Search Documents

By Kavetsky A

In surface mining operations, the ease with which a muckpile may be dug is a dominant factor in mining efficiency. Within broad limits, this is largely dependent on muckpile shape and looseness within a muckpile. Moreover, prediction of grade boundary movement and subsequent dilu- tion is of interest to most open cut mines. Such a predictive tool would enable blast design changes to be made to reduce mixing of certain grade regions. At the very least, a good knowledge of post-blast region bound- aries can reduce layout of the grade boundaries and decrease grade dis- crepancies between pit and plant. The model presented in this paper is able to consider various options for a production blast and predict consequent muckpile shapes and grade boundary movements in a three dimensional manner. The overburden geometry, hole patterns, hole inclination and charge distribution are all explicit model inputs. The model divides the rock mass into blocks and calculates a velocity for each block. Three controlling parameters are used in the model, which are related to the explosive energy per unit length of blast hole, energy attenuation of the rock mass and inter-row delay, respectively. The output of the model is a 3 dimensional muckpile shape (contour or any cross section) and a contour map of grade distribution within the muckpile. The model parameters are site specific, since their values depend on the rock condition and explosive-rock interaction which varies from site to site. In order to "tune" the model to a particular site, parameter values must be found which reproduce an observed muckpile profile shape from a given blast design. This calibration of the model requires at least one surveyed muckpile profile shape from a production blast. The model parameters which best reproduce that shape are found by a pattern search technique, and may then be used to predict the effect of changes in blast design on muckpile profile shape and grade boundary movement.

Jan 1, 1989

By D R. Seaman, R P. Tooher, B A. Seaman

"Several mine scheduling tools exist for use in optimising and scheduling the reserve according to in situ value. These tools generally value the ore on a per block basis according to regressed recovery models based on ore grade and type. It is then the task of the plant metallurgist to ensure that these ore blocks are fed to the plant in a manner that satisfies the prevailing plant constraints. This task is often completed in complex Excel® spreadsheets that are modified over time with factors applied to account for plant modifications or limitations.It is the authors’ experience that this approach is difficult to audit and replicate in a consistent manner. Furthermore, operational decisions, such as primary grind size or autoclave oxidation are made on the metallurgist or managements’ perceived optimum value and does not necessarily represent a cash flow based optimum.Many processing plants contain multiple internal constraints. As ore feed grades and properties change over the life-of-mine (LOM), processing-bottlenecks shift from one unit process to another and different operating strategies need to be adopted to maximise cash generation from the process.This paper presents a unified optimisation approach to scheduling ore based on a circuit mass balance to describe the process plant with individual models for each significant unit operation. The approach allows for optimisation over multiple input variables and the inclusion of linear and non-linear constraints.This optimisation process has been applied to different complex gold operations. The end result is an automated scheduler that selects the optimum blend of ores as well as plant operating conditions that satisfy all internal plant and shutdown constraints while maximising the objective of choice (metal production or cash flow). The models are also valuable for evaluating expansion or debottlenecking opportunities of individual processing units. These models are coded in Matlab® R2016a provided by MathWorks®, and can be deployed to end users in a compiled version with significant savings in time to run different scenarios over the Excel®-based systems.Citation:Seaman, D R, Tooher, R P and Seaman, B A, 2016. Optimisation of Complex Gold Processing Plants with Multiple Unit Operation Constraints - the Newcrest Planning Scheduler, in Proceedings 13th AusIMM Mill Operators’ Conference 2016, pp 191–198 (The Australasian Institute of Mining and Metallurgy: Melbourne)."

