused by the lower amount of
steel fibres. However, tests performed by Groth (2000), see Section 2, have shown that longer
fibres give higher toughness.

The beams reinforced with a steel bar from the mesh showed a much more ductile behaviour
compared with the steel fibre reinforced test beams. But the residual strength of the beams as
defined in Eq.(4.5) was low. One reason was the low magnitude offfl„. The second reason
was because R10,20 and R10,60 show the toughness of the beam for relatively low values of the
deflection II. R10,20 and R10,60 are determined for u in the interval 0 - 5.58„ (5.56„ 0.5 mm)
and 0 - 30.58„ (30.58, .= 3 mm) respectively, where 8„ is the first crack deflection.

'..,,Dramix 35/0.55 74.1 kg/m3

' m3
Dramix 50/0.62 65.4 'kg/ Mesh
6

Dram_ _ 0.55 —43.0

Dram ix 50/067 37:6 kg

2 4 6 8
Mid-point deflection (mm)

Figure 4-13 Load-deflection curves for the beams, each curve corresponds to the average of
three beam tests.
 76

100
StrengthFactor (%)

90
80
•
70 - x
-o
60 Ri0,20
50 - x R10,40
40
Residual

cj- R10,60
30
20 -
10
0
Dramix 35/0.55 Dramix 35/0.55 Mesh Dramix 50/0.62 Dramix 50/0.62
43.0 kg/m3 74.1 kg/m3 37.6 kg/m3 65.4 kg/m3
Reinforcement

Figure 4-14 Toughness of test beams, the residual strength factor R, according to ASTM
1018

ic-i"
o- 4.50 1
2
1 • 3
. 3.50
--g> • • •
C
11) 3.00 i - -• - fio,2o
•
(.75 2.50 -
• —9— f10,60
a)
C 2.00 1 a
e 3 - fikr
c 1.50 * -
cao) 1.00-
76 0.50
-8 0.00
LT)
a) Dramix Dramix Mesh Dramix Dramix
cc
35/0.55 35/0.55 50/0.62 50/0.62
43.0 74.1 37.6 65.4
kg/m3 kg/m3 kg/m3 kg/m3

Reinforcement

Figure 4-15 Residual bending strength

 77

4.5.5 Panel test

Method

The test rig was composed of bottom and upper frames. The bottom frame was made of 400
mm steel U beams, see Figure 4-16 , while the upper frame was made of 200 mm U beams.
The open section of the U beams was closed by steel plates, in order to increase the stiffness
of the frames. The bottom frame was 520 mm high and had eight legs. The legs were fixed to
the 900-mm thick concrete floor by eight 4)40 mm bolts. In order to match the irregularity of
the shotcrete panels, the upper frame was divided into four parts, see Figure 4-16.

b)
Upper steel frame
1500 (span)

1500 (span:

Figure 4-16 Test rig

a) Overview
b) Plan of bottom frame

To match the panel with the upper and bottom frames the gaps between the panel and the
upper and bottom frames were filled with cement mortar. The panel and the upper frame were
tightened to the bottom frame by 12 bolts. After one night (more than 12 hours) setting, the
bolts were fastened with a torque of 450 Nm.

The load was applied to the panel through the load-bearing plate (4)200 mm) by a jack placed
on the floor, see Figure 4-16. The loading speed was about 2.5 mm /min.

Result

The deflection of the panel was measured at 9 points as illustrated in Figure 4-17.
Furthermore, seven radial cracks numbered 1 — 7 around the loading point are shown in
Figure 4-17. The crack load, Fer, maximum load, Fm , the crack deflection, Scr and the
deflection at maximum load, &max, as well as the energy are summarised in Table 4-9. The
energy was calculated as the area below the load-displacement curve and was used as a
measure of the toughness. All load-displacement curves are shown in Appendix 4.1.
 78

The mesh-reinforced panels had the lowest toughness. The absorbed energy varied widely in
the interval 1350 - 3353 J. The toughness of panels reinforced with steel fibres did not vary to
much between low and high amount of steel fibres. The panels with the lower amount of steel
fibres had lower maximum load but were compensated by higher maximum deflection, see
Appendix 4.1. As for the beams it was difficult to compare the action of longer vs. shorter
fibres because of the difference in the amount of fibres. The test results showed that the
typical failure mechanism was a bending failure followed by punching failure as shown in
Figure 4-18.

Figure 4-17 Test arrangement

Table 4-9 Test results anels

Item No Amount of steel fibre fc Fcr 4, Fni. 5F., Energy Energy
Nominal Measured (MPa) (kN) (mm) (kN) (mm) (J) (Mean)
(kg/m3) (kg/m3) (J)
Panel 11 80 1.2 165 11.0 3944
Panel 12 50 43.0 52.7 90 1.5 160 11.6 5942 4710
Panel 13 80 1.2 162 8.6 4252
Panel 21 80 0.8 204 11.4 4545
Panel 22 80 74.1 40.9 75 0.7 219 11.6 5235 4590
Panel 23 80 0.9 187 13.5 3978
Panel 31 - - 139 9.3 1350
Panel 32 Mesh Mesh 49.0 50 0.6 143 18.0 3353 2340
Panel 33 45 0.5 138 16.6 2333
Panel 41 75 1.0 217 13.6 6144
Panel 42 80 65.4 47.2 20 - 191 12.1 4300 4210
Panel 43 70 1.1 190 10.4 2187
Panel 51 60 0.8 153 9.5 4131
Panel 52 50 3 7. 6 46.6 - - 156 10.1 4026 4190
Panel 53 40 0.3 169 9.4 4413
 79

Figure 4-18 Punching failure of a panel

4.5.6 Bearing capacity of the panel capacity determined by the yield line
theory

Panels made of ordinary reinforced concrete are often designed by the use of the yield line
theory because of its simplicity and a more effective use of the material compared with elastic
theory. The concept of this type of design is that the material is almost ideal elasto-plastic as
shown in Figure 4-19.

Figure 4-19 Elasto-plastic stress-strain relationship

The test results from the beam tests show (Figure 4-13), that the fibre reinforced beams do not
have an ideal elasto-plastic behaviour. In spite of the missing ideal elasto-plastic behaviour a
yield line analysis has been performed to compare the bearing capacity of the panel according
to the yield line theory with the obtained test results.

The yield line theory calculation was based on the crack pattern or yield lines as shown in
Figure 4-20. This pattern corresponds well with the crack pattern for the tested panels. For
most of the panels the apex angle e was roughly equal to 600, (the apex angle in Figure 4-20
is equal to 30°). Figure 4-20 also shows the direction of the yield line moment or ultimate
moment mr and mt caused by the yield load or ultimate load F.
 80

The moment mr per unit length occurs along the radial yield lines and the moment mt (with
opposite sign) per unit length occurs along the tangential yield lines. Also, around the circular
loading plate at the centre of the slab a set of tangential yield lines are assumed to occur.

Yield lines

L/2

Or

.)(
Section A
Figure 4-20 Yield line model

The bearing capacity of the panel according to yield line theory is derived in Appendix 4.2.
The calculation is based on the yield line mesh shown in Figure 4-20 with e = 60°. The
bearing capacity is

t n
m, (0,.)(1— k)+ m,(19r )(1+ k) a 2
= 2K (4.6)
1— k
2

where mr(a) and mr(e) is the radial and tangential moment per unit length. The yield
moments of the panels mr(Or) and mt(e) are determined by the post crack moment capacity of
the test beams m(0) as a function of angular deformation 0. For a given angular deformation
 81

was mre=mte=me, the magnitude of me is derived in Appendix 4.2. The angles, 0, and
Or are defined in Figure 4-20 and Figure 4-21.

The constant k in Eq.(4.6) is

Diameter of load bearing plate

k=
Span of panel

with the diameter of the load-bearing plate equal to 200 mm and the span of the panel equal to
1500 mm, the value of k is equal to 200/1500 = 0.13.

The relation between the angles 8 and Or is

= 20 sin cj (4.7)
2

Equation (4.7) is derived in Appendix 4.2.

Section A-A

Figure 4-21 Angular deformation Or

Figure 4-22 and Figure 4-23 show the test results and the theoretical bearing capacity of the
panels with yield line theory. It is obvious that the yield line theory based on the moment
capacity of the test beams does not predict the bearing capacity of the test panels, the
calculated bearing capacity is too low compared with the test results. The main reason is
probably the dome effect, which has been derived and discussed by for example Birke (1975).
Dome action can arise when the edges of a panel are resisted to horizontal deformations, the
panel will carry the load by compressive forces as shown in Figure 4-24

Nilsson (2000) tested the bearing capacity of circular concrete panels loaded by a point load
in the middle of the panel. The boundaries of the panels were clamped. He reported a bearing
capacity, which was much higher than the theoretical capacity according to the yield line
theory. He concluded that the dome action could be the main carrying component of the
panels.
 82

a) b)

Steel Fibre Dramix 35/0.55, nominal dosage 50 Steel Fibre Dramix 35/0.55, nominal dosage 80
kg/m', measured dosage 43.0 kg/m' kg/m3, measured dosage 74.1 kg/m'
250
250

200 - —Slab11 200 -

— Slab 21
150 - - - Slab12 150 - - - Slab 22
8 --- Slab13 100 --- Slab 23
u_
50 - ; —Yield Line Theory 50 —Yield line theory

0
O 10 20 30 40 50 0 10 20 30 40 50
Mid-point deflection (mm) Mid-point deflection (mm)

c)
Steel mesh

250 -

200 -
—Slab 31
-z 150 - - - Slab 32
8 100 - --- Slab 33
— Yield line theory
50 -

0
0 10 20 30 40 50
Mid-point deflection (mm)

Figure 4-22 Comparison between test results and numerical solutions according to yield line
theory based on test beams
a) Dramix 35/0.55, 43.0 kg/m3
b) Dramix 35/0.55, 74.1 kg/m3
c) Welded steel mesh
a) b)

Steel Fibre Dramix 50/0.62, nominal dosage 50

Steel Fibre Dramix 50/0.62, nominal dosage 80
kg/m', measured dosage 37.6 kg/m3
kg/m', measured dosage 65.4 kg/m'
250 1 250 -
1
200 H' 200 -I —Slab 51
2- —Slab 41 S -•-• Slab 52
- -. 150 -
- - Slab 42 --- Slab
Slab 53
°
•0
9 100 8 100-1
-
n
--- Slab 43 LT_ —Yield line theory
_. ---___,„_-_____
—Yield line theory 50
50
0
0
0 10 20 30 40 50
O 10 20 30 40 50
Mid-point deflection (mm) Mid-point deflection (mm)

Figure 4-23 Comparison between test results and numerical solutions according to yield line
theory based on test beams
a) Dramix 50/0.62, 65.4 kg/m3
b) Dramix 50/0.62, 37.6 kg/m3
 83

H H

Figure 4-24 Dome action, after Birke (1975)

4.6 Discussion

4.6.1 Field and laboratory tests in the Kiirunayaara mine

The results from the failure mapping shows that more than 80% of the damage was small
fallouts of plain shotcrete in areas with low or no adhesion strength. Important factors, which
may cause the failure, are rock deformation, shrinkage and dynamic loads due to blasting.

It is also important to get sufficient thickness all over the shotcreted area i.e. avoid areas with
thin shotcrete, (2 cm or less). The shotcrete thickness was 2 cm or thinner for more than 90%
of the fallouts of plain shotcrete. Furthermore, 60 % of the failures were in conjunction with
an apex on the rock surface together with a thin shotcrete layer. The results showed that in
18% of the thickness measurements the thickness of the shotcrete was less than or equal to 2
cm. If the shotcrete lining with ordered thickness of 3 — 5 cm had the same distribution of
shotcrete thickness as the tests, the supporting ability of as much as 18% of the shotcrete
lining can be considered as more or less negligible.

These examples show how important it is to get the right thickness and how important the
skill of the nozzle man is.

The growth of adhesion strength of shotcrete on a sandblasted concrete wall and the growth of
compressive strength of shotcrete show a clear correlation. The author has not been able to
compare the result with others because the author has not found any similar tests in the
literature. The adhesion strength was also relatively high (fa (28_ day) = 0.6 to 2.0 MPa)
compared to field tests in the Kiirunavaara mine.

A way to increase the adhesion strength is to use water-jet scaling as a treatment method. The
water jet-scaling machine used in these tests had a water pressure equal to 22 MPa. The result
showed that the adhesion strength was significantly higher on surfaces that had been water-jet
scaled than the ones that had not.
 84

4.6.2 Panel and beam tests

In many tests performed to investigate the behaviour of fibre reinforced shotcrete, the
concrete is poured instead of sprayed. The main disadvantage of spraying instead of pouring
the specimens is the deviation of the amount of steel fibres compared to the designed amount
of fibres. Also, differences in rebound of various types of fibres can make it difficult to
compare the bearing capacity for shotcrete reinforced with various fibre types. This was
apparent in these tests. The advantages of using shotcrete instead of poured concrete are that
the effects of accelerators, fibre directions and compaction are included in the test results.

The test results from the panel tests showed that the typical failure mechanism was a bending
failure followed by punching failure. Furthermore, the tests showed that panels reinforced
with steel fibres were more ductile and had higher punching capacity than panels reinforced
with welded steel mesh.

The yield line theory based on the rotating capacity in the yield line was not capable of
predicting the load bearing capacity. The actual load bearing capacity of the panels was much
higher than the calculated bearing capacity according to yield line theory. This can probably
be explained by the fact that dome action is the main load-carrying mechanism. The dome
effect occurs only if the edges are prevented from moving in the plane of the panel. This
constraint can be assumed to be present for a lining, because of the adhesion strength and
irregularities of the rock surface. The dome effect can, however, be low when the shape of the
lining is directed downwards as shown in Figure 4-25 a).

a) b) Rock bolt

Reinforced shotcrete

Figure 4-25 Dome action (Nilsson, 2000)

a) Low dome action when the shape of the lining is directed downwards
b) High dome action when the shape of the lining is directed upwards or is
fairly flat

To achieve interaction between bolts and shotcrete the washers of the bolt must be placed on
the surface of the shotcrete, see Figure 4-26, and not as in many cases where the bolt was
installed before the shotcrete which gave no interaction between bolt and shotcrete. The
possibility of having the shotcrete lining acting as a retaining element in this case is small or
non-existent, depending on the adhesion strength. Finally, the results from the beam tests and
the results from the panel tests can not be substitutes for each other because of the different
loading modes.
 85

a) b)

Rock bolt with washer

Figure 4-26 Placing of the washers of the bolt

a) No interaction between bolt and shotcrete
b) Possibility for interaction between bolt and shotcrete
 86

5 VIBRATION MEASUREMENTS IN THE KIIRUNAVAARA MINE

WITH BELONGING FAILURE MAPPING

5.1 General

Rock support is mostly designed for static loads. In many mines, however, the openings are
subjected to dynamic loads caused by blasting. To be able to design rock support that can
withstand dynamic loads from production blasting, vibration measurements have been done.
The vibration measurements are used as input for the analysis presented in Section 6. In order
to investigate the influence of the dynamic loading induced by the production blasting on the
stability of the roof, failure mapping close to the drawpoint has been carried out. This Section
presents the measurements of vibrations caused by blasting and the failure mapping of the
brows.

