联合粉磨系统基于T-S模糊模型广义预测控制

Predictive control for cement combined grinding system based on T-S fuzzy model

针对磨机负荷多变量、强耦合、大滞后的控制难题,利用磨机电流的历史数据进行T-S模糊建模,对所建模型设计广义预测控制器,采用对控制器模糊加权的方法调节控制器输出,通过Matlab进行仿真比较。结果表明:广义预测控制达到稳态所需时间约为25 s,而PID控制需要约90 s,广义预测控制跟踪设定值的快速性更好,能更快达到控制要求;该方法避免了反复求解不同丢番图方程的繁琐过程,简化了控制器运算步骤。

For the problem of multi-variables, strong coupling and large delay of mill load of a grinding system, a T - S fuzzy model based on historical data of mill current was built and a generalized predictive controller for the model was designed. The output of the controller was adjusted by the fuzzy weighting method. By Matlab simulation ,it is shown that the time that a grinding system controlled by the generalized predictive controller reaches a steady state is about 25 s, while the PID controller takes about 90 s. This method avoids the complicated process of solving different dioaphantine equations repeatedly and simplifies the operation procedure of the controller. The superiority of the generalized predictive controller is verified.