If you have access to OneMine as part of a member benefit, log in through your member association website for a seamless user experience.
|An important problem in mine operations is the classification of material as waste, low-grade stockpile and ore. This classification must often be made with blasthole data that are widely spaced and that have sampling errors. Geostatistical-simulation techniques combined with basic economic principles allow a procedure for classification that maximizes the expected profit. Geostatistical-simulation methods (Gaussian, indicator or annealing-based) allow the integration of hard and soft data in the creation of alternative, equally probable realizations of the mineral grades. At each location, for each realization one calculates the "profit" if the block were to be classified as ore or waste. The optimal classification is the one that maximizes expected profit (a maximum profit selec¬tion or MPS procedure). In this paper, the authors discuss the theoretical justification of the method and implementation details. The use of blasthole data from different types of mineral deposits, that is, different levels of continuity, are considered. The authors also show the efficacy of the procedure with different levels of sampling error. The increased revenue due to the MPS procedure and improved sampling is shown. The paper shows the geostatistical-simulation procedure and the uncertainty in block grades that result from incomplete and imperfect sampling. The optimal classification is presented. Optimal block classifications are transferred to realizable dig limits by hand-drawing polygonal boundaries. The results of the proposed method are compared to classification based on kriging. Limitations and areas of future work are identified.|