【作者】 沈鸿宇； 彭旭东； 徐祖华； 赵均； 邵之江；

【Author】 Hongyu Shen;Xudong Peng;Zuhua Xu;Jun Zhao;Zhijiang Shao;State Key Laboratory of Industrial Control Technology,College of Control Science and Engineering,Zhejiang University;Hangzhou Oxygen Plant Group Co.,Ltd.;

【机构】 浙江大学控制科学与工程学院工业控制国家重点实验室； 杭州制氧机集团股份有限公司；

【摘要】 批次过程具有时变特性,使得其建模过程成本高、难度大,进而增加了控制器参数整定难度。采用无模型优化方法可避免批次过程建模相关问题,提高控制器优化效率,降低优化难度。本文采用一种无模型优化算法--极值搜索算法对批次过程中的时变PID控制器参数进行寻优,分别利用线性插值和三次样条插值方法对PID控制器进行参数化,较之于多项式形式的PID控制器,插值方法不仅使得寻优参数具有物理意义,同时便于初值设置。针对极值搜索算法可能出现的发散问题,提出了非负约束和增量约束方法以保证其鲁棒性和可靠性。文章最后通过反应器温度控制实例的仿真验证极值搜索算法对时变PID控制器实时优化策略的有效性。更多还原

【Abstract】 Due to the time-varying characteristics of batch process, the modeling process is costly and difficult, which increases the difficulty of controller parameter tuning. The model-free optimization method can avoid the problems related to batch process modeling, improve the efficiency of controller optimization, and reduce the difficulty of optimization. This paper uses a mode-free optimization algorithm--extremum seeking algorithm for the time-varying PID controller parameters optimization in the batch process, respectively, using the method of linear interpolation and cubic spline interpolation to parameterization of PID controller. Compared with polynomial form for PID controller, the interpolation method not only makes the optimization parameters of physical significance, but also facilitates the initial value setting. For the divergence problem of extremum seeking algorithm, non-negative constraint and increment constraint methods are proposed to ensure the robustness and reliability of the extremum seeking algorithm. Finally, the effectiveness of the extreme seeking algorithm for the real-time optimization strategy of time-varying PID controller is verified by the simulation of a reactor temperature control example.更多还原