基于RBF神经网络的解耦方法在蒸汽管道压力控制系统中的应用

摘要

The coupling relationship between the variables of the steam pipe in the running process of the system is ana-lyzed in this paper.Aiming at the characteristics of amplitude changes,time-varying,serious coupling of the steam pressure pipe,a design scheme of decoupling PlD control is proposed for multivariable system based on RBF neural network,which used to control the two decoupling approximately independent single input-single output control systems.%分析了系统运行过程中蒸汽管道上各变量间的耦合关系；针对蒸汽管道压力变化幅值大、时变性、耦合严重的特点，提出了基于RBF神经网络的解耦控制方案。并采用基于RBF神经网络的PlD对解耦后的两个近似独立的单输入单输出系统进行控制。结果表明：基于RBF神经网络的PlD控制不仅超调量小、响应速度快，且具有很强的鲁棒性和自适应功能，取得了良好的控制效果。