Search Documents

By U König, N Norberg

Traditionally, quality control of iron sinter has relied on time-consuming wet chemistry. However, the mineralogical composition that defines the physical properties, such as strength or reducibility, is not monitored.X-ray diffraction (XRD) analysis in combination with Rietveld quantification and statistical data evaluation using partial least squares regression (PLSR) has been successfully trialled to determine the mineralogical composition and process parameters such as the FeO (Fe2+) content and basicity of iron sinter within an analysis time of less than eight minutes per sample. The results were compared with wet chemistry data.PLSR does not require pure phases, crystal structures or complex modelling of peak shapes. The analysis is therefore easy to set up and ideal if only one property needs to be analysed. PLSR results show an accuracy for the FeO content of ±0.3 per cent for the samples analysed in this case study.Both the PLSR and Rietveld methods take the full XRD pattern into account and can be simultaneously applied on the same measurement. PLSR was found to be the more robust method if only FeO and basicity results are required. The Rietveld method helps to predict other parameters, such as the compression strength of the sinter, by monitoring all existing phases (eg larnite C2S or silico-ferrite of calcium and aluminium phases).CITATION:König, U and Norberg, N, 2015. Iron sinter process control using x-ray diffraction, in Proceedings Iron Ore 2015, pp 53–60 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Jul 13, 2015

The hoisting system at the Newcrest Telfer mine is designed for a final duty of 7.35 million tonnes per annum at a hoisting distance of 1450 m. In terms of hoisting capacity (t/h) times hoisting distance it is one of the largest friction hoist installations. Initially, the hoist will be equipped for hoisting 6.5 Mtpa from 1100 m depth. This new drive technology is now well proven by ABB for mine hoist applications with the first installation in commercial operation since July 2001. To date 17 systems for medium voltage drives and another five for low voltage drives have been purchased from ABB for mine hoists. This paper will also deal with skip design and powerful data loggers with diagnostics for local and remote monitoring that are installed in the Telfer mine hoist.

Jan 1, 2005

By Van Weert G

Ferrous chloride solution pyrohydrolysis is a successful IICI regeneration process which is carried out in either a spray roaster or a fluid bed. Pyrohydrolysis is not restricted to ferrous chloride only, every metal chloride can theoretically be hydrolysed. In practice, however, the conditions for ferrous chloride pyrohydrolysis differ from those of many non-ferrous metal chlorides eg zinc, lead and cadmium. Hence, a pyrohydrolyser could function as a ferrous/non-ferrous separator, eg an iron oxide and zinc chloride separator. Most pyrohydrolysers operate on solutions, in which case more than 50 per cent of the process energy is needed for heating and evaporating water. However, this disadvantage would be diminished by feeding a slurry to the reactor. Against this background it was decided to investigate the chloridisation of zinc ferrite which is an iron bearing solid with a non-ferrous impurity, and subject to processing and disposal problems. Synthetic zinc ferrite was chloridised in a tube furnace. The gas atmosphere in the tube furnace resembled that of a fluid bed pyrohydrolyser and consisted of HCI, 1120 and nitrogen. Gas composition, temperature, residence time and sample mass were varied. The experiments showed that complete (99.6 per cent) zinc extraction from synthetic zinc ferrite was possible in a simplified pyrohydrolyser atmosphere at 1100 K in four hours. The zinc ferrite was converted into gaseous zinc chloride, and hematite and magnetite solids. It was also found that zinc removal from the zinc ferrite lattice caused not only rccrystallisation of the zinc ferrite into hematite, but also hematite crystal

