department of civil, environmental and natural resources...
TRANSCRIPT
LICENTIATE T H E S I S
Department of Civil, Environmental and Natural Resources EngineeringDivision of Mining and Geotechnical Engineering
Rock Mass Behavior and Cap Rock Stability at the Malmberget Mine
Sraj B. Umar
ISSN 1402-1757ISBN 978-91-7583-097-1 (print)ISBN 978-91-7583-098-8 (pdf)
Luleå University of Technology 2014
Sraj B. U
mar R
ock Mass B
ehavior and Cap R
ock Stability at the Malm
berget Mine
ISSN: 1402-1757 ISBN 978-91-7583-XXX-X Se i listan och fyll i siffror där kryssen är
LICENTIATE THESIS
Rock Mass Behavior and Cap Rock Stability at the Malmberget Mine
Sraj Banda Umar
Luleå University of Technology Department of Civil, Environmental and Natural Resources Engineering
Division of Mining and Geotechnical Engineering SE-971 87 Luleå
SWEDEN
Printed by Luleå University of Technology, Graphic Production 2014
ISSN 1402-1757 ISBN 978-91-7583-0 - (print)ISBN 978-91-7583-0 - (pdf)
Luleå 2014
www.ltu.se
iii
Preface
The work presented in this thesis was carried out at the Luleå University of Technology, Division of Mining and Geotechnical Engineering. The research work is part of the ongoing project named "Ground deformations around the orebodies in the Malmberget mine", which is supported by the Hjalmar Lundbohm Research Centre (HLRC) and Luossavaara-Kiirunavaara AB (LKAB).
My thanks go to my two supervisors: Adjunct Professor Jonny Sjöberg and Senior Lecturer Catrin Edelbro for their valuable advice and guidance throughout the project. I would also like to thank Professor Erling Nordlund for his guidance and previous supervision work.
I would like to thank the staff of the LKAB Malmberget Mine for their continuous contributions during this research, and in particular Thomas Savilahti, Jimmy Töyrä, Fredrik Ersholm, Linda Jonsson and Joel Andersson. I would also like to thank the core shed staff at LKAB Malmberget, that were very instrumental in making sure that the core logging process went on smoothly.
Last but not least, I would like to thank my family for their sacrifices during the period of this research, my wife Bridget and daughter Marianne-Sara.
November 2014
Sraj Umar Banda
iv
v
Abstract
The LKAB Malmberget mine is mined using sublevel caving. This mining method requires continuous caving of the hangingwall, which also result in ground deformations on the surface. The Malmberget mine is located in the municipality of Malmberget; thus well planned residential relocations have been required as the caving area has expanded and associated ground deformations developed.
This thesis aims to bring an increased understanding of the stress redistribution and likely failure mechanisms in the hangingwall of the Printzsköld orebody, one of the orebodies currently in production in Malmberget. Rock mass investigations have been conducted to gain more understanding of the rock mass. These investigations included core logging, strength testing, borehole photography, joint mapping underground, and rock mass characterization in the Printzsköld orebody. The investigations showed that the rock mass rating (RMR) for Printzsköld was in the interval of 55 – 78 and 40 – 76 respectively, indicating a fair to good rock mass, albeit with some scatter and weak units present. The work also showed that there were three main joint sets in the rock mass. A number of large-scale structural features have been inferred in a different study, and the implications of their presence were further studied in this thesis work.
Conceptual continuum and discontinuum numerical analysis was conducted for the Printzsköld orebody, in order to study potential failure behavior of the hangingwall and cap rock, as well as to study the possible effects of the inferred large-scale structures near the orebody. The results from the continuum analyses indicated high stress build-ups in the crown pillar and stress relaxation in the hangingwall of the orebody. Both shear and tensile failure were noted in the hangingwall and the cap rock. As mining proceeded to the level of 1225 meters, the stresses in the cap rock increased and the relaxation zone also increased. A parametric study of the strength parameters was conducted, which showed that lowering the cohesion had a larger effect on the stress build-ups in the cap rock and destressing in the hangingwall than lowering other parameters such as tensile strength and friction angle.
Based on the analyses, a potential failure mechanism for the hangingwall was presented. The hangingwall of the Printzsköld orebody can be considered as a beam fixed at the crown pillar and the cave bottom. Shear failure in the cap rock has the potential to cause failure in the hangingwall and thus, the "beam" is lengthened and a progressive mechanisms developing.
The discontinuum model results indicated that the presence of large-scale structures reduced the stress build-up in the crown pillar and that slip developed along these structures. A reduction in the angle of friction for the structures resulted in more slip, compared to a reduction in cohesion. The presence of large-scale structures did not affect the yielding pattern in the rock mass or the far field stress redistribution in the hangingwall and cap rock. Recommendations for future studies have been made for a better understanding of the rock mass behavior in this orebody as mining progresses.
Keywords: Strength parameters, hangingwall failure, cave propagation, failure mechanism, large-scale structures, crown pillar
vi
vii
Table of Contents
1. Introduction ......................................................................................................................... 1
1.1. Problem statement ...................................................................................................... 1
1.2. Scope and Objectives ................................................................................................. 2
1.3. Approach ...................................................................................................................... 3
1.4. Thesis outline ............................................................................................................... 3
2. Mining-Induced Ground deformation ............................................................................. 5
2.1. Continuous surface deformation .............................................................................. 5
2.2. Discontinuous surface deformation ......................................................................... 8
2.3. Deformation zones in sublevel caving ..................................................................... 9
2.4. Effects of geological structures. .............................................................................. 13
2.5. Numerical analysis methods applied to caving ..................................................... 13
3. Rock mass characterization of the Fabian and Printzsköld orebodies ...................... 15
3.1 The geology of the Malmberget deposit. ............................................................... 15
3.2 Rock mass characterization...................................................................................... 15
3.3 Results from rock mass characterization ............................................................... 17
3.3.1 Intact rock strength from point load testing ..................................................... 17
4. Numerical modeling and stress analysis of the Printzsköld orebody ........................ 25
4.1. Numerical analyses of the Printzsköld orebody ................................................... 25
4.1.1 Modeling approach ............................................................................................... 25
4.1.2 Model geometry ..................................................................................................... 26
4.1.3 Evaluation planes .................................................................................................. 28
4.1.4 Initial stresses and material parameters. ............................................................. 30
4.2. Model results .............................................................................................................. 32
4.2.1. Stress redistribution .............................................................................................. 32
4.2.2. Yielding in the rock mass and slip along the structures .................................. 36
5. Discussion ........................................................................................................................... 41 6. Conclusions ........................................................................................................................ 43 7. Recommendations for future research ........................................................................... 45 8. References ........................................................................................................................... 47
viii
Paper A
Paper B
Paper C
ix
List of papers
This thesis comprises the following journal papers:
PAPER A
Umar, S.B., Sjöberg, J. & Nordlund, E. 2013, "Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody of the Malmberget Mine, Sweden", Journal of Earth Sciences and Geotechnical Engineering, vol. 3, no. 4, pp. 147-173.
PAPER B
Umar, S.B., Sjöberg, J. & Savilahti, T. 2014, "Modeling of caving and deformation mechanisms of the hangingwall of the Printzsköld orebody in the Malmberget Mine.". Submitted to: The International Journal of Mining, Reclamation and Environment.
PAPER C
Umar, S.B. and Edelbro, C. 2014, "Influence of large-scale structures on the stability of the hangingwall in a caving mine - a modeling study". To be submitted to an international journal.
x
List of Tables
Table 4.1 Strength parameters used for elastic-perfectly plastic models (Sjöberg, 2008) ......................... 31 Table 4.2 Strength parameters used in the model as well as variations of parameters. ............................. 31 Table 4.3 Strength parameters for the large scale structures varied for sensitivity analysis in the Printzsköld orebody. ........................................................................................................................................... 32 Table 4.4 Numerical modeling results showing the calculated slip (in brown) on the three structures with the various strength parameters after mining the 1225 m level. Viewed parallel to the strike of the Printzsköld orebody. ........................................................................................................................................... 39
List of Figures
Figure 1.1 Schematic view of the orebodies at Malmberget Mine. (Courtesy of LKAB: www.lkab.com)1 Figure 1.2 Aerial view of the Malmberget Mine site. The colored objects are the horizontal projections of the various orebodies in the Malmberget mine, and the numbers following them are the current mining levels for which these orebodies are shown. The Printzsköld orebody can be seen at the center and the red patch shows the crater on the ground surface from the Fabian orebody. The ground surface is between the 150 and 220 m level in the mine coordinate system. ................................................ 2 Figure 2.1 A generalized cross section of surface deformation. Reference here is made with a flat-lying coal seam (Fonner, 1987). ..................................................................................................................................... 6 Figure 2.2 Cross section of a smooth surface deformation profile, showing various surface deformation parameters in (Brady and Brown, 1993). ..................................................................................... 7 Figure 2.3 Elements of displacement in a surface deformation zone (Shadbolt, 1975) .............................. 8 Figure 2.4 Effects of faults in the smooth surface deformation zone (Bell and Donnelly, 2006). ............ 9 Figure 2.5 Ground deformation categories in a caving zone (Gilbride et al., 2005) .................................. 10 Figure 2.6 Caving zones and behavior associated with hangingwall stability (Villegas, 2011). ................ 11 Figure 2.7 The conceptual model of the caving zone as proposed by Sainsbury (2012), based on Duplancic and Brady (1999). .............................................................................................................................. 12 Figure 3.1 Geological map of the Northern Norrbotten (Martinsson and Hansson, 2004). ................... 15 Figure 3.2 Map showing drill hole locations (red circles) with white arrows showing the azimuths of the drill holes. Also shown are the horizontal projection of drifts on the 880 m (Fabian) and 945 m (Printzsköld) mining levels. ................................................................................................................................ 16 Figure 3.3 Comparison of UCS from laboratory and the point load index (Is(50)) ...................................... 17 Figure 3.4 Intact Rock Strength isotropy in the Printzsköld orebody area, (a) and (b) represent the two drillholes (PRS01 and PRS02) from which tested samples were obtained. ................................................. 18 Figure 3.5 Printzsköld orebody RMR distribution.......................................................................................... 19 Figure 3.6 Rock core showing a zone of weakness as indicated by rock mass alteration, decomposition and weathering. .................................................................................................................................................... 22 Figure 3.7 Weak zone interval distribution along boreholes (Umar, et. al, 2013) ...................................... 22 Figure 4.1 Illustration and setting of three large-scale structures in the Printzsköld orebody shown relative to the 920 m mining level (Wettainen, 2010): (a) plan view, and (b) a longitudinal cross-section. ................................................................................................................................................................................ 25
xi
Figure 4.2 Two-dimensional model of the Printzsköld orebody (vertical cross-section with an east-west strike in the model axes). .................................................................................................................................... 26 Figure 4.3 Model geometry for the model of the Printzsköld orebody ....................................................... 27 Figure 4.4 Model set-up showing large scale structure locations in the Printzsköld orebody and the model extents........................................................................................................................................................ 28 Figure 4.5 Evaluation planes in the longitudinal view of the Printzsköld orebody. Plane L2 is about 250 m away from plane L1, while plane L3 is about 700 m away. ....................................................................... 29 Figure 4.6 Evaluation planes in the vertical cross-sectional view of the Printzsköld orebody ................ 30 Figure 4.7 Calculated major principal stress using the 3DEC and Phase2 numerical models, shown along a line traced along evaluation plane C1 in the hangingwall of the Printzsköld orebody and monitored on evaluation plane L1. The 0 m point is the starting point at the toe of the cave and the 1200 m is at the ground surface. .............................................................................................................................................. 33 Figure 4.8 Major principal stress calculated in the two-dimensional (Phase2) model. ................................. 34 Figure 4.9 Major principal stress calculated using a three-dimensional continuum model (3DEC) and shown on a longitudinal cross-section on the evaluation plane L1. ............................................................. 35 Figure 4.10 Maximum principal stress calculated using a three-dimensional discontinuum (3DEC) model and shown on the evaluation plane L1 after mining the 1225 meter level. ..................................... 36 Figure 4.11 Yielded elements of the hangingwall of the Printzsköld orebody calculated using the two-dimensional (Phase2) model after mining down to the mining level of 1225 m. ......................................... 37 Figure 4.12 Failure state calculated using three-dimensional continuum (3DEC) model and shown on the evaluation plane L1 for the base case values and for mining to the level of 1225 m. ......................... 37 Figure 4.13 Calculated slip area on the modeled large-scale structures for the three different strength cases simulated in the model of the Printzsköld orebody and for mining to the 1225 m level. .............. 38 Figure 5.1 Failure mechanism forecasted for the Printzsköld orebody for continued mining toward depth. ..................................................................................................................................................................... 42
xii
1
1. Introduction
1.1. Problem statement Caving methods are known to cause mine induced ground deformation. These deformations can take various forms starting from subtle deformations such as surface cracking to caves with volumes of large magnitudes. The Malmberget mine is a sublevel caving mine owned and operated by the Luossavaara-Kiirunavaara AB (LKAB) mining company. This iron ore mine comprises a total of 20 orebodies of which about 10 are in production today, see Figure 1.1. The total annual production from the mine amounts to 18 million metric tons of iron ore (LKAB, 2014).
Figure 1.1 Schematic view of the orebodies at Malmberget Mine. (Courtesy of LKAB: www.lkab.com)
The mine is located in the municipality of Gällivare about 100 kilometers north of the Arctic Circle; see Figure 1.2, which shows the position of the Malmberget Mine. Over the years, extraction from these orebodies has resulted in varying amounts of ground deformation on the surface. This surface deformation is associated with the mining method which undermines the hangingwall causing instability and ground deformations. As a result of this, re-location of residential areas and existing infrastructure has been required in portions of the Malmberget municipality. As mining progresses to deeper levels a larger surface deformation area is created. Caving to surface and formation of cave craters have occurred for several of the orebodies in Malmberget, e.g., Kapten and ViRi.
2
Figure 1.2 Aerial view of the Malmberget Mine site. The colored objects are the horizontal projections of the various orebodies in the Malmberget mine, and the numbers following them are the current mining levels for which these orebodies are shown. The Printzsköld orebody can be seen at the center and the red patch shows the crater on the ground surface from the Fabian orebody. The ground surface is between the 150 and 220 m level in the mine coordinate system.
Forecasting the caving processes in a caving mine is not straight-forward as it can be rapid or slow depending on the rock mass conditions, stress conditions, the mechanisms at work, etc. Making reliable ground surface deformation prognosis for non-daylighting orebodies is even more difficult. With several non-daylighting orebodies in the Malmberget mine, these difficulties are of particular importance. The mechanical behavior of the rock mass in the Malmberget deposit is also complex and requires a better quantification so that the caving mechanisms can be understood better.
1.2. Scope and Objectives
The main objective of this research is to get a better understanding of the rock mass behavior and stress distribution around the Printzsköld orebody, and to quantify potential failure mechanisms and controlling factors. This will be achieved through investigations of the stress redistributions in the hangingwall and cap rock of the Printzsköld orebody as mining deepens, and to determine their effects on the stability of the hangingwall.
The Printzsköld orebody is located in the central area of the Malmberget mine and was chosen to be the case study orebody. This orebody was chosen for two reasons: (i) it is one of the more important orebodies in terms of future production volumes in Malmberget, and (ii) it is located partly beneath the central area of the Malmberget municipality with a potentially large effect on re-location of residences and infrastructure as caving progresses. So far, there are no signs of mining-induced ground
3
deformation in the Printzsköld area, but a reliable prediction of future caving and associated ground movements is warranted.
This orebody offers a unique opportunity to study rock characteristics and evaluate the rock mass stability in relation to caving. This is so because as a non-daylighting orebody, it has developed a cave which is expected to progress upwards. While in other orebodies this process has already developed, the Printzsköld provides an opportunity for early investigations of the processes leading to caving of the ground surface. Rock mass characterization was also conducted for the Fabian orebody, but no additional analyses conducted as part of this study.
Lack of readily available geotechnical information on the rock masses in the Malmberget mine, especially for the Printzsköld orebody was one of the major limitations in this research. Another limitation was the lack of access to upper levels (in the orebody) for drift mapping as these areas had already been mined out. Groundwater conditions have not been considered in this study, but are judged to be of less significance for the failure mechanisms associated with the cave mining. Most of the analyses presented in this work were developed for hard rock mining conditions to hard rock mining, and may thus not be applicable to environments in which soft rocks are encountered.
1.3. Approach
The approach employed to achieve the aims of this study was two-staged: (i) rock mass characterization and (ii) numerical analysis to assess stress redistribution effects as mining deepens, and to gain understanding of the failure mechanisms of the exposed hangingwall and cap rock.
The investigations carried out for the rock mass characterization were focused on rock mass parameters that enabled the use of different rock mass classification systems. Geological structures and rock mass strength were investigated and presented. Discontinuity analyses were conducted using core logging, tunnel mapping and borehole photography (optical televiewer).
Two types of models were run for the Printzsköld orebody: continuum and discontinuum. A parametric analysis was conducted to study the effects of change in strength parameter values on stress redistribution and yielding. The discontinuum models were analyzed to simulate and investigate the effect of large-scale structures on stress redistribution and yielding of the hangingwall and the cap rock of the Printzsköld orebody.
1.4. Thesis outline
This thesis work has been organized in chapters as described below:
Chapter 1: This chapter includes the introduction of the study and the description of the problem this study seeks to solve.
Chapter 2: Literature review of the relevant processes that take place in a caving zone has been described here. It also includes descriptions of the effects of large-scale structures on
4
the cave shape, geometry and propagation. Literature on numerical analyses of selected cases has been presented.
Chapter 3: Descriptions of rock mass characterization methods and results in the Printzsköld orebody has been presented.
Chapter 4: Numerical analyses of the Printzsköld orebody stress redistributions have been presented. A methodological approach and results have also been presented.
Chapter 5: Discussion of the results, interpretations and implications has been presented. The results from this work have been presented in three journal papers and a technical report, see Table 1.1. However, only the three journal papers have been appended to this thesis.
Chapter 6: Conclusions of the study have been presented.
Chapter 7: Recommendations for future research have been presented
Table 1.1 Journal papers included in this thesis
Publication Title Journal name Author Paper A Rock Mass Characterization
and Conceptual Modeling of the Printzsköld Orebody of the Malmberget Mine, Sweden
Journal of Earth Sciences and Geotechnical Engineering, vol. 3, no. 4, 2013, 147-173
Umar, S.B., Sjöberg, J. And Nordlund, E.
The first author collected field data, conducted laboratory research; carried out data results interpretation, conducted rock mass characterization, conducted rock mass classification and a major part of the numerical analysis. The first author also wrote the paper.
Paper B Modeling of caving and deformation mechanisms of the hangingwall of the Printzsköld orebody in the Malmberget Mine
Submitted to The International Journal of Mining, Reclamation and Environment
Umar, S.B., Sjöberg, J. And Savilahti, T.
The first author conducted numerical analysis and wrote the paper. Paper C Influence of large-scale
structures on the stability of the hangingwall in a caving mine - a modeling study
To be submitted to an international journal
Umar, S.B. and Edelbro, C.
The first author conducted numerical analysis and interpretations of results and also wrote the paper.
5
2. Mining-Induced Ground deformation
Ground deformation induced by underground mining can be of varying categories. Two types of ground deformation have been recognized; continuous and discontinuous surface deformation. Continuous surface deformation is the smooth lowering of the ground surface making a trough-like profile with a major lowering at the center of the trough known as the maximum subsidence. Discontinuous surface deformation is characterized by large scale displacements over limited areas, evidenced by cracks, steps, sinkholes and/or chimneys (Brady and Brown, 1985). Cave mining produces a cave in the subsurface which, depending on the rock mass and presence of structures, can rapidly propagate to the surface forming a sinkhole. The material above the void caused by mineral extraction is called the cap rock. The stability of this cap rock depends on span, extraction rate and rock mass properties.
2.1. Continuous surface deformation
Continuous surface deformation is the smooth lowering of ground, associated with continuous mining methods mostly longwall mining method, which is suitable for flat lying orebodies. In longwall mining, the mined orebody leaves an empty space called the goaf and when there is failure of this goaf through convergence under stress, the overlying surface strata are affected. Whether the resulting effects will be seen on the surface or not, depends on the mined out height H, depth of the goaf, the extent or width W of the mined out area and the rock mass conditions, see Figure 2.1.
The area affected by surface deformation is usually larger than the mined out width. The extents of surface deformation are defined by using various angles measured from the corners of the working area to the surface. Extensive work was done by the National Coal Board (NCB) of the United Kingdom in defining various parameters of the surface deformation zone. They carried out investigations of many cases within the United Kingdom and reported these in the NCB’s Subsidence Engineer’s Handbook (1975), (Brady and Brown, 1985), see also Figure 2.2. Each point in the subsidence zone also undergoes vertical and horizontal displacement. The subsidence profile undergoes a differential displacement in these directions and this generates horizontal compressive or tensile strain, tilt and curvature. The tilt in the subsidence zone is the derivative of the vertical displacement with respect to the horizontal, while horizontal strain (compressive or tensile strain) is the derivative of the horizontal displacement with respect to the horizontal.
6
Figure 2.1 A generalized cross section of surface deformation. Reference here is made with a flat-lying coal seam (Fonner, 1987).
