Aalto University

School of Engineering
Degree Programme in Civil and Structural Engineering

Eero Tommila

Mining method evaluation and dilution control in

Kittilä mine

Master’s Thesis
Espoo, May 5, 2014

Supervisor: Prof. Mikael Rinne

Thesis advisor: M.Sc. André van Wageningen
 Aalto University, P.O. BOX 11000, 00076
AALTO
www.aalto.fi
Abstract of master's thesis

Author Eero Tommila

Title of thesis Mining method evaluation and dilution control in Kittilä mine
Degree programme Structural and building technology
Major/minor European mining course Code of professorship Rak-32
Thesis supervisor Prof. Mikael Rinne
Thesis advisor M.Sc. André van Wageningen
Date 05.05.2014 Number of pages 63 Language English

Abstract
Kittilä mine is a gold mine located in the Northern Finland. It began producing as an
open pit mine but nowdays all the production comes from underground mine. The
underground mining method was defined during the feasibility study with limited
information. With the experience from the underground mining and more rock
mechanical and geological data, this thesis was conducted to study if there could be
more cost effective method for mining the ore and how to reduce the dilution.
The study is divided into two parts: dilution control and mining method selection.
Dilution control was done by identifying the reasons for dilution and examining their
impact on Kittilä mine. After this, measures could be developed to decrease dilution
and improve predictability. Mining method selection was performed by doing
preliminary method selection with traditional mining method selection tools. This
yielded several methods, which applicability for Kittilä mine was assessed by terms of
safety, suitability, production, dilution & recovery and flexibility. Preliminary selection
was followed by stope design, where numerical and empirical methods were used to
find stable dimensions for stopes. Stope dimensions, different sequence patterns and
level designs were used to create mining scenarios for the deeper, undeveloped parts of
the mine. These scenarios were finally compared economically to find the most cost
effective solution. Furthermore, three different level heights were compared.
To improve the drilling accuracy it is recommended to survey all the drillhole collars
and inclination and redrill the ones that do not meet required criteria. The deviation of
the holes can be decreased by using down-the-hole hammer drills instead of top
hammer drills. The dilution prediction could be improved by gathering more rock
mechanical and geological data. The gathering of this data and utilizing it when
designing stope limits could decrease the dilution. The dilution related to overcuts and
undercuts can be prevented by systematically bolting these cuts. The proposed mining
method is cut-and-fill stoping, which technically is the same as longhole open stoping
with delayed backfill that is currently being used. The maximum width of the
longitudinal stopes should be increased from 7m to 10m. Stope length and level height
should not be changed. The sequence pattern for longitudinal stopes should be
primary-secondary-tertiary pattern with 1-4-7 sequencing. Transverse stopes should
utilize primary-secondary sequencing.

Keywords Gold, Mining, Mining method selection, Dilution control, Stope

optimization
 Aalto-yliopisto, PL 11000, 00076 AALTO
www.aalto.fi
Diplomityön tiivistelmä

Tekijä Eero Tommila

Työn nimi Kittilän kaivoksen louhintamenetelmän tarkastelu ja raakkulaimennuksen
kontrollointi
Koulutusohjelma Rakenne- ja rakennustuotantotekniikka
Pää-/sivuaine European mining course Professuurikoodi Rak-32
Työn valvoja Prof. Mikael Rinne
Työn ohjaaja DI André van Wageningen
Päivämäärä 05.05.2014 Sivumäärä 63 Kieli Englanti

Tiivistelmä
Kittilän kaivos on Pohjois-Suomessa sijaitseva kultakaivos. Tuotanto aloitettiin
avolouhintana, mutta nykyään koko tuotanto tulee maanalaisesta kaivoksesta.
Maanalaisen kaivoksen louhintamentelmä on valittu kannattavuusselvityksessä silloin
käytettävissä olleilla tiedoilla. Maanalaisesta louhinnasta saatujen kokemusten,
paremman kalliomekaanisen sekä geologisen tiedon perusteella voidaan paremmin
tutkia nykyisen louhintamenetelmän mielekkyyttä, sekä miten laimennusta voitaisiin
pienentää.
Tutkimus on jaettu kahteen osa-alueeseen: raakkulaimennuksen kontrollointiin sekä
louhintamenetelmän valintaan. Raakkulaimennuksen kontrolli on suoritettu
tunnistamalla mahdolliset syyt laimennukseen ja tutkimalla niiden vaikutusta Kittilän
kaivoksessa. Tämän jälkeen voitiin määritellä keinoja laimennuksen pienentämiseksi
ja sen arvioimisen parantamiseksi. Louhintamenetelmän valinta on tehty suorittamalla
alustava valinta perinteisillä louhintamenetelmän valintatyökaluilla. Tästä tuloksena
saatujen menetelmien käyttökelpoisuutta arvioitiin turvallisuuden, soveltuvuuden,
tuotannon tehokkuuden, laimennuksen ja saannin sekä joustavuuden suhteen.
Alustavaa valintaa seurasi louhossuunnittelu, jossa louhosten stabiilit dimensiot
pyrittiin määrittelemään numeerisilla sekä kokeellisilla menetelmillä. Saatuja
dimensioita sekä erilaisia louhintajärjestyksiä käyttäen voitiin tehdä tasosuunnitelmat
kaivoksen syväosiin, jonne ei olla vielä edetty. Näiden tasosuunnitelmien pohjalta
tehtiin useita louhintaskenaarioita, joiden taloudellisuutta verrattiin. Myös kolmea eri
tasoväliä on verrattu keskenään.
Poraustarkkuuden parantamiseksi ehdotetaan että jokaisen porareiän sijainti sekä
kaltevuus mitaittaisiin, sekä virheelliset reiät uudelleenporattaisiin. Reikätaipumaa
kyetään pienentämään siirtymällä uppovasaraporiin päältä lyövien sijaan.
Raakkulaimennuksen arviointia voitaisiin parantaa keräämällä enemmän
kalliomekaanista ja geologista dataa, sekä käyttämällä tätä dataa hyväksi louhoksia
suunniteltaessa. Ylä- ja alaperiin liittyvää raakkulaimennusta voitaisiin pienentää
systemaattisesti vaijeripulttaamalla kyseiset perät. Ehdotettu louhintamentelmä on
cut-and-fill stoping, mikä vastaa nykyistä louhintamenetelmää. Pitkittäisten louhosten
maksimileveys tulisi nostaa 7m:stä 10m:in. Louhosten pituutta tai tasoväliä ei tulisi
muuttaa. Louhosjärjetys pitkittäisille louhoksille tulisi olla kolmen vaiheen louhoksia
käyttävä 1-4-7 sekvenssi.

Avainsanat Kulta, Kaivos, Louhintamenetelmän valinta, Raakkulaimennuksen

kontrollointi, Louhosoptimointi
Foreword
This thesis was made as a last part of my master’s studies in Aalto University. Studying
and writing this thesis has been very interesting and enjoyable, mostly because of the
amazing people I have been entitled to work with and to whom I would like to express
my gratitude.
The thesis was made possible by Agnico-Eagle and Kittilä mines chief of engineering
André van Wageningen. I would like to thank André for presenting me this topic and
guiding me all the way from beginning to the very end. My thanks go to all the staff of
Kittilä mine, who were helping me with their inputs and making my stay at Kittilä mine
enjoyable and memorable. Especially I would like to thank Kyösti Huttu for providing
me with necessary information and corrections.
I would like to thank my professors Mikael Rinne, who has been very helpful and
supportive not only during this thesis but my whole studies. Mikael also told me about
European Mining Course and encouraged me to apply for this program, which resulted
in incredible year abroad, lots good friends and massive amounts of information and
experience. I would also like to thank other members of staff of the rock engineering
department and Aalto University.
I would like to thank all my friends and fellow students for making studying much more
fun and pleasant. I could have not done without you.
Lastly I would like to thank my family for supporting me in every possible way, not
only during my studies but my whole life. No words can express my gratitude.

Espoo, May 5.2014

Eero Tommila
Table of Contents
TABLE OF CONTENTS ...................................................................................... I

LIST OF FIGURES............................................................................................ III

NOMENCLATURE ............................................................................................. V

LIST OF ABBREVIATIONS .............................................................................. VI

INTRODUCTION ................................................................................................ 1
Aims and objectives ................................................................................................................................. 1
Research questions................................................................................................................................... 1

1 BACKGROUND .......................................................................................... 2
1.1 History ............................................................................................................................................... 2
1.2 Geology .............................................................................................................................................. 2
1.3 Current Operation ............................................................................................................................ 4
1.3.1 Stope Design ............................................................................................................................. 5
1.3.2 Level Design ............................................................................................................................. 5
1.3.3 Sequencing ................................................................................................................................ 5

2 MINING METHOD SELECTION ................................................................. 7

2.1 Hartman Flowchart .......................................................................................................................... 7
2.2 UBC Method...................................................................................................................................... 8
2.2.1 Input Values .............................................................................................................................. 8
2.2.1.1 Geometry of the Deposit .................................................................................................. 8
2.2.1.2 Geotechnical Parameters .................................................................................................. 9
2.2.2 Results ....................................................................................................................................... 9
2.3 Discussion on Mining Methods ........................................................................................................ 9
2.3.1 Sub-Level Open Stoping (SLOS) .............................................................................................. 9
2.3.2 Cut-and-Fill Stoping (C&FS) .................................................................................................. 10
2.3.2.1 Transverse Bench Stoping (TBS) .................................................................................. 10
2.3.2.2 Longitudinal Bench Stoping (LBS) ............................................................................... 10
2.3.2.3 Avoca ............................................................................................................................. 11
2.3.3 Vertical Crater Retreat (VCR) ................................................................................................. 12
2.3.4 Square Set Stoping and Stull Stoping ...................................................................................... 13
2.3.5 Drift-and-Fill (D&F) ............................................................................................................... 13
2.4 Comparison of Mining methods .................................................................................................... 14

3 STOPE DESIGN ....................................................................................... 15

3.1 Current Design Standards ............................................................................................................. 15
3.2 Design Parameters .......................................................................................................................... 16
3.2.1 In-Situ Stress ........................................................................................................................... 16
3.2.2 Geotechnical Parameters ......................................................................................................... 16
3.2.2.1 RQD ............................................................................................................................... 16
3.2.2.2 UCS ............................................................................................................................... 18
3.2.2.3 Joint Sets ........................................................................................................................ 18
3.3 Numerical Methods......................................................................................................................... 20
3.3.1 Input values ............................................................................................................................. 20
3.3.2 Models ..................................................................................................................................... 21
3.3.3 Results and Evaluation ............................................................................................................ 22
3.4 Empirical Methods ......................................................................................................................... 25
3.4.1 Stability Graph ........................................................................................................................ 25
3.4.2 Results ..................................................................................................................................... 26
3.5 Discussion ........................................................................................................................................ 27

I
4 DILUTION CONTROL............................................................................... 28
4.1 Definition of Dilution ...................................................................................................................... 28
4.2 Effect on Profitability ..................................................................................................................... 29
4.3 Current Situation ............................................................................................................................ 30
4.4 Reasons for Dilution ....................................................................................................................... 30
4.4.1 Drill and Blast ......................................................................................................................... 30
4.4.2 Planning................................................................................................................................... 31
4.4.3 Geotechnical ............................................................................................................................ 31
4.5 Control Measures ............................................................................................................................ 35

5 MINE DESIGN .......................................................................................... 37

5.1 Sequencing ....................................................................................................................................... 37
5.1.1 Primary-Secondary Pattern...................................................................................................... 37
5.1.2 Primary-Secondary-Tertiary Pattern (1-4-7) ........................................................................... 37
5.1.3 Center Out Pattern ................................................................................................................... 38
5.2 Level design ..................................................................................................................................... 39
5.2.1 Transverse Bench Stoping ....................................................................................................... 39
5.2.2 Longitudinal Bench Stoping .................................................................................................... 40
5.2.3 Avoca ...................................................................................................................................... 41
5.2.4 Theoretical development amounts ........................................................................................... 42
5.2.5 Implementation........................................................................................................................ 42