During the period 1998-1999 an extensive field measurement programme was carried out to
estimate to what extent the rock support in the openings was subjected to destructive dynamic
load from detonating explosives. The vibration measurements were done in co-operation
between LKAB and the Swedish Rock Engineering Research (SveBeFo).

The measurements comprised acceleration measurements in the cross cut roofs. All together
12 tests were carried out in the two cross cuts No 402 and 145 at level 740 m as shown in
Figure 5-1. A total of 76 recordings from 12 fans have been used in the analysis.
 87

a)

Y2000 Y3000 Y4000

X650
.A
.111...........11111111......111MIS nil ,. .
ZalaZZIIIIMIIIMBIBUZZZIEZIIIIMMT5,:i-Zeitäledif
--- II ,
- eidgrgew i
BINIIIEZMil IIIIMMIZZIMIIIN1111111!_ n
gleiffill1111ffignimw="2-znioniF7ponerie,Adiumargik--
X600 ü
- jegMilaiiii..ehIZArgelliiiiiidielM111Z/M
aiiiiiIIIKeZIIIWEETAtiallIZE1111311111211221
IZZlialMimmailiZZE33111111111113111M3B/EZ
111111111111113ZZEMBEIZ 1000 m all
IMBIZZMIllaZZI313 MR
X550 3311Balalnallala IIIIB
ZE1111111111131MIMIll a
IZZI11111111111Z11111ManaMM ZILIZZI
zzaizzazaretinunnozzaz Znaall
lIZEIZZUZIIRiallialn311131•11113 3/1311111

b) c)

2.4% 142
21427.88
145
,11,6.112
147
21477.38
152
11828.88
402
24026.82
405 .407
Y.6211.4
410".8{.
y4

,
14051.36
410.1
211441-30
i
.

142.1
28160.110 147.1
16117r.00
Jk I i

r[

Figure 5-1 Test area for the vibration measurements

a) Level 740 m
b) Cross cut 145
c) Cross cut 402
 88

5.2 Production blasting

Figure 5-2 shows the fans for cross cuts 145 and 402. The 4.5" (115 mm) blasting holes are
drilled upwards in a fan. All holes in a fan are within a plane, which strikes perpendicular to
the cross cut and with a dip of 800, see Figure 5-3. The distance between each fan, that is, the
burden, is 3 m. The blasting holes were charged with Kimulux R, a repumpable bulk emulsion
explosive. The length of the holes varies from 15 m for the side-holes to 40 m for holes in the
middle of the fan. The bulk emulsion is water-resistant to some extent, but if there is heavy
water flow through the hole, cartridges are used instead.
a) b)

740-402- 4 740-145 • 8
Primers with the same
45= 45- number are fired in the
same interval
40-

35 — 35-
_

307 30

25 25.1

20 20-1

15= 15

10: Charged 10=

5 Uncharged 5

11111111 I I II 11111 I I I I I I M
m
5 10 15 20 25 30 35 5 10 15 20 25 30 35

Figure 5-2 Production blasting, fans for sub level caving in the Kiirunavaara mine
a) Cross cut 402, elevation 740 (with electronic detonators)
b) Cross cut 145, elevation 740 (with pyrotechnic ms-detonators)

Charging of the holes is performed by a purpose-built truck. A chain-feeder, mounted on an

articulated boom, feeds the hose into the hole. After the hose has been fed to the top, the
automatic charging is started.
To prevent interaction between detonating holes and to keep the specific charge within limits,
a certain length is left uncharged at the end of each hole. When the hole is charged down to
this length, the operator stops the charging manually.
To prevent misfire most of the holes are initiated with two independent primers, see Figure
5-2. These sets are placed in the bottom of the explosive column and approximately 20 m
above the bottom of the explosive column respectively.
 89

Next to each blasted fan there is at least one fan fully prepared for blasting. Both explosive
agent and the ignition system are in place to avoid later problems of getting access to the
holes due to fragmented rock from the previously blasted fan. In the blast holes the explosives
have a linear density along the hole of 9 to 12.5 kWm. The specific charge for a full fan is
approximately 0.3 kg/ton.

Firing of the blasts is done electrically from central firing stations, where a mine-wide system
of fixed firing cables meet. The normal firing pattern is two holes at each interval, with a 25
ms delay to the next pair of holes. The firing starts with the two longest holes in the middle of
the fan and the duration of the whole initiation sequence is about 125 ms. Ina case of an odd
number of blast holes the firing starts at the hole in the middle, as for cross cut 402 (Figure
5-2 a).

5.3 Explosives used in the production blasting

The technical specifications for explosives used in the two cross cuts No 145 and 402 at level
740 m are shown in Table 5-1.

Table 5-1 Explosives used in production blasting

Cross cut No of No of Kimulux Density VOD a) Volume

Tests Charges (mis) Strength
(kg/In3)

145 6 10 R0000 1200 5500 115 % of

ANFO
402 6 11 R0500 850 5000 99 % of
ANFO
a) VOD = Velocity of detonation

Pyrotechnic and electronic detonators controlled the initiation sequence for the explosives in
Table 5-1. When electronic detonators were used, the time sequence was simulated to be
approximately the same as a pyrotechnic one. An exception was a case when each charge was
initiated individually to achieve a complete time separation of the waves at specified points.

5.4 Acceleration recordings in the cross cut

Recording of vibration signals in connection to blasting is often a critical task; the transducers
must be strongly fixed to the rock.

In this investigation, vibrations were recorded at four stations grouted to the rock at the top of
approximately 1 m long holes, drilled from the roof, see Figure 5-3. The holes were parallel to
and .k 4.5 m away from the blasted fan. In each hole, two accelerometers were assembled,
with the measurement directions in agreement with Figure 5-4. The holes were 3 m apart
(burden distance).
 90

The recording unit was a DAT — recorder (Digital Audio Tape recorder) with 16 channels, DC
- 20 kHz and 16 bits recording. The signals were transmitted from the stations via coaxial
cables (RG 58) and recorded at a sampling rate of 48 kS/second with time resolution At =
2.08310-5 s. Two types of transducers, with the specifications according to Table 5-2, were
used.

Table 5-2 Specifications for the accelerators

Transducer Voltage sensitivity Range Resonant Maximal amplitude
(mV/g) frequency deviation
(kHz)
(g)
PCB350 A03 0.5 ±10000 >100 -I- 3 dB 100 Hz -26 kHz
Kistler 8704 B5000 1.0 -± 5000 > 60 1- 1 dB 3 Hz —20 kHz

Accelerometers, see Figure 5-4 Fan, see Figure 5-2

Production
blasting

Accelerometer, see
Figure 5-4.

Caved rock

Bore hole for Section A Cross cut

accelerometers

Figure 5-3 Arrangement of the accelerometers in the cross cut

 91

Measurement
directions

Anchor of aluminium that the

Groutin
accelerators were mounted on.

Plastic pipe
\ Two accelerometers

Bore hole

DAT — recorder (Digital Audio

Tape recorder)

Figure 5-4 Accelerometers in the bottom of the bore hole

5.5 Results

As mentioned above two blast holes are initiated at the same time interval by using
pyrotechnic or electronic detonators. The relatively low time precision for the pyrotechnic
caps indicate that waves from blasting holes (in the same interval) randomly must overlap
each other. With respect to this, it was only possible to estimate the maximum quantity of
explosives for each interval that gave rise to the measured waveform (load). This waveform
was a superposition of waves from two individual blast holes.

The total quantity of explosives (Table 5-3 and Table 5-4) for each fan has been shown to
vary up to approximately 30 %. This deviation between planned and measured quantities of
explosives may result in a scatter in wave amplitude. For the actual fans the difference was
negative for all fans, that is, there were less explosives charged than planned.
 92

Table 5-3 The quantity of explosives, cross cut 145

Fan Type of Charged Q Theoretical Q Not possible to Difference

Explosives (kg) (kg) load a) (kg)
(kg)
17 KR0000 1760 2624 - -864
18 KR0000 2198 2628 240 -190
19 KR0500 2082 2621 13 -526
20 Mix 1704 2628 125 -799
21 KR0000 2112 2623 - -511
22 KR0000 2505 2618 - -113
a) The blast holes could not, for some reasons, be charged with explosives.

Table 5-4 The quantity of explosives, cross cut 402

Fan Type of Charged Q Theoretical Q Not possible to Difference

Explosives (Kg) (Kg) load a) (Kg)
(Kg)
20 KR0000 2700 2803 - -103
24 KR0500 2179 2804 293 -332
25 KR0500 2660 2799 - -139
26 KR0500 2338 2864 - -526
28 KR0500 2116 2810 649 -45
29 KR0500 2430 2804 - -374
a) The blast holes could not, for some reasons, be loaded with explosives.

The unloaded blast hole lengths are shown in Figure 5-5. The actual unloaded lengths were
generally shorter than planned. A few measurements have shown that no unloaded length was
less than 2 m. (In the Kiirunavaara mine the unloaded hole length in the production blasting
must be longer than 2 m (Larsson, 1998)). These measurements indicate that the charged
quantity of explosives should be larger, not smaller than Table 5-3 and Table 5-4 show.
 93

a)
Crosscut

20 - - 20
18 18
16 - - 16
Un
loa 14 - - 14
Measured- fan
de 12 - 12
F7-7-
A Measured- fan
d 10 - - 10
hol 8 - I I Measured- fan
-8
Planned
e 6- -6
len
gth 4 -4
2 .0 2
0 0
1 2 3 4 5 6 7 8 9 10
Blast

b)

Crosscut 402/740

20 71 20.0
,,
71 20.0
, 20
18 - 18
16 - - 16
eto 14 - 14 I J Measured - fan 20
12.0 12 =%4'• Measured - fan 21
ci 10 10 1 1Measured - fan 24
8 1
8 I I Measured - fan 26
.T2:3 /
6 --U Planned
4- 4
2 02
0
1 2 3 4 5 6 7 8 9 10 11
Blast hole

Figure 5-5 Planned and measured unloaded hole length

a) Cross cut 145, elevation 775
b) Cross cut 402, elevation 775

Two acceleration signals were selected for further analysis in Section 6, (direction parallel to
the hole), see Figure 5-4. They represent a more or less typical acceleration signal and have a
high magnitude. Because the quantity of explosives used was less than planned, it was
important to find signals with high wave amplitudes. Table 5-5 summarizes the main data for
the two time histories and Figure 5-6 shows the particle acceleration.
 94

Table 5-5 Load cases

Load Location Distance from the Maximum and Maximum and
case production blasting fan minimum particle minimum particle
acceleration velocity
(m)
(m/s2) (mm/s)
1 Cross cut 145, 4.5 3086/-2600 1146/-1061
ring No.18

2 Cross cut 402, 7.5 1663/-978 540/-405

ring No.26

a) b)

Load case 1
•o 4000 - 4000 Load case 2
0
3000 •-z, 3000--
• 2000 E
5 2000
• 1000 O co 1000
E 0
O ---
co E 0
•
7.5 -1000 0 50 200 250 7:5 -1000 0 50 100 150 200 250
as -2000 - • -2000 -;
a_ -3000 - -3000
Time (ms) Time (ms)

Figure 5-6 Time history with respect to acceleration

a) Cross cut 145, ring No. 18
b) Cross cut 402, ring No. 26
 95

From the two chosen time histories two load cases were extracted, load case 1 and 2, as
subsets from each time history. Figure 5-6 shows from what part of the time histories the
subsets are defined. The load cases are shown in Figure 5-7 and Figure 5-8 with respect to
particle acceleration and particle velocity. As one can see in Figure 5-9 the two subsets (load
cases) represent the chosen measured histories well with respect to the curve shape of the
acceleration. Furthermore, Figure 5-9 shows that the highest particle acceleration of the two
load cases is below 5000 Hz.

3000 —
.7, 2500 —
E 2000 — 1500 —
—

Particle velocity ( mm/s)

o 1500 — 1000 —
1000 —
500 500 --
To
0 0 0
-500i 10 20 30 30
-500
-1000 1
a. -1500 — -1000 —
-2000 -1 -1500 —
Time (ms) Time (nis)

Figure 5-7 Load case 1

a) Particle acceleration
b) Particle velocity

3000
CZ
)) 2500
sü- 2000 1500
Particle velocity (mm/s)

c 1500 1000
o
1000
500
Ioj 500
Q 0 0
a) -500
-500
-1000
co -1000
a_ -1500
-2000 -1500
Time (ms) Time (ms)
Figure 5-8 Load case 2
a) Particle acceleration
b) Particle velocity
 96

a) b)

- Measured history — Measured history

1.0 — Subset - Load case 1 1.0 —Subset - Load case 2
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0.0 0.0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Frequency (kHz) Frequency (kHz)

Figure 5-9 Normalised particle acceleration as a function of frequency

a) Cross cut 145, ring No. 18
b) Cross cut 402, ring No. 26

5.6 Failure mapping of the drawpoint

During 1999 failure mapping of the roof in connection to the drawpoint of 34 cross cuts at
740 m level was carried out, failures from 0 to 15 m away from the drawpoint were mapped.
The objective was to map typical failures in connection to the production blasting and to get
input for a verification of the model for dynamic response analysis performed in Section 6.
Table 5-6 and Table 5-7 summarise some of the results. More than 70 % of the roofs had
failures, 84 % of the failures were fallouts of wedges. Approximately 70 % of the failures
were small to medium fallouts <1 m3. The supported roofs, (76 % of the mapped roofs), had
more failures than the unsupported roofs.

Table 5-6 Failures in mapped roofs of cross cuts in connection to the drawpoint

No of Roof "without" Small to medium Larger Sum Type of failure

mapped roof failures failures <1 m3 failures failures Fallouts Other
34 9 17 8 25 21 4
- - 68% 32% 100% 84% 16%

Table 5-7 Failures in mapped roofs of cross cuts in connection to the drawpoint, failures
related to rock support.
Type of support Mapped roofs Failures in mapped roofs
No support 8 24% 4 16%
Rock bolt 9 26% 5 20%
Shotcrete 8 24% 7 28%
Rock bolt + shotcrete 9 26% 9 36%
SUM 34 100% 25 100%
 97

As shown in Figure 5-10 the plain shotcrete does not work properly in connection to areas
subjected to dynamic loads. However, if the shotcrete is reinforced the failure zone decreases.

7 Drawpoint

Reinforcement No of Failure
A
obsevations zone Lf
(m)
Plain shotcrete 13 3—9

Failed shotcrete
Intact shotcrete Welded mesh 4 2- 4

>
Steel fibre 4 3
Dramix 30 mm
(50 kg/m3)

Lf

Section A-A Roof of cross cut

close
to the drawpoint

Figure 5-10 Failure zone of shotcrete in connection to the drawpoint

5.7 Discussion

The vibration measurements show a wide scatter of magnitudes of the particle acceleration,
which was expected. More surprising was the relatively large variation of the charged
quantities of explosives compared with the planned or theoretical quantity of explosives.
Furthermore, there was a large variation of unloaded hole lengths compared to the planned
lengths.

