基于Wiener模型和Wiener-Hammerstein模型的非线性动态过程完整建模方法  

A Complete Modeling Method for Nonlinear Dynamic Processes Based on Wiener Structured Neural Network and Wiener-Hammerstein Structured Neural Network

作　　者：王珠 刘志辉 肖既磊 WANG Zhu;LIU Zhi-hui;XIAO Ji-lei(College of Information Science and Engineering,China University of Petroleum(Beijing))

机构地区：[1]中国石油大学(北京)信息科学与工程学院

出　　处：《化工自动化及仪表》2023年第5期652-659,共8页Control and Instruments in Chemical Industry

基　　金：国家自然科学基金项目(61703434)。

摘　　要：基于Wiener模型和Wiener-Hammerstein模型提出Wiener结构神经网络(WienerNN)模型和Wiener-Hammerstein结构神经网络(WHNN)模型,推导了Wiener模型和Wiener-Hammerstein模型的泛化能力,并给出了模型神经网络化的全过程。将所推导出的WienerNN模型和WHNN模型用于单变量和多变量非线性动态过程建模,验证了模型的有效性。Having Wiener model and Wiener-Hammerstein model based to propose Wiener structured neural network(WienerNN)model and Wiener-Hammerstein structured neural network(WHNN)model was implemented,including the derivation of generalization ability of the Wiener model and the Wiener-Hammerstein model,as well as the presentation of the whole process of model neural network.Applying both WienerNN model and WHNN model derived to model univariate and multivariable nonlinear dynamic processes can verify both feasibility and effectiveness of the model proposed.

关 键 词：非线性动态系统 神经网络 WIENER模型 Wiener-Hammerstein模型 

分 类 号：TP271[自动化与计算机技术—检测技术与自动化装置]