6 FINANCIAL APPRAISAL ......................................................................... 45

6.1 Input Values .................................................................................................................................... 45
6.2 Results .............................................................................................................................................. 46
6.3 Sensitivity Analysis ......................................................................................................................... 47

7 CONCLUSIONS ........................................................................................ 50

8 RECOMMENDATIONS ............................................................................. 52

REFERENCES ................................................................................................. 53

APPENDIX 1 – MODIFIED STABILITY GRAPH ................................................ I

APPENDIX 2 – DILUTION GRAPH BY CLARK ................................................ II

APPENDIX 3 – DILUTION GRAPH BY WANG ................................................ III

II
List of Figures
Figure 1.1 Central Lapland greenstone belt and gold deposits in the area (Ojala, 2008). 3
Figure 1.2 Longitudinal section of the deposit (Agnico-Eagle, 2013) ............................. 4
Figure 1.3 Cross section of the deposit, displaying rock types (Agnico-Eagle, 2013) .... 4
Figure 1.4 Example of level design in Kittilä mine (Agnico-Eagle, 2013) ...................... 5
Figure 1.5 Stoping sequence (Agnico-Eagle, 2013) ......................................................... 6
Figure 2.1 Hartman flowchart (Hartman, 1987) ............................................................... 7
Figure 2.2 Thickness of the ore defined by stope widths ................................................. 8
Figure 2.3 Sketch of LBS and TBS, showing undercuts, overcuts and three stopes ...... 11
Figure 2.4 Avoca method (Villaescusa & Kugantahan, 1998) ....................................... 11
Figure 2.5 Tight fill Avoca (Atlas Copco, 2007) ........................................................... 12
Figure 2.6 VCR (Atlas Copco, 2007) ............................................................................. 13
Figure 3.1 Distribution of RQD-values in ore zone, FW and HW, derived from drill
cores ................................................................................................................................ 17
Figure 3.2 Comparison of RQD-values distribution in the ore zone in top parts of Suuri
and Roura........................................................................................................................ 18
Figure 3.3Contour plot of joints with two main joint directions showing dip/dip
direction (Syrjänen, 2007) .............................................................................................. 19
Figure 3.4 Contour plot of joints (Storvall & Lindfors, 2013) ....................................... 20
Figure 3.5 Correlation between RQD and GSI, based on geotechnical logging of six
boreholes ......................................................................................................................... 21
Figure 3.6 Profiles used in analysis ................................................................................ 22
Figure 3.7 Different profile widths, colored by σ1 ......................................................... 23
Figure 3.8 Different stope lengths, colored by σ1........................................................... 23
Figure 3.9 Stresses of 25m high transverse stope, colored by σ1 ................................... 24
Figure 3.10 Relaxation zones around 25m high, 15m long and 10m wide stope........... 24
Figure 3.11 Supporting diagrams for stability graph (Clark & Pakalnis, 1997) ............ 26
Figure 4.1 Planned and unplanned dilution (Mitri, et al., 2010) .................................... 28
Figure 4.2 Definition of ELOS (Clark & Pakalnis, 1997) ............................................. 29
Figure 4.3 Graph showing relation of dilution and revenue ........................................... 30
Figure 4.4 Possible effects of poor drilling accuracy ..................................................... 31
Figure 4.5 Stope height effect on ELOS, based on mined stopes .................................. 32
Figure 4.6 Stope width effect on ELOS, based on mined stopes ................................... 33
Figure 4.7 Example of cavity scan of transverse stope limit in brown and cavity scan in
yellow (Agnico-Eagle, 2013) ......................................................................................... 33
Figure 4.8 Crosscut of stope looking north showing typical origins of dilution in
transverse stope shown in red ......................................................................................... 34
Figure 4.9 Comparison of ELOS and RQD of mined stopes ......................................... 34
Figure 4.10 Distribution of failed stopes in different stope categories .......................... 35
Figure 4.11 Crosscut of stope looking north showing cablebolt design for over- and
undercuts of transverse stope, cablebolts shown in blue ................................................ 36
Figure 5.1 Primary-secondary pattern (Ghasemi, 2012) ................................................ 37
Figure 5.2 1-4-7 pattern (Ghasemi, 2012) ...................................................................... 38
Figure 5.3 Center out pattern (Ghasemi, 2012) .............................................................. 39
Figure 5.4 Level design for TBS .................................................................................... 40
Figure 5.5 Level design for LBS, primary-secondary pattern ........................................ 40
Figure 5.6 Level design for LBS, 1-4-7 pattern ............................................................. 41
Figure 5.7 Level design for LBS, center out pattern ...................................................... 41

III
Figure 5.8 Amount of OPEX development in different scenarios ................................. 43
Figure 5.9 Ratio of paste- and rockfill in different scenarios ......................................... 43
Figure 5.10 Ratio of transverse and longitudinal stopes in different scenarios.............. 44
Figure 6.1 Comparison of production costs per gold ounce in different scenarios and
level heights .................................................................................................................... 46
Figure 6.2 Sensitivity plot .............................................................................................. 47
Figure 6.3 Effect of dilution on production costs for different level heights ................. 48
Figure 6.4 Effect of waste grade on production costs for different level heights........... 48
Figure 6.5 Effect of development meters on production costs for different level heights
........................................................................................................................................ 49

IV
Nomenclature
σH Major horizontal stress (in-situ)
σh Minor horizontal stress (in-situ)
σv Vertical stress (in-situ)
σ1 Major stress resultant (induced)
σ3 Minor stress resultant (induced)
kv Ratio of horizontal stress and vertical stress (σH/ σv)
kh Ratio of horizontal stresses (σH/ σh)
E Modulus of elasticity
v Poisson’s ratio
N’ Stability number
Q’ Tunneling quality index
mi Petrographic constant
D Disturbance factor

V
List of Abbreviations
AB Aktiebolaget – Limited Company
CLGB Central Lapland Greenstone Belt
D&F Drift-and-Fill
DC Diamond Core (Drilling)
DTH Down-the-hole
ELOS Equivalent Linear Overbreak Sloughing
ELRD Equivalent Linear Relaxation Depth
FW Footwall
GSI Geological Strength Index
GTK Geologian Tutkimuskeskus – Geological Survey of Finland
HW Hanging Wall
LBS Longitudinal Bench Stoping
Ltd. Limited Company
Oy Osakeyhtiö – Limited Company
RMR Rock Mass Rating
RQD Rock Quality Designation
SLOS Sub-Level Open Stoping
TBS Transverse Bench Stoping
UBC University of British Columbia
UCS Uniaxial Compression Strength
VCR Vertical Crater Retreat

VI
Introduction
Kittilä mine is a gold mine located in Northern Finland. It started producing in 2009 as
an open pit and gradually moved to underground mining. In 2013 all production came
from the underground mine. The underground mining method and design parameters
were defined during the feasibility study of the mine. The primary mining method is
longhole open stoping with delayed backfill utilizing transverse stopes. Longitudinal
stopes are used when mining narrow ores. Transverse stoping requires high amount of
development leading to high production costs, while longitudinal stoping has been
causing heavy dilution.
As there is now experience from underground mining and more rock mechanical and
geological data is available, it was decided to conduct a study to review the current
mining method and to examine how excessive dilution could be prevented. This thesis
is focused on improving the dilution control and finding the optimal mining method and
design parameters for the Kittilä mine.
Aims and objectives
The goal of this thesis can be expressed by the two following objectives:
• Determine measures to reduce dilution.
• Determine the most suitable mining method for Kittilä mine.
Research questions
As these objectives are still very broad and general, several research questions will be
developed in order to narrow down the focus of the study. These research questions also
help in reaching the objectives and defining the structure of the thesis.
Objective: determine measures to reduce dilution.
• What are the factors that can influence dilution?
• What is their impact in Kittilä mine?
• What can be done to prevent dilution?
Objective: determine the most suitable mining method for Kittilä mine.
• What are the possible mining methods for this kind of ore and geological
setting?
• What is their suitability for Kittilä mine?
• How do they compare economically?

1
1 Background
1.1 History
The first signs about mineralization in the area were received in 1986, during regional
gold exploration by GTK. At the same time the road from Kittilä to Pokka was under
construction and the outcrops revealed by road cuttings were examined by geologists.
Near Vuomajärvi, about four kilometers from the current deposit, geologists Jorma
Valkama and Pekka Puhakka saw visible gold in one of these outcrops. This particular
gold pocket revealed to be quite small, but it showed that the area has potential for gold.
GTK continued to investigate the area with low-altitude airborne magnetic and
electromagnetic surveying. Based on this information DC drillings were made in the
Suurikuusikko and Rouravaara area. The drillings proved that there was gold in the area
in relatively good grades. However, the gold turned out to occur mostly as refractory
gold, thus requiring more sophisticated processing than free gold. (Pankka, et al., 2006)
In 1998 the rights for the claim were sold to Swedish exploration company Riddarhyttan
Resources AB, which continued the drillings, mapping the resource and also test
mining. As the exploration continued the resources expanded and a mine seemed to
look possible. Riddarhyttan received mining permit in 2003. In 2005 Riddarhyttan was
acquired by Canadian company Agnico-Eagle Mines Ltd. The decision to construct the
mine came in 2006 and production from open pit began in 2008. (Riddarhyttan
Resources AB, 2003) (Agnico-Eagle, 2013)
1.2 Geology
The deposit is located within the central Lapland greenstone belt (CLGB), a belt of
Paleoproterozoic, volcanic and sedimentary rocks lying on top of the Archean
basement. The CLBG was formed roughly 2 billion years ago due to volcanic activity in
the area and has gone through several geological processes after that, like rifting and
faulting. The greenstone belts all over the world are known for their potential to host
gold deposits and CLBG is no different as can be seen from the Figure 1.1. CLGB is
still fairly unexplored in terms of mineral deposits, so there are still lots of possibilities
in the area. (Ojala, 2007)

2
 Figure 1.1 Central Lapland greenstone belt and gold deposits in the area (Ojala, 2008)
The Suurikuusikko deposit is located in the Kiistala shear zone. This shear zone is over
25 kilometers long and holds several individual gold bearing lodes along its length. The
largest concentration of these lodes occurs in the Suurikuusikko area. The gold is
mineralized in the shear zone by hydrothermal process. The strike of the mineralizations
varies from north to north-east, plunges to north and is dipping almost vertical (Figure
1.2 and Figure 1.3). The gold occurs mostly as refractory within arsenopyrite (75%) and
pyrite (21%), rest is found as free gold (4%). (Agnico-Eagle, 2013)

3
 Figure 1.2 Longitudinal section of the deposit (Agnico-Eagle, 2013)

Figure 1.3 Cross section of the deposit, displaying rock types (Agnico-Eagle, 2013)

1.3 Current Operation

The production began as open pit operation. The ore was mined consecutively from two
open pits, Suuri and Roura. The pits began to approach their design depth and
production moved gradually to underground mining. In 2012 the last blasts from the
open pits were made and in 2013 all production came from the underground mine. The
underground mine is currently being accessed from one portal. From the decline the two
operating parts, Suuri and Roura, are accessed with two ramps and series of footwall
drifts. The production areas are still relatively shallow (<500m) so all the ore is hauled
by fleet of trucks by the decline. The production is currently 1,1 Mt/a, but an expansion
of the processing plant is ongoing and will increase the annual processing capacity to
1,5 Mt.

