Search Documents

By J. Y. Foley, K. Eden, A. E. Grosz, C. J. Giles, J. E. Grosz

"Portions of the Icy Cape placer Au district have been described and exploited since the early 1900s. Industrial studies of Au-bearing garnetrich intervals established a 3.5 million troy ounce resource in the western half of the district. This estimate is not compliant with CIM and SME definition standards for Mineral Resources and Mineral Reserves. We note high variability in the textures of sediments and their heavy mineral (HM) content and composition in the district, probably as a function of the depositional environment. Sediments are medium- to fine-grained sands with a valuable HM assemblage that averages 26.0% in a range of 1.1– 72.6%. Valuable heavy minerals (VHMs) garnet, magnetite, ilmenite, rutile, and zircon, but excluding Au, average 7.9% in a range of 1.0–37.6% of the HM assemblage. Trace quantities of uraninite, thorite, monazite, xenotime, and others are noted. Garnets have an average Y content of 726 ppm with a range of 0–2150 ppm. Energy-dispersive X-ray fluorescence confirms compositions commensurate with Ca, Al, and Mg silicates. We provide empirical analytical data and note high variability in the textures, HM content, and compositions as a probable function of depositional environment. Icy Cape District land and resources are owned by the Alaska Mental Health Trust Authority and managed by the Trust Land Office (TLO). IntroductionThe Icy Cape Heavy Minerals (HM) District in Alaska subtends more than 200 km2 fronting the Pacific Ocean and backing to the foot of the Robinson Mountains at about 1700 m above sea level. We classify the district into geomorphic provinces; the eastern region as deltaic (Figure 1) and the western region as coastal plain (Figure 2), based on geomorphologic expressions. The coastal plain is entirely built up of unconsolidated outwash sediments and ancient beach sand deposits."

Jan 1, 2017

By J. Y. Foley, K. Eden, A. E. Grosz, C. J. Giles, J. E. Grosz

"Portions of the Icy Cape placer Au district have been described and exploited since the early 1900s. Industrial studies of Au-bearing garnetrich intervals established a 3.5 million troy ounce resource in the western half of the district. This estimate is not compliant with CIM and SME definition standards for Mineral Resources and Mineral Reserves. We note high variability in the textures of sediments and their heavy mineral (HM) content and composition in the district, probably as a function of the depositional environment. Sediments are medium- to fine-grained sands with a valuable HM assemblage that averages 26.0% in a range of 1.1– 72.6%. Valuable heavy minerals (VHMs) garnet, magnetite, ilmenite, rutile, and zircon, but excluding Au, average 7.9% in a range of 1.0–37.6% of the HM assemblage. Trace quantities of uraninite, thorite, monazite, xenotime, and others are noted. Garnets have an average Y content of 726 ppm with a range of 0–2150 ppm. Energy-dispersive X-ray fluorescence confirms compositions commensurate with Ca, Al, and Mg silicates. We provide empirical analytical data and note high variability in the textures, HM content, and compositions as a probable function of depositional environment. Icy Cape District land and resources are owned by the Alaska Mental Health Trust Authority and managed by the Trust Land Office (TLO). IntroductionThe Icy Cape Heavy Minerals (HM) District in Alaska subtends more than 200 km2 fronting the Pacific Ocean and backing to the foot of the Robinson Mountains at about 1700 m above sea level. We classify the district into geomorphic provinces; the eastern region as deltaic (Figure 1) and the western region as coastal plain (Figure 2), based on geomorphologic expressions. The coastal plain is entirely built up of unconsolidated outwash sediments and ancient beach sand deposits."

Jan 1, 2017

By H. Rosen

The pump is the heart of most industrial and mining plants, providing the energy to separate valuable minerals from process feed. The actual effect of the pump, both in terms of production costs and overall process efficiency, is often assumed or misunderstood by engineers due to insufficient pump performance data. Pump performance can be understood and corrected, where required, through continuous performance monitoring, leading to dramatic improvements in both pump life cycle costs and system efficiencies. The performance of every pumping application is subject to variation and the main contributing factors are as follows: ?Unavoidable deviations from pump supplier curves ?Pump wear ?Pump curve derating factors for slurry pumping ?System curve variation ?Deviation from basic principles of centrifugal pump operation. These are discussed in detail through case studies and the author?s personal experiences. This paper shows ways to minimize the effect of these variations, thereby reducing performance loss and preventing damage to the pump. The paper discusses the benefits of efficient pump operation, the requirements of pump performance monitoring, and its practical application to pump operation in the slurry and water environments. The cost of remedial actions is explored and compared against the savings in both power and plant reliability, resulting in a reduced life cycle cost of the pump.

