Search Documents

By M. C. Holton

South Africa has the world?s largest in situ deposits of platinum group metals, but economically exploitable resources are shrinking due to escalating costs and technical challenges. The South African National Research and Development Strategy comments that an increase in the knowledge intensity of the resource-based sectors is required to turn these resources to account. Highly trained human resources, continuous improvement, technological innovation, and smart acquisition of know-how will become the major differentiators in the global mining market. Mechanization and, ultimately, automation in mining have the potential to reverse the negative mining trend and resolve safety and other challenges. However, the adoption rate of mechanized mining in platinum (and hard-rock gold) has been slow for various reasons. Changing systems from conventional drill?and-blast to continuous mechanized mining is a big step-change endeavour. The problem of transformation is discussed, and the discipline of systems engineering is examined because of its successful application in aerospace and defence.

Jan 1, 2012

The depletion of favourably located and higher-grade mineral deposits, combined with more favourable metal prices and higher capital costs, has stimulated research into heap leaching of lower grade ores and wastes. This has resulted in renewed interest in heap leaching of uranium ores, and in the development of processes for the bacterial heap leaching of low-grade chalcopyrite ores and the acid heap leaching of nickel laterites. A complete programme of heap leach testwork typically involves a sequence of bottle roll tests and column tests, followed by piloting on test heaps. The initial testwork is aimed at determining the amenability of the ore to heap leaching, which depends on the characteristic of the ore with respect to porosity and permeability to leach liquor, acid consumption, metal recovery and percolation. Subsequent phases of the testwork focus on the development of the process design criteria for the treatment of the ore. Basic techno-economic models have also been developed as a tool for the initial estimation of heap leach viability. Mintek has also developed in-heap monitoring probes, coupled to an operator guidance system, in order to assist the operator in achieving optimal conditions for bacterial growth and heat preservation during pilot-plant operation.

Jan 1, 2009

By T. Orupold

This paper covers the commercial development of gravity separation of fine particles using a Lamella High Shear Rate REFLUX? Classifier (REFLUX? Classifier), focusing primarily on coal applications. The REFLUX? Classifier is a fluidized bed device that incorporates a system of closely spaced parallel inclined channels located above the fluidized bed. These channels make it possible to achieve a significant suppression of the effects of particle size, resulting in a highly effective separation on the basis of density. The improved gravity separation performance is shown to be remarkably high, with a significant reduction in the variation of separation density with particle size, and a significant reduction in the change in Ecart probable error (Ep) with size. The first full commercial-sized units of the REFLUX? Classifier were field-tested in late 2009 in coal applications. More recently, the technology has been applied in fine particle separation in minerals applications and there are a number of full-sized units operating in chrome applications in South Africa. Initially, pilot-scaled units (typically the RC?300) were trialled in iron ore, mineral sands, and manganese plants amongst other minerals, typically after other technologies failed to achieve the desired results. Currently a number of laboratories globally are carrying out more testing in minerals applications. More than 50 RC? units are now operating in coal and minerals applications. This paper introduces the REFLUX? Classifier technology, identifies commercial applications, and gives some commercial results.

Jan 1, 2014

By Willem J. Marais

The management of unpaved mine road networks on large surface mines rarely results in optimal road maintenance strategies and minimised total road-user costs. This is ascribed mostly to the complex and dynamic combination of variable road networks and loading and discharge points. In a dynamic mining environment - typically those mines in which production is managed by a centralised truck dispatch system ? there is no guarantee that a particular road maintenance intervention will contribute significantly to reducing total road-user costs or increasing productivity. Most large surface mines operating ultra-heavy mine trucks rely on an integrated on-board diagnostic data collation, communication and GPS-asset location system as a real-time fleet management tool. By extending this system to incorporate a multi-sensor analytical procedure in which specific truck vital signs are monitored and filtered, a road defect can be recognised and a trigger level set to relay the location of defects on the mine haul road to the centralised truck data management system. Previous work established the feasibility of the multi-sensor approach to road defect recognition on large mine haul trucks, and outlined the defect recognition, analytical and modelling issues and system limitations. This paper presents the development of the analytical procedure used as a basis for evaluating the truck on-board data to establish maintenance priorities amongst a network of mine roads. Following an introduction to the system architecture, the results of system field trials are analysed and the results discussed in the light of defect density and traffic volume as the primary variables in an approach to prioritising road maintenance. The paper concludes that by supplementing the existing mine communication and asset management systems, road maintenance can be managed on a near real-time basis and maintenance equipment dispatched to where most immediate benefit will realised from a maintenance intervention on the network, thereby generating the maximum improvement in service and reduction in cost per ton hauled.