The failure mapping shows that the production blasting can cause stability problems, wedges
or blocks can be ejected from the roof of the cross cut in connection to the drawpoint.
Furthermore, the failure mapping in connection to the drawpoint showed that plain shotcrete
did not work properly as support, when subjected to dynamic loads. Plain shotcrete has not
the necessary ductility to prevent wedges from falling down. But when the shotcrete was
reinforced the area of damaged shotcrete and rock falls decreased and one should remember
 98

that the rock quality was worse for roofs supported with reinforced shotcrete than roofs
supported with plain shotcrete
 99

6 THE RESPONSE OF SHOTCRETE DUE TO VIBRATIONS CAUSED BY

BLASTING

6.1 Introduction
The rock support is mostly designed for static loads. In many mines, however, the openings are
subjected to dynamic loads. For instance, rock bursts in deep mines can give serious damage to
underground structures. Another type of source of dynamic loads is detonation of explosives. To be
able to mine the right ore quality and to get a proper mass flow from the sub-level caving it is
important to keep the profile of the cross cut intact. Close to the drawpoint, the dynamic influence
from the production blasting is significant, see Figure 6-1. In this Section the influence from
blasting induced stress waves on the performance of shotcrete support is addressed.

Production blasting
Drawpoint

Stress wave from blasting

Section A

Figure 6-1 Cross cut subjected to blasting induced loading

Detonation of explosives generates stress waves in the rock mass. The stress waves propagate
outwards from the blast hole in all directions. Owing to geometrical and material damping, a stress
wave rapidly attenuates with increasing distance. The stress wave can cause shaking, which results
in relative slip on joints and ejection of joint-defined blocks as shown in Figure 6-2. The reflected
tensile pulses at the free surface of the rocks, can cause some part of the material close to the
surface to be "ripped off', this is called spalling.

In this study, the ejection of joint-defined blocks will be analysed, because many of the failures
close to the brow of the cross cuts seem to be structurally controlled, which is discussed in Section
5.6.
 100

Stress wave from blasting

Figure 6-2 Ejection of a prism

To study the dynamic response of the rock mass close to the brow, a model that consists of a 2-D
roof prism has been developed, see Figure 6-3. A single degree of freedom (SDOF) model was
established to calculate the dynamic response to a stress wave. The stress wave time history was
obtained from measurements during production blasts in the Kiirunavaara mine, see Section 5. The
vibration measurements used in this analysis were done mainly at a horizontal distance of 4.5 to 7.5
m from the blasted fan, see Figure 5-3. The stiffness of the joints and the stiffness of the support are
represented by springs in a SDOF model and the damping is represented by a dashpot.

The influence of the gas pressure is omitted in this analysis. Ouchterlony et al. (1997) and Brent &
Smith (1996) have found that measurements during bench blasting, which is similar to the
production blasting in the sublevel caving, usually gives underpressure and only in a few cases a
small overpressure in sealed boreholes behind the blast. This implies that the blast fumes do not
penetrate very far into a rock mass with tight fractures.

6.2 The dynamic behaviour of a symmetric roof prism caused by blasting

The dynamic behaviour of the symmetric roof prism is analysed using a single degree of freedom
(SDOF) system shown in Figure 6-3. The analyses were performed as transient dynamic analyses.

Figure 6-3 A diagrammatic representation of a SDOF model

 101

The load induced by the stress wave generated by blasting, was measured as particle acceleration.
The particle acceleration time history is used as the loading in the model. The effective forces
which excite the system result from the fact that the inertia force term depends on the total motion,
while the damping and elastic forces depend only on the relative motion. The equation of motion
can be written as

Mü + + K(u)u =- 0 (6.1)

where M is the mass, C is the damping coefficient and K(u) is the stiffness. The total displacement
ut and the relative displacement u is related to the ground motion ug as (Clough & Penzien, 1993)

U, = U g U . (6.2)

Using (6.2), the equation of motion (6.1) can be written in terms of relative displacements

Mü + Cz + K(u)u = —Mü g . (6.3)

The stiffness, K(u), in the SDOF system consists of two springs, where one represents the vertical
stiffness of the joints (Kj) with the bearing capacity Fi f , see Section 6.3. The second spring
represents the stiffness of the support system, K, with the yield load capacity Fsy, which is discussed
in Section 6.4. The springs act in parallel, see Figure 6-4. The total spring force is then

= F + Fs (6.4a)

where F., is the spring force of the joints and Fs is the spring force of the support. The damping
coefficient C in (6.3), can be expressed as

C=C (6.4b)

where is the damping ratio, C = 2Mco and co is the natural frequency. The material damping
ratio, , is according to SBN — S 21:64 (1970), 0.01 to 0.03 for concrete structures. In the further
analysis the damping ratio is assumed to be equal to 0.01.

Figure 6-4 The stiffness and damping of the SDOF system

 102

Equation (6.3) is solved using a linear acceleration step-by-step formulation. This method is the
well-known Newmark beta method, where fi=1/6. The method has been described by many authors,
for example Langen & Sigbjörnsson (1979) and Clough & Penzien (1993). The method is presented
in Appendix 6.1. Figure 6-5 shows a simplified flow chart for the calculation.

Input:
Dynamic load, Section 6.5
Mechanical properties of the support
Section 6.4

Fif, Kt Solving the equation of motion

and according to Section 6.2
Uff

Input:
Ground condition,
Section 6.3
Calculation of bearing
capacity Fir, stiffness
and displacement at failure Output:
z
of the joints, uif , Section 6.3 Spring force, Fspring(t) and
damping force Fd„„(t)
Displacement, u(t)
Velocity, U(t)
Acceleration ü(t)
Figure 6-5 Flow chart

6.3 The behaviour of the roof prism

The prism in Figure 6-6 represents the cross section of a long, uniform, triangular prism formed in
the crown of an excavation by symmetrically inclined joints. The behaviour of a roof prism is
dependent on the mechanical characteristics of the joints forming the prism, the shape of the prism,
the deformability of the rock, rock stresses, water pressure and the density of the prism.

The bearing capacity and the stiffness of the joints are solved by a two-stage relaxation procedure
shown in Figure 6-6. In the first stage, when the force 1/0 is determined, the joints are assumed to be
rigid. In the second stage, where the wedge is loaded with a vertical force F1, the joints are assumed
to be deformable and the rock mass is assumed to be rigid. The calculation is performed for a prism
with the thickness equal to 1 (m).
 103

b)

Figure 6-6 Loading stages;

a) Stage 1: rigid joints b) Stage 2: rigid rock

The normal and shear forces No and So, are related to the horizontal force Ho by the relation

N, =H0 cosa
(6.5)
So =H0 sin a

where ais the semi-apical angle and Ho= hue , where og is the tangential stress in the roof of the
cross cut.

When the vertical load Fj is applied, the prism is displaced vertically a distance u. The vertical
displacement u can be decomposed into the normal and shear components u„ and us respectively,

u = usina
(6.6)
= u cos a

which is illustrated in Figure 6-7.

When the prism moves away from the surrounding rock, the normal and shear forces are changed
with the increments AN and AS respectively. This increments are defined as

AN = = Ki„u sin a
(6.7)
AS Kisus = Kivu cos a

where xr,„ and Ic„ are the normal and the shear stiffnesses of the joints respectively, and defined as

b
K.= 2 k
-1 sin a "
(6.8)
b
K ft= • 2 ks
sin a
 104

where k„ and k, are the normal and the shear stiffnesses (per unit area) of the joints respectively, and
defined in Section 3.5.

Figure 6-7 Displacements and forces

The vertical load F., in terms of incremental forces, is defined as (see Figure 6-7)

= 2(AN sin a + AS cos a) (6.9)

and the vertical stiffness of the joints is defined as

F
K = . (6.10)
U

By using (6.7), (6.9) and (6.10)K, can be expressed as a function of the normal and shear stiffnesses
of the joint

1 = 2(K sin2 a + K is cos2 a) (6.11)

The vertical bearing capacity for a 2-D prism under static loading conditions and pure friction
S = N tanØ is according to Sofianos et.al (1999)

F = 2H° (K cos2 a + KJ. sin2 a)sin(0 — a) (6.12)

If D

where D -= K j, cos a cos + K j„ sin asin .

If Kjn, which is the usual condition in practice, (6.12) becomes (Brady & Brown, 1999):
 105

sin a sin(Ø —a)

Fif = 2H, (6.13)
sin 0

If Ho> 0 the condition Fi f> 0 can be satisfied only if Ø> a.

Figure 6-8 and Figure 6-9 show the variation in Fi1/2H0 as a function of the semi-apical angle, o4
the ratio kniks and the friction angle Ø. Figure 6-8 shows that the bearing capacity according to
Eq.(6.12) with k/k = 100 is almost equal to the bearing capacity according to Eq.(6.13) when
a . Figure 6-9 shows that for the chosen friction angles the bearing capacity is not sensitive
for a 20°.

0.6
0.4
0.2
0.0 04(°)
-0.2 10 20 30 60 70 90 Eq. (6.12) and kniks = 0.01
Eq. (6.12) and k„iks = 1
-0.4
(N1 —a— Eq. (6.12) and k,/k4 = 10
-0.6 Eq. (6.12) and k,/k, = 100
-0.8 0= 30° Eq.(6.13)
-1.0
-12
-1.4
-1.6

Figure 6-8 Bearing capacity of the roof prism (FAH0) as a function of the semi-apical angle, a
and the ratio k,/k.

0.4
0.2
0.0
10 20 3041% 4 70 80 90 a(o)
-0.2
-0.4 Eq.(6.12)

-Oß
__._ Ø=30°
IL -0.8 kniks= 100 0= 35°
0= 40°
-1.0
-1.2
-1.4
-1.6

Figure 6-9 Bearing capacity of the roof prism (F11/2H0 ) as a function of the semi-apical angle, a
and the friction angle Ø.
 106

Mechanical properties used in the analysis

The background for the selected properties is found in Section 3.5. Figure 6-10 shows the analysed
prism. The rock mechanical properties used to calculate kn, lcs and Fir- are summarised in Table 6-1
and Table 6-2 respectively.

Dip angle

—› (le

Friction angle, 0

Width (b) = 4500 kg/m3 (density)

Figure 6-10 The analysed roof prism

Table 6-1 Mechanical properties used to calculate kn and k,

Property Value
Joint Roughness Coefficient, JRC 6
Joint Wall Compression Strength. JCS (MPa) 100
Uni-axial compressive strength, a, (MPa) 145
Residual friction angle, 0,. (o) 28

Table 6-2 Dip angle, semi-apical angle a friction angle 0 and width b used to calculate Fif and
the mass of the prism, M

Dip a--= 90° — dip Friction angle, 0 Width, b Mass of the prism
(0) (0) (0) (m) m . b2 p
(kg)
4 tan a
31 0.4 312
60 30 0.8 1247
35 0.4 312
0.8 1247
31 0.4 495
70 20 0.8 1978
35 0.4 495
0.8 1978
 107

Simpeed calculation of the vertical joint stiffness

The data in Table 6-1 used to calculate kn and ks, according to Section 3.5, gives the ratio kiks >
190 for all ground conditions. This means that the effect of ics on 1<", is almost negligible, see
Equation (6.11). Therefore, Kj is calculated assuming k,= 0. Equations (6.8) and (6.11) and kn
according to Section 3.5 give

1
K b sin a • k„, (6.14)
2
(1 n )
„, +O

where, Vm and km are material parameters defined in Section 3.5, o, is the normal stress at the joint
and related to the tangential stress in the roof, a, , by

o- =o-o cosa (6.15)

Table 6-3 shows the stiffness Kj for various geomechanical conditions, calculated according to
Equation (6.14).

Table 6-3 Vertical stiffness of joints (Kj) for various geomechanical conditions
Semi-apical Friction Width, b Tangential stress, o-,
angle, a angle, (/) (m) (MPa)
1 10 20 30
Vertical joint stiffness, ic,
(GN/m)
30 31 0.4 7.1 154 536 1151
0.8 14.2 306 1071 2300
35 0.4 7.1 154 536 1151
0.8 14.2 306 1071 2300
20 31 0.4 5.2 121 428 921
0.8 10.4 242 854 1840
35 0.4 5.2 121 428 921
0.8 10.4 242 854 1840
 108

Failure criterion

The failure criterion for the prism expressed in forces is

> Fil (Unstable) (6.16 a)

FJ ~ FJ (Stable) (6.16 b)

For practical reasons when solving the equation of motion the failure criterion is expressed in terms
of displacements
U > uff (Unstable) (6.17 a)

u < u if (Stable) (6.17b)

where

F11
Ui f = •
(6.18)
K1

A principal load-displacement curve visualising (6.16) and (6.17) is shown in Figure 6-11. When
the joints of the prism are compressed, the behaviour of the joints is assumed to be linear elastic.

Spring force, Fi

Ulf
Displacement, u

Figure 6-11 Spring force (F) - displacement (u) characteristics for the prism

6.4 The behaviour of the shotcrete lining

Failures, which have been observed close to the drawpoint, are often structurally controlled. Tests,
which simulate this type of failure, are discussed in Section 2.7. Test results from Section 2.7
(Holmgren, 1985) are used to describe the behaviour of reinforced shotcrete. The simplified load-
displacement curve used in the analysis is shown in Figure 6-12. The yield load capacity F is
assumed to be constant until the maximum (failure) displacement usf occurs. Necessary input for the
analysis is shown in Table 6-4. Holmgren's results are obtained in static tests. The dynamic
 109

stiffness and bearing capacity can be higher compared to the static properties (Grasser et.al. 1994).
However, in this investigation the static values are used.

In addition, other types of rock support such as mesh or membrane supports as well as rock bolts
can be analysed.

