UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Real-time grade estimation and online acceptance or rejection of mined material Nayak, Preetham 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2015_may_nayak_preetham.pdf [ 3.62MB ]
Metadata
JSON: 24-1.0167668.json
JSON-LD: 24-1.0167668-ld.json
RDF/XML (Pretty): 24-1.0167668-rdf.xml
RDF/JSON: 24-1.0167668-rdf.json
Turtle: 24-1.0167668-turtle.txt
N-Triples: 24-1.0167668-rdf-ntriples.txt
Original Record: 24-1.0167668-source.json
Full Text
24-1.0167668-fulltext.txt
Citation
24-1.0167668.ris

Full Text

REAL-TIME GRADE ESTIMATION AND ONLINE ACCEPTANCE OR REJECTION OF MINED MATERIAL  by  Preetham Nayak  B-Tech., National Institute of Technology Karnataka, 2010  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF   MASTER OF APPLIED SCIENCE  in  THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Mining Engineering)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  February, 2015  © Preetham Nayak, 2015 ii  Abstract The mineral industry is currently challenged by low metal prices and depletion of high grade resources with increasing depth combined with complex geology. Achieving profits in-spite of these difficulties has been challenging and the standard industry response to tackle the situation is by pursuing economy of scale. Only large scale companies can adopt this method due to the high capital investment involved. The alternate solution is to recognize and introduce low cost mining and mineral processing methods. The current research aims at increasing profitability of a mining operation by efficiently utilizing the mineral resource and increase the recovery. The research focuses on exploiting inherent heterogeneity in an orebody with the development of a sophisticated tool that can perform real time, with online analysis and decision support systems to designate the destination of the material being mined. The research includes measuring and modeling a deposit’s heterogeneity at both BHP Billiton’s Spence and Escondida mines. This is followed by the development of two digital models to evaluate the opportunity to adopt selective partitioning of ore, prior to hauling the mined material to the mill.  An apparent increase in profitability is observed due to increased ore recovery. Determining and exploiting this heterogeneity offers different insights in orebody grade distribution, evaluating resources and reserves and production planning, all leading to more complete utilization of the resource.       iii  Preface This dissertation is original, unpublished, independent work by the author, Preetham Nayak.                    iv  Table of Contents Abstract .......................................................................................................................................... ii Preface ........................................................................................................................................... iii Table of Contents ......................................................................................................................... iv List of Tables .............................................................................................................................. viii List of Figures ............................................................................................................................... xi List of Abbreviations ................................................................................................................. xiv Acknowledgements ..................................................................................................................... xv Chapter 1: Introduction ............................................................................................................. 1 1.1 Statement of Problem ....................................................................................................... 2 1.2 Objective of the Research ................................................................................................ 4 1.3 Significance of the Research Study .................................................................................. 7 1.4 Site Details ....................................................................................................................... 7 1.4.1 Spence Mine.............................................................................................................. 8 1.4.2 Escondida Mine ...................................................................................................... 12 1.5 Research Design ............................................................................................................. 14 1.6 Summary ........................................................................................................................ 15 Chapter 2: Sorting and Preconcentration: A Literature Review ........................................ 17 2.1 Why Preconcentration at the Excavation Stage? – A Closer Look at the Importance of this Application. ........................................................................................................................ 18 2.2 Sensor Based Sorting ..................................................................................................... 22 2.2.1 Optical Sorting ........................................................................................................ 24 2.2.2 Microwave Infrared Sorting .................................................................................... 25 v  2.2.3 XRT and DEXRT ................................................................................................... 27 2.2.4 XRF ......................................................................................................................... 28 2.3 Estimation of Resources ................................................................................................. 31 2.3.1 Kriging .................................................................................................................... 34 2.3.2 Polygonal Method ................................................................................................... 35 2.3.3 Inverse Distance Weighting Method ...................................................................... 36 2.4 Impact of SMU Size on Mineral Resource Estimation .................................................. 38 2.5 Conclusion ...................................................................................................................... 41 Chapter 3: Methodology .......................................................................................................... 43 3.1 Introduction .................................................................................................................... 43 3.2 Framework ..................................................................................................................... 44 3.3 Equipment and Procedure .............................................................................................. 47 3.3.1 Jaw Crusher ............................................................................................................. 47 3.3.2 High Frequency Electromagnetic Spectroscopy (HFEMS) .................................... 48 3.3.3 X-Ray Fluorescence (XRF) .................................................................................... 50 3.4 Procedure for Classification of XRF Scanned Samples for Assaying ........................... 51 3.5 Assaying ......................................................................................................................... 52 3.6 Development of Models ................................................................................................. 53 3.6.1 Numerical Mine Model ........................................................................................... 56 3.6.2 Vulcan Model.......................................................................................................... 58 3.6.3 Trade off Study between the Two Cases in Numerical Mine Model and Vulcan Model….. ............................................................................................................................... 62 Chapter 4: Results and Discussions ........................................................................................ 64 vi  4.1 Determination of Cu Content in the Samples................................................................. 64 4.1.1 HFEMS Results ...................................................................................................... 64 4.1.2 XRF Results ............................................................................................................ 65 4.1.3 Assay Results .......................................................................................................... 71 4.2 Correlation Analysis ....................................................................................................... 71 4.2.1 Correlation between XRF and Assay Results ......................................................... 71 4.2.2 Correlation between XRF and HFEMS .................................................................. 77 4.3 Numerical Mine Model .................................................................................................. 78 4.3.1 Effect of Increasing Selectivity ............................................................................... 78 4.3.2 Comparison in Savings of Sorted Model and Composited Model ......................... 82 4.4 Vulcan Model ................................................................................................................. 83 4.4.1 Effect of Changing Block Model Size .................................................................... 83 4.5 Conclusions from Numerical Mine Model and Vulcan Model ...................................... 86 Chapter 5: Discussion............................................................................................................... 88 5.1 Introduction .................................................................................................................... 88 5.2 Analysis of Introducing Preconcentration at Excavation Stage ..................................... 88 5.2.1 Application and Limitations of using XRF and HFEMS in Ore Grade Determination ........................................................................................................................ 89 5.3 Interpretation of Numerical Mine Model and Vulcan Model ........................................ 90 5.4 Conclusion .................................................................................................................... 103 Chapter 6: Conclusions and Recommendations .................................................................. 104 6.1 Introduction .................................................................................................................. 104 6.2 Thesis Summary ........................................................................................................... 105 vii  6.3 Contributions of the Research ...................................................................................... 105 6.4 Recommendations for Future Work ............................................................................. 106 6.5 Concluding Thoughts ................................................................................................... 108 Bibliography .............................................................................................................................. 109 Appendices ................................................................................................................................. 116                    viii  List of Tables Table 1: Spence Mineral Resources ................................................................................................ 9 Table 2: Spence Mineral Reserves ................................................................................................ 10 Table 3: Escondida Mineral Resources ......................................................................................... 12 Table 4: Escondida Mineral Reserves ........................................................................................... 13 Table 5: Sensors used in mineral processing industry and corresponding minerals..................... 23 Table 6: Difference between the initial estimate and the revised estimate ................................... 38 Table 7: Evaluation Model Assumptions ...................................................................................... 54 Table 8: Number of samples for testing in every zone ................................................................. 65 Table 9: Maximum, Minimum and Average grades for different zones ...................................... 69 Table 10: Classification of samples into three categories ............................................................. 70 Table 11: Correlation coefficients between the XRF and Assay results ...................................... 73 Table 12: Results of the Numerical Mine Model .......................................................................... 79 Table 13: Results of the Vulcan Model ........................................................................................ 85 Table 14: Results of increasing selectivity for SPPP07 zone- Case 1 .......................................... 91 Table 15: Results of increasing selectivity for SPPP07- Case 2 ................................................... 92 Table 16: Results of increasing selectivity for SPPP11- Case 1 ................................................... 93 Table 17: Results of increasing selectivity for SPPP11- Case 2 ................................................... 94 Table 18: Results of increasing selectivity for Spence Mine- Case 1........................................... 96 Table 19: Results of increasing selectivity for Spence Mine- Case 2........................................... 96 Table 20: Results of increasing selectivity for Oxido Zone ......................................................... 96 Table 21: Results of increasing selectivity for Mixto Zone .......................................................... 96 Table 22: Results of increasing selectivity for Sulfuro Zone ....................................................... 97 ix  Table 23: Results of increasing selectivity for M3 Zone .............................................................. 97 Table 24: Results of increasing selectivity for Escondida Mine- Case 1 ..................................... 98 Table 25: Results of increasing selectivity for Escondida Mine- Case 2 ..................................... 99 Table 26: XRF results of SPPP07 zone ...................................................................................... 116 Table 27: XRF results of SPPP11 zone ...................................................................................... 119 Table 28: XRF results of Oxdio zone ......................................................................................... 121 Table 29: XRF results of Mixto zone ......................................................................................... 124 Table 30: XRF results of Sulfuro zone ....................................................................................... 127 Table 31: XRF results of M3 zone .............................................................................................. 130 Table 32: XRF results of Lastre zone ......................................................................................... 133 Table 33: Assay results of SPPP07 samples ............................................................................... 135 Table 34: Assay results of SPPP11 samples ............................................................................... 135 Table 35: Assay results of Oxido samples .................................................................................. 136 Table 36: Assay results of Mixto samples .................................................................................. 136 Table 37: Assay results of Sulfuro samples ................................................................................ 137 Table 38: Assay results of M3 samples ...................................................................................... 137 Table 39: Assay results of Lastre samples .................................................................................. 138 Table 40: Correlation data for SPPP07 zone .............................................................................. 141 Table 41: Correlation data for SPPP11 zone .............................................................................. 141 Table 42: Correlation data for Oxido zone ................................................................................. 142 Table 43: Correlation data for Mixto zone ................................................................................. 142 Table 44: Correlation data for Sulfuro zone ............................................................................... 143 Table 45: Correlation data for M3 zone ...................................................................................... 144 x  Table 46: Correlation data for Lastre zone ................................................................................. 144                      xi  List of Figures  Figure 1: Heterogeneity in orebody leading to underestimation and overestimation of resources 4 Figure 2: Schematic diagram showing the flow sheet of mining activities considering two mining scenarios .......................................................................................................................................... 6 Figure 3: Spence and Escondida mine location .............................................................................. 8 Figure 4: Spence mine flow sheet ................................................................................................. 11 Figure 5: Escondida mine flow sheet ............................................................................................ 14 Figure 6: Schematic representation of optical sorter .................................................................... 24 Figure 7: Thermal image and profile of microwave heated sample ............................................. 26 Figure 8: Relationship between exploration results, mineral resources and ore reserves ............ 32 Figure 9: Construction of polygons for reserve estimation by Polygonal Method ....................... 36 Figure 10: Dilution and ore loss due to inappropriate size of SMU ............................................. 40 Figure 11: Comparison of grade and contained value between 2.5m, 5m and 10m SMU-based models ........................................................................................................................................... 41 Figure 12: Jaw Crusher at MineSense Lab ................................................................................... 48 Figure 13: HFEMS mounted on the table ..................................................................................... 49 Figure 14: XRF setup for testing samples..................................................................................... 51 Figure 15: The two cases considered in the development of models ........................................... 56 Figure 16: Top view of M3 drillholes ........................................................................................... 60 Figure 17: Sulfuro drillholes ......................................................................................................... 60 Figure 18: Triangulations of the five zones at Escondida mine .................................................... 61 Figure 19: Triangulation of the two zones at Spence mine .......................................................... 62 Figure 20: Cu % in every sample of SPPP07 zone ....................................................................... 66 xii  Figure 21: Cu % in every sample of SPPP11 zone ....................................................................... 66 Figure 22: Cu % in every sample of Oxido zone .......................................................................... 67 Figure 23: Cu % in every sample of Mixto zone .......................................................................... 67 Figure 24: Cu % in every sample of Sulfuro zone ........................................................................ 68 Figure 25: Cu % in every sample of M3 zone .............................................................................. 68 Figure 26: Cu % in every sample of Lastre zone .......................................................................... 69 Figure 27: Distribution of samples into three categories based on Cu % ..................................... 71 Figure 28: Correlation between XRF and Assay results for SPPP07 samples ............................. 73 Figure 29: Correlation between XRF and Assay results for SPPP11 samples ............................. 74 Figure 30: Correlation between XRF and Assay results for Oxido samples ................................ 74 Figure 31: Correlation between XRF and Assay results for Mixto samples ................................ 75 Figure 32: Correlation between XRF and Assay results for Sulfuro samples .............................. 75 Figure 33: Correlation between XRF and Assay results for M3 samples..................................... 76 Figure 34: Correlation between XRF and Assay results for Lastre sample .................................. 76 Figure 35: Cost analysis of Composited model and Sorted model for Spence Mine ................... 82 Figure 36: Cost analysis of Composited model and Sorted model for Escondida Mine .............. 82 Figure 37: Ore- Waste distribution for SPPP07 zone- Case 1 ...................................................... 91 Figure 38: Ore- Waste distribution for SPPP07 zone- Case 2 ...................................................... 92 Figure 39: Ore- Waste distribution for SPPP11 – Case 1 ............................................................. 94 Figure 40: Ore- Waste distribution for SPPP11 – Case 2 ............................................................. 95 Figure 41: Ore- Waste recovery in Escondida Mine .................................................................... 98 Figure 42: Ore- Waste distribution for Escondida Mine- Case 1 ................................................. 99 Figure 43: Ore- Waste distribution for Escondida Mine- Case 2 ................................................. 99 xiii  Figure 44: Recovery of copper in Spence Mine ......................................................................... 101 Figure 45: Recovery of copper in Escondida Mine .................................................................... 101     xiv  List of Abbreviations HFEMS JORC NMM ROM SMU SX-EW TPD UBC XRF  High Frequency Electromagnetic Spectroscopy  Joint Ore Reserves Committee  Numerical Mine Model Run of Mine Selective Mining Unit Solvent Extraction and Electrowinning Tonnes Per Day  University of British Columbia   X-Ray Fluorescence        xv  Acknowledgements I would like to acknowledge and express the deepest appreciation to my supervisor, Dr. Michael Hitch for his guidance, assistance and motivation during my research work. I would like to thank my parents for their unwavering love and support throughout this period. Much thanks to all the students, professors and administrators of the University of British Columbia’s Norman B. Keevil Institute of Mining Engineering whose kind gestures are too numerous to count.  I am very grateful to Dr. Andrew Bamber for the support and guidance provided by him during my research study. I am also thankful for the opportunity provided by him for me to work at the MineSense laboratory. I am very grateful to Mr. Bahjat Koshaba for the training he offered me on usage of instruments working on High frequency electromagnetic sensors and X-Ray fluorescence. I owe sincere and heartfelt thanks to BHP Billiton for letting me to use the samples for the study and also being a part of the research work. Special thanks to Ansu Alex and my friends who encouraged me to stay focused and optimistic throughout this process. Finally, as always, my most sincere and deepest appreciation goes to my sisters, whom without this would be impossible. Thank you.  1  Chapter 1: Introduction Beneficiation of ore by sorting has been practiced in the mining industry since the Bronze Age, but its application to date has been intermittent and benefits not realized to its fullest extent (Manouchehri, 2003). Early sorting techniques relied on  simple methods such as hand sorting, and gradually evolved  to methods such as gravity separation, magnetic or electric separation, flotation and later with the advance of automation- sensor based sorting was introduced (Wotruba & Harbeck, 2010; Habashi, 2006). Sorting ore and waste from the mined material before hauling it to the mill is known as preconcentration (Weatherwax, 2007). Sorting has proven benefits significant to the mining industry especially when used at the early stages of mining. Successful application of sorting as preconcentration has been seen in several mines including Xstrata’s Mt Isa mine in Australia, Copper Creek mine in USA and Sullivan mine in British Columbia, Canada (Altun et al., 2013). Introduction of preconcentration at Castlemaine Goldfields in Australia showed that 48% of the feed was rejected as gangue before processing the ore in the mill. This resulted in an increase of 3.8% in recovery and 30% reduction in energy consumption (Ballantyne et al., 2012, p.2). However, preconcentration of ore has not been prevalent in the mining industry due to the misconceptions regarding the principles and benefits of the technology (Bamber et al., 2004). Studies have shown that the demand and production of metal has been on a constant rise. According to The U.S. Geological Survey (2014), the world production of Copper was 9.2 Mt in the year 1990, 13.2 Mt in the year 2000 which has increased to 16.9 Mt in the year 2012 (Kelly & Matos, 2014, p.3). Moreover metal prices have been fluctuating, and the operating costs have increased over time (Schodde, 2010). To meet this demand in response, many mining companies 2  have increased the scale of production to extract more minerals at lower unit operating cost (Bozorgebrahimi et al., 2005). Furthermore, public awareness about the mining industry has improved drastically due to which environmental regulations have been made stringent to reduce damage to the environment due to mining activities (Jenkins & Yahovleva, 2006). The above problems are addressed in this research with the development of a preconcentration technology that can minimize the  processing costs by processing lesser material in the mill,  improve the metallurgical recovery by providing a higher grade feed to the mill and reduce the  damage to the environment by generating lesser amount of tailings (Vatcha et al., 2000).   1.1 Statement of Problem The value of an open-pit mine is commonly determined in parts by dividing the mine into virtual blocks, and then, depending on the grade of each block, a value is assigned to the block (Asano & Dessureault, 2009).  The grade and value of each block is further defined by compositing and geostatistical analysis (Hekmat et al., 2013). This process has a tendency to mask the grade variability in the orebody and assigns a homogenized grade value to the block. The major problem being addressed in this research is highlighting the true inherent heterogeneity of an orebody and possibly increase its value. A mineral deposit with variation in distribution of mineralogy or variation in mineral concentration can be termed as a heterogeneous mineral deposit (Groves et al., 2005). In the current study, the use of the term ‘heterogeneous ore body’ refers to an orebody which contains a mixture of both ore and waste (Lyman, 2011). When a block of material is termed as ‘ore block’, the entire volume of the block is considered to constitute material above cut-off grade and the distribution of minerals is uniform in the entire volume of block. Similarly, a block of material termed as ‘waste block’ is considered to constitute uniformly distributed material which 3  is below the cut-off grade (Lyman, 2011; Groves et al., 2005). Whereas in reality the distribution of grade may not be always uniform and every block of material mined might be a mixture of ore and waste.  