ETD Collection
Permanent URI for this collectionhttps://wiredspace.wits.ac.za/handle/10539/104
Please note: Digitised content is made available at the best possible quality range, taking into consideration file size and the condition of the original item. These restrictions may sometimes affect the quality of the final published item. For queries regarding content of ETD collection please contact IR specialists by email : IR specialists or Tel : 011 717 4652 / 1954
Follow the link below for important information about Electronic Theses and Dissertations (ETD)
Library Guide about ETD
Browse
17 results
Search Results
Item A geometallurgical strategy for improving ore quality and mineral processing efficiency at Kansanshi Mine in Zambia(2016) Beaumont, ChristopherThe Kansanshi mine is located in north-western Zambia. It is a copper and gold bearing, vein hosted, structurally controlled, ore deposit, which is exploited using conventional open pit mining methods. A series of highly complex mineralogical suites have formed through the interaction of the in-situ geological, weathering and oxidation processes. Some of these mineralogical suites are extremely difficult for effective extraction of copper and gold. Currently ore is classified into 22 different quality categories using a system called “Mat_Type”. Only the “best quality” ore is directed to the metallurgical process plant, while the remaining “poor quality” ore is directed to long term stockpiles. These stockpiles are unlikely to be processed until the end of the life of mine despite containing metal quantities of significant value. A systematic investigation of Kansanshi’s mine value chain was carried out to determine if this value could be realised sooner. It was found that, due to a lack of integration between technical silos that form the mine value chain, the Mat_Type system does not take due consideration of geological, mineralogical nor metallurgical processes. Ore quality control factors are incorrect and economic data to determine suitable cut-off grades is both outdated and applied in an inappropriate manner. As a result 12 of the 22 Mat_Type ore categories are unnecessary, while a further six categories are inaccurately defined, leaving only four categories that can be considered to be correct. It is because of these errors, that so much ore is being directed to long term stockpiles. Through the research study presented in this report it was found that five key factors determine an effective ore classification system for Kansanshi mine. These factors can be defined as: Spatial distribution of mineralogical relationships between in-situ geological, oxidation and weathering domains; The impact of mineralogical groupings on copper recovery in each metallurgical process; The size, statistical distribution, range and accuracy of available data sets; The application of appropriate economic factors in the mineral resource to mineral reserve conversion process; and, The practicality of overcoming physical constraints at various stages of the mining process A more appropriate geo-metallurgical ore classification system can be developed that will consider the above listed factors. Three mineralogical groupings naturally exist in the deposit. Using total copper and acid soluble copper assay data, these mineralisation categories can be defined by a specific range of oxidation ratio. These ranges can be listed as: Primary copper sulphide minerals, with oxidation ratio < 0.1 ; Secondary copper sulphide minerals, with an oxidation ratio between 0.1 to 0.8; and , Primary copper oxide minerals, with an oxidation ratio between 0.8 to 1.0. Each mineralogical grouping can be assigned to a specific metallurgical process. Further subdivisions of these groupings can be made based upon economic grade ranges and appropriate metallurgical quality control factors which are linked to a specific metallurgical process. By implementing the proposed geo-metallurgical ore classification system, 12 of the erroneous ore categories in the Mat_Type system can be removed. The remaining 10 categories would be accurately and consistently defined. This will lead to a significant reduction in the quantity of ore directed to long-term stockpile, thereby releasing previously lost value. The new ore classification methodology proposed would be supported by a systematic process to manage regular, periodic updates, based on new data and developments in technical understanding across all functional areas of the mine value chain. A continued dialogue and sharing of knowledge between the main technical silos is critical when promoting a robust and integrated ore classification system. Such a system has the potential to remain relevant throughout the life of Kansanshi mine.Item An investigation of the mine call factor at Mazowe gold mine in Zimbabwe(2018) Mpofu, KennethThis study sought to establish the causes of a persistent low Mine Call Factor (MCF) at Mazowe Mine in Zimbabwe. The theory of Mine Call Factor expresses the total metal recovered and in residue, as a percentage of the metal indicated by the mine’s estimation method. From 2001 to 2016 the average MCF was 55%. Historically, the mine is characterized by a low MCF. It is estimated that over the past 50 years, unaccounted for gold amounts to 284,000 ounces. During the period under review, a total of 308,000 ounces of gold was called for, but only a total of 170,000 ounces or 55% was accounted for. This suggests that, theoretically, there was a loss