基于模糊神经网络开关磁阻电动机转矩观测

转矩脉动是开关磁阻电动机较为突出的缺点,出于电机设计和控制的目的,常需要实时在线测量电机的动态转矩,采用转矩
传感器的方法既复杂、昂贵,在高速下又不准确。利用一种新型变结构模糊神经网络对开关磁阻电动机转矩特性进行学习,之后把它用于动态转矩的实时在线测量。仿真与试验结果表明此方法能够快速、准确的测量开关磁阻电机动态转矩。
关键词　开关磁阻电动机　转矩脉动　测量　模糊神经网络
Abstract 　Switched reluctance motor (SRM) produces excessive vibration and torque ripple in low speed area1So the realtime and online measurement of SRM’s torque is required for cont rol and design purpose1However ,the measure2 ment method which using torque t ransducer is complex ,expensive and inaccurate1This paper uses a novel Fuzzy2neu2
ral networks(FNN) to approximate the nonlinear torque characteristics of SRMs as a function of position and current , and then the FNN is applied to online measurement of SRM’s torque1The simulation and experiment result s illust rate the validity of the proposed method1
Key words 　Switched reluctance motor 　Torque ripple 　Measurement 　Fuzzy2neural network