基于模糊神经解耦控制的板型板厚控制系统仿真

Simulation of Strips Gauge and Flatness Control System Based on Fuzzy Neural Network Decoupling Control

面对板型板厚控制这一复杂、多变量耦合的非线性系统,本文提出了一种两级串联结构的模糊神经网络解耦控制策略,前级为自调整模糊控制器,后级为基于动态耦合特性的自适应神经网络解耦控制器,并从理论上证明了学习算法的收敛性。实现了无模型板型板厚综合控制。仿真结果表明,该控制系统收敛性好、抗干扰性强,取得令人满意的板型板厚控制精度。

Referring to plate flatness and gauge control system with multivariable nonlinear and strong coupling aswell as uncertain parameters, the paper proposes a two stage cascade fuzzy neural network decoupling control strategy, the former is a self - tuning fuzzy controller by using the intelligent weight function rulers, the latter is a self - adaptive neural network decoupling controller based on the learning algorithm of dynamic coupling .characteristic. It theoretically proves the convergence 0f learning algorithm. _Automatic flatness control （AFC） an,1 amomatic gauge control （AGC） are achieved. The simulation result shows that this control system has good convergence and disturbance resistance and gains a satisfied control accuracy of the flatness and the gauge.