Estimation of Calorific Value of Lignite Field in Kütahya-Gürağaç (Turkey) by means of Artificial Neural Network

Artificial neural networks are generally information processing systems that mimic the working principles of human brain or central nervous system. a method gives successful results in solving many daily life problems such as classification, modeling and prediction. accomplish this by adjusting connection weights between neurons. It can solve prediction classification with back propagation algorithm, which is widely used artificial multilayer perceptron. In study, unknown calorific values were tried to be estimated using analysis (depth, ash, moisture, sulfur, value) drillings realized Kütahya -Gürağaç lignite field. An network was created for purpose. First, 8 neurons hidden layer network, 10 secondarily. learning function sigma, rate 95%, trained Levenberg-Marquardt training algorithm. The converged at desired margin error (1e-07) completed after 271 iterations. relationship actual calorie predicted reached high ratio R2=0.97. After completed, simulated estimation seams caloric values. As result, determined an average 97% confidence interval coal