The subsidence profile undergoes a differential displacement in the vertical and horizontal directions and this generates horizontal compressive or tensile strain, tilt and curvature. The tilt in the surface deformation zone is the derivative of the vertical displacement with respect to the horizontal, while horizontal strain (compressive or tensile strain) is the derivative of the horizontal displacement with respect to the horizontal. These differentials in displacement have a tendency to disrupt surface terrain and cause damage to buildings, more so when faults are reactivated (Donnelly et al., 2008).
7
Figure 2.2 Cross section of a smooth surface deformation profile, showing various surface deformation parameters in (Brady and Brown, 1993).
Shadbolt (1975) discussed the elements of displacements as shown in Figure 2.3. Some of these elements can be used to detect the onset of surface deformation damage in a mining environment. In their study of ground tilt in relation to surface deformation in longwall mining, Whittaker and Pasamehmetoglu (1981) suggested that ground tilt can be used to detect discernible effects of mining surface deformation on the surface structures (Whittaker and Pasamehmetoglu, 1981).
8
Figure 2.3 Elements of displacement in a surface deformation zone (Shadbolt, 1975)
2.2. Discontinuous surface deformation
When the surface deformation trough has been affected by geological structures, discontinuous ground deformation may occur. This type of deformation is known as abnormal subsidence (Shadbolt, 1975). Structures such as faults that have been reactivated may also result in abnormal subsidence according to Shadbolt (1975). Figure 2.4 presents a pictorial view of the effects of faults in the profile of the trough surface deformation.
9
Figure 2.4 Effects of faults in the smooth surface deformation zone (Bell and Donnelly, 2006).
This kind of surface deformation is associated with a number of mining methods including all caving methods, such as block caving, panel caving and sublevel caving. It may also develop suddenly or progressively (Brady and Brown, 1985). Tension cracks and chimneys are usually visible on the surface in the caving and deformation zones.
2.3. Deformation zones in sublevel caving
Sublevel caving is a caving mining method appropriate for steeply dipping orebodies. Once drilled and blasted the rock flows under gravity and it is drawn from the draw levels or production levels. As mining progresses downwards, continued draw from the sublevels ensures continued gravity flow of the broken ore. Depending on the fragmentation and cavability of the rock mass, in addition to ore drawing sequence, this may leave an open cavity called a cave. This mining method has a disadvantage of creating ground deformation zones as mining progresses. Ground deformations that occur in block caving are quite similar to those caused by sublevel caving. Gilbride, et al., (2005) suggested that in a block caving mining operation ground deformation can be categorized in zones as shown in Figure 2.5.
10
Figure 2.5 Ground deformation categories in a caving zone (Gilbride et al., 2005)
Lupo (1996) has described deformations zones found in a caving zone due to sublevel caving, in relation to the effects on the hangingwall as well as the cap rock. In this case the deformation zones have been categorized as caved zone (CZ); fractured zone (FZ); and continuous deformation zone (CDZ) (Lupo, 1998). Villegas (2011) illustrated these deformation zones defined by (Lupo, 1998) as shown in Figure 2.6.
11
Figure 2.6 Caving zones and behavior associated with hangingwall stability (Villegas, 2011).
These deformation zones can be described as follows:
In the caved zone a downward movement of caved material develops, due to collapse of material from the side walls and back of the cave.
The fractured zone is characterized by tension cracks, steps, fractures and sinkholes or pits distributed randomly in the caving zone. In this zone it is expected to find unstable parts as failure through toppling and shear can occur.
The continuous deformation zone is characterized by the development of continuous deformations.
Duplancic and Brady (1999) suggested a conceptual model for the caving zone, in which various regions of the zone were categorized with respect to their seismic, displacement and yielding responses. This categorization identified 5 zones, see Figure 2.7.
12
Figure 2.7 The conceptual model of the caving zone as proposed by Sainsbury (2012), based on Duplancic and Brady (1999).
The descriptions of the different zones in Figure 2.7 are (Duplancic and Brady, 1999; Sainsbury, 2012; Sainsbury, et al., 2011a):
Elastic zone
In this zone the rock mass surrounds the caving zone and behaves mostly elastically, with properties consistent with an undisturbed rock mass.
Seismogenic zone
In this region there is a concentration of seismic activities. The active seismic front that occurs here results from slip on geologic structures and failure of intact rock. This is caused by the redistribution of stresses during mining and cave advance.
The yielded zone
This is the zone in which rock has failed and cannot provide support to the overlying rock mass. It is characterized by large scale displacements of rock masses, which is subjected to significant damage.
The air gap
This is space left between the mobilized zone and yielded zone. The air gap is controlled by rate of draw from below and the bulking porosity of the broken material. If the rate of draw is very low the material may fill the cave and further yielding and cave advance upwards may be stopped.
13
Furthermore, at this point the material in the cave may provide support to the walls of the cave (Villegas, 2011).
The mobilized zone
This zone is characterized by dislodged or fallen rock blocks from the cave back. The rock usually has displacements of more than 1 - 2 m (Sainsbury, et al., 2011a).
2.4. Effects of geological structures.
The influence of geological structures has also been discussed in Sainsbury et al. (2008); Sainsbury et al. (2009); Sainsbury (2010) and Lupo (1997). In these studies it was clear that geological structures influence rock mass deformation, and the failure patterns in a caving zone. Sainsbury et al. (2011a) used a continuum condition model of FLAC3D to study the influence of large-scale structure on caving using a ubiquitous joint model approach. It was found that the orientation of large-scale structures was responsible for the cave advance direction.
In their study to investigate the influence of faults on the cave geometry, Vyazmensky et al. (2010) found the structures to be influential in the cave propagation. The direction of the cave advance was also highly influenced by the orientation of the large scale structures. It was found that joint orientation had a very important role in controlling the direction of the cave propagations towards the surface, the location of the cave breakthrough and the mechanism of the near surface rock mass failure. Furthermore, the effects of faults on surface deformation were evaluated through a series of models with an assumption that the fault was dipping towards the cave. Two different jointing conditions were used while the setting of the fault was varied through several scenarios. In these scenarios it was found that the fault location and inclination also had major influence on the surface deformation propagation. For various extractions, comparisons of surface deformation developed for various fault inclinations and varying joint set orientations were conducted. It was found that irrespective of jointing orientation caving induced failure was predominantly controlled by the plane of weakness provided by the fault.
2.5. Numerical analysis methods applied to caving
Improvements in computing power has made it possible to process and analyze rock mass behavior more accurately through numerical modeling as compared to previous analytical methods. Software has been developed, which can simulate the rock mass and its geologic structures. These software are built on principles of finite element methods (e.g. Phase2 (Rocscience, 2008)), finite difference methods (e.g. FLAC and FLAC3D (Itasca, 2009)) and discrete element methods (e.g. 3DEC (Itasca, 2013)). The above mentioned computer programs have been successfully used in simulating rock mass behavior in caving environments. For instance, Board and Pierce (2009) used the continuum model of FLAC3D to study the effects of joint orientations on the caving process. In Sainsbury (2010) FLAC3D was also used to determine the limits of the continuous surface deformation zone and the onset of the stable
14
zone. Phase2 has been used in studies such as Villegas and Nordlund (2008). Vyazmensky et al. (2010) used a hybrid finite/discrete element technique in the program ELFEN (Rockfield, 2004) and adopted a coupled elasto-plastic fracture mechanics constitutive criteria to simulate the behavior of the rock mass under the influence of structures such as faults and joints.
Often, rock mass properties are estimated using intact rock test results coupled with rock mass characterization. However, this empirical approach has the disadvantage that strength values cannot be fully validated for the full scale rock mass. Moreover, such an approach does not fully consider the influence of structures on a large scale. A synthetic rock mass modeling (SRM) methodology was developed, aiming to overcome these problems (Mas Ivars, 2010, Mas, et al., 2007, Mas, et al., 2008, Pierce, et al., 2007). Presented in Mas et al. (2007), the SRM approach uses the Itasca code PFC3D to represent the intact rock as an assembly of bonded particles, and an imbedded discrete fracture network (DFN) to represent the in-situ joints (Mas Ivars, et al., 2011).
In the SRM methodology, the intact rock mass is represented by the bonded particle model (BPM) in which the rock mass is represented as rigid circular two-dimensional or spherical three-dimensional particles bonded together at their contacts by parallel bonds (Mas Ivars, 2010). The joints are represented using the smooth-joint contact model (SJM) proposed in (Mas Ivars, et al., 2008). The behavior of the jointed rock mass is determined by the in-situ joint network within the rock mass. The discrete fracture network (DFN) has been considered to overcome the problem of representing the jointed rock mass and its behavior patterns as it can explicitly simulate the in situ joint fabric (Dershowitz and Einstein, 1988). DFNs have also been used in other studies in which jointed rock masses were represented (Board and Pierce, 2009, Vyazmensky, et al., 2010).
Building on (Mas, et al., 2007, Mas, et al., 2008, Pierce, et al., 2007, Sainsbury, et al., 2008); Sainsbury et al. (2009) presented a rock mass modeling technique in an effort to analyze the caving behavior of a jointed rock mass. In this study the SRM was used to derive the scale factors that were subsequently used with the ubiquitous joint rock mass model (UJRM) to analyze caving. This overcame the problems associated with computing power and time taken if the SRM was to be used to simulate caving for a mine-scale model. Stress paths were analyzed in which close agreements were found between those from the SRM and the UJRM.
15
3. Rock mass characterization of the Fabian and Printzsköld orebodies
3.1 The geology of the Malmberget deposit.
The Malmberget deposit is a pale Proterozoic succession of greenstones, porphyries and clastic meta-sediments which are hosted by metavolcanics that have been intruded by pegmatites and granites (Martinsson and Hansson, 2004), see Figure 3.1. The volcanic rocks have been transformed to sillimanite gneisses with quartz, muscovite, and local andalusite by the young granite intrusions. The iron ores are characterized by coarse magnetite and variable horizons of apatite with local sections rich in hematite.
Generally, the border zones of the ore are characterized by skarn zones interpreted to be related to the formation of the ore. Similar to many areas in Northern Sweden, the Malmberget deposit is characterized by NW-SE trending shear zones (Romer, 1996). These zones are thought of as resulting from a complex geodynamic evolution which included repeated extensional and compressional tectonic regimes associated with magmatic and metamorphic events (Skiöld, 1988).
Figure 3.1 Geological map of the Northern Norrbotten (Martinsson and Hansson, 2004).
3.2 Rock mass characterization
The importance of rock mass characterization in rock construction or mine design cannot be overemphasized. To adequately estimate the parameters needed to input into numerical analysis, a
16
comprehensive rock characterization practice is required. It is an integral part of mine design or rock construction using empirical approaches for estimation of the rock properties.
Many rock mass classification systems have been established over the years such as Terzaghi’s rock classification (Terzaghi, 1964), Rock Quality Designation (RQD) (Deere and Miller, 1966), Rock Mass Rating (RMR) classification (Bieniawski, 1989), the Quality index (Q) (Barton, 1974), Rock mass Index (RMi) (Palmstrøm, 1996), and the Geological Strength Index (GSI) (Hoek, et al., 2002). To adequately arrive at the appropriate results in rock mass design it is advisable to apply at least two rock mass characterization systems (Bieniawski, 1989).
Detailed rock mass characterization has been conducted in the Printzsköld and Fabian orebodies, and results presented in Umar et al. (2013). Four diamond drill holes were drilled from the ground surface, two from each orebody as shown in Figure 3.2. The drillholes were drilled at an average inclination of 70° and they averaged a length of 250 m from the ground surface. Several investigations were conducted, including core logging, point load testing, direct uniaxial compressive strength tests in the laboratory, and borehole photography (optical televiewer). Underground mapping was also conducted in which joints were investigated in the lower parts of the Printzsköld orebody (since the drillholes covered only the upper parts)
Figure 3.2 Map showing drill hole locations (red circles) with white arrows showing the azimuths of the drill holes. Also shown are the horizontal projection of drifts on the 880 m (Fabian) and 945 m (Printzsköld) mining levels.
17
3.3 Results from rock mass characterization
Even though rock mass characterization was conducted in the Printzsköld and Fabian orebodies, only results from the characterization of the Printzsköld orebody are presented in this thesis. This is because this orebody was the focus of the study. The results from the Fabian orebody can; however, be found in Banda (2013).
3.3.1 Intact rock strength from point load testing
Intact rock strengths were derived indirectly from point load testing. These strength values were usedto design a complete classification of the rock mass using the Bieniawski (1989) rock mass classification. According to the ISRM (1987), laboratory uniaxial compressive strength varies by a factor of 20-25 times the point load values at 50 mm diameter of the core (Is(50)). Figure 3.3 is a plot of the laboratory values of the UCS and the point load index Is(50) .
Figure 3.3 Comparison of UCS from laboratory and the point load index (Is(50))
The plot indicated a factor of 21, making final UCS values determined from point load tests to be factored by this value. Ratios of axial to diametral strength test results were plotted to determine the skewedness of strength isotropy in the four drill sites. Figure 3.4 represents the strength isotropy for the Printzsköld orebody. It can be seen that ratios of nearly unit were found; representing a relatively isotropic rock strength system in the entire area drilled.
UCS = 20.682Is(50)
0.0
50.0
100.0
150.0
200.0
250.0
300.0
350.0
0.0 5.0 10.0 15.0 20.0
UCS Vs Is50 plot
UCS Vs Is50 plot
Linear (UCS Vs Is50 plot)
18
Figure 3.4 Intact Rock Strength isotropy in the Printzsköld orebody area, (a) and (b) represent the two drillholes (PRS01 and PRS02) from which tested samples were obtained.
Three main joint sets were found in the hangingwall of the Printzsköld orebody. The joint conditions are summarized as follows:
Moderately weathered, predominantly rough planar joint, hard infilling, some soft infill found in some joints.
Joint spacing of 10 to 50 cm.
Predominantly hard kaolinitic infills.
Joint orientations (dip/dip-direction) were: (i) 27°/346°, (ii) 13°/097°, and (iii) 72°/173°
Rock mass rating for the Printzsköld orebody was calculated using the RMR (Bieniawski, 1989) system and was found to be in the range 55 – 78 see Figure 3.5. The geological strength index (GSI) was calculated from data from tunnel mapping of rock masses in the hangingwall of this orebody, see Table 3.1a and Table 3.1b.
00.20.40.60.8
11.21.41.61.8
15.6
31.8 53
67.8
78.6
98.5
113.
3
142.
2
159.
5
173.
3
188.
9
204
216.
6
Axi
al:D
iam
etra
l R
atio
Depth (m)
PRS01 strength isotropy ratio (axial:diametral)
PRS01ISOTROPYRATIO
0
0.5
1
1.5
2
2.5
6.5
29.1 54
72.5
91.5
119.
113
8.5
157.
517
8.75
198.
521
7.7
237.
225
8.1
279.
2
Axi
al:D
iam
etra
l R
atio
Depth (m)
PRS02 strength isotropy ratio (axial:diametral)
PRS02ISOTROPYRATIO
(b) (a)
19
Figure 3.5 Printzsköld orebody RMR distribution
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0 20 40 60 80 100
Prob
abili
ty D
istrib
utio
n
RMR
RMR Distribution - Printzsköld Orebody Area
RMR Distribution -Printzsköld Orebody Area
20
Tabl
e 3.
1a R
ock
char
acte
rizat
ion
resu
lts fo
r the
Prin
tzsk
öld
oreb
ody.
GE
OT
EC
HN
ICA
L C
OR
E L
OG
GIN
G R
ESU
LT
S
Roc
k F
orm
atio
n
Pri
ntz
sköl
d O
reb
ody
RQ
D [
%]
RM
R
UC
S [M
Pa]
Max
M
in
Av.
M
ax
Min
A
v.
Max
M
in
Av.
RLE
Re
d le
ptite
97
24
71
78
54
66
30
2 60
18
4 G
LE
Gre
y le
ptite
94
28
65
69
53
63
24
2 90
14
9
GRL
G
rey
red
lept
ite
93
46
69
70
60
67
256
120
176
SKN
Sk
arn
90
65
85
67
65
66
170
74
127
MG
N
Mag
netit
e 79
68
73
67
62
65
18
2 71
12
7 G
RA
Gra
nite
Co
mm
ent
Whe
re re
sults
are
not
pre
sent
it m
eans
the
drill
hol
e di
d no
t int
erce
pt th
e pa
rticu
lar ro
ck ty
pe.
GE
OT
EC
HN
ICA
L T
UN
NE
L M
AP
PIN
G R
ESU
LT
S (P
RIN
TZ
SKÖ
LD
OR
EB
OD
Y)
Roc
k F
orm
atio
n
945
Met
er L
evel
97
0 M
eter
Lev
el
GSI
Ran
ge
Ave
rage
C
ond
itio
n
Des
crip
tion
G
SI
Ran
ge
Ave
rage
C
ond
itio
n
Des
crip
tion
RLE
Re
d le
ptite
55
– 7
5 G
OO
D
Roug
h sli
ghtly
wea
ther
ed ir
on st
ained
Bl
ocky
55
- 70
G
OO
D
Bloc
ky, s
light
ly w
eath
ered
iro
n st
aine
d
GRL
G
rey
red
lept
ite
24 –
65
FAIR
V
ery
bloc
ky m
oder
atel
y w
eath
ered
30
- 50
FA
IR
Mod
erat
ely
wea
ther
ed a
nd
alter
ed su
rfac
es
MG
N
Mag
netit
e
20 –
60
GO
OD
Bl
ocky
schi
stoc
ity, s
light
ly w
eath
ered
40
– 6
5 G
OO
D
Roug
h an
d sli
ghtly
wea
ther
ed
BIO
Bi
otite
sc
hist
M
assiv
e, fo
liate
d.
25 –
40
POO
R
Hig
hly
wea
ther
ed su
rfac
es
and
slick
en si
des j
oint
s, se
amy
21
Tabl
e 3.
1b R
ock
char
acte
rizat
ion
resu
lts fo
r the
Prin
tzsk
öld
oreb
ody.
RO
CK
MA
SS A
ND
JO
INT
DE
SCR
IPT
ION
S (P
RIN
TZ
SKÖ
LD
OR
EB
OD
Y)
Roc
k fo
rmat
ion
R
ock
core
des
crip
tion
Join
t C
ond
itio
ns
Roc
k m
ass
des
crip
tion
s
Join
t R
ough
nes
s Jo
int
infi
llin
g
Join
t ap
ertu
re a
nd
d
escr
ipti
on
RLE
Re
d lep
tite
The
inse
t sh
ows
the
high
bre
akag
e in
red
lep
tite,
(Malm
berg
et, 2
011)
Pl
anar
, ro
ugh
undu
latin
g in
plac
es
Mos
tly n
o in
fillin
g, h
ard
infil
ling
in
plac
es
Ave
rage
se
para
tion
1.0
mm
, m
oder
ately
w
eath
ered
jo
ints
Form
ed t
he l
arge
st c
ompe
tent
par
t of
the
ove
rlyin
g ro
ck st
ruct
ure
in th
e ar
ea.
Pink
ish t
o re
d an
d br
own,
fin
e to
med
ium
gra
ined
, sil
icifie
d in
man
y pl
aces
. It
sho
ws
tend
ency
for
brit
tlene
ss,
hard
and
she
ared
es
peci
ally
along
con
tact
zon
es. 2
to 3
join
t set
s pr
esen
t in
this
form
atio
n
GRL
Gre
y
red
leptit
e
Char
acte
rized
by
dist
inct
ive
join
t, hi
ghly
brok
en
horiz
ons
Plan
ar
smoo
th
and
step
ped
in
plac
es
Soft
to h
ard
infil
ling.
Man
y un
heale
d em
pty/
clean
jo
ints
1-1.
5 m
m jo
int
sepa
ratio
n,
mod
erat
ely to
hi
ghly
wea
ther
ed
join
t sur
face
s.
It
inte
rbed
s re
d lep
tite
on
man
y de
pths
, fo
rmin
g co
ntac
t zo
nes
char
acte
rized
by
wea
ther
ing
and
high
co
re lo
ss in
man
y pl
aces
. G
rayis
h, m
ediu
m g
rain
ed, r
elativ
ely
stro
ng, w
eake
ned
mos
tly b
y th
e pr
esen
ce o
f bio
tite,
feld
spat
hic
quar
tzite
in
man
y pl
aces
and
hig
hly
folia
ted
and
mica
ceou
s. 2
to
3 jo
ints
sets
pre
sent
in m
ost a
reas
.
MG
N
Mag
neti
te
Hig
hly
brok
en a
nd lo
w R
QD
mea
sure
d, fo
rmat
iona
l co
ntac
t are
as h
ighl
y br
oken
Roug
h jo
ints
, pl
anar
, as
perit
ies
hard
blac
k in
fillin
g in
m
ost p
laces
Ave
rage
join
t se
para
tion
of
1.5-
2mm
, m
oder
ately
w
eath
ered
jo
int s
urfa
ces
A c
ombi
natio
n of
red
and
gre
y le
ptite
s. A
lso s
trong
an
d co
mpe
te i
n so
me
plac
es.
Has
a h
igh
cont
ent
of
biot
itic
min
erals
but
it is
also
stro
ng in
mos
t plac
es.
