Search Documents

By J Zhou

This paper discusses the implementation of a mixed integer linear programming model for a coal preparation plant. The plant considered consists of primary and secondary heavy media cyclones, hydrocyclones for intermediate fines and froth flotation for the very fine material. The model is designed to optimise the plant settings and product flows to maximise revenues for a given specification for two product outputs. The model can deal with bonus/penalty aspects of the product pricing for products that exceed or fail to meet specification. We report on a model with two separate coal feeds with parallel circuits for the oversize, intermediate and fines streams. The genesis of this model is more than 30 years old but recent improvements in mixed integer programming solvers make a full mixed integer formulation practical. The original model began as an enumeration of cut points with arbitrary rules for product allocation. The product allocation was improved in a linear programming formulation by Gunn (1982) and Sarker (1991 and 1995). Current mixed integer programming capabilities make it possible to replace the enumeration aspects and solve directly as a mixed integer programming problem. The current implementation uses a general purpose modelling language to formulate the model. Open source and commercial mixed integer programming solvers are compared

Sep 26, 2011

By D Misra, R Ganguli, S Bandopadhyay

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

Sep 26, 2011

By A Melkumyan, F Ramos

This paper addresses the problem of prediction of grade variables over volumes (block estimation) for an autonomous mine. Resource estimation is usually concerned with the prediction of the quality of material over volumes (blocks in resource models). This problem is tackled by developing Bayesian block estimation with the Gaussian processes (GPs) framework. GPs are commonly used for nonparametric regression and classification and offer an elegant solution to deal with incomplete knowledge and information. A GP is defined by its mean and covariance function. The parameters of the covariance function can be learnt by maximising the marginal likelihood of the data. This offers a way of automating the covariance fitting process, while at the same time allowing for an assessment of how well the covariance function fits the data. Within the GP framework the calculation of the average grade over a volume can be performed using Bayesian quadrature, which treats an integral as a random variable. First, it is demonstrated that the block kriging equations in geostatistics are the same as Bayesian quadrature. Then, it is determined when closed form analytical solutions for the integrals are possible and quadrature approaches for approximations are briefly discussed.

Sep 26, 2011

By S D. Dessureault

In one of North AmericaÆs largest mines, which contains more than six active coal-producing faces, mine controllers must meet the complex challenge of adhering to the mine plan while also ensuring customer quality. Fleet management systems (FMS) provide many different avenues to control production flows, from the faces to the various pit crushers. Nonetheless, mine controllers must balance a variety of factors such as unplanned downs, differing customer demands, their own daily key performance indicators (KPIs) related to equipment utilisation, among others. Deviating from the short-term mine plan is the most common means of balancing the other factors, since adherence to the plan is traditionally only measured during the weekly variance meeting with engineering staff. However, such deviation may be one of the most costly responses, as long-term scheduling issues may arise such as multi-face dead-heading or insufficient stripping. This paper focuses on the development of a system for real-time adherence to mine plan within a control room environment. This system is part of a larger blending-focused, multi-million dollar project with elements of a centralised remote control room, developed at the University of ArizonaÆs Mining Information Systems and Operations Management (MISOM) Mine Control Center (MC2). The mine plan (MPL) component of the overall blending system integrates both historical and real-time FMS data with mine plan requirements directly sourced from original systems (no additional input required). The measures, displayed largely as graphs, have tools to help controllers manage the production to facilitate adherence to the plan. Some measures are designed to motivate crews to perform by inciting competition between crews, softened through a study of the controller culture. The MPL component was deployed at the mine, and the before and after deployment adherence was measured. This paper discusses the technical development of the new tool, its deployment, and an ethnography (study of culture) to understand controller motivation.

Sep 26, 2011

By S Van Duin, I Porter

This paper discusses how advanced manufacturing technology (AMT), which is fundamentally reliant on and developed via multi-disciplinary skills, has benefitted non-mining industries, and what lessons can be derived from this experience to benefit the mining industry, particularly underground coal mining. Examples of AMT use in æconventionalÆ industry, some drawn from the more than 25 years experience in many different industries of the Automation and Robotics team at the University of Wollongong, are described to show typical benefits, risks and opportunities expected in using AMT. Applying AMT to mining brings many unique challenges and this paper expands on these and then provides practical examples of how a multi-disciplinary team can develop new processes and systems leading to productivity, safety and other benefits for underground coal mining.