4
1.3.1 Stope Design
The stopes are designed as 15m long along the strike of the ore and the width is
determined by thickness of the ore. The basic design height is 25m, however, in top
parts of the Roura 40m high stopes are being used. The stopes have two accesses:
overcut and undercut. Overcut is used for drilling, charging and backfilling while
undercut is used for mucking. All the stopes are backfilled, primary stopes with paste
fill or cemented rockfill and the secondary stopes with rockfill.
1.3.2 Level Design
The mining method currently used in Kittilä mine is longhole open stoping with delayed
backfill. This method will be referred as cut-and-fill stoping in this thesis. The ore is
accessed via footwall drifts that are developed in the waste, parallel to the strike of the
ore. The distance between the footwall drift and the ore is set to be at least 20 meters.
From these footwall drifts, overcuts and undercuts are driven into the ore, perpendicular
to strike. This is called transverse bench stoping (TBS). However, when the ore gets
thinner than 7m the stopes are done parallel to strike, leading to longitudinal bench
stoping (LBS). A typical level design can be seen in the Figure 1.4. In the figure the
drifts are displayed in blue, transverse stopes in purple and longitudinal stopes in cyan.
The figure also illustrates well the amount of development needed when mining
transverse stopes.

Figure 1.4 Example of level design in Kittilä mine (Agnico-Eagle, 2013)

1.3.3 Sequencing
The basis of the stope sequencing is bottom up primary-secondary sequencing. Mining
starts from the middle of the sill level and gradually expands both horizontally and
vertically. This sequencing is shown in the Figure 1.5 Stoping sequence. The yellow
stopes are primary and blue ones are secondary. The secondary stopes can be mined
only after all the primary stopes around them are mined and backfilled. Actual sequence
is affected by the grades and sizes of the stopes and the sequence pattern is not strictly
followed.

5
 Figure 1.5 Stoping sequence (Agnico-Eagle, 2013)
The longitudinal stopes are mined using retreat sequencing. This means that mining
starts from one end of the ore and progresses stope by stope along the strike to the other
end. In this kind of sequence every stope needs to be backfilled with consolidated fill.

6
2 Mining Method Selection
Selection of mining method is very important, since it is one of the most influential
factors in the success of the mine. Even though several methods are technically
available, they may result in much different economic performance. This is the reason
the selection should be done carefully and not just the pick one that seems practical.
During the history several tools for method selection have been developed. However,
they mostly rely on geometry and geotechnical properties of the deposit and these
methods should only be used as preliminary investigations in the selection procedure.
The final decision about the mining method to be used should be based on financial
analysis of the methods that are chosen in the preliminary phase.
In this thesis the preliminary mining method selection will be done by using flow chart
designed by Hartman (Hartman, 1987) and the UBC method (Miller-Tait, et al., 1995).
After this their applicability for the Kittilä mine will be discussed, ruling out the
unsuitable methods by comparing their safety, suitability, production rate, dilution,
recovery and flexibility.
2.1 Hartman Flowchart
Hartman developed a flowchart for defining the mining method. This chart is qualitative
and it is mainly based on the geometry of the deposit, with some reference to the ground
conditions. This method should only be used as an approach to the proper method
selection. Chart (Figure 2.1) is very quick and easy to use. As the open pit has already
been mined so only deep deposit options are being investigated. Then, applying the
geometrical properties of the deposit: tabular, steep and thin and moderate strength
yields the result of shrinkage stoping, cut-and-fill stoping and stull stoping. (Hartman,
1987)

Figure 2.1 Hartman flowchart (Hartman, 1987)

7
2.2 UBC Method
The UBC method (Miller-Tait, et al., 1995) is based on Nicholas method (Nicholas,
1981), difference being the use of RMR for ground conditions instead of RQD used in
Nicholas method and the depth of the deposit in the geometry section. UBC is
quantitative method, meaning it gives different methods points and then ranks them
according how many points they received. The points come from two categories: the
deposit geometry and the ground conditions, which are further divided into the strength
of hanging wall (HW), ore and footwall (FW). The points given go from one to five and
if the condition outrules a method, it receives -49 points. (Darling, 2011)
2.2.1 Input Values
2.2.1.1 Geometry of the Deposit
The geometry of the deposit is being defined by five factors. First, is the general shape,
which can be massive, platy-tabular or irregular. As the ore in Kittilä is in lenses, platy-
tabular is one to choose. Second is the thickness of the ore, this is a bit more
complicated since lenses close to each other can be considered as one making the ore
much thicker.
The thickness is defined by looking at the distribution of stope widths seen in Figure
2.2. Stopes are defined as 15m long sections along the strike of the ore. If the grade
inside the ore lens of such 15m section is above cutoff grade it is defined as mineable
stope. As 75% of the stopes are less than 10m wide, the ore is classified as narrow (3-
10m). Third factor is the plunge, as the ore dips almost vertically it is definitely steep.
Fourth represent the grade distribution. The gold grade can change very quickly over
relatively short distances so it can be considered as erratic. Last variable is the depth of
the deposit. The stopes have depths varying from 75m to 1175m so one single option
does not cover this, therefore two separate calculations need to be done for intermediate
(100m-600m) and deep (more than 600m).

Ore Thickness

11 % 5%

<5m
14 %
5m-7,5m
47 % 7,5m-10m

23 % 10m-15m
>15m

Figure 2.2 Thickness of the ore defined by stope widths

8
2.2.1.2 Geotechnical Parameters
The geotechnical information is divided into two parts: the RMR and the UCS of the
Ore. Both of these are applied to ore zone, HW and FW.
There currently is no available data about the RMR in the mine, but RQD values are
logged from the DC drilling. RQD is only one factor when calculating RMR, but since
no better data is available, RQD will be used to estimate the rock mass conditions. The
mean RQD values from the stopes mined so far and stopes to be mined in near future
are 52 for the ore, 47 for FW and 56 for HW. These will be classified as weak for HW
and medium for ore and HW.
The rock substance strength is defined by ratio of UCS and main principal stress. As the
stress is dependent on the depth, separate analysis needs to be done for the deeper and
shallower parts. The estimated major horizontal stresses are 40MPa for the deep part
and 20MPa for the shallower part (Ask, 2013). The UCSs are obtained from the
laboratory tests (Eloranta, 2012). These values result in weak rock substance strength in
all the zones in shallow case and very weak for deep.
2.2.2 Results
Parameters obtained were applied to a tool developed by EduMine (EduMine, n.d.), the
results can be seen in Table 2.1.

Intermediate (100-600m) Deep (>600m)

Cut-and-Fill Stoping (37) Cut-and-Fill Stoping (33)
Sublevel Stoping (27) Square Set Stoping (25)
Open Pit (26) Sublevel Stoping (21)
Shrinkage Stoping (26) Shrinkage Stoping (21)
Square Set Stoping (20) Top Slicing (16)
Top Slicing (14) Block Caving (-18)
Sublevel Caving (-21) Longwall Mining (-19)
Block Caving (-22) Open Pit (-22)
Longwall Mining (-22) Sublevel Caving (-22)
Room and Pillar (-33) Room and Pillar (-34)
Table 2.1 Results of UBC method (EduMine, n.d.)

2.3 Discussion on Mining Methods

2.3.1 Sub-Level Open Stoping (SLOS)
SLOS received good ranking in the UBC method being the 2nd and 3rd option for
intermediate depth and deep part. However, the Hartman flowchart does not consider
SLOS as option for narrow ore. SLOS is quite common method used in underground
mining, due to high production and being easily mechanized. Drilling is done in rings
from sublevels allowing usage of large mass blasts and mucking is done at the bottom
of the stopes. As the name implies, stopes are left open after they are mined out.
Naturally, this causes stability problems, which is why pillars need to be left between
the stopes. This leads to low initial recovery, which is not desired, since the ore in
Kittilä mine is highly valuable. Also, the large open stopes are prone to high dilution
especially in weak rock. The flexibility of this method is low since mining usually

9
progresses gradually from one end of the orebody to the other. Using remote controlled
LHDs allows the stopes to be non-entry zones, so SLOS can be considered a safe
method.
2.3.2 Cut-and-Fill Stoping (C&FS)
Cut-and-fill stoping proved to be best method according the UBC method and the
Hartman flowchart also identified it as viable method. C&FS is basically the same
method that is currently in use at Kittilä mine. The drilling is done from overcuts and
mucking from undercuts. In terms of mechanization and safety this method is very
similar to SLOS. However, when all the stopes are backfilled there is no need for pillars
and thus initial recovery is high. Stoping operations use mass blast which leads to
moderate dilution at best, depending on the stope size. C&FS is very flexible and is
suitable to orebodies of almost all shapes. There are mainly two variations for C&FS,
transverse bench stoping and longitudinal bench stoping.
2.3.2.1 Transverse Bench Stoping (TBS)
In TBS the production drifts are driven perpendicular to the strike of the ore. The stopes
are opened by doing a slot raise, usually near the back end of the stope and the
production then proceeds by ring blasting.
The TBS is a widely used method due to having several advantages. The planning is
easy since all the stopes can have very similar dimensions, thus simplifying the design.
The other important thing about planning is that TBS is highly flexible sequencing wise,
meaning that there are lots of possibilities in which order the stopes can be mined. The
similarity of the stopes also makes mechanization easier due to repetitive processes,
leading to high tonnages.
The most serious disadvantage of the TBS is the fact that every stope needs its own
draw points, thus requiring lots of development. This problem gets more severe when
mining narrow veins since the amount of supporting development stays the same while
stopes get smaller. The other set back with narrow ore is that all the stopes requires the
slot raise, as the stopes get smaller this starts to increase the drill meters per ton of ore
and cycle times of stopes.
2.3.2.2 Longitudinal Bench Stoping (LBS)
LBS is very similar to TBS but the production drifts are driven parallel to the strike of
the ore. This means that more of the production drifts are driven in the ore itself, thus
reducing the development in the waste. The production of the stope is done similarly as
in TBS, doing a slot raise and then blasting rings.
The LBS suits well for narrow veins, since the lenths of the stopes are not limited by the
thickness of the ore. This means possibly fewer slot raises, and less development in the
waste per ton of ore than in TBS, depending on the maximum stable stope length.
Planning in LBS is bit harder since the geometry of the stopes is more dependent on the
ore geometry, making the blast ring design more complicated. LBS is also less flexible
since stopes are accessed from adjacent stopes.

10
 Figure 2.3 Sketch of LBS and TBS, showing undercuts, overcuts and three stopes
2.3.2.3 Avoca
The Avoca mining method can be described as longitudinal retreat mining or LBS with
continuous backfill. The production is the same as in LBS, the difference is with the
backfilling. In LBS the stopes are mucked empty and then backfilled. Avoca introduces
backfilling at the same time as mining proceeds. The idea behind this is that there would
basically be no limit for stope length. As the stopes can be larger there will be less slot
raises to be made, thus making production faster and less drilling and faster cycle times.
In Avoca the stopes need to be accessed from both sides, the front end of the stope is
used for mucking, drilling and charging, while the back end is used for backfilling. This
introduces limitations to sequence as the mining progresses linearly from one end to
another. The other problem with Avoca is the huge amount of waste rock needed for
backfilling. If there is not enough waste from the underground development, a shaft
would be needed to bring waste from the surface.

Figure 2.4 Avoca method (Villaescusa & Kugantahan, 1998)

11
One modification for Avoca, named tight fill Avoca, is to keep the stopes almost fully
filled. As a result, the ore will be blasted against the backfill, called buffer blasting or
choke blasting. Blasting against the backfill causes extra dilution or possibly ore losses
as the ore and rockfill gets mixed. The dilution and recovery can be controlled by
systematically using cavity monitor scanning. However, this causes lot of work and
requires good coordination between planning and production. The advantage of keeping
the stope filled is that the open surface of the HW and FW is smaller, leading to less
sloughing and therefore less dilution from HW and FW.