Jan 1, 2010

By B. Talbot

How this simple plant generates a large revenue stream. The concept behind the Tailings Scavenging Plant (TSP) is simple: it is a non-complicated modular plant which is designed to be more stable than upstream plants. This, plus the use of ?high energy?, enables the plant to recover any metal missed by the upstream operation. Impala Platinum?Mineral process started pilot plant exercises in June 2003. The pilot plant ran for one year before the capital for the full-scale plant was approved. This year was used to motivate, design and cost a full-scale plant. The metal produced during the pilot and feasibility study enabled this part of the project to run at a small profit. The scale-up to full-scale was relatively simple as the pilot plant had a capacity of 180 m3/hrand the full-scale plant was designed for a capacity of 550 m3/h. The design for the full scale plant was completed in May 2003, with ground breaking taking place on 19 May 2003. The construction phase of the project went well with the commissioning of the first two modules taking place on 22 October 2003. The remaining two modules were commissioned two weeks later. The plant was at full capacity by the 20 November 2003. A review of the plants performance for the first 5 production months is as follows: the plant has required only a few minor modifications and additions, it has run well in excess of 95%availability with no major wear apparent. In addition to this, the plant has produced 3 times the metal forecast.

Jan 1, 2004

By C. K. Chabedi

"This paper discusses the difficulties associated with the potential exploitation of the deep multi-seam resources east of the Daarby fault in the Waterberg coalfield. The resources occur at a depth greater than 250 m and the thickness of the coal is roughly 110 m, but the top 50 m comprises coal intercalated with shale and the bottom 60 m contains five seams with sandstone and shale partings. Various factors affecting multiple seam mining at these great depths are discussed with reference to lessons learned from local and international experience on multi-seam mining. Field geological and geotechnical data was utilized to assess the stability of the roof of the seams. There is no specific rock mass rating for the Waterberg area, therefore approximate coal mine roof rating (CMRR) values were used to propose appropriate support strategies. Analysis of Multiple Seam Stability (AMSS) was used to analyse the strength of the parting or interburden between the various seams, the mining sequence, and the interaction between the various seams.The research indicated that it is possible to mine seams with a low CMRR at high mining rates using longwall mining, although support for gateroads is expected to be expensive, time-consuming and onerous to install, and will impact gateroad development rates. It will not be possible to simultaneously mine zones in close proximity and failure of the interburden is predicted, thus dangerous mining conditions are anticipated. However, it will be possible to mine just two of the eleven zones using longwall mining. IntroductionThis paper discusses the difficulties associated with the potential exploitation of the deep multi-seam resources east of the Daarby fault in the Waterberg coalfield using bord and pillar and longwall mining. Figure 1 illustrates the location of the Waterberg coalfield relative to other coalfields in South Africa, while Figure 2 shows the areas containing deep and shallow resources. The Waterberg resources are expected to contribute to South Africa’s future energy requirements, and are currently exploited at the Grootegeluk open pit mine. Grootegeluk produces coal mainly for a power station, with the higher quality product supplied as metallurgical coal. The deposit is technically unique and challenging, being a multi-seam coal deposit with a total of 12 seams over a thickness of 110 m as shown in Figure 3, including Zone 5. There is currently limited knowledge on the multi-seam mining of the deep Waterberg resources. Multi-seam mining utilizing the bord and pillar method has been practised in South Africa before, but at a depth of less than 100 m in the Witbank coalfields. Multi-seam mining in thin seams has also been performed in the Natal coalfields at a depth of less than 160 m, but mostly using bord and pillar mining and secondary mining (partial pillar extraction)."

Jan 1, 2016

By K. C. Sole, M. E. Mitshiabo

Although surfactants have proven to increase metal extraction by improving solution flow in porous media, these reagents have not been widely adopted in heap leaching operations because of potential incompatibilities with downstream processes that are subject to surface properties. The BASF leaching aid, LixTRA 118, was previously found to have no significant effect on post-leach operations. This study evaluated its performance in ores of different particle size distributions and the impact on acid consumption and metal recovery. Column leach studies, carried out on a copper oxide ore and mixed oxide–sulfide ore, showed that the effect of LixTRA 118 was more pronounced in ores with wider particle size distributions than in those with more uniformly sized particles. Solution percolation rates increased by up to 18% in the presence of leaching aid, and increased copper recovery and acid consumption. Improved wetting facilitated the kinetics of acid diffusion at macroand micropore level. There was no significant effect on iron recovery

