Search Documents

By B Dodd

This paper details how the Australian Securities and Investments Commission (ASIC), as a disclosure and conduct regulator, approaches regulation of continuous disclosure, capital raising and control transactions in the mining and resources sector, and the reasons for the approach. In examining this topic the paper considers the Joint Ore Reserves Committee (JORC) Code (JORC, 2004), the VALMIN Code (VALMIN Committee, 2005), the relevant ASX Listing Rules and the relevant provisions of the Corporations Act 2001 (Corporations Act). This paper discusses ASICÆs approach to expert reports and disclosure with reference to Regulatory Guides 111 Content of expert reports and 170 Prospective financial information.CITATION:Dodd, B, 2012. Issues in capital raisings and disclosure, in Proceedings International Mine Management 2012, pp 3-14 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Nov 20, 2012

Forty years ago Dr Pierre M Gy, a metallurgist, introduced variography as applied to monitoring a process. Consequently, his followers used the variogram to elaborate meaningful process control charts, especially since 1993. The name æChronostatisticsÆ was suggested several times by Gy as a logical name for this new science. The intent was to draw a parallel between the variogram identifying the variability components affecting a process parameter as a function of time, and the variogram used in geostatistics identifying the variability components of a given mineral content as a function of space. Heterogeneity problems encountered have indeed strong similarities. However, there is a fundamental difference. The variogram used in geostatistics is often stationary within a given type of mineralisation. While investigating a dynamic process through Chronostatistics, there is no such thing as a stationary variogram. Still, the evolving variogram provides essential information about subtle, often overlooked process difficulties. Chronostatistics is especially useful for management to take advantage of existing data that are often a forgotten source of valuable information to pinpoint recurring structural problems. The superiority of Chronostatistics over conventional statistical process control approaches is that it links observations to the inherent heterogeneity of materials under investigation, therefore to the sampling theory. Chronostatistics is like opening a new world to the metallurgist, making the process variability speak in a more comprehensive and in-depth way. Several case studies are presented to demonstrate the potency of this new science. For each of them, observations lead to solutions for continuous process improvement, therefore attractive economic opportunities. The philosophy is this: if a control chart shows a process is out of control, it must also suggest logical and pragmatic directions for solutions. The intent of this presentation is to vastly expand on these valuable, almost forgotten ideas.

Jan 1, 2006

By A Bossilkov

Although environmental and sustainable development reporting has now become widespread in the Australian mining and minerals processing sector, there remains a great deal of diversity in reporting contents and practice. Moreover concerns remain about the effectiveness of reporting for delivering the companyÆs business case for sustainable development as well as for communicating sustainable development performance to a range of external stakeholders. This paper therefore reviews the status of sustainable development reporting in the Australian minerals processing industry. Pertinent international and national developments are briefly captured, leading to a discussion of the framework for integrating sustainable development reporting with sustainable development management (as proposed by the World Business Council for Sustainable Development). This leads to a discussion of the results of a comparative review of sustainable development reports of a cross-section of eight Australian resource companies. Best practice elements are highlighted in regard to reporting on company context, governance, performance and assurance.

Jan 1, 2004

By B Dold

Western Europe is a dynamic region, with an emerging co-operation between the individual nations both politically, economically, technically and scientifically. The European Union, today with 15 member States and with a range of central European States applying for membership, is a catalyst for this Cupertino, using a combination of EU funds, for financing research initiatives, and legislation to enhance development of technology and environmental performance. Europe is a major net importer of metals but with a large potential for own production, in particular in the accession countries. The metals and mining sector is furthermore identified as having a large potential to generate wealth both in todayÆs EU and in the future, widened Union. This paper provides a chart of examples on the many research initiatives presently underway in Europe, addressing environmental issues related to mining. The list of programmes and projects does not claim to be complete; it is rather an illustration on the wideness of the approach, describing various ways for transboundary Cupertino. The approach taken by many of the programmes goes beyond the strict ARD scientific area and therefore only a limited number of the many programmes can be identified as ARD research programmes. Also national programmes, without EU funding, are operating, eg in Sweden, Germany, and UK. The picture is continuously changing with new programmes being commissioned and old ones being completed.