Jan 1, 2008

Toxic gases, mostly carbon monoxide (CO) and nitrous oxides (NOx), are invariably generated by commercial explosives under practical conditions of usage because of oxygen unbalance and non-ideal chemical reaction. Permitted explosives are statutorily required to meet the criteria stipulated by regulatory authority for their toxic fume quality for their safe use in underground coalmines. Solid blasting using P5 explosives contribute around 60% production from Indian underground coal mines. Low pull and yield per blast in solid blasting has been identified as a reason for low production and productivity of Indian underground coal mines. In an effort to improve performance of solid blasting, it was envisaged to apply air decking between suitable P5 explosives using high density polyethylene (HDPE) spacers under a Ministry of Coal, Government of India funded project. A non-deflagrating slurry explosive composition having high air gap sensitivity was specially developed for this purpose, so that air decked cartridges get detonated sympathetically with single priming. However, statutory authority in India apprehended that the use of oxygen negative HDPE spacers can influence the generation of toxic gases. The effect of using HDPE spacers on the generation of toxic gases was studied under simulated laboratory conditions with newly developed as well as with three commercial P5 explosives. The results of studies presented in this paper revealed that the level of carbon monoxide increases linearly and there is no significant effect on the level of oxides of nitrogen due to the use of HDPE spacers with selected explosives. Studies also revealed that this newly developed slurry explosive can be used for air decking up to 15 cm using HDPE spacers of weight not more than 21 g without exceeding the permissible limits for toxic gases.

Jan 1, 2008

A large body of work exists discussing the impact of grinding media on mineral flotation. Generally, the work indicates that a change to a less electrochemically active grinding environment has positive benefits on downstream processing. There is a fear, however, that these benefits may not be realized when treating low sulphide containing ores, typified by UG2 style deposits. A series of experiments was developed to test the flotation response of UG2 ores using the Magotteaux Mill® to determine if measurable differences in the pulp chemistry could be discerned using a range of grinding media types. Further, the flotation responses of the contained sulphides (chalcopyrite, pentlandite and pyrrhotite) were determined. The data collected provided strong evidence that the grinding environment can significantly influence the pulp chemistry and flotation characteristics of all sulphide minerals, even when the ore contains less than 0.5 per cent total sulphide. The results show that the change to an inert media type produced a substantial improvement in the flotation rate of all sulphide minerals, as well as the PGMs.

Jan 1, 2009

By M. Gericke

The microbial leaching of metal sulfides is now an established biotechnological technology. Over the past 25 years, refinements in the engineering design of bioleaching processes have paralleled advances in our understanding of the diversity and role of the micro-organisms driving the process and the mechanisms by which micro-organisms enhance metal sulfide oxidation. Commercial success started with the treatment of refractory gold concentrates using mesophilic micro-organisms, followed by the development of tank bioleaching processes for the treatment of base metal concentrates. This was, initially, a mesophilic process with limited potential for recovery of copper from chalcopyrite concentrates due to slow rates and low copper extractions. The exploitation of thermophiles represents a major breakthrough in the development of bioleaching technology for the treatment of chalcopyrite-containing ores and concentrates. This development also opened the route to heap bioleaching of chalcopyrite ores, which is now a major focus of research programmes and piloting campaigns. This paper reviews the historical development of minerals bioleaching processes and gives an update on the current status of commercial tank and heap bioleach operations around the world.

