基于MPC策略的磨矿监控系统

Supervisory System of Grinding Based on MPC Strategy

针对某选矿厂的两段磨矿回路,通过分析磨矿过程,确定了主要的输入输出变量,设计了基于模型预测控制技术的自动监控系统,其中监督层采用模型预测控制策略在线优化控制决策,控制层采用串级控制策略确保过程变量保持在设定值。工业应用表明,这种控制系统具有较好的稳定性和有效性。

In view of the two-stage grinding circuit in a concentrator, the paper analyzes the grinding process and determines the main input and output variables. An automatic monitoring system based on model predictive control （MPC） technology is designed, where the supervisory subsystem uses MPC strategy to on-line optimize the control policy-decision and the regulatory subsystem uses cascade control strategy to keep the process variables at their set values. The industrial application indicates that this control system has good stability and effectiveness.