Sep 26, 2011

Longwall mining operations have been in place in Australia since the early 1960s. This mining method is currently used to extract the bulk of coal obtained from underground mines in Australia (CSIRO, 2010). Despite the high tonnages extracted the expected utilisation of much equipment remains relatively low. To some extent this level of utilisation is influenced by the variability in conditions inherent in the mining process. This variability in production rates presents challenges to those attempting to design a coal clearance system. The difficulty is to design a system that has sufficient capacity to support production, without over capitalising or incurring high operating costs. The need to allow for a series of largely unplanned events in the design of a coal clearance system makes simulation modelling an ideal technology to employ. Simulation modelling is a methodology that allows a system with interacting components to be modelled over time, while explicitly taking into account unplanned events that may disrupt operations.This paper discusses how dynamic simulation models have been used to assist in the design of coal clearance systems. Typically this type of modelling has been used to look at the likely profile of production rates from the Longwall shearer and to determine how well various proposed combinations of belts, bins and stockpiles would support production.

Sep 26, 2011

By A Huebner, A Noack

Companies all over the world are very sensitive concerning their IT and network security. In most network connected technical installations and machines however, network security is not taken into account to a significant extent: insecure protocols are run, unencrypted information is exchanged freely and even broadcasted wirelessly and networked machines are accessible via insecure file transfer protocol (FTP).What happens in case of malfunctions of such technical installations? Who is responsible? Who can be held liable in case of accidents? The first computer worm for automation systems (æSTUXNETÆ) (Symantec, 2010) was discovered during 2010.The paper is intended to create sensitivity among decision makers for potential security and resulting safety leaks and points out the current status quo in technical network security. It outlines best practices for secure networking and procedures for remote access and remote software updates on stationary equipment and mobile mining machinery. It describes how to ensure authentic and genuine configurations and software targets and how to ensure secure communication for autonomous and remote controlled equipment. For such equipment in particular, the following statement should apply: æThere is no safety without network securityÆ.

Sep 26, 2011

By E Rubio, N Morales

A common practice in the mining industry is to decompose the planning process into different tasks, so the overall process and specific schedules can be constructed easily. Nevertheless, currently there are some adverse effects, such as not capturing the real value of a project by not taking into account the construction of several underground drifts and general infrastructure. Then the common practice is to compute mining sequence and cut-off grade profiles over the life of the mine without constitutively integrating the development schedule as a means to measure the constructability of a given mine design and sequence. Thus, the investigation summarised in this paper shows a novel, integrative way of treating an underground mine sequence and production schedule while considering drift development scheduling. Moreover, it is thought that, because of this original model, the whole underground production scheduling exercise must concentrate on combining the development and the production scheduling to compute a robust and a feasible production schedule.In order to quantify the effect of the conventional way of viewing the planning, the concept of the building information model (BIM) needs to be used, which was postulated by Charles Eastman and has been used extensively since the late 1970s. BIM takes other aspects of a project into account, considering at the same time the spatial relationships, quantities and properties of building components, but especially some kind of model visualisation, which is the advantage to be incorporated. This paper presents the basis for an optimisation model that allows for defining the sequence of the mining development so that the production plan can be achieved. For this, the model considers the space dimensions, time, pattern or strategy, cost or benefit, precedence and resources available. Tests were performed considering a block caving and sublevel open stoping method. Finally these tests showed the main conclusion, that production sequence changes when the building capacity is considered.

Sep 26, 2011

By J Ferreira

To speed up multivariate geostatistical simulation it is common to transform the set of attributes into spatially uncorrelated factors that can be simulated independently. One decorrelation method is the method of minimum/maximum autocorrelation factors (MAF), where a two structure linear model of coregionalisation is assumed. The transformation decorrelates the theoretical model exactly, but the actual data are only approximately decorrelated. The MAF transformation can therefore be seen as a special case of a non-orthogonal approximate diagonaliser of a set of symmetric matrices. A more general approach for approximate joint diagonalisation (AJD) has been developed in the context of blind source separation. For these AJD algorithms no assumptions are made beyond symmetry of the individual matrices and so they can be applied to a family of experimental semivariogram matrices. In their application there are no restrictions on the number of structures in the linear model of coregionalisation (LMC) thus removing one of the conditions placed on the subsequent modelling of the spatial structure of the factors. In this paper the general background for the U-WEDGE AJD method is presented. It is shown that the U-WEDGE method achieves better spatial decorrelation than the MAF method. The factors derived from U-WEDGE are simulated using conditional turning bands simulation and then backtransformed to attribute space.