Jan 1, 2009

By M. Nicolle

The Tati Activox® Project will be the first full-scale implementation of the patented Activox® process (The Tati Activox® project was deferred in 2008. Please refer to the Norilsk Nickel press issue on the topic for more info). The process was developed by Norilsk Process Technology and has been tested on Tati mine sulphide concentrates in laboratory, pilot and demonstration plant scales and demonstrated its viability. There are inherent risks to the final scale-up of the process from the demo plant and one of them will be investigated in this paper. Compartment 1 is designed to leach approximately 77% of the total nickel leached. For this reason agitation requirements in the first compartment of the autoclave are reviewed. An attempt is made to minimize the process and mechanical risks associated in achieving oxygen mass transfer into the slurry solution. The agitator powers for oxygen mass transfer are calculated using empirical correlations and compared to demonstration plant testwork. The resulting gassed power per unit volume (P/V) is higher than most commercial autoclaves and raises uncertainty on the viability of using such high unit power inputs. Additionally, there is concern about the ability of the autoclave shell to withstand and support the higher loading of large agitators. An alternative solution to designing for the increased P/V is assessed in which the number of compartments within the autoclave is reduced from 5 to 4 by removing the compartment wall separating compartments 1 and 2. This results in an enlarged first compartment containing 3 agitators instead of 2. Therefore the compartment 1 oxygen demand is supplied through 3 agitators, which lowers the P/V per agitator. The reduction in the number of autoclave compartments raises the potential for short-circuiting the mean flow pattern by slurry particles. Short-circuiting and low velocity zones could result in a lower recovery of metal and localized hot spots, respectively. A computational fluid dynamic (CFD) analysis was conducted to quantify these concerns and also to evaluate further design considerations. The results indicate that the proposed design change to 4 compartments affects short-circuiting. The impact of increased short-circuiting on the overall autoclave recoveries is not quantified, however; it is expected to be negligible based on testwork done in the Tati Demonstration Plant and similar modifications made to another operating autoclave. The CFD analysis also suggests that there will be no low velocity zones within the compartment.

Jan 1, 2009

By H. Gumede

The challenges of noise emission in the mining industry are well known. These challenges are exacerbated by the increasing trend towards mechanized mining, and there is general consensus that an effective industry-wide Buy Quiet initiative (BQI) will significantly assist in reducing noise-induced hearing loss by minimizing noise at its source. In its endeavours to improve the health of workers, the MOSH Noise Team has facilitated development of a proposed strategy and structure for a BQI. Extensive consultation with stakeholders resulted in support for the initiative. The MOSH Noise Team was advised to solicit the contributions of consulting mechanical and electrical engineers (CM&EEs), whose support is viewed as critical for realization of the initiative, since elimination of noise at source is an engineering challenge, and most procurement standards and specifications are set and managed by engineers. A workshop with CM&EEs and their representatives, as well as group environmental engineers (GEEs), confirmed the urgent need for the initiative and provided valuable insights. The workshop used obstacle based planning techniques to identify the factors on which the success of this BQI will depend. Most potential obstacles to the implementation revolved around proper scoping, leadership commitment, and companies? adherence to the initiative. Lack of involvement by the relevant stakeholders and economic issues were also identified as potential obstacles. Delegates at the workshop felt that the successful implementation of the industry-wide BQI, will depend on the mining industry addressing the specific obstacles. The most critical and urgent issue was identified as the formation of a properly constituted industry-wide BQI task team that includes representatives from critical stakeholder groups and which is mandated and empowered to drive the initiative.

Jan 1, 2014

By W. J. Kempson

As a result of the rising electrical energy costs in South Africa, a method was sought to reduce the overall electrical consumption of typical shaft systems. A typical shaft configuration was analysed and the primary energy consumers were identified. The ventilation fans for this system were found to consume 15% of the total energy of the shaft system. It was calculated that more than 50% of this energy is consumed by the shaft itself, more specifically, by the pressure losses that occur in the shaft as the ventilation air passes through it. In order to ensure that the theory being used for the evaluation of these shaft systems is accurate, a total of five shafts were instrumented and the actual pressure losses over the shafts plotted against time. These shafts were then analysed from a theoretical perspective. Finally, in order to ensure a thorough understanding of the behaviour of the ventilation air in shaft systems, the systems were simulated using computational fluid dynamics (CFD) techniques. There were significant discrepancies between the theoretical analysis and the CFD simulation during the initial comparisons. This discrepancy lessened as the complexity of the CFD models increased, until when the complete shaft was modelled using the full bunton sets, the pipes, and the flanges, the difference between the theoretical evaluation and the CFD simulation was small. This result demonstrates that the theory is insufficient and that the interrelated effect of the buntons and fittings has not been fully appreciated by current theory. The final phase of the work presented here was to evaluate the costeffectiveness of using different bunton shapes and shaft configurations. It is shown that the increase in the pressure losses and therefore the direct operating costs of the shaft can vary by as much as 80%, depending on the bunton configuration chosen. The placement of the piping in the shaft can increase the pressure losses, and therefore the direct operating costs of the shaft, by as much as 12%, depending on the placement of the piping in the shaft; this effect includes the use of flanges. The use of fairings on a large cage can reduce the resistance that the cage offers to the ventilation flow by as much as 30%. This, however, does not translate into a direct saving because as the cage moves through the shaft, the overall effect is transitory. These savings can be significant when the items highlighted in this work are applied correctly. Keywords shaft design, ventilation, energy consumption, modelling, CFD.

