An algorithm to simulate ore grade variability in blending and homogenization piles

Homogenization piles are largely used in the mining industry for variability reduction in the head grades of the ore feeding the processing plants. Various methods are used when designing homogenization piles, and most fail to incorporate the in situ grade variability that is intrinsic to the mineral deposit. The methodology proposed here combines longitudinal piles and geostatistical simulations to emulate the in situ variability as well as the reclaimed pile grade variability. A blending/homogenization simulator for longitudinal stockpiles has been developed, where the input to the algorithm consists of the pre-defined mining sequence, and has been applied to several simulated scenarios of the mineral deposit. This sequence is re-arranged by the algorithm, which selects the blocks that will form the pile of each simulated block model, and simulates the operation of the stacking and reclaiming equipment. The output sequence of the algorithm reflects the quality fluctuation of the grade after the blending/homogenization and reclamation. Using this methodology one can evaluate, within a certain time period, the expected grade variability for various pile sizes as well as the internal grade variability when a given pile is reclaimed. Results from a case study at two large iron mines operated by Vale S.A. proved the usefulness and functionality of the method. It is demonstrated that the rate of variability decreases as the pile size increases. It is also demonstrated that the internal grade variability decreases for a given pile size, when the number of layers in the pile is increased.

► How much the input grade variance can be reduced with different pile sizes?
► The space of uncertainty is mapped by using geostatistical simulation.
► Using a novel combination of geostatistical simulation and a blending pile emulator.
► The internal grade variability of a pile decreases when the number of layers increased.