Jan 1, 2003

By A Scott, I Onederra

Most iron ores occur in mine benches as complex mixtures of lithologies and ore types. This significantly complicates their characterisation for both blast design and fragmentation modelling. A great deal of technical and research effort has been applied to the development of models to predict fragmentation from rock blasting. A number of quite useful engineering models are currently available to blasting engineers and consultants, and more sophisticated approaches using mechanistic descriptions of rock breakage continue to be developed and applied by researchers and explosives companies.The engineering or empirical models that are in more general use within the industry, such as the Kuz-Ram model, require appropriate data if they are to generate useful fragmentation predictions. Some of these models have reached significant sophistication and have made very useful contributions to both routine blasting operations and the exploration of new blasting paradigms. However, these models struggle to adequately deal with rock that is as complex as iron ore. This paper reviews the characteristics of the most successful blast fragmentation models available to the industry and discusses approaches to the derivation of suitable rock mass properties to adequately describe blast fragmentation in a range of iron ores.CITATION:Scott, A and Onederra, I, 2015. Characterising the blasting properties of iron ore, in Proceedings Iron Ore 2015, pp 481–490 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Jul 13, 2015

By R Dimitrakopoulos, R Senécal

This paper presents a metaheuristic solution to the optimisation of open pit long-term production scheduling with a stockpile and geological uncertainty. The optimisation formulation is a two-stage stochastic integer programming (SIP) model that determines the optimal mining sequence that maximises the total discounted cash flow, while penalising for high deviations from production targets. A parallel implementation of Tabu Search is proposed to accelerate the solution of the SIP formulation and take full advantage of multicore computer processing. Different variants of the proposed algorithm are applied to a case study to assess the performance of the parallel approach. The proposed algorithm and variants are benchmarked using linear relaxation of the complete problem to determine robustness and provide a better overview of the related performance. The results show a net improvement over the sequential solution, and the new proposal seems to be promising when working with a large-scale data set.CITATION:Senécal, R and Dimitrakopoulos, R, 2014. Parallel implementation of a Tabu search procedure for stochastic mine scheduling, in Proceedings Orebody Modelling and Strategic Mine Planning Symposium 2014 , pp 405–414 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Nov 24, 2014

By J M. Wallis

Heavy mineral sand exploration drilling prior to the 1970s generally utilised open hole auger and air-blast type drilling. However, mining reconciliations were poor and limitations on accuracy and precision of drilling results were severe.Working with the exploration companies, Wallis Drilling principals designed various systems, including what was termed the ‘Eneabba System’ and then the ‘Capel Twin-Tube System’. Eventually the company with its improvements developed and patented the Aircore Drilling System.Aircore drilling has subsequently become recognised as an excellent drilling system for exploration and mine-sample drilling that gives good precision and accuracy in sampling both above and below water. Its use has also been extended to many other commodities and is used in a number of overseas countries.This paper documents anecdotally the development of the system, identifying problems encountered during the times it was evolving and of the means to overcome the problems as being the documentation of the story by the inventors and developers in Wallis Drilling Company. Aircore drilling system is a Western Australian technological development owing its existence to the designing and incorporation of a number of discrete improvements along the way and recognises the tenacity and inventiveness of all who contributed to the system, in both Wallis and also in a number of companies that gave the impetus to succeed.

Oct 5, 2011

By D J. Brake

Faced with the need to improve workplace conditions in both its development and production headings in its existing K1 orebody, as well as its new K Link and K2 orebodies, the Kencana mine in Indonesia took the novel approach of reversing the ventilation on most of its access ramp by outcasting up the ramp to the portal. Working levels were then fed with fresh air via auxiliary fans sealed into the current surface ‘egress’ intake raises (ladderways) or another small surface intake raisebore rather than from fans hung in the ramp as in the past. Outcasting the portal provided a substantial increase in the total mine exhaust capacity allowing fewer levels to be ventilated in series and significantly improving conditions, especially in the most difficult workplaces to ventilate. It also reduced the consequences if there was to be a fire on a vehicle in the ramp. The same system using the ramp as an exhaust is also intended to be used as part of the final ventilation solution for the new K2 orebody which also uses a novel single surface ventilation shaft that is vertically split into separate intake and exhaust compartments using a concrete brattice. Using the main ramp as an exhaust has potential application in many other operations in Australia using surface ramps which otherwise frequently suffer from overuse of series ventilation circuits and poor working conditions at the mine bottom.