Dec 1, 2013

By K. M. Primrose

Electrical tomography has developed rapidly over the last few years in its sophistication and practical use to enable three-dimensional information on the distribution of phases within flowing mixtures to be observed1. This paper reviews the current state of application from the industrial perspective, drawing on applications case studies based on scaling studies in the laboratory right through to installations on plant. In the laboratory, applications have looked at scale-up of mixing2 processes involving gas-liquid and solid-liquid suspensions3 with a particular interest in verifying computational fluid dynamics simulations or in checking scale-up design4,5. In the processing plant, the circumstances are usually bespoke and may range from simple detection of the degree of homogeneity of a flowing mixture through to the need to measurement a sophisticated batch process such as crystallization, sedimentation, filtering and drying. Examples drawn from the chemical and pharmaceutical industries will be presented and related to the benefits that can be gained through application of the method.

Jan 1, 2008

By N. C. Gay

The quantitative evaluation of the performance of fill materials placed in gold mine stopes is achieved by (i) measurement of the in-situ load- displacement, or stress-strain, characteristics of the materials; (ii) an objective comparison between the observed fill performance and that predicted from suitable computer modelling techniques; and (iii) objectively comparing conditions in backfilled and conventional stopes. Considerable problems have been experienced with the development of sufficiently robust instrumentation which can be installed within filled pad- docks and give reliable information over a long period of time. However, at most mines where backfill performance is being monitored, it has been possible to correlate stress measurements within the filled paddocks with strain measurements adjacent to the paddocks. The resultant stress-strain curves agree relatively well with those determined in the laboratory. Measurements in the hangingwall strata of filled and unfilled stopes indicate that the placement of fill does reduce the amount of inelastic convergence due to bed separation. Moreover, in some stopes, the rate of convergence with respect to face advance does reduce to that of the theoretical elastic convergence. Quantification of changes in the conditions at the stope face due to placement of fill has proved difficult. Observations indicate that, provided the fill is placed close to the face (- 5 m), and adequate temporary face support is used, conditions in filled panels are superior. In addition, in filled panels the damage caused by a rockburst in the stope and, particularly, in gullies is generally much less than would be expected in an unfilled stope. To quantify these improvements, parameters such as stope height, face advance and the frequency of falls of ground are being monitored.

Jan 1, 1988

Primary roof support represents the first line of defence against rockrelated falls of ground in underground mines, and improper utilization or misunderstanding of the applicability and behaviour of primary support can be costly from a safety standpoint. This is a major concern for underground mines, as roof support is the single most costly expense from a mining operational perspective. This is further backed by the evidence that, in the USA, hundreds of injuries and fatalities still occur each year because of rib, roof, and massive roof falls. Additionally, the fully-grouted passive rebar, fully-grouted tension rebar, and resin-assisted mechanical anchor bolts, which constitute a large portion (89%) of the 68 million bolts installed each year in underground mines in the USA can vary in cost quite dramatically. To mitigate this concern a study was conducted in 2010 by the National Institute of Occupational Safety and Health, in conjunction with Southern Illinois University of Carbondale, to assess the performance of primary roofbolts in underground coal mines for improved safety and cost. This was accomplished using underground roofbolt monitoring solutions, field data, and numerical modelling to better understand the quasi-static behaviour of underground coal mine roofs and the response behaviour of the bolts. In particular, over 170 instrumented extensometers, closure meters, shear meters, fully-grouted passive rebar, fully-grouted tension rebar, and resin-assisted mechanical roofbolts were installed at three coal mines across the USA. Of these three mines, two used the room and pillar extraction method and the other used the longwall extraction method. There was no evidence to indicate a difference in performance of the active primary roofbolts compared with the passive primary roofbolts. Additionally, in the initial loading phase, the active bolts showed no difference in loading, indicating that tension bleed-off is of more of a concern than originally thought. Lastly, for the initial computer modelling studies, challenges still remain in obtaining a good match to the in situ bolt measurements and replicating the discontinuous roof rock and in situ bolt behaviour over time.

