Search Documents

By K Williams, T Donohue, W Chen, T Roessler, A Katterfeld

Wet sticky iron ore poses particular difficulties in every aspect of the material handling process. Material hang-ups in bins, pipes and chutes; carry-back and spillage due to insufficient belt cleaning; clogging of screens; malfunctioning of crushers as well as rat holing of gravity reclaimed stockpiles are all predominantly due to the inherent cohesion and adhesion characteristics of the wet, sticky ore (Roberts, 1993). Limited understanding and quantitative measures on the stickiness of such a material significantly lower the mining operation efficiency.This research aims to investigate the cohesion and adhesion characteristics of the wet, sticky ore on physical modelling, experimental techniques and numerical implementation. Initially, the suitable bridging mechanism that contributes to the cohesive and adhesive forces within the iron ore material was identified (Hering, Martin and Stohrer, 1989; Rumpf, 1958). A cohesion/adhesion energy density factor based contact model was then introduced to describe the stickiness of the material. Secondly, an experimental system was proposed to directly test the stickiness of the material. Such a system is able to obtain both the cohesive/adhesive shear force and tensile force of the material at a specific moisture content. Lastly, the implementation of the proposed cohesion/adhesion model into the discrete element modelling (DEM) software was also discussed. A cohesion/adhesion calibration for modelling wet sticky iron ore in DEM was then conducted to ensure the material modelled in the simulation is reflective of the reality.From the study, understanding of the cohesion/adhesion characteristics of the wet iron ore material is significantly enhanced. Experimental and numerical outcomes can be directly applied for better designing or optimising the material handling processes that involve wet sticky iron ore. Consequently, the mining operating costs can be greatly reduced.CITATION:Chen, W, Donohue, T, Williams, K, Katterfeld, A and Roessler, T, 2015. Modelling cohesion and adhesion of wet, sticky iron ores in discrete element modelling for material handling processes, in Proceedings Iron Ore 2015, pp 387–392 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Jul 13, 2015

By S. T. Johansen, C. Colomb, B. Ravary

Combustion and thermal NOx formation have been modelled using the Computational Fluid Dynamics (CFD) commercial code Fluent 6.1. The aim was to predict and give a better understanding of the combustion in electric arc furnaces for the production of ferro-silicon and silicon metal. Another important objective was to assess the impact of a number of operating parameters on the NOx formation. The CFD tools used are de-scribed with a special focus on models for species and turbulence transport. Sensitivity studies have been carried out to study the influence of several parameters: sub-models, process gas composition and distribution at the charge surface and influence of the presence of SiO. The main conclusions are: ?Using the combination of a realistic design, boundary conditions and models, the simulated NOx level is in quantitative agreement with measurements. The presence of water vapor is particularly important because of its influence on the specific heat of the mixture and resulting temperature. ?Release of significant amounts of SiO from the charge surface results in an important increase of NOx, even if we applied a simplified combustion model for SiO gas. However, because of lack of thermal and compositional data for blowing situations (when SiO flows directly from the arc cavity through open channels to the charge surface), no quantitative validation is possible. The way SiO is distributed (uniformly or as a jet) has a limited impact on the result in our computations.

Jan 1, 2007

By Catherine Johnson, Barbara Rutter

The shock tunnel located at the Missouri University of Science and Technology Experimental Mine is 65 feet 9 inches (20.0406 m) in length. The tests performed inside of this tunnel are either arena tests or shock interaction and transfer (SIT) tests. Arena tests simulate a large open air blast, with lower explosive weights, by elongating the impulse of the blast wave. SIT tests are conducted to determine how the produced shock wave interacts and transfers its energy into a target object located at the end of the tunnel. The target used in this research is multilayered, with each layer being a different material. The targets are 2 ft. by 4 ft. (0.6096m by 1.2192m) pieces of Medium Density Fiberboard (MDF), DOW polystyrene foam insulation, and gypsum drywall. The order of these materials varies with each test. This shock tunnel is unique, because the geometry changes along its length. To fully understand how the shock waves are affected by these changes, a working model was needed. The second purpose of this model is to confirm the maximum charge size of 400 grams (14.11 oz.) of C4 that can be detonated inside the tunnel without causing damage to or failure of the tunnel, thus reducing the amount of time needed in downtime and repair. The model predicts pressures in various locations both within the tunnel and the attached target, thus assisting in determining where failures would likely occur. In the process of generating this model, several problems were encountered, including filling the tunnel with air, meshing the complex geometry, and creating the tunnel in a CAD system. This paper will discuss how these problems were solved. A working model was created by using ANSYS Explicit Dynamics and ANSYS Autodyn. The tunnel geometry file was created using SolidWorks and includes the complex geometry changes along with the target and its mount located at the end of the tunnel. The results produced by the model were compared to the field SIT tests. The comparison of the results from both the model and the experimental testing showed a similar trend in the pressure traces.

