An Artificial Neural Network for Classification a Quality of a Coal Fuel in Combustion Chambers Using FPAA

R. Suszyński, J. Marciniak, K. Wawryn (Koszalin Univ. of Techn., Poland)

A hardware artificial neural network for classification a quality of a coal fuel in combustion chambers is presented in the paper. Proposed method is based on an analysis of measured combustion process parameters in the chamber by the feedforward artificial neural network. Measured parameters have been used to train neural network weights with a help of MATLAB program. The preconditioned conjugate gradient algorithm with the Polak-Riberie formula has been used to weights training process. Calculated weights have been used to determine the quality of the coal fuel loaded into the chamber. The ANN has been tested by the MATLAB program and the FPAA implemented network. Obtained results are presented and discussed.