MINERALOGICAL ESTIMATION FOR AN IRON ORE DEPOSIT USING NORMATIVE CALCULATION FROM CHEMICAL AND DIRECT OPTICAL DATA

摘要

A geometallurgical model aims at bringing together data about mineraiogical types or minerals assemblage into a resource geological model. In case of iron ore deposits, optical mineraiogical data is available only for a few borehole samples. The principle of Mineraiogical Normative Calculation (MNC) is based on mineraiogical assemblage calculations in a closed system of mineral formation sequence. The MNC utilizes the chemical composition obtained from the samples to obtain sets of mineraiogical composition or assemblages. It is possible to obtain a relationship between optical microscopy and normative variables using multivariate linear regression. From this regression model the prediction on the mineraiogical assemblage can be made using chemical analysis results. The regression models were used to complete the database with mineraiogical assemblages predictions derived from assay readings. For this estimation it is necessary to define a group of rock types with mineraiogical similarity. The process chosen for this is cluster analysis using a Euclidian distance method. As these typological groups are defined their spatial extension is mapped using indicator kriging estimation (IK). Within each typological domain ordinary co-kriging is used to interpolate the mineraiogical proportion at each block of geometallurgical model. The result is a geometallurgical block model including mineraiogical and geometallurgical information. This model provides the means for long term mining and metallurgical processing planning. The model provides information related to ore quality and its metallurgical response during sintering or pelletizing processes.