Jan 1, 2018

This paper presents a case study where Cenozoic age detritals and an ancient buried landslide have been modelled in the final walls of an iron ore pit. Orebody 25 is located in Western Australia’s Pilbara region and is an operating pit where a redesign of the final wall cutback was required. The existing slope design utilised a simplistic geological model of mixed conglomeratic, calcretic and clay –dominated detritals overlying deformed bedrock sequences. This simplicity resulted in a conservative design that was not reflective of the slopes performance to that point. Further field investigation was carried out with the results assessed in conjunction with historic drilling, mapping and laboratory testing. A detailed geotechnical model was developed including three conglomerate and two clay dominated units and laterstage calcrete overprinting. The clay units were of significant importance to the stability assessment and resulting slope design due to the thickness and poor strength conditions indicated by the laboratory and field testing. However, the drilling investigation shed only limited understanding as to the origin of the clays and thereby how they should be treated as a component in the stability analysis. A comparison of the ratio of various geochemical elements measured by assay was made between the clays and 1) shale bedrock and 2) clays of known detrital origin from other nearby sites. This indicated the clays had a strong affinity with weathered bedrock. This and other textural features resulted in the interpretation that the clay units were relic debris preserved following ancient landslides in bedrock shale. Using laboratory testing data and calibration of results for a landslide origin, stability analyses were run, resulting in a steeper slope design being implemented. Excavation of the final cutback is now in progress with the model confirmed as appropriate.

Jan 1, 2015

By T Ma, F T. Suorineni, Q He, J Oh

Directional hydraulic fracturing was invented by Polish experts and was first applied in coal mining to prevent hard roof rock bursts. Based on its successful application in coal mining, directional hydraulic fracturing has the potential to reorientate the hydraulic fracture trajectory. Recently, this technique has been introduced into cave mining for the purpose of lowering breakdown pressure. The hydraulic fracture is much easier to initiate with the assistance of an initial notch. No systematic theoretical or experimental studies have been carried out to investigate the influencing factors in the hydraulic fracture reorientation process. In this paper, the numerical modelling software Realistic Failure Process Analysis3D (RFPA3D) – Flow Version is used to study the influence of the in situ stress condition and rock mass heterogeneity on directional hydraulic fracturing in brittle heterogeneous rock masses. Firstly, the reliability of RFPA3D in simulating hydraulic fracturing is validated by comparing its results with experimental results and the simulation results from the extended finite element method. The hydraulic fracture trajectory is then observed in a three-dimensional stress environment. The results indicate that for the in situ stress condition, when the borehole lies in the ?2–?3 plane, only the difference between ?2 and ?3 has an important influence on hydraulic fracture reorientation. Furthermore, rock mass heterogeneity is a key factor that must be accounted for when fracture reorientation is required.CITATION:He, Q, Suorineni, F T, Ma, T and Oh, J, 2016. Modelling directional hydraulic fractures in heterogeneous rock masses, in Proceedings Seventh International Conference and Exhibition on Mass Mining (MassMin 2016), pp 369–384 (The Australasian Institute of Mining and Metallurgy: Melbourne).

May 9, 2016

By S Bhandari

A model for predicting dust plume dispersion from blasting operations using the Eulerian mathematical algorithm has been developed. This model considers atmospheric stability, and wind velocity and direction for computing dust concentrations at different distances from the blast. Software has been developed which simulates the movement of a dust plume at a given horizontal distance from the blast location and time elapsed after the blast. The concentration of dust is also given in vertical planes. It takes into account ground profile and atmospheric conditions such as temperature, pressure, humidity, wind velocity, wind direction and shear. Experiments were carried out in two limestone mines to calibrate the accuracy of the predictions made by such model. High volume dust samplers were used to collect dust samples. Difficulties encountered in satisfactory dust sampling and quantification of results have been described.

