Classification of coal seams with respect to their spontaneous heating susceptibility—a neural network approach

Abstract

The paper presents the application of adaptive resonance theory of artificial neural networks (ANN) for classification of coal seams with respect to their proneness to spontaneous heating. In order to apply this technique, 31 coal samples have been collected from different Indian coalfields covering both fiery and non-fiery coal seams of varying ranks spreading over 8 different mining companies. The intrinsic properties of these samples have been determined by carrying out proximate, ultimate and petrographic analyses. The susceptibility indices of these samples have been studied by five different methods, viz. crossing point temperature, differential thermal analysis, critical air blast analysis, wet oxidation potential difference analysis and differential scanning calorimetric studies. Exhaustive correlation studies between susceptibility indices and the intrinsic properties have been carried out for identifying the appropriate spontaneous heating susceptibility indices and intrinsic properties to be used for classification of coal seams. The identified parameters are used as inputs and adaptive resonance theory of ANN has been applied to classify the coal seams into four different categories. This classification system will help the planners and practising mining engineers to take ameliorative measures in advance to prevent the occurrence of fire in mines.