一种RBF神经网络自适应PID控制器在超临界温度系统的应用研究  

Application of RBF-NN Self-Adaptive PID Controller for the Supercritical Main Steam

作　　者：李云娟[1] 方彦军[2] 

机构地区：[1]昆明学院自动控制与机械工程系,云南昆明650118 [2]武汉大学自动化系,湖北武汉430072

出　　处：《自动化技术与应用》2010年第11期1-3,10,共4页Techniques of Automation and Applications

摘　　要：将传统PID控制器参数优化和径向基函数(RBF)神经网络结合,提出了基于RBF神经网络的PID控制器。将该控制器用于某超临界电厂温度系统中,由RBF网络对温度控制系统进行在线辨识,建立其在线参考模型并为PID控制器提供信息,控制器通过在线的自学习不断进行适应性控制,从而实现参数在线自调整,优化误差性能指标。MATLAB仿真结果表明,控制器对超临界温度系统有较好的控制效果,不仅跟踪性能良好,而且抗干扰性较强,鲁棒性较好。Combined with traditional PID controller and RBF nerve network,a PID controller based on RBF never network is proposed.The controller carries on line identify for temperature control through RBF network.It buids up the reference in the line and PID control model for providing information.The controller carries adaptive control unceasingly through online study,thus realize the parameter＇ online adjustment and characteristic idex error＇ optimization.Matlab simulation results show that the controller has better control effects for super critical temperture system.It has strong anti-disturbness and good robustness.

关 键 词：径向基函数 PID控制 在线辨识 鲁棒性 

分 类 号：TP273.2[自动化与计算机技术—检测技术与自动化装置]