of 45%, which implies that about 45% of the expected revenue was not realized. This adversely impacts the mine’s ability to achieve its financial goals. On average, a total of 5,680 ounces is unaccounted for per year. The research methods included a literature review, mine visits to familiarise with the mine, collect data and observe and assess protocols and practices along the mining value chain that impact the MCF. The study revealed that the cause of the low MCF is primarily grade loss. The main sources of the grade loss are the resource estimation method which overstates the grade and the loss of high- grade fines. The loss of grade is aggravated by the lack of operational flexibility. Due to the shortage of mining faces, the mine is forced to produce from resources, rather than reserves. The bulk of the production comes from inferred resources, which are not adequately evaluated. Consequently, the grade is impacted by unplanned dilution and ore/waste misclassification. The estimation method, unplanned dilution and ore/waste misclassification contribute to the apparent loss. The real loss is constituted by the loss of gold along the mining value chain. The main contributing factor is the loss of fines, which are associated with high grades, primarily in the stopes. Whilst there are both apparent and real grade losses, it appears that the loss is primarily apparent and mainly attributable to the method of resource estimation. However, the real loss component is quite significant and largely attributable to the loss of fines. This is evident from the fact that fines reclamation has a significant positive impact on the MCF. To minimize the apparent loss of grade, adoption of geostatistical resource estimation methods, review of grade control and QA/QC protocols, and improvement of operational flexibility in order to reduce production from inferred resources, are recommended. In order to minimize real grade losses, the blasting pattern needs to be modified to optimize fragmentation and reduce the generation of fines. Fines reclamation operations must be prioritized, better planned, better resourced, ramped up and conducted in current stopes, rather than abandoned stopes as is the practice at presentItem Comparison of the genetic algorithm and linear programming approaches to optimally locate orepasses in the Venetia underground project(2019) Dabula, GangathaIn the process of optimisation and the need to get the best value from a mining operation, reducing hauling costs is a key requirement. The positioning of orepasses at Venetia Mine, which uses a sub-level caving underground mining method, will be of great importance. This will impact on the tramming distances that Load Haul Dumpers (LHDs) will have to travel from drawpoints to orepasses. This will also impact on the hauling distances for trucks from underneath the orepass to the underground crushers. In the case of Venetia Mine, tramming is planned to require diesel, tyres and other consumables. This also uses up the operational hours of machinery during its commissioned life. This study aims at analysing the current positioning of orepasses in the Venetia Underground Mine design. The current position of orepasses at Venetia Mine is decided based on experience and geotechnical constraints. This study compares this positioning to positions optimised using a linear programming approach and the genetic algorithm approach. Other optimisation techniques were available but, these two were considered because these are the best-suited optimisation techniques for solving the orepass location problem and some research on optimal facility location has been previously done using these techniques. A Microsoft Excel model was produced to calculate the equivalent cost per tonne metre for each orepass using the total tramming distance from the loading point to the orepass. This model was also able to determine the ideal orepass for each loading point by selecting the orepass with the shortest loading distance. The Microsoft Excel model had been developed in house at De Beers to test whether the orepasses had been positioned in the correct positions in the design and to also determine what the total tonnage from each orepass was during the life of mine. This excel model was then optimised using Palisade Evolver ® software. The OptQuest ® tool was used to solve the linear programming solution as one of the model constraints was not a linear function. The constraint in question is that the orepasses are not to be closer than 40m from the orebody based on geotechnical recommendations. The optimisation results showed that the genetic algorithm optimisation resulted in a 12% improvement in the total tonne metre cost for the orepasses on the K01 orebody. The linear programming solution resulted in an improvement of 10.5% on the total cost per tonne metre in the model from a base value of R 20 941 per tonne metre. The results indicated that optimisation could bring about an improvement in operating cost. However, there needs to be future work done to consider geotechnical and geological constraints which will be encountered in a real-world mine design scenario.Item Pillar stability analysis from secondary extraction at Goedehoop Colliery’s No.2 Seam(2018) Banda, ChikondiThe current outlook for coal within the global energy market compels South African based collieries to re-strategize their current methods of extraction, with the decline of coal reserves within the Witbank coalfield. An