Nov 8, 2011

By Gauti Asbjörnsson

Material handling is an essential part of the aggregate and mining industries. Mixing and blending of granular material can affect the grading and particle size distribution of the material leaving the system. Segregation and variability can cause problems in the downstream process while insufficient capacity will disturb upstream processes. Process plants utilize surge bins for handling recycle streams of material and act as a buffer for the process. The flow through the surge bin is usually controlled by a simple PID controller or PID controller in combinations with a plant-wide controller. Even though the surge bins play an important role of smoothening out the plant and creating a steady operating condition they are seldom included in plant simulations. This can lead to a number of problems in the form of decreased plant capacity, plant lockdown, and even operation units malfunctioning. The main objective of this study is to investigate the dynamic behavior in surge bins and develop a model more suitable for representing this behavior, which can occur during operation, for dynamic simulations. The proposed model adopts vertical layers which allow for flows between segments depending on the level of neighboring segments. This enables both the representing of segregation within the system and gives an indication on the effects from different inlet- and outlet placements. The proposed model was developed in MATLAB/Simulink and validated against an actual bin with the total capacity of 660 m3 in platinum ore application. With the new bin model significantly higher fidelity and more accurate dynamic, simulation results were achieved. Keywords: dynamic simulation, modeling, segregation, material handling, MATLAB/Simulink

Sep 1, 2012

By Y Wei, H-B Mnhlhaus

Particles are by far the easiest objects to simulate, because their dynamics is simple. Determining the motion of particles and their interaction, a wide variety of deformable structures can be constructed by connecting particles that have mass, position, and velocity, and that respond to forces. Starting from this simple idea we have derived what we call the particle-lattice method for the solutions of boundary value problems of deformation and fracture of solid mass. The problem of rock breaking by blasting operations as well as fracture and flow of rocks fragments involve momentum transfer between solid and fluid; gas flow and reactive mass transfer. In this paper we outline the method and describe briefly the salient features of the algorithm. We derive relationships between the contact properties for the discrete model and the properties of the underlying continuum and fluid flow coupling procedure. Finally we illustrate the potential of the method by means of example solutions related to the dynamic fracture and flow in jointed rock by blasting.

Jan 1, 1999

By J. W. A. Morris

To investigate the use of centrifugal fields in fused salt electrowinning cells, a high temperature rig has been designed, constructed and commissioned. Initially, the cell has been applied to the fused salt electrowinning of zinc. Current efficiencies of 93% have been obtained for inter electrode gaps as small as 8mm for 70mm diameter electrodes rotating at 100rpm. The construction of a semi-rigorous physical model has shown the potential exists to extend the use of centrifugal fields to the electrowinning of light metals.

Jan 1, 1994

By M. R. Jolly, J. -C. Gebelin

"A number of different types of so-called ""filters"" are used in the metal casting industries to impart some cleaning effect and flow control on the liquid metal as it passes through them. The filters range from simple planar meshes through extruded channels to reticulated foam structures and they are manufactured from a range of material going from metals to glass or ceramics. It is most common that software packages used in the industry model the filters by a simple pressure drop associated with some area fraction and permeability parameters.Recent experimental work at the IRC in Birmingham [1] has shown that filters of the same type can behave very differently depending upon the casting process in which they are employed. Modelling filter geometries for a range of different casting processes has indicated that the flow of metal and heat losses through the filters are rather complex and should be considered when using filters in the casting processes. This paper will present a number of cases of different types of filters modelled and different processes and indicate some of the sensitivities of the processes to boundary conditions imposed by the process.IntroductionFilters are commonly used in the foundry industry. The real action of the filters on the liquid metal is not really understood. Some filters are claimed to have a chemical action on the liquid metal or to be able to clean the metal by stopping the impurities or oxides that will result in inclusions in the cast part. However, some recent experiment carried out at the University of Birmingham in investment casting and results available on sand casting [2] have shown that the action of filters may be different in these two processes. As the main difference between these two processes is the mould temperature, the hypothesis of a partial solidification of the metal in the filter occurring in the sand casting process during the primary stage of the filling has been advanced, whereas there is little or no solidification in investment casting. In order to verify this hypothesis, and to reach a better understanding of how filters work, a numerical investigation of the flow and heat transfer in filters in both sand and investment casting has started at the University of Birmingham, in the framework of the FOCAST project. The FOCAST project is an EPSRC project aimed to increase the understanding of the investment casting process. Another issue is the way the simulation packages predict the influence of the filters on the metal flow. The filters are most of the time modelled as porous obstacles characterised by a pressure drop calculated from porosity and some surface ratios or proprietary coefficients. This paper will present the preliminary results of this numerical investigation. First, the different models that have been used are presented, followed by the obtained results. A discussion of these results and their perspectives conclude the paper."