Fj2 (liN)
Reinforcement
0.5 — 0.6 % Dramix ZL 45x0.35 (39 —47 kg/m3)
--- Welded steel mesh tß5c100f, =510 —560 N/mm2
_._._Welded mild steel mesh 05c1004 =220 N/rrim2

Load-displacement curve used in the

analysis

Figure 6-12 The force-displacement curve for the support, test results from Holmgren (1985)
(Not in scale)

Table 6-4 Mechanical properties for the shotcrete used in the analysis

Property Value
Yield load capacity, Fs), (kN) 60
Stiffness of the shotcrete, K, (MN/m) 60
usy= F„/ K, (mm) 1
Failure displacement, tisf (tnni) 70

6.5 Dynamic load

In Section 5 the measurements of the vibrations caused by production blasting are presented. Two
acceleration signals are used in the analysis, load case 1 and 2, and the integrated signal (particle
velocity) which are shown in Figure 6-13 and Figure 6-14 respectively. These two signals are
discussed in Section 5. Two load cases, 1 and 2, are selected since they represent typical
acceleration signals measured during production blasting and owing to the high magnitudes of the
signals.
 110

a) b)

3000
2500 - 1500 -
u> 2000 -
E
-

g 1500 7 Z' 1000 -

E
te 1000 E 500 -
500 -
0 0
-500i > -500 9 30
10 20 30 cD
c.) -1000 7 7.)
't -1000
-1500 -
-2000 -1500

Time (ms) Time (ms)

Figure 6-13 Load case 1

a) Particle acceleration, jig b) Particle velocity, ü g

a) b)
3000 -
-(7( 2500 -
0 1500 -F
2000 r
E
c 1500 -I- '7'7 1000 -
.2 1000 E

•ci) 500 >

E 500 -
0 8 o
co -500v
b' 10 20 30 10 20 30
-1000 7
-500 ä
C.)
if -1500 - -1000
-2000 o_
-1500
Time (ms)
Time (ms)

Figure 6-14 Load case 2

a) Particle acceleration, ilg b) Particle velocity, idg
 111

6.6 Results

The tangential stress 470 in the roof of the cross cut has great influence on the load bearing capacity
as well as on the stiffness of the joints. However, the stress state around a cross cut in LKAB:s
underground mines is complex and therefore difficult to calculate. Hence, the range of rock stresses
is chosen from a more general discussion:
When the tangential stress o-e in the roof of the cross cut is higher than 30 -40 MPa the main
failure is probably not ejection of a prism, but failure in the roof caused by high rock stresses
since the uniaxial compressive strength of the rock mass, o,,, is 7 — 36 MPa in the iron ore
according to Malmgren (2001).
Lundman (1998) investigated how the mining in one or two crosscut(s) affected the state of
stress in an adjacent crosscut. He concluded that mining affects the adjacent crosscut
considerably with a decrease in stress level in the roof, see Section 3.9.

For these reasons the stress magnitudes in the roof ((se) is limited to the interval 1 to 30 MPa.

6.6.1 Static bearing capacity for the roof prism

Figure 6-15 and Figure 6-16 show the bearing capacity F,f for the prism, according to Equation
(6.13) and Table 6-2. As evident from Figure 6-15 and Figure 6-16, a large increase of Fit is
obtained when the semi-apical angle a is decreased from 30° to 20°.

12000 -

2- 10000

u. 8000- a 0 =31° and b= 0.4 m

0 0 =31° and b=0.8m
6000 -
a, Ø= 35° and b = 0.4 m
4000 - = 35° and b = 0.8 m
.c
co 2000

10 20 30
Tangential stress in the roof 0-6, (MPa)

Figure 6-15 The static bearing capacity, I:if of the prism, for a semi-apical angle a = 30°
 112

12000 -
Bearing capacityFif (kN)
10000

8000 = 31° and b ---- 0.4 m

..- • -* = 31° and b = 0.8 m
6000
--s-- Ø=35° and b = 0.4 m
-- Ø=35° and b= 0.8 m
4000 -

2000 -

0
0 10 20 30
Tangential stress in the roof cse (MPa)

Figure 6-16 The static bearing capacity, Fif of the prism, for a semi-apical angle a= 20°

The bearing capacity can also be expressed as the displacement at failure of the joints, Uff, where uif
is defined in Equation (6.18). The displacement at failure of the joints, tiff is used in the analysis of
practical reasons and is shown in Figure 6-17 as a function of tangential stress in the roof, a,. As
one can see, uff decreases with increased values of (se , because the stiffness of the joints
increases faster than the bearing capacity Fif, see Equation (6.14) and (6.13) respectively.

0.1000 -1 Prism:
(mm)

a = 20° andq5=-- 35°

width 0.4 or 0.8 m
tiff

Prism:
a =20° and 0= 31°
Displacement at failure of the joints,

0.0100 width 0.4 or 0.8 m

Prism:
a = 30° and(/ 35°
width 0.4 or 0.8 m
0.0010

Prism:

0.0001
10 20
N
30
a 30° ande= 31°
width 0.4 or 0.8 m

Tangential stress in the roof cr o (MPa)

Figure 6-17 Displacement at failure of the joints, uff as function of tangential stress in the roof, o.
 113

The necessary tangential stress to carry the prism (Fif Mg) can be calculated by using Equation
(6.13) which gives

Mg tan a sin
(6.19)
o-0 >
b sin a sin(0 - a)

where g is the gravitational acceleration. Equation (6.17) with data presented in Table 6-2 shows
that the necessary tangential stress only need to be between 0.03 to 0.52 MPa to carry the prism, for
the prism geometries and the friction angles used in this study. This shows the great influence of the
rock stresses for the stability of prism.

6.6.2 Natural frequencies for the roof prism

The natural frequency due to the stiffnesses K, of the joints and Ks of the support is

1K+ s
f= (Hz) (6.20)
2ir M

where K, is defined by Equation (6.14) and the values for M and Ic are shown in Table 6-2 and
Table 6-4 respectively. The natural frequency is shown as a function of a-0 in Figure 6-18. As one
can see the natural frequency of the prism is varying from 270 to 6600 Hz. However, if failure
occurs in the joints i.e. the displacement u exceeds the displacement at failure of the joints, tif f, a
much weaker bearing system (shotcrete) replaces the original bearing system of the joints and
shotcrete = 0). In this case, the natural frequency drops to a value between 28 to 70 Hz
depending on the dip and the width of the prism.

7000
•
N 6000
>o, 5000 -
C _A- a = 30° and b = 0.4 m
g 4000 -ä-- a= 30° and b = 0.8 m
cr
47 3000 - -e-- a = 20° and b = 0.4 m
72 - cc= 20° and b = 0.8 m
2, 2000 -
a
• 1000 -

10 20 30
Tangential stress in the roof, 0-0 (MPa)

Figure 6-18 The natural frequency of the prism as a function of rock stress, dip and width of the
prism.
 114

6.6.3 Dynamic analyses

The measured acceleration time-history has a sampling time of 1/48000 s. Clough & Penzien (1993)
recommend a time increment — period ratio zit/T< 1110 to solve the dynamic equation by the used
integration method (Section 6.2). The author found that At /7' 1/15 gives reliable results for the
analysed load cases. The highest natural frequencies which could be analysed by the integration
method was 48000/15 = 3200 Hz. To allow analysis of the SDOF with higher natural frequencies,
the time increment was reduced to half of 1/48000 s —> 1/96000 s, with help of linear interpolation
according to Figure 6-19. The highest natural frequency analysed was 6200 Hz which is dt/T=
1/15. This procedure to reduce the time increment will (of course) not give any more information
about the load history, the only reason is to make it possible to use higher natural frequencies in the
analysis, see Appendix 6.2.

Load history (measured acceleration).., (üg )(üg ) ,(iig ),,-1 ,• • • •

Interpolated value, (ü,),+1/2

(üg

(ii, );_i

At/2 dt/2 At/2 At/2

72-
dt dt dt = 1/48000 (s)
-7/

Figure 6-19 Linear interpolation to reduce the time increment from 1/48000 (s) to 1/96000 (s)

In Table 6-5 and in Figure 6-20 and Figure 6-21 two typical cases are shown. The Table gives the
input parameters for the analysis and maximum calculated displacements. Figure 6-20 shows the
case where the displacement (response) u > Uff , where iv is the displacement at failure of the joints
and Figure 6-21 shows the case where the displacement u < u f i.e. elastic conditions. There is a
great difference in the response depending on if failure occurs in the joints, or not. If failure occur in
the joints, a much weaker bearing system (shotcrete) replaces the original bearing system of the
joints and shotcrete.
 115

Table 6-5 Rock mechanical properties and results for two typical cases, load case 1
Tangential Friction a Width, Prism/joint Shotcrete
stress, (To angle, 0 b system u Comments
(MPa) p a) Fe c) us), c)
(°) (°) Ziff-19

(m) (1(N) (mm) (N) (mm) (mm)

10 35 30 0.8 1040 0.0034 60 1.0 26 u>u>14,f
(plastic condition)
20 31 20 0.4 2780 0.0065 60 1.0 0.0062 u<u,f
(elastic condition)
a) Bearing capacity of the joints, according to Equation (6.13) or Figure 6-15/Figure 6-16.
b) Displacement at failure of the joints, according to Equation (6.18) or Figure 6-17.
c) Yield load capacity of the support (Fsy) and yield displacement of the support (usy), see Table
6-4.
Time (ms)
10 20 30 40 50 60
3500 - 30
Displacement 0(1)
3000
2500 Dynamic load, 25
load case 1
2000
Elastic vibration
1500
1000
500
o
-500
12 -1000 Yielding phase...
-1500
-2000
-2500 Inelastic displacement
\ Sphng force F,(t z
-3000
Elastic limit, ti "= 1 mrn
-3500
-4000 - -5

Figure 6-20 The displacement u(t) and spring force F,,,,,g(t), plastic conditions, load case 1
Time (ms)
3000 , 0.01
10 15 20 25 35 40 45
'll4 30
2500 -- •0
11
Displacement u(f)
2000 - -0.01

1500 - -0.02

1000 -0.03 E
_e 500 -0.04 i
z

-1 -0.05

-500 ro-
i5
-1000 -0.07

-1500 -0.08

-2000 -0.09

-2500 -0.1

Figure 6-21 The displacement u(t) and spring force Fsp,-,,,g(t), elastic conditions, load case 1
 116

The response (displacement or spring force) can be presented as a function of the natural frequency
of the prism. The displacement u under linear elastic conditions is shown in Figure 6-22 for load
cases 1 and 2. For linear elastic conditions, the response is governed only by the natural frequency
and the load history, which can be shown with the aid of the Duhamel integral method, see
appendix A6.3. Therefore, the response shown in Figure 6-22 is valid for all geomechanical
conditions discussed in this Section as long as the response is linear elastic. Each mark on the
curves in Figure 6-22 represents the maximum elastic response from one numerical analysis. This
type of spectra is in earthquake engineering called displacement response spectra. In fact, this
spectra shows directly the extent to which real SDOF structures respond to the input ground motion.

10.000

1.000 -

—B— Load case 1

—a-- Load case 2

0.010 -

I
0.001 -I ,
0 1000 2000 3000 4000 5000 6000 7000
Natural frequency (Hz)

Figure 6-22 Linear elastic response (displacement, u) as a function of natural frequency of the
prism, load case 1 and 2.

If the displacement, u, for load case 1 and 2 is less than the displacement at failure of the joints, uff,
the prism is stable without support. Figure 6-23 and Figure 6-24 show u and 'vas functions of the
natural frequency and the tangential stress in the roof As shown earlier (in Figure 6-18) there is a
linear relation between the tangential stress in the roof and the natural frequency. However, for a
particular tangential stress in the roof, a° , the corresponding natural frequency varies because of
different ground conditions. For example, if a, =10 MPa the corresponding natural frequency
varies from 1760 to 3520 Hz, see Figure 6-23 and Figure 6-24.

When the semi-apical angle a= 300 and the prism is subjected to load case 1 the prism is unstable
for all analysed tangential stresses because u > tiff, see Figure 6-23. For a= 20 0 the situation is more
favourable with higher values of uff as shown in Figure 6-24. The bearing capacity is sensitive for a
variation of o4 which was already shown in Figure 6-15 and Figure 6-16.
 117

a)
Tangential stress o-0 (MPa)
1 3 5 7 9 11 13 15 17 19 21
10.0000
Displacement, u, and displacement

1' 0000 -
E

Displacement (response), u, Load case 1

0.1000
-6-- Displacement (response), u, Load case 2
Displacement at failure, uff, Ø= 31 0
Displacement at failure, tiff, 0= 350
_c 0.0100-
ö

0.0010 -
ca

0.0001
1000 2000 3000 4000 5000 6000

Natural frequency (Hz)

b)
Tangential stress cre(MPa)
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
10.0000 -
Displacement, u, anddisplacement

1.0000 -
E
z'‘-• Displacement (response), u, Load case 1
ai 0.1000 - --11- Displacement (response), u, Load case 2
c
Displacement at failure, uff, Ø = 31 0
a) Displacement at failure, uff, Ø = 35°
0.0100 -
"5

'cT3 0.0010

0.0001
1000 2000 3000 4000 5000 6000

Natural frequency (Hz)

Figure 6-23 Linear elastic response (displacement, u) and displacement at failure of the joints, uff
as a function of natural frequency and tangential stress in the roof
a) Width = 0.4 m and a= 30° b) Width = 0.8 m and a= 30°
 118

Tangential stress cre(MPa)

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
10.0000

E 1.0000
-
cn
0.1000 --.— Displacement (response), u, Load case 1
-0 (I
---- Displacement (response), u, Load case 2
-- Displacement at failure, uff, Ø= 31°
0.0100 Displacement at failure, uff, 0= 35°

ca
o
=
0.0010 -
CO 3333
72.

cc

0.0001
1000 2000 3000 4000 5000 6000
Natural frequency (Hz)

b)

Tangential stress o-9 (MPa)

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
10.0000
E g*
E
'EL 1.0000
cn
'175
Displacement (response), u, Load case 1
C g
0.1000 Displacement (response), u, Load case 2
cp — Displacement at failure, uff, 0 = 31°
, Displacement at failure, uff, 0 = 35°
0.0100 ---
112
z
7). 0.0010
5 75

0.0001
1000 2000 3000 4000

Natural frequency (Hz)

Figure 6-24 Linear elastic response (displacement, u) and displacement at failure of the joints, /cif
as a function of natural frequency and tangential stress in the roof
a) Width = 0.4 m and a= 20° b) Width = 0.8 m and a= 200
 119

To make an estimate of the maximum dynamic response for load case 1 (load case 1 is the
dimensioning load case, see Figure 6-22) for various natural frequencies, a number of analyses have
been performed. Four natural frequencies (Figure 6-22), three of them corresponding to local
maximum linear responses and the fourth to a local minimum linear response (f= 1135 Hz), have
been investigated for various geomechanical conditions. The input data for the mechanical
properties of the shotcrete and the data for the prism/joint system are shown in Table 6-4 and Table
6-6 respectively. The chosen natural frequencies also represent various tangential stress conditions
in the roof, a, = 1.5 to 23.0 MPa depending on the geomechanical conditions.

Table 6-6 Input data for the numerical analysis

Semi-apical angle, a (°) 30 20

Width, b (m) 0.4 0.8 0.4 08
Friction angle 0 (°) 31 35 31 35 31 35 31 35

ao (MPa) a) 1.5 2.7 2.8 4.5

f— 91313 (Hz) KJ (GN/m) b) 10.0 39.9 15.8 63.3

Fff (km c)
18 79 66 295 390 475 1253 1526
a8 (MPa) 22 38 38 60

f--- "35 Ici (GN/m) 15.9 63.4 25.2 101

(Hz)
P (1(1\1) 26 116 89 400 529 645 1671 2035
a9 (MPa) 39 61 61 92
f= 1640 KJ (GN/m) 33.1 132 52.6 210
(Hz)
F.,./ (1(N) 46 205 143 642 836 1018 2562 3121
a, (MPa) 10.7 15.8 15.8 23.0
f=3750 Ici (GN/m) 173 692 275 1098
(Hz)
Fif (1(N) 127 568 371 1663 2200 2680 6406 7802
a) a, is the tangential stress in the roof
b) Ki is the stiffness of the joints, according to Equation (6.14)
c) Fif is the bearing capacity of the joints, according to Equation (6.13)

The results of these analyses is shown in Figure 6-25. All responses exceeded the displacement at
failure of the joints, tiff, and the original bearing system of joints and shotcrete was changed to
consist only of shotcrete because failure had occurred in the joints.