Increasing the resolution of handling an orebody may reduce internal dilution by eliminating the waste if selectively mined. Inaccuracy in identification of heterogeneity and improper classification of ore blocks and waste blocks, may cause dilution or wastage of ore.  For example, blocks classified above the cut-off grade called “ore blocks”, may include gangue or material below cut-off grade, reducing the grade and value of the block. Similarly, “waste blocks” i.e. blocks classified below the cut-off grade may include valuable ore which will be discarded and wasted. In order to appropriately classify the ore and waste during mining, the selectivity has to be improved at the digging stage. At practical levels it is challenging for a large digging equipment to precisely separate the ore block from the midst of waste and waste block from the midst of ore. It is important to note that the losses incurred are higher when dealing with marginal grade blocks (Kumral, 2012) i.e. Presence of few ore blocks marginally above the cutoff grade along with larger number of waste blocks during compositing may result in the entire mass being categorized as waste and discarded. Estimation of resources is done based on interpolation of available drillhole data. Estimation of a heterogeneous orebody can lead to underestimation or overestimation of resources as the variability in the grade may not be captured by the information generated from the drillholes (Dominy et al., 2004). Figure 1 is a hypothetical example explaining the situations leading to underestimation or overestimation of resources. The dashed lines in the figure represents the ore-waste contact zone. It can be seen that, dilution is caused due to the intrusion of waste in the 4  interpreted ore zone. If the intrusion is large, the estimated resource will be less than the actual quantity and this results in overestimation of resources. There are some zones in the figure which is ore and not included in the estimation resulting in loss of ore. This causes loss of value as it is treated as waste and this results in underestimation of resources. The error in estimation is due to the lack of appropriate tools to identify the heterogeneity in the orebody which leads to miscalculation during resource estimation resulting in underestimation or overestimation of resources.    Source: Sinclair & Blackwell, 2002, p. 17 Figure 1: Heterogeneity in orebody leading to underestimation and overestimation of resources 1.2 Objective of the Research    Estimation of resource becomes complex when the heterogeneity of the orebody increases. The existing estimation techniques masks the in-situ grade variability resulting in ore loss and dilution. 5  (Lyman, 2011). The primary objective of this quantitative experimental research study is to highlight the true inherent heterogeneity of an orebody by producing finer granular information in order to achieve higher selectivity during mining. This assists in better classification of ore and waste and may reduce dilution and wastage of ore.    The other objective of the study is, upon identifying the true heterogeneity within an orebody, introduce a selective partitioning stage to segregate the ore and waste before hauling the material to the mill and eliminate the identified waste from the mining cycle as early as possible. Elimination of waste before hauling the material to the mill reduces the quantity of the material to be processed in the mill, improves metallurgical performance of the mill and quality of the product (Salter & Wyatt, 1991). Figure 2 shows a schematic diagram of the two mining scenarios being considered in this research study. Figure 2a shows the flow sheet of mining activities currently being practiced in the mines where the excavated ore is hauled from the mine to the mill and the waste is hauled to the waste dump. All the mined ore goes to the mill and further processing happens in the mill .Figure 2b shows the flow sheet of mining activities as proposed in this research study where a classification stage is introduced immediately after excavation stage for further processing of ore before hauling it to the mill in order to discard the waste without further handling.       6   Figure 2: Schematic diagram showing the flow sheet of mining activities considering two mining scenarios 7  1.3 Significance of the Research Study The significance of highlighting the true inherent heterogeneity in an orebody is to increase the resolution of viewing an orebody in order to get a clear understanding of the grade distribution and mineralogy of the deposit. The increased resolution may assist in better classification of the ore and waste and help in better utilization of resources. It is also expected to reduce dilution and wastage of ore. Introduction of a selective partitioning stage using sensor based preconcentration is anticipated to make the process of grade determination easier and faster (Wotruba & Harbeck, 2010). The implementation of the preconcentration step immediately after excavating the mined material assists in onsite classification of ore and waste. The identified waste can be eliminated from the mining cycle and only reduced amount of ore has to be hauled and processed further. This may result in the reduction of the size of downstream processes and cause significant economic benefits.  1.4 Site Details The study was carried out on BHP Billiton’s Spence and Escondida copper mine samples in Chile. The Spence copper deposit was discovered in 1996 and is currently operated by BHP Billiton plc. It is located in the arid Atacama Desert region of northern Chile at an elevation of 1,700 m above sea level, 150 km north east of Antofagasta and 50 km south west of Calama near the small community of Sierra Gorda (Cameron et al., 2005; BHP Billiton, 2007).   The Escondida mine is also located in the Atacama Desert of Chile at an elevation of 3050m above the sea level, 160 km south east of Antofagasta (Van Eldert, 2011). The Escondida mine is the world’s lowest cost and largest producer of copper and is managed by BHP Billiton along with 8  Rio Tinto, Jeco and International Finance Corporation (BHP Billiton, 2011). The geographical location of both Spence and Escondida mines are shown in Figure 3. Spence and Escondida mines were chosen for the study as this research focuses on improving the recovery of low grade base metal mineralization and also because the sample material was readily available for the study.                                      Source: Romero et al., 2011, p.92 Figure 3: Spence and Escondida mine location 1.4.1 Spence Mine The Spence deposit is completely covered by gravels and alluvium and the mineralization is 80-100 m below the surface (BHP Billiton, 2007). It is a partially oxidized porphyry copper deposit 9  of Upper Paleocene age. The deposit is made up of both copper oxide and copper sulphide mineralization which comprises of atacamite and chalcocite minerals (BHP Billiton, 2007). The deposit has a reserve life of 10 years with 3:1 stripping ratio (BHP Billiton, 2014, p.79; BHP Billiton, 2007, p.5). With average cut-off grade of 0.3%, the proven and probable reserves are estimated to be 283 million tonnes of copper at 0.76% average grade (BHP Billiton, 2014, p. 112-114). Tables 1 and 2 shows the mineral resource and mineral reserves respectively of Spence deposit. Table 2 show the presence of four kinds of ore types. The aim of this research is to increase selectivity in mining and the presence of different ore types provides an opportunity to introduce selective mining of different ore types and treat them separately.    Table 1: Spence Mineral Resources Source: BHP Billiton Annual Report 2014, p.112 Spence Mineral Resources, as of June 2014 Ore type Resources category Copper Grade Million tons Percent total copper Percent soluble copper Oxide  Measured 49 0.85 0.53 Indicated 6.7 0.73 0.51 Low grade oxide Measured 7 0.26 - Indicated 56 0.24 - Supergene sulphides Measured 145 0.92 - Indicated 50 0.59 - Transitional sulphides Measured 24 0.75 - Indicated 3.5 0.51 - Sulphide Measured 515 0.47 - Indicated 795 0.45 -     10  Table 2: Spence Mineral Reserves Source: BHP Billiton Annual Report 2014, p. 114 Spence Mineral Reserve, as of June 2014 Ore Type Reserve Category Copper Grade Million tons Percent total copper Percent soluble copper Oxide Ore Proven 34 0.76 0.53 Probable 2.8 0.77 0.63 Oxide Low Solubility Proven 21 0.96 0.44 Probable 12 0.57 0.22    Sulphide Proven 121 0.96 0.12 Probable 32 0.64 0.11 ROM Proven - - - Probable 61 0.39 0.09 Mining at Spence is accomplished using electric shovels, front end loaders and haul trucks. The Spence deposit is a low cost open cast copper mine employing solvent extraction and electrowinning (SX-EW) extraction technology. The processing facility is comprised of crushing circuit, agglomeration circuit, leaching circuits, solvent extraction facilities and electrowinning facilities and has a capacity of 50,000 tpd (BHP Billiton, 2007, p.5). Ore from the pit is crushed, agglomerated and transferred to leach pads. The oxide and supergene sulphide mineralization are both amenable to heap leaching with the use of sulphuric acid (BHP Billiton, 2004). The ore is then processed to produce copper cathodes by SX-EW technology.  11   Source: Arriagada, 2012; p.13 Figure 4: Spence mine flow sheet  The process flow sheet of Spence mine is shown above in Figure 4. Two types of ore, sulphide and oxide are mined separately and are treated by different leaching methods to achieve higher recovery rates because of different reaction kinetics (Domic, 2007). The oxide ore is chemically leached and the sulphide ore is leached using bacteria. The copper oxide ore achieves a recovery of 82.4% and the sulphide ore achieves 80.8% (Domic, 2007, p.92). The ore then goes to the SX-EW to produce copper cathode from pregnant solution which has a capacity to produce 170 ktpa of copper cathode (BHP Billiton, 2014; p.80).  12  1.4.2 Escondida Mine Escondida deposit mineralization consists of hydrothermal sulphide ore made of brochantite and antlerite minerals, typically grading from 0.2% to 1% Cu (Padilla et al., 2001; Van Eldert, 2011, p. 25). Primary sulphide is mineralized by pyrite, chalcopyrite (50%) and bornite with covellite (10%) and chalcocite (40%) in the enriched zone. Sulphide ore forms 77% of the mineral reserves, oxide ore forms 4% and the rest 19% is low grade sulphide ore (Basto, 2012, p. 13). As of June 2014, Escandida has reserves of 7,555 million tonnes of ore at 0.63% Cu and mineral resources amounting to 8040 million tonnes (BHP Billiton, 2014; p.112). Table 3 and Table 4 shows the Mineral Resources and Mineral Reserves respectively of Escondida mine. Table 4 shows the presence of three kinds of ore types. These material have different treatment processes and different recoveries (Van Eldert, 2011). Availability of information of the ore type during mining can assist in selectively mining the material according to the ore types and then treat them separately.   Table 3: Escondida Mineral Resources Source: BHP Billiton Annual Report 2014, p.112 Escondida Mineral Resources, as of June 2014 Ore type Resources category Copper Grade Million tons Percent total copper Percent soluble copper Oxide  Measured 117 0.8 - Indicated 62 0.65 - Mixed Measured 84 0.74 - Indicated 47 0.5 - Sulphide Measured 5150 0.65 - Indicated 2580 0.52 -   13  Table 4: Escondida Mineral Reserves Source: BHP Billiton Annual Report 2014, p.114 Escondida Mineral Reserve, as of June 2014 Ore Type Reserve Category Copper Grade Million tons Percent total copper Percent soluble copper Oxide Ore Proven 92 0.88 - Probable 53 0.67 -    Sulphide Proven 3540 0.75 - Probable 1610 0.59 - Sulphide Leach Proven 1650 0.46 - Probable 610 0.40 - The Escondida mine is a conventional open pit operation processing sulphide and oxide ores. Average stripping ratio of the mine is 1.7:1 and the nominal production capacity is 85.6 Mtpa of copper concentrate and 330 ktpa of copper cathode (BHP Billiton, 2014; p. 80). The mine has three different ore types which can be sent to the mill, the acid leaching or the bio-leaching pad and the waste is sent to the waste dump. The process flow sheet of Escondida mine is shown in Figure 5. Primary crushing of ore happens with in-pit crusher and then the material is conveyed to coarse ore stockpiles. Sulphide concentrator employs crushing, milling and flotation circuits. The low grade sulphide ore is crushed and dumped on large heaps where leaching occurs by oxidation induced by bacteria (Basto, 2012). Oxide ore is crushed, agglomerated and acid leached in large heaps and then solvent extraction and electrowinning (SX/EW) technology is used to produce copper cathode from the leached ore (Basto, 2012).  14   Source: Basto, 2012; p.13 Figure 5: Escondida mine flow sheet  1.5 Research Design A case study has been undertaken using samples from BHP Billiton’s Spence and Escondida mine sites to analyze the benefits of mining a heterogeneous orebody with increased selectivity. The increase in selectivity is achieved by increasing the resolution of handling the orebody which was done by reducing the size of the selective mining unit (SMU) to the size of an excavator bucket. Selective partitioning of ore is introduced at the excavation stage where the excavator bucket is used as a sorter. The excavator bucket acts as an effective preconcentration ‘tool’ and is equipped with sensors to determine the grade of the material in the bucket.  The samples from Spence and Escondida mines were tested using HFEMS and XRF devices at MineSense Technologies laboratory in North Vancouver, Canada to determine the Cu content. The 15  samples were then sent for assaying to determine the Cu concentration by chemical analysis. Accuracy of the results obtained from XRF tests were checked by carrying out a correlation analysis between XRF and assay results. Upon achieving good correlation, the XRF results were used to develop two models, one termed as Numerical mine model and the other as Vulcan model. The Numerical mine model was a preliminary model developed to understand the scope of application of the preconcentration tool to increase selectivity in mining. Results of the Numerical mine model showed a positive impact and hence the Vulcan model was developed. The Vulcan model is an advanced model developed to represent a hypothetical mining scenario. The model evaluates the amount of mineral that could be theoretically recovered with two different SMU sizes and therefore estimate the difference in savings.   1.6 Summary Preconcentration of ore has proved to be beneficial to the mining industry, but its application in the mines is not widespread (Murphy et al., 2012). The negative perceptions regarding the application of sorting in mining such as poor reliability, high capital costs, high operating costs, need for feed preparation and loss of ore mineralization has confined the application of preconcentration in mines (Bamber, 2008). This research study focuses on introduction of a simple low cost method of preconcentration by using sensors in excavator buckets. These sensors determine the grade of the material in the bucket and hence the value of the material is known. An attempt is made to increase selectivity by producing fine granular information to an extent equal to the size of the excavator bucket. The proposed technology thus increases the resolution of viewing an orebody and identifying the inherent heterogeneity. Currently there is no literature available explaining the use of an excavator bucket as a sorter and hence this study provides a new 16  insight to increase selectivity in mining with the use of sensor equipped shovel bucket and efficiently mine a heterogeneous ore body.  The following chapters of the thesis are organized as follows: Chapter Two evaluates the previous work done in the field of preconcentration, automatic sorting and orebody estimation. Chapter Two critically evaluates the previous research carried in the field of interest and highlights its strengths and weaknesses. It provides a better understanding of the existing technology and the need for development of the proposed new technology. Chapter Three describes the research design explaining the methods adopted in testing the samples, various equipment used for the study and the procedure used in the development of models. Chapter Four describes the results of the grade determination tests carried and the outcome of the models developed. Chapter Five summarizes the results and discusses its application in the industry. Chapter Six concludes the research work carried out and highlights the theoretical and applied contribution to the mining industry and also provides recommendations for future work.           17  Chapter 2: Sorting and Preconcentration: A Literature Review Sorting of ore is a method of mineral beneficiation and is a crucial process conducted after the excavation of rock (Manouchehri, 2003). The key defining element of this process is the classification of valuable material from the mined rock, discarding the waste material (Singh & Rao, 2005). Introducing sorting of ore as early as possible in the mining cycle eliminates waste without further handling. This reduces the amount of material to be processed in the mill and thus reduces the size of downstream processes leading to a significant impact on the economic success of the mine (Murphy et al., 2012). Early introduction of sorting in the mining cycle wherein, the ore is sorted from the waste at coarse particle sizes before hauling it to the mill for further concentration is known as preconcentration (Weatherwax, 2007). There are various methods of sorting such as hand sorting, gravity concentration, dense media separation, coarse flotation or sensor based sorting (Bamber et al., 2004, p.5). Sorting has evolved over the time from simple hand sorting methods to advanced automatic sensor based sorting (Weatherwax, 2007). Automated sensor based sorting is building a reputation in different kinds of industries, but in mining it is still an emerging technology (Dalm et al., 2014). Using this technology during the excavation stage poses as a challenge to the mining industry due to the hostile mining conditions created by extremely dusty and extreme weather conditions and therefore the development of this sensor dependent application is highly limited (Brooker et al., 2007). Despite these limiting conditions, preconcentration using sensors at the excavation stage proves to be extremely worthwhile, due to the large benefits revolving around saving energy, reduction in the amount of waste generated, reduction in the amount of tailings generated, minimizing surface footprint as well as reduction of waste hauling to the mill, and finally lesser amount of material being processed in the mill (Tollinsky, 2011). This chapter evaluates the current state of the technology using the literature on 18  various available sorting methods and also focuses on the assessment and contextualization of the research topic-implementing sensor based preconcentration at the excavation stage. 2.1 Why Preconcentration at the Excavation Stage? – A Closer Look at the Importance of this Application.  As discussed earlier, preconcentration during excavation stage is designed to introduce sorting of the ore much earlier in the mining cycle as opposed to the customary methods of waste elimination. Traditionally, preconcentration of ore takes place at a location distant from the initial mining site incurring additional haulage costs due to the transportation of both ore and associated waste (Bamber et al., 2004). If liberation of the ore from the waste takes place very close to the excavation point, it is expected to be beneficial since the waste is separated from the ore immediately after excavation and discarded before the mined material undergoes further processing.  Liberation of the ore from waste when done at early stages of the mining cycle at coarse particle size improves the metallurgical performance of the mill (Bowman & Bearman, 2014). This liberation increases the quantity of the valuable components in the feed, thus increasing the quality of the concentrate (Bowman & Bearman, 2014). There are various other compelling reasons which are discussed in this chapter as to why the application of preconcentration is an advantage to the mining industry such as minimizing environmental impact, reduction in cut-off grade of the mine, decrease in capital costs and operating costs and increase in mine life (Murphy et al., 2012).  Despite strict regulations to protect the environment and constant efforts towards material recycling, the demand for metals has been on a rise to provide for the growing needs of today’s technology oriented world (Thakurdin, 2010; Kelly & Matos, 2014). This demand calls for a higher production of metal at the existing mines and an increase in production usually has a directly 19  proportional relationship with the production of waste (Thakurdin, 2010). Introduction of preconcentration at the early stages of mining may decrease the amount of tailings and other process wastes generated. Due to the reduction in the amount of waste being discharged into the environment, the size of tailing pond required reduces causing a smaller footprint. It also minimizes the risk of harmful toxic substances getting into the environment and thus minimize the damage to the environment (Weatherwax, 2007).    To meet the increasing demand of metal, companies first exploit high grade resources (Norgate & Jahanshahi, 2010). This has resulted in the depletion of high grade resources and the world currently is left with a majority of lower grade resources (Mudd, 2007). By preconcentrating the ore, waste material may be largely eliminated before processing in the mill, contributing to economic savings and thus making the low grade resources economically profitable. This can contribute in reducing the cut-off grade of the mine. Since the cut-off grade of the mine decreases, more resource can be mined profitably thus converting the previously categorized resource into reserve (Bowman & Bearman, 2014). This enhances the life of the mine (Salter & Wyatt, 1991). As discussed, production cost of metal is increasing, availability of high grade resources is decreasing while the demand for the metal is constantly on a rise (Thakurdin, 2010). To meet the demands, the mining companies have to constantly produce more ore but generating profits due to these limitations has been a challenge to the mining companies. The method easily adopted by most companies to tackle this situation is by the application of the concept of economy of scale, where the size or capacity of production is increased with the sole aim of decreasing the unit cost of production (Bozorgebrahimi et al., 2005). Instead of developing a technology to reduce production costs and increase profits, companies tend to increase the scale of production and lower 20  the unit cost of production since the fixed costs are constant irrespective of the scale of production. This results in generation of higher volumes of tailings resulting in further damage to the environment and creating larger surface footprint (Bozorgebrahimi et al., 2005). However, in order to meet production demands and targets, application of presorting technology of ore may increase the ore production without an increase in the size of the production or the mine. Advanced presorting techniques may facilitate in recovering the mineral from the waste and contribute in higher metal production (Weatherwax, 2007). Since presorting of ore discards the waste and lesser material goes into the mill, there is scope for accepting more material in the mill and thus increase production without increasing the size of the mill (Weatherwax, 2007).  Preconcentration of ROM material based on online ore composition estimation methods involves determination of the rock characteristics (Perez et al., 2011). These methods provide the mineralogical composition of the material at a very early stage in the mining cycle before the material enters the processing mill (Perez et al., 2011). Early knowledge of the feed material is always beneficial to the mill as it gives additional time and information on material characteristics and helps with the preparation at the receiving end (Caulfield, 2011). Accordingly, the process parameters of downstream processes can be changed. Caulfield states that this early knowledge increases recovery of metal by 1-2% (Caulfield, 2011, p.26). Furthermore, upon determination and classification of the ROM as waste, the waste material can be diverted immediately to the waste dump which otherwise would have been treated in the mill, wasting energy and not recovering any valuable metal. Additionally, this will reduce haulage and processing costs (Manouchehri, 2003). Preconcentration of ore assists in providing a uniform quality feed to the mill due to the elimination of unwanted materials from the ROM (Salter & Wyatt, 1991). This is crucial because the 21  uniformity of the feed contributes to higher grade of the concentrate, since the plant takes time to adjust to the characteristics of the feed material when the feed is non-uniform (Salter & Wyatt, 1991; Walker 2014). Apart from the improvement in ore quality, the quantity of the feed is significantly reduced contributing heavily to a smaller downstream process. For example, the size of the mill to recover the same amount of metal is comparatively smaller (Bamber, 2008). Also, the fleet of machinery may decrease as lower material volumes are handled in the downstream process saving capital cost, and the significant decrease in the size of tailing dam takes a much lesser toll on the environment (Bamber et al., 2004). The benefits of preconcentration as highlighted by Bamber et al. (2004), Murphy et al. (2012), McCullough et al. (1999), Salter & Wyatt (1991) and Weatherwax (2007) are well known in the mining industry but its application at practical levels is challenging. Some of the reasons contributing to the lack of applicability of the concept are due to the perception of high capital costs, high operating costs and high maintenance costs associated with preconcentration technology, requirement of feed preparation, low efficiency of the existing sorting techniques, less effective for variable ore mineralogy, low throughputs and perception of availability of cheaper sorting options (Bamber, 2008).  The previous discussion on the benefits of sorting and preconcentration suggests the necessity of introducing preconcentration as early as possible so that the capital expenditure and energy consumption is minimal. . The current research work aims to overcome the existing barriers to the application of this concept. It is believed that there is a great scope for the application of preconcentration in mining especially when the ore body is heterogeneous (Chatterjee et al., 2010). It can be said that there is scope for maximizing utilization of resources and improve the economic 22  benefits in mining by preconcentrating the ore. To deal with the problem of heterogeneity and for efficient utilization of resources, the preconcentration method has to be modified in a way such that the selectivity increases during mining, which is discussed later in this chapter.  2.2 Sensor Based Sorting  Beneficiation of ore by using sensor based sorting technology is gaining popularity in the mining industry but its application is limited to pilot plants and laboratory testing (Manouchehri, 2003; Bergmann, 2009; Wotruba & Harbeck, 2010). Modern sensor based techniques with increased throughput, lower capital investment and higher sorting efficiencies have overcome some of the limitations of preconcentration  and have found to be advantageous to be adopted in the sorting of minerals (Wotruba & Harbeck, 2010). One of the most important reasons for implementing sensor based techniques is to reduce the complexity of the grade determination process by using online tools and mineral sorting process by using automatic sorting techniques. These techniques have the ability to determine the mineralogy and rock composition without destruction of the rock sample, thus showing tremendous potential for the application in mines (Celis, 1996; Ghosh et al., 2014). Most of the sensor based techniques undergo the following five steps for sorting the ore (Manouchehri, 2003): 1. Conditioning of the material by screening to achieve appropriate size, or washing before presenting to the sensors 2. Present the material to the sensor on a particle-by-particle basis  3. Measure the properties using sensors 4. Analyze the measured property to decide if the particle has to be accepted/rejected  5. Activate a mechanical device to accept/ reject the particle based on the property  23  The working principle and application of some of the most commonly used sensors in mineral sorting have been discussed in this chapter. Some of the techniques that have been successful in achieving the goal of sorting have been listed below. Table 5 corresponds various sensor technologies to the minerals to which they are applied, on the basis of its material property.  