Jan 1, 2023

By Brian W. Mackenzie

The mineral exploration environment has observable economic characteristics, the assessment of which provides guidelines for exploration planning in the mining company. The long-term survival of the mining company is directly dependent on successful exploration. Exploration planning contributes to successful exploration through the development of exploration strategies. The objective of exploration planning is to select exploration environments which will enable the mining company to realize best possible profit and corporate risk conditions. Thus, the formulation of exploration strategies is based on profit and corporate risk criteria. Strategic considerations include the selection of commodities as exploration targets, geographic concentration of exploration effort, the realization of acceptable survival conditions, and the structuring of an exploration programme to optimize the economic benefits of the exploration investment process.

Jan 1, 2014

By D. F. Malan, D. Ile

This study explores the use of backfill in hard rock bord-and-pillar mines to increase the pillar strength and extraction ratio at depth. The use of backfill will also minimize the requirement for tailings storage on surface and the risk of environmental damage. A literature survey indicated that backfill is extensively used in coal mines, but rarely in hard rock bord-and-pillar mines. To simulate the effect of backfill confinement on pillar strength, an extension of the limit equilibrium model is proposed. Numerical modelling of an actual platinum mine layout is used to illustrate the beneficial effect of backfill on pillar stability at greater depths. The magnitude of confinement exerted by the backfill on the pillar sidewalls is unknown, however, and this needs to be quantified using experimental backfill mining sections equipped with suitable instrumentation.

May 3, 2023

By E. M. Sacris

The semi-autogenous grinding mill is the bottleneck whenever treating hard and finely fragmented ore at Benguet Corporation's Kaline Concentrator. Low milling rates result from very high and fluctuating mill power draw. Attempts to use a mill horsepower control loop to regulate feed rate were unsuccessful because of operational upsets caused by fluctuations in feed rate on downstream operations. The objective of the computer control was to maximize tonnage throughput by stabilizing the power draw at a high but safe level without upsetting subsequent operations. The semi-autogenous mill is operated in closed circuit with a stationary screen. Instrumentation consists of a feed weightometer, a power transmitter to measure mill load, a headwater flowmeter and an ultrasonic level sensor to measure the mill discharge sump level. A 64 k-byte FOX-3 Foxboro computer processes field data from various sensors and manipulates the input control variables. These variables are the fresh feed rate and the headwater flowrate. Feed rate is controlled through a supervisor-type computer program aimed at maximizing and stabilizing power draw. The control scheme was according to how the operators responded during an operator-directed automatic control mode. Mill dilution is similarly controlled. Results achieved reflect a 3% increase in tonnage milled because of the reduction in control variability that ensued.

Jan 1, 1987

By V. Jokovic

Hydrocyclones are the most commonly used classifiers in the mining industry. The reason for this is their low operating cost and a small footprint. Data from plant trials where screens replaced hydrocyclones indicated that lower bypass of fines and sharper separation could improve breakage rates, with a positive effect on flotation circuit performance throughput. Despite the importance of classification, hydrocyclones have not received as much attention as more expensive comminution units, and their technological progress has been incremental. Noticeable water split (Rf) reductions of a semi-inverted hydrocyclone have been reported in the past. According to theory, the hydrocyclone with reduced Rf will improve classification of fines. A classification study using semi-inverted hydrocyclones was conducted to determine if a hydrocyclone could be operated in a way that would minimise misclassified fines. The classification performance of a semi-inverted hydrocyclone will be presented in this paper and other potential applications will be discussed. Keywords: Semi-inverted hydrocyclone, water split, misclassification of fines