Oct 10, 2016

By B Lance, D O’Brien

The Hidden Secret ore body is located within the Mt Charlotte underground mining system in Kalgoorlie, Western Australia. Mineralisation of the Hidden Secret ore body is lode style and therefore differs from the typical economical mineralisation mined at Mt Charlotte and consequently demanded a novel approach to the mining methodology. The ore body is lens like in shape, extending 200m along strike, 210m vertically, and dipping at 55 degrees. The lode width varies from 1.5m at the extremities to over 22m in the central zone of the ore body. Average grade is 3.4 Au grams per tonne. The initial mining methodology was Modified Avoca utilising Longhole Open Stoping with pillars, in a bottom up stoping sequence. However, as Hidden Secret became better defined and wider than initially modelled, it was identified that a high portion of gold would remain in the pillars and other mining concerns emerged such as remote backfilling large open voids and potentially unsafe vertical standoff distances between levels and wide open stopes. A re-evaluation of this method was undertaken and a new approach adapting a variation of Avoca methodology was identified to remove the rib pillars and replace them with a sill pillar separating the 4 levels into 2 distinct blocks. This new approach provided an immediate advantage by removing the bogger backfilling activity from the same level of the active stope front, enabling stope drilling and prepping to occur during the backfilling cycle. Bogging productivity has also improved as the fired ore is funnelled to the draw point by the backfill rill, significantly reducing the requirement for remote bogging thus improving bogging productivity. This mining method is currently in progress in the first of the two mining blocks and is on track to realise the targeted recovery of 7,700 ounces from the rib pillars.

Mar 15, 2021

By Green S, McKernan, Skrypniuk

In November 1996, Boddington Gold Mine (BGM) began the implementation of an expert system. The objective of the expert system was to improve plant efficiency by developing an `operator best practice' control philosophy for the grinding circuit

Jan 1, 1997

By Alwan M

At Binebase mineralisation is hosted within an andesitic Neogene volcaniclastic succession on Sangihe Island, Indonesia. Two major alteration types occur; an initial widespread phyllic assemblage of quartz -illite with chlorite, smectite, gypsum and pyrite, and a superimposed quartz - barite - sulphide assemblage. The phyllic alteration arose from a neutral to weakly acidic hydrothermal solution of greater than 200 degrees C, and the quartz - barite - sulphide assemblage arose from a cooler slightly more acidic fluid. This second alteration event is interpreted to have selectively replaced permeable volcaniclastic breccias and coarse lithic tuffs within the altered sequence. Gold and silver mineralisation is associated with the latter alteration, and is locally controlled by the degree of silicification and sulphide abundance. The sulphides present are pyrite, covellite and chalcopyrite. The sulphides are very fine grained with colloform and framboidal textures and were deposited contemporaneously with the bulk silicification of the coarse sediments. A large portion of the mineralised rock has been oxidised to a quartz - barite -limonite - goethite - jarosite assemblage.Coincident gold, silver and antimony geochemical soil anomalies gave the best correlation with anomalous gold and silver grades in bedrock as identified by drilling. Resistivity proved a reliable geophysical method for delineating zones of silicification.

Jan 1, 1990

By H von Blottnitz, J L. Broadhurst

Eco-efficiency has been proposed by the World Business Council on Sustainable Development as an environmental performance indicator framework that contributes to sustainability by assisting corporate decision makers improve the environmental performance of their operations and processes, while also extracting additional economic value. Given that the recent engagement of the minerals industry with sustainability has highlighted the need for requisite tools and methodologies, it is therefore of interest to assess whether eco-efficiency indicators can drive sustainability performance improvement during process design within the minerals industry. Previous work indicates that the choice of sustainability performance indicators to be used for decision making needs to be governed by the decision context; ie the performance indicators need to be fit-for-purpose. Here, the strengths and weaknesses of eco-efficiency indicators are analysed using case studies. First, a more rigorous definition of eco-efficiency indicators for the minerals industry is proposed based on indicators currently offered in the literature. Two case studies are then selected to reflect typical decision situations encountered within minerals process design. For each case study, eco-efficiency indicators are computed and are compared to more traditional graphical representations of the economic and environmental performance of process design alternatives considered. The sensitivity of these indicators to various process design parameters is also investigated. Results confirm that the meaningful application of eco-efficiency indicators is strongly dependent on the decision context in which the process data is generated. Results also show that eco-efficiency indicators can be related to technical process design parameters to allow for richer communication of performance information for various decision trade-offs; however, in some cases, graphical approaches that provide more insights into the environmental-financial trade-offs should be preferred. The research thus contributes to guiding the selection of the appropriate tools for more environmentally sustainable design of minerals beneficiation processes.