Table 5: Sensors used in mineral processing industry and corresponding minerals Source: Harbeck & Kroog, 2008; p. 272  Wavelength [m] Sensor/ Technology Material Property Mineral Application Gamma radiation 10-12 Radiometric Natural Gamma radiation Uranium, Precious Metals 10-11 10-10 X-ray transmission Atomic Density Base/ Precious Metals, Coal, Diamonds X-ray 10-9 XRF Visible Fluorescence under X-rays Diamonds  10-8 Ultraviolet 10-7 CCD camera Reflection, Brightness, Transparency Base/ Precious Metals, Ind. Minerals, Diamonds Visible light 10-6  10-5 Photometric Monochromatic Reflection/ Absorption Ind. Minerals, Diamonds Near Infrared 10-4 Near Infrared Spectrometry Reflection/ Absorption Base metals, Industrial Minerals Infrared 10-3 Infrared cam Heat conductivity, heat dissipation Base metals, Industrial Minerals Microwaves 10-2  10-1 Microwaves Sulphides & Metals heat faster than other minerals Base/ Precious Metals Radio waves 101     102 Alternating current 103 Electro- Magnetic sensor Conductivity Base Metals Over the years, some sensor based techniques have proved to be efficient and have overcome some of the existing limitations. According to Tong (2012), the four most commonly adopted sorting 24  techniques are: Optical sorting, Microwave Infrared method, X-Ray based techniques namely X-Ray Transmission (XRT) and X-Ray Fluorescence (XRF) which are discussed in this chapter (Tong, 2012).   2.2.1 Optical Sorting Optical sorting is an ore beneficiation technique which is based on the reflectance, transparency or other visual surface characteristics of the sample. These properties are determined by photo detector or digital cameras (Tong, 2012). The detectors used for sorting under visible light are photodiodes, photomultipliers and charged coupled devices (CCD). Photodiodes and photomultipliers detect the intensity of light reflected from the material surface and CCD’s detect and analyze the colours (Tong, 2012).     Source: Grewal, 2013 Figure 6: Schematic representation of optical sorter 25  Figure 6 above explains the working of an optical sorter. The feed material falls off from the belt conveyor where the samples are exposed to a camera under visible light. Images from the camera are processed in the computer. The detected intensity or colour is compared by the data processing system with the pre-defined accept/ reject criteria and the destination of the scanned material is decided accordingly (Grewal, 2013). This method is applied on sorting of minerals such as magnesite, limestone, phosphate, gold and lignite (Gulcan & Harzanagh, 2013).  Since this method of sorting is based on surface properties of the rock, it is not able to provide information on the metal content in the sample and be utilized in the classification of ore based on cut-off grade (Gulcan & Harzanagh, 2013). This method of sorting is effective for separating various ores of different minerals as they exhibit different surface characteristics or categorizing the sample into low grade and high grade (Gulcan & Harzanagh, 2013). This method can be used as a preliminary step in preconcentration and not as a process that can replace assaying to determine the exact quantity of metal present in the ore. Another major drawback of this method is that, it takes time in the production and analysis of the image and hence its application is limited in the mining industry (Singh & Rao, 2005).   2.2.2 Microwave Infrared Sorting The next sorting technique is the microwave infrared method of ore sorting. Microwaves can be used to separate valuable material from low grade material or ores containing two different metals, based on difference in thermal absorptivity (Ghosh et al. 2014). Size reduction of ROM is done either by crushing or by heating the particles with microwaves. Microwave heating of particles generates stress within the rock particles, which causes breakage of the particles into smaller particles (Shaw & Lavin, 2012). Depending on the conducting properties of rock, the material gets 26  heated differentially. Higher metal bearing material is more susceptible to microwave energy and gets heated faster and rises to higher temperatures as opposed to low grade material (Ghosh et al., 2014). Based on the size, the material is then categorized into finer and coarser fractions and generally the finer section contains higher metal content compared to coarser fraction (Shaw & Lavin, 2012). The fines are further processed to produce concentrate by methods such as leaching, pressure oxidation or smelting depending on the kind of metal of interest, cost involved and other associated factors (Ghosh et al., 2014). The coarse fraction is subjected to thermal imaging analysis. The thermal images produced are analyzed based on the fact that particles with higher content of valuable minerals behave differently from barren material, thus helping in the classification of the material into low grade, medium grade and high grade (Ghosh et al., 2014). Hotter particles are separated from the colder particles and the hotter particles which contain higher metal content are sent to the concentrator to recover the metal. The colder particles are the waste or by-product (Shaw & Lavin, 2012). Figure 7 below shows the profile of iron ore samples heated using microwaves. The specimens with pink/ purple colours absorbed less heat since they are poor in iron content. The specimens with orange to bright yellow colour have absorbed more heat since they contain higher iron content (Ghosh et al., 2014).   Source: Ghosh et al., 2014, p.89 Figure 7: Thermal image and profile of microwave heated sample 27  However, this method of sorting does not provide information on the exact amount of metal present in the material. In order to gather information on the metal content, the samples after classification have to be sent for chemical analysis (Ghosh et al., 2014). This method can be successfully applied as a preliminary step in preconcentration but not as prominent ore sorting process. Another drawback of this method is that it requires some basic feed preparation such as cleaning of the surface since surface properties are analyzed during thermal imaging (Ghosh et al., 2014).  2.2.3 XRT and DEXRT  X-ray-based method of classification such as X-Ray transmission (XRT), Dual energy X-Ray transmission (DEXRT) and X-ray fluorescence (XRF) have proven to be effective methods in sorting of ore (Udoudo, 2010). XRT method of sorting is based on the amount of X-ray radiation absorbed and transmitted by the sample which is exposed to X-rays (Wotruba & Harbeck, 2010). When a material is exposed to X-rays, part of the energy is absorbed by the material and the rest is transmitted. The amount of absorption depends on the density of the material. Denser material absorb more energy compared to lighter material (Wotruba & Harbeck, 2010). When higher amount of energy is absorbed by the sample, the energy transmitted and received by the detector is smaller. Depending on the transmitted energy, an image is produced, based on which a transmission curve is generated which is characteristic to the specific material type (Tong, 2012). Based on these curves, rocks of different grades can be separated. A pre-defined criteria is used as the basis for sorting. High and low energy images of the material are plotted on the curve and the curve indicates the average density distribution of the material (Tong, 2012). The analysis of the transmission curves is done by software such as Mikrosort® simulation package and the material is sorted into different categories (Tong, 2012). But in case of metals with different thickness and 28  similar densities, X-rays transmitted will be of different intensities. Hence it is difficult to determine the metal content (Wotruba & Harbeck, 2010). To overcome this problem, Dual Energy XRT technique was experimented at the Delft University of Technology, Netherlands in 2010 (Mesina et al., 2007).  The working principle of DE-XRT is similar to XRT except that, instead of single energy X-rays, X-rays of two different energy levels are used to scan the ore material and the transmitted radiations are received by two line scan sensors/ detectors. This generates dual energy images of the scanned material and gives information about the average atomic number (Tong, 2012). DE-XRT has the ability to scan and sample the entire mass of the rock and produce images that are independent of the thickness of the rock sample (Lessard et al., 2014). This technique was first used in airports to scan luggage which was later extended in sorting coal (Von Ketelhodt & Bergmann, 2010). DE-XRT is now used in sorting of different kinds of base metal ores such as nickel, copper and precious metals such as gold (Allen & Gordon, 2009; Von Ketelhodt, 2009). Despite several advantages of using DE-XRT method of sorting, the throughput is very low. Even though the recovery is high, it is very important to achieve higher processing speed to meet the production demands. This method will be successfully applicable only if the economic benefits of producing lesser tonnage of high grade concentrate or finished product is higher than achieving higher quantity of output with slightly lower grade (Lessard et al., 2014).   2.2.4 XRF X-Ray fluorescence method is a non-destructive process of ore analysis and is one of the fastest and most economical online sorting technique (Celis, 1996). X-Rays when incident on a rock sample, excite the electrons causing the movement of electrons from inner shell to the outer shells 29  accompanied with the release of energy resulting in fluorescence (Fickling, 2011). Fluorescence is an exhibitive characteristic property of the metals present in the ore. Each element produces fluorescence of specific energy and wavelength. The emitted energy and wavelength helps in the prediction and estimation of mineralogy and concentration of the metal (Weltje & Tjallingii, 2008).The concentration of metal content is higher when more number of photons are emitted (Fickling, 2011).  Fickling (2011) carried out a study using Rados ore sorter to sort the valuable metal particles from the waste in the ROM. This is a simple method where ore is fed into the hopper. The particles fall from the hopper on to the belt at the rate of 8 particles per second where every particle is scanned by the XRF which determines the metal content in the sample. Based on the metal content, an electromechnical ejector is activated which decides if the particle has to be discarded or is a concentrate (Fickling, 2011). The study showed that, beneficiation of ore by XRF sorting was possible for ores with varying metal content. It was also proved that, this technique can be applied for sorting of different kinds of ore such as precious metal ores, base metal ores, ferrous metal ores, industrial minerals and rare earth minerals. XRF sorting proved to be a successful and efficient method of ore beneficiation but the technique was successful in recording only the surface readings and not the bulk grade of the sample (Fickling, 2011).    Celis (1996) carried out a study on gold ore using X-Ray fluorescence technique to analyze the gold content in the sample. When high energy X-rays were used, concentrations up to 0.5 ppm of gold were recorded. Use of high energy X-rays  along with Co- 57 radioactive isotope showed that the suggested method was economical, this method could be applied on larger samples as the absorption of radiation by the sample was low  and it achieved a sensitivity in the range of 1 ppm 30  (Celis, 1996). In order to understand the accuracy of the method with respect to particle size, Celis (1996) further carried out a correlation analysis between the XRF and assay results. The best correlation was observed for fine particle sizes and the least correlation for coarse particle sizes. This is in accordance with the results obtained by Fickling (2011) and it can be explained since X-Rays have low penetration capacity, they give better results for finer particles. Results are less accurate in case of coarse particle size since the mineral hidden inside cannot be captured (Celis, 1996). Sorting of ore by XRF technique has the potential for successful application in the mines but the limited penetration capacity of the X-rays have to be overcome.  Some of the companies producing sensor based sorting equipment’s include Tomra Sorting Solution and MineSense Technologies Limited. The equipment used by Tomra works on various sensor systems such as electromagnetic, colour, photometric, near-infrared, x-ray luminescence and x-ray transmission.  Wide range of minerals such as limestone, gold, diamond, coal etc. can be sorted using Tomra’s equipment. MineSense Technologies manufactures sorting equipment which work on X-ray fluorescence and High frequency electromagnetic spectroscopy techniques. Further discussion on the equipment developed by MineSense is carried in Section 3.2. All of the above discussed sorting methods have showed several advantages and a potential for the application in mines. Comparing all the four methods of mineral sorting, X-ray based methods seems to be the most suitable and advanced method for treating low- grade heterogeneous ore because analysis of X-ray based results helps in determining the approximate mineral content present in a rock sample unlike optical sorting techniques or microwave infrared methods of sorting. Comparing the two X-ray based methods, XRT and XRF, XRF appears to be advantageous as XRF method can process large volumes of material and achieve higher throughputs which is 31  vital for the growing production demand in the mining and mineral processing industry (De Jong & Dalmijn, 2007). 2.3 Estimation of Resources Estimation of mineral resources is calculation or interpolation of grade or value of a mineral deposit based on the available scarce information (Emery et al., 2006). Estimation of mineral resource is crucial for a mining company since most of its investment and operational decisions are dependent on the estimated resources. Based on the quantity (tonnage) and quality (grade) of the mineral resources, the development plans of the project are decided and the returns are predicted (Emery et al., 2006). Since the occurrence or development of a project depends on the estimation results, the accuracy and reliability of the results is ought to be extremely high. The accuracy of the information depends mainly on the method of estimation adopted and the drilling density (Boucher et al., 2005; Dominy et al., 2004). Generally, the estimation is better if the drilling density is higher as this provides more information about the mineral resource (Dominy et al., 2004). Based on the availability of information and confidence level, mineral resources are classified into inferred, indicated and measured resources and ore reserves are classified into probable and proven reserves (Snowden, 2001).  32     Source: JORC, 2012, p.9 Figure 8: Relationship between exploration results, mineral resources and ore reserves Figure 8 above explains that, confidence level in case of mineral resources increases from Inferred to Measured due to the increasing geological knowledge. The geological knowledge is increased mainly by increasing the number of exploratory boreholes drilled i.e. the density of drilling is least for inferred resources and maximum for measured resources (Snowden, 2001). This confirms that, with the increase in geological knowledge the error associated decreases. It is vital to increase the accuracy of reserve estimation and decrease the errors as this directly affects the profitability of the mine (Snowden, 2001). This shows that there is a necessity for increasing the geological knowledge by generating more information about the orebody. In case of ore reserves, the confidence level increases from probable to proven but even the proved ore reserves have an error associated with them and therefore the reliability is not 100% (Sinclair & Blackwell, 2002).  The error of estimation decreases with increase in information and the best possible way to increase the information is by increasing drilling density (Boucher et al., 2005).   Drilling is a very 33  expensive activity and hence the number of drillholes to be drilled has to be minimum and sufficient to enable an estimate of the tonnage and grade of the mineralization. The most important factor defining the drilling density is the financial investment involved (Boucher et al., 2005; Dominy et al., 2004). It is necessary to carry out a cost analysis between the cost savings or increase in profitability achieved due to reliable data with increased drilling and cost savings achieved with lesser drilling and less reliable data (Boucher et al., 2005). Therefore it may not be economically feasible to eliminate the estimation error by increasing drilling density. Choosing the appropriate estimation method plays a significant role in reserve estimation in order to generate the best estimate of the reserve in terms of grade and volume of the mineralization (Dominy et al., 2004). The selection of an appropriate estimation method depends on factors such as geometry of the deposit which determines the depth of details required and the kind of data, variability of grade determination which determines the amount of smoothing required and smoothing refers to generating a less variable representative value by considering and interpolating the actual variability in grade, ore boundary characteristics which is critical for the estimation of grade at boundaries where the grade zone changes, time and money which decides the extent of data generation and estimation required (Darling, 2011).  Mine design and scheduling plans during pre-production or development activities is based on modeling of the deposit which is done by estimation of resources (Luo et al., 2007). This is commonly done by methods  such as inverse distance weighted method (IDW), polygonal method or geostatistical methods such as ordinary kriging, indicator kriging, multiple-indicator kriging and disjunctive kriging (Sinclair & Blackwell, 2002; Darling, 2011). All of these methods generally have an error associated with them and the magnitude of estimation error depends on the 34  volume of material being estimated. Large blocks contain more number of known data points than smaller blocks and hence the error reduces (Hekmat et al., 2013). Higher variance in the estimated data affects the project valuation which results in poor financial gains than expected (Hekmat et al., 2013). Kriging, Polygonal and IDW methods are discussed in the following section. This helps in understanding the application of the methods, the advantages and disadvantages of every method which will ultimately help in choosing the appropriate method for application.   2.3.1 Kriging Kriging is a geostatistical method of reserve estimation used to estimate the value of a point or a block. Based on the information of known data points (scattered or organized), the value of an unknown data point is determined. It is a statistical approach of estimation and hence gives the measure of the variance of estimation (Bohling, 2005). The steps involved in kriging of drill data are: statistical analysis of data, variogram modeling and creating the kriged surface (ArcGIS resources, 2011).  Kriging has several advantages compared to other estimation techniques (Sinclair & Blackwell, 2002; Bohling, 2005; Darling, 2011).  Kriging quantifies the errors when two cases are being compared   This method calculates variance along with the estimated grade and gives estimate of the estimation error   The quantity of data points is considered while kriging   Weights are assigned based on the distance from the known data point and it decreases with increase in distance  35   Conditional bias implies that the estimated block value differs from the true block value generally due to underestimation of low grade points or overestimation of high grade points.  Conditional bias is not entirely corrected but minimized by kriging and this method is considered to be one of the least biased method of estimation.  This method compensates the effect of data clustering by assigning less weight to the data points within a cluster than compared to the individual points.  The disadvantages of kriging is that it smooths the original heterogeneous data and produces a smooth kriged surface, the variograms do not represent the mineralized zone accurately due to lack of data and every estimated block has an associated error with it (Bohling, 2005; Darling, 2011; Sinclair & Blackwell, 2002).  2.3.2 Polygonal Method Polygonal method is a simple method of resource estimation where, a polygon is generated by drawing a perpendicular bisector from the line joining two drill holes (Darling, 2011). Figure 9 shows an example of generating polygons around drillholes. The area of the polygon is the area of influence of the enclosed drillhole. The average grade is calculated by weighting the sample grade by the corresponding polygonal area. The tonnage of each polygon is then calculated based on the density and volume (Darling, 2011). The application of this method is simple and the construction of polygons is arbitrary and can be adapted to meet the geological requirements (Darling, 2011). Disadvantages of this method are, the area of the polygons generated do not represent the exact area of influence, the method is conditionally biased and this method tends to overestimate the resources (Darling, 2011). 36   Source: Wellmer et al., 2008, p.28 Figure 9: Construction of polygons for reserve estimation by Polygonal Method 2.3.3 Inverse Distance Weighting Method Inverse Distance Weighting method is an interpolation technique, applied to estimate the grade of a point or a block based on neighboring data points. The estimation depends on the distance between the known data points and the point or block whose value is to be determined. The value is weighted average of known points and the weights are inversely proportional to the distance between them (Sinclair & Blackwell, 2002; Lu & Wong, 2008). The weights are defined by the formula (Sinclair & Blackwell, 2002, p.19): 𝑤 =(1𝑑𝑖𝑥)∑(1𝑑𝑖𝑥) 37  Where, w is the weight, d is the distance between the two points, x is the arbitrary power generally 1, 2 or 3 and i is an integer. This method can be chosen since it is a simple method and easy to apply. Disadvantages of the method are the value of ‘x’ is not determined empirically but it is arbitrary, this method cannot estimate the variance of predicted values and this is a deterministic method, i.e. the distribution of data is not considered during estimation and hence clustering of data doesn’t have any effect on the estimated value (Lu & Wong, 2008). All of the above methods have both pros and cons and their accuracy depends on the geological conditions. Hence no method can be considered better than the other (Dominy et al., 2004). All of these methods have a limitation which is, the estimated value have an error associated with them and elimination of the error is almost impossible but can be reduced (Dominy et al., 2004). The associated error causes variance between the actual reserves and estimated reserves. The amount of variance can be reduced only by increasing the available information which means there is necessity of high resolution data for accurate estimation (Dominy et al., 2004).   Mine planning and scheduling is done considering the occurrence of estimation errors and hence an allowance is taken into account for mining loss and dilution (Snowden, 2001). Even though the allowance is considered the errors can sometime lead to significant variance in estimated and actual data.  Dominy et al., (2004) gives an example of an underground gold mine which on estimating the resources and reserves twice, showed variation as shown in Table 6. The resources and reserves were overestimated during first estimation.     38  Table 6: Difference between the initial estimate and the revised estimate Source: Dominy et al., 2004, p.84 Parameter Mineral Resource Ore Reserve Tonnes -40 % -74 % Grade -36 % -34 % Contained Gold -61 % - 83% The difference could be due to low density of the data, inappropriate estimation method used, or wrong perception of continuity which led to the over estimation of resources (Dominy et al., 2004).  Reserve estimation have shown errors in the magnitude of ± 50% to ± 80% which is risky for a mining project since it effects the cash flow and changes the entire project dynamics (Dominy et al., 2004). Apart from the errors, the above discussed methods of estimation do not provide selectivity in mining due to the limitation of availability of data. One of the possible solutions to increase selectivity and reduce the variance is by increasing the density of drilling data either by increased drilling or by new simpler and cheaper methods of grade determination. This can increase selectivity, efficiently mine a heterogeneous orebody and reduce dilution.     2.4 Impact of SMU Size on Mineral Resource Estimation An ore deposit is converted into ‘virtual’ mineable volumes usually by diving the orebody into blocks (Hekmat et al., 2013). This is essential to divide the deposit into mineable volumes. A grade and value is assigned to every block which is determined by grade estimation techniques and then block modeling is done (Emery et al., 2006; Hekmat et al., 2013). The smoothing of the data by interpolation techniques to estimate the mineral resources and ore reserves is known as block modeling (Hekmat et al., 2013). Some of the techniques used in the process of modeling are already discussed in Section 2.3. In this section, the focus is on increasing the accuracy and 39  reliability in the estimation of grade, value or reserves prior to mining by varying the SMU size (Hekmat et al., 2013). According to the study carried out by Jara et al., (2006) and Szmigiel (2005), reducing the size of SMU increases selectivity in mining. Since smaller blocks are mined, it is easy to segregate between the ore and unwanted material during mining. The study carried out by Szmigiel (2005) on a hard rock deposit showed that, reduction in the SMU size reduces dilution in mining and also improves the recovery of the orebody. Mining of an ore deposit by smaller SMU’s increases the resolution of viewing the orebody and due to this more detailed information of the orebody is gained which is very essential for planning and scheduling purposes (Boland et al., 2009). Also, since the selectivity while mining is higher, it is believed to reduce wastage of ore. If the size of the SMU is smaller, the resolution of viewing the orebody increases and hence the ore-waste contact zone can be easily distinguished and categorized accordingly (Darling, 2011). The resolution can be increased by increasing the drilling density and this reduces uncertainty in estimation (Hekmat et al., 2013). The small amount of ore shown as ‘ore lost’ in Figure 10, can be mined without the need of mining the associated waste with smaller SMU. If larger SMU is considered, the waste has to be mined along with the small quantity of ore resulting in dilution and this might make the entire block economically unworthy. The two cases are explained with the help of Figure 10. Thus an attempt is made in this research study to reduce the size of SMU’s in order to reduce dilution in mining.  40      Source: Darling, 2011, p.215 Figure 10: Dilution and ore loss due to inappropriate size of SMU Reduction in the SMU size also increases mineral recovery from the orebody. According to the study carried out by Jara et al., (2006) considering blocks of three different sizes 2.5m x 2.5m x 2m, 5m x 5m x 4m and 10m x10m x 8m, it showed that to extract the same quantity of metal, the amount of material to be mined is least for block of size 2.5m x2.5m x 2m and it increases for 5m x 5m x 4m with maximum tonnage for 10m x 10m x 8m. For example, as highlighted in Figure 11, in order to extract 1.2 million tonnes of metal, the amount of material to be mined is 78 million tonnes with larger SMU’s and 55mt with 2.5m x 2.5m x 2m SMU.  This favors the discussion of employing smaller SMU size for mining. The increased selectivity reduces dilution, ore loss and helps in extracting more metal with lesser material being mined (Jara et al., 2006).  