基于时变ARX模型的自适应预测函数控制  

Adaptive Predictive Functional Control Based on Time-varying ARX Model

作　　者：符莎 李平[1] 

机构地区：[1]辽宁石油化工大学信息与控制工程学院,辽宁抚顺113001

出　　处：《控制工程》2017年第4期756-760,共5页Control Engineering of China

基　　金：国家自然科学基金(612030021);辽宁省自然科学基金项目(2013020024)

摘　　要：针对非线性系统提出了一种基于时变ARX模型的非线性自适应预测函数控制算法。该算法通过T-S模糊将非线性系统转换为时变ARX模型,其模型的参数利用遗忘因子法在线辨识获得。然后与预测函数控制方法相结合,采用状态空间方法来重新构造时变ARX模型,从而减少了计算量,并且利用实际增益与模型增益相等来递推预测控制量的最优解。以磁悬浮系统为研究对象,在模型匹配和失配的情况下,将所提的控制算法与直接自适应预测控制算法进行对比研究。仿真结果表明,所提方法具有更好的跟踪性、鲁棒性和抑制干扰的能力。In this paper an algorithm about adaptive predictive functional control for nonlinear systems based on time-varying ARX model is proposed. The nonlinear system is converted to the ARX model by T-S fuzzy, the algorithm carries out on-line identification of model parameters with the forgetting factor method. Then combined with the predictive functional control theory, the state space method is used to reconstruct the ARX model, which reduces the amount of computation, and the optimal solution of the control volume is predicted by making the actual gain equal to the model gain. For the magnetic suspension system, the control algorithm and the direct adaptive predictive functional control algorithm are compared in the case of model matching and mismatching. The simulation results show that the proposed control strategy has better tracking, robustness and rejection of disturbance ability.

关 键 词：ARX模型 遗忘因子 自适应预测函数控制 状态空间 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]