The maximum response (displacement) varied from 8 to 42 mm for the various natural frequencies.
However, in the case when the semi-apical angle was equal to 20° and the natural frequency was
equal to 3750 Hz, (Figure 6-25 b), the response was lower than the yield displacement of the
shotcrete lining (usy). For the other cases, the yield displacement was exceeded and a yielding phase
of the shotcrete lining occurred. In no cases was the maximum calculated u (for all analysed geo-
mechanical conditions) greater than the failure displacement of the reinforced shotcrete, usf = 70
mm. Figure 6-26 shows the force-displacement curve for the shotcrete, same curve as in Figure
6-12.
 120

a) b)
Failure displacement of the shotcrete lining, tie 70 mm

Yield displacement of the support (shotcrete lining),

u, = 1.0 min
100 - 100 -2 111 = 900 (Hz), load case 1
E 10 -
E
— 10
z 1 - 1 O f = 1135 (Hz), load case 1
a) 0.1 - 0.1
E
0.0 1 0.01 - o f = 1640 (Hz), load case 1
a
m 0.001 0.001
2 0.0001 0.0001 • f = 3750 (Hz), load case 1
= 31° 35° 31° 35° 31° 35°
Width = 0.4 (m) Width = 0.8 (m) Width = 0.4 (m) Width 0.8 (m)

Lower and upper boundary for displacement at failure of

the joints, uff, depending on the geomechanical conditions

Figure 6-25 Maximum dynamic response for various natural frequencies and geo-mechanical
conditions
a) Semi-apical angle, a= 300
b) Semi-apical angle, a= 20°

Fs), = Yield load capacity of the

shotcrete lining
usy = Yield displacement of the
F, = 60 kN shotcrete lining
usf = Yield displacement of the
shotcrete lining

Displacement, u

Figure 6-26 Force-displacement curve for the support (shotcrete lining)

 121

6.64 Shotcrete as rock support after failure has occurred in the rock joints

The contribution of shotcrete to the bearing capacity is negligible as long as the displacement u is
less than the displacement at failure of the joints, uff. This depends on the great difference in
stiffness, Ki>> lc, where K and K, is the stiffness of the joints and of the support (shotcrete lining)
respectively. However, if u > uli only the shotcrete will support the prism.

If an event has occurred and u > uff , the shotcrete is the only support for the prism. If u in addition
has exceeded the yield line displacement for the shotcrete lining, usy, a permanent displacement has
occurred, see for example Figure 6-20. Furthermore, if a second vibration event occurs with u >
usy then a new inelastic deformation will be added to the first one. Figure 6-27 shows a case when
the prism is exposed to a second event equal to load case 1, failure has already occurred in the joints
during the first event, and the only bearing member is the shotcrete. The calculation was done with
the assumption that the stiffness of the joints, Ki and the bearing capacity of the jonts F1 were equal
to zero, also in compression.
Time (ms)
0 10 20 30 40 50 60 70
1500 13

-7/ 12
Displacement u(t)
11
1000
10
Elastic vibration
Dynamic load 9
500
Yielding phase 8

7
E
6

inelastic displacement 5 '5

a. 4 :•3
-500

-1000

-1500
Elastic limit, u„ = 1 mm

-2

-3

Mechanical properties of shotcrete, see Table 6-4

Prism, width = 0.4 m, semi-apical angle = 30°.

Figure 6-27 Results from an analysis where the prism is exposed to a second event equal to load
case 1, failure has already occurred in the joints during the first event, and the only
bearing member is the shotcrete.

In Table 6-7 the responses from a second event equal to load case 1 and Fif = K., = 0 are summarised
for all investigated geomechanical conditions. The total displacement is the sum of the maximum
inelastic displacement from the first event (approximately equal to the maximum displacement =42
mm, see Figure 6-25 b) and the displacement of shotcrete from the second event (12 mm), see Table
6-7. It is not likely that the shotcrete will be exposed to more than two dynamic events of the
 122

magnitudes that load case 1 represents, since the stress waves induced by blasting rapidly attenuate
with increasing distance from the source. In any case, the total displacement equal to 54 mm shows
that the shotcrete has necessary yielding capacity for two events with the same magnitude as load
case 1, since 54 mm < usf = 70 mm, where usf is the displacement at failure of the shotcrete lining.

Table 6-7 Maximum calculated displacement (from a second event) for shotcrete carrying a
prism (i.e. Fit. = K = 0), load case 1. (Mechanical properties for the shotcrete
according to Table 6-4)

Friction angle, Ø (°) 31 35

Width (m) 0.4 0.8 0.4 0.8
a (0) 30 20 30 20 30 20 30 20
Maximum u (mm) 10 12 12 12 10 12 12 12

6.6.5 Ductility of the shotcrete lining

Rock support for dynamic events must be ductile to be effective in absorbing energy. The joint
failure is brittle but the reinforced shotcrete has the necessary plastic capacity to absorb the energy
from the dynamic event. The elastic energy, which the joints can absorb without failing is

1 F2
We = (6.21)
``'"` 2 IC

and the energy, which the reinforced shotcrete can absorb, is

W = F7 (11 sf — 0 .5117 ) (6.22)

The maximum value of Werostic for any of the analysed prism is 28 Nm, F1= 10.2 MN (Figure 6-16)
and KJ = 1840 GPa (Table 6-3). The energy, W, for the rock support is 4170 Nm, see Table 6-4.
This shows why the reinforced shotcrete is much more effective than the joints.

Plain shotcrete

The elastic energy, which the plain shotcrete can absorb, is much less than the energy which the
reinforced shotcrete can absorb. If we assume that the crack load (Fsc) is 2 times the yield load
capacity Fe and assume the same stiffness 1C, the maximum elastic energy is

1 F2 1 (2 60 .103 )2
=120 Nm « 4170 Nm
2 K, 2 60000•10

that is, 4 times higher than the elastic energy of the prism.
 123

This simple calculation shows that use of plain shotcrete as rock support subjected to dynamic loads
is not recommendable.

6.7 Discussion

A single degree of freedom (SDOF) model of a symmetric roof prism subjected to stress waves
induced by blasting has been developed. The model consists of a mass and two springs which
model the behaviour of the joints and the support. Furthermore, the damping is represented by a
dashpot. The load induced by the stress wave generated by blasting, is measured as particle
acceleration. The vibration measurements used in this analysis were done at a horizontal distance of
4.5 to 7.5 m from the blast holes.

The results from the linear elastic analyses, presented as a displacement response spectra, showed
that a prism can be ejected from the roof as a result from the production blasting, since the
displacement was larger than the displacement at failure of the joints of the prism. This is also
observed by the failure mapping of the roof of the cross cuts close to the drawpoint, see Section 5.

Furthermore, the non-linear analysis showed that shotcrete reinforced with approximately 40 kg/m2
Dramix steel fibres or stress-relieved anneal steel mesh (ø6c150 fy=220 N/mm2) has the necessary
bearing capacity to support prisms with shape and mechanical properties typical for the
Kiirunavaara mine exposed to a load induced by the stress waves with a magnitude of 1.1 m/s.
Because of the brittle behaviour of plain shotcrete is it not recommendable to use plain shotcrete as
rock support under dynamic loads.
 124

7 DISCUSSION WITH CONCLUSIONS AND PROPOSAL FOR

FURTHER RESEARCH AND DEVELOPMENT

7.1 General

The objectives for the thesis are to

increase the understanding of shotcrete as rock support
determine the load conditions caused by production blasting
investigate how shotcrete acts as a rock support when it is subjected to dynamic loading
induced by the production blasting.

To fulfil the objectives a number of activities have been performed. The result of the work to
fulfil the first objective is discussed in Section 7.2 and the result of the work to fulfil the two
remaining objectives is discussed in Section 7.3. The proposal for further research and
development is presented in Section 7.4.

7.2 Shotcrete - field and laboratory test

7.2./ Field and laboratory tests performed in the Kiirunavaara mine

The results from the failure mapping shows that more than 80% of the damage was small
fallouts of plain shotcrete in areas with low or no adhesion strength. Important factors, which
may cause the failure, are rock deformation, shrinkage and dynamic loads due to blasting.

It is also important to get sufficient thickness all over the shotcreted area i.e. avoid areas with
thin shotcrete, (2 cm or less). The shotcrete thickness was 2 cm or thinner for more than 90%
of the fallouts of plain shotcrete. Furthermore, 60 % of the failures were in conjunction with
an apex on the rock surface together with a thin shotcrete layer. The results showed that in
18% of the thickness measurements the shotcrete thickness was less than or equal to 2 cm. If
the shotcrete lining with ordered thickness of 3 —5 cm had the same distribution of shotcrete
thickness as the tests, the supporting ability of as much as 18% of the shotcrete lining can be
considered as more or less negligible.

The growth of the adhesion strength of shotcrete on a sandblasted concrete wall and the
growth of the compressive strength of shotcrete show a clear correlation. The author has not
been able to compare the result with other similar test results, because the author has not
found any similar tests reported in the literature.

A way to increase the adhesion strength is to use water-jet scaling as a treatment method. The
water jet-scaling machine used in the tests presented in this thesis used a water pressure equal
to 22 MPa. The results showed that the adhesion strength was significantly higher on surfaces
that had been water-jet scaled than those treated in other ways. Another water-jet scaling test
was done at a tunnelling project in Hallandsås in the Southern part of Sweden. Lundmark and
Nilsson (1999) reported similar results from this test. However, the scatter in the results was
so wide that one could not from a statistical analysis conclude if the adhesion strength was
significantly higher on water-jet scaled surfaces than on surfaces scaled by mechanical scaling
and cleaned by spraying water on it. Furthermore, they reported shorter scaling time with
water-jet scaling especially if the number of underbreaks was minimised. However, should
 125

any underbreaks exist there was a disadvantage because mechanical scaling must be done
before the water-jet scaling. In our tests, only the adhesion strength was investigated.

To control surface treatment and thickness of the shotcrete the author believes that water-jet
scaling in combination with a shotcrete robot which can control the thickness of the shotcrete
layer will help to increase the quality of the shotcrete. To be able to remove smaller blocks the
water-jet aggregate should be equipped with a small scaling hammer. Small, because a drift or
tunnel in hard rock should not be drifted by scaling. To minimise the number of overbreaks
this concept could be combined with cautious blasting.

7.2.2 Panel and beam tests

To find a test method to model the interaction between shotcrete and rock is very difficult.
This is mainly because of the heterogeneous properties of the rock mass but there are also
difficulties to include the effect of rock stresses.

The test model used in this study consisted of a square panel with clamped edges, loaded in
the centre by a point load. The main reason for using a point load is to investigate the
behaviour of the shotcrete around the rock bolt.

hi many tests performed to investigate the behaviour of fibre reinforced shotcrete, the
concrete is poured instead of sprayed. The main disadvantage of spraying instead of pouring
the specimens is the deviation of the amount of steel fibres from the planned amount of fibres.
Also, differences in rebound for various types of fibres can make it difficult to compare the
bearing capacity for shotcrete reinforced with various fibre types. This was apparent in these
tests where all test specimens were shotcreted. The advantages of using shotcrete instead of
poured concrete are for instance that the effects of accelerators, fibre directions and
compaction are included in the test results.

The test results indicate that the typical failure mechanism was a bending failure followed by
punching failure. Furthermore, the tests showed that panels reinforced with steel fibres were
more ductile and had higher punching capacity than panels reinforced with steel mesh.

The yield line theory based on the rotating capacity in the yield line was not capable of
predicting the load bearing capacity. The actual load bearing capacity of the panels was much
higher than the calculated bearing capacity according to yield line theory. This can probably
be explained by the fact that dome action is the main load carrying mechanism. The dome
effect occurs only if the edges are prevented from moving in the plane of the panel. This
constraint can be assumed to be present for a lining, because of the adhesion strength and
irregularities of the rock surface. The dome effect can, however, be low when the shape of the
lining in the roof is directed downwards. Finally, the results from the beam test and the results
from the panel tests can not be substitutes for each other because of the different loading
modes.
 126

7.3 Shotcrete subjected to dynamic loads from production blasting

Close to the drawpoint, the dynamic influence from the production blasting is significant. In
this study it was decided to analyse the ejection of joint-defined blocks because many of the
failures close to the brow of the cross cuts seem to be structurally controlled.

The model used to study the dynamic response of the prism was a Single Degree of Freedom
(SDOF) model with non-linear behaviour. The stress wave time history was obtained from
measurements during production blasts in the Kiirunavaara mine. In addition to the vibration
measurements 34 drawpoints were failure mapped.

The vibration measurements showed a wide scatter of magnitudes of the particle acceleration,
which was expected. The maximum particle velocity was approximately 1.1 m/s at a
horizontal distance of 4.5 m away from the blast holes.

The results from the linear elastic analyses, presented as a displacement response spectra,
showed that a prism can be ejected from the roof as a result of production blasting, since the
displacement was larger than the displacement at failure of the joints of the prism. This was
also observed by the failure mapping of the roof of the cross cuts close to the drawpoint.

No vibration measurement has been done close to the blasted fan than at a horizontal distance
of 4.5 m, for practical reasons. The non-linear numerical analysis showed that analysed
reinforced shotcrete has the necessary bearing capacity to support actual prisms for load
histories measured at a horizontal distance of 4.5 m away from the blasted fan. This was
confirmed by the failure mapping which showed that reinforced shotcrete was more and less
intact in the roof of the cross cut at a horizontal distance of 1 to 4 m away from the drawpoint.

Furthermore, the analysis showed that plain shotcrete is not recommendable because of its
brittle behaviour. The observations from the failure mapping showed that the area or zone of
failed shotcrete was much larger in areas with plain shotcrete compared to areas with
reinforced shotcrete. Plain shotcrete has not the necessary ductility to prevent wedges from
falling down.

The SDOF model is of course a simplification of the interaction between the shotcrete and the
rock mass. However, to analyse the ejection of joint-defined block as symmetric prism, is an
example where a SDOF model could be used without too many simplifications.

Test results presented in Section 2.2 (Tannant, 1997) showed that mesh or steel fibre
reinforced shotcrete can maintain its supporting function for peak particle velocities of up to
1.5 to 2 m/s at the drift surface when the rockmass itself is not severely damaged. The
numerical analysis and the failure mapping show an agreement with Tannant's results.