economically feasible alternative would be secondary extraction methods that are relatively low in financial and safety risk. However, efficient mining can only occur where optimal rock engineering practices are used. The No. 2 Seam (thickness of about 6m) has always been a favoured deposit for extraction within Goedehoop colliery, consisting of ‘good quality’ economical coal. Where previously mined at about 4.5m, a possible 1.5m remains either in the roof or floor for extraction (when mined in conjunction with the No.1 seam in areas of thin seam parting). Despite the remaining coal having the potential to be included in future plans, the current concern is the stability of pillars within panels associated with the progressive failure due to minimal roof support and poor rock engineering design, without accounting for the risk of failure. If the failure of the pillars were to occur, it will cause the sterilization of the No.4 seam that is currently being mined above the No. 2 Seam. A bottom/top coaling case study was looked at Goedehoop colliery with the aim to evaluate the stability of pillars within the panel/s using stability criterions (Factor of Safety, Probability of Failure and Probability of Survival). Using the pressure arch theory and applicable strength equations to the No.2 seam, the stability of individual pillars was evaluated, where previously coaling had taken place. The results estimated the likelihood of failure of the pillars within the selected panel. The results from the study showed that if and where coaling has taken place it is likely that the pillars would remain stable well above the FOS value limit of 1.6 and resides within the 99% POS limit of production panels.Item Yielding pillar design in South African collieries(1997) Oldroyd, OldroydThree cases of pillar failure on Southern African Collieries have been studied to analyse the behaviour of both the pillars and the overlying strata. Each of the cases shows a different type ofpillar and strata behaviour during failure and thus provides an opportunity for back analysis. In the first case, pillars failed in a controlled fashion while the overlying strata behaved in an elastic fashion. In the second pillars failed in a controlled fashion while the surrounding strata behaved inelastically. In the third failure was initially controlled but became uncontrolled. Computer models have been run to determine the theoretical critical post peak pillar slopes and the results of these models have been compared with the actual pillar behaviour as derived from in-situ measurements during failure and that which might be predicted from the theory of controlled and uncontrolled pillar failure. Comparisons are also made with the expected behaviour implied from the results of in-situ strength tests carried out on small coal pillars to ascertain their load deformation characteristics. The results indicate that the behaviour of the pillars more closely resemble that predicted from in-situ tests carried out by Van Heerden4• The results also indicate inadequacies in using elastic methodologies to determine whether pillar failure will be controlled or uncontrolled.Item Conceptualization of a mining information model (MIM) using real time information for smart decisio making: a smart economics approach(2018) Javaid, FaiqSouth Africa is a mineral resource-rich country with the largest concentrations of gold and platinum in the world. Yet the South African mining industry is facing an economic crisis. Some of the reasons for this crisis are: Low commodity prices, escalating production costs, depletion of economically ore grades, volatile currency, volatile exchange rates, difficulty to compete in the international markets, increased concerns of deteriorating relationships with mine-worker unions and the South African Government’s Department of Mineral Resources (DMR) etc. Thus there is a critical need to develop an information and decision-making system that will cater for modern-era needs. Such a system would need to optimize production cost, while properly linking it to current and expected market conditions to enable synchronised and timely decision-making. This can only be done via a framework that is supported by relevant and timely information. This will need to include the following mine and market data (in both current- and anticipated-forms): Production rates; assessment of what is going on underground; incident reporting; scheduling; costing; market updates; inventory management; life cycle management. And such a system, named Building Information Modelling (BIM), was developed for the construction industry. This indicates that development of a Mining Information Modelling (MIM) may also address above-mentioned aspects, allowing maximum production, optimal cost, less uncertainty and more efficiency – something that is difficult to attain via existing mining-software. The purpose of this research was to investigate and present a conceptual framework for the development of MIM and a smart, real-time decision-making tool. The study reveals that MIM can be achieved via the combination of software from a number of providers, together with some additions. Ideally, the combination should cover the entire mining value chain. With MIM, a decision-making tool (with the support of MIM) can either be developed as a separate software or can be integrated with MIM. Such system will cater for modern-era needs, thereby enhancing mining capabilities without affecting job-creation.Item