Jan 1, 2009

By S J. Schafrik, Z Agioutantis, E C. Jong, K Luxbacher

Real-time monitoring of mine operating parameters is available at most mines. However, access to these data for an individual operation is typically underutilised by mine managers and technical personnel. Such monitoring systems are typically designed to provide data streams and alarms to a responsible person, but they are not utilised to extract underlying trends or correlate different data groups. Furthermore, multiple monitoring systems are usually implemented at a given site without interconnection between vendors’ systems or connection to any central management, and mine personnel are only exposed to a minimal subset of such data. Thus, it is not easy to analyse and correlate individual data streams and generate information to enhance day-to-day operational decision-making and improve safety and efficiency.This paper addresses the primary challenges associated with centralising data management and effecting greater utilisation of the vast amount of collected data, while at the same time reducing the time and effort required to manage such large data sets. Furthermore, it presents the conceptual design of a database application that can organise, present and analyse data collected by multiple sensors in different environments. The database is designed to operate in a passive manner so that active mine data management systems are not impacted. An example is given where this database is used for managing atmospheric data collected at an underground coalmine operation in the US. These techniques have a universal application and can be incorporated in a wide range of mining and construction projects.CITATION:Agioutantis, Z, Schafrik, S J, Jong, E C, Luxbacher, K and Karmis, M, 2015. A platform for managing and analysing atmospheric data in underground coalmines, in Proceedings Third International Future Mining Conference, pp 3–10 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Nov 4, 2015

By M I. Pownceby, S Bhargava, F Charalambous, S McMaster, R Ram, J Tardio

"A key component of any geometallurgical investigation involves being able to predict element or mineral recoveries from simplified input data such as bulk chemistry and/or mineralogy. This relies on the metallurgist having access to accurate information regarding the chemistry of the ore and gangue minerals as well as the likely impact on processing/extraction caused by the impurities (present either as gangue or as substitutional impurities). In this study the chemical properties of two commercially important uranium minerals, brannerite (U4+Ti2O6), and uraninite (U4+O2), were characterised using X-ray diffraction (XRD), scanning electron microscopy (SEM), and electron probe microanalysis (EPMA). Significant levels of impurities were present in both samples with the uranium minerals deviating significantly from their ideal chemical formulae. Differences in chemistry were due to: cationic substitution decay of the uranium isotopes into radiogenic lead and other daughter products mixed uranium oxidation states post-depositional alteration.As well, the brannerite sample showed evidence of significant radiation-induced metamictisation. Leach tests to measure the ease of extraction of uranium from both samples were conducted over a wide range of conditions (temperature, acid concentration and oxidant concentration). For uraninite, the lower U dissolution rates observed in the natural sample compared to a synthetic UO2 sample were directly attributed to the presence of impurities. For brannerite, significant metamictisation of the natural sample led to enhanced U extraction rates compared to a heat treated fully crystalline sample. This study highlights the importance of detailed mineralogical characterisation studies in any uranium process mineralogy investigation as the compositional, textural and microstructural aspects of the main uranium phases present will likely play an important role in any process to extract uranium.CITATION:Ram, R, Charalambous, F, McMaster, S, Pownceby, M I, Tardio, J and Bhargava, S, 2013. Chemical and microstructural characterisation studies on naturally occurring uranium minerals - implications for processing, in Proceedings The Second AusIMM International Geometallurgy Conference (GeoMet) 2013, pp 325-334 (The Australasian Institute of Mining and Metallurgy: Melbourne)."

Sep 29, 2013

By S Mitra, D O’Dea, N Barrett, T Honeyands, P Zulli

The steel industry continues to explore and demonstrate technologies that decrease its greenhouse gas emissions, especially in blast furnace ironmaking where alternate operating conditions will be required. As the furnace has been optimised over many years, it is approaching the minimum fuel rate possible with coke and Pulverised Coal Injection (PCI). Alternative operating conditions are therefore required to further decrease greenhouse gas emissions. Because of this, hydrogen has been proposed as a reducing agent in the blast furnace. Injection of hydrogen through the tuyeres was assessed using a comprehensive 2D blast furnace global model. The model was calibrated to a blast furnace operating with PCI as a base case and various approaches to hydrogen injection assessed, with a particular focus on displacing PCI. The maximum hydrogen injection rate of 19.5 kg H2 per tonne of hot metal was predicted, limited by decreases in the raceway adiabatic flame temperature and top gas temperature. Internal blast furnace conditions, such as gas composition and reduction degree were also predicted. Holistic analysis was performed to estimate the effectiveness of hydrogen injection on reduction of greenhouse gas emissions. Based on these results, the changes in blast furnace operation and emissions under hydrogen injection are discussed.