SKN
Sk
arn
May
sh
ow
pres
ence
of
m
agne
tites
, lep
tites
in
fe
ldsp
athi
c m
atrix
Roug
h jo
int,
undu
latin
g in
som
e pl
aces
. Pl
anar
in
man
y pl
aces
Har
d bl
ack
infil
l in
man
y pl
aces
Ave
rage
se
para
tion
is 1m
m,
mod
erat
ely
wea
ther
ed
join
t sur
face
s
A h
ighl
y w
eak
form
atio
n, i
nter
bedd
ing
red
leptit
e, us
ually
coa
rse
grain
ed w
ith f
eldsp
athi
c qu
artz
ite i
n so
me
plac
es, p
itted
, sili
cifie
d, m
ay h
ave
high
RQ
D b
ut
wea
k, h
ighl
y fo
liate
d an
d w
eak
along
thes
e fo
liatio
ns
(M(M(M(MM(MMM(((((aaaaaaalllmmmm
bmmm
bm
berrergggegegegggg
t,t,t,,,2222220
100001111)1)1)1))))
22
Some of the weak zones in the two areas were found to be characterized by lithological rock formation contacts. Other weak zones comprised low strength zones due to weathering, decomposition, leaching, etc. The most notable weakness zones were the highly decomposed and broken or leached segments as shown in Figure 3.6.
Figure 3.6 Rock core showing a zone of weakness as indicated by rock mass alteration, decomposition and weathering.
There were no established weak zone patterns in the two areas due to distances separating the boreholes. However core logging revealed weak zones in individual boreholes which were categorized as (i) highly fractured and (ii) weathered and low strength zones as shown in Figure 3.7.
Figure 3.7 Weak zone interval distribution along boreholes (Umar, et. al, 2013)
Further studies were necessary to get a better understanding of the rock mass behavior as mining progressed in this orebody. A numerical modeling approach was adopted to test the influence of
23
various rock mass parameters and large-scale structures. The following chapter presents the numerical modeling activities that were carried out to investigate the Printzsköld orebody hangingwall and surrounding rock mass behavior due to mining.
24
25
4. Numerical modeling and stress analysis of the Printzsköld orebody
4.1. Numerical analyses of the Printzsköld orebody
4.1.1 Modeling approach
To understand the stress redistribution and hangingwall stability of the Printzsköld orebody as mining progresses to lower levels, three different numerical simulations were conducted. A two-dimensional finite element continuum conceptual model was conducted using Phase2 (Rocscience, 2002). This was followed by two three-dimensional modeling studies – one continuum and one discontinuum – both using 3DEC (Itasca Consulting Group Inc., 2013).
The two-dimensional model was used to investigate stress redistribution and hangingwall failure mechanisms through elastic and elasto-plastic model runs. Furthermore, a parametric analysis was conducted for the estimated strength parameters for the rock mass around the Printzsköld orebody.
To overcome the possible limitations associated with the two-dimensional model analysis, a three-dimensional model was conducted as described above. Similar input parameters and model dimensions were used and the results compared. A parametric sensitivity analysis was also conducted to be compared with the two-dimensional model results.
The third model simulation involved explicitly including large-scale structures in the model. The inferred structures from Mattsson and Magnor (2010) were included in this model, see Figure 4.1. This was done to investigate their potential effect on the stress redistribution and rock mass yielding as mining deepened.
Figure 4.1 Illustration and setting of three large-scale structures in the Printzsköld orebody shown relative to the 920 m mining level (Wettainen, 2010): (a) plan view, and (b) a longitudinal cross-section.
26
4.1.2 Model geometry
The Printzsköld orebody is of fairly complex geometry and assumptions had to be made to come up with a workable model to simulate mining of the orebody. A representation of the two-dimensional model as shown in Figure 4.2 was made in Phase2, which became the basis for evaluation planes where stresses were studied in more detail in the three-dimensional models. The two-dimensional model size was set to a length of 3.2 km and a depth of up to 1.7 km to accommodate the entire Printzsköld orebody. This size was chosen to reduce possible boundary effects.
Elastic constants used in the two-dimensional model were taken from Sjöberg (2008). The Young's modulus and Poisson’s ratio used for both the footwall and hangingwall rock mass was 70 GPa and 0.27, respectively. For the orebody these values were set to 65 GPa and 0.25, respectively. The density of the host rock mass was 2700 kg/m3 while that for the orebody was 4700 kg/m3. The boundary conditions in the two-dimensional model were that of zero velocity on the bottom with roller boundaries on the vertical boundaries.
Figure 4.2 Two-dimensional model of the Printzsköld orebody (vertical cross-section with an east-west strike in the model axes).
The three-dimensional models were set to a size 2800 m x 3629 m x 2050 m to reduce possible model boundary interference, see Figure 4.3. A graded discretization was used with fine meshing close to the orebody and a gradually coarser mesh toward the model boundaries. The boundary conditions were that the model was fixed at the bottom, and with roller boundaries (zero horizontal velocity) on all
27
vertical boundaries. The top surface was modeled as a free surface. As mentioned previously, it was assumed that neighboring orebodies did not affect the Printzsköld orebody; hence, only the Printzsköld orebody was included in the model. This assumption needs to be verified in future mine-scale models, but initial stress models by Perman et al. (2011) indicated that interaction effects may be limited. Figure 4.4 shows the model set up of the discontinuum model of 3DEC in which large-scale structures were included.
Figure 4.3 Model geometry for the model of the Printzsköld orebody
28
Figure 4.4 Model set-up showing large scale structure locations in the Printzsköld orebody and the model extents.
The maximum virgin horizontal stress in the Malmberget mine is believed to be oriented at 131° (Sjöberg, 2008), thus being nearly perpendicular to the strike of the Printzsköld orebody. In these models the model axes were rotated so that the x-axis was aligned with the maximum horizontal stress direction. This eliminated the need to translate all stress calculations from the model axes to the mine axes; thus, no shear stresses had to be applied to the boundaries of the model to replicate the virgin stress.
In both types of models (two and three-dimensional models) caving was not explicitly simulated. The current cave was simulated as a void that has progressed to about 300 m below the ground surface, corresponding to the cave front location as of 2014 (determined from probe drilling and micro-seismic monitoring by LKAB). Lastly, cave mining was simulated by removing one sublevel at a time, and as mining progressed a void was left. Additional caving of the existing void was not simulated in these models.
4.1.3 Evaluation planes
Six evaluation planes (C1, C2, C3 and L1, L2, L3) were selected on which the results were compared for the three models see Figure 4.5 and Figure 4.6. The plane, on which the Phase2 model was constructed in Figure 4.2, was selected as one of the primary evaluation planes in 3DEC and was denoted by L1. The plane, on which the query line passes for stress evaluation in the Phase2 model, was selected in 3DEC and denoted C1 as a primary evaluation plane see Figure 4.5 and Figure 4.6.
Printzsköld orebody (red)
Structure ID DZ 050 Structure ID DZ 031
Structure ID DZ 032
2050
m3629 m
2800 m
N
29
Figure 4.5 Evaluation planes in the longitudinal view of the Printzsköld orebody. Plane L2 is about 250 m away from plane L1, while plane L3 is about 700 m away.
0 200m
L3 L2 L1
30
Figure 4.6 Evaluation planes in the vertical cross-sectional view of the Printzsköld orebody
Other evaluation planes were also used to investigate trends in stress redistribution especially in the three-dimensional continuum model. The mining sequence simulated was on level basis (one mining level at a time and the model was re-run).
4.1.4 Initial stresses and material parameters.
Initial stresses were obtained from Sjöberg (2008) in which the vertical stresses were set equal to the pressure from overburden material and the horizontal stresses were found to be (through stress calibration):
zH 0358.0 (4.1)
zh 0172.0 (4.2)
in which z is the vertical depth in meters and all stresses are in MPa. Here, H is the maximum horizontal stress and h is the minimum horizontal stress with H having an orientation of 131° from the local north of the mine. The boundary conditions used were zero-velocity ("roller") boundaries on the vertical boundaries of the model, as well as for the bottom boundary. The ground surface was modeled as a free surface. The strength parameters used for elastic-perfectly plastic models are shown in Table 4.1.
C3 C2 C1
Hangingwall
Cave
Footwall
Printzsköld Orebody
0 200m
970 m level
1225 m level
31
Table 4.1 Strength parameters used for elastic-perfectly plastic models (Sjöberg, 2008)
Area Strength parameter
Value
Hangingwall
c [MPa] 5.18 [°] 50.7
tm [MPa] 0.71 Orebody c [MPa] 4.81
[°] 50.7
tm [MPa] 0.48 Footwall c [MPa] 6.67
[°] 52.9
tm [MPa] 1.3 c = cohesion; = internal friction angle; tm = tensile strength
For plastic analysis, three scenarios were analyzed to assess the effects of varying cohesion and tensile strength of the hangingwall rock material. The three scenarios were base case, low values and high values for cohesion and tensile strength, respectively. The friction angle was kept at a value of 50.7° for the sensitivity analysis, see Table 4.2. The cohesion value was varied by ± 3 MPa for the low and high cohesion cases. For the tensile strength, zero tensile strength was chosen for the low strength case, and an approximately twice as high tensile strength (1.5 vs. 0.7 MPa) for the high strength case.
Table 4.2 Strength parameters used in the model as well as variations of parameters.
Parameter Base case values
High tensile strength
Low tensile strength
High cohesion Low cohesion
Cohesion c [MPa]
5.18 5.18 5.18 8.18 2.18
Tensile strength
t [MPa] 0.71 1.5 0 0.71 0.71
In the discontinuum model, the large scale structures were modeled with strength parameters presented in Table 4.3. The strength parameters were varied to investigate their sensitivity on the hangingwall, crown pillar and footwall stability.
32
Table 4.3 Strength parameters for the large scale structures varied for sensitivity analysis in the Printzsköld orebody.
Case Cohesion (MPa)
Friction angle (°)
Tensile strength t (MPa)
Shear stiffness ks
(GPa)
Normal stiffness kn
(GPa) Base case 0.64 37 0 10.3 28.1 Low cohesion 0 37 0 10.3 28.1 Low friction angle 0.64 27 0 10.3 28.1
4.2. Model results
4.2.1. Stress redistribution
Conceptual continuum model results from Phase2 and 3DEC indicated stress build-up in the cap rock. A trace line along evaluation plane L1 in the cap rock was evaluated to further examine this behavior. In Phase2 the maximum value of major principal stress along this line was calculated to be 100 MPa while in 3DEC it was 20 MPa at the same location. However, the maximum calculated value for the major principal stress in the entire crown pillar (regardless of location) for the Phase2 model was 168 MPa; whereas for the 3DEC continuum model the maximum stress was 84 MPa.
Stresses were also analyzed along the line which traces the C1 evaluation plane. Higher stresses were calculated from the Phase2 model compared to the 3DEC model. Stress variation graphs from 3DEC and from Phase2, see Figure 4.7, show similar trend in the behavior of the stresses along the C1 trace line in the hangingwall for all mining levels tested.
33
Figure 4.7 Calculated major principal stress using the 3DEC and Phase2 numerical models, shown along a line traced along evaluation plane C1 in the hangingwall of the Printzsköld orebody and monitored on evaluation plane L1. The 0 m point is the starting point at the toe of the cave and the 1200 m is at the ground surface.
One reason for this seemingly large difference is that the two-dimensional model does not include abutment effects from the third dimension, which would lead to stresses being over-estimated compared to a three-dimensional model. Also it should be noted that zone sizes are larger in the 3DEC model, which further reduces the stress concentrations compared to those in the Phase2 models.
In both models it was found that as mining deepened towards the 1225 meter level the stress was redistributed such that a destressed zone was created in the hangingwall of the Printzsköld orebody. The destressed zone extended to approximately 200 meters into the hangingwall from the cave. The results from the two modeling approaches compared well with respect to stress redistribution patterns, see Figure 4.8 and Figure 4.9. It should be noted that the sign conventions for the two programs are different. In 3DEC (Figure 4.9) compressive stress is negative while in Phase2 it is positive (Figure 4.8). Sensitivity analyses in the Phase2 model revealed that cohesion was the most sensitive parameter to the increased yielding in the rock mass.
34
Figure 4.8 Major principal stress calculated in the two-dimensional (Phase2) model.
35
Figure 4.9 Major principal stress calculated using a three-dimensional continuum model (3DEC) and shown on a longitudinal cross-section on the evaluation plane L1.
The discontinuum model results were compared to the continuum model results for 3DEC, see Figure 4.10. Stresses in the cap rock in the discontinuum model were lower than those found in the continuum model. In the cap rock, the maximum value of the major principal stress along the line tracing the evaluation plane L1 was about 72 MPa in the continuum model while on the same location the discontinuum model revealed only 48 MPa.
In the cave bottom the maximum value of the major principal stress was 192 for the continuum model compared to 144 MPa for the discontinuum model results. It was inferred that this result was due to the presence of large-scale structures, with the reduction in the stress buildup in various places resulting from slip developing along the large-scale structures. These structures thus acted as stress release areas for the highly stressed regions.
36
Figure 4.10 Maximum principal stress calculated using a three-dimensional discontinuum (3DEC) model and shown on the evaluation plane L1 after mining the 1225 meter level.
4.2.2. Yielding in the rock mass and slip along the structures
Elasto-plastic analyses in the continuum models (Phase2 and 3DEC) showed a mixture of shear and tensile yielding mechanisms in the hangingwall. In the Phase2 model, tensile yielding dominated most of the destressed zones in the hangingwall as shown in Figure 4.11. There was more shear and tensile yielding in the cap rock as well as in the lower parts of the hangingwall, see Figure 4.12. In the three-dimensional continuum model using 3DEC, the reduction of cohesion by 3 MPa from 5.18 MPa was evaluated on the L1 evaluation plane and showed that the area of yielding (combination of shear and tensile) increased by more than 100,000 m2. The rest of the strength parameters variations did not have such a response in the model. Large-scale structures did not influence the amount of yielding in the rock mass.
37
Figure 4.11 Yielded elements of the hangingwall of the Printzsköld orebody calculated using the two-dimensional (Phase2) model after mining down to the mining level of 1225 m.
Figure 4.12 Failure state calculated using three-dimensional continuum (3DEC) model and shown on the evaluation plane L1 for the base case values and for mining to the level of 1225 m.
Predominantly shear yielding
Mined to the 1225 m level 0 200
Level 1225m
38
Slip was analyzed in the discontinuum model and it was found on all the three large scale structures simulated in the Printzsköld orebody, but to varying extent. The majority of slip movements along the structures were confined to the cave boundaries. The sensitivity of amount of slip with respect to a reduction in cohesion and friction angle was also studied. The analysis showed that by lowering friction angle, there was a large increase in slip on structures DZ031 and DZ050 compared to that arising from lowering only the joint cohesion. The amount of slip on structure DZ032 stayed almost the same when the two strength parameters were reduced, as presented in Figure 4.13.
Figure 4.13 Calculated slip area on the modeled large-scale structures for the three different strength cases simulated in the model of the Printzsköld orebody and for mining to the 1225 m level.
In Table 4.4 results are presented from modeling with discontinuum condition. The changes in slip were monitored with variations in friction angle and cohesion as outlined in Table 4.3. The two structures (DZ031 and DZ050) for which large joint slip area was recorded were dipping at a more vertical angle compared to DZ032. They also exhibited a larger slip area due to the orientation of the in situ stresses, which was found to be more parallel to the two large-scale structures and thus provided more shear force on the two structures than on DZ032.
Base Case
Low cohesion
Low friction
0
50000
100000
150000
200000
250000
300000
350000
400000
450000
500000
Jnt IdDZ031 Jnt Id
DZ032 Jnt IdDZ050
Are
a un
der s
lip (m
2)
Large Scale Structure Id
Base Case
Low cohesion
Low friction
39
Table 4.4 Numerical modeling results showing the calculated slip (in brown) on the three structures with the various strength parameters after mining the 1225 m level. Viewed parallel to the strike of the Printzsköld orebody.
case Joint DZ031 Joint DZ032 Joint DZ050 Base case
Low cohesion
Low Friction angle
40
41
5. Discussion
Core logging provided information on the number of joint sets, joint infilling and joint wall conditions. In this characterization the rock mass was assumed to be in dry conditions. However, one of the limitations in core logging was that the core was not oriented. This made it difficult to relate joint set orientations. It was however possible to infer the number of joint sets in a given piece of rock core by investigating neighboring joints whose relative orientation could only be referenced with respect to the orientation of the core. Borehole photography provided more accurate information on the joint orientations and confirmed the three joint sets.
Drift mapping was conducted only in the Printzsköld orebody. The Geological Strength Index (GSI) was estimated from information collected in this activity. There were no triaxial data for the rock mass in the orebody and so the mi values were estimated from the Hoek-Brown tables using the descriptions from tunnel mapping. Rock mass characterization alone is not enough to make informed decision on how to deal with the rock mass especially when caving is involved. Further investigations are necessary using numerical analysis to simulate rock mass behavior before, during and after mining to various levels.
While the two continuum condition model addressed stress redistributions and compared well, the nature of sensitivity analysis conducted in these two models were slightly different. Firstly in the Phase2 model, three strength parameters friction angle, cohesion and tensile strength were tested for sensitivity in the model. The 3DEC model was used to test for the effects of varying cohesion and tensile strength with regard to yielding in the hangingwall and cap rock. The model results were more sensitive to variations in cohesion compared to the other strength parameters in the Phase2 model.
The destressed zone in the true hangingwall can be thought of as a beam "fixed" to the cave back and the toe of the hangingwall-cave boundary (see Figure 5.1). Bending of this "beam" will induce tensile failure in the hangingwall. The cave back, on the other hand, is likely to fail in shear, due to high compressive stresses, and thus advance upwards. This would lead to an extension of the exposed hangingwall "beam", which in turn will promote additional tensile failure, thus resulting in progressive cave growth.
42
Figure 5.1 Failure mechanism forecasted for the Printzsköld orebody for continued mining toward depth.
The analyses showed that the large-scale structures provided slip surfaces as mining progressed to the 1225 meter level. Slip acted to reduce the stress build-ups observed in the continuum models. By implication, the reduced stress build-up in the crown pillar reduces failure by shear and the beam effect is compromised. The presence of weak zones increases the chances of these large-scale structures to intersect these weak zones and cause further instability in the hangingwall. Also, from the observed slip it could be said that caving progression may be influenced by the structures intersecting the cave boundary, as the structures provide the likely caving direction through this intersection. This may progress through slip. This trend is thought to repeat as the cave boundaries change.
The beam effect experienced in the hangingwall. The hangingwall is likely to fail in tensile
Toe of the cave
Ground surface
Stage one Rock mass e Rock mass
Mined out void
Ground surface
Tensile and shear failure will lead to extended cave in to the hangingwall.
Stage two
Cave back advances upwards due to shear failure
43
6. Conclusions
Based on the work presented in this thesis, the following conclusions may be drawn:
The rock mass in the hangingwall of the Printzsköld orebody is competent but contains weak zones which need further characterization. The rock mass can likely withstand high stresses but as mining progresses to lower levels the hangingwall would eventually fail in shear and/or tensile yielding.
The cohesion was the most sensitive parameter in the strength parametric analysis done in Phase2. Lowering the rock mass cohesion resulted in a larger area of yielding (on the evaluation plane).
Shear and tensile yielding mechanisms were found to be dominant in the hangingwall. The lower parts of the hangingwall were found to be in shear whilst the middle part was primarily failing in tension. Numerical analyses have shown that mining to the level 1225 meters will likely cause shear failure in the crown pillar and as this happens the hangingwall will also fail.
Large-scale structures did not affect the far field stresses and rock mass yielding. However, there was slip recorded along these structures and the slip was predominantly confined to the cave boundaries. Changing the friction angles resulted in larger amount of slip on two of the three large scale structures. It was inferred that the structure in which there was less slip was oriented in a more perpendicular direction with respect to the orientation of the major principal virgin stress.
The reduction of stress build-up in the cap rock in the discontinuum model was concluded to be due to the slip developing on the large-scale structure surfaces.
A potential failure mechanism has been described in which the hangingwall is assumed to form a beam fixed in the cap rock and the cave bottom. Failure of the cap rock has been described to potentially cause the failure of this ‘beam’ causing a progressive failure of the hangingwall.
44
45
7. Recommendations for future research
Based on the work presented in this thesis, the following recommendations for continued work are proposed:
It is recommended that using a local model approach, a numerical analysis investigation be conducted to simulate the effects of ore extraction and its rate on the stress redistribution and yielding in the hangingwall.
A detailed damage mapping is recommended in the Printzsköld orebody. This will enable a systematic review of rock mass behavior in relation to the ore extraction as mining deepens. The results from the damage mapping may also be used to validate numerical models
More tunnel mapping should be carried out to establish discontinuity and rock mass characteristics. Through this tunnel mapping and perhaps more drilling; a detailed study of the biotite schist zones be conducted to ascertain the biotite schist extents in the currently mined areas of the Printzsköld orebody. This is important, as the biotite schist is the weakest rock unit, thus possibly affecting cave propagation significantly if present near the orebody.
The inferred large-scale structures in the Printzsköld orebody, and probably the entire mine; need further verification of their existence and locations. Moreover, a better estimate of their mechanical properties is warranted. It is also recommended that future work should include studying the mining-induced seismicity in the cap rock and hangingwall. A mine-wide seismic network monitoring system has been in place at the Malmberget mine for several years and the collected data could be used to analyze patterns and identify linkages that may increase the understanding of both cave progression and how caving is affected by the structures.
