| 引用本文: | 陈登乾,王昕,王振雷.多模型混合H∞型鲁棒控制[J].控制理论与应用,2018,35(8):1074~1082.[点击复制] |
| CHEN Deng-qian,WANG Xin,WANG Zhen-lei.Multi-models Mixing H∞ Theory Robust Control[J].Control Theory and Technology,2018,35(8):1074~1082.[点击复制] |
|
| 多模型混合H∞型鲁棒控制 |
| Multi-models Mixing H∞ Theory Robust Control |
| 摘要点击 2181 全文点击 1383 投稿时间：2016-09-27 修订日期：2018-01-10 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI编号 10.7641/CTA.2018.60718 |
| 2018,35(8):1074-1082 |
| 中文关键词 多模型 混合 H∞理论 鲁棒性 聚合釜 |
| 英文关键词 multi-models mixing H∞ theory robustness polymerization kettle |
| 基金项目 国家自然科学基金资助项目（ 21376077）；国家自然科学基金优秀青年基金资助项目（ 61422303）； 上海市“科技创新行动计划”研发平台建设项目（ 13DZ2295300）；上海市自然科学基金资助项目(14ZR1421800)； 流程工业综合自动化国家重点实验室开放课题基金资助项目(PAL-N201404). |
|
| 中文摘要 |
| 基于多模型控制器设计方法中需要建立大量离线子模型，导致计算负担过大的问题，本文提出基于 H∞理论的多模型混合鲁棒控制器设计方法。该方法首先对参数变化区间进行分解，根据得到的子区间建立子模型，从而获得多模型集。然后依据 H∞理论设计子模型鲁棒控制器，具有良好的抗干扰型，并能有效地减少子模型数量；其次，针对各鲁棒控制器，设计混合信号实现混合输出，可以很好地解决切换多模型控制中存在的持续切换问题。最后，对本文所提方法进行仿真和聚合釜温度系统的应用研究，结果表明当被控对象参数存在不确定性和存在输入扰动的时候，该方法可以在较少子模型数量的前提下保证系统具有良好的鲁棒性。 |
| 英文摘要 |
| To solve the problem that leads to establish a large number of off-line sub-model and increase computational burden
obviously because of the traditional methods of controller design, a new idea that means controllers design of multi-models
mixing control based on H∞ control theory is proposed. Firstly, the variable parametric interval is decomposed, then the model
set are established according to the sub intervals. Secondly, according to the H∞ theory the robust sub-controllers are designed,so this method own good anti-disturbance performance, and can effectively reduce the number of sub-models; Next the control output of the object is obtained by mixing the mixing signal and the output of robust sub-controllers, which can deal with the continuous switching problem of multi model switching control. Finally, numerical simulation and the polymerization kettle temperature system are researched on the basis of the proposed method. The results indicate that when the controlled object have the features of uncertain parameters and input disturbances, the method can guarantee excellent robustness under the condition of a less number of sub-models. |