41   Source: Jara et al., 2006, p.207 Figure 11: Comparison of grade and contained value between 2.5m, 5m and 10m SMU-based models 2.5 Conclusion The focus of this research study is to efficiently mine a heterogeneous orebody with increased profitability. A heterogeneous orebody is made of both ore and waste and it is very challenging to identify and segregate the ore and waste appropriately which can result in dilution or wastage of ore (Yamamoto, 1999). Increasing the resolution of viewing a mineral deposit may produce detailed information about the deposit which can assist in identifying and segregating the ore and waste. The resolution of defining an orebody improves when more number of selective mining blocks (SMU) are used and increase in the number of blocks results in smaller size. Reducing the size of the SMU blocks may assist in obtaining finer mineralogical information within the ore body and help in identifying the inherent heterogeneity (Boland et al., 2009). On identifying the heterogeneity, waste can be segregated from the ore, reducing the chances of dilution of ore 42  (Darling, 2011; Jara et al., 2006). Based on the literature and discussion carried out in this chapter, there is a need to introduce a preconcentration technique at the early mining stage which can segregate the ore from the waste with increased selectivity. In the next chapter a methodology is presented to employ a sensor equipped excavator bucket to act as a sorter at the excavation stage and preconcentrate the ore. The selective mining unit is the sensor equipped excavator bucket thus resulting in increased selectivity. Also, this results in generating very finer granular information which assists in high resolution deposit modeling. Successful application of this concept will result in maximum extraction of ore from the mineral resource and improve the economics of the mine.               43  Chapter 3: Methodology 3.1 Introduction This chapter focuses on the correlation of the research problem at hand and the associated research design. The proposed research design is structured to answer the problem statement with experimental data collected, along with the discussion of various methods utilized to approach this research problem. The benefits and limitations of all the methods chosen are also discussed. The research objective is twofold: identify and highlight the inherent heterogeneity in an orebody, and secondly the introduction of selective partitioning tool to segregate the ore and waste from the identified heterogeneous orebody.  Studies on the benefits of increasing selectivity and preconcentration have been carried out earlier. To further understand this concept and to meet the objectives of the research, a case study was carried out to investigate a real-life situation using samples from different zones of Spence and Escondida mines. To highlight the presence of heterogeneity in the orebody, the Cu content in the samples was determined and the distribution of grade in the orebody was presented graphically by plotting the Cu content against the sample.  The Cu content in every sample is known and this helps in understanding the variability of metal content in the orebody. In order to understand the effect of identifying heterogeneity on mine planning, two digital models were developed: Numerical mine model and Vulcan model. The numerical mine model was a preliminary model developed to understand the difference in savings achieved by mining a heterogeneous ore deposit with varying selectivity. The Vulcan model is an advanced model developed after the numerical mine model since the numerical mine model showed a potential for improving the savings by increasing selectivity. The Vulcan model involves high resolution deposit modeling using the 44  information generated for testing of increasing selectivity. This model demonstrates the difference in the amount of mineral that can be extracted from a heterogeneous orebody by varying selectivity.  The methods adopted in achieving the objectives of the research work are explained in detail in the following sections.   3.2 Framework The technology proposed in this research was developed by MineSense Technologies Limited based in Vancouver, British Columbia, Canada. MineSense is actively involved in the development and marketing of sensor based sorting equipment. The sensor technology used by MineSense includes XRF and HFEMS which have been used in the current study and their working principle is explained in Section 3.3. The company is also currently working on the development of LASER and hyperspectral techniques. The sorting equipment’s developed by the company includes BeltSense™, ShovelSense™ and SortOre™. The current research work deals with the application of ShovelSense™ in mines to increase selectivity and introduce selective partitioning of ore. The ShovelSense™ systems includes mounting of sensors in the bucket of excavator to determine the grade of the material in the excavator bucket. This helps in instantaneous grade determination and it also includes a decision support system which decides the destination of the material in the bucket (MineSense, 2015).  To achieve the objectives of the study, samples from BHP’s low grade copper (Cu) deposits of Spence and Escondida mines in Chile were considered. The samples from Spence and Escondida mines are best suited for the research study because of the abundance of low grade resources in the world along with increasing demand for base metals and also because the samples were readily available. Samples from seven different zones, two zones from Spence mine namely, SPPP07 and 45  SPPP11 and five zones from Escondida mine namely, Sulfuro, Oxido, Mixto, M3 and Lastre were considered. In order to estimate the amount of mineral content present in every zone an orebody was modeled with available information. Since no actual drillhole data was available, the drillholes were synthetically constructed using individual rock samples. Every sample was considered to represent one drillhole and every zone represents an orebody. A block of size 5m x 5m x 5m was developed around every drill hole and the grade of the sample represented the bulk grade of the block.  The block size was chosen which would closely represent the volume of an actual shovel bucket. In this study, the P&H 4100 XPC shovel is considered which has a bucket capacity of 120 tonnes. The block size closely matching this capacity is 4 m3 but for calculation purposes and easy understanding 5 m3 blocks were chosen. Assuming the specific gravity of the material to be 2.3 g/cm3, mass of one block is equal to 287.5 tonnes. Hence 500 blocks represent a mass of 143,750 tonnes. The specific gravity of the material at Escondida mine is between 2.42 g/cm3- 2.57 g/cm3 but since exact information was not available it is assumed to be 2.3 g/cm3 for both ore and waste (Alpers & Brimhall, 1989). The specific gravity for both ore and waste is considered same for easier understanding of the model when other parameters are varied. The Spence mine which is made of two zones, thus has a mass of 287,500 tonnes and Escondida mine which is made of four zones has a mass of 575,000 tonnes excluding the Lastre zone which is waste and is not considered in the development of models. This mass was sufficiently large to study the objectives of the research study.  The samples were sieved and classified into three categories, large samples which are greater than 3 inches/76.2 mm, medium sized samples whose sizes are between 1-3 inches/ 25.4 mm-76.2 mm and fines whose size is less than 1 inch/25.4 mm. In some instances imperial units are used for 46  easy representation during the study. The testing of samples was carried out at the MineSense laboratory and the grade determination techniques available are based on HFEMS and XRF technology. Each of the samples from all the seven zones were first scanned on HFEMS and then on XRF device to determine the Cu content in the samples. The working principle of HFEMS and XRF are explained in Section 3.3.2 and 3.3.3 respectively. After determining the grade by using sensors, the samples were crushed and then sent for assaying as per the procedure described in Section 3.5. Even though the grade of the sample is determined using sensors, assaying is done to re-check the results. To ensure and confirm the reliability of results generated from XRF and HFEMS tests, correlation study was carried out before proceeding with further experiments. First correlation analysis was done between assay results and XRF results and the second analysis was carried out between XRF results and HFEMS results. Zone wise results of the correlation analysis between XRF and assay results are presented in Table 11. XRF and assay results showed good correlation with correlation co-efficient varying between 0.548 and 0.827. Correlation co-efficient of 1 indicates highest level of correlation and 0 indicates that there is no correlation between the two sets of results obtained (Prion & Haerling, 2014). The results of the correlation analysis are discussed and interpreted in Section 4.2. However, no good correlation was observed between HFEMS and XRF results and hence the results from HFEMS tests were not considered for further tests in the study and only the results from XRF tests are considered. The reason for unreliable data received from HFEMS could be due to insensitivity of the device to the mineralogy of the sample under consideration. Furthermore, in terms of size, best correlation was observed for 25.4 mm-76.2 mm size range and hence this size was used in the development of models.  47  To analyze the outcome of introducing increased selectivity in mining and also the effect of reducing SMU size on the amount of mineral extracted and cost savings achieved, the results from XRF tests are used in the development of two digital models called Numerical mine model and Vulcan model. The development of models is discussed in Section 3.6.       3.3 Equipment and Procedure The following instruments were used in the current research work. The working principle of the instruments is explained in detail in the following section.  3.3.1 Jaw Crusher Jaw crusher is used in reducing the size of large rocks with the application of compressive force. In the current study it is used in the initial breaking of oversized rocks and it is also used for crushing the scanned ore samples before being sent for assaying. The samples which are to be assayed have to be crushed to finer sizes of less than 25.4mm. Even though in a real mining situation there is no uniformity of particle size, the crushing and grouping of samples in necessary to test the sensitivity of the grade determining devices or methods to the particle size. The jaw crusher used for the study is shown in Figure 12.  48   Figure 12: Jaw Crusher at MineSense Lab The jaw crusher has two jaws made of cast steel. One of the jaw reciprocates and one is stationary. When rocks are placed between the two jaws, due to the application of force the rock gets crushed (Georget & Lambrecht, 1982). The rock samples are fed from the top and the crushed sample is obtained at the bottom.  3.3.2 High Frequency Electromagnetic Spectroscopy (HFEMS) The High frequency electromagnetic spectroscopy sensor unit which is used in scanning of rock samples was developed by MineSense Technologies Ltd® at North Vancouver, British Columbia, Canada.  HFEMS is a ground penetrating technology which integrates with material handling equipment to measure the mineral content in the ore. It is used to identify the mineral content in 49  the rock sample by studying the conductivity and magnetic permeability of the rocks (Kamenetsky, 2009). Figure 13 is an image of the HFEMS setup at MineSense laboratory.  Figure 13: HFEMS mounted on the table The HFEMS setup consists of two blocks: reference block and sensor block. The reference block acts as a standard reference which is kept isolated from the sample, and the sensor block is the block on which the rock samples are scanned. Low power, high frequency alternating current is passed through both the blocks which generates an oscillating electromagnetic field over the coil. When the sample to be analyzed is brought near the sensor block, it alters the coil’s electrical and magnetic properties. The variation depends on the conductive material present in the sample. The difference in reading between the standard reference and the sensor is measured with an electrical bridge network. The readings are digitally analyzed and the response is correlated with a database of prerecorded values of metal content from samples already tested (Richards, 2011). Thus the 50  mineral content in the sample is determined. The sample is flipped and the procedure is repeated to get results from at least four different faces of the rock. The samples when scanned for six or eight faces showed the same result as when scanned for four faces and hence readings from only four faces were considered. The steps associated with one sample involves, scanning the rock sample, recording the readings and calibrating the device which approximately took two minutes during the experiment. This methods helps in instantaneous determination of the grade of the material with less capital investment. The drawback of this method is that it calculates the grade based on pre-recorded values of other samples and not independently using the sample under test (Caulfield, 2011).    3.3.3 X-Ray Fluorescence (XRF) High speed X-Ray fluorescence device was also developed by MineSense Technologies® to provide real-time grade of the mineral for sorting applications at MineSense laboratory, North Vancouver, British Columbia, Canada. It is widely used in determination of the composition of rocks (Weltje & Tjallingii, 2008). X-rays when incident on a rock sample excite the electrons and the electrons from inner shell occupy the outer shells with the release of energy. The emitted energy and wavelength helps in the prediction and estimation of mineralogy as they are the characteristic property of atoms of specific elements (Weltje & Tjallingii, 2008).  51   Figure 14: XRF setup for testing samples The device used in the experiment is shown above in Figure 14. The rock sample being scanned is placed in the grey coloured box which is an X-ray protected chamber. The computer attached to the unit records the reading generated by scanning every side of the rock sample and then produces an average outcome of all the readings. The approximate time involved to scan and record the result is 3-4 minutes. This test helps in predicting the mineralogy of the samples and calculate the Cu grade in the samples.  3.4 Procedure for Classification of XRF Scanned Samples for Assaying  The classification of XRF scanned samples is done in order to check the sensitivity of the grade measuring tools to be uniform for all the grade ranges. The samples were classified into three categories based on the grade and 10 samples from each category were sent for assaying. The procedure followed in selection of samples for assaying is explained in Section 3.5. All the samples 52  of size 25.4 mm-76.2 mm/ 1-3 inch were weighed on a weighing scale and then based on the XRF result, the samples were classified into three categories: Category 1 with samples less than 0.3% Cu (low grade), category 2 with samples in the range of 0.3-0.59% Cu (medium grade) and category 3 with more than or equal to 0.6% Cu (high grade). The distribution of samples in every zone was plotted on a pie chart. There are 500 samples in every zone and the total weight of the samples in every category was measured and then the weight percentage of every category was represented on a pie chart. The samples thus chosen for assaying are not biased and represent the entire grade range for the zone. Due to financial restrictions the numbers of samples sent for assaying was limited which otherwise would have given an opportunity for detailed analysis of XRF and HFEMS results.  3.5 Assaying During the study, the samples were chemically assayed at ALS Laboratory, North Vancouver, British Columbia, Canada. As explained in Section 3.4, the samples were divided into three categories: Cu content less that 0.3% (low grade), 0.3%-0.59% (medium grade) and greater than or equal to 0.6% (high grade). Only samples weighing more than 50 grams were considered as per the requirement of ALS Laboratories since the samples below this weight were very small and not sufficient for the chemical analysis. Ten samples each from all the three categories were randomly selected, but in some zones the number of samples chosen varied slightly. The variation was due to insufficient number of samples being available meeting all the requirements or the samples were fragmented unintentionally during the tests. The damage to the samples occurred during the movement of samples. The average grade of samples in Oxido, M3 and Sulfuro zone was higher and hence the grade range considered was less than 0.5%, 0.5%-1% and greater than or equal to 53  1%. The total number of samples chosen for assaying from every zone were 30 except for Lastre where only 15 samples were available above 50 grams. All the chosen samples were then crushed in the jaw crusher and then sent for assaying. Samples from different grade range were chosen in order to analyze and ensure the sensitivity of the XRF device to varying mineral grade in the sample. The purpose of assaying is to measure the mineral content in the sample by an established reliable method and to recheck the results generated by XRF device. To ensure the reliability of XRF generated results, a correlation analysis is carried out between XRF generated results and assay results which is discussed in Section 4.2.1. On establishing a good correlation i.e. correlation co-efficient of greater than 0.35, the XRF results are used in the development of the Numerical mine model and Vulcan model to study the effect of introducing increased selectivity in mining (Prion & Haerling, 2014).  3.6 Development of Models To analyze the outcome of reducing the SMU size and introducing increased selectivity in mining, the results obtained from XRF tests were used in the development of two digital models. One model is called the Numerical mine model and the other is called as the Vulcan model. The numerical mine model is a preliminary mathematical model developed to analyze the savings that can be achieved by mining a heterogeneous orebody with increased selectivity. The development of the model is explained in detail in Section 3.6.1.  The Vulcan model differs from the numerical mine model and is developed to analyze the difference in the amount of Cu that can be recovered from the concentrate by increasing selectivity while mining a heterogeneous orebody. The increase in selectivity is achieved by reducing the size of SMU’s. The development of Vulcan model is 54  explained in Section 3.6.2. The models are a flow sheet of mining activities which includes excavation of ore, sorting the ore, handling of the ore and then processing it in the mill.  A cost factor is associated with every stage and hence an economic model is built for the Numerical mine model based on the results obtained from the XRF scanning of the samples. The cost factors are included in this study only for highlighting the savings that can be achieved by increasing selectivity and easier understanding of the models. The assumptions considered in the development of the models are listed in Table 7. The assumptions included in Table 7 are not specific to any mine. The variable parameters included in the assumptions are listed along with their corresponding value.  The cut-off grade at Escondida mine for oxide ore is 0.2% and sulphide ore is 0.3%. Hence, in the current research study, the cut-off grade assumed is 0.25% Cu which is the average of the two values since ores of different minerals are considered in the study and the cut-off grade is kept constant (BHP Billiton, 2014). Table 7: Evaluation Model Assumptions Parameter Value Copper price $7000/tonne1 Cut-off grade 0.25%  Specific gravity of ore 2.3 g/cm3 Specific gravity of waste 2.3 g/cm3 Mining cost of ore $3/tonne Mining cost of waste $2/tonne Processing cost of ore $10/tonne Other costs $3/tonne Classifying cost $1/tonne Recovery 82% @ 0.6% Cu                                                            1 Average price of copper in August 2014 (London Metal Exchange, 2014) 55  Two cases are considered in each model which represents two different kinds of mining scenarios and hence the cost associated with it varies which changes the savings achieved in the two cases. The two cases considered under both the numerical mine model and Vulcan model are explained with the help of a schematic diagram shown in Figure 15. The amount of Cu shown in every block is hypothetical for easy understanding. The case 1 is depicted as composited model and case 2 as discrete model. The discrete model is made of smaller SMU’s of size 5m x 5m x 5m aiming to achieve higher selectivity and the composited model is made of SMU’s of size 25m x 5m x 5m. These SMU sizes were chosen to closely match the SMU volume with the volume of P&H 4100 XPC shovel bucket. The closest size would be 4m3 but for simpler representation of the models and easier understanding of the calculations involved, 5m3 is chosen in developing the two models. Five smaller SMU’s from discrete model are combined by averaging and a bigger SMU is developed in composited model. Hence, lesser selectivity is achieved in case of composited model. But to analyze the effect of varying the amount of selectivity on resource modeling and metal extraction, the two models are developed and discussed.    56   Figure 15: The two cases considered in the development of models                                            3.6.1 Numerical Mine Model The Numerical mine model (NMM) is a simple mathematical model developed using Microsoft Excel. The aim of this model is to highlight the effect of introducing increased selectivity in mining. The selectively mined blocks are then sent to the appropriate destination depending on the mineral content in the blocks. 500 samples from every zone are used in the development of the model. Every sample is converted into a block of size 5m x 5m x 5m as explained in Section 3.2 and the grade of the sample represents the bulk grade of the block. The two mining scenarios considered in the numerical mine model are the composited model and the discrete model. The SMU size under consideration in composited model is 25m x 5m x 5m which is developed by merging of five smaller SMU’s of size 5m x 5m x 5m. Hence there are 100 blocks in this model. The weight of one block is 1437.5 tonnes and the total amount of ore being mined is 143,750 57  tonnes. The composited model considers the application of grade estimation technique which in this case is an arithmetic averaging of grades. The grade of the SMU block in composited model is the average grade of five SMU’s of discrete model.  On the other hand, the discrete model is made of SMU’s of size 5m x 5m x 5m and hence there are 500 blocks in this model. No estimation of grade is done in this model and every block is treated separately in order to achieve higher amount of selectivity. The weight of one block in this case is 287.5 tonnes and the total amount of ore being mined in this case is also 143,750 tonnes.  Assuming a cut-off grade of 0.25% Cu, every block being mined is classified into ore or waste. Any block above cut-off grade is considered as ore block and is hauled to the mill for processing and any block below the cut-off grade is the waste block and is eliminated from the mining cycle without processing it in the mill. Every block being mined has an associated mining cost, processing cost and other miscellaneous costs such as administrative costs or maintenance costs and, this is termed as “other costs” in the models. An additional cost of $3/tonne is added under other costs to compensate for the missing cost factors. Any block below the cut-off grade is not processed and hence there is no processing cost incurred. The discrete model differs from the composited model because in the discrete model the mined material is sorted into ore and waste before processing it in the mill and, thus has an additional sorting cost associated with it. This costs an additional $1/tonne which is known as the classifying cost. The inclusion of cost factors is done in this model only to highlight the difference in savings that can be achieved by considering the two models and not for detailed economic analysis. The total cost is the sum of all the associated costs. The costs are calculated as shown: Total cost for Discrete model ($) = Mining cost + Sorting cost + Processing cost + Other costs 58  Total cost for Composited model ($) = Mining cost + Processing cost + Other costs Cu recovered from every block (Tonnes) = (Tonnage of the block x Grade x Recovery)/ 100 Cost value of every block ($) = Cu recovered x Market price of Cu Savings from every block ($) = Cost Value – Total costs It should be noted that every waste block being mined incurs cost and hence it is very important to identify it and eliminate it from the mining cycle as early as possible. 3.6.2 Vulcan Model The Vulcan model is developed to transform the raw mining data into 3D models. The model is developed using Maptek’s Vulcan mine planning tool. Estimation of mineral resources is done and the ore reserves are calculated using the Vulcan model. This model is an advanced model compared to the numerical mine model as estimation of resources is done by inverse distance weighting technique due to which the accuracy of the results is higher and also the reliability of the model. Similar to the numerical mine model, two cases are considered to study the impact of the size of block model on the amount of Cu that can be recovered from the concentrate. The two cases differed from each other in the size of SMU considered. In Case 1, the SMU size considered is 25m x 5m x 5m and in Case 2, the SMU size is 5m x 5m x 5m.   Using the available  information from Spence and Escondida mines, the models are developed by transforming the  500 samples into 500 data points as transformed during the development of Numerical mine model which is explained in Section 3.6.1. The transformation helps in easier representation which will help in better understanding of the blockmodel developed. Availability of more data points and information on their location helps in development of a mine plan and not just a bench. Since the information on location of the samples was not known, synthetic drillhole 59  pattern was developed and every drillhole represented one sample. Assuming the co-ordinates of the first hole of Sulfuro mine is at (5, 0, 15) the co-ordinates for the rest of the samples are defined and it extends from (5, 0, 15) to (500, 0, 15) along x-direction and (5, 0, 15) to (5, 20, 15) along y-direction. The spacing and burden is kept constant at 5m and hence the drill pattern looks like a bench as shown in Figures 16 and 17. The depth of the holes is same for all the holes and it is 15m which can be seen in Figure 17. The depth is kept uniform so that all blocks are of same volume and the grade of every sample represents a constant volume of material. The drillholes are assumed to be regular in pattern for simple representation which is shown in Figures 16 and 17. . Using the information from the drillhole, the raw mining data in converted into a 3D model. In this research study, since the drill holes are regularly patterned and the distribution of data is not dispersed, inverse distance weighting method of interpolation is chosen for resource and reserve estimation. This method is simple in application compared to other methods and also the accuracy of estimation is reliable when the drillholes are regularly spaced (Sinclair & Blackwell, 2002; Lu & Wong, 2008). This method is described in detail in Section 2.3.3. 60   Figure 16: Top view of M3 drillholes  Figure 17: Sulfuro drillholes 61  Triangulations are done and blockmodels are constructed around the drillholes with two different sizes of SMU’s. Every zone is designed to represent a rectangular bench. The drillholes from 5 zones of Escondida mine and 2 zones from Spence mine are patterned in such a way that they look like a giant mass of ore as shown in Figure 18 and Figure 19. The same procedure is followed in estimating and calculating the reserves for all the zones. Figures of only selected zones are shown as they all appear to be same.    