Figure 2.5 Tight fill Avoca (Atlas Copco, 2007)

2.3.3 Vertical Crater Retreat (VCR)
Shrinkage stoping was placed 4th in the UBC method and was also included in the
Hartman flowchart. VCR is a modernized version of shrinkage stoping. The main
difference between the two is that in VCR drilling and charging is done from the
overcut making it much safer than shrinkage stoping, where this is done from inside the
unsupported stope. This also makes mechanization easier, thus making VCR more
productive.
In typical stoping ore is blasted in rings, extracting vertical slices from the orebody. In
VCR this is done by blasting craters to the bottom part of the ore, extracting horizontal
slices. The main benefit from this is that no slot raises are required since the void for
expansion is below the ore, undercut in the first blast and the open stope area in the
subsequent blasts. This lack of slot raises reduces the amount drillmeters needed for the
stope. Other benefit is that the blasts will be done as soon as there is enough void for
one slice to expand. The stope is emptied only after all the ore is blasted thus keeping
the time stope is open at minimum, increasing the stability.
The downside of the VCR is the complexity of blasting. Since the same drillholes are
used for every blast, it is vital for optimal performance to know exactly where to place
charges and holes need to stemmed below and above the charge. The spherical charges
used in cratering technique usually have the length of six times the hole diameter, so the
height of the extracted slice is directly proportional to the hole diameter. This is why the
diameters used in VCR are usually 140mm or 165mm.

12
The mine design of VCR is very similar to TBS, therefore they have mostly the same
strengths and weaknesses. The main difference is that VCR is not very well suited for
narrow ore due to utilization of large diameters for drilling.

Figure 2.6 VCR (Atlas Copco, 2007)

2.3.4 Square Set Stoping and Stull Stoping
Square Set Stoping proved to be the 2nd best option for the deep parts in the UBC
method, while Stull stoping appeared in the Hartman flowchart. Both of these methods
are artificially supported methods, usually with timber. Both are very labor intensive
and have low degree of mechanization, resulting in low production rates. The
advantages of these methods are that they can follow the orebody closely allowing high
selectivity and low dilution. Building timbered support requires workers to enter the
unsupported stopes making them unsafe methods.
2.3.5 Drift-and-Fill (D&F)
D&F is not mentioned in neither of the used mining method selection tools, but it is one
possible method for this type of ore. D&F is a mining method that was frequently used
before the longhole drills came popular. This method is suitable for steeply dipping
narrow veins hosting high grade ore. The production is done from drifts driven into the
ore rather than stopes. This minimizes the development in the waste since there is no
need for the footwall drifts. Although, the ore in Kittilä mine is very disseminated

13
which would require extracting lot more waste. As all the ore comes from drifting the
blast are small, which enables low dilution and high selectivity. However, the
production is lower due to less ore from single blast, usually leading to higher cost per
ton than with stoping methods. The flexibility of D&F is not very good, since there are
only few production areas.
2.4 Comparison of Mining methods
The comparison of mining methods was done in terms of safety, suitability, production
rate, dilution & recovery and flexibility. The method was considered unsafe if the
workers have to work in areas of unsupported rock. Suitability means if the method can
be effectively applied to this kind of ore. Production rate is defined by how labour
intensive the method is. Dilution refers to planned and unplanned dilution that will end
up in the mill feed while recovery means how much of the ore cannot be mined or
mucked, including possible pillars. Methods are given colors to measure their
applicability: green for applicable, red for not applicable and yellow for neutral. For
further elaboration see 2.3 Discussion on Mining Methods.

Stull
Sublevel Cut and Vertical stoping/
Drift and
open fill crater Square
fill
stoping stoping retreat set
stoping
Safety
Suitability
Production rate
Dilution & recovery
Flexibility
Table 2.2 Comparison of mining methods
Table 2.2 presents the comparison of the mining methods. From this table it can be
concluded that Cut-and-fill stoping would be the most suitable mining method for
Kittilä mine. From this point onwards, all the design is done based on this method.

14
3 Stope Design
Stope design is an important part in the mine planning since its influence on overall
success is quite influential. The basic idea is to create as large stopes as possible to
minimize the cost per ton. However, if the stopes are too large the problems with
stability can cause excess dilution, ore losses and delays, thus retarding profitability.
3.1 Current Design Standards
The current design for stopes is based on three studies made about the dimensions of the
stopes. These are the original underground geotechnical study by Piteau Associates
(Rose, 2000), Kittilä underground mine stability analysis by WSP Gridpoint (Syrjänen,
2007) and the rock mechanics study of Kittilä mine deep expansion by KMS Hakala
Oy (Hakala, 2009).
In the Piteau study it was concluded that stopes of 25m high and 30m wide should be
possible. However, even in the report itself it was noted that the parameters used in this
study do not have the sufficient accuracy to provide reliable information and that further
investigations should be done. This study was based on stability graph method, which
will be explained in more detail in part 3.4 Empirical Methods. (Rose, 2000)
The second study was performed by WSP Gridpoint in 2007. This study also relied on
the stability graph method, with stress analysis made with Examine 3D program. The
background information was obtained from geological block model, made by Gridpoint
Finland Oy. This study was investigating stopes 15m long, 12m wide and the heights of
25m, 35m and 45m. The recommendation from this report was to use 25m high stopes
below level 400. Above this level, 35m and even 45m high stopes may be possible.
(Syrjänen, 2007)
The third study was made by KMS Hakala Oy in 2009 to ensure stope stability when
proceeding deeper. Stability graph method was used, together with Examine 3D
program for stress calculations. The input parameters were obtained from drillhole data.
The stope dimensions used were 15m wide, 15m long and heights of 25m and 50m for
transverse stopes and 7m wide, 25m high and lengths of 25m and 50m for longitudinal
stopes. The recommendations were not to increase the stope dimensions. However,
increasing dimensions might be possible in good rock, but further investigations are
needed. (Hakala, 2009)
Based on these studies it was decided to use 25m sub-level height except for the top
parts of Roura where 40m is the height. The transverse stopes are designed to be 15m in
lentgh and maximum width of 35m. Longitudinal stopes have been tested with lengths
up to 60m with varying success. The optimum length is not yet determined.
All of the studies implied that the rock mechanical data available was not sufficient to
perform reliable calculations. Especially the lack of stress measurements was seen as a
major setback.

15
3.2 Design Parameters
3.2.1 In-Situ Stress
The lack of stress measurements has a high influence on creditability of rock
mechanical calculations. To overcome this problem, a study was made in 2013 in order
to get better idea about the stress field. The study was performed by Pöyry SwedPower
AB (Ask, 2013) and was done by hydraulic breaking method. The results can be seen in
Table 3.1.

Depth 375m 650m

σv (MPa) 9,0-9,7 18
σh (MPa) 9,3-10,3 13,9
σH (MPa) 12,1-24,2 26,6
σH trend 79-83° 75°
Table 3.1 Results of stress measurements (Ask, 2013)
During the testing there were difficulties due to stability problems, high water pressure
and high fracture frequency. This limited the amount of data, thus reducing the
reliability of these results. The assumptions for calculations can be seen below, where kv
is the ratio of vertical stress and major horizontal stress and kh is the ratio of horizontal
stresses.
• Gradient for σv is 0,029MPa/m
• Direction of σH is 75°
• Dip of horizontal stresses is 0º.
• kv value is 1,5
• kh ratio is 2
Vertical stress is directly proportional to the weight of the rock above so the gradient is
derived from density of 2,9t/m3. The direction of the major horizontal stress, kh ratio and
the kv value are from the measurements in the depth of 650m as this was considered
more reliable than the measurements in the depth of 375m.
3.2.2 Geotechnical Parameters
In order to calculate stope stabilities, certain parameters are required. Acquiring the
information is an important part in the design, since wrong base values will lead to
inaccurate results, rendering them unreliable if not useless. So far, very little mapping
has been done at the mine so most of the needed rock mechanical data will be acquired
from the drill cores and block model.
3.2.2.1 RQD
RQD is basically an indicator of fracture spacing and is currently the only available
information about the rock mass condition that has sufficient coverage. It has strong
influence on both Q’ and GSI, rock mass rating systems that will be used in the
calculations later on.
The RQD values for ore zone, FW and HW are derived from the block model or drill
cores. Ore zone is defined by all the stopes that are on the current life of mine and FW
and HW are 5m wide zones in both sides of these stopes. The distribution of these
values can be seen in the Figure 3.1.

16
 35%

30%

25%
Proportion
20%
ORE
15%
FW
10% HW

5%

0%
0-20 20-40 40-60 60-80 80-100
RQD

Figure 3.1 Distribution of RQD-values in ore zone, FW and HW, derived from drill cores
From the figure it can be seen that the distribution of RQDs is quite similar, FW being
bit worse than HW and ore zone. At the Table 3.2 average RQD-values for different
areas are presented, as well as portion of weak rock in these areas. Rock is defined weak
when its RQD-value is less than 40.

Average RQD/Percentage of weak rock

Suuri Roura Rimpi
Top 56/25% 55/32% 51/30%
Deep 69/8% 62/17% 80/1%
Table 3.2 RQD-values in different areas, derived from block model
The average values are moderate in top parts and good in the deep parts. The amount of
weak rock also indicates that the rock mass quality is better in the deep areas. These
values are acquired from the block model, which may cause that they are biased towards
the mean values, since smaller sections of good and poor rock are being evened out
within the blocks. This can be seen in the Figure 3.2, where the RQD-values obtained
from drill cores and block model are being compared. This figure indicates that the
portions of weak rock shown in Table 3.2 are probably lower than in reality. The
difference between average RQDs from drill cores and block model was insignificant.

17
 35%

30%

25%

20%
Drillcores
15% Block Model

10%

5%

0%
0-20 20-40 40-60 60-80 80-100

Figure 3.2 Comparison of RQD-values distribution in the ore zone in top parts of Suuri and Roura
All the values are derived from the drillcores. This may cause slight error, since the
RQD logging is done from the drillcores used for exploration purposes rather than
geotechnical cores. This kind of coring may cause some additional breaking, thus
increasing the portion of low RQDs.
3.2.2.2 UCS
Series of laboratory tests were conducted in Aalto University to find UCS, E and v for
the most common rock types of Kittilä mine. The ore zone, HW and FW consist mostly
of mafic metaigneous rocks (MML, MVX and MDY), these make up about 80%. The
last 20% consists of felsic metaigneous (FIN), metavolcanogenic sedimentary (AVS)
and metasedimentary (CHT) rocks. In the weak rock graphitic zones (GFZ) are also
present, occurring mostly in sections of about 20cm, but even zones of several meters
exists. GFZ is clearly the weakest rock with UCS of only 14,1 MPa and E of 1,4 GPa.
(Eloranta, 2012) (Agnico-Eagle, 2013)
For the modeling purpose, values for weak and moderate rock mass were estimated
using data from the drillcores from top parts of Suuri and Roura. As a result, following
values will be used:
• Moderate rock mass
o UCS = 132 MPa
o E = 69 GPa
o v = 0,25
• Weak rock mass
o UCS = 89 MPa
o E = 61 GPa
o v = 0,24
3.2.2.3 Joint Sets
From logging of six geotechnical boreholes it was noted that in most cases there were
two joint sets plus random. The categories three plus random and four or more joint sets
consisted 31% of all the samples. Near the ore zones the rock mass seems to get more
jointed, which would be logical since the ore is located in the shear zone. This data is

18
based on six boreholes, of which only four intersects the ore zone, therefore it is not
very reliable. The joint orientations can be seen in two contour plots in Figure 3.3
produced by WSP (Syrjänen, 2007) and Figure 3.4 by Itasca consultants (Storvall &
Lindfors, 2013). The plots show only low concentrations of joints but indicate that the
joints with east or west dip direction are more common. Plot made by Syrjänen also
shows increased density of vertical and almost vertical joints. Plot made by Itasca
shows no concentration of vertical joints, which may be result of plotting of only one
hole, which was also vertical. These concentrations seem logical, since they are parallel
to the shear zone in which the ore body lies.

