论文部分内容阅读
针对低速运行时电动机存在转矩脉动大的实际问题,在传统电动机直接转矩控制的基础上,采用改进RBF神经网络控制器替代PID控制器,利用GA找出RBF网络的最优参数值,得到最优的网络结构,用该网络结构进行电动机直接转矩控制。实验结果表明,该方法克服了网络参数选择的随机性,具有更强的适应能力;同时能降低电动机的转矩脉动,改善磁链的波形。
Aiming at the practical problem of large torque ripple in low speed motor, based on the traditional direct torque control, an improved RBF neural network controller is used instead of the PID controller, and the optimal parameter value of RBF network is obtained by using GA. Optimal network structure, using the network structure of the motor direct torque control. Experimental results show that this method overcomes the randomness of network parameter selection and has more adaptability. At the same time, it can reduce the torque ripple and improve the flux linkage waveform.