Can the stability problem in connection to the drawpoint primarily be caused by stress
changes instead of the impact from blasting? Lundman's result (1998) from deformation
measurements indicated no increased deformation within 10 m of the drawpoint. However,
the results from the failure mapping in this thesis shows that stability problems occur. The
small measured deflection and the author's observation of failures close to the drawpoint
indicate that stability problems are caused by the production blasting, maybe in interaction
 127

with unloading of the roof But the influence on the stability caused by stress changes must be
studied in detail.

For static loading conditions, it is interesting to note that the prism is stable for very low
tangential stresses in the roof of the cross cut or tunnel. Necessary stresses are calculated to lie
in the range of 0.03 to 0.53 MPa with rock mechanical properties typical for the Kiirunavaara
mine.

7.4 Proposal for further research and development

7.4./ Research

In this study failure mapping together with laboratory tests have been performed. To be able
to understand the interaction between a surface support as shotcrete and rock and rock bolts
further studies must be done. More specifically, the following issues need to be addressed.

The shotcrete lining follows the contour of the tunnel which means the irregularities of the
tunnel shape will influence the structural behaviour of the shotcrete lining. Chang (1994)
has done studies on this phenomenon, but further studies have to be done to clarify the
impact of irregularities on the structural behaviour of the shotcrete lining.
In tunnelling the largest rock stress changes mostly occur during the construction time.
Myrvang (1982) concluded that shotcrete sprayed on the rock surface was almost
stressless, while the stresses in the rock behind the shotcrete could have high magnitudes.
This fact can indicate that the stiffness and strength of the shotcrete lining is not mobilised
until the rock stress changes are done (especially in hard rock). In mining, however, the
rock stresses increase when mining goes deeper and openings are subjected to a varying
rock stress state caused by the mining. In this cases also the shotcrete lining will be
influenced by the varying rock stress state. How much depends on the stiffness ratio for
shotcrete-rock mass and the strength. But the zone close to the opening, disturbed by the
blasting, has other properties than the rock mass outside the disturbed zone, this will
influence the behaviour of the rock-shotcrete interaction when rock stresses are changed.
The main failure in the panel tests was punching failure, how does the dome effect
influence the punching failure mechanism and how does the effect of irregularities
influence the dome effect?
The influence of creep and shrinkage of both shotcrete and the rock mass on the behaviour
of shotcrete is not very well understood.
 128

7.4.2 Development

The following issues are more suited for development i.e. how to improve the method or
equipment for industrial purposes.

The influence on the behaviour of shotcrete when using water jet scaling is tested and
discussed in this study. However, to implement the water-jet scaling methods requires
development of the equipment, for example what is the right water pressure for the rock or
should the pressure be varied for various types of geology? What type of strategy for
construction of tunnels and drifts should be used when using water-jet scaling.
The thickness of the shotcrete is a very important parameter both for the technical
function and for the cost of the support. Semi-automatic or automatic robots, which spray
shotcrete with a more uniform thickness, compared to what is done today may be a
solution. These robots are already in limited use. The author is convinced that
development of these robots would help the industry to increase the productivity and
quality of the lining.
To prevent fallouts of loose rock material, the rock requires immediate support,
unfortunately the use of shotcrete in short term stabilisation is clearly inappropriate.
During the last years alkali free accelerators have been used and one important effect is
faster early strength development, see Section 2.3. In spite of that, in many situations the
short term stabilisation is too low, therefore development of more effective accelerators or
changes to the clinker phases in the cement to increase the development of strength are
ways to increase the performance of shotcrete. The hydration of cement is discussed in
Section 2.3
 129

REFERENCES

Anse11 A. 1999 Dynamically Loaded Rock Reinforcement Doctoral Thesis, Department of

Structural Engineering, Royal Institute of Technology (KTH), Stockholm, Sweden

ASTM C1018 1994. Standard Test Method for Flexural Toughness and First Crack Strength
of Fiber-Reinforced Concrete (Using Beam With Third-point Loading)

Ay L. 1999. Using Prestressed Steel Fiber Reinforced High Performance Concrete in the
Industrialization of Bridge Structures. Licentiate Thesis, Department of Structural
Engineering, Royal Institute of Technology (KTH), Stockholm, Sweden

Bandis S.C., A.C. Lumsden and N. Barton 1983 Fundamentals of Rock Joint Deformation.
International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts,
Vol. 20 No.6, 1983, p.249 —268

Banthia N., J-F. Trottier, D. Beaupre and D. Wood 1994. Properties of steel fiber reinforced
shotcrete. Canadian Journal of Civil Engineering 21, 1994, p. 564-575

Barton N. 1988. Project OSCAR — LKAB/LuTH Estimation of shear strength parameters for
mine design studies. Norwegian Geotechnical Institute (NGI), report No. 87659-1, 2 June
1988

Barton N. and V. Choubey 1977. The Shear Strength of Rock Joints in Theory and Practice.
Rock Mechanics 10, p.1-54 (1977), by Springer Verlag

Barton N., R.lien and J.Lunde 1974. Engineering Classification of Rock Masses for the
Design of Tunnel Support. Rock Mechanics 6, p.189-236, 1974, Springer Verlag

BBK 94 1995. Boverkets handbok om betongkonstruktione, band 1, konstruktion. Boverket

(in Swedish)

Becket D. and J. Humphreys 1989. Comparative tests on plain, fabric reinforced and steel
fibre reinforced ground slabs. Report No. TP/B/1, School of Civil Engineering Thames
Polytechnic, Oakland Lane, Dartford, UK

Bergan P., P.K. Larsen and E. Mollestad 1981 Svingning av konstruksjoner. Tapir forlag (In
Norwegian)

Bernard E.S. 1997. The Influence of Edge Restraint on Flexural Behaviour in Square SFRC
Slabs. Engineering report No. CE5, Department of Civil and Environmental Engineering,
Facultyof Engineering, University of Western Sydney, Nepean Kingswood NSW Australia,
July 1997, ISSN 1326-0693

Betonghandboken 1997. Material (Concrete Handbook, Material), in Swedish

Beton_grapport nr 4 1995. (Concrete report No. 4) Svenska Betongföreningen/Swedish

Concrete Association (in Swedish)
 130

Bieniawski Z.T. 1976. Rock mass classification in rock engineering. Proceeding of the
Symposium on Exploration for Rock Engineering, Johannesburg, South Africa, November
1976, p.97-106

Birke H.. 1975 Kupoleffekt vid betongplattor Doctoral Thesis, Royal Institute of Technology
(KTH), Stockholm, Sweden, report No 108, 1975

Blom G. 1970 Sannolikhetsteori och statistikteori med tillämpningar Studentlitteratur (in

Swedish)

Brady B.H.G.and E.T. Brown 1999. Rock mechanics for underground mining Second edition
reprinted with correction, Chapman&Hall

Brantmark J. 1997. Rock support in weak rock — a study based on the Uri project. Licentiate
Thesis, Division of Soil and Rock Mechanics, Department of Civil and Environmental
Engineering, Royal Institute of Technology (KTH), Stockholm, Sweden

Brent G.F. & G.E. Smith 1996 Borehole pressure measurements behind blast limits as an aid
to determining the extent of rock damage Proceedings of the fifth international symposium on
rock fragmentation by blasting — Fragblast-5 Montreal, Quebec, Canada, 25-29 August 1996,
p.103-112

CEN/TC104/WG 10 N 62 1999. Test Methods for sprayed concrete — Part 3: Detemination of

flexural and residual strength (TTT-2). Secretariat NABau, DIN Deutsches institut far
normung e.V.

CGE 1995 Vidhäftningsmätare modell CGE 101. Swedish Concrete and Grouting
Equipment AB (operating manual for the adhesion equipment, in Swedish) 1995

Chang Y. 1994 Tunnel Support with Shotcrete in Weak Rock — A Rock Mechanics Study.
Doctoral Thesis, Division of Soil and Rock Mechanics, Royal Institute of Technology (KTH),
Stockholm, Sweden

Clements M. 1999. The use of shotcrete in Australian underground hardrock mines. Third
International Symposium on Sprayed Concrete, Gol, Norway September 26-29, 1999, p.150-
160

Clough R.W. & J. Penzien 1993 Dynamics of structures. Second edition, McGraw Hill
international editions

Grasser E. et.al. 1994 Concrete Structures, Euro Design Handbook, 1994-1996, Ernst & Sohn

Durrheim R., A.T. Haile and W.D. Ortlepp 1998 Observation of shotcrete behaviour.
Shotcrete and its application, 25-26 February 1998. The South African Institute of Mining and
Metallurgy

EFNARC 1999. European Specification for Sprayed Concrete.European Federation of

Producers and Applicators of Specialist Products for Structures
 131

Ellison T. 2000. Vidhäftningsprovning Södra länken Stockholm. Personal communication (in

Swedish)

Enger J. Statistisk analys av mätdata. Utgåva 4 (in Swedish)

Erlien 0. 1999. On site experiences with alkalifree shotcrete accelerators in Norway. Third
International Symposium on Sprayed Concrete, Gol, Norway September 26-29, 1999, p. 227-
236

Esplay S., C.C. Langille and D. McCreath 1995 Innovative liner support in underground hard
rock mines. OM Bulletin, July-August 1995, Volume 88, No 992, p.66 — 74

Examine313 User's Manual, version 4.0, Rock Engineering Group, University of Toronto,
Canada, Rocscience, Inc.

Franzén T. 1992 Shotcrete for underground support — A state of art report with focus on steel
fibre reinforcement. Proceedings of the International Symposium on Rock Support, Sudbury,
Canada, 16-19 June 1992, A.A. Balkema, p.91 - 104

Fredriksson A. and H. Stille 1992. Bergförstärkningsprinciper för olika typfall i svenska

gruvor Gruvteknik 2000, report No 92:07 (in Swedish)

Golser, J. 1995. History, Definiton, Principles. NATM, Sommercourse July 2 — 8, 1995,

IACES

Grimstad E.1997. Rock support in hard rock tunnels under high stress International
Symposium on Rock Support — Applied solutions for underground structures, June 22-25,
1997, Lillehammar, Norway, p.504 - 513

Groth P. 2000 Fibre reinforced Concrete — Fracture Mechanics Methods Applied on Self-
Compacting Concrete and Energetically Modified Binders. Doctoral Thesis, Department of
Civil and Mining Engineering, Division of Structural Engineering, Luleå University of
Technology, Sweden

Halm T. 1983 Sprutbetongens vidhäftning mot olika bergytor. Swedish Rock Engineering
Research Foundation — BeFo and Royal Swedish Fortifications Administration — FortF,
Report No BeFo 55:1/83 or FortF C 206, (in Swedish)

Hamrin H. 1986 Guide to underground mining, methods and applications. Atlas Copco MCT
AB

Hansågi I. 1965. Praktisk bergmekanik och bergförstärkning. Almkvist & Wiksell,

Stockholm, Sweden

Hansson K-E 1999 History and geology of the Kiruna Area. Internal LKAB-report

High Performance Concrete Structures, Design Handbook 2000. Svensk Byggtjänst, Sweden

Hilti AG A combined Method for Determining the Early Strength of Sprayed Concrete,
operating manual (draft), HILTI AG
 132

Hoek E. and E.T. Brown 1980. Underground Excavations in Rock. Institution of Mining and
Metallugy

Hoek E. and E.T. Brown 1997. Practical estimates of Rock Mass Strength. International
Journal of Rock Mechanics and Mining Sciences Vol. 34 No.8, 1997, p.1165 - 1186

Hoek E., P.K. Kaiser and W.F. Bawden 1995. Support of Underground Excavations in Hard
Rock. A.A. Balkema

Holmgren J. 1979 Shotcrete, Punch-loaded shotcrete linings on hard rock Dissertation. Royal
Institute of Technology (KTH), Stockholm, Sweden, Swedish Rock Engineering Research
Foundation — BeFo report No. 7:2/79

Holmgren J. 1985. Bolt Anchored Steel Fibre Reinforced Shotcrete Linings. Swedish Rock
Engineering Research Foundation — BeFo and Royal Swedish Fortifications Administration —
FortF, Report No BeFo 73:1/85 or FortF A 2:85

Holmgren J. 1992 Bergförstärkning med sprutbetong. ISBN 91-7186-298-6 (in Swedish)

Holmstedt A. 1999. Personal communication

Holmstedt A. 1994 Bergmasseklassificiering av Kiirunavaara malmkropp. Internal LKAB

report No. 94-788, December 13, 1994 (in Swedish)

Holmstedt A. 1995 Bergmasseklassificiering av liggväggen i KUJ ovanför nivå 740. Internal

LKAB report No. 95-782, September 27, 1995 (in Swedish)

Jodl H.G., J. 1995. Construction Method, NATM 1995. NATM, Sommercourse July 2 — 8,
1995, LACES

Kaiser P.K. and D.D. Tarmant 1997. Use of shotcrete to control rockmass failure International
Symposium on Rock Support — Applied solutions for underground structures, June 22-25,
1997, Lillehammar, Norway, p.580-595

Karlsson L. 1980. Sprutbetongens vidhäftning mot bergunderlag. Byggforslcningsrädet, report

No R69:1980. (in Swedish)

Kirsten H.A.D, W.D. Ortlepp and T.R. Stacey 1997. Performance of reinforced shotcrete
subjected to large deformations. 1st Southern African engineering Symposium, South African
National Group of ISRM, Johannesburg 1997

Kirsten H.A.D. 1993. Equivalence of mesh- and fibre-reinforced shotcrete at large delections.
Canadian Journal of Civil Engineering 30, 1993, p. 418-440

Kirsten H.A.D. 1998. Design criteria for shotcrete as support membrane. Shotcrete and its
application, 25-26 February 1998. The South African Institute of Mining and Metallurgy

Langen I. & R. Sigbjörnsson 1979 Dynamisk analyse av konstruksjoner. Trondheim,

Norway, Tapir (in Norwegian)
 133

Larsson M. 1998 Instruktion för rasborrningsplanering. Internal LKAB report, in Swedish

Laubscher D.H. and H.W. Taylor 1976 The importance of geomechanical classification of
jointed rock masses in mining operations. Proceeding of the Symposium on Exploration for
Rock Engineering, Johannesburg, South Africa, November 1976, p.119-128

Laubscher D.H. 1984 Design aspects and effectiveness of support system in different mining
conditions Trans. Instn Min. Metall (Section A:Min. industry), 93, April 1984, p. A70-A81

LKAB 1999. Provning av sprutbetong — vidhäftningshållfasthet. Internal report, the report

was performed by Vattenfall Utveckling AB, report No 19078-14. (in Swedish)

LKAB's image archive 2001, LKAB's home page: www.lkab.com

Lukas W. and W. Kusterle 1990. The influence of water glass on the technological parameters
of shotcrete. Shotcrete for Underground Support V, Proceedings of Engineering Foundation
Conference, Uppsala, Sweden, 1990. p. 197-212

Lundman P. 1998. The effect of changing stresses on the stability of underground openings —
a studyof the behaviour of underground tunnels. Licentiate Thesis, Department of Civil and
Mining Engineering, Division of Rock Mechanics, Luleå University of Technology, Sweden

Lundmark T. and L. Nilsson 1999. Vattenskrotning vid sprutbetongarbeten. Master thesis,

examensarbete nr 119, Department of Structural Engineering, Royal Institute of Technology
(KTH), Stockholm, Sweden (in Swedish)

Magnor B. and Mattson H. (2000). Strukturgeologisk modell över Kiimnavaara, slutrapport.