Mar 21, 2011

By D J. Williams

The approach to the disposal and storage of mine tailings is slowly evolving. Tailings disposal has been based on minimising short-term capital and operating costs, with rehabilitation costs discounted by the application of a high discount factor in a net present value approach to accounting. This has led to the widespread adoption of surface tailings storage facilities (TSFs) to store slurried tailings delivered by robust and cheap centrifugal pumps and pipelines. Depending on the site climate, rate of rise of the tailings, nature of the tailings and the extent to which supernatant water is removed from the surface TSF, the tailings may or may not dewater, densify and gain strength. High rates of rise of dilute tailings to a facility used to store both tailings and supernatant water results in a wet, low-density, soft tailings deposit that not only occupies a large storage volume but also presents rehabilitation difficulties. A low tailings density results in high containment wall costs, very high rehabilitation costs and a low potential for future land use or ecological function. Overall, the cost of slurried tailings disposal, thought to be low, is actually high. However, low rates of rise, even of dilute tailings, in a dry climate with efficient supernatant water removal can produce a dewatered, high-density, stiff tailings deposit that may occupy a large footprint but to shallow depth that is relatively easy to rehabilitate. There are advantages in dewatering tailings in the processing plant to recover process water and chemicals, facilitate tailings disposal in a more solid form and facilitate rehabilitation. However, the cost of dewatering in-plant can be high, and the process water may not lend itself to reuse. There are also advantages in removing supernatant water to facilitate solar and wind drying of exposed tailings, although water is lost to evaporation and the dried tailings may best be harvested prior to inundation by fresh tailings or rainfall run-off. This paper discusses the background and trends towards the elimination of conventional surface slurried TSFs, with examples presented of the management of black coal and other tailings, including those that are clay mineral-rich.CITATION:Williams, D J, 2015. Towards the elimination of conventional surface slurried tailings storage facilities, in Proceedings Tailings and Mine Waste Management for the 21st Century , pp 11–24 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Jul 27, 2015

Bacterial leaching of sulfide minerals are fairly well established in heap, dump and in situ leaching and the primary bacterial treatment of refractory gold concentrate using CSTR have gained industrial prominence in recent years. Three cultures, a mesophile T ferrooxidans (DSM583), a moderate thermophile Sulfobacillus acidophilus (strain YTF1) and a extreme thermophile (Sulfolobus strain BC65) were used for the biooxidation (pre-treatment) of the Porgera gold-bearing pyrite concentrate under optimum O2 and CO2 enriched air (30 per cent O2 and ten per cent CO2) based on volume ratios. The positive effect of using oxygen and carbon dioxide enriched air for the biooxidation of gold-bearing pyrite mineral was clearly demonstrated where an increase in Fe release rate by a factor of 2.8, 1.8 and 1.5 for DSM583, YTF1 and Sulfolobus respectively were observed compared with the shake flask tests. The lag phase observed with the other test conditions such as in shake flask or with air decreased by about 80 per cent for all three cultures. The average observed dissolved oxygen concentrations in the pulp for these tests were 8.7, 5.8 and 5.0 mg/L for DSM583, YTF1 and Sulfolobus respectively, indicating that O2 was not limiting in these tests. The pulp redox potential was observed to increase with time for all tests coincident with the leaching rate and reaching a maximum of 663 mV, however the maximum value differs for each strain, with DSM583 exhibiting the highest followed by YTF1 and Sulfolobus. Biooxidation of pyrite under optimum gas enrichment at different pulp densities showed near total Iron extraction occurred for pulp densities of three, ten and 20 per cent (w/v). At high pulp densities, more so at 40 per cent than 30 per cent (w/v), the leaching rate decreased rapidly for tests with YTF1 and Sulfolobus. Monitoring of dissolved O2 shows that O2 demand increased with increasing pulp density and the average dissolved O2 concentration observed for YTF1 and Sulfolobus at 40 per cent solids concentration was 1.0 mg/L and below, suggesting limiting oxygen and carbon dioxide conditions for these two cultures. The decrease in leaching rate observed at high pulp densities (30 and 40 per cent) under optimum gas enrichment was accompanied by low metal dissolution and low redox potential values which was attributed to shear stress due to attrition of cells by high solids concentration, toxicity of metal ions especially Ag and Fe and the formation of precipitates and jarosites as confirmed from XRD analysis of leach residues hence impeding the leaching process. Limiting conditions due to high pulp density and viscosity were only observed for YTF1 and Sulfolobus. The subsequent cyanide leaching for precious metals from the pre-treated pyrite concentrate in all cases shows the extent of gold and silver recovery increased with the degree of pyrite oxidation and Au and Ag recoveries in excess of 80 per cent were achieved compared with only 18.1 per cent and 35.9 per cent for gold and silver respectively when treated by direct cyanide leaching on the original, untreated concentrate.

