The Research And Application Of Advanced Control Algorithm Of Ball Mill System

The stability and economic operation of ball mill system is of great importance for improving the economic benefits of the thermal power plant. However, the ball mill system is a control object with the characteristics of strong-coupling, high-inertia, large-delay, and strong-disturbance, and it’s difficult to implement conventional control algorithm in it. Therefore, it’s necessary to study advanced control algorithm for ball mill system in order to improve its control quality and economic performance.An improved decoupling generalized predictive controller for the ball mill system is proposed based on the analysis of dynamic characteristics of each control loops. Then, the scheme of advanced control and optimization of ball mill system is designed and applied in concrete object successfully. Practical operating results show that this scheme can effectively inhibit the coupling effects between various control loops, strengthening system’s ability to restrain disturbance, improve the control precision, and reduce the power consumption of unit coal pulverization.The research work in this dissertation mainly includes four aspects as follows:(1) An improved feed-forward decoupling internal model controller is proposed. By adding a delay-compensation unit and feedback channels, the capability of restraining disturbance and robust stability of the decoupling controller are effectively improved. Theoretical analysis shows that this proposed decoupling controller can achieve completely decoupling under the condition of model matching. The simulation results of ball mill system with white noise disturbance and multiplicative output uncertainty show that the decoupling controller has good performance.(2) An improved stair-like generalized predictive controller for each control loops of the decoupled ball mill system is proposed. The high-inertial and large-delayed characteristic of the ball mill system will lead to obvious oscillations of control variables during the transient process, which would affect the regulation time of the control system and cause frequent movements of actuators, increasing the abrasion. Therefore, this dissertation adjusts the value of output prediction by introducing an output weighting factor. Simulation results show that this control algorithm can effectively abate the oscillations of control variables, decrease the regulation time and reduce the overshoot of the control system.(3) An advanced control and optimization scheme of the ball mill system is proposed based on the improved decoupling generalized predictive controller, taking the characteristics of the ball mill system into account. The ball mill system is ameliorated by modifying the logic and configurations of the original DCS control platform. The excogitation includes:logical judgment of the advanced control on/off switch, schematic design of on-line optimization, estimation of blocking and coal plugging of the ball mill system and its emergency processing procedure. This control system with excellent reliability and friendly HMI realized the functions of automatic detection, bump-less switch, judgment of the condition of operation and tripping, real-time monitoring. Practical operating results show that the quality of ball mill system is significantly improved and the power consumption of unit coal pulverization is substantially reduced.