Once more rock mass and discontinuity characterization has been done, it is recommended that a discrete fracture network (DFN) be constructed which could best describe the Printzsköld orebody and its surrounding rock masses for future numerical analyses. This would enable e.g., synthetic rock mass approaches to potentially be employed in future work.
46
47
8. References
Aksoy, C.O., Kose, H., Onargan, T., Koca, Y. & Heasley, K. 2004, "Estimation of limit angle using laminated displacement discontinuity analysis in the Soma coal field, Western Turkey", International Journal of Rock Mechanics and Mining Sciences, vol. 41, no. 4, pp. 547-556.
Banda, U.S. 2013, Rock Mass Characterization of the Printzsköld and Fabian orebodies at the Malmberget Mine, Luleå University of Technology, Luleå, Sweden.
Bell, F.G. & Donnelly, L.J. 2006, Mining and its Impact on the Environment, CRC Press, London.
Bieniawski, Z.T. 1989, Engineering Rock Mass Classifications, 251, John Wiley & Sons, New York.
Board, M. & Pierce, M. 2009, "A review of recent experience in modeling of caving." 43rd U.S. Rock Mechanics Symposium, June 28, 2009.
Brady, B.H.G. & Brown, E.T. 1985, Rock Mechanics for Underground Mining, George Allen and Unwin (Publishers) Ltd, London.
Brady, B.H.G. & Brown, E.T. 1993, Rock Mechanics for Underground Mining, 2nd Ed, London; New York: Chapman & Hall.
Cai, M., Kaiser, P.K., Uno, H., Tasaka, Y. & Minami, M. 2004, "Estimation of rock mass deformation modulus and strength of jointed hard rock masses using the GSI system", International Journal of Rock Mechanics and Mining Sciences, vol. 41, no. 1, pp. 3-19.
Deere, D.U. & Miller, R.D. 1966, Engineering classification and index properties for intact rock, University of Illinois, Illinois.
Dershowitz, W. & Einstein, H. 1988, "Characterizing rock joint geometry with joint system models", Rock Mechanics and Rock Engineering, vol. 21, no. 1, pp. 21-51.
Donnelly, L.J., Culshaw, M.G. & Bell, F.G. 2008, "Longwall mining-induced fault reactivation and delayed subsidence ground movement in British coalfields", Quarterly Journal of Engineering Geology and Hydrogeology, vol. 41, no. 3, pp. 301-314.
Duplancic, P. & Brady, B.H. 1999, "Characterization of caving mechanisms by analysis of seismicity and rock stress", 9th International Congress on Rock Mechanics, A.A. Balkema, pp. 1049.
Fonner, R.F. 1987, 1999-last update, Mountain State Geology [Homepage of West Virginia Geological and Economic Survey], [Online]. Available: http://www.wvgs.wvnet.edu/www/geohaz/geohaz3.htm.
Gilbride, L.J., Free, K.S. & Kehrman, R. 2005, "Modeling block cave subsidence at the Molycorp, Inc., Questa Mine", 40th U.S. Symposium on Rock Mechanic for Energy. American Rock Mechanics Association (ARMA). June 25-29, Anchorage, Alaska, USA. 05-881.
48
Hamrin, H., 2001, "Underground mining methods and applications. Underground Mining Methods: Engineering Fundamentals and International Case Studies". Society of Mining, Metallurgy and Exploration, 8307 Shaffer Parkway, Littleton, CO 80127, USA, 2001., pp. 3-14.
Hoek, E. 1974, "Progressive caving induced by mining an inclined orebody", Trans Instn Min Metall, Sect A: Min Industry, vol. 83, pp. A133-139.
Hoek, E., Carranza-Torres, C. & Corkum, B. 2002, "Hoek-Brown Failure Criterion - 2002 Edition. Proc.", North American Rock Mechanics Society meeting in Toronto in July 2002.
Itasca Consulting Group Inc. 2013, 3DEC (3-Dimensional Distinct Element Code), Version 5.0. www.itascacg.com, Minneapolis, Minnesota, USA.
Itasca, F. 2009, Fast Lagrangian Analysis of Continua in 3 Dimensions, Version 4.0. www.itascacg.com, Minneapolis, Minnesota, USA.
Jing, L. & Hudson, J.A. 2002, "Numerical methods in rock mechanics", International Journal of Rock Mechanics and Mining Sciences, vol. 39, no. 4, pp. 409-427.
Laubscher, D.H. 1994, "Cave mining - the state of the art", Journal of South African Institute of Mining and Metallurgy, vol. 94, no. 10, pp. 279-293.
LKAB 2014, , Malmberget Mine Level [Homepage of LKAB], [Online]. Available: https://www.lkab.com/en/Future/Urban-Transformations/Why/New-Mine-Level/Malmberget-Mine-Level/.
Lupo, J.F. 1998, "Large-scale surface disturbance resulting from underground mass mining", international Journal of Rock Mechanics and Mining Sciences., vol. 35, no. 4-5.
Lupo, J.F. 1997, "Progressive failure of hanging wall and footwall Kiirunavaara Mine, Sweden", International Journal of Rock Mechanics and Mining Sciences, vol. 34, no. 3–4, pp. 184.e1-184.e11.
Marinos, V., Marinos, P. & Hoek, E. 2005, "The geological strength index: applications and limitations", Bulletin of Engineering Geology and the Environment, vol. 64, no. 1, pp. 55-65.
Martinsson, O. & Hansson, K.E. 2004. "DAY SEVEN FIELD GUIDE, Apatite Iron ores in the Kiruna Area", Department of Geosciences, Luleå University of Technology.
Mas Ivars, D., Potyondy, D., Pierce, M. & Cundall, P. 2008, "The smooth-joint contact model", Proceedings of 8th World Cong. Comp. Mech./5th Eur. Cong. Comp. Mech. Appl. Sci & Eng, Venice, Paper a 2735.
Mas Ivars, D., 2010. "Bonded particle model for jointed rock mass". PhD thesis. Stockholm, Sweden: Royal Institute of Technology (KTH); TRITA-LWR PHD 1058. <http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-12055>.
49
Mas Ivars, D., Pierce, M.E., Darcel, C., Reyes-Montes, J., Potyondy, D.O., Paul Young, R. & Cundall, P.A. 2011, "The synthetic rock mass approach for jointed rock mass modelling", International Journal of Rock Mechanics and Mining Sciences, vol. 48, no. 2, pp. 219-244.
Mas, I.D., Deisman, N., Pierce, M. & Fairhurst, C. 2007, "The Synthetic Rock Mass approach–a step forward in the characterization of jointed rock masses", Proceedings 11th Congress of the International Society for Rock Mechanics, Ribeiro e Sousa, Ollala and Grossmann (editors), Taylor and Francis, Leiden, pp. 485.
Mas, I.D., Pierce, M., DeGagné, D. & Darcel, C. 2008, "Anisotropy and scale dependency in jointed rock mass strength–A synthetic rock mass study", Proceedings of the First International FLAC/DEM Symposium on Numerical Modeling, pp. 231.
Palmstrøm, A. 1996, "Characterizing rock masses by the RMi for use in practical rock engineering: Part 1: The development of the Rock Mass index (RMi)", Tunnelling and Underground Space Technology, vol. 11, no. 2, pp. 175-188.
Perman, F., Sjöberg, J. & Dahnér, C. 2011, "Detailed three-dimensional stress analysis of complex orebody geometry–model setup and results for the Malmberget Mine", Continuum and distinct element numerical modeling in geomechanics Itasca international Inc., Minneapolis.
Pierce, M., Mas Ivars, D., Cundall, P. & Potyondy, D. 2007, "A synthetic rock mass model for jointed rock", Rock Mechanics: Meeting Society's Challenges and Demands, 1st Canada-US Rock Mechanics Symposium, Vancouver, pp. 341.
Rockfield 2004, ELFEN 2D/3D Numerical Modelling Package, Rockfield Software Ltd. www.rockfield.co.uk, Swansea.
Rocscience. 2008, PHASE2 Version 6.028. Finite Element Analysis and Support Design for Excavations., Rocscience Inc. www.rocscience.com, Toronto, Canada.
Romer, R.L. 1996, "U-Pb system of stilbite-bearing low-temperature mineral assemblages from the Malmberget iron ore, Northern Sweden", Geochimica et Cosmochimica Acta, vol. 60, no. 11, pp. 1951-1961.
Sainsbury, B., Pierce, M. & Mas, I., D. 2008, "Simulation of rock mass strength anisotropy and scale effects using a Ubiquitous Joint Rock mass (UJRM) model.", Proceedings of the 1st International FLAC/DEM Symposium on Numerical Modeling. Itasca, August 25-28, 2008.
Sainsbury, B.L. 2010, "Sensitivities in the numerical assessment of cave propagation in Caving 2010:” Second International Symposium on Block and Sublevel Caving. Australian Centre for Geomechanics, 20-22 April, 2010.
Sainsbury, B.L., Sainsbury, D.P. & Pierce, M.E. 2011a, "A historical review of the development of numerical cave propagation simulations", Proceedings of the 2nd International FLAC/DEM Symposium in Numerical Modelling, Feb 14-16.
50
Sainsbury, B., Pierce, M. & Mas Ivars, D. 2009, "Analysis of Caving Behavior Using a Synthetic Rock Mass—Ubiquitous Joint Rock Mass Modeling Technique", Proceedings of the 1st Southern Hemisphere International Rock Mechanics Symposium (SHIRMS), Y. Potvin, J. Carter, A. Dyskin and R. Jeffrey (eds), pp. 16.
Sainsbury, D., Sainsbury, B., Board, M. & Loring, D. 2011, "Numerical Back-Analysis of Structurally Controlled Cave Initiation and Propagation at the Henderson Mine", 45th US Rock Mechanics/Geomechanics Symposium.
Shadbolt, C. 1975, "Mining subsidence and protective measures for surface subsidence", Chartered Surveyors Land Hydrogeology Minutes Quarterly, vol. 3, pp. 29-32.
Sjöberg, J. 2008, "Three-Dimensional Unit Stress Tensor Modeling of Complex Orebody Geometry", The 42nd U.S. Rock Mechanics Symposium. American Rock Mechanics Association (ARMA). June 29 – July 2, San Francisco, CA, USA. 08-124.
Skiöld, T. 1988, "Implications of new U-Pb zircon chronology to early proterozoic crustal accretion in northern Sweden", Precambrian Research, vol. 38, no. 2, pp. 147-164.
Terzaghi, K. 1964, "Rock Defects and Loads on Tunnel", Rock Tunneling with Steel Supports, eds. R.V. Proctor & T. White, Commercial Shearing Co, Youngstown, Ohio, USA, pp. 15–99.
Umar, S.B., Sjöberg, J. & Nordlund, E. 2013, "Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody of the Malmberget Mine, Sweden", Journal of Earth Sciences and Geotechnical Engineering, vol. 3, no. 4, pp. 147-173.
Umar, S.B., Sjöberg, J. & Savilahti, T. 2014, "Modeling of caving and deformation mechanisms of the hangingwall of the Printzsköld orebody in the Malmberget Mine.", submitted to: The International Journal of Mining, Reclamation and Environment.
Villegas, T. & Nordlund, E. 2008, "Numerical analysis of the hangingwall failure at the Kiirunavaara mine", Proceedings of the MassMin 2008, Luleå University of Technology , pp. 867-876.
Villegas, T., Nordlund, E. & Dahnér-Lindqvist, C. 2011, "Hangingwall surface subsidence at the Kiirunavaara Mine, Sweden", Engineering Geology, vol. 121, no. 1–2, pp. 18-27.
Vyazmensky, A., Elmo, D. & Stead, D. 2010, "Role of Rock Mass Fabric and Faulting in the Development of Block Caving Induced Surface Subsidence", Rock Mechanics and Rock Engineering, vol. 43, no. 5, pp. 533-556.
Wettainen, T. 2010, "Analys och prognostisering av uppblocning i Printzsköld", Master’s Thesis. Luleå University of Technology. ISBN 1402-1617.
Whittaker, B.N. & Pasamehmetoglu, A.G. 1981, "Ground tilt in relation to subsidence in Longwall mining", International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, vol. 18, no. 4, pp. 321-329.
51
Woo, K., Eberhardt, E., Rabus, B., Stead, D. & Vyazmensky, A. 2012, "Integration of field characterisation, mine production and InSAR monitoring data to constrain and calibrate 3-D numerical modelling of block caving-induced subsidence", International Journal of Rock Mechanics and Mining Sciences, vol. 53, pp. 166-178.
Woo, K., Eberhardt, E., Elmo, D. & Stead, D. 2013, "Empirical investigation and characterization of surface subsidence related to block cave mining", International Journal of Rock Mechanics and Mining Sciences, vol. 61, pp. 31-42.
APPENDED PAPERS
PAPER A
Umar, S.B., Sjöberg, J. & Nordlund, E. 2013, "Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody of the Malmberget Mine, Sweden", Journal of Earth Sciences and Geotechnical Engineering, vol. 3, no. 4, pp. 147-173.
Journal of Earth Sciences and Geotechnical Engineering, vol. 3, no. 4, 2013, 147-173ISSN: 1792-9040 (print), 1792-9660 (online)Scienpress Ltd, 2013
Rock Mass Characterization and Conceptual Modeling of
the Printzsköld Orebody of the Malmberget Mine,
Sweden
Sraj Banda Umar1, Jonny Sjöberg2 and Erling Nordlund3
AbstractThe LKAB Malmberget Mine is mined using sublevel caving. This mining method is cost-effective but results in successive caving of the host rock and mining-induced ground deformations. Consequently, re-locations of residential areas have been in progress inMalmberget ever since iron ore extraction on industrial scale commenced about a century ago. This study seeks to increase the understanding of the intrinsic characteristics of the rock mass governing deformation and caving activities. Rock mass characterizations were done in two selected orebodies — Printzsköld and Fabian. Two drill holes were drilled in each orebody from the surface. Geotechnical core logging was performed using the RMR system. Weakness zones were categorized to determine what role they played in the caving process. Point load testing was conducted for a sampling interval of about 5 m and selected uniaxial compressive strength tests were conducted to calibrate the point load index. Tunnel mapping was conducted in the hangingwall of the Printzsköld orebody. The finite element modeling code Phase2 was used for a sensitivity analysis of rock strength parameters and to study factors that may influence initiation of caving of the hangingwall.
Keywords: Mining-induced subsidence, rock strength, numerical analysis, weak zone characterization
1 Introduction1.1 BackgroundThe Malmberget mine is operated with large-scale sublevel caving. The mine comprises a total of 20 orebodies of which about 10 are in active production today. The mine is located in the municipality of Gällivare, see Figure 1 [1].
1Luleå University of Technology, Sweden. 2Itasca Consultants AB, Luleå, Sweden.3Luleå University of Technology, Sweden.
148 Sraj Banda Umar, Jonny Sjöberg and Erling Nordlund
Over the years, extraction from these orebodies has created varying subsidence problems on the ground surface, thus affecting residential areas and existing infrastructure. This subsidence is associated with the mining method which undermines the hangingwall causing instability and ground deformations. As mining progresses to deeper levels a larger subsidence area is created. The subsidence zone on the surface is characterized by cracks, sinkholes and steps in many areas. Forecasting the subsidence processes is not straight-forward as it can be rapid or slow depending on the rock mass conditions, stress conditions, the mechanisms at work, etc. Moreover, several of the orebodies are non-daylighting, which makes reliable subsidence prognosis even more difficult. Studies have been conducted [2, 3, 4, 5] in which many of these processes have been targeted and investigated in Malmberget mine; however, a deeper and detailed understanding of the subsidence processes is required. The present study was designed to improve the understanding of the subsidence mechanism of the Printzsköld orebody hangingwall. The Printzsköld orebody is located in the central area of the Malmberget mine [2], and was chosen to be the case study orebody. This orebody was chosen for two reasons: (i) it is one of the more important orebodies in terms of future production volumes in Malmberget, and (ii) it is located partly beneath the central area of the Malmberget municipality with associated large impact, if caving to the ground surface developed. Since mining started in this orebody there has not been any subsidence on the ground surface, but a reliable prediction of future caving and associated ground deformations is lacking. This study comprised a rock mass characterization campaign and a conceptual numerical modeling of the Printzsköld orebody.
Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody 149
Figure 1: Geological map of the Northern Norrbotten Region (Sweden), (from [1])
2 Conceptual Model for Caving Analysis2.1 Cave MiningLaubscher [6] defines cave mining as all mining operations in which natural caving of the orebody is encouraged through undercutting [6]. It includes mining methods such as sublevel caving, block and panel caving, and inclined draw caving. These mining methods allow for the bulk beneficiation of large orebodies at a low cost [7]. Since their introduction in the early 20th century it has been very important for the mine operators to predict the cave propagation.In [6], Laubscher also posits that this type of mining is the lowest-cost underground mining method as long as draw sizes are designed to equitable requirement for the material cavability. Among the 25 factors that Laubscher pointed out to be of primary consideration are surface subsidence and induced cave stresses [6].For cave propagation to be successfully initiated a well-developed low-dip joint must be present which interacts with two steeply dipping joints to create a free falling block [8].
150 Sraj Banda Umar, Jonny Sjöberg and Erling Nordlund
This assertion was earlier stated by [9] who suggested that when tangential stresses are low or tensile in nature free blocks may be able to slide on inclined discontinuities and fall by gravity in what is called gravity caving. These fallouts can also be augmented by the entire domains of weakness due to other factors such as shearing, weathering and dissolution. In [10] it is suggested that fallout conditions may develop when horizontal in situ stresses are low such as in those cases where slots and early mining have relieved the stresses or redistributed them away from the block [10].Duplancic and Brady [11] developed a conceptual model for caving by analyzing the seismic responses in the vicinity of a cave. They found that a caving zone can be characterized by the zones shown in Figure 2.
Figure 2: Conceptual model of the caving zone as proposed by [11].
The zones are described briefly here:i. Caved or mobilized zone
This zone is made up of rocks that have been mobilized or broken. The material in this zone is considered cohesion less. This material also provides stability to the walls of the cave.
ii. Air gapThe extraction of the caved material affects the height of the air gap. This is the gap that is left at the back of the cave.
iii. Zone of discontinuous deformation
Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody 151
In [11] this region was characterized as one that no longer provides support to the over-lying rock masses. Large scale movements of the rock mass have occurred in this region while no seismic activities are recorded.
iv. Seismogenic zoneIn this region a seismic front occurs due to brittle failure of joints and their slip on joints. The behavior is attributed to the changing stress regime and the propagation of caving.
v. Elastic zone
This is the zone furthest from the cave. It is composed of intact rock mass and has an elastic behavior towards deformation. It is a region ahead of the seismic front.Sublevel caving practiced at the Malmberget mine in the Printzsköld orebody fits into this model, since there is still a cap rock above the cave. Seismic monitoring in this orebody revealed trends of seismic activities in the hangingwall as well as the cap rock, which were consistent with the ones described in [11]. Also, the drill hole PRS01, drilled from the surface above the Printzsköld orebody in the hangingwall intercepted a void at a hole length of about 282 m, as shown in Figure 3.
Figure 3: Seismic monitoring and drilling in the Printzsköld orebody conforming to the conceptual model of caving by [11] (courtesy of Malmberget Mine).
152 Sraj Banda Umar, Jonny Sjöberg and Erling Nordlund
3 Rock Mass Characterization3.1 Local Mine GeologyThe Malmberget deposit is a paleproterozoic succession of greenstones, porphyries and clastic meta-sediments which are hosted by metavolcanics that have been intruded by pagmatites and granites [1] see Figure 1. The volcanic rocks have been transformed to sillimanite gneisses with quartz, muscovite, and local andalusite by the young granite intrusions. The iron ores are characterized by coarse magnetite and variable horizons of apatite with local sections rich in hematite. Generally, the border zones of the ore are characterized by skarn zones interpreted to be related to the formation of the ore [5]. Similar to many areas in Northern Sweden, the Malmberget deposit is characterized by NW-SE trending shear zones [5]. These zones are thought of as resulting from a complex geodynamic evolution which included repeated extensional and compressional tectonic regimes associated with magmatic and metamorphic events [12].Two of the four diamond drill holes used, were placed in the hangingwall of the Printzsköld orebody with borehole numbers PRS01 and PRS02 and their respective azimuths indicated by the white arrows as shown in Figure 4. The two drill holes were drilled to about 280 mand 302 m respectively. The rock mass was mainly dry during drilling.
3.2 Characterization of the Malmberget Mine.In the Malmberget Mine, many studies have been carried out aimed at characterizing the rock mass [3, 4]. Debras [3] tried to characterize the rock mass hosting the Printzsköld orebody, and offered a comprehensive petrologic and geologic description. Wänstedt, [4] carried out a rock mass characterization of the Malmberget rock using geophysical borehole logging [4]. He demonstrated the variations in rock mass properties, including rock mass strength, based on observations and deductions of rock densities from the geophysical electromagnetic methods.However, the results of the previous work were not geotechnically useful as no explicit characterization of geotechnical parameters was done. The lack of geotechnical characterization of the Printzsköld and Fabian orebodies in Malmberget necessitated a more thorough rock mass characterization study.