Jan 1, 2004

By P D. Munro

"This is a reprise of the paper ‘Back to the Future – Why Change Doesn’t Necessarily Mean Progress’ presented to the Tenth Mill Operators’ Conference in 2009. It was observed that the ‘boom and bust’ cycles of the industry, together with trends in tertiary education and the effects of fly-in fly-out (FIFO) operations raised serious questions about the sustainability of human capital in the mineral processing sector.In 2016 we find ourselves in a ‘bust’ after enjoying a decade of the longest commodity price ‘boom’ in a century. In the ‘boom’ did we:make any technological breakthroughs?put measures in place to improve our human capital, especially in the area of ‘professional formation’?The paper makes observations on the current performance of mineral processing engineers and the organisational milieu in which they operate. It questions whether mineral processing engineers are actually ‘adding value’ at some sites given their organisational and operational practices.Some technical trends are examined and what it might mean for mineral processing engineers.Given the above and demographic trends for mineral industry professionals, changes to the seemingly current laissez-faire model of professional development are proposed.Citation:Munro, P D, 2016. Back to the Future - Still on the Dark Side, in Proceedings 13th AusIMM Mill Operators’ Conference 2016, pp 341–348 (The Australasian Institute of Mining and Metallurgy: Melbourne)."

Oct 10, 2016

By D J. Brake

In most underground mines, the ventilation system is a significant portion of the total mine capital and operating cost and the largest consumer of underground electrical power. The effectiveness of the ventilation system has a major bearing on the safety, health, comfort and morale of underground personnel, and has a critical role in meeting production targets and therefore business revenue. Perhaps the most important component of the primary ventilation system is the primary fan. Correct performance testing of the primary fan during the warranty period is important not only to ensure the mine has obtained the asset that it paid for and can therefore deliver on the fanÆs crucial safety and economic role in the future of the mine, and to provide a justification for releasing final performance warranty payments to the supplier, but also to provide a benchmark or baseline against which to assess future performance issues. As there are at least three potential international standards available to measure fan performance, there has been considerable confusion regarding which of these to use for testing primary mine fans, and also the correct allowances and tolerances to use. This paper provides guidance on which standards to use for different applications, how the standards should be interpreted, what allowances and tolerances are reasonable, and other important issues to be resolved with the fan supplier prior to making a contract award. The paper also discusses practical issues and methodology for in situ primary fan performance testing.CITATION:Brake, D J, 2013. Performance and acceptance testing of main mine ventilation fans, in Proceedings The Australian Mine Ventilation Conference , pp 141-148 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Jul 1, 2013

By X Hill

Throughout the 2016/2017 financial year, the NSW Resources Regulator conducted an assessment of each underground metalliferous operation across the state to assess how each operation managed the principal mining hazard of ground and strata control. The assessment considered all aspects of the mining process, including all aspects of managing geotechnical risk within drawpoints. This assessment is ongoing and, therefore, this paper is limited to the first six mines assessed in the program.In 2015 there were three fatalities in and around underground drawpoints in Australian metalliferous mines. Whilst many of the systems of work around underground development are consistent across the industry, there are large discrepancies around how each underground mining operation manage the risks associated with drawpoints.For drawpoints to be effectively managed, their management needs to be considered at the earliest stages of the mine planning process. By using downholes and adopting remote systems of work – systems that are continuing to improve with technology – the risks drawpoints posed can be greatly reduced.The assessment by the NSW Resources Regulator revealed a number of common failings across the state in the management of the risks that drawpoints pose. The assessment also observed a number of innovative practices that particular operations have adopted in the wake of serious incidents occurring within drawpoints at their operation; however, these practices have not been communicated to other operations.It is hoped the learnings from the assessment can be considered by industry and reduce the risks drawpoints pose in underground metalliferous mines.CITATION:Hill, X, 2017. Learnings from review of drawpoint management across New South Wales mines, in Proceedings 13th AusIMM Underground Operators’ Conference 2017, pp 143–150 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Oct 16, 2017