Unlocking value through improved production decision making : a trackless mining systems analysis(2018) Mukonoweshuro, ChristopherThis study was based on the hypothesis that there are opportunities to maximize production outputs in many existing underground hard rock trackless mining systems using the same or less resources by improvement in decision making paradigms. This is very important in the current operating environments of uncertainties and continued drop in metal prices. The project main goal was thus to carry out a detailed investigation of trackless mining production systems and test how to maximize output by focusing on three objectives, namely: analyzing key technical factors that impact the production rates in terms of tons per hour, identifying major operational activities which impact effective equipment operating hours, and identifying decision support systems (DSS) to improve operational decision making. Regarding the first objective (production rates), through the analysis of trackless mining as a serial production system, it was shown that production rates could be increased by focusing at system level, process level and work station/equipment level decisions. System level decisions must minimize the total residence time of the material (ore) in transit or work in process(WIP). This will open capacity for generating more ore. Process level decisions must reduce the gross cycle times at the work stations to equal or be below the Takt times inorder to smoothen production flow. Takt time is an important factor in a production system which shows the maximum cycle time allowed to meet the daily demand. The third level focuses on the capability of the mining equipment itself through decisions that improves the reliability, maintainability and capacity. Decision tables based on reducing the equipment failure rates (λ), improving the repair rates (µ) and the cycle times were developed to aid in making the reliability, maintainability or capacity decisions. For the second objective (operational activities), the focus is to maximize effective operating times of the equipment through reduction of delays. The study shows this can be achieved through use of real-time decision support systems (DSS) for better control of the operations. The third objective was able to identify functional modern DSS that can be implemented in trackless mining. Effectively, the study was able to highlight opportunities of generating extra capacity for trackless mines at same or less resources by focusing on the above three objectives.Item Concept development of a smart rock engineering system for real-time decision-making and risk minimization in deep level hard rock mines: a digital mining approach(2018) Kwiri, JosephHuge milestones have been achieved in an attempt to reduce rock related accidents in the South African mining industry. However, ZERO HARM has not yet been accomplished. Despite the mining industry being active for years, accident statistics are still high compared to the mining industry in Australia and United States of America, in particular gold mining. Gold deposits occur in deep and ultra-deep levels where the stress levels and rock related risks are high making mining, more difficult and riskier. This is further exacerbated by the intensive labour force at such mining depths. Risk management strategies have been formulated and refined over the years and their success is not questionable as there have been, significant reduction of fatalities over the years. However, the number of fatalities over the years is still too high. As an example, fatal accidents exceed twenty from January 2018 to July 2018. These unfortunate statistics indicate that existing strategies to reduce rock engineering risks, have limitations. Surface mining, civil and petroleum engineering have more advanced risk management technologies compared to underground mining. The attributes of these technologies can be used to develop better underground risk management strategies. Such attributes include remote operation, integrated sensor system, and the ability to predict impending danger. Some of these attributes can be compared to human body and brain as it is an ideal system that smart systems should mimic in particular decision making and actioning of decisions made as well as the ability of skin to sense, repair itself and insulate. Used with advanced material science, such properties of the skin (sense, repair itself and insulate) can be used to make a type of support that can reduce support replacement costs and ventilation related operating costs. The conceptual smart rock engineering system developed consist of sensors, expert system for data analysis and decision making, wireless communication system and an emergency and preparedness response plan through automated alerts which are received by the miner in the stope area. The system should measure a number of parameters including stress, water level, convergence, face advance, face profile, loading of support and ground movement. The results are displayed concurrently on a video wall in the control room area. A case study was done to compare what is currently available in the Digital Mine Laboratory and the conceptual smart rock engineering. From this case study, opportunities to improve the system installed in the Digital Mine Laboratory were identified. The Botswanan earthquake, which occurred on the 3rd of April 2017 at 17:40 Coordinated Universal Time (UTC) was a critical event to test