O uso da Cokrigagem Co-Locada (collocated cokriging) na integração de dados químicos e mineralógicos no minério de ferro

Abstract

This paper presents a coestimation technique - Collocated Cokriging - applied for integrating mineralogical and chemical data from iron ore. Adopting, as an assumption, a Markov Model for making the covariances modelling easier, the suggested procedure is made up of coestimating the mineralogical variable Goethite having as auxiliary variable Loss on Ignition. The application of this technique is based on: a) The existing statistical correlation between the two variables; b) Existence of a large amount of Loss on Ignition samples, making its direct covariance modelling easier. The proposed method has, as the most important characteristic, operational simplicity, emphasizing the easily direct covariance modelling of Goethite through the Markov Model 2. The technique was applied to iron ore minerals from a deposit located in the Iron Quadrangle (MG). Considering that the proportion of Goethite affects the comminution, pelletizing and reduction processes, the method is suitable for characterizing the iron ore in both resources modelling and mining control phases.

kriging; coestimation; mineralogy; data integration; collocated cokriging