3.3 Data Collection and ResultsGeotechnical core logging and tunnel mapping was the major source of classification information for the various rock formations intercepted. Data was collected and processed in accordance with the Bieniawski (1989) Rock Mass Classification (RMR) system [13].
Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody 153
Figure 4: Drill hole locations for the Malmberget rock mass characterization. The orebody (magnetite in blue) and the footwall rock mass (Red Leptite in yellow, Red-Grey Leptite in brown) are shown as horizontal projections from the 945 m mine level (Courtesy of LKAB,
Malmberget Mine, 2011).
Core logging consisted of physical inspection, measurement and observation of the diamond drill cores. This approach enabled the collection of information such as RQD length, natural breaks, rock mass formation and description, see example in Figure 5. Table 1 shows the rock formations intercepted in drilling and their descriptions
Table 1: Rock formations intercepted in the Printzsköld orebody
Rock formation Abbrev Description of rock
Red Leptite RLE Reddish-brown, medium to coarse grained feldspathic quartzite matrix, hard.
Grey Leptite GLE Grey medium grained, partly micaceous
Red-Grey Leptite RGL Pinkish grey, medium to coarse grained, hard, micaceous in many places
Magnetite MGN Greenish grey, dark patches, medium grained, micaceous.
Skarn SKN Dark, green, coarse grained.
154 Sraj Banda Umar, Jonny Sjöberg and Erling Nordlund
Figure 5: Example of rock core from the Printzsköld orebody drill holes.
Tunnel mapping was undertaken in the Printzsköld orebody on 970 and 945 m levels see Figure 6. Other characteristics investigated were the geological strength index (GSI) [14], estimated from the rock mass characteristics in the field; joint orientations, and general rock mass descriptions. Table 2 shows the GSI values obtained from tunnel mapping of the 945 and 970 m levels in the Printzsköld orebody.
Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody 155
Figure 6: The 945 m level in the Printzsköld orebody in which underground mapping wascarried out. (Courtesy of LKAB, Malmberget Mine, 2012)
Table 2: GSI values as obtained from mapping the 945 and 970 m levels in the Printzsköld orebody
Rock FormationGSI
Conditions Comment
Red Leptite (RLE) 68 GOOD Blocky, slightly weathered iron stainedGrey Leptite (GLE) 64 GOOD Blocky, slightly weathered joints, iron stained.Red-Grey Leptite(RGL) 40 FAIR moderately weathered and altered surfacesMagnetite (MGN) 65 GOOD rough and slightlyweathered
BiotiteSchists (BIO) 20 POORhighly weathered surfaces and slicken sides joints, seamy
3.4 Point Load Strength TestingSampling of rock cores for intact rock strength was conducted at an interval of 5 m along the borehole. Samples were prepared for axial as well as diametral point load strength testing as described in [15]. Care was taken to apply the force at the rate of 10-60 seconds to the break of the sample in either direction. Laboratory direct uniaxial compressive strength test results from ten samples were used to calibrate the point load index as shown in Figure 7, which was found to be 21 MPa. The resulting intact rock strengths derived directly from the point load testing were subsequently used in the determination of the RMR values for this area.
156 Sraj Banda Umar, Jonny Sjöberg and Erling Nordlund
Figure 7: Point Load Index Calibration using Direct Uniaxial Compressive Strength (UCS) test results.
Intact rock strength results showed that most of the rock mass is highly competent and hard. Table 3 shows UCS strength results for the Malmberget area for each rock formation intercepted. The disparities in minimum and maximum values in these values were due to various weak zones.
Table 3: Uniaxial compressive strength values for the Printzsköld orebody, Malmberget.
Rock Formation UCS* [MPa]Max Min Average
Red Leptite (RLE) 302 60 184Grey Leptite (GLE) 242 90 149Red-Grey Leptite (RGL) 256 120 176Skarn (SKN) 170 74 127Magnetite (MGN) 182 71 127*values from point load tests
3.5 Rock Strength IsotropyThe rock units around the Printzsköld orebody showed considerable strength isotropy. The axial-to-diametral strength ratio is shown in Figure 8 (a and b).
UCS = 20.682Is(50)R² = 0.4033
0.0
50.0
100.0
150.0
200.0
250.0
300.0
350.0
0.0 5.0 10.0 15.0 20.0
UCS
(MPa
)
Point Load Index (Is(50)) (MPa)
UCS Vs Is50 plotLinear (UCS Vs Is50 plot)
Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody 157
Figure 8: Intact rock strength isotropy ratio in the Printzsköld orebody.
With the exception of a few outliers, the majority of the test data indicated isotropy ratios of between 0.6 and 1.1 for the Printzsköld orebody, see also Figure 9. This suggests that this is a fairly isotropic rock with respect to strength. Anisotropy was found to be characteristic of biotite schist zones and areas generally categorized as weak zones as observed from core logging and tunnel mapping.
158 Sraj Banda Umar, Jonny Sjöberg and Erling Nordlund
Figure 9: Normal distribution of the intact rock strength isotropy in the Printzsköld orebody, as inferred from point load tests.
3.6 Rock Mass Rating (RMR) and Rock Quality Designation (RQD)Tables 4 and 5 show value ranges for rock quality designation (RQD) [16]; and rock mass ratings (RMR) [13] for the Printzsköld orebody rock units. Rock mass rating showed that most of the rock mass was classified as good rock
Table 4: RQD values for the Printzsköld orebody, Malmberget mine.
Rock Formation RQD%Max Min Average
RLE 97 24 71GLE 94 28 65RGL 93 46 69SKN 90 65 85MGN 79 68 73
Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody 159
Table 5: RMR for the Printzsköld orebody, Malmberget mine.
Rock Formation RMR Class (based onAverage)Max Min Average
RLE 78 54 66 IIGLE 69 53 63 IIRGL 70 60 67 IISKN 67 65 66 IIMGN 67 62 65 II
The distribution of RQD and RMR has been graphically presented for boreholes PRS01 and PRS02 in Figure 10. The RMR in the orebody hangingwall fluctuates between 65 and 75, with many rock formations reaching a maximum of 80.
Figure 10: RMR and RQD distribution for the Printzsköld area. Both boreholes PRS01 (a) and PRS02 (b) showed similar value ranges.
0
20
40
60
80
100
120
0 50 100 150 200 250 300
Para
met
er R
atin
g
Borehole Depth (m)
RQD
RMR
0
10
20
30
40
50
60
70
80
90
100
0 50 100 150 200 250 300 350
Para
met
er R
atin
g
RQD
RMR
(a)
(b)
160 Sraj Banda Umar, Jonny Sjöberg and Erling Nordlund
3.7 Joint DescriptionsThree main joint sets were found in the hangingwall of the Printzsköld orebody. The joint conditions are summarized as follows:
Moderately weathered, predominantly rough planar joint, hard infilling, some soft infill found in some joints. Joint spacing of 10 to 50 cm. Predominantly hard kaolinitic in fills.Joint orientations (dip/dip-direction) were: (i) 27°/346°, (ii) 13°/097°, and (iii) 72°/173°
3.8 Weak ZonesWeak zones were encountered in the Printzsköld hangingwall. These zones had properties different from the host rock and as such they needed to be systematically evaluated to determine their effect on overall stability of the rock mass. It was found that the weak zones had an impact on the rock strength isotropy in the Printzsköld orebody as seen from the distorted ratio plots for isotropy comparisons in Figure 6. Weak zones were divided into two categories: (i) highly fractured zones, and (ii) weathered/low strength zones.The highly fractured zones comprised rocks characterized by many fractures and they showed disking in some places due to stress concentrations. These were commonly observed in rock formations such as red leptite (RLE) in the hangingwall.The weathered low strength zones were made up of rocks that exhibited weaknesses due to material types, alteration and weathering. The material weaknesses were characteristic of all biotie zones found mostly in grey leptites as well as marking the contacts with the orebody magnetite (MGN). Figure 11 shows weak zones intercepted by each drill hole and their distributions.
Figure 11: Weak zone intervals presented for each borehole
Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody 161
4 Numerical Modeling of the Printzsköld Orebody4.1 ApproachThe objective of the conceptual modeling of the Printzsköld orebody was to assess the sensitivity of strength parameters. The aim was also to provide insight into possible failure mechanism(s) of the hangingwall. Phase2, a Finite Element Modeling (FEM) program from Rocscience Inc. [17] was used to analyze the stress re-distribution of the Printzsköld orebody hangingwall. The Printzsköld orebody has a complex geometry which require some geometrical simplifications. In this initial work, only a two-dimensional analysis was conducted, in which a vertical cross-section of the orebody and hangingwall was modeled, see Figure 12. This rather severe simplification was judged acceptable to study, conceptually, stress distribution and possible failure mechanisms of the hangingwall. Three-dimensional modeling, enabling including the plunge of the orebody, is obviously required in future work. However, the vertical cross-section perpendicular to the orebody strike was believed to, at least in some aspects, be justified in a two-dimensional model, see Figure 12. Caving was not explicitly simulated. Rather, the caved rock was simulated as a void, starting from the current situation, in which caving has progressed to about 300 m below the ground surface. The conceptual model was aimed at investigating the stability of the existing cave and factors that may trigger additional cave growth. The hangingwall response in this case was analyzed using both elastic and plastic material models.
4.2 Model Set-UpMining of the Printzsköld orebody started at the 780 m level. With continued mining toward depth, the cap rock caved and the cave advanced to the current depth of about 300 m below ground surface as of 2012, as shown in Figure 13. This mining and cave development has not been simulated in this model. Rather, simulation started with the extraction of 920 m level (mined in 2011). Sublevel heights in the mine are 25 m, but were slightly simplified in the model so that each mining level was set at 25 to 30 m from the sublevel below, and mining was simulated down to the 1052 m level (a total of six mining stages).
162 Sraj Banda Umar, Jonny Sjöberg and Erling Nordlund
Figure 12: 3D model of the Printzsköld orebody. Section line A-A indicates where the cross-section of the numerical model has been taken from.
The model length was set to 3.2 km and the depth to 1.7 km, to accommodate the entire Printzsköld orebody, both the caved and previously mined parts as well as the un-mined delineated orebody. The model size was chosen to minimize possible boundary effects. A query line for interpretation was offset at about 10 – 15 m from the hangingwall boundary, see Figure 13.
4.3 Mechanical PropertiesThe elastic constants used in this model were derived from [20] and they were Young’s
hangingwall). For the orebody the values were set at 65 GPa and 0.25 respectively. The density of the host rock was set at 2700 kg/m3 and that for the orebody was set to 4700 kg/m3.
Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody 163
Figure 13: Cross-section of the Printzsköld orebody in the Malmberget mine.
The rock strength parameters were also taken from [18]. Table 6 shows the strength parameter values in previous studies for the Malmberget area.
Table 6: Typical strength parameters for the Malmberget Mine. (from [18])Unit c [MPa] [°] tm [MPa]Global, mine-scale modelHangingwall 5.18 50.7 0.71Orebody 4.81 50.7 0.48Footwall 6.67 52.9 1.30Local, drift-scale modelFootwall — Low 4.55 50.3 0.37Footwall — High 6.67 54.8 1.04
The joint strength parameters were obtained from [19] as follows:Normal joint stiffness: 110 GPa/m.Shear joint stiffness: 9 GPa/m.Joint friction angle: 35°.Joint cohesion: 0 MPa.
The selected input parameters for the plastic models are shown in Table 7. This model was run using an elastic-perfectly plastic Mohr-Coulomb material model.
164 Sraj Banda Umar, Jonny Sjöberg and Erling Nordlund
Table 7: Strength parameters used for the plastic modelsArea Strength parameter Peak value Residualvalue
Hangingwallc [MPa] 5.15 5.18
[°] 50.7 50.7tm [MPa] 0.71 0.71
Orebody c [MPa] 4.81 4.81[°] 50.7 50.7
tm [MPa] 0.48 0.48Footwall c [MPa] 6.67 6.67
[°] 52.9 52.9tm [MPa] 1.3 1.3
tm = tensile strength
The sensitivity analysis performed using the elastic model was based on the estimated input strength values obtained above. Parameters were varied over the ranges according to Table 8.
Table 8: Ranges of strength parameters used in the modelsStrength Parameter Valuec [MPa] 4 6 8 10
[°] 30 40 60tm MPa 0 0.5 1 1.5
4.4 In-Situ StressesThe rock stresses used in this model were derived from those used in [18], which comprised a numerical stress calibration against conducted measurements. The vertical stresses were considered to equal the pressure of overburden material at a given depth. The principal horizontal stresses were given by the equations below [19]:
zH 0356.0 (1)zh 0172.0 (2)
wherez is the vertical depth below ground surface in m, and H has an orientation of 130.6° clockwise from local north.
4.5 Model ResultsLinear elastic continuum models were first analyzed and the resulting stress distribution studied. Subsequently, the strength factors (essentially factors of safety) for various combinations of strength parameters were calculated. Variations in cohesion, internal friction angle and tensile strength showed that the rock mass behavior was most sensitive to changes in the cohesion. The distribution of strength factors in the elastic models provided an indication for low or high values of tensile or compressive stresses. Strength factors that reached zero were tensile stressed regions. In the plastic analysis it was possible to show the yielding in these zones and identify possible failure mechanisms in the hangingwall.
Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody 165
The elastic models showed that there were extensive areas of low stresses in the hangingwall as shown in Figures 14 and 15. A zone of relatively high compressive stresses was observed in the cap rock. However, low stresses were observed close to the ground surface as shown on point X.
Figure 14: Distribution of 1 after mining the 1052 level.
Figure 15: Distribution of stress 3 for mining level 1052 m.
166 Sraj Banda Umar, Jonny Sjöberg and Erling Nordlund
On the query line, the value of strength factor can be seen at point X for both cohesion and internal friction angle variations see Figures 16 and 17. The strength factors have high values because in these areas, the stresses are low, but because of low confinement, the strength is also low. In relative terms (strength factor defined as a ratio) the strength factor may be high, but even a slight increase in stress may cause the strength factor to drop significantly. In essence this can give an implication of adequate factor of safety in as far as failure is concerned. As will be shown in the plastic model this whole region is susceptible to failure.
Figure 16: Strength factor associated with variations in cohesion on the 10-15 m offset query line
Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody 167
Figure 17: Strength factor variations on 10-15 m offset query line due to variations in internal friction angle.
A ubiquitous joint post analysis investigation for joint orientations (anti-clockwise from the positive x-axis) of -72° and 27° (defined anti-clockwise from the positive x-axis) was performed. For the 72° joint set the zone of tensile strength in the hangingwall remained relatively smaller. For the 27° joint set the strength factors changed significantly and the indicated zone of tensile failure increased outwardly.
4.6 Yielding and Failure MechanismsA plastic model was used to determine possible yielding and failure mechanisms for the hangingwall. Figure 18 shows yielded elements of the plastic model indicating areas of shear and tensile stresses. The zones of yielding correlate, qualitatively, with the regions observed earlier in the elastic models as having low strength factors. In this model the entire hangingwall showed tensile yielding while a mixture of shear and tensile yielding was observed in the back of the caved room. Shear failure regions were also observed towards the bottom of the caved area.
168 Sraj Banda Umar, Jonny Sjöberg and Erling Nordlund
Figure 18: Yielded elements of the hangingwall of the Printzsköld orebody at the mining level of 1052 m.
The results also showed that as the mining depth increased the boundary of the yielded zone in the hangingwall increased progressively. Figures 19 through 21 show progressive yielding with deepened mining from 920 to 1052 m.
Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody 169
Figure 19: Yielding for mining of the 920 m level in the Printzsköld orebody.
Figure 20: Yielding for mining of the 996 m level in the Printzsköld orebody.
170 Sraj Banda Umar, Jonny Sjöberg and Erling Nordlund
Figure 21: Yielding for mining of the 1052 m level in the Printzsköld orebody.
5 DiscussionThe Printzsköld orebody is hosted within a relatively strong and competent rock mass. Rock core logged from this orebody showed five main rock formations: Red Leptite, Grey-Red Leptite, Grey Leptite, Magnetite (orebody formation) and Skarn. The length of the drill holes did not enable granite rock formation present in this region to be core drilled for examination; however, its presence has been inferred from other studies such as [2]. There was also biotite schist zones observed in the contacts between Grey Leptite and Magnetite. These biotite zones were characterized as part of the weak zones of the Printzsköld orebody. The average RMR range for the competent rock masses was 63 to 67. Individual rock formations exhibited rock mass ratings as high as 80. Biotite schist zones present a geotechnical challenge in this orebody due to their weakness. Tunnel mapping showed that most of the areas containing biotiteschists had an effect of lowering the RMR to as low as 30, while the GSI was found to be reduced to around 20. The rock mass also exhibited a relatively high strength isotropy in all rock masses. The loss of isotropy was observed in parts characterized as weak zones. It was not possible to fully describe the weak zones in the Printzsköld orebody as data obtained from the two diamond drill holes was too far apart to allow any correlation. The sensitivities of the hangingwall and cap rock to variations in rock mass strength parameters have been studied. The strength factor was most sensitive to variations in cohesion.
Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody 171
The plastic model showed a yielding in the hangingwall and a major part of the hangingwall was exposed to tensile stresses, giving us a tensile failure mechanism of the rock mass into the mined out area. It also indicated that a likely progressive failure under the mechanisms given above could ensue as the depth of mining increases. While this model offers a starting point for the evaluation of stress redistribution in the hangingwall in the cross section of the Printzsköld orebody, it does not fully capture an accurate description of stresses and failure distribution since failure is taking place in a three dimensional system.As a preliminary analysis of this orebody, fairly large simplifications have been employed to make the model as simple and understandable as possible. Some challenges were met pertaining to the numerical code (Phase2) output of strength factors in the post analysis interpretation. High values of strength factors were found in areas with very low confining stresses. These values do not necessarily reflect stability but rather that the strength was also low. The high values of strength factor could be misinterpreted as safe areas.
6 Conclusions and RecommendationsFrom this study the following conclusion can be drawn:
The results of the rock mass characterization showed that the rock mass strength in the Printzsköld area is generally high and competent. The averaged RMR across the rock formations ranged between 63 and 67, while the RQD had an average value in the interval of 71 to 85 %. The rock mass also had a GSI range of 60 to 75. GSI values were found to be very low in areas characterized by biotite schist.Numerical elastic models of the hangingwall of the Printzsköld orebody showed that the sensitivity of the hangingwall behavior was strongest for variations in cohesion.
The following recommendations are given:It is recommended that more joint orientation mapping be carried out in the lower levels of the Printzsköld orebody. More geotechnical drilling is required to ascertain rock mass geotechnical parameters for the lower levels of the Printzsköld orebody area which can be compared to those of upper levels obtained in this study. Drilling will also provide more information in the analysis of weak zones, their spatial distributions and their categorizations.A three dimensional model approach needs to be undertaken to account for the whole orebody geometry and stress interactions.
ACKNOWLEDGEMENTS: The authors would like to thank LoussavaaraKirunavaara AB (LKAB), The HjalmarLundbohm Research Centre (HLRC) and the Centre for Applied Mining and Metallurgy (CAMM) at LTU for funding this research. Special thanks go to Tomas Savilahti at the LKAB Malmberget Mine, for his continuous assistance for the smooth running of the study. Thanks also go to the entire research group members: Linda Jacobsson, Jimmy Töyrä, Fredrik Ersholm and Joel Andersson, all of LKAB, for their guidance, assistance and support during this research.
172 Sraj Banda Umar, Jonny Sjöberg and Erling Nordlund
References[1] Martinsson, O. and K. E. Hansson. 2004. Day seven field guide, Apatite Iron ores in
the Kiruna Area.[2] Wettainen, T. 2010. Analys och prognostisering av uppblockning i Printzsköld.
Licentiate Thesis. Luleå, Sweden: LuleåUniveristy of Technology. ISSN: 1402-1617.[3] Debras, C. 2010. Petrology, geochemistry and structure of the host rock for the
Printzsköld ore body in the Malmberget deposit. Master Thesis. Luleå: Luleå university of Technology: ISSN: 1653-0187.
[4] Wänstedt, S. 1991. Geophysical borehole logging in Malmberget. Technical Report. Luleå University of Technology; 14T.
[5] Romer, R.L. 1996. U-Pb system of stilbite-bearing low-temperature mineral assemblages from the Malmberget iron ore, Northern Sweden. GeochimCosmochimActa 6; 60(11):1951-1961.
[6] Laubscher, D.H. 1994. Cave mining - the state of the art. J South Afr Inst Min Metall; 94(10):279-293.
[7] Sainsbury, B.L., D.P. Sainsbury, and M.E. Pierce. 2011. A historical review of the development of numerical cave propagation simulations. Proceedings of the 2nd International FLAC/DEM Symposium in Numerical Modeling; Feb 14-16.
[8] Kendorski, F.S. 1979. Cavability of ore deposits: Min Engng, V30, N6, June 1978, P628–631. International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts. 2; 16(1):A8.
[9] Mahtab, M.A., D.D. Bolstad, and F.S. Kendorski. 1973. Analysis of the geometry of fractures in San Manuel Copper Mine, Arizona; 7715.
[10] Brown, E.T. 2003. Block Caving Geomechanics, The International Caving Study Stage 1 1997-2001, Australia: JKMRC, University of Queensland.
[11] Duplancic, P. and B.H. Brady. 1999. Characterization of caving mechanisms by analysis of seismicity and rock stress. 9th International Congress on Rock Mechanics Characterization of caving mechanisms by analysis of seismicity and rock stress.A.A. Balkema.