Jan 1, 1993

By K C. Corin, C T. O’Connor

In recent years it has been shown that the concept of using partial molar excess Gibbs energy (G̅ex) to explain the behaviour of collectors in the recovery of minerals can be usefully applied to explain flotation performance. Fundamentally this is because hydrophobicity of a mineral particle on which a collector is adsorbed may be strongly related to the fact that the alkyl chain associated with the collector behaves in the same way as any hydrocarbon would in an aqueous medium. This hydrocarbon-water system represents a classical non-ideal mixture, and such mixtures are quantified thermodynamically by determining the activity coefficient of the hydrocarbon moiety of the collector and hence its G̅ex in the mixture. It has been shown that G̅ex is a more rigorous indicator of the contributions of various alkyl groups of a collector than HLB (Corin and O’Connor, 2014) and was also able to explain aspects of froth behaviour (Harris and O’Connor, 2017) and hydrophobicity (O’Connor, 2021). The present paper is aimed at extending the above concepts in two different ways. Firstly, a study is undertaken of a range of results previously reported in which various reagents containing different alkyl groups were analysed in terms of their effect on contact angle and their performance as flotation reagents in terms of recovery. These results sustain the previous observations that there is a robust positive relationship between hydrophobicity as indicated by flotation recovery and contact angle and G̅ex of the alkyl group associated with the collector. Secondly, on the basis of the observation that increases in G̅ex are indicative of increases in the hydrophobicity of the mineral particle-collector aggregate, a study was undertaken to examine the effect of ions typical of electrolytes in water as well as the effect of the presence of alcohols and of pH on G̅ex of the collector alkyl group. In the former two instances the absence of any correlation between G̅ex and recoveries supported the proposition of the respective authors of these studies that the major effect of these electrolytes and added alcohols may be on particle–bubble attachment rather than on the hydrophobicity of the mineral/collector in the aqueous system. On the other hand, the effect of pH on G̅ex correlated reasonably well with changes reported in the literature on the effect of pH on recovery and contact angle.

Aug 24, 2022

By E Nettleton, A Melkumyan

Identification of geology and mineralogy in situ on the mine face is essential if mining processes are to be automated for tasks such as exploration, grading and reconciliation. Hyperspectral data acquired from field based platforms are ideally suited to these tasks, and are set to revolutionise the automated mining industry. Changing conditions of solar illumination introduces variation into the hyperspectral signal, which is unrelated to geology, complicating analysis of data. Thus, algorithms must be able to deal with changing illumination conditions and operate in a consistent and probabilistic framework, which is a prerequisite for data fusion with other sensors. To produce accurate classification maps of geology, algorithms must be developed that are effective under different conditions of illumination.The efficacy of three algorithms to classify hyperspectral data for geology mapping on vertical mine faces was assessed. These algorithms are the spectral angle mapper (SAM), and two machine learning methods operating within a fully probabilistic Gaussian process (GP) framework û the squared exponential (SE) and the observation angle dependent (OAD) covariance functions (kernel). Two experiments were done to assess the classification performances. The first determined the performance of the algorithms by classifying hyperspectral data of known rock samples acquired using a field spectrometer. The GP-based methods had a similar performance, with average accuracies and F-scores of 97 and 88 per cent (GP-OAD) and 95 and 81 per cent (GP-SE), respectively. Average accuracies and F-scores for SAM were significantly smaller (80 and 40 per cent respectively). The algorithms were then applied to hyperspectral imagery acquired from a vertical mine face at different times of the day, under very different conditions of illumination. In contrast with experiment 1, SAM and GP-OAD gave results which were superior to GP-SE and were entirely consistent with geology mapped in the field. Compared with SAM, GP-OAD was less sensitive to the effects of shadow and produced classified images that were less noisy. The GP-SE method failed to delineate even gross changes in geology on the mine face. These results show conclusively that in order to use independent training data (spectral library) to classify imagery acquired under a different set of conditions, algorithms must take into account variability in incident illumination. Measures based on spectral distance are inferior to those based on spectral angle.The OAD kernel within the GP framework (GP-OAD) is superior to the other two methods because:it provides a probabilistic framework;it is insensitive to variability in illumination within and between imagery acquired under different conditions; andit can identify and map geological zones on a vertical mine face using an independent spectral library, without a priori knowledge.

