基于RBF神经网络的开关磁阻电机瞬时转矩控制

开关磁阻电机(SRM)因其结构简单、工作可靠、效率高、成本低等优点使之成为当前极具竞争力的一种调速电动机。但由于电机本身的非线性电磁特性,导致了其转矩脉动比其他传动系统严重。如何更好地对开关磁阻电机的转矩进行控制,抑制转矩脉动也成为了近年来研究的热点。针对这一问题,提出了一种基于基于径向基函数(radial basis function,RBF)神经网络的开关磁阻电机瞬时转矩控制方法。利用从SRM动态模型仿真中产生的数据来对RBF神经网络进行离线训练,使之学习不同转速和转矩下的优化电流波形,再将训练好的RBF网络用于电机的转矩控制中,完成不同转速下,转矩、位置到电流的非线性映射。最后通过瞬时电流跟踪控制使电机电流跟踪参考电流,完成电机的转矩控制。该控制方法充分利用了RBF神经网络逼近、泛化能力强,运算速度快的优点,且控制过程简单,网络无需在线训练。实验结果证明,该控制策略能有效减小开关磁阻电机的转矩脉动,具有控制精度高、能适应转速变化等优点。

Instantaneous Torque Control of Switched Reluctance Motors Based on RBF Neural Network

The inherent simplicity,ruggedness,great efficiency and low cost of a switched reluctance motor(SRM)make it a potential candidate for adjustable speed application.However,owing to its nonlinear electromagnetism characteristic,the torque ripple of motor is much more severe than other drive system.For that reason,the torque ripple minimization in SRM has obtained great attention.To solve this problem,this paper presents a scheme of instantaneous torque control for switched reluctance motors based on radial basis function(RBF)neural network.The off-line training of RBF network is based on the data acquired by dynamic modeling.After trained from the optimized current profiles in different speed and torque,RBF network could be able to achieve the nonlinear mapping between torque,position and current.The approach of transient current tracing control is implemented in this system to adjust the current in order to control the torque of the motor.This method which has taken advantages of RBF neural network in approximation,generalization and calculation is simple and needs no on-line training.The results of experiment prove that this method could reduce SRM's torque ripple effectively and is adaptive to speed variety.