Ignition temperature and activation energy of power coal blends predicted with back-propagation neural network models

Abstract

Back-propagation (BP) neural network models were developed to accurately predict the ignition temperature and activation energy of 16 typical Chinese coals and 48 of their blends. Pearson correlation analysis showed that ignition temperature and activation energy were most relevant to the moisture, volatile matter, fixed carbon, calorific value and oxygen of coals. Accordingly, three-layer BP neural network models with five input factors were developed to predict the ignition characteristics of power coal blends. The BP neural network for ignition temperature gave a relative mean error of 1.22%, which was considerably lower than 3.7% obtained by the quadratic polynomial regression. The BP neural network for activation energy gave a relative mean error of 3.89%, which was considerably lower than 10.3% obtained by the quadratic polynomial regression. The accuracy of the BP neural network was significantly higher than that of traditional polynomial regression.