Sep 26, 2011

By Golding D

The Charters Towers goldfield is located in the centre of a Lower to Mid-Palaeozoic igneous complex where schistose, amphibolite-grade metamorphic rocks were intruded by multiple phases of medium grained granodioritic rocks. Several generations of mafic and felsic dykes also occur. Two distinct mylonites cut all but the youngest granitoids. Gold mineralisation occurs in quartz veins located in brittley reactivated mylonite and in sympathetic fissures in the undeformed granitoids. The gold-quartz veins occur in complex, gouge-filled, 2.5m thick zones, with minor sulfides, surrounded by zoned phyllic and propylitic alteration assemblages. Four main vein systems are each restricted to separate granitoids or contain oreshoots near granitoid contacts. No genetic relationship between the granitoids and the vein systems can be documented. The early mylonitic fabrics and the granitoid contacts played an important role in ground preparation and in the localization of oreshoots. Non mylonitized granitoids also served as favourable hosts for many of the lodes. Whole rock values of 180 of the granitoids range between 6.5 and 7.9 per mil. Hydrothermal vein quartzes have 180 values between 12.2 and 15.7 per mil. Mylonite quartzes have values of 8.4 per mil. Variation in 180 values of quartzes is in part related to granitoid host rock type.

Jan 1, 1987

A field trial of the prompt neutron-gamma SIROLOG technique was undertaken by the CSIRO - Division of Geomechanics at Callide Coalfields during July 1989. The trial investigated whether the sonde could reliably be used for the determination of Ash and Iron Oxide (Fe2O3) in coal seams, which had a highly variable iron content. Ten boreholes were logged, being water filled of a nominal diameter of 120 mm. The logging probe employed a 1.6 ¦g 252Cf neutron source and a 51 x 51 mm BGO detector. The source to detector distance was 15 cm. The results from one exploration borehole at the Dunn Creek Mine are discussed in the paper. The rms deviations for ash and iron determination between the neutron-capture predictions of the geophysical logs and the laboratory - chemical assays were 1.9 per cent Ash and 0.88 per cent Fe respectively. Using existing empirical equations which relate iron oxide concentration in coal ash to Initial Deformation Temperature (reducing conditions) at Callide, it may be possible to indirectly determine ash fusion properties as least as accurately, on a more detailed ply-by-ply basis, and more cost competitively than current laboratory techniques allow.

Jan 1, 1991

By G M. Zhomin, A S. Arkhipin, A Kondratiev

Viscosity of multicomponent slags is one of the crucial physicochemical properties for many industrial processes and technologies. While the viscosity-temperature relationship of completely molten slags can well be described by an Arrhenian-type law, the compositional dependence of viscosity is still not fully understood, especially in compositional regions with the lack of experimental data (eg at the glass-former concentration of 90 molar per cent or higher). In the present work major compositional relationships of the activation energy of viscous flow have been revised for numerous binary and ternary slag systems on the basis of experimental data collected and stored in the viscosity database OxiVis. For example, it has been found that the activation energy relevant to the polymerisation effect in binary systems of the MO-GFO type, where MO is the modifier oxide (ie monovalent K2O, Na2O, Li2O and bivalent CaO, MgO, BaO, MnO, PbO etc) and GFO is the glass-forming oxide (SiO2 or B2O3) is not always proportional to the third power of XGFO (contrary to the Urbain formalism) and can vary significantly dependent on the slag system. Also, the activation energy of the charge compensation effect has been revised for ternary systems of the AO-MO-GFO and AO-MO type, where AO is the amphoteric oxide (Al2O3, Fe2O3 or other). A simple polynomial model has been proposed on the basis of the carried-out analysis for estimation of viscosities of multicomponent slag systems. The model has been optimised against the experimental viscosities in the selected binary, ternary, and higher-order oxide systems and has been shown reasonable agreement with the experimental data.

Aug 21, 2024

By Q Qian, M Karakus, J Wang, N Melkoumian, D Goodchild, Q Sui, C Wang

Probabilistic analysis is widely used in slope stability assessment because of its capability to take into account uncertainties associated with geotechnical parameters. There are, however, still many unresolved practical issues to the application of the method. One of these issues is the integration of assessments when multiple sets of fractures are considered for the entire slope. In the present study, this issue is addressed within the framework of Monte Carlo simulation. Probability of failure for pairs of fracture sets are assessed independently and the final assessment of the probability of failure for the entire slope is done using conditional probabilities. Prominent Hill open pit, in particular, the problematic fresh Bulldog shale formation, was used as a case study to test the method proposed. The SWEDGE program of the RocScience software package was used for the analysis.