Jan 1, 2001

By W Valery, S Morrell

In the mining industry blasts are usually designed to fracture the in situ rock mass and prepare it for excavation and subsequent transport. The run of mine (ROM) fragmentation is considered good when it is fine enough and loose enough to ensure efficient digging and loading operations. Mining optimisation strategy is hence usually focussed on minimising total mining costs and maintaining these ROM fragmentation characteristics. Although this approach ensures an efficient mining operation it ignores the potential impact on crushing and grinding. Investigations by several researchers have shown that designing blasts to produce ROM fragmentation to optimise crushing and grinding performance enhances the overall efficiency and profitability (Eloranta, 1995; Kojovic et al, 1995; Bulow et al, 1998; Kanchibotla et al, 1998; Scott et al, 1998; Simkus and Dance, 1998). Field experimentation in this area is often found to be difficult because of high implementation costs and insufficient understanding about the effect of fragmentation at different stages in a comminution circuit. Modelling and simulation of blasting and comminution processes provide a more economic alternative to explore the impact of blast design changes on the down stream operations (Kanchibotla et al, 1998; Kojovic et al, 1998; Adam and Siddall, 1998; Bulow et al, 1998). However, the reliability of such simulations depends on:the ability of the models to reproduce the physical processes that take place; characterisation of rock mass in terms of its blastability and comminution characteristics; and in the case of blast modelling the ability to predict fines (-10 mm) has been found to be particularly important as this fraction has a significant effect on mill throughput.

Jan 1, 1999

By A Tawadrous, B Wheeler, D S. Preece, S A. Silling

The two key results of rock blasting are fragmentation and movement of geomaterials. Movement/flow of the blasted rock is often referred to as heave. Moving rock with explosives rather than mechanical equipment can improve mining efficiency. Such is the case with coalmine cast blasting. Controlling and directing rock movement through explosive deck delay timing can greatly improve ore recovery by segregating ore and waste. In both cases, understanding, predicting and controlling heave is an important element of successful blasting.Computer programs for predicting 2D blasting-induced heave have been used in mining for decades, helping to drive the innovation of new and more efficient mining methods, ie cast blasting. The past few years have seen the advent of 3D distinct element modelling of blast-induced heave. This is an important advancement because most blasting situations are three-dimensional in nature, especially when controlling the direction and intensity of rock movement through explosive deck delay timing.This paper documents the first full-scale blasting simulations in 3D with distinct element models that include hundreds of blastholes and in excess of a million particles, made possible with parallel processing. Surface mineral blasting is the focus of this study which treats the effect of blasthole delay timing on ore recovery. Ore waste and dilution resulting from typical row-on-row delay timing will be demonstrated. Also demonstrated will be the ore recovery possible with delay timing designs that significantly enhance ore/waste segregation. This type of high fidelity modelling and the use of electronic detonators makes possible the sculpting of rock blasting muck into separate piles of ore and waste.CITATION:Preece, D S, Tawadrous, A, Silling, S A and Wheeler, B, 2015. Modelling full-scale blast heave with three-dimensional distinct elements and parallel processing, in Proceedings 11th International Symposium on Rock Fragmentation by Blasting, pp 127–136 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Aug 24, 2015

By S Khosrowshahi, A Weeks, W J. Shaw

"To provide an indication of expected performance of the mine and mill, feasibility studies should include an assessment of the expected day-to-day variability in process plant feed. Attempts to do this based on conventional resource models will be flawed as such studies fail to consider and account for the in-built smoothing (averaging) of input data.This can lead to one of two situations: an expensive, over-engineered process plant that can cope with every possible situation, or operational strategies that fail to manage the natural variability within the orebody and the impact of day-to-day operational decision-making.In both cases the value of the project diminishes through excessive capital costs or reduced throughput or recovery during operations.Unlike conventional estimation techniques, the geostatistical method of conditional simulation can reproduce the inherent variability of the deposit. Once a valid model has been developed, it is possible to analyse the model using a transfer function that simulates the proposed mining processes and provides a representation of the likely variation of mill feed characteristics at any time scale.The case study demonstrates how the proposed mining strategy would result in mill feed being considerably more variable than expected, resulting in poor recovery, low throughput and potential breaches of environmental conditions. Mitigation measures included budgeting for additional mining equipment to enable greater flexibility during mining and revised stockpiling and crusher feed strategies.Awareness of the benefits of modelling the variability of geometallurgical parameters will increase the use of this approach in producing more realistic financial models of expected mine performance at the design stage.CITATION:Shaw, W J, Khosrowshahi, S and Weeks, A, 2013. Modelling geometallurgical variability - a case study in managing risk, in Proceedings The Second AusIMM International Geometallurgy Conference (GeoMet) 2013, pp 247-252 (The Australasian Institute of Mining and Metallurgy: Melbourne)."