Figure 18: Triangulations of the five zones at Escondida mine 62   Figure 19: Triangulation of the two zones at Spence mine On completion of the block model, reserves are calculated for the two models developed. A comparison is done between Cu extracted in case 1 and case 2 of Vulcan model in order to understand the effect of SMU size on recovery of Cu.   3.6.3 Trade off Study between the Two Cases in Numerical Mine Model and Vulcan Model A trade-off study is carried out when one situation leads to multiple solutions and the best solution is chosen which depends not only on the financial gain but also on other factors. Apart from financial savings, it must have minimum impact on environment, maximum metal recovery with minimal expenditure of energy, high amount of safety to the manpower and cause minimum damage to the society.  63  The study is carried out considering the cost savings achieved by increasing selectivity in numerical mine model and the difference in amount of Cu extracted with smaller SMU’s in Vulcan model. A trade off study is carried out between the composited model (case 1) and sorted model (case 2) in numerical mine model. The comparision is between the savings achieved by adopting higher selectivity during mining as per the sorted model against lesser cost involved due to no sorting in the composited model. In the sorted model, higher selectivity is achieved due to the additional sorting process taking place which has an associated cost factor. In the composited model, no sorting takes place and the sorting cost is eliminated but the disadvantage is that lesser selectivity is achieved. The best case scenario will be specific to the mine and it may not be possible to arrive at a general conclusion. The two cases in Vulcan model draws a comparison with the amount of Cu extracted due to varying the SMU size. In this model the emphasis is on the amount of mineral extracted and savings achieved.    Results generated from the models is presented in the next chapter. Interpretation of the results is carried out and the probable outcomes are discussed considering the Numerical mine model and Vulcan model and also the impact of reducing the size of the blockmodel on metal recovery is presented.         64  Chapter 4: Results and Discussions  4.1 Determination of Cu Content in the Samples The determination of Cu content in the samples is the primary step in achieving the objective of the research study. In order to identify the inherent heterogeneity in the orebody, it is necessary to understand the distribution of grades in every zone. This is accomplished by first scanning the samples on HFEMS and XRF devices which helps to determine the Cu content electronically using sensors. XRF and HFEMS scanning of samples were done at MineSense Technologies laboratory. The samples were then chemically assayed at the ALS laboratory. The results of the HFEMS tests, XRF tests and assaying are presented in the following sections.  4.1.1 HFEMS Results The samples from the Spence and Escondida mines were scanned on HFEMS device as per the procedure explained in Section 3.3.2. Results were generated at every 100 kHz interval for 14 values of magnitude and 14 values of phase. When two samples of same Cu percentage from the same zone were scanned the HFMES results showed no correlation. The same observation was noticed for many samples. The reason for this could be that the HFEMS reading was getting influenced by the presence of other minerals in the sample. Since the metal of interest in this case was Cu, the HFEMS tests didn’t generate results which helped in estimating the mineral content in the samples. Moreover, the correlation of HFEMS results with XRF results also did not help in arriving at any conclusion which is discussed in Section 4.2.2. It was not possible to confirm the reliability of the instrument used based on the results and hence the HFEMS results were not used in the development of the models. 65  4.1.2 XRF Results As per the procedure explained in Section 3.3.3, 1000 samples from the Spence mine and 2215 samples from Escondida mine are scanned on XRF device to determine the Cu content. The results are graphically represented which shows the approximate Cu percentage in every sample. The number of samples used for the test from every zone is presented in Table 8.  Table 8: Number of samples for testing in every zone Mine Zone Number of samples  Spence SPPP07 500 SPPP11 500   Escondida Oxido 500 Mixto 500 Sulfuro 500 M3 500 Lastre 215 Figures 20 to 26 shows the results of XRF tests for all the zones of Spence and Escondida mine. The Cu content in every sample is shown in the figures and the variation in the grade can be seen for every zone. Detailed results of XRF tests are presented in Appendix A.        66  1. SPPP07  Figure 20: Cu % in every sample of SPPP07 zone 2. SPPP11  Figure 21: Cu % in every sample of SPPP11 zone 00.250.50.7511.251.51.7521 26 51 76 101 126 151 176 201 226 251 276 301 326 351 376 401 426 451 476Cu (%)Sample NumberSPPP-0700.250.50.7511.251.51.7522.252.51 26 51 76 101 126 151 176 201 226 251 276 301 326 351 376 401 426 451 476Cu (%)Sample NumberSPPP-1167  3. Oxido  Figure 22: Cu % in every sample of Oxido zone 4. Mixto  Figure 23: Cu % in every sample of Mixto zone 00.250.50.7511.251.51.7522.252.52.7533.253.53.751 26 51 76 101 126 151 176 201 226 251 276 301 326 351 376 401 426 451 476Cu (%)Sample NumberOxido01234567891011121314151617181 26 51 76 101 126 151 176 201 226 251 276 301 326 351 376 401 426 451 476Cu (%)Sample NumberMixto68  5. Sulfuro  Figure 24: Cu % in every sample of Sulfuro zone 6. M3  Figure 25: Cu % in every sample of M3 zone 00.511.522.533.544.555.566.51 26 51 76 101 126 151 176 201 226 251 276 301 326 351 376 401 426 451 476Cu (%)Sample NumberSulfuro00.250.50.7511.251.51.7522.252.51 26 51 76 101 126 151 176 201 226 251 276 301 326 351 376 401 426 451 476Cu (%)Sample NumberM369  7. Lastre  Figure 26: Cu % in every sample of Lastre zone In the Figures 20-26 variation of Cu percentage in every zone is shown. This shows that the XRF device was sensitive to varying Cu concentration in the sample. Table 9 shows the maximum, minimum and average grades for every zone.  Table 9: Maximum, Minimum and Average grades for different zones Zone Max. grade Min. grade Avg. grade SPPP07 1.96 0.03 0.30 SPPP11 2.40 0.01 0.28 Oxido 3.74 0.02 0.29 Mixto 17.45 0.03 0.59 Sulfuro 6.32 0.01 0.58 M3 2.34 0.03 0.41 Lastre 0.10 0.01 0.02 00.010.020.030.040.050.060.070.080.090.11 26 51 76 101 126 151 176 201Cu (%)Sample NumberLastre70  4.1.2.1 Classification of XRF scanned samples for assaying   The XRF scanned samples were classified into three categories based on the Cu percentage in every sample before sending the samples for assaying. Since assaying is an expensive procedure, only few samples from the zone were sent for grade determination. The samples were classified into three categories: Cu content less than 0.3%, Cu content between 0.3% -0.59% and Cu content greater than or equal to 0.6% as explained in Section 3.4. The classification is important so that samples from different grade range were chosen which helps in understanding the sensitivity of the XRF device to varying metal content in the sample. The zone wise distribution of samples based on mineral content is shown in Figure 27. The distribution of samples into three categories according to the Cu percentage is tabulated and presented in Table 10.  Table 10: Classification of samples into three categories Zone Percentage of samples <0.3% Cu 0.3%-0.59% Cu >=0.6% Cu SPPP07 67% 25% 8% SPPP11 70% 20% 10% Oxido 72% 17% 11% Mixto 78% 6% 16% Sulfuro 36% 29% 35% M3 45% 37% 19% Lastre 100% 0% 0%  71   Figure 27: Distribution of samples into three categories based on Cu % 4.1.3 Assay Results Representative samples were chosen from every zone as per the procedure explained in Section 3.5. 30 samples each from the two zones of Spence mine, 30 samples each from the four zones of Escondida mine and 15 samples from the Lastre zone of Escondida mine were sent for assaying. The results of assaying are tabulated and presented in Appendix B. The results provide the concentration of Cu in the samples. The assay results are generated in order to check the accuracy of the results generated from XRF tests. This is done by carrying out a correlation analysis between the XRF generated results and the assay result which is discussed in Section 4.2.  4.2 Correlation Analysis 4.2.1 Correlation between XRF and Assay Results Correlation between XRF and assay results was carried out by graphically plotting the values of both the tests and then generating a correlation co-efficient for every zone. The correlation 0100200300400500SPPP07 SPPP11 Oxido Mixto Sulfuro M3 LastreNumber of samplesZone NamesClassification of samples  <0.3% Cu 0.3%-0.59% Cu >=0.6% Cu72  coefficient generated is the Pearson correlation coefficient which was generated in Microsoft Excel and is denoted by ‘r’. The value of ‘r’ is a measure of the strength of a linear relationship between two variables. It is common to use r2 during correlation analysis but this is used mainly to know the proportion of variance. In the current study, the value of ‘r’ is more significant as the purpose is to analyze the relationship between the two variables (McClure, 2005). Figures 28 to 34 shows the correlation between XRF and assay results for all the seven zones. Correlation co-efficient determines the degree of correlation between the two set of results. Correlation co-efficient of 1 indicates highest level of correlation and 0 indicates that there is no correlation between the two sets of results obtained. Typically, correlation coefficient of 0-0.20 is negligible, 0.21-0.35 is weak correlation, 0.36-0.67 is moderate, 0.68-0.9 is strong and 0.91-1 is considered very strong correlation (Prion & Haerling, 2014).  The correlation co-efficient for every zone is presented in Table 11. It can be seen from the analysis that a high degree of correlation exists between the results obtained from XRF and assaying method of grade determination. High correlation co-efficient is observed for all the zones with the least value being 0.548 for Oxido zone and the highest being 0.827 for Lastre zone. The high degree of correlation suggests that the accuracy of the results generated by XRF device is convincingly good. The XRF results can hence be used in the development of models.     73  Table 11: Correlation coefficients between the XRF and Assay results Zone Correlation coefficient (r) SPPP07 0.805 SPPP11 0.707 Oxido 0.548 Mixto 0.750 Sulfuro 0.639 M3 0.650 Lastre 0.827 The correlation plots of XRF and assay results for the seven zones are as follows: 1. SPPP07   Figure 28: Correlation between XRF and Assay results for SPPP07 samples    00.511.522.530 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8Assay Cu (%)XRF Cu (%)XRF vs Assay- SPPP07 r= 0.805 74  2. SPPP11   Figure 29: Correlation between XRF and Assay results for SPPP11 samples 3. Oxido   Figure 30: Correlation between XRF and Assay results for Oxido samples 00.511.522.533.540 0.5 1 1.5 2 2.5Assay CU (%)XRF Cu (%)XRF vs Assay- SPPP11r= 0.707-505101520250 0.5 1 1.5 2 2.5Assay Cu (%)XRF Cu (%)XRF vs Assay- Oxidor= 0.54875  4. Mixto  Figure 31: Correlation between XRF and Assay results for Mixto samples 5. Sulfuro  Figure 32: Correlation between XRF and Assay results for Sulfuro samples 05101520250 2 4 6 8 10 12Assay Cu (%)XRF Cu (%)XRF vs Assay- Mixto 00.511.522.530 0.5 1 1.5 2 2.5 3 3.5 4Assay Cu (%)XRF Cu (%)XRF vs Assay- Sulfuror = 0.639r= 0.750 76  6. M3  Figure 33: Correlation between XRF and Assay results for M3 samples 7. Lastre   Figure 34: Correlation between XRF and Assay results for Lastre sample 00.20.40.60.811.21.41.61.820 0.5 1 1.5 2 2.5Assay Cu (%)XRF Cu (%)XRF vs Assay- M3r= 0.65000.010.020.030.040.050.060.070 0.01 0.02 0.03 0.04 0.05 0.06 0.07Assay Cu (%)XRF Cu (%)XRF vs Assay- Lastrer = 0.82777  From the Figures 28-34, it is observed that Oxido zone has the lowest correlation coefficient of 0.548 between the XRF and assay result. This is due to the presence of sample number 66 whose assay grade is 23% Cu and XRF grade is 2.2% Cu. Exclusion of this sample increases the correlation coefficient to 0.74. Repetition of both XRF and Assay tests on this sample did not vary the earlier results. The reason for low reading by XRF method was not very clear. The highest correlation was observed for Lastre zone since all the samples had mineral concentration in a very narrow grade range. Sulfuro and M3 zone had samples with varying Cu content and the correlation coefficient is moderate for both the zones. Hence, it can be said that, when the variability in the grades of the samples is higher the correlation is lower.    4.2.2 Correlation between XRF and HFEMS Since the XRF results showed reliability by demonstrating good correlation with assay results, the HFEMS results were plotted against XRF results and not assay results. Moreover, since both the XRF and HFEMS results are available for every sample unlike the assay results, it is easy to understand the behavior of the curve. All the 500 samples for every zone can be plotted against each other and a better comparison can be drawn. For every sample, HFEMS is measured for 14 values of magnitude and 14 values of phase at every 100 kHz interval ranging from 0-1400 kHz. Since there are 1000 Spence samples and 2215 Escondida samples, the results generated are difficult to present.  It was difficult to arrive at any conclusion since every value of magnitude and phase showed different degrees of correlation which even varied as the sample zone changed. When the HFEMS data of 14 different values of phase was compared with XRF data, every curve behaved in a different manner. Similar observations were noticed when the XRF data was compared with 14 different magnitude values. This behavior was same for all the seven zones. 78  Only for some samples the device showed sensitivity when the grade in the sample varied but when other samples from the same zone were scanned, the behavior was not similar. This could be because the samples may be having low electrical and magnetic conductivity. The mineralogy of the samples can be a reason due to which a correlation couldn’t be established with the Cu content in the sample. Good correlation was not observed between the HFEMS and XRF results and hence for the current research study only the XRF and Assay results were used.  4.3 Numerical Mine Model 4.3.1  Effect of Increasing Selectivity Table 12 summarizes the numerical mine model for different zones of Spence and Escondida mines and draws the comparison of adopting sorted model against traditional composited model. Detailed zone wise results of the numerical mine model is presented in Appendix C. The calculation of all the values in Table 12 are as per the assumptions in Table 7 in Section 3.6. As mentioned in Section 3.6.1, the cost factors are included only to highlight the difference in savings and no detailed economic analysis is done. The values in column ‘savings’ are rounded-off to the nearest thousand value. Spence mine is divided into two zones: SPPP07 and SPPP11 and Escondida mine is divided into five zones: Oxido, Mixto, Sulfuro, M3 and Lastre. Since Lastre consists of samples only below the cut-off grade, it is not considered in the development of the model. Two cases are considered for every zone, where case 1 has block models of size 25m x 5m x 5m and case 2 has blocks of size 5m x 5m x 5m. Case 1 has been termed as composited model and case 2 has been termed as sorted model. NSMU refers to the number of blocks being mined which is 100 in case of composited model and 500 in case of sorted model. 79  Table 12: Results of the Numerical Mine Model Mine Zone Case Block (t) SMU (t) NSMU % Ore blocks Cu (%) in Ore blocks % Waste blocks Cu (%) in Waste blocks Cu recovered in concentrate (t)  Savings ($) Spence SPPP07 1 143750 1437.5 100 60 0.36 40 0.20 249.00 76,000 2 143750 287.5 500 43 0.48 57 0.15 254.72 171,000 SPPP11 1 143750 1437.5 100 55 0.38 45 0.17 238.30 80,000 2 143750 287.5 500 37 0.55 63 0.13 237.70 219,000 Spence Total 1 287500 2875 200 57.5 0.37 42.50 0.18 487.30 156,000 2 287500 575 1000 40 0.51 60 0.14 492.42 391,000 Escondida Oxido 1 143750 1437.5 100 51 0.41 49 0.17 237.88 140,000 2 143750 287.5 500 35 0.61 65 0.12 247.85 325,000 Mixto 1 143750 1437.5 100 60 0.88 40 0.15 607.50 2,585,000 2 143750 287.5 500 24 2.08 76 0.12 588.72 2,879,000 Sulfuro 1 143750 1437.5 100 94 0.60 6 0.22 651.40 2,355,000 2 143750 287.5 500 69 0.79 31 0.13 634.89 2,497,000 M3 1 143750 1437.5 100 91 0.43 9 0.20 447.10 972,000 2 143750 287.5 500 66 0.53 34 0.16 417.01 1,010,000 Escondida Total 1 575000 5750 400 74 0.58 26 0.18 1943.88 6,052,000 2 575000 1150 2000 48.5 1.00 51.5 0.13 1888.47 6,711,000 80  It can be seen from Table 12 that the amount of ore being mined in both the cases of Spence mine is 287,500 tonnes each. In case of composited model, there are total of 57.5% ore blocks with 0.37% Cu and 42.5% waste blocks with 0.18% Cu. The amount of Cu recovered is 487.3 tonnes making a savings of $156,000. In the case of sorted model, there are 40% ore blocks of 0.51% Cu and 60% waste blocks with 0.14% Cu. The amount of Cu recovered is 492.42 tonnes making a savings of $391,000.  It is observed that, the sorted model has 17.5% lesser ore blocks compared to the composited model, but the Cu content is more by 0.14%. Furthermore, there are 17.5% more waste blocks in sorted model compared to the composited model but with 0.04% lesser Cu content. This clearly indicates that to recover the same amount of Cu, 17.5% lesser material has to be processed than the traditional method. The difference in the processing costs can be seen with the increase in savings from $156,000 to $391,000.    Escondida mine also showed similar results where the total amount of material being mined is 575,000 tonnes in both the sorted and composited model. In case of composited model, there are 74% ore blocks of 0.58% Cu and 26% waste blocks with 0.18% Cu. The amount of Cu recovered is 1943.88 tonnes achieving a savings of $6,052,000. In case of sorted method, there are 48.5% ore blocks of 1% Cu content and 51.5% waste blocks of 0.13% Cu. The amount of Cu recovered is 1888.47 tonnes making a savings of $6,711,000. An increase of $659,000 in savings is observed in case of sorted model but the amount of Cu recovered is lesser. The rise in savings achieved is mainly due to the higher grade of the ore blocks and also due to the lesser amount of material to be processed in the mill.  From Table 12 it can be observed that, the difference in savings achieved in case of Case 2 (sorted model) is higher compared to Case 1 (composited model) for all the six zones. In case of sorted 81  model, the size of the SMU blocks is smaller i.e. 5m x 5m x 5m. Since every block is 287.5 tonnes, to mine 143,750 tonnes of material, 500 blocks have to be mined. This gives a freedom of inspecting every 287.5 tonnes of material and categorize it into ore or waste. Thus, the heterogeneity in the orebody is being captured increasing the resolution. The hidden waste within a larger SMU block is exposed and when there is presence of waste material, it is being discarded which otherwise would have been processed in the mill incurring additional costs. The amount of material to be processed in the mill in case of composited model is higher compared to sorted model in order to recover the same amount of Cu. Since the size of SMU is five times larger than the SMU size considered in sorted model, the categorization and classification of material happens for every 1437.5 tonnes of material. Lesser selectivity is available compared to the sorted model. The presence of unwanted material within the SMU block cannot be captured and heterogeneity within the block is being encapsulated. The valuable ore which is present within the block can be segregated only after processing in the mill and not in the initial stages, thus compelling to process higher amount of material in the mill.  The difference in savings achieved between the two cases considered in Numerical mine model is presented in Section 4.3.2. Based on the cost factors considered in the development of the model, the difference in cost incurred in every stage of mining is presented in Figure 35 for Spence mine and Figure 36 for Escondida mine.      82  4.3.2 Comparison in Savings of Sorted Model and Composited Model 1. Spence Mine  Figure 35: Cost analysis of Composited model and Sorted model for Spence Mine 2. Escondida Mine  Figure 36: Cost analysis of Composited model and Sorted model for Escondida Mine  - 200,000 400,000 600,000 800,000 1,000,000 1,200,000 1,400,000 1,600,000 1,800,000Mining Sorting Processing Other TotalSpence MineSPPP07-1 SPPP07-2 SPPP11-1 SPPP11-2Cost ($)Process Type0500,0001,000,0001,500,0002,000,0002,500,000Mining Sorting Processing Other TotalEscondida MineOxido-1 Mixto-1 Sulfuro-1 M3-1 Oxido-2 Mixto-2 Sulfuro-2 M3-2Cost ($)Process Type83  In Figures 35 and 36, a comparison of mining cost, sorting cost, processing cost and other associated costs considered in the development of the model has been done. As earlier, case 1 represents composited model and case 2 represents sorted model. The suffix ‘1’ represents case 1 and suffix ‘2’ represents case 2 in the Figures 35 and 36. It is observed that the mining costs and other costs are same for both the cases. An additional sorting cost is associated with the sorted model which is a very small component compared to the rest of the cost components involved. Since there is no sorting happening in case of composited model, no additional costs are added to that stage. A significant difference is observed between the processing costs for the two models. The processing cost in case of composited model is much higher compared to the processing cost of sorted model which is directly affecting the total savings and hence, making the sorted model more resourceful.  4.4 Vulcan Model 4.4.1 Effect of Changing Block Model Size Table 13 represents the summary of results generated from the Vulcan model. Zone wise results of the Vulcan model are presented in Appendix D. The calculation of all the values in Table 13 are as per the assumptions in Table 7 in Section 3.6. The Spence mine is divided into two zones: SPPP07 and SPPP11 and Escondida mine is divided into five zones: Oxido, Mixto, Sulfuro, M3 and Lastre as done in the development of Numerical mine model. Two cases are considered for every zone where, case 1 has SMU of size 25m x 5m x 5m and case 2 has SMU of size 5m x 5m x 5m. The development of the Vulcan model differs from the Numerical mine model in two aspects. The estimation of resources for both the cases in Vulcan model is done by inverse distance weighting technique where as in the case of Numerical mine model, the resources are estimated 84  mathematically. The other difference is that no sorting of ore takes place in the Vulcan model as considered in the case of Numerical mine model. Hence the aim of the Vulcan model is to use the data generated by increasing selectivity in the development of high resolution deposit modeling.    85  Table 13: Results of the Vulcan Model Mine Zone Case Block (t) SMU (t) NSMU % Ore blocks Cu (%) in Ore blocks % Waste blocks Cu (%) in Waste blocks Cu recovered in concentrate (t) Spence SPPP07 1 345000 625 240 57.08 0.38 42.92 0.20 52.42 2 345000 125 1200 57.00 0.39 43.00 0.19 265.34 SPPP11 1 345000 625 240 45.42 0.39 54.58 0.17 42.78 2 345000 125 1200 47.42 0.42 52.58 0.17 236.17 Spence Total 1 690000 1250 480 51.25 -  48.75 -  95.20 2 690000 250 2400 52.21 -  47.79         - 501.51 Escondida Oxido 1 362250 625 252 51.59 0.49 48.41 0.17 63.25 2 355350 125 1236 45.31 0.45 54.69 0.17 250.81 Mixto 1 362250 625 252 50.40 0.94 49.60 0.15 119.87 2 355350 125 1236 49.11 0.98 50.89 0.14 591.53 Sulfuro 1 345000 625 240 89.58 0.62 10.42 0.22 134.27 2 355350 125 1236 88.43 0.62 11.57 0.19 674.50 M3 1 362250 625 252 90.87 0.45 9.13 0.20 103.18 2 355350 125 1236 86.49 0.44 13.51 0.20 466.16 Escondida Total 1 1431750 2500 996 70.38 -  29.62 -  420.57 2 1421400 500 4944 67.33        - 32.67 -  1983.00 86  The effect of SMU size on the amount of Cu extracted is evident for both the Spence and Escondida mines. Considering Spence mine, the amount of material mined is same in both the cases which is 690,000 tonnes. Since similar grade estimation techniques are applied in both the cases, the number of ore blocks and waste blocks are almost the same but with significant difference seen in the amount of Cu recovered. It is seen that in Case 1 only 95.2 tonnes of Cu is recovered but in case of case 2, 501.51 tonnes is recovered which is more profitable than case 1. In case of Escondida mine, 420.57 tonnes of Cu is recovered in Case 1 and 1983 tonnes of Cu in case 2. The amount of material mined again is almost similar but the Cu recovered is more than 5 times higher. The effect of the block model size is clearly visible and it can be said that, as the size of the block model decreases, more amount of mineral can be recovered. Also, with the decrease in the size of blockmodel, information about very small mass of material is known which helps in classification of that smaller mass into ore or waste. Material identified as waste can be immediately discarded and only the valuable ore will be treated in the mill. Thus, reduction in size of SMU has several advantages and the main aim of the research study, to efficiently mine a heterogeneous orebody, is achieved.   4.5 Conclusions from Numerical Mine Model and Vulcan Model The interpretation of Numerical mine model shows that, reducing the size of selective mining unit increases the selectivity while mining. This helps in selective partitioning of the mined material. Segregation of ore and waste immediately after excavation of ore, facilitates in early elimination of waste without further handling and processing. A significant improvement in processing costs was observed thus proving the importance of increasing selectivity during mining. The Vulcan model showed that, reducing the SMU size while mining improves the selectivity and contributes 87  in increased recovery of mineral from the ore. The reduced SMU size helps in capturing the small amount of ore present and recover it, thus helping in maximum utilization of resources.    Comparing case 1 of Numerical mine model with case 2 of Numerical model and case 1 of Vulcan model with case 2 of Vulcan model, it was observed that, the amount of Cu recovered in both cases of Numerical mine model differs by a very small amount. However, in Vulcan model the difference between the two cases is more than 4 times. In numerical mine model, almost same amount of ore is mined in both the cases and same amount of Cu is recovered, but the effect of smaller SMU size is seen on the savings achieved. The savings varies from $6,208,000 to $7,102,000. The additional savings achieved in Case 2 is because lesser material is processed in the mill. Smaller SMU size facilitates higher selectivity which helps in discarding the unwanted material without processing in the mill. In the sorted model (case 2), the number of ore blocks is lesser but of higher grade and number of waste blocks is more but of lower grade. This is because, the SMU size is smaller and the heterogeneity in the orebody is being captured, thus reducing dilution.  The difference in values seen between the two cases of Vulcan model is due to the difference in selectivity in the two cases. Case 2, with smaller SMU’s is highly selective and hence the amount of Cu recovered is more compared to Case 1. The higher recovery of Cu in Case 2 compared to Case 1 is because of the decrease in SMU size, which increases the selectivity. Thus it can be inferred from the study that, smaller size of SMU helps in identification of the heterogeneity in a mineral resource and captures the smallest amount of ore hidden within the blocks. This makes the process more efficient and economically profitable with higher metal recovery.     88  Chapter 5: Discussion 5.1 Introduction The results presented in Chapter 4 are further discussed and interpreted in this chapter, which explains and supports the importance of this research work in the mining industry. Firstly, the results on the possibility of the application of preconcentration stage is discussed. This is followed by the analysis of the results generated through the development of two digital models- Numerical mine model and Vulcan model. Finally, the interpretations and limitations of various methods adopted in this research study is presented. The chapter concludes with the summarization of the scope of application of the proposed technology giving an overview of the advantages the research poses to the mining industry.  