This study aims to predict the elemental compositions of 104 thermal coals with artificial neural network(ANN), which is based on proximate analysis properties. ANN model was optimized with six activation functions and four hidden layers after evaluat...
This study aims to predict the elemental compositions of 104 thermal coals with artificial neural network(ANN), which is based on proximate analysis properties. ANN model was optimized with six activation functions and four hidden layers after evaluating various performance. Best topology was set, and the relative impact on calculation accuracy was derived as a result to estimate chemical composition. performance was improved by 4.71%~0.91% from previous research. Then relative impact based on coal rank devision and optimized model was derived.