Nov 8, 2021

Diamonds have been known in Australia for many years, especially in association with alluvial gold and tin occurrences. Exploration for primary sources in the Kimberely area of Western Australia has been carried out since 1968, culminating in the discovery, in 1979, of the Argyle Kimberlite No 1 (AK1) kimberlite pipe. Various techniques have been used in the search for diamonds such as geologic interpretation, aerial photography and Landsat, soil geochenical techniques, geophysics, and stream sediment sampling. Stream sediment sampling, a method developed in South Africa, using the detection of indicator minerals derived from kimberlites in gravel samples, has had wide application and success in Australia. Nevertheless, although Australian diamond bearing kimberlite and lamproite rocks contain some of the classic indicator minerals, their concentrations are generally much lower than in South Africa and Russia. Despite this paucity of indicators, stream sediment sampling is still the best known and most commonly used method of diamond prospecting in Australia. Small samples are treated by hand but larger samples require the use of diamond pans, jigs or heavy media separation plants. Each of these methods has advantages and disadvantages. The final concentrate is passed over a grease table or subjected to x-ray sorting.

Jan 1, 1984

By R G. Barker, N Mortimer

In New Zealand, platinum group metals (PGM) occur in three main geological environments of economic interest: layered mafic igneous complexes, the Dun Mountain Ophiolite Belt, and in association with placer gold deposits. Most of the layered mafic complexes with PGM potential occur in the Median Batholith. These include the Longwood (Permian), Bluff (Permian), Riwaka (Devonian) and Rotoroa (Jurassic) igneous complexes. Exploration of the Longwood Complex discovered PGM associated with gabbroic rocks. Best assays to date include a 1 m drill intersection of 1 g/t Pt + Pd and float rock with 3.0 ppm Pt and 3.3 ppm Pd. The Greenhills complex at Bluff consists of basal dunite overlain by wehrlite and gabbro. PGM occur with chrome spinel in the basal dunite and Pt geochemical anomalies have been identified associated with the wehrlite-gabbro contact zone. The Greenhills complex has some similarities to Alaskan-Ural type complexes.Reconnaissance exploration in the Dun Mountain Ophiolite Belt has identified PGM geochemical anomalies, but no follow-up work has been carried out. During the late 1800s and early 1900s, PGM were recovered as a byproduct of placer gold mining in Southland. Placer gold prospects with PGM potential include the Waiau River, deposits flanking the Longwood Complex and some Southland beaches.

Jan 1, 2002

By C Catania

Mining companies and operations across the globe typically target best practice financial management, along with a focus on economically driven mine plans. Not only is this inherent in running a successful mining business, but also critical in addressing investor scrutiny. The standards for social and environmental practices is less consistent than this, with most companies adopting standards that are driven through legislative and compliance requirements of the nations in which they operate. Although there is still room for improvement, Australian is a country that stands at the forefront of social and environmental legislation, driving more sustainable planning practices for mining operations, with an ever-increasing focus on mine closure considerations. Of the many nations yet to reach such a legislative position, Kazakhstan is a place where some companies are driving towards world class standards beyond what is government mandated. CITATION:Catania, C, 2018. World-class mine planning for a sustainable future in Kazakhstan Mining, in Proceedings Life-of-Mine 2018, pp 6–7 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Jul 25, 2018

By R Paterson, A Christensen

The routine use of gravity gradiometry for the detection of orebodies with no outcrop is now being undertaken. The commodities and orebody styles range from banded iron ore formations to volcanogenic massive sulfide (VMS) style polymetallic deposits. This recent innovation of flying gravity gradiometry was pioneered by BHP Billiton and Lockheed Martin with the Falcon technology (Lee, 2001), and now two additional systems, also using Lockheed Martin instruments, are deployed. In South Africa, magnetic gradiometry is being developed and deployed with the same objectives in mind. In this case, a system from Institute of Photonic Technology (IPHT) is being used. The detectable wavelengths from such systems have decreased to less than 200 m in the last year, and it is expected in 2011 that wavelengths to less than 100 m, with anomaly signal strengths of less than 2 Eotvos for gravity gradiometry will be viable. The key to these advances lies in the improvements of processing software and geological modelling. In particular, methods using all the tensor components are starting to be routinely deployed to filter, level and grid the observed signal. These methods do, however, require full insight and understanding of the reference coordinate systems used by the various instrument designers and survey operators. The digital terrain models need also to be observed to an accuracy of better than 20 m cell sizes, using such techniques as Lidar. Then the terrain effects can be modelled. The topographic surface represents the largest and most proximal density contrast encountered in an airborne survey. Hence terrain effects can have significant impact on airborne gravity gradiometry data. A decorrelation of the terrain effect with the observed signal requires a method of best fitting with a regional surface density. A novel technique has been developed for this purpose. This technique uses a tensor algebra to estimate a density between 1 - 4 g/cc that explains the average observed signal. Armed with this result, the anomalous buried bodies can be revealed by subtraction.

