Rem: Revista Escola de Minas
Print version ISSN 0370-4467
BERETTA, Filipe Schmitz; COSTA, João Felipe Coimbra Leite and KOPPE, Jair Carlos. Reducing coal quality variability with size optimization of blended piles. Rem: Rev. Esc. Minas [online]. 2011, vol.64, n.1, pp. 85-90. ISSN 0370-4467. http://dx.doi.org/10.1590/S0370-44672011000100011.
The in-situ grade variability imposes a series of problems during ore processing. Blending is commonly used to reduce this variability. The main factor to reduce head grade variability is the pile mass and size. The larger the pile, the lesser the reclaimed grades vary. However, there are practical limits to the pile length and mass. An incorrect design leads to financial losses due to either high capital costs for oversized piles or losses in plant recovery due to undersized piles incapable of attenuating grade fluctuations. Block estimation by traditional methods, such as ordinary krigging, does not properly incorporate the uncertainty associated with the estimates. Therefore, geostatistical simulation was selected. as it is the best suited to predict local grade fluctuations. The variables used to illustrate the proposed methodology were ash and sulfur content, for which the reduction of the fluctuation was analyzed for different pile sizes.
Keywords : Blending; piles; geostatistical simulation.