By Chitombo GPF

The dilution of blasted ore with cemented backfill materials has been recognised as a significant factor influencing the economics of some underground metalliferous mining operations. Several mechanisms have heen identified as possible causes of fill failures. The purpose of this study was to determine the possibility of damage as the result of one mechanism, that of production blasting adjacent to bodies of cemented hydraulic fill (CHF). A vibration amplitude transfer function has been developed to model the transmission of blast vibrations from the source, through ore, across the ore to CHF interface and into the CHF mass. The model may be applied in the near field. The model has permitted estimation of vibration amplitude at a point in a CHF mass, to which damage criteria may be applied. A number of assumptions and simplifications have been made to render the problem relatively tractable. Ore and CHF material characteristics recorded by a number of authors were utilised to formulate preliminary damage estimations. A simple application of the amplitude transfer model indicated that a zone of plastically deformed CHF would be produced, and would extend to a distance of approximately 0.5 metres from the interface into the CHF. A sensitivity analysis was also conducted, that indicated that the zone of plastic deformation would be more extensive given appropriate conditions. Recommendations for further model development and future laboratory and field investigations have been made.

Jan 1, 1998

By Tuckey SH, Wake BA, Ariti J

Astute exploration on the northern Minahasa Peninsula, North Sulawesi, resulted in the discovery of the Toka Tindung, Pajajaran, Blambagan and Araren low-sulphidation quartz-adularia vein and stockwork gold-silver deposits. Combined, these deposits have estimated mineral resources in excess of I Moz gold at an average grade of 2.8 gAu/t and 7 gAg/t and are known as the Toka Tindung Gold Project. Gold typically occurs as fine grained electrum and is often associated with acanthite. The structurally controlled mineralised veins are hosted by Late Pliocene plagioclase and pyroxene-phyric basaltic andesite volcaniclastics and flows which form part of the Early Miocene to Pleistocene Sulawesi-East Mindanao Arc. Silica-adularia alteration dominates the volcaniclastic host of the Toka Tindung deposit. K-Ar dating of vein adularia indicates a 2.16 ¦ 0.04 Ma mineralisation age. Relatively narrow envelopes of silicification and argillisation surround the chalcedonic silica, quartz and adularia veins, stockworks and breccias at the other three deposits. Remnant hydrothermal breccias, chalcedony and anomalous arsenic, antimony and mercury indicate these gold deposits were part of a mineralised geothermal system.As of June 1997, Aurora Gold Ltd is finalising a feasibility study on exploitation of the four deposits and treatment of ore through a centrally located conventional CIP plant. Subsea tailings disposal is anticipated. Favourable metallurgy of the ore and logistics of the project area assist in minimising capital requirements.

Jan 1, 1997

By G P. F Chitombo

The dilution of blasted ore with cemented backfill materials has been recognised as a significant factor influencing the economics of some underground metalliferous mining operations. Several mechanisms have been identified as possible causes of fill failures. The purpose of this study was to determine the possibility of damage as the result of one mechanism, that of production blasting adjacent to bodies of cemented hydraulic fill (CHF). A vibration amplitude transfer function has been developed to model the transmission of blast vibrations from the source, through ore, across the ore to CHF interface and into the CHF mass. The model may be applied in the near field. The model has permitted estimation of vibration amplitude at a point in a CHF mass, to which damage criteria may be applied. A number of assumptions and simplifications have been made to render the problem relatively tractable. Ore and CHF material characteristics recorded by a number of authors were utilised to formulate preliminary damage estimations. A simple application of the amplitude transfer model indicated that a zone of plastically deformed CHF would be produced, and would extend to a distance of approximately 0.5 metres from the interface into the CHF. A sensitivity analysis was also conducted, that indicated that the zone of plastic deformation would be more extensive given appropriate conditions. Recommendations for further model development and future laboratory and field investigations have been made.