Jan 1, 2014

By J. B. A. Solar Menéndez, M. A. Martínez Fidalgo, L. A. Alvarez-Santullano

1. ANTECEDENTS This present communication refers to the results of the research project entitled "Training in the correct use of chemical oxygen self rescue apparatus and also an analysis of apparatus autonomy", carried out under the auspices of the Specific Collaboration Agreement signed between the Industry Council of the Principality of Asturias (Spain) and INSALUD (National Silicosis Institute, Spain). The antecedents of this project lie in work undertaken by the National Silicosis Institute aimed at optimising efficiency in the use of chemical oxygen self rescuers through improvements in worker training. To be precise, by means of Agreements drawn up between the above-mentioned parties, a simulator was designed and incorporated as the basis of a training station within the " Fundación Barredo" of the Industry Council of the Principality of Asturias. This has been operative since the year 2001 and, up to the present, 2300 Asturian coal mine workers have been trained there. What prompted this activity was the observation of various errors in the use of the self rescuers, deriving from their innate characteristics, especially with respect to the sensation of discomfort - the high temperature and humidity of the breathed oxygen, resistance and inconvenience - and limited autonomy. It was precisely this insight that, in critical situations, lack of familiarity with the conditions of discomfort likely to be experienced during the evacuation process might lead to rejection of the equipment, with fatal consequences, that has justified the setting up of this training and simulation centre. The centre is able to provide all mine workers with the necessary training, and experience of the characteristic sensations experienced when using such equipment (Fig.. 1).

Jan 1, 2003

By D. E. Sweeney

In May of 1980, a mine operated by Belmoral Mines Ltd experienced a failed crown pillar, resulting in the death of eight miners and putting sixteen others in harm?s way. The Ferderber mine was located in the Val d?Or region of the province of Quebec in Canada, and had been under development for 19 months and production for 9 months, since the opening of the portal. The crown pillar separating the 1?7 level from overlying overburden ultimately failed, permitting an estimated 1.5 million cubic feet of liquefied sediment to gain entry into the mine. This paper back-analyses the cave-in by applying decision error theory, in an examination of the organizational culture and the collective decisions contributing to the cave-in disaster. A case will be made for a greater understanding of mining methodology, geomechanics and risk assessment when operating in such challenging geologic conditions. Boundary condition and decision (BCD) analysis will be introduced as an innovative new method for the analysis of mine accidents and incidents. It will also be shown that BCD analysis can be applied to any event in an enterprise for which standards, norms or legislation exists. A cognitive profile will be presented that provides insight into the safety ethos existing within the Ferderber mine organization at the time of the tragedy.

Jan 1, 2007

By E. Dadzie

This paper will review some methods used to manage geotechnical risk at Damang Pit Cutback. The pit depth is designed up to 250m with push back on the eastern and western walls of the existed and depleted Railway Cutback (RCB) pit. The geotechnical design was presented by SRK Perth, focussing on the optimisation of the bench face angle which finally evolved with ?steep slope strategy?; 24m at 800 and 24m at 750 on the eastern and western walls respectively. The east wall was designed strategically to optimise the proximity of the tailings dam to the final pit crest of the push back. The pit has experienced failure in the weathered profile transgressing through the transitional material into the fresh rock at different location of the east wall since development in July 2005. To ensure stability and to minimise ore loss, the east wall has been redesigned using 700 batter angles and reduced berm width of 4m compared to 800 and 8.5 berm width of 24m bench height. The paper will also discuss how slope failures to the east wall had been sustained, managed and monitored to enhance production. Emphasis is placed on slope management at Damang which includes the principles of wall control blast, visual inspection, slope depressurisation and the state-of-the art of slope monitoring using Autoslope automated prism monitoring. Future slope monitoring is considering the use of 3D Laser and Radar to provide early warning so that evacuation could be effected. The intension is to provide mine personnel the means to identify risk areas during mining activities.

Jan 1, 2008

By G. P. Dyke

The MRMR classification system was developed specifically for mining applications, namely caving operations, and is one of three rock mass classification systems used in the South African mining industry today. Increased usage of the MRMR classification system has raised concerns that it does not adequately address the role played by discontinuities, veins and cemented joints in a jointed rock mass. To address these concerns, Laubscher and Jakubec introduced the In-Situ Rock Mass Classification System (IRMR) in the year 2000. Although the IRMR system is more applicable to a jointed rock mass than the MRMR system, a quantitative comparison of the MRMR and IRMR classification systems indicates that there is not a significant difference between the resultant rock mass rating values derived from the two classification systems.