CTMG report 00001 (in Swedish)

Magnusson 1987. Mätningar på järnvägsbro i Gnosjö, Småland, Sweden. Stencil, Department

of Structural Engineering, Royal Institute of Technology (KTH), Stockholm, Sweden (in
Swedish)

Malmberg B. 1993a. Shotcrete for Rock Support Guidelines and Recommendations — A

compilation. Swedish Rock Engineering Research Foundation, BeFo 463:1/93 (in Swedish)

Malmberg B. 1993b. Sprutbetong — uppföljning av sprutbetongprovningar på Grödingebanan.

Banverket No T93-913/34 (in Swedish)

Malmberg B. 2001. Personal communication

Malmgren L. and T. Svensson 1999 Investigation of Important Parameters for Unreinforced

Shotcrete as Rock Support in the Kiirunavaara Mine, Sweden. Proceedings of the 37th U.S.
rock mechanics symposium, Vail, Colorado, USA, 6-9 June 1999, p.629-635

Malmgren L. 1999. Sprutbetong — Provning av plattor och balkar. Internal LKAB report, No
99-704 (in Swedish)

Malmgren L. 2001. Internal notes

 134

Martin C.D., D.R. McCreath and M. Stocbmal 1997 Estimating support demand-loads caused
by stress-induced failure around tunnels International Symposium on Rock Support — Applied
solutions for underground structures, June 22-25, 1997, Lillehammar, Norway, p.235-248

Melbye T.A. 2001 Modem advances and applications of shotcrete. Proceedings of the
international conference on engineering developments in shotcrete; Hobart, Tasmania,
Australia, 2 —4 April 2001, p. 7 —29.

Morgan D.R. and D.N. Mowat 1984. A Comparative Evaluation of Plain, Mesh and Steel
Fiber Reinforced Concrete Proceedings of International Symposium on Fibre Reinforced
Concrete, American Concrete Institute (Ad), p.307-324

Morgan D.R., R. Heere , N. McAskill and C.Chan System Ductility of Mesh and Synthetic
Fibre Reinforcement Shotcretes Third International Symposium on Sprayed Concrete, Gol,
Norway September 26-29, 1999, p. 358 -372

Myrvang A.1982 Stresses in shotcrete. Fjellsprengningstelmik, Bergmekanikk, Geoteknikk

p.26.1 — 26.9 (in Norwegian)

Myrvang A. and A. Davik 1997 Heavily spalling problems in road tunnels in Norway — long
time stability and Derformance of sprayed concrete as rock support. International Symposium
on Rock Support — Applied solutions for underground structures, June 22-25, 1997,
Lillehammar, Norway, p.751-764

Mörch A. 1993 Large scale testing of shotcrete Shotcrete for underground support VI,
Proceedings of the Engineering Foundation Conference, Niagara-on-the-Lake, Canada, May
2-6, 1993, p.41 - 48

NCA No 7 1999. Sprayed Concrete for Rock Support. Norwegian Concrete Association

Nilsson U. 2000. Load bearing capacity of steel fibre shotcrete linings. Licentiate Thesis,
Department of Structural Engineering, Royal Institute of Technology (KTH), Stockholm,
Sweden

Nordlund E., G. Rådberg and J. Sjöberg 1997. Bergmekanikens grunder. Department of Civil
and Mining Engineering, Division of Rock Mechanics, Luleå University of Technology,
Sweden, (in Swedish)

Nordström E. 1996. Sammanställning av sprutbetongprovningar. Vattenfall Utveckling AB,

Sweden (in Swedish)

Nordström E. 2000 Swedish viewpoints on test methods for steel fibre reinforced sprayed
concrete. Letter to CEN/TC 104/WG 10/TG 3 att. Convenor Josef, January 25, 2000

Ortlepp W.D. and T.R. Stacey 1997. Testing of tunnel support: Dynamic load testing of rock
support containment system. Final project report from SIMRAC, Project No. GAP221,
August 1997.
 135

Ortlepp W.D., T.R. Stacey and H.A.D. Kirsten 1999 Containment support for large static and
dynamic deformations in mine, Proceedings of the International Symposium on Ground
Support, Rock Support and Reinforcement Practice in Mining, Kalgoorlie, Western Australia,
15-17 March, 1999, p.359 - 364

Ouchterlony F., S. Nie U. Nyberg and J. Deng 1997 Monotoring of large open cut rounds by
VOD, PPV and gas pressure measurements. Fragblast — International Journal of Blasting and
Fragmentation 1 (1997):3-25, p.5-26

Paganus T. and 0. Stephansson 1990. Projekt Oskar — Bergmekanik, slutrapport. Gruvteknik

2000, report No 90:13 (in Swedish)

Quintero C. 1998. Internal LKAB report (in Swedish)

Reynolds D.O. 1998 Mix Design and Additives. Shotcrete and its application, 25-26 February
1998. The South African Institute of Mining and Metallurgy

Sandström D. 1999 Utvärdering av genomförda bergspänningsmätningar i Kiirunavaara.

M.Sc. Thesis. Department of Civil and Mining Engineering, Division of Rock Mechanics,
Luleå University of Technology, Sweden, (in Swedish)

Savilahti T. 1995. Use of steel-fibre reinforced shotcrete at Malmberget mine. Internal LKAB
paper

SBN — S 21:64 1970 Vindlastens dynamiska inverkan. Publikation nr 31, Statens planverk (in
Swedish)

Silfwerbrand J. 1987 Effekter av differenskrypning, krypning och fogytans egenskaper på

bärförmågan hos samverkansplattor av gammal och pågjuten betong. Meddelande nr 147,
Department of Structural Engineering, Royal Institute of Technology (KTH), Stockholm,
Sweden (in Swedish)

Silfwerbrand J. 1988 Lagning av betongskador med sprutbetong, Försök med balkar under
statisk och utmattande last. Meddelande nr 153, Department of Structural Engineering, Royal
Institute of Technology (KTH), Stockholm, Sweden (in Swedish)

Sjöberg J., P. Lundman and E. Nordlund 2001. Analys och prognos av utfall i bergschakt,
KUJ 1045. internal LKAB-report (in Swedish)

Sofianos A.I., P. Nomikos and C.E. Tsourelis 1999 Stability of symmetric wedge formed in
the roof of a circular tunnel: non-hydrostatic natural stress fields. International Journal of
Rock Mechanics and Mining Sciences Vol. 36 No.5, 1999, p.687-691

Spearing A.J.S. 1998. Practical guidelines on shotcrete application Shotcrete and its
application, 25-26 February 1998. The South African Institute of Mining and Metallurgy

SS 13 72 07 Concrete testing — Hardened concrete — Compressive strength-Conversion

factors. 1988 Swedish Standard, in Swedish
 136

SS 13 72 20 Concrete testing — Hardened concrete — Compressive strength of shotcreted test

specimens. 1984 Swedish Standard, in Swedish

SS 13 72 43 Concrete testing — Hardened concrete — Adhesion strength. 1987 Swedish

Standard, in Swedish

Stephansson 0., A. Myrvang and P.Särkkä 1987 Fennoscandian Rock Stress Data Base —
FRSDB, research report, TULEA 1987:06, Luleå University of Technology, Sweden

Stiglat K., R. Linder and H. Peters 1985. Nass-spritzbeton im Dichtrom —

Betontechnologische Einflussgrössen. Beton- und Stahlbetonbau, August 1985 (in German),
p.201 to 204

Stille H., T. Groth and A. Fredriksson 1982 FEM analysis of rock mechanical problems with
JOBFEM. Swedish Rock Engineering Research Foundation — BeFo Report No BeFo 307.1/82
(in Swedish)

Stille H. 1992. Rock support in theory and practice Proceedings of the International
Symposium on Rock Support, Sudbury, Canada, 16-19 June 1992, A.A. Balkema, p.421-438

Swan G., L. Ng and R. Brummer 1996. Shotcrete as primary support system for cut-and-fill
mining. 2nd North American Rock Mechanics Symposium, Montreal, June 1996. A.A.
Balkema, p. 173-180

Svensson T. 1998 Shotcrete thickness in the Kiirunavaara mine. (internal LKAB report, in
Swedish)

Tannant D.D. and Kaiser P. 1997. Evaluation of shotcrete and mesh behaviour under large
imposed deformations. International Symposium on Rock Support — Applied solutions for
underground structures, June 22-25, 1997, Lillehammar, Norway, p.782 —792

Tannant T.T. 1997 Shotcrete performance near blasts. Geomechanics Research Centre,
Laurentian University, Sudbury, Ontario, Canada

Torsteinsson H.P. and R. Kampen 1986 High strength reinforced shotcrete, new types of
fibres and test results. Fjellsprengningstelmik, Bergmekanikk, Geotelcnikk p.34.1-34.40

Vandewalle, Marc 1994. Tunnelling the World

 Appendix 4.1-1

APPENDIX 4.1 Results - panel tests

The results from the panel tests are summarised in Table A4.1-1 and in Figure A4.1-1 to
A4.1-5.

Table A4.1-1 Test results, panels

Item No Amount of steel fibre fc F„ ,. F„,,„ oF„,a, Energy Energy
Nominal Measured (MPa) (k) (mm) (kN) (mm) (J) (Mean)
(kg/m3) (kg/m3) (-1)
Panel 11 80 1.2 165 11.0 3944
Panel 12 50 43.0 52.7 90 1.5 160 11.6 5942 4710
Panel 13 80 1.2 162 8.6 4252
Panel 21 80 0.8 204 11.4 4545
Panel 22 80 74.1 40.9 75 0.7 219 11.6 5235 4590
Panel 23 80 0.9 187 13.5 3978
Panel 31 - - 139 9.3 1350
Panel 32 Mesh Mesh 49.0 50 0.6 143 18.0 3353 2340
Panel 33 45 0.5 138 16.6 2333
Panel 41 75 1.0 217 13.6 6144
Panel 42 80 65.4 47.2 70 1.0 191 12.1 4300 4210
Panel 43 70 1.1 190 10.4 2187
Panel 51 60 0.8 153 9.5 4131
Panel 52 50 37.6 46.6 - - 156 10.1 4026 4190
Panel 53 40 0.3 169 9.4 4413

Steel Fibre Dramix 35/0.55, nominal dosage 50

kg/m3, measured dosage 43.0 kg/m3

250

-Panel 11

- - Panel 12

--- Panel 13

0 10 20 30 40 50
Mid-point deflection (mm)

Figure A4.1-1 Panel 11 to 13

 Appendix 4.1-2

Steel Fibre Dramix 35/0.55, nominal dosage 80

kg/m3, measured dosage 74.1 kg/m3
250

200
Panel 21
-2 150 -
a) - - Panel 22
g 100
u_
50 --- Panel 23

0
0 10 20 30 40 50
Mid-point deflection (mm)

Figure A4.1-2 Panel 21 to 23

Steel mesh

250

Panel 31
- - Panel 32
- Panel 33

0 10 20 30 40 50
Mid-point deflection (mm)

Figure A4.1-3 Panel 31 to 33

 Appendix 4.1-3

Steel Fibre Dramix 50/0.62, nominal dosage 80

kg/m3, measured dosage 65.4 kg/m3
250

200
Panel 41
150 -
a)
- - Panel 42
(k-2) 100 -
u_
50 --- Panel 43

0
0 10 20 30 40 50
Mid-point deflection (mm)

Figure A4.1-4 Panel 41 to 43

Steel Fibre Dramix 50/0.62, nominal dosage 50

kg/m3, measured dosage 37.6 kg/m3
250 -

200 - —Panel 51
150 - •-• Panel 52

L_ --- Panel 53
50 -

0
0 10 20 30 40 50
Mid-point deflection (mm)

Figure A4.1-5 Panel 51 to 53

 Appendix 4.2-1

APPENDIX 4.2 Yield line theory and the moment capacity of

test beams expressed as a function of angular
deformation, 0

A4.2.1 The bearing capacity of a point-loaded slab according to the yield

line theory

Figure A4.2-1 shows the yield line model used to predict failure loads. The moment mr per
unit length occurs along the radial cracks in the figure, and the moment mt (with opposite
sign) per unit length along the tangential crack lines. A set of tangential crack lines has been
assumed to devop around the circular loading plate at the centre of the slab.

L/2

o
j k•
,

Section A

Figure A4.2-1 Yield line model

 Appendix 4.2-2

The length r defined in Figure A4.2-1 is

L e
r = — cos — (A4.2.1)
2 2

and the deflection can be expressed as

L 9
(A4.2.2)
2 2

where the angle et and kis defined in Figure A4.2-1. The external and the internal work are
respectively

W = Fy 8 Fy0,(1- k) L- cos g;' (A4.2.3)

2

27r L 91
= U (A4.2.4)
2 2 2 2 2

where 09, is defined in Figure A4.2-2.

(A4.2.3) is equal to (A4.2.4) gives

mrO, (1 -k)+ m9 2(1+ k) sin

2 (A4.2.5)
0, (1 - k) cos g2
2

The relationship between Or and ei is derived with help of Figure A4.2-2 and A4.2-3.
 Appendix 4.2-3

Section A-A
Figure A4.2-2 Angular deformation «.

Figure A4.2-3 Geometrical condition

The angle e is defined as

9 8
o = = 9, cosf (A4.2.6)
d 2

where d is defined in Figure A4.2-3. The deflection d can be expressed as

L.g 0
A = 00c = 0, cos — (2 sm sin ) (A4.2.7)
2 2 22
 Appendix 4.2-4

where c is defined in Figure A4.2-3 and

L 97
b = 2—sin—cos--- (A4.2.8)
2 2 2

the angular deformation 9r is

(A4.2.9)
2 b

(A4.2.9) together with (A4.2.7) and (A4.2.8) give

9 L 9 9
, Of cos — 2 - sin - sin —
ur 2 2 2 2 -= 0 sin 9 (A4.2.10a)
2= L op 9 t 2•
2 sin cos-
2 2 2

Eq. (A4.2.10a) can be expressed as

= 20 sin 9 (A4.2.10b)
2

(A4.2.5) and (A4.2.10b) give

tan te
r (0,)(1— k) + mr (6 )(1 + k) 2
Fy = 27r m (A4.2.11)
1— k 4;1
2

where mr(0) and ?me are In,. and mt expressed as a function of angular deformation Or and a
respectively.
 Appendix 4.2-5

A4.2.2 The bearing capacity of test beams expressed as a function of

angular deformation, 6+

To calculate the bearing capacity of the slabs according to (A4.2.11) the post crack moment
capacity of the test beams m(0) as a function of angular deformation 8 is used to calculate
mr(Or) and md9d.

