灰色预估神经网络在时滞系统控制中的应用  

Research on Neural Network Controller with Grey-Model Prediction Applied in Controlling Systems with Uncertain Time Delay

作　　者：赵仕俊[1] 李逊[2] 左光远[2] 

机构地区：[1]北京科技大学信息工程学院,北京100083 [2]中国石油大学计算机与通讯工程学院,山东东营257061

出　　处：《微电子学与计算机》2008年第2期25-27,共3页Microelectronics & Computer

摘　　要：针对过程控制中由于系统时滞引起的建模和控制方面的困难,提出了利用RBF神经网络对系统时滞进行辨识,采用灰色预估器对系统输出进行预测,并以此替代系统未来的实际输出反馈给输入端,最后利用RBF神经网络控制器对系统进行控制,有效地提高了控制器对时滞变化的自适应性。通过在一个具有时滞的加热炉系统中的应用,说明此方法能够克服不定时滞的影响,使控制目标具有良好的动态和静态性能。In order to solve modeling and controlling difficulties caused by time delay in process control, this article proposed a scheme that use RBF neural networks to recognize system time delay as well as Grey-Model Predictor to forecast system outputs, then passes this result to inputs of predictor as feedbacks, which instead of actual output, finally control system by RBF neural network controller. It improves adaptability to time delay of controller. The application in a heating oven with time delay shows that it can overcome uncertain time delay affect, and has good dynamic and static performance on controlling aim.

关 键 词：径向基(RBF)神经网络 灰色预估理论 时滞 加热釜 

分 类 号：TP273.5[自动化与计算机技术—检测技术与自动化装置]