Sep 26, 2011

LESLIE BRADFORD GOLDEN JUBILEE ORATION "Not only a metallurgist but a scientist of the highest order" Leslie Bradford and the Australian Metallurgical Industry

Jan 1, 1987

By Y Durkheim, P Saunier, L Ameglio

"In the last decade Afitex developed the Draintube™ geocomposite, which provides simultaneous drainage and waterproofing functions. This new type of drain tubes planar geocomposites (DTPG) differs from other geocomposites as the drainage core is composed of multiple corrugated and perforated pipes instead of biaxial or triaxial nets. It is furthermore associated with needle-punch stitched layers of non-woven geotextiles, which act either as capillary mediums or as filters.In this paper, the structure of the Draintube™ drainage composite is presented along with its key properties and the drainage mechanism associated with its particular structure.The relevance of this kind of structure is then reviewed based on past and current laboratory bench evaluations and field installations for mining applications such as the covering of acid rock drainage (ARD) tailings for the Central Manitoba Mine (Canada) site-rehabilitation; and as a pregnant solution collection layer in heap leach pads.Some of the demonstrated critical advantages of Draintube™ in the mining industry are as follows: reduction of the thickness, hence lower volume of granular drain layers improvement of fluid and gas collection and reduction of the hydraulic load reduction of secondary collector network mechanical protection of the sealing membrane fast and economical installation of interlifts stability under extreme conditions (compression, temperature, pH) increased stability of entire covering system and reinforcement of protection of dykes filtration of the covered material customised drainage positive connection of the geocomposite to collector pipes enable leaks detection (using a conductive textile layer) with an electrical test.CITATION:Saunier, P, Ameglio, L and Durkheim, Y, 2015. Tubular drainage and lining geocomposite for mine tailings and heap leach pads, in Proceedings Tailings and Mine Waste Management for the 21st Century, pp 297–304 (The Australasian Institute of Mining and Metallurgy: Melbourne)."

Jul 27, 2015

A Feasibility Study to determine the future of ZC Mines was completed in mid 1987. Early in this study, which lasted two years, it was identified that significant reductions in treatment costs would result if the Zinc Corporation Concentrator was closed. All of the production from the existing two mines, The Zinc Corporation Limited (ZC) and New Broken Hill Consolidated Limited (NBHC), would then be treated in the concentrator at NBHC.After a partial upgrade of the NBHC concentrator, the ZC concentrator was finally closed and all of the equipment disposed of. This closure highlighted the need to concentrate all ore haulage in one shaft to further reduce operating costs and remove the cost of surface transport of ZC production to the improved NBHC concentrator. This paper outlines the changes that were made to improve the haulage system for ZC Mines. After a number of options were investigated, it was finally decided that the upgrade of the existing NBHC haulage shaft was the best option even though some production would be lost while the upgrade was in progress.

Jan 1, 1988

By Perring CS, Vearncombe J. R, Ho SE, Groves DI

Large Late Archaean greenstone belts in Western Australia, Canada and Zimbabwe, contain abundant mesothermal gold, volcanogenic massive base-metal sulphide, and komatiite-associated Ni-Cu mineralization. These greenstone belts are similar in many respects to convergent and accretionary plate margins found in the Pacific rim; modem environments which are cor-respondingly rich in gold and base-metal mineralization. The giant gold placer deposits of the Late Archaean Witwatersrand Basin (the world's largest gold producer) formed in a foreland or retro-arc basin; in a similar tectonic setting to much modem placer mineralization around the Pacific. Gold and volcanogenic base-metal sulphide mineralization are abundant in Late Archaean terranes and the Pacific rim because1. gold mineralization is a product of convergent margin tectonics, and 2. accretion in convergent margin orogens provides the best chance of preserving base-metal mineralization related to submarine volcanism.It is therefore likely that the abundance of gold and base-metal sulphide mineralization in the Late Archaean reflects the widespread growth and sta-bilization of continental crust via tectonic processes broadly similar to those operating in the present Pacific basin.

