开关磁阻电机RBF神经网络电流控制  

Current Control Method of SRM Based on RBF Neural Network

作　　者：孙鹤旭[1] 李鹏[1] 董砚[1] 

机构地区：[1]河北工业大学电气与自动化学院,天津300130

出　　处：《电气传动》2010年第1期59-62,共4页Electric Drive

基　　金：国家863计划重点项目资助(2006AA040306);河北省自然科学基金项目资助(E2007000084)

摘　　要：开关磁阻电机(SRM)具有结构简单、成本低、控制灵活等优点,尤其组成的调速系统具有交、直流调速系统所没有的优点。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,因此如何控制好转矩成为关键,而转矩控制最终要通过控制电流来实现。对8/6结构SRM的绕组磁场特性及电感进行分析,构建了基于3层结构的径向基函数(RBF)神经网络的SRM电感模型,该模型算法简单并能较好地反映SRM电感非线性模型;依据该模型提出了一种自调节的电流控制方法,该方法通过已建立的SRM电感模型动态调节PWM的占空比,克服电感对电流的影响。实验结果证明,该方法使实际电流很好地跟随给定电流,有效减小了电流波动,取得了良好的电流控制效果。Switched reluctance motor （SRM） take on the merits such as simple structure, low-cost, flexible control, etc. switched reluctance drives systems （SRD） has the advantages that conventional AC or DC drivers system may not have. However, owing to its nonlinear electromagnetism, the torque ripple of motor is much more severe than other driver system, so it is pivotal to control the torque. The torque of switched reluctance motors （SRM） can be controlled by controlling the stator current. The winding magnetic properties and inductance of 8/6 SRMs were analyzed and built inductance model based on three-layer structure radial basis function （RBF） neural network. This algorithm is simple and can better reflect the nonlinear model of SRM inductance. A self-regulation of the current control method can be get based on this model. According to the tendency of current acquired by the RBF, the duty of PWM should be adjusted so as to obtain the optimized current control. The results of experiment prove that this method could make actual current follow given current perfectly.

关 键 词：开关磁阻电机 径向基函数 神经网络 两相励磁 

分 类 号：TM3[电气工程—电机]