5.2 Analysis of Introducing Preconcentration at Excavation Stage A new method of ore beneficiation technology is proposed and introduced in this research study. Since this proposal is for the application of a preconcentration step at the excavation stage, all the benefits of preconcentration are accomplished in addition to the benefits of eliminating waste from the mining cycle at the point of excavation. As per the technology proposed in this research study, an excavator bucket is used as the sorting equipment. This is done by affixing sensors to the excavator bucket which can determine the mineral content of the material in the bucket. Since the mineralogical composition of the material in the bucket is known, the excavator bucket is proposed to be the new SMU and the size of the SMU will be equal to the excavator bucket size. This makes the process of grade determination faster, easier and also eliminates the risk of assaying samples getting adulterated. Moreover, since the grade of the material is known immediately after scooping, it can be decided if the scooped material is below or above the cutoff grade. The 89  destination of the material can be immediately decided without further handling. Preconcentration or sorting of the mined material at excavation stage results in reduction of haulage costs, processing costs, capital costs and lesser generation of tailings minimizing the impact on environment. However, precisely separating the ore and waste in a mine using an extremely big excavator seems to be impossible. For example the bucket size of P&H 4100 XPC shovel is 65m3. Accurate cutting of wedges in the ore-waste contact zone is practically challenging and not feasible. Dilution caused due to this is unavoidable and it always seems to pose a challenge to the mining industry.  The scope of application of the sensor techniques used in this research are discussed in the next section.   5.2.1 Application and Limitations of using XRF and HFEMS in Ore Grade Determination The sorting techniques used in this research are HFEMS and XRF. X-ray fluorescence has been widely used in the determination of metal content from various kinds of minerals. The results obtained from the scanning of Spence and Escondida samples showed a potential for the application of this technology. However, the success of this technology depends on two factors: accuracy of results generated and feasibility of application in mines. The accuracy of the XRF results were verified by chemically assaying the samples already tested on XRF device and then, carrying out a correlation analysis between the two set of results. The summary of correlation coefficient generated was presented in Table 11 in Section 4.2.1. This showed that the correlation coefficients varied from 0.548 to 0.827. The correlation coefficient for more than four out of seven zones was above 0.7 which shows the reliability of XRF technology. An interesting observation is that, when there was a deflection in results, mineral content determined by XRF was sometimes lower and sometimes higher than the assay readings. This 90  suggests that the variance could be due to the XRF device or some factors associated with assaying. The other challenge lies in the practical application of sensors in excavator buckets. Sensors are very sensitive devices and an excavator is always exposed to harsh mining conditions which makes the sensors prone to damage. Since the accuracy of the results from sensors is important, it is vital to protect the sensors and ensure its sensitivity to metal content in the material. The application of this technology by installing sensors in the excavator bucket is challenging but it may not be too difficult to overcome this problem with the advancement of technology. This research study showed a potential for the application of XRF technology in grade determination, but further research work is necessary in order to attain highest level of accuracy.  The application of High frequency electromagnetic sensors (HFEMS) based on current study did not yield results to support the application of this technology. No correlation could be established with the XRF results or the assay results. Literature has been presented on the successful application of this technology but from the current study no conclusion could be drawn. Hence it is suggested to continue the research on application of HFEMS by varying the experimental conditions such as using samples with changing mineralogy or concentration of metal content.   5.3 Interpretation of Numerical Mine Model and Vulcan Model Numerical mine model and Vulcan model help in understanding the effect of reducing SMU size to increase the selectivity in mining, and enhances the accuracy of resource estimation.  The impact of increasing selectivity can be explained with Numerical mine model. In this model the estimation of resources is done by mathematical calculations and no interpolation techniques are applied. Thus the total amount of mineral that can be extracted from an orebody is almost similar when the SMU size is changed. This model helps in understanding the effect of increasing selectivity in 91  mining due to the simple calculation methods adopted. Higher selectivity is achieved by the application of a sorting stage after mining. Since the excavation happens before the sorting stage with the same amount of material being mined in both the cases, the mining costs are same for both the cases. The “other costs” also remains the same because this is dependent on the total material mined. The difference is visible in the case of sorting and processing costs. The advantage of increasing selectivity is attained by the application of a sorting process. Hence there is an additional cost associated with the sorting process as well. The outcome of increasing selectivity is better explained with the help of pie charts considering the two cases of SPPP07 and SPPP11 zone. Case 1 for SPPP07 zone is summarized in Table 14. Table 14: Results of increasing selectivity for SPPP07 zone- Case 1 Ore Blocks Waste Blocks Total tonnage Cu recovered (t) Savings ($) 60 40 143750 249 76,000   Figure 37: Ore- Waste distribution for SPPP07 zone- Case 1 Ore blocks -60%, 0.36% CuWaste blocks -40%, 0.2% CuSPPP07- Case 1Ore blocksWaste blocks92  Figure 37 above shows the distribution of ore and waste in the SPPP07 zone of Spence mine for a SMU size of 25m x 5m x 5m. 60% of the material is ore at 0.36% Cu and 40% of the material is waste at 0.2% Cu. The model showed that, 249 tonnes of Cu can be extracted achieving a savings of $76,000. The same orebody was mined by reducing the SMU size from 25m x 5m x 5m to 5m x 5m x 5m. The observations made are tabulated in Table 15 and represented in Figure 38. Table 15: Results of increasing selectivity for SPPP07- Case 2 Ore Blocks Waste Blocks Total tonnage Cu recovered (t) Savings ($) 217 283 143750 245.72 171,000   Figure 38: Ore- Waste distribution for SPPP07 zone- Case 2 It is seen that the amount of Cu recovered with reduced SMU size is 245.7 tonnes. Interestingly the amount of Cu recovered in both the case is almost similar. The cost of production is $1.5M and returns earned is $1.7M. Hence the savings is $171,000. This was 2.25 times more than the savings in case 1 with larger SMU. The pie chart analysis for case 2 showed that, the amount of ore decreased from 60% to 43%, but the average grade increased from 0.36% Cu to 0.48% Cu. On Ore Blocks -43%, 0.48% CuWaste Blocks -57%, 0.15% CuSPPP07- Case 2Ore BlocksWaste Blocks93  the other hand, the quantity of waste increased from 40% to 57% and the grade decreased from 0.2% Cu to 0.15% Cu. Thus introduction of a sorting process helps in eliminating the extra waste identified without processing in the mill. Thus lesser material was processed in the mill which resulted in lesser processing costs. The processing costs decreased from $863,000 in Case 1 to $624,000 in Case 2. Even though there was an additional sorting cost of $1/tonne in Case 2, the operation turns out to be much more economical. The savings in Case 2 was $171,000 as opposed to $76,000 in Case 1. The increase in savings was mainly because lesser amount of material was processed in the mill due to increased selectivity.   Similar observations were made for SPPP11 zone of Spence mine. The conditions for the two cases are same as explained earlier for SPPP07 zone. In case 1, there are 55% ore blocks weighing 79,062.5 tonnes at 0.38% Cu and 45% waste blocks weighing 64,687.5 tonnes at 0.17% Cu. The cost of production is $1,588,000 and revenue generated is $1,668,000. The savings achieved is $80,000. Summary of Case 1 for SPPP11 zone is presented in Table 16 and the pie chart distribution is shown in Figure 39.   Table 16: Results of increasing selectivity for SPPP11- Case 1 Ore Blocks Waste Blocks Total tonnage Cu recovered (t) Savings ($) 55 45 143750 238.3 80,000  94   Figure 39: Ore- Waste distribution for SPPP11 – Case 1 In case 2, there are 37% ore blocks at 0.55% Cu weighing 52,900 tonnes and 63% waste blocks at 0.13% Cu weighing 90,850 tonnes. The cost of production is $1,301,000 and the revenue earned is $1,664,000. The savings achieved is $219,000. A summary of case 2 in SPPP11 zone is presented below in Table 17 and the pie chart distribution is shown in Figure 40. Table 17: Results of increasing selectivity for SPPP11- Case 2 Ore Blocks Waste Blocks Total tonnage Cu recovered (t) Savings ($) 184 316 143750 237.7 219,000  Ore Blocks -55%, 0.38% CuWaste Blocks -45%, 0.17% CuSPPP11- Case 1Ore BlocksWaste Blocks95   Figure 40: Ore- Waste distribution for SPPP11 – Case 2 As seen in case of SPPP07, similar results were observed for SPPP11 mine. The quantity of ore decreased from 55% to 37% but the average grade increased from 0.38% Cu to 0.55% Cu. The quantity of waste increased from 45% in Case 1 to 63% in Case 2 and the grade decreased from 0.17% Cu to 0.13% Cu. It is evident from the study that, preconcentration of the mined material by reduction in SMU size increases the profits earned due to more waste being identified very early in the mining cycle and eliminating it without further treatment. Moreover, the amount of mineral content being discarded in the waste is decreased, thus making better utilization of resources. A summary of Spence mine considering both SPPP07 and SPPP11 zone is presented below. Table 18 shows the results for case 1 of SPPP07 and SPPP11 zone. Table 19 shows the results for case 2 of SPPP07 and SPPP11 zone.   Ore Blocks -37%, 0.55% CuWaste Blocks -63%, 0.13% CuSPPP11- Case 2Ore BlocksWaste Blocks96  Case 1: Table 18: Results of increasing selectivity for Spence Mine- Case 1 Ore Blocks Waste Blocks Total tonnage Cu recovered (t) Savings ($) 57.5% (0.37% Cu) 42.5% (0.18% Cu) 287,500 487.4 156,000 Case 2: Table 19: Results of increasing selectivity for Spence Mine- Case 2 Ore Blocks Waste Blocks Total tonnage Cu recovered (t) Savings ($) 40% (0.51% Cu) 60% (0.14% Cu) 287,500 483.4 391,000 To strengthen the observations made from Spence mine, similar studies were carried for the four zones of Escondida mine. Tables 20 to 23 describes the distribution of ore and waste in every zone of Escondida mine and the difference in recovery of Cu for the two cases of mining considered. The table also shows the difference in savings achieved in all the zones comparing the two cases.  Oxido zone: Table 20: Results of increasing selectivity for Oxido Zone  Ore Blocks Waste Blocks Total tonnage Cu recovered (t) Savings ($) Case 1 51% (0.41% Cu) 49 (0.17% Cu) 143,750 237.88 140,000 Case 2 35% (0.61% Cu) 65% (0.12% Cu) 143,750 247.85 325,000 Mixto zone: Table 21: Results of increasing selectivity for Mixto Zone  Ore Blocks Waste Blocks Total tonnage Cu recovered (t) Savings ($) Case 1 60% (0.88% Cu) 40 (0.15% Cu) 143,750 607.5 2,585,000 Case 2 24% (2.08% Cu) 76% (0.12% Cu) 143,750 588.72 2,879,000    97  Sulfuro zone: Table 22: Results of increasing selectivity for Sulfuro Zone  Ore Blocks Waste Blocks Total tonnage Cu recovered (t) Savings ($) Case 1 94% (0.6% Cu) 6 (0.22% Cu) 143,750 651.4 2,355,000 Case 2 69% (0.79% Cu) 31% (0.13% Cu) 143,750 634.89 2,497,000 M3 zone: Table 23: Results of increasing selectivity for M3 Zone  Ore Blocks Waste Blocks Total tonnage Cu recovered (t) Savings ($) Case 1 91% (0.43% Cu) 9 (0.2% Cu) 143,750 447.1 972,000 Case 2 66% (0.53% Cu) 34% (0.16% Cu) 143,750 417.01 1,010,000 A case wise analysis of Cu recovery is presented in Figure 41 graphically. A zone wise comparison of the two cases described in the Numerical mine model with distribution of ore blocks and waste blocks in presented. Considering ore blocks, Case 1 with larger SMU’s has more amount of recoverable Cu compared to case 2. But the average grade of Cu in case 2 is higher than case 1. In waste blocks it is observed that, case 1 with larger SMU’s has lesser waste than case 2 with smaller SMU’s but the average grade of Cu in case 2 is lesser than case 1. It can be seen that the amount of ore has always decreased in Case 2 and amount of waste has increased. An interesting observation to be made here is that, even though the quantity of ore is decreasing from case 1 to case 2, the Cu concentration is increasing. Similarly, the quantity of waste material is increasing from case 1 to case 2 but with reduction in Cu concentration. This signifies that, some amount of Cu which was earlier categorized as waste is now identified as ore. This is seen with the increase in ore grade and decrease in waste grade. This doesn’t interpret that the amount of recoverable Cu has decreased, but since the amount of ore has decreased the amount of material to be processed has decreased in order to recover the same amount of Cu. This analysis proves the benefits of implementing smaller SMU’s in mining. 98   Figure 41: Ore- Waste recovery in Escondida Mine Presenting all the four zones as one mine, the following observations are made: Case 1 for Escondida Mine: Table 24: Results of increasing selectivity for Escondida Mine- Case 1 Ore Blocks Waste Blocks Total tonnage Cu recovered (t) Savings ($) 74% (0.58% Cu) 26% (0.18% Cu) 575,000 1943.88 6,052,000  0102030405060708090100Ore Waste Ore Waste Ore Waste Ore WasteOxido Mixto Sulfuro M3Ore-Waste recoveryCase 1 Case 2Zones in Escondida Weight %99   Figure 42: Ore- Waste distribution for Escondida Mine- Case 1 Case 2 for Escondida Mine: Table 25: Results of increasing selectivity for Escondida Mine- Case 2 Ore Blocks Waste Blocks Total tonnage Cu recovered (t) Savings ($) 48.5% (1% Cu) 51.5% (0.13% Cu) 575,000 1888.47 6,711,000   Figure 43: Ore- Waste distribution for Escondida Mine- Case 2 Ore blocks -74%, 0.58% CuWaste blocks-26%, 0.18% CuEscondida Mine- Case 1Ore blocksWaste blocksOre blocks-48.5%, 1% CuWaste blocks-51.5%, 0.13% CuEscondida Mine- Case 2Ore blocksWaste blocks100  Table 24 and Figure 42 shows the distribution of ore-waste in Escondida mine for case 1. Table 25 and Figure 43 shows the distribution of ore and waste in Escondida mine for case 2. The cost savings achieved for Escondida mine are substantial. Savings of $6,052,000 was generated in Case 1 and $6,711,000 was generated in case 2. A rise in savings of 10% is achieved. It is very interesting to note that the quantity of ore blocks reduced from 74% to 48.5% with increase in grade from 0.58% Cu to 1% Cu. Initially there were only 26% waste blocks at 0.18% Cu which after the application of sorting by reducing SMU size increased to 51.5% at 0.13% Cu. The results of Escondida mine are clearer and more convincing when compared to Spence mine since there is a significant uplift of grade in the ore blocks.  The Vulcan model showed the effect of changing SMU size on the amount of Cu extracted. When larger SMU size of 625 m3 was considered, the Cu that could be extracted from Spence mine was 95.20 tonnes and from Escondida mine was 420.57 tonnes. When the SMU size was decreased to 125 m3, the amount of Cu that can be extracted from Spence mine was 501.51 tonnes and from Escondida mine was 1983 tonnes.  A graphical representation of Cu recovered for every zone from Spence mine and Escondida mine is presented and the two cases as explained are shown in Figure 44 and 45. The Cu recovered increased from case 1 to case 2 for both the Spence and Escondida mines.      101  Cu recovered from Spence Mine:  Figure 44: Recovery of copper in Spence Mine Cu recovered from Escondida Mine:  Figure 45: Recovery of copper in Escondida Mine 050100150200250300SPPP07 SPPP11Cu Recovery-Spence MineCase 1 Case 2ZonesCu recovereed (t)0100200300400500600700800Oxido Mixto Sulfuro M3Cu Recovery- Escondida MineCase 1 Case 2Cu recovereed (t)Zones102  The reduction in SMU size increased the amount of recoverable copper by 4.8 times considering both the Spence and Escondida mines together. The method of resource estimation adopted in both the cases is same i.e. Inverse distance weighted method. Hence the difference in reserves estimated is due to the increase in SMU size. This is because, increase in SMU size increases the zone of influence of given known data points. Interpolation by inverse distance weighted method depends on grade of neighboring points and all the neighboring points influence in estimating the reserves of the orebody. The presence of low grade zone in the neighborhood may lead to underestimation of resources or the presence of high grade resources might lead to overestimation. Thus some parts of the orebody which is made of valuable material might be regarded as waste and some parts of waste material might be considered to be valuable. This problem can be overcome by generating fine granular information, which depends on the reduced spacing of the drill holes generating more information. From the literature it is clear that, increased information results in better prediction or estimation of resources. Thus, decrease in the SMU size helps in better estimation of resources and assists in increased extraction of mineral. The research findings are in accordance with most of the other previous findings. Reduction in SMU size is believed to benefit the mining industry by improving recovery but in order to reduce the SMU size it is necessary to generate the required additional drill hole data before resource estimation (Jara et al., 2006). The technology of using sensors in excavator bucket proposed in this research overcomes this problem and generates value and information of every bucket of material being mined. This also provides the mineralogy of the material in the bucket instantaneously. Thus, when every bucket of material is mined, the material can be immediately classified as ore or waste. Depending on the size of the excavator bucket, the SMU size changes. This helps in generating 103  information of a very small mass of material being mined to the extent of the size of an excavator bucket. With this high resolution information of an orebody, the distribution of ore and waste is very clear. Every block of ore present can be mined and recovered. Thus, heterogeneity in an orebody is approached and properly handled. The problem of dilution also reduces and the chances of waste getting mixed with ore and ore getting mixed in waste reduces.  Thus reduction in SMU size seems to be beneficial to the mining industry. The objective of identifying inherent heterogeniety in order to reduce dilution of ore and better metal extraction can be achieved.   5.4 Conclusion The Numerical mine model clearly signifies the need to increase selectivity in mining and the Vulcan model signifies the need to generate more data for better reserve estimation. Both the needs are served with reduction in SMU size. This research study demonstrates a great scope for improving the project economics particularly in low grade mines. With reduction in cost and lesser harm to the environment, efforts should be taken to improve the technology and use it in industrial scale.          104  Chapter 6: Conclusions and Recommendations  6.1 Introduction The approach of selective partitioning of ore in order to achieve, higher selectivity in mining heterogeneous ore bodies is suggested as a potential alternative in mines to improve mine valuation. The research study focused on achieving higher selectivity and practicing high resolution deposit modeling for resource estimation. The deposits most benefitting from this innovation is, low grade heterogeneous ore bodies. The introduction of a selective partitioning tool can successfully identify the inherent heterogeneity within an orebody and segregate the ore and waste. The technology proposed in this research work to achieve the objective was with the help of a sensor mounted excavator bucket which can scan and determine the grade of the material in the bucket instantaneously. The sensor enabled excavator bucket, thus has the ability of decision making and acts as a sorter. The sorter is also the new SMU which is considered in resource estimation. This contributes to high resolution deposit modeling. The results from the case study carried out indicate that the method of preconcentrating the ore immediately after excavation before further handling is beneficial than the traditional approach. The process of increasing selectivity in mining through the introduction of a new method of preconcentration generating high resolution data is believed to contribute for better utilization of mineral resources and make a mining project more profitable. This chapter comprises the thesis summary, the contributions of the research and recommendation for future work. Finally it provides a conclusion for the research study.         105  6.2 Thesis Summary The research work was divided into two phases. In the first phase, rock samples from two mines were tested to determine the metal content in the samples electronically using XRF and HFEMS sensors. The samples were then chemically assayed to check the accuracy of the sensors used and  XRF method was concluded as the reliable method. It helped in understanding the variability in the distribution of grade in the orebody.  The second phase of the research work was to use the sensor generated data in resource modeling. To check the effect of increasing resolution of data generated for resource estimation and increasing selectivity in mining, two models were developed. The models showed that increasing the resolution by reducing the SMU size contributes for better utilization of resources and higher recovery of mineral. Increasing the selectivity contributes in elimination of waste, reduces dilution in mining, enhances metal concentration and improves the profits. Thus the research study was successful in achieving the objectives of the study.       6.3 Contributions of the Research The use of the technique proposed in this research work using the tool which was developed by MineSense Technologies Ltd®, is an automatic and instantaneous method of grade determination of the material mined and the innovation is believed to be a significant contribution to the mining industry. Particular benefits have been identified for optimization of low grade heterogeneous ore bodies. The approach is also an ore beneficiation method through preconcentration. According to this approach, ROM is classified into ore and waste immediately after excavation. The study showed that 18-25% of ROM can be rejected as waste and this waste is eliminated from all the downstream processes. This results in 33% decrease in processing cost, 13% decrease in operating 106  cost, decrease in capital costs, haulage costs and resultant improvement in savings of around 13%. The approach improves the metallurgical recovery compared to unsorted ore, reduces the cut-off grade of the mine leading to enhanced mine life, reduces the amount of tailings generated which results in reduced harmful effects to the environment and causes smaller surface footprint. Increasing the resolution of handling an orebody helps in identifying the heterogeneity, reduces dilution and wastage of ore, increases ore reserves due to better utilization of resources. The current study proves that, reduction in the SMU size increases selectivity in mining and contributes in better utilization of resources by capturing the heterogeneity of the rocks.     6.4 Recommendations for Future Work The case study of Spence and Escondida mines showed several benefits to the mining industry by the proposed technology. However, substantial amount of further research work is required to practically implement the system in mines. Based on the study, the following recommendations were suggested. The results of XRF tests for particle sizes greater than 76.2 mm were not well correlated with the assay results. This is because, X-Rays have low penetration capacity. In case of larger rocks, only the outer layer mineralogy is known but the inner hidden minerals cannot be determined. A more advanced method of grade determination which has higher penetration capacity is sought. The XRF results for 25.4 mm-76.2 mm particle size showed good correlation which was useful for the current study but higher degree of correlation is necessary for practical applications. Application of HFEMS for determining the mineral content in the rock samples requires extensive research work. In the current study, the range of magnitude and phase was from 0-1400 kHz. Since no conclusions could be drawn for the range used, it is suggested to use different values of magnitude and phase. Also, the experimental conditions can be changed from the current 107  conditions using samples of different mineralogy, samples of different mineral concentration in order to check the sensitivity of the sensors.  There is a potential to develop the Numerical mine model and Vulcan model and evaluate the impact of using increased selectivity in orebody modeling, mine planning and scheduling. A trade-off study can be carried out to evaluate the way ore control and mine dispatch system works. Particular consideration can be given to the dispatching of trucks. Instead of scheduling the trucks to destination based on the mine plan, the trucks can be scheduled to various destinations such as the leach pad, processing mill, mill feed stockpile or waste dump based on the grade measured by sensors. This has the potential to considerably reduce the misallocations of ore to wrong destination and hence improve the overall grade of the ore delivered to various destinations. Based on the feasibility of logistics two scenarios can be considered: a) The ore is discriminated from the waste on a truck by truck basis b) Double-sided trucks are considered where one truck is loaded with only ore and other with only waste i.e. discrimination of ore is based on shovel by shovel basis. A study should be carried out to compare the costs incurred in the proposed method of truck dispatch model and the current method being practiced in the mines.   A case study can be carried out using actual geological data from mine to design the mine plan and manage mine scheduling.   Based on the results obtained from the current research study, the two models can be further developed with different SMU sizes to analyze the relationship between SMU size and mineral extraction. The impact of varying the selectivity can also be analyzed to improve mine valuation. The accuracy of the methods adopted in resource estimation can be increased with the application of statistical analysis and geostatistical estimation techniques.  