Automatic classification of iron ore lithologies using petrophysical and geochemical data

Abstract

We use Fuzzy Inference Systems on a combination of petrophysical and geochemical data to automatically classify iron ore lithologies. Our results show that only two measurements are needed to group the data according to the major rock class and grade. Either Fe and Al, or Fe and Natural Gamma logs may be used, where the Al or gamma log are indicative of shale units. We propose a method to gather all data necessary for iron ore classification in a single down-hole logging run using Spectral Gamma-Gamma to provide a real-Time update of the iron ore resource model.