一种基于时频控制方法的控制器设计与实验
On the design and physical validation of a time-frequency controller
摘要点击 1444  全文点击 1879  投稿时间:2016-05-19  修订日期:2017-03-13
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2017.60334
  2017,34(3):329-336
中文关键词  时频控制  小波分析  自动寻优  电液系统  硬件在环
英文关键词  time-frequency approach  wavelet analysis  automatic optimization  electro-hydraulic hybrid system  hardware-in-the-loop
基金项目  北京市科技创新基地培育与发展基金专项(Z151100001615017)资助.
作者单位E-mail
武秀恒 中国农业大学工学院 wxh599@126.com 
杜岳峰 中国农业大学工学院  
C. Steve Suh 德克萨斯ADdDdM大学机械工程系  
毛恩荣 中国农业大学工学院  
宋正河 中国农业大学工学院 songzhenghe@ cau.edu.cn 
中文摘要
      为提升被控系统在整个工作频域内的动态性能, 本文提出一种时频(time-frequency, TF)控制方法的理论框架, 该框架通过应用小波分析技术以提高传统控制器性能的方式达到上述目标. 在阐述了此框架下TF控制机理的基础上, 结合传统PI控制方法设计了一种时频PI(time-frequency PI, TFPI)控制器, 并对TFPI控制器中时频参数的设计原理进行了分析. 针对TFPI控制器中参数众多所导致的整定工作量增大的问题, 结合最优控制理论中积分型性能指标的概念, 建立了以单纯形优化方法为手段的参数自动寻优策略. 最后以电磁阀控制液压缸系统模型为例进行了控制器硬件在环仿真测试. 阶跃响应结果显示, 在超调量不超过1%的要求下, 相对传统PI控制器0.15 s左右的响应时间, TFPI控制器响应时间可达0.06 s, 在不失稳定性的情况下显著提升系统的动态性能; 此外, 与文献中其他两种TF控制算法亦进行了仿真比较, 结果进一步证明了该控制方法在系统动态性能提升方面的有效性.
英文摘要
      A general framework feasible for time-frequency (TF) controller design is presented in order to improve the dynamic characteristics in all working frequency domain of systems. And the effort is the application of wavelet analysis to enhance the performance of the traditional controller to reach the goal. Based on the principle, a new controller incorporates traditional PI controller into TF control to realize a novel concept properly termed as the time-frequency PI (TFPI) approach. Although optimal control performance can be achieved by tuning PI gains, however, the large number of parameters to tune in TFPI can be overwhelming. To address the particular issue, the ISTE criteria and the simplex optimization method is followed to optimize the gains. The TFPI concept is successfully applied to an electro-hydraulic hybrid system using a hardware-in-the-loop test rig to demonstrate feasibility. The performance of the TFPI controller is favorably evaluated against the conventional PI controller. It is found that TFPI controller can decrease the step response time of the system from 0.15 s that is the best result of the traditional PI controller to 0.06 s on the premise of the overshoot less than 1%. Besides, the comparison between this method and the other two control algorithm in literatures is conducted and the results prove further that the TFPI controller can improve the dynamic performance markedly and maintain the stability of the system at the same time.