Figure 3.3Contour plot of joints with two main joint directions showing dip/dip direction (Syrjänen,
2007)

19
 Figure 3.4 Contour plot of joints (Storvall & Lindfors, 2013)

3.3 Numerical Methods

Numerical modeling of the stopes is performed using program Examine 3D by
rocscience. This program uses boundary element method to study three dimensional
excavations. The main reason for building numerical models is the stress analysis. The
goal is to provide information about induced stresses for the stability graph method and
to determine the relaxation zones around the stopes.
3.3.1 Input values
The calculations were made for two different rock types: weak and moderate. Failure
criterion that is used is Hoek-Brown (Hoek, et al., 2002). This criterion utilizes intact
rock strength parameters and scales them with GSI, petrografic constant mi and
disturbance factor D. These factors are easier to obtain than friction angle and cohesion
used in Mohr-Coulomb failure criterion, which is why Hoek-Brown is used. Strength
parameters for intact rock were given in the section 3.2.2.2. GSI values are estimated
from the RQD values. In Figure 3.5 are RQD and GSI values from six boreholes, where
geotechnical logging has been made. From the graph a correlation between the two can
be made and is expressed as follows:
𝐺𝑆𝐼 = 0,33 ∗ 𝑅𝑄𝐷 + 37
The average RQD-value for weak rock is 14,2 and for moderate 49,8. Using the formula
above these would render to GSI values of 42 and 53. The mi is determined to be 17, a
value typical for similar rocks found in the Kittilä mine. Disturbance factor is set to 0,7
due to production blasting of the stopes.

20
 90
80
70
60
50
GSI
40
30
20
10
0
0 20 40 60 80 100
RQD

Figure 3.5 Correlation between RQD and GSI, based on geotechnical logging of six boreholes
The constants for calculating stresses were defined in section 3.2.1. As these derived
constants are not very reliable, several different values for kv and kh are studied in order
to investigate the effect of stress. The direction of major horizontal stress is
perpendicular to the strike of the longitudinal stopes and parallel to transverse ones in
all cases. Calculations will be performed to stopes at depth of 1000m, following values
will be used:

Case 1 2 3
σv 30 MPa 30 MPa 30 MPa
σH 44 MPa 60 MPa 60 MPa
σh 22 MPa 22 MPa 30 Mpa
Table 3.3 Stress resultants for numerical models

3.3.2 Models
In the analysis total of 9 different profiles were used. These are divided between three
different level heights: 30m, 25m and 20m and three different profile widths: 5m, 10m
and 15m. All the profiles used can be seen in the Figure 3.6. All the profiles were
further analyzed in three different lengths: 15m, 30m and 60m. The orientation of the
stopes was mostly done so that the strike of the stope is perpendicular to the major
principal stress, this means longitudinal stopes. For transverse stopes, 15m wide and
15m long with heights of 20m, 25m and 30m are used.

21
 Figure 3.6 Profiles used in analysis
3.3.3 Results and Evaluation
Rock quality had no visible effect on the stress distribution around stopes. This leaves
four different variables: profile width, length of the stope, height and stress
combination. Figure 3.7 and Figure 3.8 show comparison of different profile widths and
different stope lengths. There are several things to be noticed in these figures. First, the
side walls (HW and FW) become relaxed, meaning that σ3 is less than zero. Second,
highest stresses are induced in the roof and become even greater when length is
increased. Third, profile width has little effect on stresses on roof or side walls. Fourth,
the stresses in the end walls are affected only by profile width, the narrowest profile
inducing higher stresses than the rest.

22
 Figure 3.7 Different profile widths, colored by σ1

Figure 3.8 Different stope lengths, colored by σ1

In Figure 3.9 the stresses of a transverse stope can be seen. The most notable difference
is that the stress induced to the roof is much lower. Otherwise the induced stresses are
quite similar: FW and HW become relaxed while stress in the walls parallel to σH has
induced stress of same magnitude as in parallel stopes.

23
 Figure 3.9 Stresses of 25m high transverse stope, colored by σ1
The relaxation zone was defined to be areas in the rock where σ3 < 0. Predicting these
areas is important since they are very vulnerable to failures since there is no confining
stress to keep joints compressed. In the Figure 3.10 example of these relaxation zones
can be seen.

Figure 3.10 Relaxation zones around 25m high, 15m long and 10m wide stope
While analyzing the models, few features about the relaxation zones were noticed. It
was noted that the shape and size of these zones is not dependent on the magnitude of
stresses, but only their relation to each other. Furthermore, difference between major
and minor stress seemed to be more influential than the difference of major and
intermediate stress. In order to make reasonable comparison between different profiles,

24
heights and widths, term equivalent linear relaxation depth (ELRD) is being used
(Wang, 2004). ELRD can be expressed as average relaxation depth and is calculated by
following formula:
𝑉𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑟𝑒𝑙𝑎𝑥𝑒𝑑 𝑟𝑜𝑐𝑘 𝑓𝑟𝑜𝑚 𝑐𝑒𝑟𝑡𝑎𝑖𝑛 𝑠𝑖𝑑𝑒 (𝑚3 )
𝐸𝐿𝑅𝐷(𝑚) =
𝑆𝑡𝑜𝑝𝑒 𝑙𝑒𝑛𝑔𝑡ℎ (𝑚) × 𝑆𝑡𝑜𝑝𝑒 ℎ𝑒𝑖𝑔ℎ𝑡 (𝑚)
The volume of the relaxed rock mass is calculated by the program.
In terms of ELRD the most influential factor was the kv-value, then length of the stope,
height of the stope and lastly the width of the profile. The width of the profile had
surprisingly little effect since wider profiles have better geometry, in some cases the
narrowest profile gave highest ELRD.
3.4 Empirical Methods
Empirical methods are quick and easy to use and their reliability gets better with time
due to larger sample size. One of the most used and acknowledged method is the
stability graph by Mathews (Mathews, et al., 1980) which was extended by Potvin
(Potvin, 1988). This is the method that will be used in this thesis.
3.4.1 Stability Graph
Stability graph is used to examine if the wall or roof of the stope will be stable or not.
This is done by comparing the hydraulic radius of the wall or roof with the stability
number N’. The hydraulic radius can be calculated by following formula:
𝐴𝑟𝑒𝑎 𝑜𝑓 𝑠𝑢𝑟𝑓𝑎𝑐𝑒
𝐻𝑦𝑑𝑟𝑎𝑢𝑙𝑖𝑐 𝑟𝑎𝑑𝑖𝑢𝑠 =
𝑃𝑒𝑟𝑖𝑚𝑒𝑡𝑒𝑟 𝑜𝑓 𝑠𝑢𝑟𝑓𝑎𝑐𝑒
The stability number is derived from modified tunneling quality index Q’ by Barton
(Barton, 1974), which is then multiplied by factors A (rock stress factor), B (joint
orientation adjustment factor) and C (gravity adjustment factor). Criteria for these
factors can be seen in Figure 3.11.
In the analysis two different Q’ will be used: 1,4 for the weak rock mass and 6,8 for the
moderate rock mass. When defining the rock stress factor, the UCSs for these categories
are 89 MPa and 132 MPa. In most cases this results in A factors 0,1 for the roof and the
end walls except bit higher values for end walls at 500m level. As the side walls (HW
and FW) are relaxed, they receive A factor value of 0,7 (Stewart & Trueman, 2004).
The joint orientation factor is set 0,3 for all surfaces (joints parallel to surface). The
gravity adjustment factor C is 8 for the end walls and 2 for the roof and side walls.

25
 Figure 3.11 Supporting diagrams for stability graph (Clark & Pakalnis, 1997)
3.4.2 Results
Using the parameters defined in the previous chapter, the stability numbers for the
surfaces can be calculated. Results can be seen in the Table 3.4. The variation in the end
walls is due to difference in induced stresses in different in-situ stresses.

Moderate Weak
Roof 0,4 0,1
End walls 1,6-5,3 0,3-0,6
Side walls 2,8 0,6
Table 3.4 N’ for stope surfaces for moderate and weak rock mass
From the stability graph maximum hydraulic radii for these N’ values can be obtained.
The results are that all the roofs up to 10m wide are within the “stable with support” -
area and all the end walls in the “stable” –area. Hydraulic radius can be used to
calculate the possible stope lengths. Maximum stope lengths with stable side walls are
presented inTable 3.5 Maximum stable side wall lengths.

Stope Maximum stope length

Height Moderate Weak
30m 20m 12,5m
25m 23m 14m
20m 30m 17m
Table 3.5 Maximum stable side wall lengths for moderate and weak rock mass

26
Based on these results it is not suggested to increase the length of the stopes from the
current 15m. It could be done in the moderate rock mass, but even though the quality
gets better deeper, the weak rock still makes up about 13% of the stopes in deeper parts
of Suuri and Roura.
The different in-situ stress combinations have no effect on stope stabilities on 1000m
level in stability graph analysis, since even with the lowest estimations end walls and
roof receive stress factor A of 0,1 and side walls always receive 0,7.
3.5 Discussion
The stability graph is like other empirical methods, it works best on the area where most
of the cases studied are located. In stability graph most of stopes used for the study had
larger dimensions than what was examined in this thesis and were located in rock with
better quality than what is found in Kittilä mine. Numerical models do not depend on
the scale of the excavations, but they take poorly into account the effect of joints, faults
and graphitic zones.

27
4 Dilution Control
4.1 Definition of Dilution
Dilution refers to material below cutoff grade that gets blended with ore, thus reducing
the grade of excavated material. Dilution in general is impossible to avoid in stoping
due to geometries of the orebodies and it is therefore divided into planned and
unplanned dilution. Planned dilution is the waste material that is necessary to extract the
ore. The unplanned dilution is waste material that originates from outside the stope
boundaries. This can be seen in the Figure 4.1.

Figure 4.1 Planned and unplanned dilution (Mitri, et al., 2010)

Dilution can also be defined in few different ways depending on what the information is
used on. When estimating plant feed or monitoring production over longer periods of
time, it makes most sense to compare the ratio of waste and ore. However, this may not
be the best way to express dilution when evaluating the success of stoping. If comparing
only the ratio of waste and ore, even terribly failed stope can have low dilution if
collapsed material happens to be ore from adjacent stopes.
One way to quantify dilution is ELOS (Equivalent Linear Overbreak Sloughing), which
indicates the volume of material from outside the stope boundaries from FW and HW
divided by the area of particular wall, seen in Figure 4.2. This is a good way to examine
success of stoping as it is not dependent on the volume of the stope, only the surface
area of the walls. ELOS can be calculated using the following formula:
𝑉𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑜𝑣𝑒𝑟𝑏𝑟𝑒𝑎𝑘 𝑓𝑟𝑜𝑚 𝑐𝑒𝑟𝑡𝑎𝑖𝑛 𝑠𝑖𝑑𝑒(𝑚3 )
𝐸𝐿𝑂𝑆 (𝑚) =
𝑠𝑡𝑜𝑝𝑒 𝑙𝑒𝑛𝑔𝑡ℎ (𝑚) × 𝑠𝑡𝑜𝑝𝑒 ℎ𝑒𝑖𝑔ℎ𝑡(𝑚)

28
 Figure 4.2 Definition of ELOS (Clark & Pakalnis, 1997)
In this thesis several figures for dilution will be used. The definition for terms to be used
can be seen in the list below.
• Waste: material from outside the stope boundaries that is not considered as ore,
includes backfilling material. Does not include planned dilution.
• Extra ore: material from outside the stope boundaries that has sufficient grade
to be considered as ore.
• Ore loss: material from inside stope boundary that cannot be recovered.
• Dilution: percentage of waste in all excavated material.
• ELOS: All the material from outside stope boundaries from HW or FW,
includes waste and extra ore.
4.2 Effect on Profitability
The dilution affects profitability in two ways. First is the direct cost from handling more
material. Second is the loss in gold production since the capacity of the plant is fixed
and with higher dilution less ore will be processed and thus fewer ounces produced. As
the waste and ore are difficult to distinguish from each other visually, it is very likely
that all the diluting material will go through the processing plant. The relation between
dilution and profit can be seen in the Figure 4.3. This figure was calculated using
current operating costs, grades, recoveries and gold price.