Sep 26, 2011

The Nickel Rim Depth Cu-Ni-PGE deposit, discovered by Falconbridge Limited in 1993, illustrates several trends in exploration in the Sudbury area. Recent discoveries tend to be deep, high grade, footwall hosted, close to or straddle property boundaries and made using a multi-disciplinarian exploration approach that integrates economic considerations with geology, metal geochemistry and geophysics (borehole EM). The Nickel Rim deposit occurs down dip of Inco's Victor Deep deposit (4.2 million tonnes at 2.25 per cent Ni, 7.30 per cent Cu) in the eastern margin of the Sudbury Igneous Complex (SIC) at a depth of 2650 m. Before commencing drilling an economic evaluation was done to determine the minimum size needed to be economic at the expected depths and determine the expected value of an exploration venture. The subsequent surface exploration program discovered a drill indicated, undiluted mineral inventory of 1.6 million tonnes of 1.58 per cent Ni, 10.13 per cent Cu, 4.17g/t Pt and 3.50 g/t Pd including 519 000 tonnes of 4.13 per cent Ni, 25.53 per cent Cu, 9.76 g/t Pt and 8.42 g/t Pd. The deposit is hosted by gneiss, migmatite, gabbro and pseudotachylite (Sudbury Breccia) 100 m below the base of the SIC.

Jan 1, 1996

By A Kobzev, C Robben

Sensor-based particle sorting of gold ores is the masterclass of preconcentration. The technical challenges of a grade in ppm scale makes a direct detection with currently available technology in milliseconds impossible. Consequently, the detection and classification of particles into ore and waste classes relies on detectable proxies which are the foundation for the classification. These must be reliable and repeatable across the ore bodies processed in future. A sensor-based particle ore sorting equipment must therefore be calibrated systematically by understanding the gold deportation in different lithological groups. Also, it must be in best operational state to minimise possible misplacement. The technical challenges are met by a stringent financial regime governing the business case. Many gold ore sorting projects require a stable grade of gold in the rejected waste significantly below 1gr/t. Two examples are introduced in this paper:  The pilot results and operational experience of X-ray transmission (XRT)-based particle ore sorting for pebble sorting of sulphide associated ore;  Laser-based particle ore sorting for pebble circuit of an orogenic, quartz associated gold ore. Usually, sensor-based particle sorters are installed in the crushing section of a plant. However, for large capacity plants, the number of equipment required, and the associated material handling costs result in significant capital and operating expenses for plants with 1200tph run-of-mine capacity. As a conclusion, it was decided to investigate sensor-based particle ore sorting applied to the pebble stream of two existing semi autogenous mills.

Nov 10, 2020

By Ward B

Central Colliery was developed to exploit large underground reserves in the German Creek lease. It was planned from the outset as a retreating longwall operation and is the first longwall mine in Queensland. Longwall mining is not new to Australia but all previous experience has been confined to the coalfields of New South Wales. There was no precedent for longwall extraction in the markedly different geological setting of the Bowen Basin in Central Queensland. Various investigations were thus carried out during the initial development of Central Colliery, prior to the installation of the longwall, to ascertain the geotechnical characteristics of the coal seam and the adjacent strata and to predict the performance of the heading and pillar design as mining proceeds to depth. Laboratory tests showed the coal to be weak with an unconfined compressive strength of 9.6 MN/m2. Site inspection and geophysical correlation revealed a shear zone and extensive intraformational shear planes in the seam. The latter were found to exert a controlling influence on rib reaction in the headings. This behaviour was predicted by finite element modelling of the development headings which indicated that mining induced stresses in the coal would be relieved by block movement of the ribs. A chain pillar width versus depth curve was established for yielding pillar design based on empirical and computational methods. This gave pillar widths from 25m to 48m over the proposed depth range of 150m to 300m. Finite element modelling of the 25m pillars chosen for the first gateroads confirmed the overall stability of the design but suggested that rib spall could be significant in the adjacent tailgate. (1) Chief Mining Engineer, Capricorn Coal Management Pty. Ltd., German Creek Mine, Queensland (2) Senior Engineering Geologist, Coal Division, The Shell Company of Australia Limited, Melbourne

