引用本文:孙浩杰,邹涛,张鑫,惠存万.基于Hammerstein-Wiener逆模型补偿的预测控制 非线性变换策略[J].控制理论与应用,2020,37(4):705~712.[点击复制]
SUN Hao-jie,ZOU Tao,ZHANG Xin,HUI Chun-wan.Nonlinear transformation strategy of predictive control based on Hammerstein-Wiener inverse model compensation[J].Control Theory and Technology,2020,37(4):705~712.[点击复制]
基于Hammerstein-Wiener逆模型补偿的预测控制 非线性变换策略
Nonlinear transformation strategy of predictive control based on Hammerstein-Wiener inverse model compensation
摘要点击 1776  全文点击 801  投稿时间：2019-06-05  修订日期：2019-07-11
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2019.90426
  2020,37(4):705-712
中文关键词  Hammerstein-Wiener模型  逆模型  非线性系统  模型预测控制
英文关键词  Hammerstein-Wiener model (H-W)  Inverse model  nonlinear system  model predictive control  
基金项目  国家重点研发计划（2016YFF0101700-04）,国家自然科学基金（61773366,61533015）,
作者单位E-mail
孙浩杰 中科院沈阳自动化研究所 sunhaojie@sia.cn 
邹涛* 中科院沈阳自动化研究所 zoutao@sia.cn 
张鑫 中科院沈阳自动化研究所  
惠存万 中科院沈阳自动化研究所  
中文摘要
      针对一类Hammerstein-Wiener模型描述的非线性控制系统，提出一种基于逆模型补偿的预测控制策略。在控制优化计算中，利用Wiener非线性环节的逆模型分别对系统输出设定值和采样值进行变换；在控制实施过程中，将控制器输出操作量经过Hammerstein静态非线性环节模型逆变换后施加到实际被控对象上，通过两次逆变换，使得标称模型下控制器输出与闭环系统中线性环节的输入相一致。通过非线性变换补偿将非线性过程的控制转化为线性系统控制，避免了对非线性模型进行优化计算量大及预测不准确的问题。最后通过仿真验证了所提方案的可行性及有效性。
英文摘要
      This work focuses on the nonlinearity control system which is descripted by Hammerstein-Wiener model, and proposes a model predictive control strategy bases on compensation of inverse model. During the calculation of optimal control, an anti-model of Wiener nonlinearity unit is used to set the output setting values and the sample values, and in the control process, the controller output is applied to the actual controlled object after the inverse transformation by the static nonlinear Hammerstein link model. Through the above two inverse transformation, which could ensure the output of the controller consistent with the input of the linearity unit in the closed-loop system. Nonlinear transform compensation method is utilized to transform nonlinear process control into linear system control, which avoid large computation and inaccurate prediction in optimizing the nonlinear model directly. Finally, the feasibility and effectiveness of the proposed scheme are verified by simulation.