29
 120,00

100,00

80,00

60,00
Profit (%)

40,00

20,00

0,00
0 10 20 30 40 50
-20,00

-40,00
Dilution

Figure 4.3 Graph showing relation of dilution and revenue

4.3 Current Situation

Currently the unplanned dilution in Kittilä mine is 19%. The average ELOS for HW and
FW is 1,2m, this consists all the material outside stope boundaries. Of all the waste
around three quarters originates from HW and FW resulting in 14% dilution. The
remaining 5% is from the stope roofs, floors, shoulders and backfilling from
neighboring stopes. Not all the material from HW and FW is waste, 13% of this is
classified as ore. However, this makes up only 21% of all the extra ore. This means that
reducing ELOS would have significant reduce in dilution.
From dilution graphs by Clark (Clark, 1998) and Wang (Wang, 2004) and the numerical
models it was estimated that the ELOS for 25m high and 15m long stope should be
0,6m. If this ELOS could be achieved it would bring dilution down to 13%. This
number was achieved by decreasing the HW and FW dilution while keeping the dilution
from other sources constant. The ratio of extra ore to waste in FW and HW is also kept
the same.
4.4 Reasons for Dilution
There are many different possible reasons for dilution and usually the dilution is result
of these reasons acting together. This makes it difficult to examine the influence of
single variable for dilution. Basically the reasons can be divided into three categories:
Drill and blast issues, errors in planning and geotechnical issues.
4.4.1 Drill and Blast
Issues in drill and blasting consist of drilling accuracy and powder factor. Drilling
accuracy is result from errors in location of collar, wrong inclination, hole deviation and
incorrect hole length. The effect of drill and blast issues to dilution is difficult to predict
since there is not enough data to compare drilling accuracy and ELOS. Also, the poor
performance in drilling can lead the wide variety of results, including poor
fragmentation, loss of blast holes, overbreak or underbreak and increase or decrease in
spacing/burden. In Figure 4.4 Possible effects of poor drilling accuracy are shown, in
blue are the planned blast holes and in red are the blast holes due to possible errors. As
these errors can and will cumulate, the symmetry of the blast is no longer what is

30
planned and may lead to dilution or ore loss. The effect of powder factor was not
investigated due to lack of information.

Figure 4.4 Possible effects of poor drilling accuracy

4.4.2 Planning
Dilution due to planning means that the stope boundaries are designed so that dilution is
almost inevitable. If the stope would have been designed differently the outcome could
still be the same in terms of volume excavated, but waste could have been included in
internal dilution. There are few aspects where this makes a difference. First is that it
distorts statistics by showing higher dilution than what should really be expected.
Second and more important is that it might make some stopes unfeasible. If the stope
boundaries are increased the grade gets lower and can fall under cutoff grade. This can
lead to situation where badly planned stope may seem feasible but results in operating
loss as it gets diluted. Decreasing the stope size can lead to ore loss, which is of course
not desired.
Stope limits are currently designed differently for transverse and longitudinal stopes. In
transverse stopes the width of the stope is defined by burden of the rings and length is
set to 15m. This kind of design leads often to box-type stopes, which may not result in
the most feasible stope economically, since the stope width is limited to intervals equal
to burden. In longitudinal stopes the boundaries of stope come from the limits of ore.
This design easily leads to stope limits that cannot be met due to geometry being too
complex, leading to dilution or ore losses. Neither of the designs takes into account the
geological formations near the boundaries such as joints, faults and shears.
It is difficult to estimate how much effect errors in planning has since distinguishing
planning errors from poor performance in drill and blast is hard.
4.4.3 Geotechnical

31
Geotechnical reasons include the influence of stresses and the condition of the
surrounding rock mass. Failures are usually caused by interaction of these two factors,
like relaxation of poor rock. Geotechnical issues goes hand in hand with planning, since
with good planning these issues can be overcome, or at least made less severe.
In the numerical modeling it was noted that stope length and height have significant
effect on the ELRD of the stope, which indicates more dilution for higher and longer
stopes. In Figure 4.5 and Figure 4.6 the ELOS of the stopes are being compared with
stope heights and widths. The ELOS in these figures are calculated by comparing the
scanned cavities and designed stope limits, example of cavity scan can be seen in Figure
4.7 From the Figure 4.5 it can be noted that as the stope height increases so does ELOS.
The ratio of ELOS for different heights is what is predicted in empirical and numerical
methods, but the values are higher. Stope lengths are not being examined since vast
majority of the stopes are 15m long and there is not enough stopes longer than this to
make reasonable comparison. However, the stope width seems to have larger effect on
ELOS than what was expected from the numerical models, this can be seen in Figure
4.6.
6,0

5,0

4,0
ELOS (m)

3,0

2,0

1,0

0,0
15 25 35 45
Height of the stope (m)

Figure 4.5 Stope height effect on ELOS, based on mined stopes

32
 6,0

5,0

ELOS (m) 4,0

3,0

2,0

1,0

0,0
0,0 5,0 10,0 15,0 20,0 25,0 30,0 35,0 40,0
Width of the stope (m)

Figure 4.6 Stope width effect on ELOS, based on mined stopes

To better understand the effects of geotechnical issues, failed HWs and FWs were
investigated. The wall was considered as failed when ELOS was over 1m. The
investigation was done by comparing stope limits and cavity scans to locate the failures.
Example of cavity scan can be seen in Figure 4.7. The failures were later compared with
RQDs from boreholes. The total number of failed walls was 60 out of 132 inspected
walls. Of the failed walls 19 were HW and 41 FW. Typical locations for failures were
related to overcuts, undercuts and backbreak; these can be seen in Figure 4.8.

Figure 4.7 Example of cavity scan of transverse stope limit in brown and cavity scan in yellow (Agnico-
Eagle, 2013)

33
Figure 4.8 Crosscut of stope looking north showing typical origins of dilution in transverse stope shown
in red
The draw points are systematically cable bolted which should decrease the amount of
dilution from FW undercut. However, there are cases where cable bolts have not been
installed or installed but missing face plates. There is no data about whether certain
undercut has been bolted or not, so no correlation between bolting and failures can be
made. The HW undercut is shotcreted, overcuts are not reinforced. Influence of over-
and undercuts could explain the ELOS difference between HW (0,9m average) and FW
(1,4m average) since the HW cuts continue shorter distance into the waste and have
smaller effect on stability.
7,0

6,0

5,0
ELOS (m)

4,0

3,0

2,0

1,0

0,0
0 20 40 60 80 100
RQD

Figure 4.9 Comparison of ELOS and RQD of mined stopes

In Figure 4.9 comparison of RQD and ELOS can be seen, there seems to be no clear
correlation between the two. However, the RQD values are from block model and

34
therefore give out the average RQD of the area and do not indicate whether there are
weak joints, fault or graphitic zones. When comparing failures and RQDs from
drillholes, it was noted that of all the failures 60% were related to RQD less than 25 and
total of 90% to RQD less than 50. These were usually short sections in drill cores
meaning that the average RQD of the wall may be good but failure occurred along joint
or fault.
100 %
90 %
80 %
70 %
60 %
50 % Stable
40 % Failed
30 %
20 %
10 %
0%
Primary Secondary Transverse Longitudinal

Figure 4.10 Distribution of failed stopes in different stope categories

Figure 4.10 shows the distribution of failed and stable stopes for primary, secondary,
transverse and longitudinal stopes. This indicates that secondary stopes are more prone
to failure than primary stopes and that transverse stopes are more prone than
longitudinal stopes. These assumptions confirm what was noticed when comparing the
stope limits and cavity scans and the numerical modeling in two ways. First, since
backfilled stopes do not convoy stresses, secondary stopes will have higher ELRD,
which most likely leads to higher ELOS. Second, longitudinal stopes do not have over-
or undercuts, which seem to be causing lot of failures. However, the sample size is still
quite small and there are lots of factors that can influence the dilution, so these might
just be coincidences. Also, longitudinal stopes have higher ore loss, which may indicate
that the deviation is affected by different planning in transverse and longitudinal stopes.
4.5 Control Measures
Drilling accuracy can be increased by systematically surveying the drillholes and re-
drilling the holes that do not meet required standards. This way the operator of the drill
will be more careful so he does not have to re-drill the holes. To decrease the deviation
of the holes one solution would be to change top hammer drill to DTH (down the hole
hammer), as they have less deviation: 5-10% for top hammer and 1% for DTH
(Corcoran, 2011). Increasing the diameter will also reduce the deviation, but this might
be problem in narrow stopes due to increase in spacing and burden.
Stope designs could be improved by taking geology and rock mechanical information
into account during the process of defining the stope boundaries. This requires
systematical mapping of joints, faults and shears near the stope boundaries and then
either adjusting the stope limits, modify drill plans or design reinforcement if needed. It
should be noted that RQD is not sufficient parameter when assessing the stability of the
stope, Q’, RMR or GSI might prove otherwise. The method of designing stope limits for
longitudinal stope should be changed so that the geometries are simpler and easier to

35
drill. In some cases this could increase the planned dilution but it would decrease ore
loss and improve the predictability and success of stoping.
Decreasing the stope size should effectively result to lower dilution. However, lower
level height will result in more development, which will be assessed in the financial
appraisal. Dividing wide stopes into to two separate stopes might have positive effect on
dilution, but would result in slower cycle times due to more slot raises required and
delays due to increase in curing time of paste needed for two stopes instead of one. This
design would also expose backfill from the first stope when the second one is being
mined, causing extra dilution from backfill, which could neglect the benefits of having
lower dilution from HW and FW.
The dilution related to over- and undercuts could be reduced by cable bolting. As the
draw points will already be bolted, extra bolting would consist of two overcuts and HW
undercut. Calculations were done by comparing the costs of cable bolting and direct
costs from the dilution. It was noticed that three rings of cable bolts for each cut would
cost less than ELOS of 0,2m, so even systematically bolting all the cuts would most
likely be feasible. This was calculated by comparing the direct costs of cablebolting and
production costs of milled material. However, this would increase the cycle time for
each stope. Example of such design can be seen in Figure 4.11. The shown stope is 10m
wide and 25m high with 7m long cablebolts in blue.