[12] Skiöld, T. 1988. Implications of new U-Pb zircon chronology to early proterozoic crustal accretion in northern Sweden. Precambrian Res 2; 38(2):147-164.
[13] Bieniawski, Z.T. Engineering Rock Mass Classifications, 251. New York (1989), John Wiley & Sons.
[14] Hoek, E., P. Marinos, and M. Benissi. Applicability of the geological strength index (GSI) classification for very weak and sheared rock masses. The case of the Athens Schist Formation. Bulletin of Engineering Geology and the Environment; 57(2), (1998), 151-160.
[15] Broch, E. and J.A. FranklinThe point-load strength test. International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts (1972), 11; 9(6):669-676.
[16] Deere, D.U. and R. D. Miller. Engineering classification and index properties for intact rock, AFWL-TR (1966), 65.116.
[17] Rocscience Inc. Phase2 Version 6.028 - Finite Element Analysis for Excavations and Slopes. www.rocscience.com, Toronto, Ontario, Canada.2008.
[18] Sjöberg, J. Three-Dimensional Unit Stress Tensor Modeling of Complex Orebody Geometry. American Rock Mechanics Association (ARMA).2008.
Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody 173
[19] MalmgrenL., and E. Nordlund.Interaction of shotcrete with rock and rock bolts—Anumerical study. Int J Rock Mech Min Sci, 6;45(4), (2008), 538-553
PAPER B
Umar, S.B., Sjöberg, J. & Savilahti, T. 2014, "Modeling of caving and deformation mechanisms of the hangingwall of the Printzsköld orebody in the Malmberget Mine.", submitted to: The International Journal of Mining, Reclamation and Environment.
Modeling of caving and deformation mechanisms of the hangingwall of the
Printzsköld orebody in the Malmberget Mine
Sraj Banda Umar1*, Jonny Sjöberg2 and Thomas Savilahti3
- 1*Corresponding author: Department of Civil, Enviornmental and natural resources
engineering, Luleå University of Technology, SE-97187, Luleå, Sweden, email:
[email protected]; 2Itasca Consultants AB, Luleå, Sweden,email: [email protected];3Loussavaara-Kiirunavaara AB (LKAB) Malmberget Mine, 3DECGällivare, Sweden,
email: [email protected].
Modeling of caving and deformation mechanisms of the hangingwall of the
Printzsköld orebody in the Malmberget Mine
Abstract
The sublevel caving in Malmberget mine caused mining induced surface
deformations. One of the currently mined orebodies is the Printzsköld orebody. As
mining deepens there is need to assess the behavior of the cave formed in the
subsurface above this orebody. Numerical analysis was used to assess effects of
extraction to deeper levels and perform strength parametric studies. Stress
redistribution was studied and results show high stress build-ups in the hangingwall
and the crown pillar. Two failure mechanisms have been identified; shear and tensile.
Reducing cohesion by 50 % increased the yielded zone to more than 100% in the
hangingwall.
Keywords: Numerical modeling; Printzsköld orebody; caving mechanism; strength
parametric study; stress redistribution; hangingwall stability
1 Introduction
1.1 Problem description
The Malmberget mine is owned and operated by the Luossavaara-Kiirunavaara AB (LKAB)
mining company. Iron ore is extracted with large-scale sublevel caving. The mine comprises a
total of 20 orebodies of which about 10 are in production today, see Figure 1. The mine is located
in the municipality of Gällivare about 100 kilometers north of the Arctic Circle, see Figure 2.
Figure 1. Schematic view of the orebodies at Malmberget Mine. (Courtesy [1])
The success of sublevel caving as a mining method requires progressive caving of the
orebody after initial blasting. This results in caving of the hangingwall, sometimes all the way to
ground surface with surface deformations as a side-effect. As mining progresses to deeper levels a
larger area is affected by these deformations. Over the years, extraction from the Malmberget
orebodies has led to deformations on the ground surface, thus affecting residential areas and
existing infrastructure.
Forecasting the caving process is not straight-forward as it can be rapid or slow depending
on the rock mass conditions, stress conditions, the failure mechanisms at work, etc. Moreover,
several of the orebodies in Malmberget are non-day-lighting, which makes reliable caving analysis
even more difficult, primarily since there is less available experience from such cases. The present
study was designed to improve the understanding of the caving and deformation mechanisms at
Malmberget and more particularly for one of the orebodies called Printzsköld.
Figure 2. Aerial view of the Malmberget Mine site. The blue- and magenta-colored objects are the
horizontal projections of the various orebodies in the Malmberget mine, and the numbers following
them are the current mining levels for which these orebodies are shown. Blue is magnetite ore and
magenta denotes hematite ore. The red patch shows the newly developed cave crater on the ground
surface from mining of the Fabian orebody. The Printzsköld orebody can be seen at the center and
the black line shows the current fence delimiting the mining area. The ground surface is between
the 150 and 220 m level in the mine coordinate system.
Mining of the non-daylighting orebodies in Malmberget, including Printzsköld, will leave a
cap rock, or crown pillar, between the extracted areas and the ground surface. As mining goes
deeper in the Printzsköld orebody, caving will gradually progress upward in the crown pillar and
the hangingwall. The caving propagation can lead to ground surface deformations, dependent on
the extent and rate of caving, and whether break-through develops or not. It is therefore essential to
investigate, describe and quantify this phenomenon and the governing factors.
The rock mass in the Printzsköld orebody was described in [2] as highly jointed in many
places. This rock mass is characterized by fractures, faults and weak zones. The ore is quite
competent but it is affected by the weak schistocity of biotite schists in some places on the
boundary between magnetite (ore) and the grey leptite horizon.
The caving is believed to be controlled by, among other factors, the stress conditions in the
rock mass. Umar et al. (2013) presented an investigation of the stress redistribution in the
hangingwall of the Printzsköld orebody. In this work a two-dimensional conceptual model was
constructed using the continuum code Phase2 [3]. A sensitivity analysis of strength parameters was
performed in which cohesion (c); tensile strength tm), and the angle of internal friction ( ) of the
rock mass were varied. The results from this study showed that cohesion was the most sensitive
strength parameter with respect to the stability of the hangingwall rock mass. However, the two-
dimensional simplification was judged to be rather severe and a three-dimensional model study is
thus warranted. For the Printzsköld orebody no three dimensional stress modeling has previously
been conducted with the aim of studying failure mechanisms to understand the caving process.
1.2 Objectives and scope of paper
This study was focused on the Printzsköld orebody, selected because of: (i) its importance as one
of the main production areas for future planned mining at Malmberget, and (ii) the importance of
better understanding future caving behavior and possible ground deformations in the central area
of Malmberget. The paper presents a conceptual model intended to study stress redistribution and
yielding in the hangingwall of this orebody as a function of mining at depth and the caving
activities. However, the caving process itself was not explicitly simulated in this study. Rather, the
study was focused on the specific objectives outlined below:
i. To provide an understanding of the rock mass behavior in the crown pillar of a non-
daylighting orebody as mining progresses deeper.
ii. To assess the sensitivity of rock mass strength parameters with respect to possible
caving and deformation mechanisms.
iii. To identify the failure mechanism likely to affect the hangingwall and the crown pillar
of the Printzsköld orebody, as mining deepens.
2 Numerical Modeling of the Printzsköld Orebody
2.1. Modeling approach
Three-dimensional numerical modeling was employed using the Three Dimensional Distinct
Element Code (3DEC) [4]. The rock mass was assumed to be an equivalent continuum (fractures
and joints thus not explicitly simulated) and both elastic and elasto-plastic analyses were
conducted. In the elastic analysis, stress patterns were evaluated for effects of mining at certain
depths. It was necessary to determine where stress build-ups developed in relation to the crown
pillar and the hangingwall.
In the elasto-plastic analysis, extent of yielding as a function of mining was investigated.
Moreover, a parametric study was conducted in which strength parameter values were changed and
the effect on yielding and stress distribution was studied.
2.2. Model geometry
The Printzsköld Orebody has a complex geometry. Extraction from this orebody requires a
controlled approach in which stresses are redistributed uniformly to avoid high stress build-ups in
certain mining areas as well as the hangingwall. Mining with sublevel caving in this orebody
started at the 780 m level. As mining progressed toward depth, an upward cave developed, which
as of 2012 had propagated up to approximately 300 m below the ground surface (at about the 500
m level in the local mine coordinate system). This orebody has a plunge hangingwall which is
defined by the flat-lying cave back, see Figure 3. In [2], only a vertical cross section of the
Printzsköld orebody was examined in a two-dimensional model, thus leaving out much of the
effects of the plunge hangingwall. However, a two-dimensional model of a longitudinal section of
the orebody was also modeled, but not reported. This work was included in the present study to
compare with the three-dimensional model.
Figure 3. The Printzsköld orebody simplified geometry showing mine production levels with a
downward mining sequence.
Another simplification that has been made concerns the mining sequence of the Printzsköld
orebody in the current model, as shown in Figure 3. The draw schedule within each sublevel or at
different sublevels simultaneously, has not been simulated; instead, each sublevel has been
excavated in full for each mining step in the model.
Mining was simulated to progress downwards on levels separated by 25 to 30 meters
(corresponding to actual sublevel height). This mining simulation started from the current active
production level at 920 m and progressed to the future 1225 m level with a total extraction of
approximately 70 Mt planned by November 2030. It should be noted that the possible continued
caving as mining goes deeper was not explicitly simulated in the model. Rather, the cave back that
945m 970m
996m 1023m
1051m 1080m
1109m
1138m 1167m
1196m
1225m
Caved zone with an upward cave
propagation
45m70m70
96m9623m1m
0mm
m
Mining levels with a simplified downward mining sequence
Plunge Hangingwall
True Hangingwall
had developed up until the year 2012 was kept at the same position and the possible behavior of the
rock around the cave analyzed as mining progressed downward.
The model size was set to 2800 m x 3629 m x 2050 m to reduce possible model boundary
interference, see Figure 4. A graded discretization was used with fine meshing close to the orebody
and a gradually coarser mesh toward the model boundaries. The boundary conditions were that the
model was fixed at the bottom, and with roller boundaries (zero horizontal velocity) on all vertical
boundaries. The top surface was modeled as a free surface.
It was further assumed that neighboring orebodies did not affect the Printzsköld orebody,
hence, only the Printzsköld orebody was included in the model. This assumption needs to be
verified in future mine-scale models, but initial stress models by [5] indicated that interaction
effects may be limited.
Figure 4. Model geometry for the model of the Printzsköld orebody.
2.3. Initial stresses and material parameters.
Initial stresses were obtained from [6] in which the vertical stresses were set equal to the pressure
from overburden material and the horizontal stresses were found to be (through stress calibration):
zH 0358.0 (1)
zh 0172.0 (2)
in which z is the vertical depth in meters. Here, H is the maximum horizontal stress and h is the
minimum horizontal stress with H having an orientation of 130.6° from the local north of the
mine. It should be noted that the model boundaries were oriented parallel and perpendicular,
respectively, to the initial major horizontal stress, to avoid having to apply shear stresses to the
model boundaries. The strength parameters used for elastic-perfectly plastic models are shown in
Table 1. These values were taken from [2], and are based on a study by [6]. These values are
considered equivalent strengths for the rock mass, when modeled as a continuum.
Table 1. Strength parameters used for elastic-perfectly plastic models (adapted from [6])
Area Strengthparameter
Value
Hangingwallc [MPa] 5.18
[°] 50.7
tm [MPa] 0.71Orebody c [MPa] 4.81
[°] 50.7
tm [MPa] 0.48Footwall c [MPa] 6.67
[°] 52.9
tm [MPa] 1.3c = cohesion; = internal friction angle; tm = tensile strength
For plastic analysis, three scenarios were analyzed to assess the effects of varying cohesion
and tensile strength of the hangingwall rock material. The three scenarios were base case, low
values and high values for cohesion and tensile strength, respectively. The friction angle was kept
at a value of 50.7° for the sensitivity analysis, see Table 2. The cohesion value was varied by
±3MPa for the low and high cohesion cases. For the tensile strength, zero tensile strength was
chosen for the low strength case, and an approximately twice as high tensile strength (1.5 vs. 0.7
MPa) for the high strength case.
Table 2. Strength parameters used in the model as well as variations of parameters
Base case values
High tensile strength
Low tensile strength
High cohesion
Low cohesion
Cohesion c [MPa] 5.18 5.18 5.18 8.18 2.18
Tensile strength t [MPa] 0.71 1.5 0 0.71 0.71
2.4. Evaluation
Comparisons with the 2-D models were made for vertical and longitudinal cross-sections of the
orebody. The mining stages corresponding to levels 970 and 1225 meters were selected for
interpretation and evaluation. The 970 meter level was selected as this represents an area of
mining in the near future. The 1225 m level was selected as it represented a significant difference
in the amount of overall extraction compared to that of the current mining state.
Three query lines were set in the vertical cross-section along the hangingwall to study
parameter effects on the model in comparison to the two-dimensional model analysis presented in
[2]. These lines represented three evaluation planes that were established in the vertical cross-
section view. Three evaluation planes were also established in the longitudinal cross-section of the
Printzsköld orebody. These planes were used to monitor and evaluate the stress redistribution as
mining progressed. Three cross-sectional view planes - C1, C2 and C3 - were chosen to evaluate
how far into the hangingwall the stress redistribution reached. Plane C1, placed about 100 m from
the cave boundary, was set as a starting point and established for comparison with the two-
dimensional Phase2 model stress evaluations reported in [2], see Figures 5 and 6. Evaluation plane
C2 was 50 m away from plane C1 and C3 was placed 50 m away from C2.
Figure 5. Evaluation planes C1, C2 and C3 in the vertical cross-sectional view of the Printzsköld
orebody
The evaluation planes in Figure 5 were situated in the true hangingwall while the
longitudinal evaluation planes in Figure 6 cut across the orebody. Planes L1 and L3 in Figure 6
were set-up to evaluate the behavior in the cap rock and plunge hangingwall, respectively. On each
plane, principal stresses and yielding were evaluated. L2 was used to give infill stress distribution
information in the plunge hangingwall to ascertain the trend of failure as stresses build-up in the
cap rock.
C3 C2 C1
Hangingwall
Cave
Footwall
Printzsköld Orebody
0 200m
970 m level
1225 m level
Figure 6. Evaluation planes in the longitudinal view of the Printzsköld orebody. Plane L2 is about
250 m away from plane L1, while plane L3 is about 700 m away.
The results are presented for both elastic and elasto-plastic analyses. A comparison of these
results has been made with those obtained in the two-dimensional analysis using Phase2 [2]. The
level 1109 meters was used as a control, to observe the trends and tendencies of rock mass
behavior. It was found that the rock mass behavior when mining level 1109 m did not depart much
from that found when mining the 1225 m level.
The sign convention for stress results in 3DEC is negative for compressive stresses and
positive for tensile as mentioned in section 2.3. However, for the Phase2 results the convention is
0 200m
L3 L2 L1
the opposite of that found in 3DEC, which is positive for the compressive stresses and negative for
tensile stress.
3 Results
3.1. Elastic modeling
Results for the 970 m and 1225 m level are shown in the vertical cross-section and the longitudinal
section in Figures 7 through 9. In the longitudinal section, a build-up of the major stresses in the
cap rock and the plunge hangingwall can be observed. For Figures 7 and 8, (a) represents results
evaluated on a plane that cuts through the orebody, while (b), (c), and (d) represent the major
principal stress in various parts of the model on the evaluation planes (C1, C2, C3) for the 970 and
1225 meter levels. A high concentration of compressive stress was seen to develop in the cap rock
when mining reached the level 1225 m.
Figure 7. Calculated major principal stress (negative stresses are compressive in units of Pa) for
mining down to the 970 m level in the Printzsköld orebody, shown on a longitudinal cross-section
for the evaluation planes (C1, C2 and C3) shown in Figure 5.
Evaluation plane C1
Evaluation plane C2 Evaluation plane C3
Plane across orebody
Figure 8. Calculated major principal stress (negative stresses are compressive in units of Pa) for
mining down to the 1225 m level in the Printzsköld orebody, shown on a longitudinal cross-section
for the evaluation planes (C1, C2 and C3) shown in Figure 5.
Evaluation plane C1
Evaluation plane C2 Evaluation plane C3
Plane across orebody
Figure 9. Calculated major principal stress (negative stresses are compressive in units of Pa)
compared between mining of 970 and 1225 m level in the Printzsköld orebody, shown on a vertical
cross-section the evaluation plane L1 as shown in Figure 6.
In the cross-sectional direction (shown in Figure 6) three monitoring planes (L1, L2 and L3)
were used which enabled assessment of the stress redistribution in the cap rock, true hangingwall
and plunge hangingwall. The major principal stress on these evaluation planes are shown in Figures
9 through 11.
The evaluation plane L1 in Figure 9 cuts through the roof of the cave and the Printzsköld
true hangingwall. It reveals a compressive stress regime build-up in the roof of the cave. A
destressed region is observed on the boundary of the cave going outwards into the hangingwall to
Evaluation plane L1
Evaluation plane L1
(a)
(b)
about 100 meters. Evaluation plane L2 (Figure 10) also shows a high stress region in the upper part
of the plunge hangingwall.
Figure 10. Calculated major principal stresses (in Pa) on the evaluation plane L2 in the cross
section of the Printzsköld orebody.
Figure 11 shows the monitoring plane L3 which addresses the plunge hangingwall on the
fringes of the orebody. The major principal stress was also analyzed along a line traced through
evaluation plane L1 (from Figure 6) and monitored results from evaluation planes C1, C2 and C3.
This was to determine how the major principal stress changed for the various distances of the
planes from the cave boundary.
Evaluation plane L2
Evaluation plane L2
(a)
(b)
Figure 11. Calculated major principal stress (negative stresses are compressive in Pa) compared
between mining to 970 and 1225 m level in the plunge hangingwall of the Printzsköld orebody,
shown on a vertical cross-section for evaluation plane L3 as shown in Figure 6.
Figure 12 shows the major principal stresses plotted against distance for the three
evaluation planes. The bottom of the cave is represented at the origin in the graph while the surface
is at 1100 m (upper part of the graph). Figure 12 also shows zones of high stressing and destresed
along each evaluation plane.
Evaluation plane L3
Evaluation plane L3
Figure 12. Major Principal Stress plot on line traced along the evaluation plane L1 and evaluated
on planes C1, C2 and C3.
A stress variation evaluation was undertaken along a line traced along the evaluation plane
C1 in Figure 5. The results were compared to the stress variations observed along the
corresponding evaluation (or query) line analysis from the two-dimesional model analyzed in
Phase2 [2], shown in Figure 13.
0
200
400
600
800
1000
1200
0 5 10 15 20 25 30
Dist
ance
alo
ng th
e ha
ngin
gwal
l on
plan
e (m
)
Stress (MPa)
C1
C2
C3
High stressing region in the cap
Figure 13. Phase2 model showing results for the major principal stress for comparison with the
3DEC results in Figure 9 (adapted from [2]).
Figure 13 from Phase2 compares well with Figure 9b, in which case both figures have been
taken showing the stress after mining the 1225 m level. However, larger stresses were found from
the Phase2 model compared to the 3DEC model. Stress variation graphs from 3DEC and from
Phase2, see Figure 14, show similar trend in the behavior of the stresses along plane C1 in the
hangingwall for all mining levels. However, it should be noted that the maximum stress value in
the 3DEC model in the cap rock is 20 MPa as shown in Figure 3.8 while the Phase2-model shows a
maximum stress just above 100 MPa at the same location, (both for the 1225 m mining step). One
reason for this seemingly large difference is that the two-dimensional model does not include
abutment effect from the third dimension, which would lead to stresses being overestimated
compared to a three-dimensional model Also it should be noted that zone sizes are larger in the
3DEC model, which further reduces the stress concentrations compared to those in the Phase2
models.
Figure 14. Comparisons between 3DEC and Phase2 Major Principal Stress distribution on a line traced along evaluation plane C1 in the hangingwall of the Printzsköld orebody and monitored onevaluation plane L1. The 0 m point is the starting point at the toe of the cave and the 1200 m is at the ground surface.
For both the 3DEC and the Phase2 models, the 970 m level stress behavior shows a high
stress build-up at the 300 m distance mark because this is the toe of the cave adjacent to the
position of the line in the hangingwall. For the 1225 m level this point has become destressed and
so the high stresses have been redistributed to lower levels. Destressed regions were observed in
the hangingwall on all the three evaluation planes in the longintudinal and cross sectional
directions. In the cap rock, as seen from Figures 9 and 14, there does not seem to be much change
in the major principal stress when mining from the 970 to the 1225 m level. This could be because
the thickness of the crown pillar remains the same when mining to the deeper levels.
Further stress evaluation was conducted in the hangingwall and the cap rock of the
Printzsköld orebody. Two evaluation lines were established, one starting from the cave back going
veritcally upwards to the ground surface with length of about 300 m (A) and the other from the
cave boundary generated after mining the 1225 m level to about 1000m into the hangingwall
oriented in the north-eastern direction and inclined at a dip of approximately 30° (B), see Figure
15. The graphs in Figure 15 also show the plots of major and minor principal stresses against
distance along the evaluation plane into the hangingwall. These plots illustrate the variations
between the two principal stresses along the lines A and B.