108  6.5 Concluding Thoughts This research identifies the benefits of application of sensors in excavator buckets to preconcentrate the ore in early mining stages and also the benefits of employing high resolution deposit modeling. The technique introduced in this research is a real-time ore grade determination and decision making system. The study demonstrated significant economic and environmental benefits but the practical application of this technique might be challenging in the mines in terms of both technical and economic feasibility. The growing demand for metals and rise in price of metals motivates further work towards the application of this technique. The research can be further developed by carrying out additional case studies to increase the confidence and range of applicability. This study can be carried out on other low grade base metals or precious metal ores since the potential for upgradation is very high.              109  Bibliography Allen, A., & Gordon, H. (2009). X-Ray sorting and other technologies for upgrading nickel ore. In: ALTA 2009 Nickel- Cobalt, Copper and Uranium conference, Perth. Alpers, C.N., & Brimhall, G.H. (1989). Paleohydrologic evolution and geochemical dynamics of cumulative supergene metal enrichment at La Escondida, Atacama Desert, northern Chile. Economic Geology, 84, 229-255. Altun, N.E., Weatherwax, T., & Klein, B. (2013). Upgrading Valuable Mineralization and Rejecting Magnesium Silicates by Pre-Concentration of Mafic Ores. Physiochemical Problems of Mineral Processing, 50 (1), 203-215. ArcGIS Resources (2011). Geoprocessing. ArcGIS Resource Center retrieved on November 9, 2014 from http://help.arcgis.com/en/arcgisdesktop/10.0/help/index.html#//009z00000076000000.htm Arriagada, I. (2012). Pampa Norte Spence site tour. Retrieved on October 12, 2014 from http://www.bhpbilliton.com/home/investors/reports/Documents/2012/121002_Spence Site Visit Presentation.pdf  Asano, A.T., & Dessureault, S.D. (2009). Block model reconstruction using truck fleet management system data. Mining Technology, 118 (2), 53-58. Ballantyne, G.R., Hilden, M., Powell, M.S. (2012). Early rejection of gangue -how much energy will it cost to save energy? In Comminution ’12, ed. Wills, B. Mineral Engineering, Capetown, South Africa. Bamber, A. (2008). Integrated mining, pre-concentration and waste disposal systems for the increased sustainability of hard rock metal mining. PhD thesis. University of British Columbia, Vancouver. Bamber, A., Klein, B., Morin, M., & Scoble, M. (2004). Reducing Selectivity in Narrow-Vein Mining through the Integration of Underground Pre-concentration. In Proceedings IV CANMET International Symposium on Narrow Vein Mining Techniques. Basto, E. (2012). Escondida site tour. Retrieved on October 12, 2014 from http://www.bhpbilliton.com/home/investors/reports/Documents/2012/121001_Escondida Site Visit Presentation.pdf Bergmann, C. (2009). Developments in ore sorting technologies- Mintek. Council for mineral technology developments in ore sorting technologies. .  110  BHP Billiton (2004). Spence Briefing retrieved on July 3, 2014 from http://www.sec.gov/Archives/edgar/containers/fix031/811809/000081180904000090/spencebriefing.pdf BHP Billiton (2007). BHP Billiton Inaugurates Spence Copper Operation. Engineering and Mining Journal, 208(4), 4-5.  BHP Billiton (2011). BHP Billiton approves USD 554 Million investment at Escondida retrieved on August 8, 2015 from http://www.bhpbilliton.com/home/investors/news/Pages/Articles/BHP%20Billiton%20Approves%20USD554%20Million%20Investment%20at%20Escondida.aspx BHP Billiton (2014). BHP Billiton Annual Report 2014. Retrieved on September 28, 2014 from http://www.bhpbilliton.com/home/investors/reports/Documents/2014/BHPBillitonAnnualReport2014_interactive.pdf  Bohling, G. (2005). Kriging. Retrieved on June 27, 2014 from http://people.ku.edu/~gbohling/cpe940/Kriging.pdf Boland, N., Dumitrescu, I., Froyland, G., & Gleixner, A. M. (2009). LP-based disaggregation approaches to solving the open pit mining production scheduling problem with block processing selectivity. Computers & Operations Research, 36(4), 1064-1089. Boucher, A., Dimitrakopoulos, R., & Vargas-Guzmán, J. A. (2005). Joint simulations, optimal drillhole spacing and the role of the stockpile. In Geostatistics Banff 2004. Springer Netherlands, 35-44. Bowman, D. J., & Bearman, R. A. (2014). Coarse waste rejection through size based separation. Minerals Engineering, 62, 102-110. Bozorgebrahimi, A., Hall, R., & Morin, M. (2005). Equipment size effects on open pit mining performance. International Journal of Surface Mining, Reclamation and Environment, 19(1), 41–56.  Brooker, G., Hennessey, R., Lobsey, C., Bishop, M., & Widzyk‐Capehart, E. (2007). Seeing through dust and water vapor: millimeter wave radar sensors for mining applications. Journal of Field Robotics, 24(7), 527-557. Cameron, E, M., & Leybourne, M, I. (2005). Relationship between groundwater chemistry and soil geochemical anomalies at the Spence copper porphyry deposit, Chile. Geochemistry - Exploration, Environment, Analysis, 5(2), 135-145 Caulfield, P. (2011). Upfront processing: Bigger not always better. CIM Magazine, 6 (1), 26-27. 111  Celis, B. (1996). X-ray fluorescence analysis of gold ore. Applied spectroscopy, 50(5), 572-575. Chatterjee, S., Bhattacherjee, A., Samanta, B., & Pal, S. K. (2010). Image-based quality monitoring system of limestone ore grades. Computers in Industry, 61(5), 391-408. Dalm, M., Buxton, M. W., van Ruitenbeek, F. J., & Voncken, J. H. (2014). Application of near-infrared spectroscopy to sensor based sorting of a porphyry copper ore. Minerals Engineering, 58, 7-16. Darling, P. (2011). Mineral Resource Estimation (3rd Ed.). SME Mining Engineering Handbook. United States: SME, 203-217. De Jong, T., & Dalmijn, W. (2007). DEXRT Imaging For Automatic Particle Sorting. Domic, E. M. (2007). A review of the development and current status of copper bioleaching operations in Chile: 25 years of successful commercial implementation. In Biomining, Springer Berlin Heidelberg, 81-95. Dominy, S. C., Noppé, M. A., & Annels, A. E. (2004). Errors and uncertainty in mineral resource and ore reserve estimation: The importance of getting it right. Exploration and Mining Geology, 11(1-4), 77-98. Emery, X., Ortiz, J. M., & Rodríguez, J. J. (2006). Quantifying uncertainty in mineral resources by use of classification schemes and conditional simulations. Mathematical geology, 38(4), 445-464. Fickling, R. S. (2011). An introduction to the Rados XRF ore sorter. In 6th Southern African Base Metals Conference, Phalaborwa, South, Africa, 99-110. Georget, J. P., & Lambrecht, R. (1982). U.S. Patent No. 4,361,289. Washington, DC: U.S. Patent and Trademark Office. Ghosh, A., Sharma, A. K., Nayak, B., & Sagar, S. P. (2014). Infrared thermography: An approach for iron ore gradation. Minerals Engineering, 62, 85-90. Grewal, I. (2013). Separation and concentration techniques. Mineral Processing Introduction. Retrieved on January 19, 2015 from http://met-solvelabs.com/library/articles/mineral-processing-introduction/  Groves, D. I., Vielreicher, R. M., Goldfarb, R. J., & Condie, K. C. (2005). Controls on the heterogeneous distribution of mineral deposits through time. Geological Society, London, Special Publications, 248(1), 71-101.  112  Gulcan, E., & Harzanagh, A. (2013). Optical sorting of iron ore tailings. Paper presented at the 15th Balkan Mineral Processing Congress- BMPC 2013, Bulgaria. Habashi, F. (2006). A Short History of Mineral Processing. In: Proceedings XXIII International Mineral Processing Congress, 3–8. Harbeck, H. & Kroog, H. (2008). New Developments in Sensor Based Sorting, Montan University Loeben, Austria, p. 272.  Hekmat, A., Osanloo, M., & Moarefvand, P. (2013). Block size selection with the objective of minimizing the discrepancy in real and estimated block grade. Arab J. Geosci. 6, 141- 155. Jara, R. M., Couble, A., Emery, X., Magri, E. J., & Ortiz, J. M. (2006). Block size selection and its impact on open-pit design and mine planning. Journal of The South African Institute of Mining and Metallurgy, 106, 205–212. Jenkins, H., & Yakovleva, N. (2006). Corporate social responsibility in the mining industry: Exploring trends in social and environmental disclosure. Journal of Cleaner Production, 14 (3), 271-284. JORC (2012). Australian Code for Reporting of Exploration Results, Mineral Resources and Ore Reserves. Report of the Joint Ore Reserves Committee of the Australian Institute of Mining and Metallurgy, Australian Institute of Geoscientists and Minerals Council of Australia (JORC), p. 9. Kamenetsky, F.M. (2009). Equations of the electromagnetic field in dispersive media, unpublished. Kelly, T.D., & Matos, G.R. (2014). Historical statistics for mineral and material commodities in the United States: U.S. Geological Survey Data Series 140. Retrieved on January 24, 2015 from http://minerals.usgs.gov/minerals/pubs/historical-statistics/ Kumral, M. (2012). Production planning of mines: Optimisation of block sequencing and destination. International journal of mining, reclamation and environment, 26 (2), 93-103. Lessard, J., de Bakker, J., & McHugh, L. (2014). Development of ore sorting and its impact on mineral processing economics. Minerals Engineering, 65, 88-97. London Metal Exchange (2014). LME Copper. Retrieved on January 15, 2014 from http://www.lme.com/metals/non-ferrous/copper/#tab4 Lu, G. Y., & Wong, D. W. (2008). An adaptive inverse-distance weighting spatial interpolation technique. Computers & Geosciences, 34(9), 1044-1055. 113  Luo, Z. Q., Xiao-ming, L., Jia-hong, S., Ya-bin, W., & Wang-ping, L. (2007). Deposit 3D modeling and application. Journal of Central South University of Technology, 14, 225-229. Lyman, G.J. (2011). In situ and particulate material heterogeneity. Proceedings of 5th World Conference on Sampling and Blending (WCSB5), Santiago, Chile, 3-21.  Manouchehri, H.R. (2003). Sorting: possibilities, limitations and future. In: Proc. of Conference in Mineral Technology, Lulea, 1–17. McClure, P. (2005). Correlation statistics: review of the basics and some common pitfalls. Journal of Hand Therapy, 18(3), 378-380.  McCullough, W.E., Bhappu, R, B., & Hightower, J.D. (1999). Copper ore preconcentration by heavy media separation for reduced capital and operating costs, Proceedings Copper 99- Cobre 99, 13-27.  Mesina, M. B., De Jong, T. P. R., & Dalmijn, W. L. (2007). Automatic sorting of scrap metals with a combined electromagnetic and dual energy X-ray transmission sensor. International Journal of Mineral Processing, 82(4), 222-232.  MineSense (2015). MineSense Technologies Limited. Retrieved on January 20, 2015 from http://minesense.com/ Mudd, G. (2007). The sustainability of mining in Australia: key production trends and their environmental implications for the future. Research report, October 2007. Retrieved on June 13, 2014 from http://nowandfutures.com/d2/SustMining-Aust-aReport-Master.pdf Murphy, B., van Zyl, J., & Domingo, G. (2012). Underground Preconcentration by Ore Sorting and Coarse Gravity Separation. In Narrow Vein Mining Conference, 26-27. Norgate, T., & Jahanshahi, S. (2010). Low grade ores–Smelt, leach or concentrate? Minerals Engineering, 23(2), 65-73. Padilla, R., Titley, S.R., & Pimentel, F. (2001). Geology of the Escondida porphyry copper deposit, Antofagasta region, Chile: Economic Geology, 96, 307–324. Perez, C. A., Estévez, P. A., Vera, P. A., Castillo, L. E., Aravena, C. M., Schulz, D. A., & Medina, L. E. (2011). Ore grade estimation by feature selection and voting using boundary detection in digital image analysis. International Journal of Mineral Processing, 101(1), 28-36. Prion, S., & Haerling, K.A. (2014). Making Sense of Methods and Measurement: Pearson Product- Moment Correlation Coefficient. Clinical Simulation in Nursing, 10, 587-588. 114  Richards, G. (2011). Sorting out low-grade nickel ores. Materials World Magazine. Retrieved on November 13, 2014 from http://www.iom3.org/news/sorting-out-low-grade-nickel-ores Romero, B., Kojima, S., Wong, C., Barra, F., Véliz, W., & Ruiz, J. (2011). Molybdenite Mineralization and Re‐Os Geochronology of the Escondida and Escondida Norte Porphyry Deposits, Northern Chile. Resource geology, 61(1), 91-100. Salter, J.D., & Wyatt, N. P. (1991). Sorting in the Minerals Industry: Past, Present and Future, 4(d), 779–796. Schodde, R.C. (2010). Setting the Stage: Global Context for Mining and Extractive Industries. First International Mining Conference, Phnom Penh, Cambodia.  Shaw, R. W., & Lavin, B. (2012). Sorting mined material. U.S. Patent No. 8,240,480. Washington, DC: U.S. Patent and Trademark Office. Sinclair, A.J. & Blackwell, G.H. (2002). Mineral Inventory: An Overview. Applied Mineral Inventory Estimation. Cambridge: Cambridge University Press, 1-30.  Singh, V., & Rao, S. M. (2005). Application of image processing and radial basis neural network techniques for ore sorting and ore classification. Minerals Engineering, 18(15), 1412-1420. Snowden, D. V. (2001). Practical interpretation of mineral resource and ore reserve classification guidelines. Mineral resource and ore reserve estimation–the AusIMM guide to good practice, Melbourne, AusIMM, 643-652. Szmigiel, P. (2005). Discovery of an optimum selective mining unit for a high quality selective deposit. PhD thesis. University of Missouri-Rolla. Thakurdin, K. (2010). Geometallurgy is a rapidly growing discipline. Retrieved on January 21, 2014, from http://www.miningweekly.com/print-version/geometallurgy-is-a-rapidly-growing-discipline-2010-06-11 Tollinsky, N. (2011). CEMI raises interest in ore pre-concentration. Sudbury mining solutions journal. Retrieved on October 15, 2014 from http://www.sudburyminingsolutions.com/cemi-raises-interest-in-ore-pre-concentration.html Tong, Y. (2012). Technical amenability study of laboratory-scale sensor-based ore sorting on a Mississippi Valley type lead-zinc ore. Master’s thesis, The University of British Columbia, Canada. Udoudo, O. B. (2010). Modelling the efficiency of an automated sensor-based sorter. Doctoral dissertation, The University of Exeter. 115  Van Eldert, J. (2011). Stochastic Open Pit Design with a Network Flow Algorithm: Application at Escondida Norte, Chile. PhD thesis, TU Delft, Delft University of Technology. Vatcha, M.T., Cochrane L.B., & Rousell D.H. (2000). Pre-concentration by magnetic sorting of Ni–Cu ore at Whistle mine, Sudbury, Canada, Mineral Processing and Extractive Metallurgy, 109 (3), 156–160. Von Ketelhodt, L. (2009). Viability of optical sorting of gold waste rock dumps. In World Gold Conference, 271-277. Von Ketelhodt, L., & Bergmann, C. (2010). Dual energy X-ray transmission sorting of coal. Journal of the South African Institute of Mining & Metallurgy, 110 (7), 371. Walker, S. (2014). Ore sorters and analyzers: the technology develops. Engineering and mining journal. Retrieved on November 16, 2014 from http://www.e-mj.com/features/3897#.VNuwPPnF_wk Weatherwax, T.W. (2007). Integrated mining and preconcentration systems for nickel sulfide ores. Master’s thesis, The University of British Columbia, Canada. Wellmer, F. W., Dalheimer, M., & Wagner, M. (2008). First estimates of grade and tonnages and potential grade and tonnages. Economic evaluations in exploration, (2nd Ed.). Berlin: Springer, 21-40. Weltje, G. J., & Tjallingii, R. (2008). Calibration of XRF core scanners for quantitative geochemical logging of sediment cores: Theory and application. Earth and Planetary Science Letters, 274(3-4), 423–438.  Wotruba, H., & Harbeck, H. (2010). Sensor-Based Sorting. In Ullamann’s Encyclopedia of Industrial Chemistry (32nd Ed.). Yamamoto, J. K. (1999). Quantification of Uncertainty in Ore-Reserve Estimation : Applications to Chapada Copper Deposit, State of Goias, Brazil, 8(2), 153–163.      116  Appendices Appendix A XRF Results A.1 SPPP07 Table 26: XRF results of SPPP07 zone Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 1 0.282 101 0.350 201 0.146 301 0.076 401 0.324 2 0.662 102 0.353 202 0.267 302 0.247 402 0.110 3 0.116 103 0.288 203 0.113 303 0.311 403 0.673 4 0.474 104 0.154 204 0.574 304 0.172 404 0.154 5 0.391 105 0.523 205 0.180 305 0.328 405 0.483 6 0.167 106 0.176 206 0.167 306 0.227 406 0.572 7 0.239 107 0.136 207 0.115 307 0.163 407 0.390 8 0.494 108 0.305 208 0.327 308 0.196 408 0.327 9 0.269 109 0.411 209 0.278 309 0.215 409 0.181 10 0.394 110 0.206 210 0.255 310 1.456 410 0.228 11 0.296 111 0.208 211 0.372 311 0.221 411 0.083 12 0.132 112 0.096 212 0.174 312 0.609 412 0.149 13 0.717 113 0.176 213 0.109 313 0.112 413 0.195 14 0.222 114 0.096 214 0.244 314 0.266 414 0.248 15 0.078 115 0.456 215 0.211 315 0.427 415 0.187 16 0.591 116 0.280 216 0.307 316 0.250 416 0.084 17 0.274 117 0.242 217 0.131 317 0.146 417 0.077 18 0.450 118 0.415 218 0.195 318 0.278 418 0.572 19 0.134 119 0.263 219 1.956 319 0.396 419 0.482 20 0.771 120 1.320 220 0.223 320 0.117 420 0.176 21 0.371 121 1.012 221 0.082 321 0.320 421 0.236 22 1.073 122 0.070 222 0.346 322 0.810 422 0.189 23 0.178 123 0.132 223 0.225 323 0.254 423 0.898 24 0.301 124 0.230 224 0.345 324 0.660 424 0.320 25 0.519 125 1.706 225 0.356 325 0.271 425 0.111 26 0.516 126 0.339 226 0.326 326 0.224 426 0.182 27 0.236 127 0.398 227 0.127 327 0.482 427 0.256 28 0.114 128 0.247 228 0.347 328 0.510 428 0.104 29 0.168 129 0.201 229 0.493 329 0.363 429 0.224 30 0.108 130 0.748 230 0.162 330 0.911 430 0.210 117  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 31 0.159 131 0.317 231 0.277 331 0.167 431 0.107 32 0.130 132 0.520 232 0.177 332 0.400 432 0.284 33 0.109 133 0.185 233 0.282 333 0.194 433 0.125 34 0.123 134 0.461 234 0.232 334 0.432 434 0.162 35 0.254 135 0.237 235 0.052 335 0.164 435 0.411 36 0.507 136 0.099 236 0.111 336 0.282 436 0.495 37 0.591 137 0.167 237 0.389 337 0.375 437 0.147 38 0.135 138 0.293 238 0.128 338 0.081 438 0.538 39 0.254 139 0.271 239 0.172 339 0.109 439 0.048 40 0.126 140 0.123 240 0.360 340 0.298 440 0.107 41 0.398 141 0.116 241 0.441 341 0.217 441 0.135 42 0.133 142 0.092 242 0.155 342 0.223 442 0.105 43 0.580 143 0.094 243 0.299 343 0.226 443 0.297 44 0.271 144 0.850 244 0.541 344 0.171 444 0.101 45 0.254 145 0.148 245 0.115 345 0.157 445 0.236 46 0.085 146 0.172 246 1.030 346 0.223 446 0.233 47 0.238 147 1.057 247 0.106 347 0.115 447 0.213 48 0.214 148 0.282 248 0.129 348 0.128 448 0.173 49 0.228 149 0.651 249 0.064 349 0.287 449 0.166 50 0.123 150 0.192 250 0.130 350 0.214 450 0.211 51 0.129 151 0.132 251 0.162 351 0.095 451 0.236 52 0.117 152 0.279 252 0.417 352 0.172 452 0.394 53 0.317 153 1.404 253 0.127 353 0.174 453 0.341 54 0.122 154 0.361 254 0.520 354 0.131 454 0.093 55 0.224 155 0.180 255 0.126 355 0.147 455 0.355 56 0.158 156 0.182 256 0.826 356 0.198 456 0.033 57 0.294 157 0.419 257 0.166 357 0.199 457 0.723 58 0.062 158 0.181 258 0.764 358 0.932 458 0.170 59 0.292 159 0.345 259 0.194 359 0.696 459 0.179 60 0.099 160 0.190 260 0.224 360 0.223 460 0.260 61 0.140 161 0.369 261 0.208 361 0.137 461 0.285 62 0.175 162 0.238 262 0.231 362 0.928 462 0.419 63 0.188 163 0.157 263 0.580 363 0.081 463 0.435 64 0.276 164 0.429 264 0.230 364 0.145 464 0.171 65 0.389 165 0.716 265 0.186 365 0.084 465 0.286 66 0.167 166 0.061 266 0.414 366 0.341 466 0.482 67 0.226 167 0.083 267 1.115 367 0.106 467 0.465 68 0.202 168 0.198 268 0.305 368 1.170 468 0.182 118  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 69 0.241 169 0.088 269 0.324 369 0.073 469 0.434 70 0.072 170 0.194 270 0.176 370 0.424 470 0.117 71 0.310 171 0.175 271 0.169 371 0.096 471 0.142 72 0.249 172 0.136 272 0.378 372 0.565 472 0.422 73 0.222 173 0.094 273 0.090 373 1.051 473 0.355 74 0.141 174 0.269 274 0.445 374 0.746 474 0.077 75 0.064 175 0.162 275 0.052 375 0.187 475 0.112 76 0.086 176 0.424 276 0.367 376 0.385 476 0.062 77 0.173 177 0.123 277 0.313 377 0.131 477 0.413 78 0.273 178 0.360 278 0.240 378 0.312 478 0.253 79 0.712 179 0.389 279 0.368 379 0.634 479 0.330 80 0.205 180 0.112 280 0.135 380 0.187 480 0.103 81 0.476 181 0.339 281 0.221 381 0.087 481 0.070 82 0.269 182 0.150 282 0.363 382 1.027 482 0.108 83 0.571 183 0.262 283 0.105 383 0.095 483 0.415 84 0.127 184 0.339 284 0.118 384 0.322 484 0.053 85 0.354 185 0.299 285 0.238 385 0.091 485 0.147 86 0.080 186 0.410 286 0.057 386 0.102 486 0.240 87 0.131 187 0.387 287 0.210 387 0.188 487 0.248 88 0.155 188 0.220 288 0.249 388 0.301 488 0.326 89 0.198 189 0.307 289 0.434 389 0.062 489 0.328 90 0.192 190 0.181 290 0.806 390 0.156 490 0.252 91 0.105 191 0.060 291 0.087 391 1.671 491 0.121 92 0.255 192 0.400 292 0.208 392 0.256 492 0.126 93 0.437 193 1.223 293 0.154 393 0.204 493 0.088 94 0.251 194 0.077 294 0.228 394 0.147 494 0.284 95 0.127 195 0.227 295 0.183 395 0.175 495 0.472 96 0.295 196 0.158 296 0.158 396 0.099 496 0.513 97 0.051 197 0.228 297 0.175 397 0.285 497 0.209 98 0.419 198 0.088 298 0.126 398 0.301 498 0.635 99 0.382 199 0.135 299 0.211 399 0.377 499 0.165 100 0.831 200 0.153 300 0.520 400 0.886 500 0.323   119  A.2 SPPP11 Table 27: XRF results of SPPP11 zone Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 1 0.2312 101 0.3674 201 0.0536 301 0.0368 401 0.1594 2 0.3149 102 0.4373 202 0.2428 302 0.031 402 0.1648 3 0.229 103 0.3621 203 0.0519 303 0.1008 403 0.1339 4 0.5221 104 0.188 204 0.114 304 0.2405 404 0.1779 5 0.2504 105 0.336 205 0.0974 305 0.0933 405 0.2518 6 0.3483 106 0.1214 206 0.1519 306 2.1734 406 0.0721 7 0.088 107 0.2018 207 0.1191 307 0.2426 407 0.2703 8 0.1639 108 0.4495 208 0.1225 308 0.1658 408 0.051 9 0.0572 109 0.2312 209 0.5397 309 0.3832 409 0.3524 10 1.0025 110 0.0292 210 0.1226 310 0.096 410 0.0609 11 0.1729 111 0.2754 211 0.5765 311 0.6133 411 0.1502 12 0.0988 112 0.1325 212 0.3669 312 0.3428 412 0.1875 13 0.4097 113 0.0841 213 0.2849 313 0.0641 413 0.2328 14 0.4937 114 0.0706 214 0.2515 314 0.2075 414 0.0605 15 0.4085 115 0.2505 215 0.208 315 0.0749 415 0.6452 16 0.56 116 0.2443 216 0.3234 316 0.2161 416 0.3196 17 0.8558 117 0.0716 217 0.3067 317 0.2729 417 0.7171 18 0.027 118 0.095 218 0.1621 318 0.793 418 0.1627 19 0.0741 119 0.1109 219 0.0918 319 0.3365 419 0.0491 20 0.057 120 0.043 220 0.3315 320 0.1518 420 0.0555 21 0.2115 121 0.445 221 0.3084 321 0.1223 421 0.2546 22 0.7655 122 0.8745 222 0.1514 322 0.1284 422 0.1334 23 0.1334 123 0.3806 223 0.2292 323 0.4282 423 0.071 24 0.1748 124 0.513 224 0.1884 324 0.6017 424 0.2006 25 0.1622 125 0.1332 225 0.0571 325 0.0377 425 0.1401 26 0.1651 126 0.1865 226 0.1512 326 0.2306 426 0.3415 27 0.3151 127 0.0277 227 0.0239 327 0.1551 427 0.0331 28 0.4677 128 0.2801 228 0.1334 328 0.1802 428 0.1616 29 0.0233 129 0.1736 229 0.1587 329 0.0648 429 0.3092 30 0.2798 130 0.1165 230 0.2909 330 0.3193 430 0.1155 31 0.1008 131 0.3709 231 0.5737 331 0.1566 431 0.1147 32 0.2484 132 0.205 232 0.2729 332 0.3946 432 0.0617 33 0.0812 133 0.4614 233 0.2384 333 0.2268 433 0.3958 34 0.0635 134 0.3153 234 0.2163 334 0.1289 434 0.161 35 0.1683 135 0.0501 235 0.2403 335 1.7844 435 0.302 120  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 36 0.0966 136 0.2578 236 0.2134 336 1.0659 436 0.3255 37 0.2237 137 0.1187 237 0.1576 337 0.2028 437 0.0738 38 0.1453 138 0.12 238 1.0141 338 0.0219 438 0.2488 39 0.3677 139 0.0769 239 0.3664 339 0.4299 439 0.0603 40 0.1485 140 0.0606 240 0.0357 340 0.5575 440 0.1824 41 0.1213 141 0.0477 241 0.0751 341 0.6529 441 0.1529 42 2.1062 142 0.0849 242 0.3536 342 1.6193 442 0.0715 43 0.8419 143 0.2489 243 0.1611 343 0.1151 443 0.4646 44 0.245 144 0.4702 244 0.1304 344 0.4633 444 0.1725 45 0.2228 145 0.5567 245 0.0448 345 0.2502 445 0.8352 46 0.0766 146 0.0696 246 0.0949 346 0.2915 446 0.0559 47 0.1636 147 0.4656 247 0.4502 347 0.1639 447 0.2234 48 0.1098 148 0.1736 248 0.0569 348 0.0728 448 0.0329 49 0.1383 149 0.1467 249 0.9818 349 2.3964 449 0.104 50 0.1418 150 0.4219 250 1.3261 350 0.0551 450 0.4879 51 0.4161 151 0.1827 251 0.4503 351 1.673 451 0.8012 52 0.0464 152 0.0304 252 0.0868 352 0.3205 452 0.1962 53 0.2734 153 0.4549 253 0.2396 353 0.2141 453 0.1305 54 0.1504 154 0.1733 254 1.0857 354 0.1962 454 0.1529 55 0.0924 155 0.3841 255 1.8368 355 0.1049 455 0.2947 56 0.1476 156 0.3013 256 0.4929 356 0.129 456 1.1888 57 0.1692 157 0.1025 257 0.1563 357 0.0315 457 0.2111 58 0.4329 158 0.1043 258 0.0412 358 1.3683 458 0.6409 59 0.3748 159 0.0583 259 0.0433 359 0.2944 459 1.8338 60 0.0811 160 0.2763 260 0.0502 360 0.1544 460 0.0831 61 0.3891 161 0.0848 261 0.1352 361 0.1678 461 0.2766 62 0.0621 162 0.382 262 0.7686 362 0.0598 462 0.2335 63 0.1773 163 0.1678 263 0.2195 363 0.4785 463 0.4333 64 0.1368 164 0.6312 264 0.1372 364 0.5041 464 0.1706 65 0.2767 165 0.1758 265 0.7457 365 0.5493 465 0.177 66 0.1467 166 0.0752 266 0.2854 366 0.2321 466 0.1901 67 0.0975 167 0.0974 267 0.6578 367 0.1813 467 0.3506 68 0.2749 168 0.0685 268 0.1829 368 0.4298 468 0.168 69 0.0962 169 0.0795 269 0.1895 369 0.3807 469 0.0636 70 0.1493 170 0.5038 270 0.1769 370 0.4269 470 0.0497 71 0.2013 171 0.1756 271 0.0198 371 0.4702 471 0.0901 72 0.2891 172 0.2088 272 0.0642 372 0.1042 472 0.113 73 0.0583 173 0.1874 273 0.0987 373 0.154 473 0.5273 121  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 74 0.4617 174 0.1411 274 0.1709 374 0.1853 474 0.0272 75 0.355 175 0.1402 275 0.0827 375 0.6393 475 0.8656 76 0.1098 176 0.0487 276 1.2963 376 0.0659 476 0.2526 77 0.0793 177 0.2135 277 0.1621 377 1.5833 477 0.0653 78 0.3428 178 0.5896 278 0.0567 378 0.1058 478 0.4469 79 0.0974 179 0.4442 279 0.086 379 0.1128 479 0.0867 80 0.1456 180 0.1097 280 0.1969 380 0.7261 480 0.2815 81 0.1172 181 0.1104 281 0.1461 381 0.3222 481 0.4359 82 0.1287 182 0.3978 282 0.298 382 0.2016 482 0.1026 83 0.0477 183 0.0511 283 0.1383 383 0.9131 483 0.3031 84 0.0239 184 0.084 284 0.2007 384 0.1752 484 0.1801 85 0.0341 185 0.142 285 0.1696 385 0.1939 485 0.3356 86 0.0121 186 0.2659 286 0.3695 386 0.0869 486 0.2836 87 0.052 187 0.295 287 0.1472 387 0.0933 487 0.1508 88 0.0648 188 0.1469 288 0.2533 388 0.0706 488 0.231 89 0.0771 189 0.4335 289 0.0385 389 0.054 489 0.3424 90 0.1703 190 0.2757 290 0.2089 390 0.0565 490 0.0744 91 0.1234 191 0.7422 291 0.6005 391 0.2856 491 0.5456 92 0.1529 192 0.1897 292 0.3115 392 0.3461 492 0.1021 93 0.0322 193 0.1243 293 0.1868 393 0.6015 493 0.0404 94 0.154 194 0.1084 294 0.0743 394 0.3945 494 0.1388 95 0.1659 195 0.5311 295 0.18 395 0.3768 495 0.0493 96 0.7233 196 0.1219 296 0.7725 396 0.0787 496 0.243 97 0.18 197 0.1233 297 0.1143 397 0.6061 497 0.6299 98 0.1791 198 0.4167 298 0.0903 398 0.7779 498 0.2046 99 0.0935 199 0.0474 299 0.0839 399 0.2042 499 0.7421 100 0.5172 200 0.5743 300 0.2704 400 0.3096 500 0.7421 A.3 Oxido Table 28: XRF results of Oxdio zone Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 1 0.1034 101 0.1415 201 0.1081 301 0.2502 401 0.9371 2 0.1446 102 0.1717 202 0.1041 302 0.2372 402 0.1995 3 0.0747 103 0.1612 203 0.2475 303 0.0913 403 0.1477 4 0.065 104 0.7257 204 0.1271 304 0.5597 404 0.0563 5 0.1203 105 0.0935 205 0.5151 305 0.6048 405 0.1076 122  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 6 0.4106 106 0.0578 206 0.1058 306 1.922 406 0.0608 7 0.0161 107 0.2708 207 0.2575 307 0.3236 407 1.6411 8 0.2698 108 0.2392 208 0.0924 308 0.1852 408 0.8199 9 0.8521 109 1.1839 209 0.5747 309 0.2055 409 0.0408 10 0.0638 110 0.0589 210 0.4462 310 0.0698 410 0.2848 11 0.2619 111 0.3143 211 0.1967 311 0.3825 411 0.3784 12 0.1198 112 0.0893 212 0.1359 312 0.0783 412 1.2415 13 0.0863 113 0.1295 213 0.4112 313 0.3031 413 0.4222 14 0.8596 114 0.1174 214 0.2192 314 0.1983 414 0.1127 15 0.602 115 0.48 215 0.1956 315 0.7992 415 0.2706 16 0.0258 116 0.3425 216 0.3165 316 0.9236 416 1.6701 17 0.6914 117 0.3909 217 0.035 317 0.1032 417 0.1071 18 0.221 118 0.3295 218 0.1318 318 0.6851 418 0.1633 19 0.4552 119 0.0825 219 0.1683 319 0.1286 419 0.2913 20 0.2303 120 0.0733 220 0.0695 320 0.0948 420 0.0956 21 0.1233 121 0.0682 221 0.0955 321 0.0791 421 0.1913 22 0.044 122 0.4242 222 0.0466 322 0.1263 422 0.1589 23 0.0474 123 0.0747 223 0.057 323 0.0893 423 0.2829 24 0.0639 124 0.5444 224 0.1639 324 0.371 424 0.1273 25 0.5767 125 0.2411 225 0.0978 325 0.1085 425 0.1497 26 0.3133 126 0.2639 226 0.1474 326 0.2767 426 0.1058 27 0.2403 127 0.0639 227 0.0929 327 0.0663 427 0.0562 28 0.0791 128 0.1924 228 0.0881 328 0.1363 428 0.7244 29 0.0826 129 0.0898 229 0.0719 329 0.155 429 0.1397 30 0.1253 130 0.8742 230 0.2557 330 0.111 430 0.2851 31 0.0544 131 0.1054 231 0.1868 331 0.1011 431 0.1885 32 0.1533 132 0.0756 232 0.4855 332 1.4224 432 0.0801 33 0.0967 133 0.0769 233 0.2069 333 1.6437 433 0.5481 34 0.5689 134 0.1459 234 0.4964 334 0.0577 434 0.432 35 0.0835 135 0.6442 235 0.3644 335 0.1426 435 0.0464 36 0.0567 136 0.0751 236 0.8455 336 0.4256 436 0.1222 37 0.1382 137 0.3779 237 1.5424 337 0.6345 437 0.117 38 0.1471 138 0.1754 238 0.3342 338 0.0814 438 0.1179 39 0.0624 139 0.2187 239 1.0776 339 0.5521 439 0.1366 40 0.0605 140 0.1988 240 0.2243 340 0.3539 440 0.2159 41 0.0671 141 0.1197 241 0.1395 341 0.162 441 0.4593 42 0.1322 142 3.7396 242 0.0714 342 0.1813 442 0.1062 43 1.8802 143 0.5127 243 2.065 343 1.2338 443 0.0892 123  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 44 0.097 144 0.4695 244 0.1976 344 0.0984 444 1.2149 45 0.884 145 0.4282 245 0.1128 345 0.1261 445 0.1937 46 0.0335 146 0.1905 246 0.1927 346 0.2232 446 0.0354 47 0.1222 147 0.1354 247 0.148 347 0.0741 447 0.0737 48 0.1146 148 0.2896 248 0.1244 348 0.07 448 0.2013 49 0.1069 149 0.0602 249 0.2895 349 0.1079 449 0.2502 50 0.0209 150 0.0769 250 0.1316 350 0.0679 450 0.2446 51 0.3222 151 0.1532 251 0.2228 351 0.1466 451 0.6829 52 0.3046 152 0.1171 252 0.5613 352 0.2128 452 0.1227 53 0.2964 153 0.0897 253 0.2393 353 0.7991 453 0.3946 54 0.0825 154 0.0593 254 0.5553 354 0.2722 454 0.0819 55 0.0784 155 0.385 255 0.3683 355 0.056 455 1.2372 56 0.369 156 0.0437 256 0.9377 356 0.225 456 0.2348 57 0.0826 157 0.0648 257 1.2477 357 0.3183 457 0.1559 58 0.3804 158 0.8593 258 0.069 358 0.3809 458 0.4836 59 0.4787 159 0.0955 259 0.1272 359 0.0364 459 0.1612 60 0.1993 160 0.0989 260 0.0856 360 0.636 460 0.169 61 0.0334 161 0.0304 261 0.4199 361 0.1163 461 0.1181 62 0.0427 162 1.0272 262 0.086 362 0.1623 462 0.1722 63 0.1103 163 0.1064 263 0.2161 363 0.2361 463 0.0327 64 0.0484 164 0.3701 264 0.1089 364 0.2198 464 0.2098 65 