Figure 4.11 Crosscut of stope looking north showing cablebolt design for over- and undercuts of
transverse stope, cablebolts shown in blue

36
5 Mine Design
5.1 Sequencing
Sequencing is the guideline for deciding in which order the stopes should be mined. The
sequence pattern has effect mainly on four things: flexibility, stress distribution and
concentration, backfilling and draw point frequency.
5.1.1 Primary-Secondary Pattern
This is the currently used sequencing and very typical in TBS. In this design half the
stopes are primary and half are secondary. The sequencing can be seen in the Figure 5.1.
The main advantage of this design is the high level of flexibility and ability to use
unconsolidated rockfill for secondary stopes. The disadvantage is stress concentration
on secondary stopes. (Ghasemi, 2012)

Figure 5.1 Primary-secondary pattern (Ghasemi, 2012)

5.1.2 Primary-Secondary-Tertiary Pattern (1-4-7)
This pattern adds tertiary stopes to the design. The difference to primary-secondary
pattern is that more stopes can be mined at the same time. However, the stress
concentration will be even higher and both primary and secondary stopes need to be
backfilled with consolidated fill. In tertiary pattern the sequence should be strictly
followed to prevent too high lifts and too long exposure time, this of course makes the
pattern less flexible. Figure 5.2 shows one example of tertiary pattern, 1-4-7 sequence.
The name of the pattern comes from the primary stopes located in the profiles 1, 4 and
7. The numbering in the figure is based on sequencing, not profile numbers. (Ghasemi,
2012)

37
 Figure 5.2 1-4-7 pattern (Ghasemi, 2012)
5.1.3 Center Out Pattern
Center out pattern or pyramid retreat is a sequence where mining progresses stope by
stope both upwards and sideways. This means that every stope to be mined has recently
filled stope next to it, leading to two undesired things: every stope needs to be filled
with consolidated fill and the next stope can be mined only when the previous one has
been cured. To reduce cycle time of each stope, rapid curing fill can be used, but this
would increase the costs. The Avoca method does not suffer from this problem since
stopes are filled with rockfill at the same time they are advancing. Stress wise this
pattern provides best stress distribution, concentrating them evenly on the edges of the
pyramid. However, as the length of strike of backfilled stopes increase, so does the
depth of relaxation since backfilling material conveys stresses very poorly. When
increasing the length of the strike the induced stresses on the roofs also become larger,
increasing the possibility of failure. These can be prevented by leaving pillars, which
can be recovered later on. The center out pattern can be seen in the Figure 5.3.
(Ghasemi, 2012)

38
 Figure 5.3 Center out pattern (Ghasemi, 2012)

5.2 Level design

The level design will be done in two parts. First, the theoretical design to clarify the
principles of design. Second is the implantation of these designs to develop different
scenarios that can be compared. The theoretical designs are derived from the sequencing
pattern for idealized section of the orebody.
5.2.1 Transverse Bench Stoping
For TBS it does not make sense to use other sequence patterns than what is currently
used. As every stope needs its own draw point, different sequences offer no difference
to amount of development that is needed. The primary-secondary pattern will be used
due to its flexibility and possibility to backfill every other stope with rockfill. The
design can be seen in Figure 5.4, where yellow stopes are primary and blue secondary.
Each secondary stope can be backfilled with rockfill.

39
 Figure 5.4 Level design for TBS
5.2.2 Longitudinal Bench Stoping
In LBS the sequencing plays much more important role in terms of draw points. Unlike
in TBS, where draw points also serve as over- and undercuts, the over- and undercuts of
LBS are accessed through neighboring stopes. This introduces some limitations for
draw point frequency, since backfilled stopes cannot be used for accessing other stopes.
What this means is that in primary-secondary pattern draw points are developed to
secondary stopes, leading to design where every other stope needs a draw point. This
design can be seen in Figure 5.5, where yellow stopes are primary and blue are
secondary, secondary stopes can be backfilled with rockfill. If using the 1-4-7 pattern
the draw points are developed to tertiary stopes, requiring draw point for every third
stope. The design for 1-4-7 pattern can be seen in Figure 5.6, primary stopes in yellow,
secondary in red and tertiary in blue. In this design primary and secondary stopes are
backfilled with pastefill and tertiary stopes with rockfill.

Figure 5.5 Level design for LBS, primary-secondary pattern

40
 Figure 5.6 Level design for LBS, 1-4-7 pattern
In the center out pattern the draw point frequency is not limited by sequencing and in
theory the whole level could be operated through two draw points. However, this kind
of design makes little sense, due to increased mucking/hauling distance. For center out
method the distance between draw points was decided to be 60 meters, meaning every
fifth stope. This was done to keep mucking distance reasonable and to provide reliable
production. Stopes where the draw points are developed can also serve as pillars to
convoy stresses. These pillars can be recovered later since the adjacent stopes are
backfilled with pastefill. The design can be seen in Figure 5.7.

Figure 5.7 Level design for LBS, center out pattern

5.2.3 Avoca
The level design for Avoca method is basically the same as for the center out method in
LBS. The difference is that in Avoca there is no possibility for leaving pillars that can
be recovered later. This means that Avoca can be used when the strike of adjacent
stopes is limited, since leaving permanent pillars is very expensive. The length of this
strike is dependent on side wall stability.

41
5.2.4 Theoretical development amounts
In the Table 5.1, theoretical amount for development and backfilling are presented for
different sequence patterns. These values are calculated for average stope width of 10m
and distance between the footwall drift and ore of 25m.

Primary- Center
TBS 1-4-7
Secondary out
Opex development in ore 27 % 48 % 53 % 72 %
Rockfill 50 % 50 % 33 % 20 %
Table 5.1 Stastics for different patterns
When the draw point frequency is decreased and the amount of longitudinal stopes is
increased the development is significantly reduced and more of the development can be
done in the ore. However, this comes at the cost of having to use more consolidated
backfill, which is more expensive than rockfill.
5.2.5 Implementation
As the geology can change over short distances and there are parallel lenses with
varying widths, the actual level design is more complex than what was presented in the
figures earlier. In order to make reasonable comparison, level designs are made for
levels 700-1050 in Suuri and Roura orebodies. These levels were chosen since they are
the most probable locations where changes for the current mining method could be
conducted. The designs done are based on the sequence patterns presented earlier and
with longitudinal stopes of widths up to 10m and 15m. In total seven different scenarios
will be made, these can be seen in Table 5.2

Scenario 1 2 3 4 5 6 7
Max width of
7m 10m 15m 10m 15m 10m 15m
longitudinal stopes
Sequence pattern Retreat Primary-Secondary 1-4-7 Center out
Table 5.2 Scenarios used in level designs
Whenever parallel lenses occur that have less than 25m between them, pastefill is used
for the first lens to prevent the collapse of rock between lenses. In these scenarios the
whole area is considered to be one mining panel. This means that there are no sill levels
which require pastefill for the whole level. The other factor that could affect the ratio of
paste- and rockfill is the availability of waste rock. As the last stopes are mined there
will be no development done and no waste rock is produced for rockfill. The waste rock
could be provided from the surface via shafts but shaft sinking would probably be too
expensive to justify the cheaper backfill. In order to determine the sill levels or
availability of waste rock, life of mine calculations would need to be done, which is out
of scope of this study.
The main purpose of these level designs was to provide information for the financial
analysis. Since the stopes are same in every case, the only real change is in amount of
development and the ratio of paste- and rockfill. In Figure 5.8 the amount of
development in different scenarios can be seen, the development was considered to be
in the ore when it has a stope above, i.e. undercuts. This means that overcuts that do not
serve as undercut for any stope are classified as development in waste. Figure 5.9 shows

42
the ratios of paste- and rockfill. Figure 5.10 shows the ratio of transverse and
longitudinal stopes.

Figure 5.8 Amount of OPEX development in different scenarios

Figure 5.9 Ratio of paste- and rockfill in different scenarios

43
 Figure 5.10 Ratio of transverse and longitudinal stopes in different scenarios
The graphs show that the difference between the sequencing patterns is lot smaller than
what was anticipated in the theoretical design and there is more development in waste.
Especially worth noting is that center out and 1-4-7 patterns have almost the same
amounts of development while center out still requires more consolidated fill to be used.
The amount of development required seems to be closely related to ratio of transverse
and longitudinal. The ratio of paste and rockfill is more determined by the sequence
pattern used.

44
6 Financial Appraisal
6.1 Input Values
To make a reasonable comparison between the scenarios it is important that the input
values are as close to reality as possible. Table 6.1 shows constants used in calculations.
Production costs were separated to development, stoping, services, pastefill, rockfill and
processing. Values for these cost units were derived from 2014 budget.

Constant Value Source or explanation

Ore grade (g/t) 4,6 Average grade of reserves
Waste grade (g/t) 0,5 Estimation
Mining recovery (%) 84 Reconciliation of mined stopes
Processing recovery (%) 89 Cut off calculation
Dilution (%) 16 Estimation based on 0,6m ELOS
2 ramps with 350 vertical meters with 1:7
Ramp length (m) 4900
steepness
Table 6.1 Constants for financial analysis
The same cost of development is applied for all development: CAPEX, OPEX and
ramp. Dimensions for these drifts is 5m by 5m, the ore that will be recovered from
development is removed from stoping tons to avoid double calculation. No level designs
were created for 20m and 30m level heights, so amount of development is adjusted by
adding or subtracting 20%. This 20% is directly derived from the difference in level
heights. Ramp length stays the same in every case.
Stoping cost is the direct costs from mining. This consists of drilling, blasting, mucking
and hauling. Stoping costs depend on tonnage and is applied also on the waste mined.
The services costs include costs that are not directly related to mining like dewatering,
ventilation and infrastructure. These were set as fixed cost, so it is the same sum in
different level heights and scenarios rather than being dependent on tonnage.
The 16% dilution was calculated from 0,6m ELOS. The difference in dilution between
the scenarios and already mined stopes is due to different width of the stopes. The
average width for mined stopes was 12,4m while the same figure for stopes in the
scenarios is 9,3m, thus higher dilution with same ELOS. As was noted during numerical
analysis and dilution control, dilution is dependent on the level height. Therefore, it was
estimated that 30m level height will result in 20% higher dilution, while 20m level
height 15% lower. These estimations are based on results from numerical methods,
dilution graphs and the data from already mined stopes. Different sequencing and the
amount of longitudinal stopes is not assumed to have effect on dilution on these
calculations.
The cost of hauling in the analysis is included in the variable for stoping, which is
derived from the 2014 budget. This figure is calculated for hauling by truck from levels
200-500, which will be in production in 2014. The cost for hauling or hoisting from
levels 700-1050 would be different. Also, this cost is applied for all excavated material
while some of it can be used in backfilling, thus reducing the hauling costs. However,
the cost of hauling is only 3% of the total production costs, so the difference would not
be significant.

45
6.2 Results
The production cost per ounce of gold is used in comparing the scenarios. The results
can be seen in Figure 6.1. The base case is the scenario with currently used parameters
and is set to 100% production cost. All the other scenarios show the production cost in
relation to the base case.