Jan 1, 2020

By J. D. T. Steyl, G. B. Grobbelaar

Two examples are presented to show how first-principles – physical-dynamic, here termed behavioural modelling, is used to build digital twins of unit processes or plant sections. A digital twin equips the metallurgist, engineer, or manager with a powerful tool to conduct model-based design, solve process problems and improve operations. It is particularly relevant in the metallurgical industry where plant behaviour is highly integrated, rates respond to particle dynamics in complicated ways and mineral flows across flowsheet boundaries are variable. The two examples used to illustrate the behavioural modelling methodology are 1) a flotation column unit block and 2) a multi-compartment pressure oxidation (POX) autoclave. It simultaneously accounts for the causality of material and heat flow, takes care of mineral particle dynamics, physical and chemical reaction rates, the plant control system behaviour and operating heuristics. The basic modelling building blocks are briefly reviewed to show how the behavioural modelling attributes constrain the mass and heat flow dynamics to physically realistic values. This is the key aspect that allows project value to be created with minimal input data. Behavioural modelling is especially valuable where the process dynamics are poorly understood or controlled, where data-driven models fail to cope with complex and highly integrated plant behaviour, where quality data is lacking or limited funds are available to generate the data required to fill all the knowledge gaps. The paper uses the flotation column example to show how behavioural modelling and simulation can be used to optimise plant performance, while the multi-compartment POX autoclave model shows how it can be used to diagnose process problems. It also provides a solid basis to add a data-driven component – in other words, it makes a strong case for a hybrid first-principle– data-driven approach to solve process problems, especially where microscopic-level understanding is poor. Keywords: Digital twin, behavioural modelling, simulation, flotation column, pressure oxidation

Jan 1, 2020

By S. T. Johansen, B. T. Løvfall, T. Rodriguez Duran

In this paper we develop a physics-based model for lining erosion in steel ladles. The model predicts the temperature evolution in the liquid slag, steel, refractory bricks, and outer steel shell of the ladle. The flow of slag and steel is due to forced convection induced by inert gas injection, vacuum treatment (extreme bubble expansion), natural convection, and waves caused by the gas stirring. The lining erosion takes place by dissolution of refractory elements into the steel or slag. The mass and heat transfer coefficients inside the ladle during gas stirring are modelled based on wall functions which take the distribution of wall shear velocities as a critical input. The wall shear velocities are obtained from computational fluid dynamics (CFD) simulations for a sample of scenarios, spanning out the operational space, and a model is developed using curve fitting. The model is capable of reproducing both thermal and erosion evolution well. Deviations between model predictions and industrial data are discussed. The model is fast and has been tested successfully in a ‘semi-online’ application. The model source code is available to the public at [https://github.com/SINTEF/refractorywear].

Apr 16, 2024

By H. S. Sichel

This study deals with the development of some new theory concerning the bias which will result from the employment of the t-estimator if in fact the underlying distribution of ore values is not lognormal. It is found that the t-estimator is robust as long as the logarithmic variance is small and the shape factors of the transformed distribution are not unduly large. The choice of the threshold parameter [t] for the t-estimator is important, and some hints are given how to estimate it if the sample size is small, as en- countered in borehole valuation problems.

Jan 1, 1987

"The management of outlier grades in positively skewed gold distributions is a contentious issue. Incorporating outliers in standard ordinary kriging (OK) estimation procedures in a way that honours the data without smearing extreme grades into surrounding areas has been problematic. Cutting or capping of outliers to mitigate their influence in estimation techniques is common practice, while methods that manipulate the OK system of equations fail to honour the data. We propose a method of post-processing of kriging weights that provides realistic OK estimates and mitigates smearing without manipulating kriging equations or changing the original grades. The method requires that the data is not clustered, is approximately equally spaced, and is of the same support. Positively skewed data is ordered on attribute grade and nonlinearly transformed to a Gaussian histogram of categorical bins whose frequency is based on their likelihood of occurrence and location in the sample distributions. Factors that restrict kriging weights are calculated by dividing the percentage frequency of data in each bin by the percentage frequency of data in the bin with the highest frequency. Restriction factors applied to the kriging weights in the OK estimation restrict the range of influence in proportion to their probability of occurrence in the distribution. Smear reduction post-processing is easy to implement and addresses issues arising from negative kriging weights while considering the spatial location of samples, the sample grades, and their probability of occurrence. The method mitigates both smoothing and conditional bias.IntroductionOutliers are sample observations that deviate considerably from the standard or expected (Hawkins, 1980). These deviations can be high or low in grade depending on the distribution. The presence of outliers in a mine sampling campaign could provoke mixed reactions. Mining engineers might view high-grade outliers as a promise of achieving higher grades, while the geostatisticians view outliers as affecting the accuracy and precision of the overall estimate, among other issues. Having identified outliers, the geostatistician must decide how to account for their presence and treat them so that true underlying estimates are not distorted. Advanced knowledge of the deposit allows the geostatistician to identify and treat outliers acceptably, especially where samples are limited in number and the true grade distribution is uncertain. If ignored, or treated incorrectly, outliers are likely to lead to smearing of extreme grades into the surrounding estimates, thereby steering regions towards potential under- or over-estimation."