Sep 29, 2013

By Harald Ulrik Sverdrup, Anna Hulda Olafsdottir

The long-term supply of nickel to society was assessed with theWORLD7 model for the global nickel cycle, using new estimates of nickel reserves and resources, indicating that the best estimate of the ultimately recoverable resources for nickel is in the range of 650–720 million ton. This is significantly larger than earlier estimates. The extractable amounts were stratified by extraction cost and ore grade in the model, making them extractable only after price increases and cost reductions. The model simulated extraction, supply, ore grades, and market prices. The assessment predicts future scarcity and supply problems after 2100 for nickel. The model reconstructs observed extraction, supply and market prices for the period 1850–2020, and is used to simulate development for the period 2020–2200. The quality of nickel ore has decreased significantly from 1850 to 2020 and will continue to do so in the future according to the simulated predictions from the WORLD7 model. For nickel, extraction rates are suggested to reach their maximum value in 2050, and that most primary nickel resources will have been exhausted by 2130. After 2100, the supply per capita for nickel will decline towards exhaustion if business-as-usual is continuing. This will be manifested as reduced supply and increased prices. The peak year can be delayed by a maximum of 100 years if recycling rates are improved significantly and long before scarcity is visible.

Jan 26, 2021

Metsim is proving to be a powerful tool when designing a heap leach operation. The most useful outputs arc determining the fluctuations in pregnant solution recovery rate, make-up water rate, gold recovery rate. and pond levels. Detailed data are needed for input: a pad loading schedule, a mining rate schedule for each ore type, column leach test data for each ore type, column tests on water dynamics (ie hourly flow-out versus-in during saturation and drain-down), monthly rainfall and evaporation data, and proposed pond dimensions. These data allow 'what if' studies on application rates, leach rates, loading rates and other variables of interest. Examples will he given of modelling both a rectangular pad and a valley fill operation.

Jan 1, 1994

By J. B. Boisvert, W. Board, C. R. Mooney

"Traditional geostatistical modelling methods are difficult to apply to heavy-tailed deposits because the extreme positive skew of the grade distribution leads to bias. The focus here is on an innovative method of modelling the high-grade population by conditionally simulating discrete volumes (nuggets) of gold with marked Poisson and Cox processes. Capping and other ad-hoc manipulation of the assay data is not required; however, an intensity map for the point processes is required. The method is demonstrated on a dataset from an epithermal deposit and demonstrates an improvement in reducing smearing and controlling the skew of the grade distribution.RÉSUMÉ Les méthodes traditionnelles de modélisation géostatistique sont difficiles à appliquer dans des gisements dont la distribution de la teneur est à queue lourde car l’extrême asymétrie positive de la distribution de la teneur génère des biais. Dans le présent article, l’accent est mis sur une méthode novatrice de modélisation de la population à haute teneur par la simulation conditionnelle des volumes discrets d’or (pépites) au moyen de procédés marqués de Poisson et de Cox. La limite d’utilisation et d’autres manipulations ad hoc des données ne sont pas requises; toutefois, il faut une carte des intensités pour le processus ponctuel. La méthode est démontrée sur un ensemble de données d’essais provenant d’un gisement épithermal et elle montre une amélioration dans le lissage du flou et dans le contrôle de l’asymétrie de la distribution de la teneur."

