基于遗传神经网络的超声波电机转速控制系统  

Ultrasonic Motor Speed Control System Based on Genetic Neural Network

作　　者：陈欢 史敬灼[2] 

机构地区：[1]武汉船用电力推进装置研究所,湖北武汉430064 [2]河南科技大学,河南洛阳471003

出　　处：《电机与控制应用》2010年第6期38-41,共4页Electric machines & control application

摘　　要：介绍一种超声波电机(USM)自适应控制策略,系统采用双闭环控制,内环通过对弧极电压进行采样、判断,补偿USM谐振频率的漂移;外环利用动态递归神经网络控制器调节两相驱动交变电压的相位差,实现转速的自适应控制。DRNN的结构和初始参数通过混合递阶遗传算法离线训练获得。试验结果证明,该控制系统具有较高的控制精度和较好的稳定性。An ultrasonic motor adaptive control strategy was introduced. Two closed loops were constructed here. The inner loop compensate USM resonant fl＇equency drift by sampling, judging isolated voltage. Driving phasedifference was regulated in the other loop by dynamical recurrent neural networks （DRNN） controller whose parameters were adjusted by the use of hybrid hierarchical genetic algorithm. With the proposed method, high precision and good robustness were obtained by experiments.

关 键 词：超声波电机 自适应控制 遗传神经网络 

分 类 号：TM301.2[电气工程—电机]