Apr 1, 2019

By P. N. Erasmus

This paper considers the fundamental concepts of support design for underground openings in highly stressed ground. It concludes that the conventional deformed reinforcing bar, woven mesh and cable lacing support methods that have been widely used in deep gold mines are quite in appropriate for the support of highly stressed tunnels. The following is suggested as the criterion for adequate dynamic support under rockburst conditions: the energy absorbed by a support unit (yielding through some acceptable limited displacement) should exceed the kinetic energy generated by the mass of rock that is intended to be retained by that unit. Duraset in conjunction with SRK Consulting has developed a yieldable steel tendon called the Durabar®. It is unique in the sense that it cleverly utilizes the property of the ductility of ordinary low-carbon steel from which it is made to accommodate the excessive displacement it is subjected to in the case of rockbursts or swelling ground.

Jan 1, 2009

By G. Chitiyo

Prediction of pothole characteristics is a challenging task, confronting production geologists at the platinum mines of the Bushveld Complex. The frequency, distribution, size, shape, severity and relationship (FDS3R) of potholes has a huge impact on mine planning and scheduling, and consequently cost. It is with this in mind that this study was initiated. Quantitative analysis of potholes indicates that pothole size (area covered) can be described by two partly overlapping lognormal distributions. These are referred to as Populations A (smaller) and B (larger). The range of observed pothole sizes conforms to a simple double exponential growth model based on Newton's Cooling Curve. A third size range of very large potholes (Population C) that could not be modelled properly within the proposed growth model is interpreted as to represent a very early phase of aggressive regional thermo-chemical erosion and potholing. In general, potholes are dominantly quasi circular with only a subordinate tendency of elongation. In the UG2, potholes with elongated forms are more prevalent in the size range between 20 and 500 m diameter. In the Merensky Reef elongated potholes are more common in the size range above 50 m in diameter. Although not distinctly visible, elongation of potholes in both the UG2 and the Merensky Reef show a net north/south orientation. Using shapes of pothole rims and floor areas, documented for the UG2 in the Waterval Shaft area, it is demonstrated that the northern, down-dip edges of potholes have steeper dips than the southern, up-dip edges. Spatial distribution studies using a uniform quadrant method suggest that in both the UG2 and the Merensky Reef, the potholes are randomly distributed, with a tendency towards clustering. Clustering appears to be more prevalent in the smaller Population A potholes. Bearing all findings in mind, it is found that the model of thermo-chemical erosion of the cumulus floor by new influxes of superheated magma best explains the observed data. Partial to complete melting of the cumulate floor occurred in three phases. The first represents the emplacement of hot magma. This magma, due to turbulent flow and high chemical and physical potential, aggressively attacks the existing floor (crystal mush on the magma/floor interface). Regional erosion is manifested by large, often coalescing potholes. During the second phase, when the magma emplacement process ceased and cooling in situ started, two distinct periods of pothole formation ensued. The first is related to rapid cooling along the relatively steep part of the Newton Cooling Curve, when Population B potholes nucleated randomly and grew rapidly with concurrent convective overturn and largely laminar flow condition. The second period of cooling occurred on the shallow-dipping part of the Newton Cooling Curve. Population A pothole growth became more subdued and nucleation appears to have been, at least locally, clustered. The final phase of this proposed 'super magma cycle' was introduced when chromitite crystallization and precipitation terminated pothole formation. This was followed sequentially by pyroxene and plagioclase crystallization to form the typical cyclic units of the Critical Zone. No suitable proxy could be found for the prediction of pothole density of potholes associated with the UG2 (likely due to limited spatial data coverage), and these are best predicted by extrapolation. The Merensky Reef slope index, in contrast, provides a proxy for pothole density, enabling prediction with reasonable confidence. This predictive model has been verified using underground information. The findings made during the course of this investigation have, when implemented, significant impact. Successful implementation will not only allow enhanced resource and reserve definition, but also better mine planning and scheduling.