Sep 26, 2011

By F Grobler

Many projects fail because project performance in unfavourable economic conditions was not anticipated at the investment decision stage. This type of failure may have been avoided through the selection of a strategy that enables the project to perform well under a range of possible future economic conditions (a ærobustÆ strategy); rather than an æoptimalÆ strategy for expected future conditions that will never eventuate. This paper describes a method of rapidly generating mine plans for a number of project strategies, and then evaluating it against a representative range of equally likely economic conditions. The key performance indicators derived for each strategy are the mean value and the distribution of values (variance). A prudent investor will identify prospective strategies not only on the merit of highest mean values (as a proxy for economic reward), but also on the probability of not meeting specified investment hurdles (a proxy for risk). A robust strategy will achieve its financial objective even under adverse economic variations. This paper focuses on the effect of commodity price uncertainty; particularly around the choice of mining and processing capacity levels. Results from enumerating 2000 prices scenarios against 63 strategy combinations show how processing and mining capacity can be manipulated to manage the risk-reward trade-off.

Sep 26, 2011

By H Khoshrou, A Siamaki

Blast designs can have productive and non-productive impacts on downstream stages, mine productivity and operating costs. On the other hand, ground vibration, fragmentation, and back break caused by blasting impose damage and financial penalties and which must be controlled by blast design. One of the significant variables in blast design is the burden. In this study, the potentials of artificial neural network are investigated in prediction of burden in the Sungun copper open-pit mine. Input data were assembled through 18 blasting blocks according to different levels and experimental geomechanical investigation. To construct the model blastability index, uniaxial compression strength, hole diameter, specific weight of rock, rock quality designation and cohesion strength are taken as input parameters, whereas, burden is considered as an output parameter. Mean square error was used as the performance function and back propagation algorithm as the training function, containing four hidden layers and 14 data sets. Four sets of data were used to make sure that correct training had been carried out. This produced the coefficient correlation of 0.662.

Sep 26, 2011

By B Olsson, R Norman, M Volz

Australia has seen unprecedented demand for coal exports in the last ten years and this demand is set to expand further still according to many analysts. This rapid increase in demand has placed pressure on multi-user coal chains to guarantee the capability to meet the required demand, in terms of both current operations and capacity expansion.The projects required to deliver these forecast tonnages take place in an increasingly complex commercial and technical environment. These challenges include brownfield and greenfield projects, multiple rail and terminal service providers as well as a myriad of mine operators, all needing to operate in an increasingly integrated manner to get the most out of expensive shared assets.Increasing complexity and integration drives the need for more holistic and sophisticated analysis tools. The capacity of coal chain subsystems, such as mine, rail and port, can no longer be determined effectively in isolation. Understanding the nature of the interactions between these subsystems and the resultant capacity of the integrated system is the key to maximising value. TSG advocates the use of a total integrated system approach to understand the behaviour and performance of complex coal chains.Recently, TSG has worked in conjunction with the Integrated Logistics Company and a number of stakeholders in the Dalrymple Bay Coal Chain to deliver the first independent system master plan. This master plan has provided a quantitative and rational assessment of coal chain capacity in its current state and a prediction of its capacity once a pipeline of committed capacity improvements is in place. It is underpinned by a highly detailed discrete-event simulation model of the complete coal chain as an integrated system, including mine load-outs, rail, terminal and port operations.This paper provides an overview of the study and describes the differences between subsystem and total integrated system model performance.

