基于RBF神经网络的开关磁阻电机转矩脉动控制

摘要

开关磁阻电机的双凸极结构和高度的非线性电磁特性会引起严重的转矩脉动.针对这一问题,提出了一种基于径向基函数神经网络的瞬时转矩控制方法.利用径向基函数神经网络较强的泛化和逼近能力,结合开关磁阻电机样本数据和控制要求,设计转矩观测器,实现电流、角度到转矩的非线性映射.然后将转矩作为转矩内环的反馈,直接控制瞬时转矩跟踪转速外环输出的参考转矩,再结合转矩滞环控制器完成电机的转矩控制.仿真和实验结果表明,所提控制策略具有响应速度快、控制精度高、可适应转速变化等优点,能有效地减小开关磁阻电机的转矩脉动.%Torque ripple is a primary disadvantage of switched reluctance motor(SRM)owing to its doubly salient struc-ture and obvious nonlinear electromagnetic characteristics.To solve this problem,a scheme of instantaneous torque con-trol method based on radial basis function(RBF)neural network is presented.With the combination of requirement for sample data and control,a torque observer is designed based on the advantages of RBF neural network including its gen-eralization and approximation ability,which realized a nonlinear mapping of current and angle to torque.Then,torque inner loop is constituted by a torque hysteresis-controller,and torque is used as feedback directly. The proposed ap-proach accomplished the torque control of the motor by constructing transient torque to track the reference torque, which is the output of speed outer loop.Simulation and experimental results demonstrated that the proposed control strat-egy can effectively reduce the torque ripple of SRM,and have the advantages of fast response,high control accuracy and adaptability to speed variation.