Jan 1, 1986

By D Weatherley, T Baumgartl, L Huang, Q Shao

The stabilities of rehabilitated landforms are often threatened by the altered surface hydrology, especially the excessive surface run-off, which may cause severe water erosion and associated pollutant transport to surrounding areas. These effects would lead to pollution in waterways and render the failure of newly established vegetation. An accurate quantitative evaluation of run-off problems are consequently essential for the proper rehabilitation designs which can help minimise these harmful consequences. For this purpose, in this study a novel run-off model, RunCA, was developed based on Cellular Automata (CA), integrating fundamental hydraulic principles. In this model, a series of rectangular cells were employed to cover the study area. Run-off production within each cell was simulated by determining the cell state (water surface elevation) that consists of both cell altitude and water depth. The distribution of water flow among cells was determined by applying CA transition rules based on the improved minimisation-of-difference algorithm and the calculated flow travelling time. RunCA was applied to various aspects in mined land rehabilitation planning, including landform designs for both hill-slopes and tailings dams, revegetation strategies and cover constructions. Simulation results showed that surface run-off increased with slope gradient and more run-off was produced on convex and straight slopes than on concave and S-shape slopes. Both mounds and rip lines constructed on the tailings dam surface could help reduce the water discharge. Vegetation planted in rows were more effective than those in columns or patches in controlling run-off. Increased cover thickness had the potential in reducing run-off rates, and the covers with an increasing thickness from top to bottom of the slope tended to be more efficient than those with a constant thickness. All these results have demonstrated the efficacy of RunCA as a simulation tool in assessing the hydrologic behaviours of different rehabilitation designs. Compared to the traditional hydrologic models, most of which were developed for the catchment scale studies, RunCA has the potential advantages in capturing the dynamic and spatially varied run-off behaviours at different spatial and temporal scales.CITATION:Shao, Q,?Baumgartl, T, Huang, L and Weatherley, D, 2014. A cellular automata-based run-off model and its application in mined land rehabilitation designs, in Proceedings Life-of-Mine 2014 , pp 81–96 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Jul 16, 2014

Statistical frequency distributions may be used to characterize the overall variation of grade in orebodies. Special types of distribution (e.g. lognormal, log-type A) are necessary to describe skewness inthe case of rare metals. Also needed (but for the most part hitherto neglected) is the serial correlation function which describes the tendency for grades near together in an orebody to be closer together than grades at points relatively far apart.The sampling problem for orebodies is interpreted in statistical terms. Different methods are reviewed, and the effect of serial correlation discussed.A discussion is given, based on experiments and observations made by the author, of biases arising in the minei:l material by such methods as chip sampling and grab sampling due to such causes as nonrepresentative selection by the sampler.INTRODUCTIONMiners have always recognized the large element of chance in their operations. Chance is the raw material of statistical theory and, provided there is some recognizable pattern or underlying regularity in a set of data arising from the operation of chance events, statisticians are able to formulate mathematical models to study them. Mining engineers, geologists, and statisticians have gone a long way between them towards finding general statistical rules which describe the distribution of minerals and metals in orebodies, and even the distribution of orebodies over the earth's surface.These rules do not have the force of physical laws, though they obviously derive from them and have been used to throw light on them. From statistical theory it is possible to find methods of estimating from sampling some of the characteristics of orebodies which are of importance to...