Jan 1, 1990

By E Widzyk-Capehart, T Bednarz, O Daud, N Cutmore

The interest of industry, in particular future mining, is pointing towards the use of novel visualisation systems to improve the operational efficiency of user interfaces. While the technologies used to enhance operators’ experience are advancing, the implementation of novel visualisation paradigms within mining operations is trailing behind other industries, particularly health and aviation. In this work, we present a number of initiatives and use cases of immersive virtual and augmented reality systems that can fit within the mining context. These ideas represent a step towards new platforms that might increase the efficiency and safety of mining operations through applications such as 3D panoramic video monitoring technology with annotated information coming from big data analytics and visual analytics processing engines. However, while the technologies are used in a variety of training settings, their application in the working environment is lagging behind. This paper presents examples of scientific visualisations that could be applied in a mining context to extend the possibilities of operators’ cognitive abilities and working efficiency. They are intended to foster a discussion within the mining community to move beyond the training realm and into the operational environment.CITATION:Bednarz, T, Widzyk-Capehart, E, Daud, O and Cutmore, N, 2015. Effective immersive visualisation systems for mining applications and beyond, in Proceedings Third International Future Mining Conference, pp 31–36 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Nov 4, 2015

By H Law, M Muller

Mount Isa Mines (MIM) Copper Concentrator currently utilises Agitair 120 banks for its roughing flotation circuit. Apart from a rotor change, the only other significant work completed to the Agitair banks since their installation and commissioning over 40 years ago has been a like for like refurbishment in 2016. The Agitair banks at the MIM Copper Concentrator currently utilise the Pipsa rotor design which progressively replaced the original Chile X rotor design in the 1980’s and 90’s in the copper concentrator. Since the development of the Pipsa rotors, modern rotors such as the Dorr-Oliver (DO) high shear rotors have been released to the market with proven recovery and concentrate grade improvements over previous rotor designs. This paper aims to detail the installation, testing and comparison of Pipsa versus Dorr-Oliver rotors in two different flotation sections of the MIM Copper Concentrator.CITATION:Muller, M and Law, H, 2018. High shear rotor trials in the Mount Isa Mines Copper Concentrator, in Proceedings 14th AusIMM Mill Operators' Conference 2018, pp 393–402 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Aug 29, 2018

By K Jupp, T J. Howard

Cliffs Natural Resources Pty Ltd (CNR) operates iron ore mines in the Koolyanobbing region of Western Australia. Ore is mined from three locations, separated by many kilometres. The ore is stockpiled at these locations, according to an in-house ore classification system (based on grade and source) and then trucked to the crushing and screening plant at Koolyanobbing. Lump and fines products are railed to Esperance for ship loading and export to Asian customers. CNR prides itself on the relatively low inter-shipment grade variability of the products. The Koolyanobbing crusher is fed using a daily blend plan, generated to maintain lump and fines product grades within acceptable tolerance ranges around targets. Achieving low variability requires predicting lump and fines grades as accurately as practical from the estimated head grades of the ROM ores potentially going into the blend. The grade prediction model may be either a direct prediction of crushed lump and fines grades and lump percentage, or be split into two stages: the bias between blasthole estimated head grade and crusher head grade and then the lump-fines algorithm for splitting the head grade between lump and fines products. The lump-fines algorithm comprises the percentage of lump and the difference between the lump and fines grades. We describe a weighted least squares (WLS) regression model for predicting crusher grades and lump proportion from the estimated head grade for Fe, P, SiO2, Al2O3, Mn and S. The method is applied to CNR production data, where the regression model explains approximately 60 per cent of the variance in the crushed ore grade, for both lump and fines. A further small but significant improvement in prediction can be achieved by including the ore classification data in the model. The regression errors exhibit strong positive serial correlation, indicating trends in grade error across multiple blend records. To compensate for the longer-term error, an exponentially smoothed model was developed and applied to the daily grade blend errors. This gave an increase in the longer-term variance explained and therefore an improvement in grade prediction.

Jan 1, 2009