Jan 1, 2015

By D. Hodouin

Jan 1, 1978

By M Campbell, A Jeffcoate, J-C Vannay, B Perriaux

Phyllite waste rock at the Simandou Pic de Fon and Oueleba Iron Ore deposits (Guinea, West Africa) variably weather to form clay-rich rocks. This weathering modifies the mineralogical, geochemical and mechanical properties of the phyllites. Final pit walls will largely be formed in phyllite, so it is critical that the slope design takes account of the distribution and variability of weathering within this material. Phyllite intact material strength has been correlated with geochemistry (ie mineralogy) facilitating the establishment of domains of intact rock strength (IRS). In order to calibrate this geochemically-based strength model, core from diamond drilling (DD), comprising a variety of phyllite material types (from soft clay to moderate strength rock), was geotechnically logged to assess intact strength; further validation was achieved by uniaxial compression (UCS) testing and point load tests (PLT). The correlation between geochemistry and strength offers a potentially powerful tool for differentiating basic mechanical properties of the rock mass, especially at Simandou where a significant proportion of the resource drilling is done by reverse circulation (RC). Unlike DD, reliable measurements of material strength from RC samples can prove difficult, especially in clay-rich material. Therefore, gaining information from routine geochemical assay results, independent of the sampleÆs physical integrity, is a valuable and cost-effective source of reliable empirical data. Data acquired using this approach provides an opportunity to develop strength domains, which are integrated into mine planning block models. In combination with other geotechnical characteristics, this can be a valuable input to the development of geotechnical models. This technique has allowed spatially representative rock strength data to be derived and offers the potential for optimising pit slope angles in areas where hitherto conservative design criteria had been assumed.CITATION:Jeffcoate, A, Campbell, M, Vannay, J-C and Perriaux, B, 2013. Modelling intact rock strength of clay-rich rocks with geochemistry at the Simandou iron ore deposit, Guinea, West Africa, in Proceedings Iron Ore 2013 , pp 193-202 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Aug 12, 2013

By Frédéric Flament, Mary Goldman, Gille Barbery

"Pilot plant testwork on autogenous and semi-autogenous grinding was conducted at the Centre des Recherches Minérales du Québec (CRM) to investigate the effect of ball addition, solids concentration in the feed, specific energy consumption and feed size distribution.A mathematical model that incorporates breakage and surface mechanisms of loss of mass was developed to predict load size distribution of autogenous and semi-autogeneous mills. An empirical model was used to predict product size distribution of autogenous and semi-autogeneous mills.IntroductionAutogenous (AG) and semi-autogenous (SAG) grinding circuits offer potential capital and operating cost savings when compared with conventional grinding. Empirical relationships have been developed to calculate specific power consumption for SAG mills using pilot plant test results (Mular et al., 1989; Barrett and Allan, 1986). Empirical models can be obtained from plant practice data but are applicable only to the particular ore and set of operating conditions used in the model. Grinding mechanisms have to be taken into account in order to develop a model that can be applied to different ores and mills.In AG and SAG, the ore itself acts as grinding media, with the finer sizes being crushed by the coarser ones. Surface mechanisms of loss of mass, in which the inner core of the particle remains intact, also take place. In SAG milling, the effect of balls also has to be taken into account. The role of each one of these mechanisms changes with particle size and the model has to account for such changes."

Jan 1, 1991

By P Dowd, A Adeli, C Xu, X Emery

Geometallurgical modelling is increasingly being incorporated into mineral resource modelling and estimation as a means of increasing efficiency, decreasing operating costs and reducing risk in mining operations. Traditional geostatistical workflows for geometallurgical modelling face several difficulties, in particular, related to the non-additivity of process response variables, to highly heterotopic data sets and/or to preferential sampling designs, all of which may explain why most attempts to estimate geometallurgical response variables from assay data and mineralogy are still based on statistics and machine learning. The aim of the work presented here is to demonstrate the applicability of geostatistics to estimate copper recovery from copper sulfide ores at the Prominent Hill Iron Oxide Copper-Gold (IOCG) deposit in Australia. As recovery is non-additive, the preferred workflow is the joint estimation of the masses of metal in the feed and in the concentrate, both of which are additive variables. While the mass of metal in the feed can be estimated from abundant online assay analyses, the mass of metal in the concentrate can only be estimated from a very limited number of laboratory-scale batch flotation tests. Traditional approaches to estimating the masses of metal in the feed and in the concentrate, such as simple and ordinary co-kriging, are described and discussed. A modified version of co-kriging is introduced, which incorporates linear relationships between the mean values of the input variables, in addition to a spatial correlation model. Co-kriging with related means is shown to outperform traditional simple and ordinary co-kriging in terms of precision, accuracy and consistency of the estimates, and offers a practical and efficient tool to deal with highly heterotopic and preferential sampling designs.

May 24, 2023