Jan 1, 2008

By F. Smit

Mining and plant practices can benefit from preconcentration of UG2 platinum ores in a dense medium separation (DMS) plant. The main benefit of DMS is the selective discard of waste from the feed stream, thus improving project economics while reducing the power and water requirements for processing. On the mining side, with the correct combination of variables, implementing DMS as a pre-concentration step can reduce cut-off grades and, as a result, increase ore reserves. In addition, the use of DMS technology gives the mine team the opportunity to consider alternative mining methods for an operation. The challenges of balancing the mining tonnage, methods, and resulting grades with the process plant size, recoveries, and DMS discard volumes have resulted in the development of a model to optimize value on new projects or existing operations. This paper will focus on these benefits from a mining and plant perspective, quantifying the water and power savings, and look at the economic benefits of preconcentration and the effect of metal prices, plant feed grade, and DMS efficiency in different mining environments.

Jan 1, 2013

By Alexandra Amiel

Tarkwa Gold Mine is owned by Gold Fields Ghana, IAMGOLD and the Government of the Republic of Ghana in a 71%, 19% and 10% ratio. The mine is operated by Gold Fields Ghana. This large, low-grade, open pit, gold mining operation is located in Tarkwa, in the Western Region of Ghana, approximately 320 kilometres by road, west of Accra, the capital of Ghana, West Africa. In the Financial year ended June 2007, the operation mined a total of 108 million tonnes of which 22 million tonnes was ore. The mined ore is processed through three gyratory crushers supplying two heap leach pads (North and South) and a carbon-in-leach (CIL) plant. Before the 6 Sigma project commenced, the current fragmentation was resulting in sub-optimal feed size to these crushers. It was decided to use 6 Sigma methodology to improve the fragmentation. A defect was defined as a fragmentation analysis indicating less than 95% passing 750 mm. The objective was to improve fragmentation in order to achieve a constant throughput to the north heap leach crusher of 1500 t/hr, a 5% increase in pit loading efficiency and a 50% decrease in rock breaker hours, equipment damage costs and in-pit sheeting costs respectively. However, this was not to be achieved at the expense of grade control as Tarkwa is exploiting a low grade, high volume deposit and therefore ore dilution must be kept to a minimum. This paper is intended to take the reader through the various 6 Sigma phases of the project (Define ? Measure ? Analyse ? Improve ? Control) which have resulted in a 10% reduction in loading times, of which half has been attributed to improved fragmentation, as well as a substantial increase in blasted stocks. The focus has been on both drilling and blasting parameters as both processes had to be taken into account. These parameters are technical but personnel factors were also identified and dealt with as they arose. In order to measure the impact of any change in the process parameters on the fragmentation size distribution, cameras have been acquired. One has been installed looking into the bowl of the north crusher and the other one was dedicated to field observation, of individual blasts. Both are used for on-going fragmentation analysis. At Tarkwa, mining efficiency is strongly dependent on good fragmentation as well as the presence of un-blasted ?toes? on the pit floors. Crusher throughput depends on an acceptable ore fragment size. Compliance to plan with regards to actual hole depths was temporarily monitored by manual hole dipping, but, this is now being replaced with a fully automated Drill Data Management system developed by Sandvik and Modular Mining. The Six Sigma methodologies, using control charts based on data collected in the field, allowed the mine to improve methods with regards to the control of blasted material. It also enforced the implementation of a quality control programme encompassing the entire Drill and Blast process.

Jan 1, 2008

By L. Cortez, Á. Magalhães, F. Durão

Methods based on regionalized variables theory, commonly known as geostatistics, have been widely used since more than three decades to predict/estimate spatial distributed data, namely mineral resources, with quite good results. However, some restrictive theoretical assumptions, such as the need of some kind of stationarity of data and the linearity of the predictor/estimator, are often wrong options in many applications. An alternative approach based on a finite mixture modelling, or cluster weighted modelling (CWM) technique, is presented and compared, in its theoretical and practical aspects, with the geostatistical methodology. CWM is a general non-linear framework based on the estimation of joint density functions allowing the calculation of the conditional probability function of the target values given their locations. Validation tests are performed in a 2D case study, where testing data proceeds from a soil contaminated by a heavy metal. The results of ordinary kriging and CWM estimations show a very good agreement with actual data. Keywords: Non-linear regression, Geostatistics, Ordinary kriging, Finite mixture modelling, Cluster weighted modelling.

Jan 1, 2003