基于RBF的无位置开关磁阻电机控制系统

摘要

与传统交流电机驱动系统相比,开关磁阻电机( SRM)具有结构简单,成本低,效率高等优点.但由于电机本身的电磁特性呈现复杂的非线性,采用传统的线性控制方法如PI,PID控制策略很难取得较好的控制效果.基于局部逼近神经网络-径向基函数(RBF)网络,建立了SRM磁特性模型,设计了一套以18.5 kW三相(12/8)SRM为控制对象,基于TMS320F2812型DSP控制器的无位置SRM系统.实验结果表明,RBF神经网络位置检测使SRM具有较好的动态特性和较高精度,方案简单可靠,且有效可行.%Compared with conventional motors,switched reluctance motor(SRM) has merits of simple structure,low-cost, flexible control,etc.But based on SRM high nonlinear electromagnetism characteristics,the traditional linear control methods adopts PI,PID control strategy which is difficult to obtain better control effect.Based on partial approaches neural network-the radial basis function (RBF) network, a switched reluctance motor magnetic property model is established and a set of 18.5 kW three-phase(12/8) SRM which is the object of control, based on DSP TMS320F2812 controller of position sensorless SRD system is designed.The experimental results show that the RBF neural network position detection makes SRD have good dynamic performance and higher accuracy and sheme is simple, reliable, effective and feasible.