基于多信号源的神经模糊Hammerstein-Wiener模型及其应用研究.docVIP

 基于多信号源的神经模糊 Hammerstein-Wiener 模型及其应用研究# 贾立1，杨爱华1，邱铭森2* 5 10 15 （1. 上海大学机电工程与自动化学院自动化系，上海市电站自动化技术重点实验室， 上海 200072； 2. 新加坡国立大学工程学院，新加坡 119260） 摘要：面对复杂工业过程控制的需求，设计一种结合数据信息的特殊模型结构，在保证控制 系统有效性的前提下通过模型的结构来简化控制器的求解是亟待解决的问题。为此，本文提 出一种基于多信号源的神经模糊 Hammerstein-Wiener 模型。突破传统的迭代分离方法，通 过组合式多信号实现 Hammerstein-Wiener 模型中的非线性环节和线性环节的分离，同时设 计了神经模糊 Hammerstein-Wiener 模型的参数非迭代优化算法，将研究结果拓广到分段非 线性系统，不仅保证了模型的精度，还保证了模型参数的收敛性。该方法具有逼近较强非线 性过程的能力，改善了模型的适用范围。在此基础上设计了神经模糊 Hammerstein-Wiener 模型的控制系统，利用模型特殊的结构将非线性系统的控制问题简化为线性系统的控制问 题，如采用传统的 PID 控制器便能达到较好的控制效果。仿真结果验证了该方法的有效性。 关键词：Hammerstein-Wiener 模型；神经模糊系统；非线性系统；信号分离 中图分类号：TP273 20 Multi-signal Based Neuro-fuzzy Hammerstein-wiener Model and Its Application JIA Li1, YANG Aihua1, CHIU Minsen2 (1. Shanghai Key Laboratory of Power Station Automation Techonlogy, Department of 25 30 35 40 45 Automation, College of Mechatronics Engineering and Automation, Shanghai University, ShangHai 200072; 2. Faculty of Engineering, National University of Singapore, Singapore 119260) Abstract: For demands of the complex industrial process control, it is a important to design a special structure model combined with data information to simple the solution of controller at the premise of guaranteeing the effectiveness of systems. Thus a multi-signal based neuro-fuzzy Hammerstein-wiener model is proposed in this paper. Break the traditional iterative separation method, the nonlinear and linear part of the Hammerstein-wiener model comes true by the combination of multi-signals, and the parameter noniterative optimization algorithm of the neuro-fuzzy Hammerstein-wiener model is designed to expand the research results to the subsection nonlinear system, not only guaranteeing the precision of the model, but also the convergence. On this foundation, the control system of neuro-fuzzy, Hammerstein-wiener model is designed to simple the control problem of the nonlinear system to the problem of linear system by the special structure of the model, then traditional PID