Figure 15. Evaluation lines in the cap rock and hangingwall of the Printzsköld orebody.
Major and minor principal stresses were also plotted along the two lines, with the results
shown in Figures 16 and 17. In the cave back along line A, both minor and major principal stresses
show a steady decrese when moving upward from the boundary of the cave. The stress build-up in
the immediate cave back is accompanied by high confinement. In the hangingwall (along line B),
the minor principal stress is about 5 MPa less than the major principal stress. The graph for line B
shows a lack of confinement at the cave boundary in the hangingwall.
,,
020040
060080
0100012
00
-1
0
1
2
Dist. from hangingwall (
m)
13
050
100150200250300350
0 10 20 30
Dist
. fro
m c
ave
back
(m)
1 3
Figure 16. Major and minor principal stress profiles in the cap rock of the Printzsköld orebody
along line A.
Figure 17. Major and minor principal stress distribution along line B from the cave boundary to
about 1000 m into the hangingwall.
0
50
100
150
200
250
300
350
0 10 20 30 40
Dist
ance
from
Cav
e ba
ck
Stress [MPa]
Major
Minor
surface
0
200
400
600
800
1000
1200
-5 0 5 10 15 20
Dis
tanc
e fr
om o
rebo
dy b
ound
ary
[m]
Stress [MPa]
Major
Minor
1000 m into the hangingwall
Plots of the major and minor principal stresses were used to check for confnement in the
crown pillar and the hangingwall. It was evident that there was a high stress confinement at the
cave back along line After mining simulated for the 1225 m level, the hangigwall showed an
increase in confinment along line B from about 300 m to 800 m from the cave boundary.
Thereafter there was a reduction in confinment at about 1200 m into the hangingwall.
3.2. Plastic modeling
3.2.1 The base case
The base case scenario produced results shown in Figures 18 for mining of levels 970 and 1225 m,
respectively. The results showed that tensile failure is dominating in the hangingwall, whereas
shear failure develops below the active mining level. As mining goes deeper, the area of yielding
increases. Also, mixed shear and tensile yielding is developing in the lower part of the
hangingwall, and shear failure can also be noted in the footwall. The area of hangingwall yielding
in this cross-section for mining the 970 m level is 35 600 m2 compared to 54 055 m2 for the
mining of the 1225 m level.
Figure 18. Evaluation plane L1 showing the element failure state for the base case values and for
mining to the 970 m level (top figure) and 1225 m level (bottom figure).
Figure 19 shows yielding patterns in the true and plunge hangingwall sections after mining
level 1225 m. Shear and tensile failure are the main failure mechanisms identified in the
hangingwall on plane C1. The plunge hangingwall at this level fails primarily in shear. On the
evaluation planes C2 and C3 similar tensile failure patterns were observed.
Figure 19. Longitudinal section of the true and plunge hangingwall of Printzsköld on plane C1, 50
meters from the cave boundary after complete mining of the orebody (i.e. 1225 meters level).
3.2.2 Effects of varrying cohesion
The sensitivity of the model to changing strength parameters was studied. This rock mass has
several weak zones and biotite schists, which affect the rock mass strength. In this initial study, the
strength was changed for the entire hangingwall rock mass, to increase the understanding of what
strength parameters are important to quantify in the continued work. Figures 20 and 21 show the
results from the cases of high and low cohesion (cf. Table 2), respectively, which may be
compared to the base-case shown above. The results showed a marked difference in the extents of
volumes of yield in the hangingwall with a larger yielding obtained for the case of the lower
cohesion value.
Plunge Hangingwall
True Hangingwall
Near cave back
Failure State: Base Case
Level 1225 m
Ground surface
Figure 20. Yielding resulting from high cohesion value in the hangingwall rock mass shown on a
vertical cross-section along the evaluation plane L1, for mining to the 1225 m level.
Figure 21. Yielding resulting from low cohesion value in the hangingwall rock mass shown on a
vertical cross-section along the evaluation plane L1, for mining to the 1225 m level.
3.2.3 Effects of varying tensile strength
Tensile strength was increased to twice that of the base-case values for the high value case and
reduced to zero for the low value case, cf. Table 2. Figures 22 and 23 show the results for the two
cases. It can be seen that the overall area of yielding in the hangingwall and cap rock does not
significantly change as the tensile strength is changed. However, the failure mechanism changes
from shear to tensile failure, as could be expected. Moreover, for the low strength case, some
yielding near the ground surface can also be observed.
Figure 22. Yielding results from high tensile strength value in the hangingwall rock mass evaluated
on the vertical cross section along evaluation plane L1, for mining to the 1225 m level.
Figure 23. Yielding results from low tensile strength value in the hangingwall rock mass evaluated
on the vertical cross section along evaluation plane L1, for mining to the 1225 m level.
3.2.4 Comparison of sensitivity between cohesion and tensile strength changes
The changes observed for the variations of cohesion are higher in terms of area of yielding,
compared to the cases with varied tensile strength. However, changing the tensile strength affected
the yielding mechanism. It was clear from model simulations that cohesion played an important
role in the amount of yielding of the hangingwall rock mass. The type of yielding for low cohesion
was more of shear while that for high cohesion was predominantly tensile. Evaluation plane L1
was used to estimate the areal changes on this plane, which represented a volume variation along
the evaluation plane. The variations shown in Figure 24 show the differences in the area of the
hangingwall affected by the cohesion changes (in terms of tensile or shear yielding) on the L1
evaluation plane.
Figure 24. Histogram showing the amount of yielding in the hangingwall in response to changes in
the cohesion values.
The plastic analysis results in 3DEC were further compared to those presented in Umar et
al. (2013). It was found that the results from the Phase2 model presented more conservative results
in the hangingwall yielding than those obtained in 3DEC, (compare Figure 25 and Figure 18
above). The results showed that there was a larger area of yielding noted in the Phase2 model
compared to that in 3DEC. As stated in the section 3.1, the plane strain assumption used in Phase2
will generally result in exaggerated yielding compared to a 3D-model.
Area Change inHWArea of yielding
Figure 25. Yielded elements in the hangingwall for mining of levels 970 and 1225 meters
calculated in the Phase2-model (Umar, et al., 2013).
4 Discussion
The results from the three-dimensional modelling presented in this paper have been compared with
results from two-dimensional modelling presented in [2] using Phase2. Elastic modelling showed a
good match with the Phase2 results especially regarding the stress redistribution pattern. However,
there is a difference in the values of the stress build-up especially in the cave back. The two-
dimensional Phase2-model gave higher stresses compared to those found from the 3DEC-model.
The primary reason for this is that the two-dimensional analysis assumes infinite cross sections
and so it eliminates the influence of the effect of abutment from the walls perpendicular to the
model plane. Hence, it can be reasonably assumed that the 3DEC model yields more realistic
stress values for a complicated orebody geometry such as the one for the Printzsköld orebody.
Plastic modelling revealed two failure types in the hangingwall; tensile and shear failure
mechanisms. The base case scenario showed that failure at the hangingwall was mainly tensile
followed by shear in some specific places.
Varying strength parameter values led to changes in the behaviour of the model in the
hangingwall. It was found that lowering cohesion in the hangingwall caused an increase in the
overall volume of yielding in either tension or shear failure. Since there are weak zones in this rock
mass, this finding can be used to estimate failure extents within the hangingwall rock mass with
low cohesion such as biotite schist zones found in the Printzsköld orebody.
The destressed zone in the true hangingwall can be thought of as a beam “fixed” to the cave
back and the toe of the hangingwall-cave boundary (see Figure 26). Bending of this "beam" will
induce tensile failure in the hangingwall. The cave back, on the other hand, is likely to fail in shear,
due to high compressive stresses, and thus advance upwards. This would lead to an extension of the
exposed hangingwall "beam", which in turn will promote additional tensile failure, thus resulting in
progressive cave growth.
Figure 26 Failure mechanism forecasted for the Printzsköld orebody for continued mining toward
depth.
The beam effect experienced in the hangingwall. The hangingwall is likely to fail in tensile
Toe of the cave
Ground surface
Stage one Rock mass e Rock mass
Mined out void
Ground surface
Tensile and shear failure will lead to extended cave in to the hangingwall.
Stage two
Cave back advances upwards due to shear failure
5 Conclusions and recommendations
The hangingwall of the Printzsköld orebody was analyzed numerically using elastic and plastic
material models. The following conclusions can be drawn from this study:
As mining deepens there are high stress build-ups in the cave back. Lowering the cohesion
has an overall effect of increased yielding in the hangingwall. Simulations with high
cohesion values produces very little yielding n the hangingwall as expected.
The tensile strength variation had a limited effect on yielding in the hangingwall compared
to that from variations in cohesion.
The failure pattern observed in the hangingwall was a mixture of tensile failure and shear
failure. Shear failure was observed mostly near the crown pillar and also at the toe of the
hangingwall.
It is recommended that for future work more efforts should be made in the simulation of
caving, cave geometry and cave advance in the Printzsköld orebody.
This simulation has been undertaken without considering the geologic structures that may
be present in the rock mass. It is therefore envisaged that further analysis of this orebody
should incorporate potential large-scale geological features to address their possible effects
on the stability of the hangingwall.
Acknowledgements
This paper was written as part of an ongoing research project for Loussavaara Kirunavaara AB
(LKAB) Malmberget Mine, Sweden. The project is funded through the Hjalmar Lundbohm
Research Centre (HLRC). The support from LKAB and the Centre for Applied Mining and
Metallurgy is acknowledged. The authors would also like to thank the following members of the
research team for their various roles they played in making this research a smooth run: Jimmy
Töyrä, Linda Jonsson, Joel Andersson and Fredrik Ersholm, all from LKAB. Also, thanks go to the
assistant supervisor of the first author, Catrin Edelbro, as well as to Professor Erling Nordlund for
their guidance and support during this research.
References
[1] LKAB., 2014. Malmberget Mine Level. Luleå: LKAB. Available from:
https://www.lkab.com/en/Future/Urban-Transformations/Why/New-Mine-
Level/Malmberget-Mine-Level/.
[2] S.B. Umar, J. Sjöberg, and E. Nordlund, Rock Mass Characterization and Conceptual
Modeling of the Printzsköld Orebody of the Malmberget Mine, Sweden. Journal of Earth
Sciences and Geotechnical Engineering. 3 (2013).
[3] Rocscience Inc., Phase2 Version 6.028 - Finite Element Analysis for Excavations and
Slopes, 2008; Software available at: www.rocscience.com, Toronto, Ontario, Canada.
[4] Itasca Consulting Group Inc., 3DEC (3-Dimensional Distinct Element Code), Version 5.0.
Minneapolis: ICG, 2013
[5] F. Perman, J. Sjöberg, and C. Dahnér, Detailed three-dimensional stress analysis of
complex orebody geometry–model setup and results for the Malmberget Mine. In:
Proceedings of the 2nd International FLAC/DEM symposium. Melbourne, 2011. Paper N0.
02-04.
[6] J. Sjöberg, Three-Dimensional Unit Stress Tensor Modeling of Complex Orebody
Geometry. In Proceedings of: The 42nd U.S. Rock Mechanics Symposium (ARMA). 29
June-2 July, San Francisco, CA, USA, 2008
PAPER C
´
Umar, S.B. and Edelbro, C. 2014, Influence of large-scale structures on the stability of the hangingwall in a caving mine - a modeling study. To be submitted to an international journal.
Influence of large-scale structures on the stability of the hangingwall in a caving mine - a modeling study.
Sraj Banda Umar1 and Catrin Edelbro2
Abstract
The Printzsköld orebody in the Malmberget mine, located in the Northern Sweden in the Municipality of Gällivare, is currently under production and has a central location in the mining area. Recent caving activities in neighboring orebodies have prompted studies to determine the effects of the on-going sublevel cave mining in this orebody, initially with emphasis on the effects on stress redistribution and potential rock mass yielding. Previous continuum numerical modeling analysis has shown that as mining deepens, a mixture of shear and tensile failure mechanisms is expected to occur in the surrounding rock. In the present study, three-dimensional discontinuum modeling of the Printzsköld orebody is presented, in which the effect of three pre-existing large-scale geological structures have been studied. The results showed that stress build-up reduced in the crown pillar and the cave bottom, compared to the results from the continuum model, due to shear slip developing along the structures. However, the presence of large-scale structures has no significant effects on the far-field stresses as shear slip along these structures is confined to the cave boundaries. A sensitivity analysis of the shear strength of the structures showed that reducing their friction angle had a larger effect on the slip along the structures compared to reducing the cohesion.
Keywords: large-scale structures, shear slip, hangingwall stability; Printzsköld orebody; failure mechanism; numerical analysis
1 Corresponding Author: Luleå University of Technology. Email: [email protected] Luleå University of Technology. Email: [email protected]
1. Introduction
1.1. Caving due to mining
Caving of the rock mass is a necessary component in cave mining. In sublevel caving, the mining method is based on the principle of gravity flow of the blasted ore rock (Brunton, et. al., 2010). To understand the processes that take place during caving, Duplancic and Brady (1999) suggested a conceptual model for the caving zone, in which various regions of the zone were categorized with respect to their seismic and discontinuity responses, see Figure 1.
Figure 1 The conceptual model of the caving zone as proposed by Sainsbury (2012), courtesy of Duplancic and Brady (1999).
The descriptions of the different zones in Figure 1 are (Duplancic and Brady, 1999; Sainsbury, 2012):
Elastic zoneIn this zone the rock mass surrounds the caving zone and behaves mostly elastically, with properties consistent with an undisturbed rock mass.
Seismogenic zoneIn this region there is a concentration of seismic activities. The active seismic front that occurs here results from slip on geologic structures and failure of intact rock. This is caused by the redistribution of stresses during mining and cave advance.
The yielded zone.
This is a zone in which rock has failed and cannot provide support to the overlying rock mass. It is characterized by large scale displacements of rock masses, which is subjected to significant damage (Sainsbury, 2012).
The air gap
This is space left between the mobilized zone and yielded zone. The air gap is controlled by rate of draw from below and the bulking porosity of the broken material. If the rate of draw is very low the material may fill the cave and further yielding and cave advance upwards may be stopped. Furthermore, at this point the material in the cave may provide support to the walls of the cave (Villegas, 2011).
The mobilized zone
This zone is characterized by dislodged or fallen rock blocks from the cave back. The rock usually has displacements of more than 1 - 2 m (Sainsbury, et al., 2011a).
The ground surface above the caving zone suffers large scale deformations. These deformations can be evidenced by cracking, stepping, sinkhole formation, etc. Categorization of these zones was done by e.g. Herdocia (1991) and Lupo (1998). In sublevel caving, the deformation zones differ as suggested by Lupo (1997) in relation to the effects on the hangingwall. In this paper the following descriptions of the deformation zones have been adopted (Lupo, 1998):
In the caved zone a downward movement of caved material is present which has been formed by collapse of material from the side walls and back of the cave.The fractured zone is characteristic of tension cracks, steps, fractures and sinkholes or pits distributed randomly in the caving zone. In this zone it is expected to find unstable parts as failure through toppling and shear can occur.The continuous deformation zone is characterized by the development of continuous deformation. The variation in ground level and coordinates of positions can be detected through surveying activities.
1.2. Influence of large scale structures on caving
Large scale structures are believed to influence the cap rock stability as well as the direction of caving and the propagation and shape of the cave formed. Vyazmensky et al. (2010) found that the
direction of the cave propagation is highly influenced by the orientation of large-scale structures, such as faults. Following from the conceptual model of caving by Duplancic and Brady (1999),Sainsbury et al. (2011) used FLAC3D (Itasca, 2009) to determine the influence of large-scale structures and it was found that the crater formed during caving and its progression followed the orientation of faults, if present in the cave area. Despite it being a continuum approach Sainsbury et al. (2011) used FLAC3D to construct a ubiquitous joint model, in which the influence of geologic structures were simulated in a caving zone. This result is important in relating cave advance with respect to the orientation of large scale structures. The influence of geologic structures has also been discussed in Sainsbury et al. (2008); Sainsbury et al. (2009); Sainsbury (2010) and Lupo (1997).
1.3. The Printzsköld orebody
The Printzsköld orebody is one of approximately 20 orebodies making up the Malmberget mine in Sweden, out of which about 10 are currently in production. The mining method in this orebody is sublevel caving with the current mining level being the 970 m level. The orebody is located almost in the middle of the mining area, does not daylight and strikes at about 40° from the magnetic north with a dip of approximately 60° towards the south (Wettainen, 2010). The average thickness of the orebody is 50 meters and is situated about 600 meters below the ground surface (780 m level in the mine coordinate system). The average strike length on the upper part of the orebody is around 400 meters while for the lower part, the orebody is more than 900 m in length along strike.
The deposit is a paleproterozoic succession of greenstones, porphyries and clastic meta-sediments which are hosted by metavolcanics that have been intruded by pegmatites and granites. The volcanic rocks have been transformed to sillimanite gneisses with quartz, muscovite, and local andalusite by the young granite intrusions. The iron ores are characterized by coarse magnetite and variable horizons of apatite with local sections rich in hematite. Generally, the border zones of the ore are characterized by skarn zones interpreted to be related to the formation of the ore (Romer, 1996). Locally named rock formations found in the Printzsköld orebody host rock are red leptite; grey leptite, red-grey leptite, magnetite and granite. There is also a band of biotite schist in the boundary between the orebody (magnetite) and grey leptite.
Similar to many areas in Northern Sweden, the Malmberget deposit is characterized by NW-SE trending shear zones (Romer, 1996). These zones are thought of as resulting from a complex geodynamic evolution which included repeated extensional and compressional tectonic regimes associated with magmatic and metamorphic events (Skiöld, 1988).The rock mass of the Printzsköld orebody has also been described in e.g., Martinsson and Hansson (2004) and Umar (2013).
Magnor and Mattsson (2010), identified a number of potential large-scale structures in the Malmberget area. Table 1 shows descriptions of the large scale structures inferred in the vicinity of the Printzsköld orebody, adapted from Mattsson and Magnor 2010, as reported in (Wettainen, 2010).
Table 1 Classification of the large-scale structures influencing rock mass stability in the Printzsköld orebody, adapted from Wettainen, 2010.
Structure ID Classification Dip/DD*[°] Description
DZ050 Brittle structure 70/315Indicated by magnetic and gravimetric measurements and core drillholes. Length estimated at 400 meters and width at 5 meters.
DZ031Probably brittle structure
80/003Indicated by magnetic measurements, resistivity and seismic reflection. Length estimated at 900 meters and width at 15 meters
DZ032Brittle fracture (high fracture frequency)
68/167Indicated by field observations above ground and by resistivity. Length estimated at 500 meters and width at 5 meters.
*DD = Dip Direction
An illustration of the large-scale structures in relation to the 920 m mining level in the Printzsköld orebody is shown in Figure 2. These large-scale structures are expected to affect the hangingwall and the cave that develops as mining progresses toward depth.
Figure 2 Illustration and setting of three large-scale structures in the Printzsköld orebody in relation to the 920 m mining level (Wettainen, 2010) shown in a) a plan view, and b) a longitudinal cross-section
1.4. Problem statement
Two conceptual models have been analyzed for the Printzsköld orebody to determine the stress patterns in the hangingwall as mining deepens. The first, a conceptual two-dimensional analysis was run in which strength parameter sensitivities were also evaluated by Umar et al. (2013). The second model was a three-dimensional model, in which further studies of stress distribution and rock mass yielding was conducted (Umar, et al., 2014.). These two models were carried out using a continuum approach. However, the presence of the inferred large-scale structures (described above) may affect the rock mass behavior and stress re-distribution around the orebody. In order to take into account the large-scale structures a discontinuum approach is required.
1.5. Objectives and scope of this paper
The objective of this paper is to analyze the effect of the large scale structures on failure mechanism and the cap rock stability of the Printzsköld orebody. Hence, this study seeks to investigate the characteristics and type of yielding in the crown pillar and the hangingwall. The paper focusses on the large-scale structures which have been inferred for the Printzsköld orebody (Magnor and Mattsson, 2010). Small scale structures have not been considered because of the model size and that those structures are not persistent enough in comparison to the model scale.
2. Numerical modeling
2.1 Modeling approach, set-up and input data
In this study, the three-dimensional discrete element code 3DEC (Itasca Consulting Group Inc., 2013) was used as the modeling software. The choice of modeling tool was governed by the ability of 3DEC to model discontinuum behavior, including rotation, separation and slip between blocks, as well as deformation and yielding within the rock blocks. The model is shown in Figure 3, in which the red part is the Printzsköld orebody and the purple planes are the large-scale structures included in this model from the specifications as described in Table 2. The model is of mine-scale, covering 5000 by 3000 by 2150 meters.
Several assumptions were made with regard to the modeling of the Printzsköld orebody. It was assumed that the mining taking place in the nearby orebodies did not have an effect on the stress redistribution in the Printzsköld orebody. The current cave was simulated as a void from the current situation, which has progressed to about 300 m below the ground surface. Lastly, cave mining was simulated by removing one sublevel at a time, and as mining progressed a void was left in mined areas.
Figure 3 Model set-up showing large scale structure positions in the Printzsköld orebody.
The elastic constants and the strength parameters for the block material (between the structures) were taken from Sjöberg (2008), as shown in Table 2. A Mohr-Coulomb perfectly-plastic material model was used in this study.
