引用本文:丁海旭,汤健,夏恒,乔俊飞.基于TS-FNN的城市固废焚烧过程MIMO被控对象建模[J].控制理论与应用,2022,39(8):1529~1540.[点击复制]
DING Hai-xu,TANG Jian,XIA Heng,QIAO Jun-fei.Modeling of MIMO controlled object in municipal solid waste incineration process based on TS-FNN[J].Control Theory and Technology,2022,39(8):1529~1540.[点击复制]
基于TS-FNN的城市固废焚烧过程MIMO被控对象建模
Modeling of MIMO controlled object in municipal solid waste incineration process based on TS-FNN
摘要点击 1242  全文点击 388  投稿时间：2021-06-18  修订日期：2022-04-02
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2022.10524
  2022,39(8):1529-1540
中文关键词  城市固废焚烧  被控对象模型  多输入多输出  模糊神经网络  过程控制
英文关键词  municipal solid wastes incineration  model of controlled object  multi-input multi-output  fuzzy neural networks  process control
基金项目  国家自然科学基金项目(62021003, 61890930, 62073006), 北京市自然科学基金项目(4212032, 4192009), 科学技术部国家重点研发计划项目 (2018YFC1900800–5), 矿冶过程自动控制技术国家(北京市)重点实验室项目(BGRIMM–KZSKL–2020–02)资助.
作者单位E-mail
丁海旭 北京工业大学 dinghaixu@emails.bjut.edu.cn 
汤健 北京工业大学  
夏恒 北京工业大学  
乔俊飞* 北京工业大学 junfeiq@bjut.edu.cn 
中文摘要
      针对城市固废焚烧(MSWI)过程中因机理反应复杂、不确定性严重等原因导致被控对象模型难以建立的问 题, 设计了一种基于(T-S)型模糊神经网络(FNN)的多输入多输出(MIMO)模型. 首先, 描述了MSWI过程的核心工艺 流程并分析了模型的影响因素; 接着, 设计了面向过程控制的被控对象建模策略, 其由工况识别模块、数据预处理 模块、特征约简模块、被控对象模型训练模块与被控对象模型测试模块组成; 最后, 通过实验表明了所构建模型的 有效性, 为研究MSWI过程的优化控制算法奠定了基础.
英文摘要
      Aiming at the problem that the model of controlled object is difficult to establish due to the complicated mechanism reaction and serious uncertainty in the municipal solid wastes incineration (MSWI) process, a multi-input multioutput (MIMO) model based on Takagi-Sugeno (T-S) fuzzy neural network (FNN) is designed. First, the core process flow of MSWI process is described and the influence factors of model are analyzed. Then, a modeling strategy for controlled object of process control is designed, which is composed of an operating condition recognition module, a data preprocessing module, a feature reduction module, a training module for the model of controlled object, and a testing module for the model of controlled object. Finally, the effectiveness of the constructed model is verified by experiments, which lays the foundation for studying the optimal control algorithm of MSWI process.