Estimation of Calorific Value of Coals Using BP and Elman Neural Networks

Abstract

Coal quality analysis is an important part in the fuel management of power system. The coal calorific value (CCV) is the basis to appraisal of coals heat balance calculation, coal consumption, heat efficiency and improving of heat utilization. There exist many methods on estimation of CCV. Artificial neural networks have many advantages in this area. This paper described the network structure, the mathematical model and the algorithm flow of BP neural network (BPNN) and Elman neural network (ENN) which are both used for the CCV estimation. Results show that both BPNN and ENN can well reflect the non-linear relationship between CCV and other factors of coal quality. And ENN is more accurate to predict CCV, with smaller absolute error.