For a given angular deformation 8 is m,.(0) = mr (61) = m(0) (moment per unit length), and
m(8) is
h 2 1
)M (0) (A4.2.12)
Jib bb

where hp and hb is the height of the panel and the test beam, bb is the width of the test beam
and M(6) is

L Fto,(19)
M(0) (A4.2.13)
3 2

where L is the span of the four point test beam and Ft0t(8) is the total applied load on the
beam.

The height, width and length of the beams and the height of the panels were

h„ = 100 mm
hb = 75 mm
bb= 125 mm
L= 450 mm
 Appendix 4.2-6

The moment m(0) used in the analysis was calculated from a regression analysis of the
measured test results of the beam tests. Figure A4.2-4 shows the moment as a function of
angular deformation.

8-

7
E
z 6

Dramix 35/0.55 - 43.0 (kg/m3)

Dramix 35/0.55 - 74.1 (kg/m3)
Mesh
Dramix 55/0.62 - 65.4 (kg/m3)
Dramix 55/0.62 - 37.6 (kg/m3)

0 0.01 0.02 0.03 0.04 0.05

Angular deformation 0 (rad)

y=-111.76x+6.1767 Dramix 35/0.55 -43.0 (kg/m3)

R2 = 0.9761

y= -174.87x+ 9.4541 Dramix 35/0.55 -74.1 (kg/m3)

R2 = 0.9739

y= 172953x3 - 15448)(2 +478.84x + 1.1352 Mesh

R2 = 0.9978

y = -73.028x+ 6.2724
R2 = 0.9962 Dramix 55/0.62 - 65.4 (kg/m3)
y= -24.465x+ 3.8257
Dramix 55/0.62 - 37.6 (kg/m3)
R2 = 0.9382

Explanation to regression curves:

Y=ine
and
x=0

Figure A4.2-4 Moment capacity as a function angular deformation

 Appendix 6.1-1

APPENDIX 6.1 Numerical integration

The equation of motion can be written as

Mü + Cü + K(u)u = —Müg (A6.1.1)

In this thesis Eq.(A6.1.1) is solved using a linear acceleration step-by-step formulation. The
method has been described by many authors, for example Langen & Sigbjörnsson (1979) and
Gough & Penzien (1993). This presentation is mainly based on the work of Gough &
Penzien (1993). The method is an integration method and the essential concept is represented
by the two following equations

ü, =i, + fü(r)ch- (A6.1.2)

and

u, =u, + Jû(v)dr (A6.1.3)

which express the final velocity and final displacement in terms of initial values of these
quantities and an integral expression. The method with linearly varying acceleration is only
conditionally stable, if At IT 3 /7r = 0.55. However, this is seldom important in an
analysis of a SDOF system because a shorter time increment than this must be used to obtain
a satisfactory representation of the dynamic input and response. This integration method with
linear acceleration during a time increment correspond to the Newmark beta method, where ß
1/6 and y= 1/2. The constants ß = 1/6 and y= 1/2 give linearly varying acceleration and no
artificial damping respectively.

For non-linear problems it is assumed that the physical properties remain constant only for
short increments of time, therefore it is convenient to reformulate the response in terms of
incremental equations of motion. In this thesis, only the stiffness K(u) is changed by the
displacement, the mass M and the damping C of the mechanical system are constant. The
properties of the system, K(u), C and M are shown in Figure A6.1-2 a) and the forces acting
on the mass are shown in Figure A6.1-2 b). The (general) non-linear behaviour of the spring
and the arbitrary external load are shown in Figure A6.1-2 c) and Figure A6.1-2 d)
respectively. The equilibrium of forces acting on the mass at the time t = to gives

Fol + Fs9 +Fos = F0 (A6.1.4)

where F0-1 , Fo° and Fos are the inertial force, the damping force and the spring force at the
time to respectively. F, is the applied load at the time to. A short time later, a time increment
At = t,— t,, the equilibrium of forces gives
 Appendix 6.1-2

F,' + F ° F = F (A6.1.5)

b)

C Fs(t) FD(t)

u(t)
F(t) F(t)

c) d)

F(t) Initial tangent slope, K0

AFS

Average
Fs.91—
\ Slope

u(t)

Figure A6.1-2 Definition of a non-linear system, after Clough and Penzien (1993)
a) SDOF system
b) Force equilibrium
c) Non-linear stiffness
d) Applied load

Subtracting Eq.(A6.1.5) from (A6.1.4) yields the incremental equations of motion

AF' + AF D +AFs = AF (A6.1.6)

where
AF' =F11 —F01 =MAü
AFD FD FD =ceü
(A6.1.7)
AFS =F15 —F05 =K(t)Au
 Appendix 6.1-3

Substituting Eq(A6.1.7) into (A6.1.6) gives the incremental equilibrium equation

MACt + CAU + K(t)Au -= AF (A6.1.8)

In Eq. (A6.1.8) K(t) represents the average stiffness that may vary during the time increment
At as indicated in Figure A6.1-2 c). In practice, the average stiffness could only be calculated
by iteration. Therefore, the initial tangent slope

dF
K(t)= ( '), = (A6.1.9)
du
is used as stiffness in this analysis. This method to calculate properties which change with
time is quite normal according to Gough and Penzien (1993). The incremental motions are
shown in Figure A6.1-3, the acceleration, velocity and displacement during the time interval
At can be expressed as

Aü
u(r)=u0 +(—Yr
z2
2.1(7) = + idol" ( ) 2 (A6.1.10)
At

Aü 73
U(r) Uo +1:10T mo 2 ± At 6

Eq.(A6.1.10) and r = At give the incremental velocity and displacement

. At
= üo At + (A6.1.11)
2
2 &2
+ ü0 -+ . (A6.1.12)
2 6
 Appendix 6.1-4

ü(r)

ü(r)

At
Aü üo At +Aü ü
2 °

u(r)

At 2 t2
At+ü
" 2 6

Figure A6.1-3 Incremental motion based on linearly varying acceleration, after Clough and
Penzien (1993).

In general, the implicit formulation in Equation (A6.1.11) and (A6.1.12) is inconvenient to

use because iteration is required at each time increment to determine the acceleration at the
end of the step. The equations can be converted to an explicit form. If Eq. (A6.1.12) is re-
written, Aü can be expressed as

. 6 6.
= Au — 3u (A6.1.13)
At2 At

and Eq.(A6.1.11) and Eq.(A6.1.13) give

3
AU = Au — 3u, — —u,At (A6.1.14)
At 2
 Appendix 6.1-5

Substitution of Aü and Aü from Eq. (A6.1.13) and (A6.1.14) respectively, into Eq. (A6.1.8)
and with rearrangement of the terms gives the resulting incremental equilibrium equation

(A6.1.15)

in which the effective stiffness is

= 6- M + 3 C + K (A6.1.16)
At' At °

and the effective loading increment is

6 At
AT' AF + M( it, +3ii0)+ C(3ü, + iio). (A6.1.17)
At 2

A û,
When the displacement Au has been calculated according to Eq (A6.1.15) the velocity,
can be calculated according to Eq. (A6.1.14).

Summary of the procedure

1. Use the initial displacement and velocity, uo and ü, , which are known either as values
from initial conditions or as values from the preceding time increment.

2. Calculate the spring and damping forces Fos and Fop , with help of /40 and f40 and the
spring and damping coefficients, Ko and Co, to be used over the interval.

3. Generate the initial acceleration of the interval using the equation of motion in the form

= (A6.1.18)

4. Compute the effective stiffness k (Eq.6.16) and the effective loading increment AT'
(Eq.6.17).

5. Determine the displacement and velocity increment according to Eq (6.15) and (6.14)
respectively.

6. Finally, evaluate the velocity and displacement at the end of the time increment using
= Z o + Aü
(A6.1.19)
141 = U0 + Au

When step 6 has been completed, the calculation for this time increment is finished and the
analysis may be stepped forward to the next time increment.
 Appendix 6.2-1

APPENDIX 6.2 Shorter time increment for the numerical

integration

The measured acceleration time-history has a sampling time of 1/48000 s. Gough and
Penzien (1993) recommend a time increment — period ratio

&/T •1/10 (A6.2.1)

to solve the dynamic equation by the used integration method. At is the time increment and T
is the natural period. The author found that

AtIT 1/15 (A6.2.2)

gives reliable results for the analysed load cases. This means that the maximum natural
frequency, jr, of the analysed model must be less or equal to 48000/15 = 3200 Hz, (the natural
frequency, f = VT). To allow analysis of a model with higher natural frequencies, the time
increment was reduced to half of 1/48000 s, that means, 1/96000 s, with help of linear
interpolation

(iig ) j+1
(jig)j+1 / 2 - g j (A6.2.3)
2

where (Üg) j+1 / 2 (jig )i and(üg )1+, are the accelerations after Llt/2, at the beginning and at the
end of dt respectively.

Interpolated value, (iig ),±1/ 2

j -1
\ Load history (measured acceleration)
•• • •• (ii g) 1-13 (ii g) i , (jig).1+1 ,
9

dt/2 dt/2 dt/2 dt/2

dt dt At = 1/48000 (s)

Figure A6.2-1 Linear interpolation to reduce the time increment from 1/48000 (s) to
1/96000 (s)
 Appendix 6.2-2

The highest natural frequency analysed was 6200 Hz which fulfils the condition

dt/T =(1/96000)/(1/6200)< 1/15.

This procedure to reduce the time increment will (of course) not give any more information
about the load history. The only reason is to be able to analyse models with higher natural
frequencies. This procedure will not give any disturbances to the solution, which will be
shown in the following.

Numerical integration

Consider a SDOF system illustrated in Figure A6.2-2 a) exposed to the dynamic load F(t)
(Figure A6.2-2 b). The response (deflection) is calculated by numerical integration (see
Appendix A6.1). Necessary input for the analyses is shown in Figure A6.2-2 and in the
following.

a) b)

F(t)
1980 kN

u(t) Time
Duration, t = 1.48 .10-1 ms

Figure A6.2-2 Analysed model

a) Model
b) Load

Input data is

K = 6.38 •10' kNim

= 0.01 (Damping ratio)
M= 1980 kg

The natural frequency is

K
= 904 Hz.
m
The applied load is (see Figure A6.2-2) a triangular shaped load with the maximum
magnitude of 1980 kN, the duration of the load is 1.48 -10-' ms (a short impulse only 1/7.5 of
the natural period T)
 Appendix 6.2-3

In the first analysis the time increment was equal to At and in the second analysis the time
increment was equal to At / 2, see Figure A6.2-3 a) and A6.2-3 b) respectively. The time
increment At is equal to half of the duration of the load F(t) and At I T =1/15 which fulfils
the requirement in Eq.(A6.2.2).

a) b)

1980 lcN 19801(N

dt dt dt/2 dt/2 dt/2 dt/2

Duration Duration

Figure A6.2-3 Time increment for the dynamic load

a) Time increment = At
b) Time increment = At / 2

The results of the two analyses are shown in Table A6.2-1. The displacement is almost equal
for the two analyses and Figure A6.2-4 shows that the shapes of the displacement curves of
the analyses are very similar. The analyses show that the calculated displacement and the
shape of the displacement do not differ if the increment is equal to At or equal to At / 2. The
result is valid for linear elastic conditions and the time increment must fulfil the condition in
Eq. (A6.2.2).

Table A6.2-1 Calculated maximum deflection

Time increment (ms) u„,„,„ (mm) t at u„(ms)
First analysis At = 7.38 •10-2 1.251.10-2 0.36
Second analysis At/2 = 3.69-10-2 1.245 -10-2 0.33
 Appendix 6.2-4

0.015 2000
,L 1800
g•
-t 0.005 --)" i• 1400

: 0.000 — 1000 -0
ci \ -T 800 2
g. -0.005 i
.0 'il — 600
° -0.010 1 400
Time increment= At = 7.38 .10-2 (ms)
-0.0151 0 Time increment= At / 2 = 3.69 .10-2 (ms)
Applied load
0.00 0.50 1.00 1.50 2.00
Time (ms)

Figure A6.2-4 Displacement and applied load

When a failure occurs in the joints the stiffness K will decrease from KJ + K, to K,, which is a
dramatic decrease in stiffness and the natural period will be controlled by the stiffness of the
support K„ which has a natural period» dt. The response will for this reason not be
influenced if the time increment is dt or dt/2.

The result from this discussion is, that if the time increment is changed from dt to dt/2 and the
load Mi g is determined by linear interpolation according to (A6.2.2) when t = V2.At, the
response will not be significantly different for the two cases.
 Appendix 6.3-1

APPENDIX 6.3 The Duhamel integral equation

For linear elastic condition the dynamic response can be calculated with help of the Duhamel
integral (Bergan et.al.,1981)

u(t)= 1 — f F (r) sin co° — r)ch- (A6.3.1)

Mco

the dynamic load F(t) analysed in this thesis can be expressed as

F(t)= —Mü g (A6.3.2)

where üg is the measured acceleration. Eq.(A6.3.1) and (A6.3.2) give

u(t) üg sin co(t r)dr (A6.3.3)

= co-1

This shows that for linear elastic conditions, only the natural frequency and the load history
govern the displacement.
 NR: 2001:64
LULEÅ ISSN: 1402-1757
TEKNISKA 1SFtN: LTU-LIC--01/64—SE
UNIVERSITET

Utbildning
Licentiate thesis
lietitution Upptaga
Väg- och vattenbyggnad 150
Avdelning Datum
Bergmekanik 2001-11-13
Titel
Shotcrete rock support exposed to varying load conditions
Författare Språk
Lars Malmgren Engelska
Sammanfattning
Field and laboratory tests and analyses

This Licentiate thesis deals with the function of shotcrete as rock support. An extensive
failure mapping of shotcrete in the Kiirunavaara mine has been conducted to improve the
understanding of the performance of shotcrete. Furthermore, the adhesion strength
obtained for different types of scaling/cleaning methods and the relationship between
adhesion and compressive strength were investigated in field tests in the mine.

The failure mapping showed that most of the observed failures of shotcrete are in areas
with a thin shotcrete layer (<2 cm) together with a low adhesion strength. Because of that,
it is very important to have a well- cleaned surface and to get sufficient thickness all over
the shotcreted area i.e. to avoid areas with thin shotcrete. Furthermore, the growth of the
adhesion strength of shotcrete on a sandblasted concrete wall and the growth of the
compressive strength of shotcrete showed a clear correlation. The results from the field
tests showed that the adhesion strength was significantly higher on rock surfaces that had
been water-jet scaled (with 221VIPa water pressure) than on those treated by mechanical
scaling followed by cleaning of the rock surface (water pressure 0.7 MPa).

Higher rock stresses in the Kiirunavaara mine will probably require a more ductile surface
support, which will increase the use of reinforced shotcrete. Panels and beams of
reinforced shotcrete have been tested. The (cont.)

Granskare/Handledare
Erling Nordlund
GAL: http://epubliuth.seJ1402-1757/2001/64
Universitetstryckeriet, Luleå