Table 2 Input parameters used for the elastic-perfectly plastic model (Sjöberg, 2008).
Area Strengthparameter
Value
Hangingwall
c [MPa] 5.18
[°] 50.7
tm [MPa] 0.71
Orebody
c [MPa] 4.81
[°] 50.7
tm [MPa] 0.48
Footwall
c [MPa] 6.67
[°] 52.9
tm [MPa] 1.3
c = cohesion; = internal friction angle; tm = tensile strength
N
The structures were simulated as joint planes and strength parameters obtained through an internal mine study of the joint shear strength in the Printzsköld orebody. A Coulomb slip model was assumed for the joints. Three simulations were conducted; the first of which was the base case. Cohesion and internal friction angle were varied as shown in Table 3 for the remaining two model simulations. This was used to determine their effect on the stability of the hangingwall and the crown pillar as mining proceeded to lower levels.
Table 3 Strength parameters for the large scale structures in the Printzsköld orebody obtained (from shear tests of the Printzsköld orebody joints).
Case c (MPa) (°) tm (MPa) ks(GPa) kn (GPa)
Base case 0.64 37 0 10.3 28.1Low cohesion 0 37 0 10.3 28.1Low friction angle 0.64 27 0 10.3 28.1
c = cohesion; = internal friction angle; tm = tensile strength; ks = shear stiffness; kn = normal stiffness
Initial stresses used were taken from a study by Sjöberg (2008) in which the vertical stresses were set equal to the pressure from overlying rock material and the horizontal stresses were found to be (through stress calibration against conducted stress measurements):
zH 0358.0 (2.1)
zh 0172.0 (2.2)
z is the vertical depth in meters and all stresses are in MPa. Here H is the maximum horizontal stress and h is the minimum horizontal stress with H having an orientation of 130.6° from the local north of the mine. The boundary conditions used were zero-velocity ("roller") boundaries on the vertical boundaries of the model, as well as for the bottom boundary. The ground surface was modeled as a free surface.
2.2 Evaluation planes.Six evaluation planes (C1, C2, C3 and L1, L2, L3) were used in this study, see Figure 4 and 5. The locations of the “C and L” evaluation planes were chosen based on the need for comparisons of stress redistributions in both the hangingwall and cap rock. The evaluation of the result were conducted in a similar way as in Umar et al., (2014), in order to be able to compare the results from continuum and discontinuum models. Yielding was studied on two evaluation planes C1 (situated about 50 m from the cave back) and L1 (cutting through the cave back).
Figure 4 Evaluation planes C1, C2 and C3 in the vertical cross-sectional view of the Printzsköld orebody.
Figure 5 Evaluation planes in the longitudinal view of the Printzsköld orebody.
C3 C2 C1
Hangingwall
Cave
Footwall
Printzsköld Orebody
0 200m
970 m level
1225 m level
0 200m
L3 L2 L1
3. Model results
The results from the discontinuum modeling in this study were compared with previous continuum model studies to determine the possible effect of including large-scale structures in the model. These comparisons were made on the orebody axis, and the C1 and L1 evaluation planes. Both stress distribution and rock mass yielding was studied. Additionally, the developed slip along the large-scale structures was analyzed and the potential effects on the stability of the hangingwall and crown pillar of the orebody assessed.
3.1. Stress analysis
Three discontinuum cases were simulated and as an example the stress redistribution results for evaluation plane L1 after mining the 1225 m level are presented in Table 4. Presented along in Table 4 are continuum modeling stress results from Umar et al. (2014). The evaluation plane L1 represents the vertical cross-section along the Printzsköld orebody at which maximum stress redistribution in the cap rock and hangingwall would be more evident compared to L2 and L3 planes. There were no significant differences observed between the three cases. However, lower stress build-up was observed in the discontinuum model compared to the continuum models.
The maximum stresses in the crown pillar and cave bottom area was calculated in each of the models and compared. Note that these values are not from the exact same location, but represent the maximum stress within the volume above the cave top and below the cave bottom, respectively. The maximum stresses in the two areas were 72 and 192 MPa, respectively for the continuum model, and 48 and 144 MPa respectively for the discontinuum model. These stresses have been presented in Table 5. The two areas were selected for stress observation because they were the most sensitive sections to stress changes in the simplified geometry of the Printzsköld orebody as mining progressed. The continuum models have higher stresses in the cap rock and the cave bottom compared to those calculated in the discontinuum model. The stresses were calculated from the cap rock roof and the cave bottom and evaluated on the L1 evaluation plane. The difference in stress build-up is attributed to the inclusion of the structures in the discontinuum model. Comparisons of the destressed regions in the hangingwall did not show significant differences in terms of area of destressing.
Table 4 Results for the continuum and discontinuum models, on evaluation plane L1, for the Printzsköld orebody after mining the 1225 meters level.
Case description
Structure strength parameters used
Model results
Base case discontinuum model
cj = 0.64 MPaj = 37°tj = 0 MPa
Low cohesion discontinuum model
cj = 0 MPaj = 37°tj = 0 MPa
Low friction angle discontinuum model
cj = 0.64 MPaj = 27°tj = 0 MPa
Base case continuum model
No structures included in the model
cj = Cohesion [MPa]j = Friction angle [°]tj = Tensile strength [MPa]
Table 5 Major principal stresses calculated from the 3DEC model in the cap rock roof and cave bottom; and results comparison for the continuum and discontinuum models
Orebody area under comparison
Continuum model stress recorded (MPa)
Discontinuum model stress recorded (MPa)
Crown pillar 72 48Cave bottom 192 144
An evaluation along a longitudinal section parallel to orebody strike and passing through the orebody axis was done and a comparison of stress levels in the plunge hangingwall conducted. The results are presented in Figures 6 (a) and (b). There is a slightly reduced build-up of stresses in the plunge hangingwall in the discontinuum model (Figure 6) as compared to that seen in the continuum model (Figure 6).
Figure 6 Stress patterns in the longitudinal section calculated along the center plane of the orebody (a) in the discontinuum model and (b) in the continuum model, after mining the 1225 meter level.
(a) Discontinuum Model Results
(b) Continuum Model Results
3.2. Joint slip and rock mass yielding
It was observed that there was slip on all the large-scale structures especially around the cave boundaries. The calculated slip for the three different strength cases is presented in Table 6. It can be seen that changing the strength parameters has an effect of increased slip area along two of the three large scale structures. This is believed to be linked to the setting of the structures. These two structures have in common the near parallel strike to that of the maximum horizontal stress direction compared to DZ032. Furthermore, DZ032 has a shallower dip compared to DZ031 and DZ050. From Figure 7 it is also clear that changing the friction angle was much more sensitive than changing the cohesion with respect to area of slip on the structures. The effects of this are evident for the two structures, DZ031 and DZ050. Thus, it appears that the normal stress is a major factor, with the orientation of the structures relative to the largest stress determining the amount of slip and the sensitivity to strength changes.
Table 6 Numerical modeling results showing the calculated slip (in brown) on the three structures with the various strength parameters after mining the 1225 m level. Viewed from the south and parallel to the strike of the Printzsköld orebody.
Case Joint DZ031 Joint DZ032 Joint DZ050Base case
Low cohesion
Low Friction angle
Figure 7 Calculated slip areas on the modeled large-scale structures for the three different strength cases simulated in the model of the Printzsköld orebody and for mining to the 1225 m level.
Except in the cases where the strength parameters were changed, slip was observed to be confined mainly to the mine openings (cave). There was also tensile failure observed along all the structures in both the hangingwall and footwall, see Figure 8.
Base Case
Low cohesion
Low friction
0
50000
100000
150000
200000
250000
300000
350000
400000
450000
500000
Jnt IdDZ031 Jnt Id
DZ032 Jnt IdDZ050
Are
a un
der s
lip (m
2)
Large Scale Structure Id
Base Case
Low cohesion
Low friction
Figure 8 Calculated slip along the large-scale structures after mining down to 1225 m. Red spots indicate tensile failure along the structures.
Yielding in the rock mass was evaluated on the C1 and L1 evaluation planes, see Figures 9 and 10. Comparisons were made on the yielding pattern in the rock mass between the results obtained from the Umar et al., (2014) continuum model and the current discontinuum model. However, it was found that there were no significant differences between the continuum and discontinuum yielding results, for both the vertical and longitudinal evaluation sections.
DZ050 DZ031
DZ032
Figure 9 Calculated failure in the Printzsköld orebody evaluated on the L1 evaluation plane using 3DEC from the continuum model in Umar et al. (2014) in (a) and from the current discontinuum model in (b), after mining to the 1225 meter level.
DZ050
DZ032
DZ031
1225 m level
1225 m level(a)
(b)
Figure 10 Calculated failure in the Printzsköld orebody evaluated on the L1 evaluation plane using 3DEC from the continuum model in Umar et al. (2014) in (a) and from the current discontinuum model in (b), after mining to the 1225 meter level
Near cave back
Plunge hangingwall
Truehangingwall
(a)
DZ050
DZ031DZ032
1225 meter level(b)
4. Discussion
Results from the discontinuum model showed that there was no effect from the large-scale structures on the far-field stress regime in the Printzsköld orebody. The lack of effect on the yielding pattern in the hangingwall, despite the inclusion of these structures, could be associated in part with the joint strength parameters used to simulate these large-scale structures such as joint strength. This is because was difficult to assign actual large-scale structure strength parameters and the ones used were calculated for single selected joints in an internal mine study.
Cohesion and friction angles were varied and their contribution to slip monitored. The large area of slip observed along structures DZ031 and DZ050 resulted from the orientation of these structures as compared to the DZ032 structure. DZ050 and DZ031 have a near vertical dip and an orientation almost parallel to that of the maximum horizontal stresses. This setting gives a higher possibility for shear than when the structure is set at an oblique angle, similar to DZ032.
As shown in Figure 10 (b), all the three large-scale structures intersect the orebody at various positions. As they slip, the slipping surfaces present points of failure genesis in the caving of the hangingwall. It can be considered that the structures provide surfaces which can shape and affect the cave propagation and cave geometry (in the event of caving through shear failure) and advance through a failure cycle defined by the persistence of these structures. This is likely to happen in the structures where slip is higher such as DZ050 and DZ031. When the current cave boundary has failed, the process is likely to begin again as mining deepens and slip will still be confined to the cave opening until it fails in shear. This trend can go on as long as there is slip along the structures and as long as stress imbalances occur.
The result from the comparison between the continuum and discontinuum model can be interpreted in the following way: (i) there is a loss of stress build-up in the evaluated areas (cap rock and cave bottom). From this perspective, there is a reduction in the likelihood of failure induced as a result of the stress build-ups in these areas. (ii) the stress build-up lost has been due to the observed slip along the large scale structures. There is however, a likelihood that the slip along the structures can provide failure direction or caving advance especially that this slip is mostly confined to the cave boundaries.
5. Conclusions and recommendations
It has been found in this study that the large scale structures have no significant effect on the yielding of the hangingwall but that this effect is confined to the cave boundary. Stress build-up was reduced in the discontinuum model, compared to the previous continuum models. This can be attributed to the shear slip that takes place along the structures and plastic yielding of the rock mass has largely remained unaffected by the large scale structures.
Lowering the strength parameters of the large-scale structures resulted in an increased area of slip along all the three structures. However, structure DZ032 showed fairly little increase in slip area.
DZ031 and DZ050 had larger slip areas, because their strikes were near parallel with the direction of the maximum horizontal stress, thus resulting in higher shear stress and lower normal stress acting on these two structures
The inferred large-scale structures in the Printzsköld orebody, and probably the entire mine; need further verification of their existence and locations. Moreover, a better estimate of their mechanical properties is warranted. It is also recommended that future work should include studying the mining-induced seismicity in the cap rock and hangingwall. A mine-wide seismic network monitoring system has been in place at the Malmberget mine for several years and the collected data could be used to analyze patterns and identify linkages that may increase the understanding of both cave progression and how caving is affected by the structures.
Acknowledgements
This paper was written as part of an ongoing research project for Loussavaara Kirunavaara AB (LKAB) Malmberget Mine, Sweden. The project is funded through the Hjalmar Lundbohm Research Centre (HLRC). The support from LKAB and the Centre for Applied Mining and Metallurgy is acknowledged. The authors would also like to thank the following members of the research team for their various roles they played in making this research a smooth run: Jimmy Töyrä, Tomas Savilahti, Linda Jonsson, Joel Andersson and Fredrik Ersholm, all from LKAB. Also, thanks go to the Principal Supervisor of the first author, Adjunct Professor Jonny Sjöberg as well as to Professor Erling Nordlund for their guidance and support during this research.
References
Aksoy, C.O., Kose, H., Onargan, T., Koca, Y. & Heasley, K. 2004. Estimation of limit angle using laminated displacement discontinuity analysis in the Soma coal field, Western Turkey, International Journal of Rock Mechanics and Mining Sciences, vol. 41, no. 4, pp. 547-556.
Board, M. and Pierce, M., 2009. A review of recent experience in modeling of caving. InProceedings of: 43rd U.S. Rock Mechanics Symposium. June 28, 2009.
Cai, M., Kaiser, P.K., Uno, H., Tasaka, Y. and Minami, M., 2004. Estimation of rock mass deformation modulus and strength of jointed hard rock masses using the GSI system. International Journal of Rock Mechanics and Mining Sciences, 1, vol. 41, no. 1, pp. 3-19. Available from: http://www.sciencedirect.com/science/article/pii/S136516090300025X ISSN 1365-1609. DOI 10.1016/S1365-1609(03)00025-X.
Dershowitz, W. and Einstein, H., 1988. Characterizing rock joint geometry with joint system models. Rock Mechanics and Rock Engineering, vol. 21, no. 1, pp. 21-51.
Duplancic, P. and Brady, B.H., 1999. Characterization of caving mechanisms by analysis of seismicity and rock stress. In Proceedings of: 9th International Congress on Rock Mechanics. 1049-1053. Rotterdam.
Herdocia, A., 1991. Hanging Wall Stability of Sublevel Caving Mines in Sweden. Luleå, Sweden: Luleå University of Technology.
Hoek, E., 1974. Progressive caving induced by mining an inclined orebody. Trans Instn Min Metall, Sect A: Min Industry, vol. 83, pp. A133-139.
Itasca Consulting Group Inc., 2013. 3DEC (3-Dimensional Distinct Element Code), Version 5.0. Minneapolis: ICG.
Itasca, 2009. Fast Lagrangian Analysis of Continua in 3 Dimensions, Version 4.0.
Laubscher, D.H., 1994. Cave mining - the state of the art. Journal of South African Institute of Mining and Metallurgy, vol. 94, no. 10, pp. 279-293.
Lupo, J.F., 1998. Large-scale surface disturbance resulting from underground mass mining.International Journal of Rock Mechanics and Mining Sciences., vol. 35, no. 4-5.
Lupo, J.F., 1997. Progressive failure of hanging wall and footwall Kiirunavaara Mine, Sweden.International Journal of Rock Mechanics and Mining Sciences, 0, vol. 34, no. 3-4, pp. 184.e1-184.e11. Available from: http://www.sciencedirect.com/science/article/pii/S1365160997002773 ISSN 1365-1609. DOI http://dx.doi.org/10.1016/S1365-1609(97)00277-3.
Magnor, B. and Mattsson, H., 2010. Strukturgeologisk modell över Malmbergets gruvområde.Rapport. Hifab, Geovista.
Martinsson, O. and Hansson, K.E., 2004. DAY SEVEN FIELD GUIDE, Apatite Iron ores in the Kiruna Area. Department of Geosciences, Luleå University of Technology and LKAB.
Mas, I.D., Deisman, N., Pierce, M. and Fairhurst, C., 2007. The Synthetic Rock Mass approach-astep forward in the characterization of jointed rock masses. In Proceedings of: Proceedings 11th Congress of the International Society for Rock Mechanics, Ribeiro e Sousa, Ollala and Grossmann (editors), Taylor and Francis, Leiden. 485-490.
Mas, I.D., Pierce, M., DeGagné, D. and Darcel, C., 2008. Anisotropy and scale dependency in jointed rock mass strength-A synthetic rock mass study. In Proceedings of: Proceedings of the First International FLAC/DEM Symposium on Numerical Modeling. 231-239.
Perman, F., Sjöberg, J. and Dahnér, C., 2011. Detailed three-dimensional stress analysis of complex orebody geometry-model setup and results for the Malmberget Mine. Continuum and Distinct Element Numerical Modeling in Geomechanics Itasca International Inc. Minneapolis.
Rocscience., 2002. PHASE2. Finite Element Analysis and Support Design for Excavations. Toronto: Rocscience Inc.
Romer, R.L., 1996. U-Pb system of stilbite-bearing low-temperature mineral assemblages from the Malmberget iron ore, Northern Sweden. Geochimica Et Cosmochimica Acta, 6, vol. 60, no. 11, pp. 1951-1961 ISSN 0016-7037. DOI 10.1016/0016-7037(96)00066-X.
Sainsbury, B., 2010. Sensitivities in the numerical assessment of cave propagation in Caving 2010: Second International Symposium on Block and Sublevel Caving. 20-22 April 2010. Australian Centre for Geomechanics.
Sainsbury, B., 2012. A Model for cave propagation and subsidence assessment in jointed rock masses. Doctoral Thesis. The University of New South Wales. Australia.
Sainsbury, B.L., Sainsbury, D.P. and Pierce, M.E., 2011a. A historical review of the development of numerical cave propagation simulations. In Proceedings of: Proceedings of the 2nd International FLAC/DEM Symposium in Numerical Modelling. Feb 14-16. Melbourne, Australia.
Sainsbury, B., Pierce, M. and Mas Ivars, D., 2009. Analysis of Caving Behavior Using a Synthetic Rock Mass-Ubiquitous Joint Rock Mass Modeling Technique. In Proceedings of: Proceedings of the 1st Southern Hemisphere International Rock Mechanics Symposium (SHIRMS), Y. Potvin, J. Carter, A. Dyskin and R. Jeffrey (eds). 16-19.
Sainsbury, B., Pierce, M. and Mas Ivars, D., 2008. Simulation of rock mass strength anisotropy and scale effects using a Ubiquitous Joint Rock Mass (UJRM) model. In Proceedings of: Proceedings First International FLAC/DEM Symposium on Numerical Modelling. 25-27.
Sainsbury, D.P., Sainsbury, B.L., Board, M.P. and Lorig, D.M., 2011b. Numerical Back-Analysis of Structurally Controlled Cave Initiation and Propagation at the Henderson Mine. In Proceedings of: 45th US Rock Mechanics / Geomechanics Symposium. June 26-29.
Sjöberg, J., 2008. Three-Dimensional Unit Stress Tensor Modeling of Complex Orebody Geometry. In Proceedings of: The 42nd U.S. Rock Mechanics Symposium (ARMA). San Francisco, CA, USA. 08-124
Skiöld, T., 1988. Implications of new U-Pb zircon chronology to early Proterozoic crustal accretion in northern Sweden. Precambrian Research, 2, vol. 38, no. 2, pp. 147-164 ISSN 0301-9268. DOI 10.1016/0301-9268(88)90089-7.
Umar, S.B., 2013. Rock Mass Characterization of the Printzsköld and Fabian orebodies at the Malmberget Mine. Luleå, Sweden: Luleå University of Technology ISBN 978-91-7439-821-2.
Umar, S.B., Sjöberg, J. and Nordlund, E., 2013. Rock Mass Characterization and Conceptual Modeling of the Printzsköld Orebody of the Malmberget Mine, Sweden. Journal of Earth Sciences and Geotechnical Engineering, vol. 3, no. 4 ISSN 1792-9040.
Umar, S.B., Sjöberg, J. and Savilahti, T., 2014. Modeling of caving and deformation mechanisms of the hangingwall of the Printzsköld orebody in the Malmberget Mine. Submitted to: The International Journal of Mining, Reclamation and Environment.
Villegas, T., Nordlund, E. and Dahnér-Lindqvist, C., 2011. Hangingwall surface subsidence at the Kiirunavaara Mine, Sweden. Engineering Geology, vol. 121, no. 1-2, pp. 18-27 ISSN 0013-7952. DOI 10.1016/j.enggeo.2011.04.010.
Vyazmensky, A., Elmo, D. and Stead, D., 2010. Role of Rock Mass Fabric and Faulting in the Development of Block Caving Induced Surface Subsidence. Rock Mechanics and Rock Engineering, vol. 43, no. 5, pp. 533-556 ISSN 0723-2632. DOI 10.1007/s00603-009-0069-6.
Wettainen, T., 2010. Analys och prognostisering av uppblockning i Printzsköld. Examensarbete. 2010:143 CIV, Luleå University of Technology ISBN 1402-1617.
Woo, K., Eberhardt, E., Elmo, D. and Stead, D., 2013. Empirical investigation and characterization of surface subsidence related to block cave mining. International Journal of Rock Mechanics and Mining Sciences, vol. 61, pp. 31-42.
Woo, K., Eberhardt, E., Rabus, B., Stead, D. and Vyazmensky, A., 2012. Integration of field characterization, mine production and InSAR monitoring data to constrain and calibrate 3-Dnumerical modelling of block caving-induced subsidence. International Journal of Rock Mechanics and Mining Sciences, 7, vol. 53, no. 0, pp. 166-178 ISSN 1365-1609. DOI 10.1016/j.ijrmms.2012.05.008.