Sep 26, 2011

By A Galli

Traditionally pit optimisation and scheduling is carried out assuming that the resources, the commodity prices and production costs are all known. Once the pit contour and the mining schedule have been determined, geostatistical simulations can be used to assess the impact of the uncertainty in the resources on the pit design and the schedule, and likewise stochastic simulations can be used to test its robustness to fluctuations in costs and prices. But it would be more interesting to have a dynamic procedure for optimising the pit as the resources become better known and as economic factors such as commodity prices and costs evolve over time. This paper presents a novel approach for doing this. Rather than working with millions of small selective mining units, it was decided to work with large blocks, called macro-blocks. Mining sequences are defined as the quantities to be extracted from each macro-block in each time period. Feasible mining sequences are mining sequences that respect the accessibility constraints, and also the geometric constraints on the amounts that can be extracted from any one block over time, or from all the active blocks at a given time. Having defined feasible mining sequences, their mathematical properties were studied. It was found that natural breakpoints occur in the sequences when one set of blocks have been completely mined out, thereby opening up access to others. It was discovered that families of sequences exist, and that they are convex and closed under certain types of row permutation operations. Convexity is an advantage when optimising. The closure property opens up the possibility of mixing and matching subsequences of rows in the sequences. Another interesting property is that the sequences form branching families, rather like the binomial and trinomial trees used to model prices in finance and in real options.It is easy to generate large numbers of feasible mining sequences randomly. A two-step procedure is used to find the best one (or the best few) using three economic criteria: expected NPV,the probability of a negative NPV, and the percentage of years when the cash flow is negative. The procedure was tested on a synthetic gold mine based on the characteristics of the Essakane mine in Burkina Faso. The results were encouraging.

Sep 26, 2011

By C Niemann-Delius

The problem of long-term open pit mine planning is a large combinatorial problem, which canÆt be solved easily by mathematical programming models because of the large problem size. In this paper a new metaheuristic algorithm, which has been developed based on the ant colony optimisation (ACO) and its efficiency, have been discussed. To apply the ACO process to a mine planning problem, a series of variables is considered for each block as the pheromone trails that represent the desirability of the block for being the deepest point of the mine in that column for the given mining period. During implementation several mine schedules are constructed in each iteration. Then the pheromone values of all blocks are reduced to a certain percentage and additionally the pheromone value of those blocks that are used in defining the constructed schedules are increased according to the quality of the generated solutions. By repeated iterations, the pheromone values of those blocks that define the shape of the optimum solution are increased, whereas those of the others have been significantly evaporated. By using this algorithm a high degree of complexities could be considered in objective function and constraints of optimisation model. In addition it can improve the value of the initial mining schedule up to 34 per cent for some cases in a reasonable computational time.

Sep 26, 2011

By E Rubio, N Morales

Currently, mine plans are optimised by using several criteria as objective functions, like profit, life-of-mine, concentration of some pollutants, mining costs, confidence level or ore resources, with consideration of constraints related to production rates, plant capacities and grades. Whilst this approach is successful in terms of producing high value production schedules, it uses a static opening sequence of the drawpoints and therefore the optimisation is made within the level of freedom left by the original opening schedule, and as a result it is far from the true optimal value of the project. This paper presents a model to optimise the sequence of drawpoint opening over a given time horizon. The emphasis is in the precedence constraints that are required to produce meaningful operational sequences considering the exploitation method (panel or block caving), physics considerations and logical rules. Further on, while it applies the standard approach of maximising net present value (NPV), it considers other targets for optimisation, like the robustness and constructability of the plans. Finally, it applies the model to real data to obtain feasible plans by means of this model. The first tests were done varying the production capacity of the mine.

Sep 26, 2011

By I Douglas, L Allen

Localised indicator kriging was developed following a similar approach used by Marat Abzalov (2006) in his paper titled æLocalised uniform conditioning (LUC): A new approach for direct modelling of small blocksÆ. Abzalov introduced the idea of mapping panel distributions from uniform conditioning onto selective mining unit-sized blocks that were created using ordinary kriging. This concept of remapping panel distributions was applied to multiple indicator kriging histograms.At Newmont a program was developed that made the remapping of MIK histograms to SMU sized blocks extremely efficient and was subsequently tested on several gold deposits. The results have been encouraging as the localised indicator kriging (LIK) method has eliminated many issues that have plagued both ordinary kriging and multiple indicator kriging. Use of localised indicator kriging methodology limits the smearing of high-grade and the over-smoothing of grade that is common and difficult to control in OK but retains the ability for localising grades. LIK also simplifies change-of-support corrections, eliminates un-mineable slivers of low or high-grade when dealing with small tonnages, and allows for the mapping of linear and non-linear estimates into the same block.

