引用本文:何德峰,余世明.有界扰动多变量Hammerstein系统输入到 状态稳定模型预测控制[J].控制理论与应用,2019,36(4):605~612.[点击复制]
HE De-feng,YU Shi-ming.Input-to-state stabilizing model predictive control of multi-variables Hammerstein systems with bounded disturbances[J].Control Theory and Technology,2019,36(4):605~612.[点击复制]
有界扰动多变量Hammerstein系统输入到 状态稳定模型预测控制
Input-to-state stabilizing model predictive control of multi-variables Hammerstein systems with bounded disturbances
摘要点击 1978  全文点击 1109  投稿时间：2017-01-06  修订日期：2018-06-27
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2018.70012
  2019,36(4):605-612
中文关键词  HAMMERSTEIN模型  模型预测控制  约束控制  输入到状态稳定性  L2增益
英文关键词  HAMMERSTEIN models  model predictive control  constrained control  input-to-state stability  L2-gain
基金项目  国家自然科学基金,省自然科学基金
作者单位E-mail
何德峰* 浙江工业大学 hdfzj@zjut.edu.cn 
余世明 浙江工业大学  
中文摘要
      考虑具有状态和控制约束的有界未知扰动多变量HAMMERSTEIN系统，提出一种具有输入到状态稳定和有限L2-增益性能的鲁棒非线性模型预测控制策略。基于多变量线性子系统H?控制律，滚动预测非线性代数方程的解算误差，继而在线优化计算满足系统约束条件的预测控制量。利用输入到状态稳定性概念和L2-增益思想，建立闭环系统关于该扰动信号具有鲁棒稳定性和L2-增益的充分条件，使闭环系统不仅满足系统约束，而且对不确定扰动输入和解算误差具有鲁棒性。最后以工业聚丙烯多牌号切换过程控制为例，仿真验证本文算法的有效性。
英文摘要
      A robust nonlinear model predictive control strategy, with guaranteed input-to-state stability and finite L2-gain, is proposed for multi-variable HAMMERSTEIN systems subject to unknown but bounded disturbances and constraints on the state and control. Based on H? control laws of the multi-variable sublinear system of the HAMMERSTEIN model, the solution errors of the nonlinear algebra equation are predicted over a receding horizon. Then the predictive control actions are computed by online optimization. Using the concepts of input-to-state stability and L2-gain, some sufficient conditions are estabilished to guarantee the robust stability and L2-gain properties of the closed-loop system with respect to the disturbance signal. This implies that the closed-loop system not only statisfies the constraints but also has robustness with respect to the disturbace input and the solution errors. Finally, the exampel of industrial polypropylene multi-grade transition control is used to illustrate the effectiveness of the strategy presented here.