Semi-automated petrographic assessment of coal by coal grain analysis

Abstract

A new classification method, coal grain analysis, which uses optical imaging techniques for the microscopic characterisation of the individual grains present in coal samples is discussed. This differs from other coal petrography imaging methods in that a mask is used to remove the pixels of mounting resin to obtain compositional information of the maceral (vitrinite, inertinite and liptinite) and mineral abundances on each individual grain within each image. Experiments were conducted to establish the density of individual constituents in order to enable the density of each grain to be determined and the results reported on a mass basis. The grains were sorted into eight grain classes of liberated (single component) and composite grains. By analysing all streams (feed, concentrate and tailings) of the flotation circuit at a coal washing plant, the flotation response of the individual grain classes was tracked. This has implications for flotation process diagnostics and optimisation.