Figure 6.1 Comparison of production costs per gold ounce in different scenarios and level heights
In general the 25m level height seems to have the lowest production costs. In 30m level
height the reduction of development meters is not enough to compensate the higher
dilution, while 20m level height has too much development to benefit from lower
dilution.
There is a clear correlation between the ratio longitudinal stopes to transverse stopes
and production cost. However, the center out pattern has more longitudinal stopes and
less development than 1-4-7 pattern but still having very similar production costs. This
is due to more pastefill required in center out pattern.
To determine the effect of using Avoca method, one more scenario was developed
based on scenario of 10m maximum longitudinal width and 1-4-7 sequence. This
scenario had Avoca applied in areas where only one ore lens was present and the strike
was long enough for it to make sense. The result of this analysis was that the Avoca
method proved to be more profitable in case the extra dilution from backfill is less than
6%. Besides this, the cycle time of the stopes should be taken into consideration when
deciding whether or not to use this method. Test mining of Avoca should be performed
as soon as possible to determine its applicability.
It must be noted that these calculations are only true for the selected area, levels 700-
1050 of Suuri and Roura orebodies. Rimpi ore body has different characteristics and
therefore might yield different results. The maximum width of the longitudinal stopes
should be the same but level height could be adjusted. The bottom part of the Rimpi ore
body has wider stopes and more competent rock. Increasing the level height here could
be profitable since most of the stopes will be transverse, thus lots of development needs
to be made. This can be seen in Figure 6.1 when comparing the level heights of the

46
scenario with 7m max width for longitudinal stopes. As the stopes get wider, the
increase in ELOS has smaller effect on dilution than with narrow stopes and the
decrease in development could lead to lower production costs.
6.3 Sensitivity Analysis
Many of the figures and constants used in the calculations are estimations or
assumptions and may have fluctuation. As these have different effects on costs and
profitability, it is essential to investigate their influence in the final figures.
Sensitivity plot seen in Figure 6.2 shows the effects of several variables on production
cost per ounce. This plot was generated for scenario with 10m maximum width for
longitudinal stopes, 1-4-7 sequence and 25m level height.
102%

101%
Production cost

Development
Stoping
100%
Rockfill ratio
Dilution
99% Grade
Waste grade
98%
-10% 0 10%
Change in variable

Figure 6.2 Sensitivity plot

From the sensitivity plot it can be seen that the grade is the most influential parameter,
followed by development, stoping and dilution. The grade is the same in every scenario
and stoping tons have very little difference.
The sensitivity plots are different for each level height, therefore some further
comparison needs to be made. Figure 6.3 shows the effect of dilution on cost per ounce
for different level heights, Figure 6.4 shows the effect of waste grade and Figure 6.5 the
effect of development meters. 20m level height is the most profitable option if the base
case dilution increases to 18%. If the waste grade gets higher than 0,81 g/t the 30m level
height becomes more profitable than the 25m. Within the reasonable fluctuation of
development the 25m level height is always most profitable. The figures are based on
scenarios with 10m maximum longitudinal stope width and 1-4-7 sequence for
longitudinal stopes.

47
 104%

103%
Production cost

102% 20m
25m
30m
101%

100%
16 17 18 19 20
Base case dilution (%)

Figure 6.3 Effect of dilution on production costs for different level heights

104%

103%

102%
Production cost

101%
20m
100%
25m
99% 30m

98%

97%
0 0,5 1
Waste grade (g/t)

Figure 6.4 Effect of waste grade on production costs for different level heights

48
 103%

102%

Production cost
101%
20m
100% 25m
30m
99%

98%
-10% 0 10%
Development

Figure 6.5 Effect of development meters on production costs for different level heights

49
7 Conclusions
Conclusions are drawn by answering the research questions that were stated at the
beginning of this thesis.
What are the factors that can influence dilution?
The factors that influence to dilution are divided into three groups: Drill and blast
issues, errors in planning and geotechnical reasons. Drill and blast issues include
inaccurate drilling and powder factor. Errors in planning consist of badly designed stope
limits and difficult geometries of the stopes. Geotechnical reasons are the effect of stress
and rock quality.
What is their (factors influencing in dilution) impact in Kittilä mine?
The effect of drill and blast could not be quantified, due to lack of information. The
possible effect of stope limit design can be seen in the Figure 4.10 in the difference
between transverse and longitudinal stopes, as they are designed differently. The effect
is that longitudinal stopes have fewer failures than transverse stopes. This difference
might also be due to lack of under- and overcuts in the longitudinal stopes. Under- and
overcuts were noticed to have an effect on dilution when comparing cavity scans and
stope limits. The effect of under- and overcuts could also explain the higher dilution
from FW than HW, since the cuts continue longer into the FW than HW. Figure 4.10
also presents the difference between primary and secondary stopes, showing that
primary stopes are more stable than secondary stopes. This indicates that the larger
relaxation zone around secondary stopes cause more dilution. Figure 4.5 and Figure 4.6
show the correlation between stope height, stope width and ELOS indicating that higher
level height and wider stopes cause higher dilution. Figure 4.9 shows that there is no
correlation between RQD and ELOS.
What can be done to prevent dilution?
The drilling accuracy could be improved by systematically surveying the collar location
and inclination of each hole and redrilling the ones that are faulty. Drillhole deviation
can be decreased by utilizing DTH drills instead of top hammers and larger drill
diameters. To improve the designing of stope limits more geological and rock
mechanical data should be collected and this data should be taken into account when
designing stope limits. Dividing wide stopes to two narrower stopes could decrease
dilution from HW and FW, but causes more dilution from backfill and slower
production. The dilution related to under- and overcuts could be decreased by
cablebolting.
What are the possible mining methods for this kind of ore and geological setting?
Preliminary mining method selection was done to find possible methods. The following
methods were suggested by traditional mining method selection tools:
• Sublevel open stoping
• Cut-and-Fill stoping
• Vertical crater retreat
• Stull stoping/square set stoping
• Drift-and-Fill

50
These methods were compared in terms of safety, suitability, production rate, dilution &
recovery and flexibility.
What is their (mining methods) suitability for Kittilä mine?
The result of comparing the methods was that cut-and-fill stoping was seen as the most
suitable method for Kittilä mine. No other methods were seen as plausible and were
thus discarded. Several scenarios based on cut-and-fill stoping method were developed
for economical comparison.
How do they (mining methods/scenarios) compare economically?
The economical comparison of these scenarios can be seen in Figure 6.1. 25m level
height proved to be the most cost effective in all scenarios except with current design
parameters, where 30m level height has slightly lower production costs. 1-4-7 and
center out sequence have almost the same production costs, which is lower than in
primary-secondary sequence. 1-4-7 is more suitable sequence since it allows much more
flexibility than center out sequence. Scenarios with higher maximum width for
longitudinal are more profitable but the stope design suggested that the width should be
limited to 10m.

51
8 Recommendations
The recommendations are done by assessing the objectives of the thesis
Determine measures to reduce dilution.
To reduce dilution following measures are suggested:
• Surveying collars and inclination of all drillholes and redrilling the ones that do
not meet the requirements
• Use of DTH hammers instead of top hammers to reduce the hole deviation.
• Gathering more geological and rock mechanical data. Q’, GSI or RMR instead
of just RQD and mapping of joints, faults, shears and weakness zones near the
stopes. and to use this data in the in stope design process.
• Systematically cablebolting of all under- and overcuts.
Determine the most suitable mining method for Kittilä mine.
The most suitable mining method for Kittilä mine was determined to be cut-and-fill
stoping, which is the one currently in use. For the design parameters for Suuri and
Roura levels 700-1050 following aspects are suggested:
• Keep the level height at 25m
• Not to increase the stope length of 15m
• Increase the maximum width of longitudinal stopes from 7m to 10m.
• Utilize 1-4-7 sequencing for longitudinal stopes and primary-secondary for
transverse stopes.
• Perform test mining of Avoca to determine its applicability and amount of extra
dilution from the backfill.

52
References
Agnico-Eagle, 2013. Kittilä: s.n.
Ask, D., 2013. Hydraulic stress measurements in boreholes STEC12107, STEC12109,
and STEC13101, Kittilän Kaivos, Northern Finland, s.l.: Pöyry SwedPower AB.
Atlas Copco, 2007. Mining Methods in Underground Mining. 2nd ed. Örebro: Atlas
Copco.
Barton, N., 1974. Engineering classification of rock masses for the design of tunnel
support. In: Rock Mechanics 6. s.l.:Springer-Verlag, pp. 189-236.
Clark, L., 1998. Minimizing dilution in open stope mining with a focus on stope design
and narrow vein longhole blasting, s.l.: University of British Columbia.
Clark, L. & Pakalnis, R., 1997. An empirical design approach for estimating unplanned
dilution from stope hanginwalls and footwalls, Montreal: Canadian Institute of Mining
and Metallurgy.
Corcoran, P., 2011. In the hole (ITH) vs. top hammer drilling in underground
applications. [Online]
Available at: http://www.miningcongress.com/pdf/presentations-downloads/Cubex-
Peter-Corcoran.pdf
[Accessed February 2014].
Darling, P., 2011. SME Mining Engineer Handbook. 3rd ed. USA: SME.
EduMine, n.d. [Online]
Available at: http://www.edumine.com/tools/mining-method-selection/
[Accessed 2013].
Eloranta, P., 2012. Kittilän kaivos, Kivinäytteiden lujuuskokeet, Espoo: Aalto
University.
Ghasemi, Y., 2012. Numerical studies of mining geometry and extraction sequencing in
Lappberget, Garpenberg, Luleå: Luleå University of Technology.
Hartman, H. L., 1987. Introductory Mining Engineering. 1st ed. New York: Wiley.
Hoek, E., Carranza-Torres, C. & Corkum, B., 2002. Hoek-Brown failure criterion -
2002 Edition. [Online]
Available at: http://www.rockscience.com/assets/files/uploads/7715.pdf
[Accessed December 2013].
KMS Hakala Oy, 2009. Rock Mecahnics Study of Kittilä Mine Deep Expansion, s.l.:
KMS Hakala Oy.
Mathews, K. E., Hoek, E., Wyllie, D. C. & Stewart, S. B., 1980. Prediction of stable
excavation spans for mining depths below 1,000 meters in hard rock. Ottawa:
Department of Energy and Resources.
Miller-Tait, L., Pakalnis, R. & Poulin, R., 1995. UBC mining method selection. In:
Mine planning and equipment selection International symposium; 4th, Mine planning
and equipment selection. s.l.:s.n., pp. 163-168.
Mitri, H. S., Hughes, R. & Lecomte, E., 2010. Factors influencing unplanned ore
dilution in narrow vein longitudinal mining, Phoenix: SME.

53
Nicholas, D. E., 1981. Method selection. In: D. Stewart, ed. A Numerical Approach,
Design and Operation of Caving and sublevel Stoping Mines. New York: SME-AIME,
pp. 39-53.
Ojala, J., 2008. Geologian tutkimuskeskus. [Online]
Available at:
http://www.gtk.fi/export/sites/fi/_system/PressReleases/kuvat/geol_yleistetty_color_juh
ani_lehdisto_1.pdf
[Accessed September 2013].
Ojala, J. V., 2007. Gold in the Central Lapland Greenstone Belt, Espoo: Geological
Survey of Finland.
Pankka, H., Keinänen, V. & Valkama, J., 2006. Kaarestunturin kultakonglomeraateista
Suurikuusikon kultaesiintymään - Ilkka Härkösen panos kultageologina. Geologi, pp. 4-
8.
Potvin, Y., 1988. Empirical open stope design. s.l.:University of British Columbia.
Riddarhyttan Resources AB, 2003. Riddarhyttan Resources AB, Press Release. [Online]
Available at:
http://en.gtk.fi/export/sites/en/informationservices/explorationnews/2003/rr_040203.pdf
[Accessed October 2013].
Rose, N. D., 2000. Underground geotechnical design study, North Vancouver: Piteau
Associates Engineering Limited.
Stewart, P. C. & Trueman, R., 2004. Quantifying the effect of stress realaxation on
excavation stability. Mining Technology, Volume 113, pp. A107-A117.
Storvall, E. & Lindfors, U., 2013. Analysis of Shaft Stability at the Kittilä Mine, Luleå:
Itasca Consultants AB.
Syrjänen, P., 2007. Kittilä underground mine stability analysis, Helsinki: WSP
Gridpoint.
Wang, J., 2004. Influence of Stress Undercutting, Blasting and Time on Open Stope
Stability and Dilution, Saskatoon: University of Saskatchewan.
Villaescusa, E. & Kugantahan, K., 1998. Backfill fo Bench Stoping Operations.
Brisbane: s.n.

54
Appendix 1 – Modified stability graph

Modified stability graph (Clark & Pakalnis, 1997)

I
Appendix 2 – Dilution Graph by Clark

Dilution graph by Clark (Clark, 1998)

II
Appendix 3 – Dilution graph by Wang

Dilution graph by Wang (Wang, 2004)

III