Jan 1, 1998

By Wright S

Single pellets of mixed lead-zinc concentrate containing up to 20 % lead were roasted in the temperature range of 800-1100¦C. Variables studied included pellet size (from 9-8 to 3-2 mm diameter), gas flowrate past the pellet (from 0.05 to 5 ms t), composition of the atmosphere and the effect of additives such as zinc oxide and silica. After defined periods of time, the pellets were removed from the furnace, quenched and submitted for chemical, XRD, SEM and microprobe analysis. The oxidation of the zinc sulphide followed a shrinking core model. Temperature measurements showed that very high pellet temperatures were experienced at the start of an experiment but after 5-10 minutes these approached a constant temperature. The lead loss from the pellets was then constant with time. Thermodynamics predicted the presence of two bands of lead material in the reacted layer of the pellet. These were found in practice. The lead loss from the pellet occurred from the outer oxysulphate layer via the vaporization of lead oxide polymers. The vapour pressure of PbO, and the loss of lead, was dependent on temperature and bulk gas concentrations of SO2 and 02. At 1100¦C the addition of silica greatly reduced the loss of lead from the pellet. Physical factors such as pellet diameter and gas flowrate past the pellet affected the particle dynamics. These results were found to vary in accordance with the predictions of the Froessling equation.

Jan 1, 1989

In 1985 BMR assessed Australia's economic demonstrated resources of mineral sands as follows: rutile 8.0 Mt, ilmenite 41.4 Mt, zircon 11.5 Mt, and monazite 229 000 tonnes. These resources are located along the east coast of Australia, in the southwest corner of Western Australia, and in the Eneabba region, north of Perth, WA. However, over 40% of Australia's rutile and zircon resources, particularly those on the east coast, are presently not available for mining because of environmental considerations. As well, east coast ilmenite has been considered unsuitable for use by sulphate route pigment producers because of its relatively high chrome content. World demonstrated resources of rutile and zircon are estimated by BMR as 29 Mt and 38 Mt respectively. Demonstrated resources of ilmenite amount to about 1 billion tonnes. These occur predominantly in hard rock deposits in Norway, Canada, USSR and China, as well as alluvial deposits in Australia, South Africa and India. Quantitative data for world monazite resources is sparse. However, world production of rare earth oxides is now mainly from bastnaesite, a hard rock resource. Australia is presently the world's principal producer of rutile and zircon, and also provides a major share of the world's ilmenite. With the expected increase in Australia's rate of output of these commodities, combined with restrictions on mining along parts of the east coast, its position could diminish substantially over the next 25 years.

Jan 1, 1986

By Rabie A. S, Fouad K. M, Ammar A. A

The existence of Gabal El Kahfa ring complex, located in the Eastern Desert of Egypt, has been ascertained from aeromagnetic and aeroradiometric survey data as well as by aerial photography. Two maps for -the net aerial radiometric measurements reduced to ground level and aerial magnetic data corrected for the regional normal gradient of the earth's magnetic field have been constructed. The aeroradioactivity data have been interpreted geologically and analyzed statistically to outline various radiometric units and compute their charac- teristic statistics. The ring complex showed a circular radiometric feature and yielded a radiometric mean background and a standard deviation of 6.48 and 0.89 pR/h respectively. Besides, the observations showed a normal distribution. The aeromagnetic data have been reduced to the north magnetic pole, then filtered to produce the regional-and residual-component maps. Statistical trend analysis was conducted for the tectonic linea- ments resulting from the three magnetic maps, to define the structural framework of the area under study. This statistical study proved the exis- tence of trends in the east-west, northwest-south- east, northeast-southwest and north-south directions. The ring complex is characterized by a circular magnetic feature, whose average relief is 2150 nT. It is included within a high east-west trending magnetic zone. The latter represents a deep-seated uplift or anticline which is bordered from all sides by some faults of different trends.

Jan 1, 1988

By C Aldrich

In this paper, the use of topology preserving projections (Sammon maps) and neural networks are considered as a basis for the construction of multivariate statistical process control charts for a semi-autogenous grinding circuit on a gold plant. Comminution circuits are often operated in a quasi-steady state mode, where the mills tend to move rapidly from one state to another. This makes them difficult to monitor by conventional means designed for continuous steady state processes. As a first step, process data reflecting normal operating conditions are mapped with SammonÆs algorithm. Only one feature is extracted at a time and each mapping is captured by a neural network model to allow online application. Reverse neural network models are used to reconstruct the data prior to the extraction of subsequent features from the data residuals. The ensemble of neural network models can then be used for both fault detection and identification. The approach compared favourable with both linear and non-linear principal component models. It was able to detect two simulated fault conditions both with a reliability of 100 per cent, as opposed to 57.8 per cent and 95.2 per cent detected by an approach based on the use of linear principal component analysis and 84.1 per cent and 98.0 per cent detected by an approach based on non-linear principal components extracted with auto-associative neural networks.

Jan 1, 2007