Sep 26, 2011

By S D. Dessureault

The largest surface coal mine in North America requires complex materials handling infrastructure to fulfil the highly variable quality specifications of its contracts. These diverse demands, reflected in the arrival of numerous customer trains every day, necessitate the integration of information in every stage of production. To meet this goal, quality measurement, train line-up and mine planning processes are being integrated into local and remote control rooms. The effectiveness of such a system can only be measured by comparing decisions that are made with, and without, each of the individual integrated control systems. Traditionally, mine controllers use their experience, intuition or a variety of trial-and-error spreadsheets to determine the ideal materials handling flows needed to achieve the correct quality. These balanced decisions are complex since many factors must be considered such as the mine plan, fleet costs, quality variability and others; consequently mine controllers usually have widely varying success between crews.An evaluation model was developed to assess the impact of a decision support system (DSS) integrated in a control room for controlling the coal blending process. To this end, an ARENA-based simulator representing the whole production process can be used to pull performance and current materials handling settings directly either from the real-time systems in the mine, or from historical data. In this way, the same shift (breakdowns, cycle rates, etc) can be simulated and performance compared. Different settings of the DSS, in terms of what information is displayed and the methods for doing so, can be tested for effectiveness. The historical data are extracted directly from a corporate data warehouse, without altering the everyday working databases. At the same time, the simulator writes its results back into the same data warehouse structure, allowing the same KPI reports to be used to measure simulated shifts.

Sep 26, 2011

By A Noack

Today underground infrastructures, like tunnels and especially wide spread underground mines, are often equipped with ethernet based communication systems. These networks are used for the unified communication of data, voice and video. As these networks stretch along the major tunnels and in increasing scale also into the production areas, a high degree of coverage can be reached. For wireless coverage, standard IEEE 802.11 WLAN access points are being used. In order to provide a maximum availability and reliability, the active components are equipped with battery backup power supplies and the network infrastructure is set up in a redundant way using ring or meshed topologies. This provides a much higher availability compared to traditional phone systems or dedicated data links using a conventional star topology.Distributed computing intelligence that is installed on active network nodes throughout the whole network, additionally supported by a central server system, creates the basis for new mine safety functions introduced in this paper for both coal and non-coal mining. These new safety functions cover the detection of potential emergencies, using the status of the network as a æsensorÆ, as well as dynamically guiding people to safe exits or rescue shelters in case of an emergency. Furthermore, these functions are of important benefit for improving the efficiency of search and rescue operations following a major incident or accident underground.

Sep 26, 2011

In this paper a quantitative deterministic risk modelling approach (Quadrim) is introduced. The mathematical framework and equations for quantifying uncertainty within and aggregating across value chain nodes from error propagation techniques is proposed and discussed. A case study and worked examples, based on a hypothetical mine are then presented. This illustrates how uncertainty from the mining and economic value chains can be handled in a typical net cash flow calculation, and how this can be translated to a risk model. It is then illustrated that current valuation techniques do not just result in an information gap for decision makers. Traditional approaches such as discounted cash flow (DCF) analysis and truck and shovel productivity analysis, which do not take into account uncertainty in input values, can result in potential bias in estimates such as net present value (NPV) or equipment hours

Sep 26, 2011

By J V. Mehrabani, A R. Sayadi

Adequate estimation of equipment costs is a key factor in a feasibility study and in evaluation of design alternatives for mining and mineral processing projects. Inaccurate early cost estimation could lead to the elimination of investment benefits. However, this estimation is complex and time consuming because of data accessibility problems and consequently access to a rapid, up-to-date and accurate tool is desirable for carrying out cost estimations.This paper describes the development of models for estimating capital and operating costs for a jaw crusher, which is used as the primary crusher for a wide variety of mineral beneficiation plants, applying single and multivariate regression analyses, together with the design of experiment approach, based on principal component analysis. Explanatory variables include gape size, length of feed opening and power as numeric (quantitative) variables and toggle type (single or double) and the existence of extra heavy duty construction as categorical (qualitative) variables. The performance of each model has been investigated and comparing the estimated R-squared (R2) of each model, the best was selected and its validity examined. The cost models include the capital and operating costs along with the operating cost components. The results indicate that the suggested best cost model is sufficiently accurate for use in feasibility studies.

Sep 26, 2011