基于RBF神经网络的无刷直流电机转矩波动抑制

摘要

在分析换相转矩波动形成原因的基础上,利用RBF神经网络的非线性控制特性优化PWM抑制法,提出了基于RBF神经网络的无刷直流电机转矩波动抑制新策略.该策略根据采样得到的电机转速和相电流,实时修正神经元之间的连接权值,并且调节换相过程中开关管的开通和关断时间,使换相电流的下降和上升速率近似相等,从而基本保持非换相电流恒定.仿真和实验表明:与传统的PWM抑制法相比,该控制策略有效抑制了换相转矩波动,具有良好的控制性能.%Having analyzing the causes of commutation torque ripple , together with the help of nonlinear control from RBF neural network, which could optimize pulse width modulation( PWM) method, a new torque ripple restrain method of brushless DC motor based on RBF neural network was presented. In this method, according to the speed of motor and phase current, connection weight of neural network was revised in time. Meanwhile, the opening and closing time of switch tube was adjusted in the commutation process, in order to equal the declining rate with the inclining one of commutation current and keep non-commutation current stable. Simulation and experimental results indicated that the proposed method was effective to reduce commutation torque ripple and had high control performance, compared with the conventional PWM method.