Elman neural network based temperature prediction in cement rotary kiln calcining process

Cement rotary kiln calcining process is a kind of functional equipment for fuel combustion, heat exchange, and chemical reaction. A complex succession of chemical reactions takes place as the temperature rises. One can not establish a precise mathematical model of rotary kiln, so it is difficult to achieve its optimal control. In order to accurately reflect the system dynamic characteristics, we use Elman neural network to establish the model, because Elman network has the superiority to approximate delay systems and adaptation of a time-varying characteristics. We first in-depth analyze mechanism and working parameters correlation to determine factors of the yield and quality as the model input variables; then use Elman network construction rotary model, and compare the method with ordinary BP method. The results show that, Elman network has a faster convergence speed and high precision of the model; it can solve the problem of modeling for the cement kiln.