Jan 1, 1962

By L Lachance, F Flament, D Leroux, S Gariépy

The pervasiveness of poor sampling practices in and around mineral processing plants could seem intriguing in view of the various documented evidences both for causes and consequences. To this extent, this paper revisits the well-documented but lesser-known fact that failure to care sufficiently about how samples are selected and managed is the mother source of almost all sampling biases. The authors believe that the potentially huge leverage between causes and consequences as well as the diversity of people involved in sampling call for a globally integrated solution rather than for a point solution. Such a globally integrated solution has already been put forward by the AMIRA P754 Metal accounting – code of practice and guidelines (AMIRA, 2007) that inherently requires accuracy to be targeted, measured, reported and managed. Furthermore, always in order to avoid a point solution, several layers of protection against biases must be put in place with each one involving a different ownership whenever possible. This is exactly the role intended by the supervisory bias detection method described in this publication. It is based on the continuous monitoring of the expected value of reduced residuals obtained from statistical data reconciliation conducted in a metal accounting context. This supervisory bias detection method is developed in full details after which Monte Carlo simulations are done on a mineral processing plant example for showing how it can be successfully applied for detecting hidden biases.CITATION:Lachance, L, Leroux, D, Gariépy, S and Flament, F, 2014. Detecting sampling biases in metal accounting, in Proceedings Sampling 2014 , pp 109–120 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Jul 29, 2014

By P J. Sack, L V Danyushevsky

"An EXTENDED ABSTRACT is available for download. A full-length paper was not prepared for this presentation. Regional-scale exploration for sediment-hosted gold deposits is hampered by the challenge of evaluating the fertility of ancient sedimentary sequences that host these deposits. Recent developments in laser ablation inductively coupled plasma mass spectrometry allow for the evaluation of the gold fertility of sedimentary sequences using chemistry of early-formed sedimentary pyrite. We present a case study from the Selwyn basin area of Yukon, Canada, where significant sediment-hosted gold exploration has been ongoing for the last five years. Data from sedimentary pyrite in some Neoproterozoic shales are anomalous in Au and As compared to a global data set of unmineralised sedimentary pyrite. If at least some of the gold and associated trace elements in sediment-hosted gold deposits are derived from recrystallisation of early sedimentary pyrite, it follows that strata containing anomalous sedimentary pyrite should be more prospective. Based on this premise, our data clearly show that Yusezyu Formation Neoproterozoic shales in the Selwyn basin area of Yukon are a fertile source rock for sediment-hosted gold deposits, while pyrite from Nadaleen formation shales are strongly anomalous in Hg and Tl, but not Au. We suggest that the area warrants continued exploration for sediment-hosted gold.CITATION:Sack, P J, Gregory, D D, Large, R R and Danyushevsky, L V, 2015. Using sedimentary pyrite chemistry to inform regional exploration for sediment-hosted gold deposits – a gold fertility case study from the Selwyn basin area, Yukon , in Proceedings PACRIM 2015 Congress, pp 69–76 (The Australasian Institute of Mining and Metallurgy: Melbourne)."

Mar 18, 2015

By A L. de Ambrosis, F M. Weir, M Fowler

Traditional kinematic and statistical analyses of geological structure for geotechnical engineering applications, while considered industry standard practice, have some limitations and unrealistic assumptions. An improved approach involves a discrete fracture network (DFN), a numerical model that explicitly represents how defects may be spatially distributed through the rock mass. The DFN technique utilises the structural data in stochastic analyses to give a range of possible models used for visualisation and stability analyses.The early development of the DFN approach was largely associated with modelling fractured hydrocarbon reservoirs and nuclear waste isolation programs. With recent improvements to computational power it has become recognised internationally as a valuable tool for modelling some geomechanical problems encountered when designing excavations in fractured rock masses. For tunnelling or slope stability projects DFN modelling provides visualisation of a rock mass and identification of likely failure mechanisms.This paper presents a suite of DFN modelling undertaken for the design of a tunnel within the Ashfield Shale unit of the Sydney Basin, Australia. The necessary input parameters and derivation of these, along with the importance of both an accurate engineering geology model and validation of the synthetic fracture network are demonstrated. Multiple realisations of the DFN were analysed to provide an indication of the number and volume of unstable blocks for the proposed tunnel geometry, enhancing support design. One fracture network realisation was sliced along multiple sections for use in finite element modelling.CITATION:Weir, F M, de Ambrosis, A L and Fowler, M, 2014. Discrete fracture network modelling for tunnel design in the Sydney Basin, in Proceedings 15th Australasian Tunnelling Conference 2014 , pp 219–226 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Sep 17, 2014