0.2924 165 0.1617 265 0.3242 365 0.1335 465 0.1184 66 2.2026 166 0.0752 266 0.0754 366 0.1163 466 0.3626 67 0.3325 167 0.182 267 0.1826 367 0.1959 467 0.0589 68 0.674 168 0.2647 268 0.5328 368 0.1403 468 0.2639 69 0.1436 169 0.025 269 0.1019 369 0.1978 469 0.0432 70 0.1711 170 0.0268 270 0.1485 370 0.2047 470 0.7009 71 0.746 171 0.0157 271 0.9708 371 0.0946 471 1.0348 72 0.185 172 0.3447 272 0.2125 372 0.1159 472 0.103 73 0.2271 173 0.2943 273 0.1403 373 0.1325 473 0.049 74 0.2534 174 0.0238 274 0.2482 374 0.9284 474 0.0696 75 0.1571 175 0.1341 275 0.1259 375 0.4337 475 0.7509 76 0.1404 176 0.1255 276 0.0871 376 0.2782 476 0.158 77 0.3044 177 0.3818 277 0.141 377 0.2552 477 0.0742 78 0.1543 178 0.083 278 0.06 378 0.4925 478 0.0836 79 0.5442 179 0.3152 279 0.0575 379 0.311 479 0.1515 80 0.2208 180 0.3083 280 0.0466 380 0.2493 480 0.0707 81 0.4551 181 0.1527 281 0.0591 381 0.1191 481 0.1083 124  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 82 0.0921 182 0.5983 282 0.4241 382 0.4171 482 0.5387 83 0.1522 183 0.4528 283 0.1793 383 0.1395 483 0.1491 84 0.9343 184 0.1935 284 0.1384 384 0.0328 484 0.0624 85 0.1611 185 0.5187 285 0.053 385 0.1022 485 0.1752 86 0.1965 186 0.0592 286 0.1324 386 0.2926 486 0.1186 87 0.8723 187 0.2224 287 0.4138 387 0.256 487 0.0945 88 0.516 188 0.2208 288 0.2318 388 0.2075 488 0.3657 89 0.1443 189 0.122 289 0.0522 389 0.1449 489 0.4104 90 0.0546 190 0.3029 290 0.2512 390 0.1652 490 0.2528 91 0.1635 191 0.3052 291 0.1218 391 0.0785 491 0.0966 92 0.0752 192 0.073 292 0.2591 392 0.5295 492 0.1345 93 1.4107 193 0.1218 293 0.1016 393 0.428 493 0.073 94 0.0991 194 0.3102 294 0.1222 394 0.2155 494 0.0432 95 0.1144 195 1.3602 295 0.1042 395 0.0755 495 0.2387 96 0.1101 196 0.1402 296 0.4519 396 0.1367 496 0.2217 97 0.4121 197 0.2222 297 0.0563 397 0.3561 497 0.0765 98 0.0778 198 0.6657 298 0.1504 398 1.2038 498 0.1084 99 0.3667 199 0.0544 299 0.146 399 0.1807 499 0.291 100 0.9812 200 0.4717 300 0.0739 400 0.741 500 0.1002 A.4 Mixto Table 29: XRF results of Mixto zone Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 1 14.3104 101 0.1019 201 1.0936 301 0.1299 401 0.0756 2 0.095 102 0.9584 202 0.098 302 0.0634 402 0.8922 3 0.2088 103 0.4398 203 0.0991 303 0.1488 403 1.1379 4 0.0633 104 0.1476 204 0.0747 304 0.0464 404 0.0971 5 0.0537 105 0.3597 205 0.1628 305 0.0884 405 0.0782 6 0.1714 106 0.122 206 0.246 306 0.1026 406 0.2255 7 0.8436 107 1.4623 207 0.4904 307 0.1381 407 0.0947 8 0.1018 108 0.7779 208 0.1055 308 0.0951 408 0.1038 9 0.0898 109 0.2055 209 0.1253 309 0.0858 409 17.4466 10 0.0637 110 0.084 210 0.1484 310 0.1151 410 0.2136 11 0.085 111 0.1137 211 0.0981 311 0.1279 411 0.187 12 0.0795 112 0.0777 212 0.0759 312 0.0952 412 0.2135 13 0.0757 113 2.835 213 0.4316 313 0.0911 413 0.0796 125  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 14 0.1124 114 0.0934 214 1.4941 314 0.1603 414 0.151 15 0.3441 115 0.0875 215 0.0644 315 0.1162 415 0.1191 16 0.1246 116 0.1098 216 0.1925 316 0.3404 416 2.6418 17 0.4226 117 0.1385 217 0.4475 317 0.1609 417 0.1335 18 0.1121 118 0.1584 218 0.1095 318 0.1571 418 1.9143 19 0.1288 119 0.0665 219 0.109 319 0.1004 419 0.1887 20 0.1431 120 8.6345 220 0.1599 320 0.1589 420 0.0532 21 0.0948 121 0.0778 221 0.1389 321 0.1405 421 0.1079 22 0.096 122 2.3331 222 0.726 322 0.1114 422 0.0977 23 0.1259 123 1.4485 223 0.1345 323 0.1163 423 2.5298 24 0.0685 124 0.0956 224 0.1552 324 0.1394 424 0.0836 25 2.6115 125 0.091 225 0.0863 325 7.1734 425 0.0664 26 0.1092 126 0.086 226 0.0731 326 1.3544 426 0.0851 27 0.0825 127 0.0347 227 0.0738 327 0.2102 427 0.0973 28 0.2116 128 0.0976 228 0.1132 328 0.253 428 0.071 29 0.2131 129 0.0484 229 0.1087 329 0.1093 429 1.1775 30 0.0399 130 1.2244 230 0.1588 330 0.0923 430 0.1372 31 0.4316 131 0.1158 231 0.9477 331 0.0963 431 0.1034 32 0.1475 132 1.75 232 0.1137 332 0.1052 432 0.1255 33 0.0672 133 0.1094 233 0.6201 333 0.1842 433 0.1157 34 0.1858 134 0.0547 234 1.4136 334 1.0506 434 0.0847 35 0.196 135 0.1708 235 0.3126 335 0.2178 435 0.104 36 0.072 136 0.2622 236 1.6633 336 0.1237 436 0.1098 37 0.1264 137 0.1398 237 1.1988 337 0.1066 437 0.1118 38 0.0815 138 0.1617 238 0.1661 338 0.5031 438 0.828 39 0.631 139 0.1454 239 0.1685 339 6.3722 439 3.8698 40 0.0663 140 0.097 240 0.1075 340 0.1189 440 0.1477 41 0.1467 141 0.0668 241 0.0818 341 0.2093 441 0.1279 42 0.2966 142 0.133 242 0.5016 342 2.3198 442 0.5324 43 2.4884 143 1.5635 243 0.1065 343 0.1584 443 0.1185 44 0.1036 144 0.3021 244 0.1303 344 0.9086 444 0.0699 45 0.0827 145 0.1089 245 0.1274 345 0.0822 445 0.0725 46 0.0738 146 0.0629 246 0.1474 346 1.4454 446 0.2007 47 0.188 147 0.064 247 0.1024 347 0.0827 447 0.0933 48 4.059 148 0.093 248 0.1004 348 0.147 448 0.0646 49 0.164 149 0.0904 249 0.218 349 0.1262 449 1.9148 50 0.0809 150 0.0787 250 0.1043 350 0.0881 450 0.1585 51 0.2744 151 0.0633 251 0.0876 351 0.1139 451 0.5564 126  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 52 1.724 152 0.9906 252 1.5615 352 0.4973 452 0.2216 53 0.1081 153 0.133 253 0.1402 353 0.1441 453 0.0701 54 0.0858 154 0.1177 254 0.131 354 0.113 454 0.0876 55 0.0783 155 0.0938 255 0.1895 355 0.3686 455 1.6653 56 0.0772 156 1.2628 256 0.1203 356 1.7895 456 2.4811 57 0.0895 157 0.1018 257 0.1143 357 0.1085 457 0.1825 58 0.1657 158 0.1021 258 0.1032 358 0.149 458 6.1032 59 0.1021 159 1.1167 259 0.1069 359 0.1195 459 0.2592 60 0.115 160 0.1087 260 0.1148 360 0.3839 460 0.2569 61 0.0664 161 0.1634 261 0.4096 361 0.1117 461 1.7021 62 0.3824 162 0.21 262 0.1044 362 4.9221 462 0.2036 63 0.2487 163 0.1689 263 0.0665 363 0.2988 463 0.1153 64 0.0745 164 0.0829 264 1.3583 364 0.0931 464 0.108 65 0.1204 165 0.132 265 0.1269 365 0.0864 465 10.6608 66 0.0809 166 0.1146 266 0.9081 366 0.0879 466 0.1685 67 0.0792 167 0.0995 267 0.1224 367 0.0998 467 0.1001 68 0.0965 168 0.0934 268 0.2073 368 0.101 468 0.0953 69 0.1043 169 0.1451 269 0.1513 369 0.1213 469 0.1162 70 0.0336 170 0.1939 270 0.0628 370 0.1319 470 0.0785 71 0.0532 171 0.1477 271 0.1031 371 0.1373 471 0.0975 72 11.959 172 0.1255 272 0.1178 372 0.0665 472 0.0801 73 0.068 173 0.1027 273 0.2063 373 0.1767 473 0.2172 74 0.032 174 0.3102 274 0.0819 374 0.0932 474 5.4828 75 0.0591 175 0.078 275 0.1176 375 0.691 475 0.2181 76 1.1229 176 0.1037 276 0.1885 376 0.083 476 0.0616 77 0.1218 177 0.1043 277 0.1348 377 0.1765 477 0.7404 78 0.2148 178 0.0826 278 0.3383 378 0.0951 478 0.0556 79 0.1159 179 0.1087 279 0.1132 379 5.5303 479 0.1242 80 0.2882 180 0.2432 280 0.1476 380 0.1669 480 1.4818 81 0.0863 181 0.1173 281 0.3275 381 0.085 481 0.058 82 0.1668 182 0.1059 282 0.0972 382 0.0725 482 0.1466 83 6.7065 183 0.0949 283 0.6546 383 0.4323 483 0.2041 84 0.0991 184 3.3986 284 0.2608 384 0.0745 484 1.0571 85 0.1472 185 3.2815 285 0.6915 385 0.0591 485 0.1185 86 0.1005 186 0.1568 286 0.437 386 0.0606 486 0.8728 87 0.2057 187 0.1108 287 1.5614 387 1.1166 487 0.0762 88 11.4844 188 0.091 288 1.6836 388 0.132 488 0.2253 89 0.1237 189 0.1201 289 0.1677 389 0.2913 489 0.9212 127  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 90 0.1287 190 0.0705 290 0.1017 390 0.4951 490 0.1134 91 0.082 191 0.106 291 0.1093 391 0.1025 491 0.3695 92 0.1343 192 0.0794 292 0.1313 392 0.0959 492 0.3661 93 0.0761 193 1.9131 293 0.0988 393 0.1021 493 0.1537 94 0.1134 194 0.0751 294 0.2135 394 1.0735 494 0.1196 95 0.1058 195 0.0681 295 0.1262 395 0.1138 495 0.1107 96 0.0733 196 0.1505 296 0.1017 396 0.1121 496 10.4208 97 0.0929 197 0.1143 297 0.1121 397 0.4432 497 0.1346 98 1.8783 198 0.0756 298 0.1273 398 0.0817 498 0.2195 99 0.1922 199 0.1103 299 0.1318 399 3.0034 499 0.1669 100 0.2414 200 0.2106 300 5.5967 400 0.3676 500 0.0665 A.5 Sulfuro Table 30: XRF results of Sulfuro zone Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 1 0.0262 101 0.4116 201 0.149 301 0.4462 401 0.4153 2 0.6165 102 0.2709 202 0.1107 302 0.2191 402 0.3264 3 0.3292 103 0.5521 203 0.1398 303 2.3415 403 1.095 4 1.1383 104 0.1786 204 0.7604 304 1.4085 404 0.3938 5 0.2772 105 0.2794 205 1.1826 305 0.3012 405 1.5336 6 0.4818 106 0.1862 206 0.3465 306 0.2748 406 0.6806 7 0.453 107 0.2837 207 0.2942 307 0.1236 407 0.1659 8 0.3461 108 0.0939 208 0.2281 308 0.436 408 0.1523 9 0.2205 109 0.3369 209 0.1745 309 0.2425 409 0.2146 10 1.085 110 0.6585 210 0.1595 310 0.0349 410 0.1175 11 0.1489 111 0.4911 211 0.7934 311 0.3031 411 1.6254 12 1.2148 112 0.0763 212 0.1914 312 0.5815 412 0.7137 13 0.7126 113 0.5869 213 1.362 313 0.4911 413 0.1707 14 0.9742 114 0.4086 214 0.4538 314 1.2443 414 0.6812 15 0.3188 115 0.775 215 0.4005 315 0.3897 415 1.9075 16 0.8225 116 0.1416 216 0.1334 316 0.2915 416 0.6333 17 1.3554 117 0.2273 217 2.0165 317 0.3476 417 0.2795 18 0.0521 118 0.4422 218 0.6902 318 0.2916 418 0.477 19 0.5138 119 0.1594 219 0.7691 319 0.0843 419 1.1422 20 0.1653 120 0.0967 220 0.7103 320 0.1495 420 0.3373 21 0.317 121 1.014 221 0.4375 321 0.7164 421 0.5102 128  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 22 0.2969 122 0.1542 222 0.3713 322 0.6657 422 0.3142 23 0.3293 123 0.0562 223 0.3903 323 1.5263 423 0.1927 24 0.8788 124 1.8609 224 0.4806 324 0.8784 424 0.7506 25 0.1975 125 0.1868 225 0.8249 325 1.0567 425 0.6102 26 0.0992 126 0.2653 226 1.4202 326 0.5364 426 0.5683 27 0.171 127 0.4705 227 0.2093 327 1.5982 427 0.0784 28 0.1331 128 1.099 228 0.9024 328 0.1787 428 0.2136 29 0.3815 129 0.5828 229 0.0296 329 0.2491 429 0.1178 30 0.5889 130 0.7482 230 1.0168 330 1.2685 430 1.4817 31 1.3564 131 0.166 231 1.1221 331 0.1835 431 0.5221 32 0.9431 132 1.3299 232 0.5309 332 0.1364 432 0.3597 33 0.6708 133 0.6366 233 0.4015 333 0.0657 433 0.0535 34 0.5997 134 1.2507 234 1.3256 334 0.0733 434 0.4312 35 0.7701 135 2.2591 235 0.5544 335 0.4949 435 0.3715 36 0.6932 136 0.0644 236 0.0405 336 0.5422 436 0.7955 37 0.1401 137 0.9733 237 1.6747 337 0.1541 437 0.0987 38 0.7179 138 0.1511 238 0.2995 338 1.3498 438 0.3004 39 0.9805 139 1.1402 239 0.1732 339 0.2459 439 0.3072 40 0.8768 140 1.1562 240 0.6556 340 0.1144 440 0.2813 41 0.8444 141 0.3545 241 0.2147 341 0.9362 441 0.4026 42 0.6788 142 0.8683 242 0.147 342 0.0811 442 0.0928 43 0.0477 143 2.2446 243 0.545 343 0.4571 443 0.7124 44 0.3004 144 0.0301 244 0.9203 344 0.3946 444 0.3363 45 1.4461 145 0.6833 245 0.5429 345 0.3252 445 0.1689 46 0.4545 146 0.3619 246 0.9144 346 1.1552 446 0.6978 47 0.2416 147 0.2399 247 0.029 347 0.1744 447 0.1817 48 0.8627 148 1.6252 248 0.356 348 0.4703 448 3.0437 49 0.4046 149 1.0538 249 0.1775 349 0.9771 449 0.5507 50 0.0989 150 0.0469 250 0.0952 350 0.111 450 0.2045 51 0.3458 151 2.7891 251 0.0482 351 0.1428 451 0.2095 52 0.1274 152 0.3212 252 0.1642 352 0.28 452 0.8031 53 0.3701 153 3.6263 253 1.5243 353 0.0827 453 0.4794 54 0.634 154 0.3134 254 0.4458 354 0.7095 454 1.0337 55 0.4438 155 0.3205 255 0.6079 355 2.2475 455 0.6108 56 0.2044 156 0.3879 256 0.5589 356 0.1153 456 0.0558 57 0.0488 157 0.0759 257 0.0255 357 0.189 457 0.4247 58 0.587 158 0.762 258 0.24 358 0.6217 458 0.9289 59 0.3426 159 0.3357 259 1.5054 359 0.6709 459 0.6376 129  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 60 0.3403 160 0.1084 260 0.2826 360 1.0317 460 0.0853 61 1.1405 161 0.0519 261 1.2637 361 0.1456 461 0.6191 62 0.6191 162 0.7329 262 0.432 362 0.5949 462 0.6588 63 0.5023 163 0.0479 263 1.1016 363 0.1057 463 1.7202 64 1.3897 164 0.4305 264 0.0631 364 0.4245 464 0.2488 65 0.3431 165 0.5574 265 0.5277 365 0.0189 465 0.1215 66 0.3173 166 0.4685 266 0.6118 366 0.347 466 0.3917 67 0.6311 167 0.8968 267 0.0692 367 0.3318 467 0.5613 68 1.9845 168 0.2008 268 0.1947 368 0.9905 468 0.6252 69 0.1066 169 0.1889 269 0.8977 369 0.1943 469 0.7707 70 1.0192 170 0.6477 270 0.0359 370 0.0725 470 0.8721 71 0.1567 171 0.3469 271 0.9457 371 1.3823 471 0.9673 72 0.0877 172 0.5873 272 0.0111 372 1.4563 472 1.6541 73 0.2359 173 0.9972 273 0.2936 373 0.3469 473 0.1712 74 0.3008 174 0.0727 274 0.0289 374 0.0436 474 0.0641 75 0.1698 175 0.3885 275 1.1056 375 0.4207 475 1.2719 76 0.4141 176 0.4614 276 0.8153 376 0.9763 476 0.7297 77 0.0173 177 1.0382 277 1.5924 377 0.1331 477 1.5095 78 0.5998 178 0.0119 278 0.5163 378 0.3637 478 0.4667 79 1.5392 179 0.3777 279 0.4994 379 0.2875 479 0.4159 80 0.4704 180 0.5166 280 0.2092 380 0.1157 480 0.0575 81 0.048 181 0.3813 281 0.3545 381 0.111 481 0.5485 82 0.0154 182 0.2919 282 0.8637 382 0.0322 482 0.6118 83 0.1398 183 0.8335 283 1.3219 383 0.1852 483 0.9446 84 0.1544 184 0.3465 284 2.4714 384 0.2122 484 0.2974 85 1.5779 185 0.716 285 0.1777 385 0.7419 485 0.446 86 0.0886 186 0.3174 286 0.9254 386 0.3475 486 2.2521 87 0.4293 187 0.3656 287 0.071 387 1.2159 487 0.9297 88 0.1293 188 0.3285 288 0.2039 388 0.2195 488 0.4941 89 0.2436 189 0.38 289 0.6822 389 0.7176 489 0.3756 90 0.4742 190 0.6354 290 0.2305 390 0.3492 490 1.1515 91 1.0744 191 0.4059 291 0.1339 391 0.3574 491 1.2087 92 0.2053 192 0.6372 292 0.0489 392 0.5955 492 0.5077 93 0.9561 193 0.0459 293 0.4383 393 0.8233 493 0.4967 94 0.9885 194 0.176 294 3.4887 394 0.0346 494 1.5318 95 0.2828 195 0.5919 295 0.6302 395 2.612 495 1.3425 96 0.6693 196 0.1627 296 0.4345 396 0.3106 496 0.8363 97 0.0988 197 0.8401 297 0.4113 397 0.2918 497 1.3851 130  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 98 6.3176 198 0.5715 298 0.3714 398 0.5883 498 0.7531 99 0.5224 199 0.1771 299 0.6708 399 0.2115 499 0.5392 100 0.0377 200 0.122 300 0.0264 400 1.2987 500 0.5437 A.6 M3 Table 31: XRF results of M3 zone Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 1 0.2063 101 0.3083 201 0.2557 301 0.1573 401 0.2979 2 0.1019 102 0.3783 202 0.2567 302 0.4925 402 0.5536 3 0.7638 103 0.0747 203 0.4598 303 0.1826 403 0.2066 4 0.2011 104 1.02 204 0.2878 304 0.3493 404 0.1972 5 0.5005 105 0.4711 205 0.3525 305 0.4389 405 0.2794 6 0.4741 106 0.3555 206 0.6248 306 0.0917 406 0.1525 7 0.1609 107 0.1975 207 0.1191 307 0.2181 407 0.1583 8 0.2901 108 0.7469 208 1.0394 308 0.1742 408 0.1261 9 0.106 109 0.0673 209 0.2825 309 0.6535 409 0.5834 10 0.2427 110 0.0689 210 0.9084 310 0.216 410 0.3828 11 1.3538 111 0.4571 211 0.226 311 0.4958 411 0.3362 12 0.5789 112 0.1725 212 0.4174 312 0.1914 412 0.6607 13 0.101 113 0.1705 213 0.9199 313 0.1215 413 0.7709 14 0.5814 114 0.3586 214 0.21 314 0.3465 414 0.4009 15 0.43 115 0.2675 215 0.8352 315 0.4488 415 0.4937 16 0.3444 116 0.4118 216 0.9399 316 0.1928 416 1.1045 17 0.2042 117 0.2384 217 0.4336 317 0.3282 417 0.5066 18 0.3286 118 0.3027 218 0.3491 318 0.2529 418 0.5112 19 0.1445 119 0.3565 219 0.3521 319 0.2831 419 0.2893 20 2.2712 120 0.8256 220 0.3496 320 0.1317 420 0.0883 21 0.364 121 0.2217 221 0.303 321 0.4871 421 0.1683 22 0.5354 122 0.2506 222 2.3391 322 0.2583 422 0.3679 23 0.3264 123 0.1313 223 0.2412 323 0.1216 423 0.098 24 0.5754 124 0.7457 224 0.4742 324 0.1739 424 0.6378 25 0.1875 125 0.1607 225 0.1927 325 0.1499 425 0.4329 26 0.149 126 0.4137 226 0.1377 326 0.9283 426 0.2062 27 2.0165 127 0.256 227 0.4143 327 1.4751 427 0.6555 28 0.5286 128 0.1574 228 0.293 328 0.4039 428 0.0725 131  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 29 0.7931 129 0.1084 229 0.7437 329 0.2973 429 0.6822 30 0.8374 130 0.4145 230 0.3973 330 0.3023 430 0.3428 31 0.1112 131 0.6718 231 0.6641 331 0.4369 431 0.3277 32 0.4402 132 0.5499 232 0.3724 332 0.1886 432 0.1946 33 0.1427 133 0.2343 233 0.7372 333 0.1418 433 0.2217 34 0.2963 134 0.3871 234 0.3214 334 0.2963 434 1.0498 35 0.0867 135 0.3486 235 0.2499 335 0.3585 435 0.8123 36 0.6215 136 0.3725 236 0.1278 336 1.1352 436 0.4002 37 0.2843 137 0.5361 237 0.1137 337 0.3994 437 0.4473 38 0.4582 138 1.0426 238 0.181 338 0.2862 438 0.8112 39 0.1652 139 0.1304 239 0.5323 339 0.3667 439 0.0816 40 0.3517 140 0.8128 240 0.2878 340 0.3085 440 0.3807 41 0.2544 141 0.1083 241 0.2528 341 0.6102 441 0.425 42 0.4742 142 0.3603 242 0.2235 342 0.4951 442 0.1404 43 0.7231 143 0.3288 243 1.5088 343 0.2627 443 0.1469 44 0.4983 144 0.424 244 0.6549 344 0.2799 444 0.5975 45 0.0322 145 0.1605 245 0.2016 345 0.2693 445 0.3895 46 0.2266 146 0.1594 246 0.2311 346 0.453 446 0.4667 47 0.1151 147 0.2998 247 0.4691 347 0.3056 447 0.3096 48 0.1625 148 0.8244 248 0.2899 348 0.7728 448 0.3301 49 0.1601 149 0.2101 249 0.5329 349 0.4124 449 1.0777 50 0.2654 150 0.2319 250 0.2515 350 0.1261 450 0.0826 51 0.7883 151 0.5457 251 0.4123 351 0.4333 451 0.3716 52 0.7155 152 0.673 252 0.4071 352 0.2376 452 0.7866 53 0.2352 153 0.3493 253 0.2869 353 0.1413 453 0.9586 54 1.0415 154 0.6741 254 0.6152 354 0.4994 454 0.5184 55 0.202 155 0.4409 255 0.2147 355 0.362 455 0.1606 56 0.2013 156 0.2646 256 0.2722 356 1.0935 456 0.7866 57 0.6806 157 0.3826 257 0.1804 357 0.7633 457 0.5051 58 0.3733 158 1.0161 258 0.1787 358 0.2067 458 0.7588 59 0.339 159 0.2199 259 0.1012 359 0.3462 459 0.261 60 1.7647 160 0.158 260 0.2144 360 0.2611 460 0.1726 61 0.5852 161 1.0133 261 0.9649 361 0.5266 461 0.1107 62 0.0734 162 0.3988 262 0.4383 362 0.2962 462 0.2116 63 0.3251 163 0.7044 263 0.1119 363 0.6404 463 0.1707 64 0.4615 164 0.3125 264 0.1825 364 0.2689 464 0.768 132  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 65 0.0474 165 0.491 265 0.6501 365 0.2144 465 0.9092 66 0.0472 166 0.3131 266 0.2807 366 0.3272 466 0.7275 67 0.164 167 0.1433 267 1.7525 367 0.3792 467 0.3574 68 0.438 168 0.3594 268 0.4015 368 0.4153 468 0.6475 69 0.4584 169 0.4619 269 0.4077 369 1.0966 469 0.3876 70 0.6041 170 0.3484 270 0.2653 370 0.8673 470 0.38 71 0.3156 171 0.1032 271 0.4195 371 0.3861 471 0.392 72 0.1883 172 0.1468 272 0.2291 372 0.2216 472 0.1889 73 0.0427 173 0.3262 273 0.5096 373 0.4864 473 0.2661 74 0.4703 174 0.7981 274 0.1551 374 0.3426 474 0.4158 75 0.3914 175 0.277 275 0.3396 375 0.3477 475 0.7124 76 0.151 176 0.4089 276 0.2068 376 0.897 476 0.3878 77 0.0566 177 0.5457 277 0.2656 377 0.4233 477 0.2922 78 0.4494 178 0.3824 278 0.29 378 0.2663 478 0.5477 79 0.1343 179 0.2572 279 0.1868 379 0.0995 479 0.2711 80 0.1175 180 0.1076 280 0.4079 380 0.2342 480 0.1102 81 0.0948 181 0.2011 281 0.0931 381 0.054 481 0.2707 82 0.2927 182 0.246 282 0.6342 382 0.3821 482 0.1982 83 1.4183 183 0.2528 283 1.1263 383 0.1133 483 0.3845 84 0.4606 184 0.2809 284 0.538 384 0.1481 484 0.3107 85 0.4616 185 0.2991 285 0.4182 385 0.1078 485 0.1104 86 0.3472 186 0.2539 286 0.3296 386 0.6482 486 0.1085 87 0.4354 187 0.0501 287 0.753 387 0.6673 487 0.137 88 0.3651 188 0.0912 288 0.3301 388 0.3328 488 0.7324 89 0.3981 189 0.1884 289 0.4129 389 1.0462 489 0.1292 90 0.0653 190 0.0737 290 0.2837 390 0.1859 490 0.2095 91 0.3756 191 0.2251 291 0.084 391 0.8444 491 0.5224 92 0.1102 192 0.4517 292 0.5516 392 1.4813 492 0.2991 93 0.5242 193 0.0695 293 0.2462 393 0.7766 493 0.2481 94 0.7128 194 0.1664 294 0.1787 394 0.1917 494 0.1699 95 0.3766 195 1.0261 295 0.6444 395 0.3405 495 0.1483 96 0.5457 196 0.4067 296 0.1664 396 0.6904 496 0.5248 97 0.0653 197 0.216 297 0.758 397 0.5831 497 0.1682 98 0.1968 198 2.0596 298 0.2026 398 0.3319 498 0.2192 99 0.3388 199 0.6158 299 0.4626 399 0.1969 499 0.4555 100 0.1251 200 0.469 300 0.4077 400 0.0662 500 0.4062 133  A.7 Lastre Table 32: XRF results of Lastre zone Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 1 0.0121 55 0.0095 109 0.0149 163 0.0084 2 0.0239 56 0.0068 110 0.0106 164 0.0987 3 0.0146 57 0.0171 111 0.0239 165 0.009 4 0.0356 58 0.013 112 0.0114 166 0.0282 5 0.0117 59 0.0101 113 0.0374 167 0.0092 6 0.0087 60 0.0205 114 0.0937 168 0.0073 7 0.024 61 0.0298 115 0.0325 169 0.0151 8 0.0215 62 0.0143 116 0.0163 170 0.0464 9 0.0134 63 0.0135 117 0.0118 171 0.0207 10 0.0117 64 0.0158 118 0.0431 172 0.0125 11 0.0113 65 0.0091 119 0.0103 173 0.0062 12 0.0181 66 0.062 120 0.0133 174 0.0253 13 0.01 67 0.0064 121 0.0081 175 0.0201 14 0.009 68 0.0127 122 0.0114 176 0.0086 15 0.0572 69 0.0056 123 0.0111 177 0.0169 16 0.0078 70 0.0213 124 0.0263 178 0.007 17 0.0196 71 0.0088 125 0.0132 179 0.0119 18 0.0119 72 0.0091 126 0.0118 180 0.0294 19 0.0105 73 0.0151 127 0.0069 181 0.0351 20 0.0151 74 0.0195 128 0.0086 182 0.0256 21 0.0127 75 0.0203 129 0.03 183 0.0088 22 0.0125 76 0.0092 130 0.022 184 0.0328 23 0.014 77 0.0117 131 0.0187 185 0.0096 24 0.0241 78 0.0161 132 0.0575 186 0.0104 25 0.0131 79 0.0078 133 0.0117 187 0.0363 26 0.0441 80 0.0151 134 0.0164 188 0.0269 27 0.0274 81 0.013 135 0.0653 189 0.015 28 0.0327 82 0.0187 136 0.0138 190 0.0092 29 0.0095 83 0.0081 137 0.0079 191 0.0209 30 0.0107 84 0.0225 138 0.0083 192 0.0087 31 0.0096 85 0.0482 139 0.0136 193 0.0099 32 0.0247 86 0.0165 140 0.0571 194 0.0402 33 0.0155 87 0.0276 141 0.0268 195 0.0129 34 0.0095 88 0.0096 142 0.0155 196 0.0079 35 0.0092 89 0.02 143 0.0063 197 0.0212 134  Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) Sample No. Cu (%) 36 0.0103 90 0.0147 144 0.0181 198 0.0148 37 0.0083 91 0.0156 145 0.0704 199 0.0379 38 0.0065 92 0.0079 146 0.0086 200 0.0146 39 0.0133 93 0.0106 147 0.0208 201 0.0147 40 0.0107 94 0.0063 148 0.0783 202 0.024 41 0.0834 95 0.0094 149 0.0111 203 0.0169 42 0.0102 96 0.0456 150 0.0074 204 0.0169 43 0.0128 97 0.0201 151 0.0078 205 0.0121 44 0.0075 98 0.0083 152 0.0227 206 0.081 45 0.009 99 0.0125 153 0.015 207 0.0209 46 0.0109 100 0.0114 154 0.0102 208 0.0141 47 0.01 101 0.0121 155 0.012 209 0.0115 48 0.006 102 0.0247 156 0.0338 210 0.0105 49 0.018 103 0.0134 157 0.0306 211 0.0113 50 0.0072 104 0.0067 158 0.0325 212 0.0167 51 0.0106 105 0.0096 159 0.0066 213 0.0186 52 0.0085 106 0.0061 160 0.0074 214 0.0125 53 0.0131 107 0.0073 161 0.0131 215 0.0692 54 0.0155 108 0.0145 162 0.008              135  Appendix B  Assay Results B.1 SPPP07 Table 33: Assay results of SPPP07 samples Sample No. Assay Cu (%) Sample No. Assay Cu (%) 7 0.45 202 0.39 13 0.57 249 0.09 16 0.76 252 0.47 17 0.34 256 0.71 22 0.59 260 0.50 26 0.43 268 0.31 27 0.47 300 0.85 65 0.44 330 0.73 94 0.23 332 1.31 97 0.02 358 0.99 100 1.30 380 0.21 105 0.57 419 0.60 109 0.51 423 0.56 121 0.96 456 0.02 125 2.71 487 0.37 B.2 SPPP11 Table 34: Assay results of SPPP11 samples Sample No. Assay Cu (%) Sample No. Assay Cu (%) 17 1.28 267 0.44 22 0.45 276 3.44 29 0.01 280 0.50 42 3.66 318 0.50 60 0.01 380 0.00 110 0.00 383 0.52 122 1.30 390 0.24 127 0.03 409 0.00 140 0.36 416 1.52 200 0.94 425 0.51 227 0.00 450 0.95 235 0.31 451 0.43 236 0.19 459 0.62 136  Sample No. Assay Cu (%) Sample No. Assay Cu (%) 238 2.37 462 0.63 242 1.14 463 0.92 B.3 Oxido Table 35: Assay results of Oxido samples Sample No. Assay Cu (%) Sample No. Assay Cu (%) 1 0.17 195 2.62 14 1.86 198 1.23 29 0.08 223 0.12 30 0.34 224 0.19 31 0.04 237 0.58 40 0.08 243 3.53 43 1.07 271 0.99 53 0.92 316 0.35 66 2.20 339 0.91 94 0.32 401 1.09 105 0.10 408 1.11 124 0.94 416 1.50 143 1.25 516 0.95 148 0.787 524 0.701 158 0.706 525 0.188 B.4 Mixto Table 36: Assay results of Mixto samples Sample No. Assay Cu (%) Sample No. Assay Cu (%) 120 6.60 385 0.04 129 0.04 386 0.05 207 0.43 390 0.25 242 0.46 397 0.37 270 0.06 420 0.06 300 2.37 442 0.40 325 10.10 451 0.88 338 0.70 458 19.10 339 0.13 465 9.33 352 0.50 474 2.51 360 0.35 476 0.06 137  Sample No. Assay Cu (%) Sample No. Assay Cu (%) 362 1.70 478 0.06 372 0.04 481 0.04 379 5.38 496 10.25 383 0.532 501 0.1655 B.5 Sulfuro Table 37: Assay results of Sulfuro samples Sample No. Assay Cu (%) Sample No. Assay Cu (%) 1 0.02 376 1.80 14 0.83 382 0.04 18 0.05 394 0.10 269 1.53 395 2.62 270 0.02 411 2.35 274 0.03 415 0.63 286 0.92 433 0.11 292 0.07 448 0.99 294 1.04 458 1.21 303 1.57 463 0.87 310 0.06 471 0.84 341 0.56 472 2.55 349 1.21 483 0.66 355 0.973 486 1.57 365 0.0141 487 1.915 B.6 M3  Table 38: Assay results of M3 samples Sample No. Assay Cu (%) Sample No. Assay Cu (%) 198 1.83 428 0.34 267 0.99 434 0.21 282 0.25 438 0.31 327 0.78 439 0.22 336 0.40 448 0.41 368 0.28 449 0.42 370 1.00 453 0.44 391 0.37 455 0.21 392 0.49 456 0.43 138  Sample No. Assay Cu (%) Sample No. Assay Cu (%) 409 0.24 457 0.20 415 0.77 462 0.60 416 0.49 465 0.35 422 0.18 479 0.17 423 0.535 488 0.212 424 0.181 489 0.263 B.7 Lastre Table 39: Assay results of Lastre samples Sample No. Assay Cu (%) Sample No. Assay Cu (%) 126 0.009 193 0.008 130 0.011 195 0.008 132 0.066 198 0.016 134 0.013 202 0.008 143 0.003 208 0.006 153 0.005 210 0.006 161 0.008 212 0.011 170 0.016             139  Appendix C Numerical Mine Model Results Development of Numerical mine model and the results are presented as supplementary files. The supplementary files can accessed using the link http://hdl.handle.net/2429/52134                     140  Appendix D Vulcan Model The data used in the development of Vulcan model is included in the appendix. The data files includes collar, assay and survey data for every zone. The summary of resource estimation results generated are included along with the Vulcan model results and are presented as supplementary files. The supplementary files can accessed using the link http://hdl.handle.net/2429/52134                   141  Appendix E Correlation Data of XRF and Assay Results E.1 SPPP07 Table 40: Correlation data for SPPP07 zone Sample No. XRF Cu (%) Assay Cu (%) Sample No. XRF Cu (%) Assay Cu (%) 7 0.239 0.451 202 0.2668 0.391 13 0.7171 0.565 249 0.0638 0.0867 16 0.5914 0.763 252 0.417 0.471 17 0.2738 0.34 256 0.8262 0.711 22 1.0725 0.59 260 0.3181 0.501 26 0.5156 0.432 268 0.3049 0.313 27 0.2363 0.472 300 0.7561 0.849 65 0.3889 0.435 330 0.9107 0.725 94 0.2512 0.23 332 0.4004 1.31 97 0.0509 0.0237 358 0.9323 0.988 100 0.8311 1.3 380 0.1556 0.211 105 0.5233 0.574 419 0.7345 0.602 109 0.4108 0.505 423 0.8984 0.557 121 0.4969 0.955 456 0.0331 0.0157 125 1.7062 2.71 487 0.2478 0.373 E.2 SPPP11 Table 41: Correlation data for SPPP11 zone Sample No. XRF Cu (%) Assay Cu (%) Sample No. XRF Cu (%) Assay Cu (%) 17 0.8558 1.28 267 0.4954 0.444 22 0.7655 0.447 276 1.2963 3.44 29 0.0233 0.0065 280 0.1194 0.501 42 2.1062 3.66 318 0.793 0.498 60 0.0773 0.0075 380 0.0736 0.0044 110 0.0292 0.0047 383 0.9131 0.521 122 0.5234 1.295 390 0.0565 0.236 127 0.0277 0.511 409 0.3524 0.0029 140 0.3025 0.363 416 0.3196 1.52 200 0.5765 0.938 425 0.1401 0.508 142  Sample No. XRF Cu (%) Assay Cu (%) Sample No. XRF Cu (%) Assay Cu (%) 227 0.0239 0.0031 450 0.4879 0.953 235 0.2403 0.308 451 0.8012 0.429 236 0.2134 0.185 459 1.8338 0.619 238 1.0141 2.37 462 0.3314 0.631 242 0.3386 1.135 463 0.4333 0.921 E.3 Oxido Table 42: Correlation data for Oxido zone Sample No. XRF Cu (%) Assay Cu (%) Sample No. XRF Cu (%) Assay Cu (%) 1 0.1034 0.165 195 1.3602 2.62 14 0.8596 1.855 198 0.6657 1.23 29 0.0826 0.083 223 0.057 0.124 30 0.1253 0.34 224 0.1639 0.194 31 0.0544 0.04 237 1.5424 0.581 40 0.0605 0.084 243 2.065 3.53 43 1.8802 1.07 271 0.9708 0.993 53 0.2964 0.919 316 0.9236 0.347 66 23 2.203 339 0.5521 0.914 94 0.0991 0.318 401 0.9371 1.085 105 0.0935 0.104 408 0.8199 1.11 124 0.5444 0.943 416 1.6701 1.5 143 0.5127 1.245 516 1.1201 0.953 148 0.2896 0.787 524 0.1273 0.701 158 0.8593 0.706 525 0.1497 0.188 E.4 Mixto Table 43: Correlation data for Mixto zone Sample No. XRF Cu (%) Assay Cu (%) Sample No. XRF Cu (%) Assay Cu (%) 120 8.6345 6.6 385 0.0591 0.0386 129 0.0484 0.0449 386 0.0606 0.0524 207 0.4904 0.429 390 0.4951 0.252 242 0.5016 0.456 397 0.4432 0.369 143  Sample No. XRF Cu (%) Assay Cu (%) Sample No. XRF Cu (%) Assay Cu (%) 270 0.0628 0.0576 420 0.0532 0.0594 300 5.5967 2.37 442 0.5324 0.399 325 7.1734 10.1 451 0.5564 0.878 338 0.5031 0.701 458 6.1032 19.1 339 6.3722 0.1305 465 10.6608 9.33 352 0.4973 0.504 474 5.4828 2.51 360 0.3839 0.351 476 0.0616 0.0562 362 4.9221 1.7 478 0.0556 0.0606 372 0.0665 0.0401 481 0.058 0.042 379 5.5303 5.38 496 10.4208 10.25 383 0.4323 0.532 501 0.0392 0.1655  E.5 Sulfuro Table 44: Correlation data for Sulfuro zone Sample No. XRF Cu (%) Assay Cu (%) Sample No. XRF Cu (%) Assay Cu (%) 1 0.0262 0.0231 376 0.9763 1.8 14 0.9742 0.827 382 0.0322 0.0354 18 0.0521 0.0505 394 0.0346 0.0963 269 0.8977 1.525 395 2.612 2.62 270 0.0359 0.0229 411 1.6254 2.35 274 0.0289 0.0276 415 1.9075 0.631 286 0.9254 0.92 433 0.0535 0.1075 292 0.0489 0.0731 448 3.0437 0.988 294 3.4887 1.04 458 0.9289 1.21 303 2.3415 1.565 463 1.7202 0.874 310 0.0349 0.0603 471 0.9673 0.842 341 0.9362 0.559 472 1.6541 2.55 349 0.9771 1.21 483 0.9446 0.659 355 2.2475 0.973 486 2.2521 1.57 365 0.0189 0.0141 487 0.9297 1.915 144  E.6 M3 Table 45: Correlation data for M3 zone Sample No. XRF Cu (%) Assay Cu (%) Sample No. XRF Cu (%) Assay Cu (%) 198 2.059 1.83 428 0.072 0.336 267 1.752 0.99 434 1.049 0.213 282 0.634 0.25 438 0.811 0.309 327 1.475 0.775 439 0.081 0.219 336 1.135 0.402 448 0.33 0.41 368 0.415 0.28 449 1.077 0.415 370 0.867 0.995 453 0.958 0.439 391 0.844 0.373 455 0.16 0.214 392 1.481 0.487 456 0.786 0.427 409 0.583 0.237 457 0.505 0.195 415 0.493 0.772 462 0.211 0.603 416 1.104 0.494 465 0.909 0.354 422 0.367 0.1815 479 0.271 0.1675 423 0.001 0.535 488 0.732 0.212 424 0.003 0.181 489 0.129 0.263 E.7 Lastre Table 46: Correlation data for Lastre zone Sample No. XRF Cu (%) Assay Cu (%) Sample No. XRF Cu (%) Assay Cu (%) 126 0.0118 0.0087 193 0.0099 0.0083 130 0.022 0.011 195 0.0129 0.0081 132 0.0575 0.0663 198 0.0148 0.0161 134 0.0164 0.0125 202 0.024 0.008 143 0.0063 0.0031 208 0.0141 0.0056 153 0.015 0.0051 210 0.0105 0.0056 161 0.0131 0.0077 212 0.0167 0.0111 170 0.0464 0.0159 - - -  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0167668/manifest

Comment

Related Items