The Inferential predictive control method of coal mill

Abstract

This paper presents the shortcomings of traditional PID method in coal mill control, such as multivariable coupling, nonlinearity and large-time delay, and give a new method to solve the problems. Firstly, the principles of coal mill is presented, and a coal mill model is established. Then we use prediction error method to identify the parameters of the model based on datas. Next, the paper estimates the coal outlet mass flow of the coal mill by using extended Kalman filter (EKF), and the result is compared with the coal flow instruction. Finally, we take the estimation of coal outlet mass flow as the new controlled variable and use inferential predictive control method to solve the coal mill control problems. The simulation results show that the new control method can shorten the delay time of the coal preparation system, and speed up the response of the boiler to AGC command.