Jan 1, 2008

By G. Akdogan, I. J. Geldenhuys, Q. G. Reynolds

Efficient recovery of titanium from titaniferous magnetite ores continues to be a promising, yet unrealised opportunity, despite significant occurrences of titaniferous magnetite deposits worldwide and the existence of established industrial-scale practices. Current smelting processes only recover iron and vanadium, while the titanium is discarded in the waste slag. These slags are stockpiled and present massive environmental challenges, while still containing value as residual iron, vanadium and titanium. The complex mineralisation of these deposits does not allow for a significant opportunity to separate the titanium and magnetite minerals through normal physical upgrading processes. An upgrading process is therefore required to derive economic value. An efficient process technology could create an opportunity to unlock all three valuable commodities. South Africa’s Bushveld Complex, the largest known deposit of its kind in the world, is known for the high vanadium content but is also associated with relatively high concentrations of titaniferous magnetite. The quality of the Bushveld Complex was one of the drivers for the establishment of the Highveld Steel and Vanadium Company in the mid-1960s. The unfortunate demise of Highveld Steel and Vanadium created an opportunity to review current practices and evaluate the best available process for this complex ore in the context of the three elements of interest. Past and current processes, the nature of these deposits and the potential to recover titanium are assessed to highlight the potential of fluxless smelting of titaniferous magnetite to produce a titania-rich slag as feedstock to the pigment, titanium metal industry in conjunction with recovery of iron and vanadium. Ilmenite smelting practices are proposed as a suitable alternative to achieve the goal, namely sustainable processing. Keywords: Titaniferous magnetite, titanomagnetite, smelting, titania slags, comprehensive processing

Jan 1, 2020

By M. Z. Naghadehi, E. Emami, J. Sattarvand, R. Battulwar

In this paper, a methodology for a computerised recognition of joint sets on 3D point cloud models of rock masses is presented. The process starts with classifying joints on a 3D rock mass surface through training a deep network architecture and validated using manually labeled datasets. Then, individual joint surfaces are identified using the Density-Based Scan with Noise (DBSCAN) clustering algorithm. Subsequently, the orientations of the identified joint surfaces are computed by fitting least-square planes using the Random Sample Consensus (RANSAC). Finally, the joint planes are classified into different joint sets, and the dip direction and dip angle for each set are calculated. The performance of the proposed methodology has been evaluated using a case study. The results show that the presented procedure is fast, accurate, and robust.

Sep 1, 2021

By S. Banisi, A. Ghasemi, A. Fazeli, M. Rezaei, M. Maleki-Moghaddam

The efficiency of high capacity SAG mills which consume large amounts of energy depends on the grinding and slurry discharge systems. A 9.7 m × 4.8 m SAG mill with radial pulp lifters is used at the Sarcheshmeh copper complex grinding circuit. The process audit indicated that the pulp discharge system was not efficient, judging by the amount of slurry flow back and carry over. These events limited the discharge rate and increased the wear rate of the lifters and grates. With the objective of finding a new design for pulp lifters, a 3D DEM software (KMPCDEM ©), along with the plant operational data and photographic evidence was used. The KMPCDEM © software was used to compare the efficiency of the radial and curved pulp lifter design based on the material flow. The slurry flow was analysed in terms of ease of discharge and lifter wear rate. It was found that changing the radial pulp lifters to curved pulp lifters could improve the discharge efficiency,and decrease wear rate of the pulp lifter in particular at the base of the pulp lifters. The simulation results showed that if curved lifters with a radius of 0.3×mill radius are used, the amount of slurry flow back and carry over will be negligible. It is predicted that by installing the curved pulp lifters, the capacity of the mill increases by 10% and the pulp lifter service life improves by at least 20%. Keywords: SAG mill, radial pulp lifter, DEM,curved pulp lifter, Sarcheshmeh

Jan 1, 2020

By M. A. Mahboob, T. Celik, B. Genc

Mineral Resources estimation plays a crucial role in the profitability of the future of mining operations. The conventional geostatistical methods used for grade estimation require expertise, understanding and knowledge of the spatial statistics, resource modelling, geology, mining engineering as well as clean validated data to build accurate block models. However, the geostatistical models are sensitive to changes in data and would have to be rebuilt on newly acquired data with different characteristics, which has proved to be a time-consuming process. Machine learning methods have in recent years been proposed as an alternative to the geostatistical methods to alleviate the problems these might suffer from in Mineral Resource estimation. In this paper, a systematic literature review of machine learning methods used in Mineral Resource estimation is presented. This has been conducted on such studies published during the period 1990 to 2019. The types, performances, and capabilities, of several machine learning methods have been evaluated and compared against each other, and against the conventional geostatistical methods. The results, based on 31 research studies, show that the machine learningbased methods have outperformed the conventional grade estimation modelling methods. The review also shows there is active research on applying machine learning to grade estimation from exploration through to exploitation. Further improvements can be expected if advanced machine learning techniques are to be used.

Nov 30, 2022