By M D. Taylor, G Duclaux, K Atkinson, L A. Fisher, A R. Latham, M F. Gazley, S J. de Beer, C M. Tutt

Geological logging by mine and exploration geologists involves the application of a logging code that requires a geologist to make a subjective decision to classify a particular rock unit. This will most likely result in inconsistencies in logging that will be difficult to monitor or assess over time. A potential solution to this problem is to have an objective method to determine the true nature of a given rock unit. This can be achieved using portable X-ray fluorescence technology (pXRF), as large geochemical data sets can be collected rapidly and at low cost. Plutonic Gold Mine (Plutonic), Western Australia, has a growing data set of more than 150 000 pXRF multi-element analyses from core and underground face samples, reconciled to Au fire-assay concentration, and logging codes. This study investigates and statistically validates the consistency with which the logging codes have been applied at Plutonic for the last four years, with respect to both Au distribution and other elements. In doing so, we have developed a schema that allows objective logging to occur once core has been sent for assay and analysed by pXRF. While far from replacing the geologist in the field, this is a complementary method that adds significant value to a data set.CITATION:Gazley, M F, Tutt, C M, Latham, A R, Duclaux, G, Fisher, L A, Taylor, M D, de Beer, S J and Atkinson, K, 2013. Objective geological logging using geochemical multi-element data - Plutonic Gold Mine, Western Australia, in Proceedings World Gold 2013 , pp 455-462 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Sep 26, 2013

By D J. Williams, A Shokouhi, A Naderian

Extracting alumina from bauxite involves the production of fine-grained waste known as red mud. The highly alkaline red mud is usually neutralised by applying sea water, thickened to between 25% and 30% solids by mass, before disposal to a Tailing Storage Facility (TSF). Hydraulic sorting, settling, self-weight consolidation, and desiccation are the processes contributing to the overall volume change of the red mud after deposition. Investigating the effect of these processes on the volume change behaviour of the fine-grained red mud is an important step in the estimation of the required storage volume of the storage facilities. This paper presents the results of a series of laboratory consolidation tests in a slurry consolidometer, and comparisons are made with field testing and observations.CITATION:Shokouhi, A, Naderian, A and Williams, D J, 2018. Assessing the volume change behaviour of red mud through laboratory and field testing, in Proceedings Mine Waste and Tailings Stewardship Conference 2018, pp 309–316 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Jul 23, 2018

By D Misra, R Ganguli, S Bandopadhyay

The objective of this paper is to assess the accuracy and applicability of recent state-of-art methods such as the machine learning algorithms (MLA) for mineral resource estimation. Conventional methods such as geostatistics have been predominantly used in the mining industry for this purpose with varying success. Recent advances in the use of MLA have provided a fresh approach to obtain improved accuracy in the estimation of mineral resources. Two MLA methods: the neural network (NN) and the support vector machine (SVM) have been used to estimate a polymetallic ore deposit in Alaska.The general characteristic of SVM and NN emphasises the fact that they can approximate any multivariate non-linear relation among variables. Although neural network models are able to capture the nonlinear spatial relationships that may be present in the data, they are usually difficult to optimise under sparse data settings. Of the various NN alternatives, despite their effectiveness, the model selection and estimation process is typically difficult as it involves solving complex integration and optimisation of parameters. In support vector modelling (SVM), estimating of parameters involves optimisation of a convex cost function. The working principle of SVM makes it robust against noisy and extreme value data. At the same time, it can capture the spatial distribution of ore grade more effectively with careful modelling and selection of SVM parameters.The outcome of the MLA methods were compared to those of the geostatistical ordinary kriging (OK) application to study their generalisation ability. The comparison was made using the following goodness of fit criteria: mean-squared error; mean absolute error; mean error; and coefficient of determination. The results indicated that the MLA methods may improve the predictability of ore deposits and thereby reduce the inherent risk in mineral resource estimation.

Sep 26, 2011