CMAC交流励磁水轮发电系统(CMAC+神经元的复合).pdfVIP

第20 卷第7 期 　控　制　与　决　策 2005 年7 月 Vol. 20 No .7 　Control and Decision　 　 July 2005 : 1001-0920(2005) 07-0778-04 CMAC 李　辉 ( a. , b. , 400044) : 在分析常规CMAC 结构的基础上, 针对一类非线性、参数时变和不 定的控制系统, 提出了 一种自适应 CM AC 神经网络的控制器. 该控制器以系统动态误差和给定信号量作为CMAC 的激励信号, 并与自适应线性神经 元网络相结合构成系统的复合控制. 为了验证其有效性, 将其应用到交流励磁水轮发电机系统的多变量非线性控制 中, 并与常规的PID 控制效果进行了比较. 仿真结果表明, 该控制器具有较强鲁棒性和自适应能力, 控制品质优良. : CMAC 神经网络; 自适应控制; 交流励磁; 水轮发电机 : T P27 : A Simulation Research of an Adaptive CMAC for Hydrogenerating System with AC Excitation LI H ui ( . , a T he Key Laboratory of High Voltage Engineering and Elect rical New Technology in M inistry of Education b. College of Electrical Engineering, Chongqing University , Chongqing 400044, China. E -mail: cqulh@ 16 . com) Abstract: Based on the convent ional structure of the cerebellar model art iculation controller ( CMAC) , an adapt ive - CM AC neural netw ork controller is present ed for a class of nonlinear uncertain systems w it h time varying parameters. Combining it w ith the adaptive linear neuron netw ork, the mult iplex control strategy takes the dynamic errors and given signals of the system as input signals to the CM AC neural netw ork. T he digit al sim ulation of multivariable and nonlinear cont rol for the hydrogenerat or syst em w it h AC excit ation is studied and the conventional . PID controller is used t o compare w it h it The simulation results show t hat the proposed adaptive CM AC is of good robustness, and adaptive ability and excellent control quality. Key words: CMAC neural netw ork; Adaptive control; AC excitation; Hydrogenerator 1 引　　言