the functionality of the system installed in the Digital Mine Laboratory. Recordings from the Digital Mine Laboratory were compared with other ground sensing technologies (United States Geological Survey (USGS) and GEOFOrschungsNetz Global Seismic Network (GEOFON) systems). which also captured the Botswana earthquake event USGS and GEOFON systems recordings suggested that there are a number of possible mechanisms that could have resulted in the earthquake. Compared to the USGS and GEOFON systems, Digital Mine Laboratory system could not provide a self-analysis data that could be used to determine the source and source mechanisms. The Digital Mine Laboratory Botswana earthquake event is an indication that the system can be developed and or improved. For a comprehensive analysis, more data from various sensors needs to be collected, for example, by connecting to the national seismic monitoring system or to the local mines seismic monitoring system. Such connections will enable the development of a better self-analysis system and possibly prediction of future events within the Digital Mine Laboratory.Item The application of ash adjustable density in the evaluation of coal deposits(2017) Roux, LeonThe initial evaluation of a coal deposit often raises uncertainty with regard to the accuracy of the reported resources and reserves. Difficulty is experienced in reconciling tonnages produced during mining and beneficiation with the original raw field data. The credibility of resource and reserve estimations, which form the basis on which an entire mining enterprise is motivated, funded and established as a commercially viable proposition, is of paramount importance. In essence, this research has sought to establish and validate a more realistic and accurate method for (i) coal resource and reserve estimation and (ii) the reconciliation of saleable tonnages produced following beneficiation. Previous research undertaken by this author resulted in the formulation of a methodology to provide a more accurate assessment of a coal body by using the dry density of the coaly material derived from proximate analytical data for the ash content for float fractions obtained from float sink analysis. The determination of the dry density was obtained through the application of the ash adjusted density algorithm derived from the regression of the median proximate ash values at fixed float densities in the range 1.35 g/cc to 2.20 g/cc. The derived density results were validated against laboratory pycnometer determined densities and found to be applicable to both of the two major geological stratigraphic units in the Waterberg Coalfield. This resulted in significantly more accurate predictions of coal product tonnages from the Waterberg Coalfield. In the current research, this methodology has been applied to cover the entire coal value chain, from exploration through to final products. The primary purpose was to ascertain the correct resource and reserve values relative to that originally reported using conventional methods and to match those values to actual saleable tonnages produced down the line. Density is the key factor underpinning such calculations and this varies not only due to geology, and specifically coal rank, type and grade, but also to the method used for its measurement. It plays a major role in the estimation of reserves and in the beneficiation process because density is the primary separation medium utilized in coal beneficiation. Coal plies and particles have different relative densities and physical properties, as determined by their maceral composition, rank, mineral (ash) and moisture contents. The relationship between such parameters, as measured by ash, moisture content, matrix porosity and density, was found to play an even greater critical role in establishing the correct tonnage of coal at any single point in the value chain. A combination of theoretical, empirical and reconciliatory evaluations of the available data from the exploration phase through the mining process to final production has shown that an integrated approach using the ash adjusted density methodology provides more accurate and credible results with a higher degree of confidence at all stages across the coal value chain than is currently possible using conventional practices.Item The appropriateness of the localised uniform conditioning technique for high-nugget Birimian-style gold deposits(2017) Maritz, EmmarentiaThe localized uniform conditioning (LUC) technique converts conventional Uniform Conditioning (UC) grade-tonnage curves into single grade values attached to each smallest mining unit (SMU). This is achieved by ranking the SMUs within a panel in increasing order of their grade based on the local grade patterns predicted by direct kriging of the SMUs. However, the quality of this localization process will depend heavily on the validity of the predicted grade patterns. A study was undertaken to determine how valid the predicted grade patterns of a typical Birimian-style gold deposit (with high nugget effect and strong short-range variability) might be expected to be. The direct SMU kriging rankings (based on sparse data) were compared with the grade control model ranking (based on close-spaced data and the best available estimate of the deposit). The results showed a satisfactory correlation and relationship between these rankings. It was concluded that the application of the LUC technique is still useful and appropriate for this style of deposit.