神经网络模糊直接转矩控制在感应电机中的应用研究

Research of DTC for induction motor based on neural network and fuzzy logic

针对感应电机直接转矩控制中存在较大转矩脉动的特点,提出基于神经网络与模糊逻辑的感应电机直接转矩控制策略,用训练好的神经网络建立定子磁链观测器,将模糊控制算法引入逆变器状态开关选择器,通过对转矩误差、定子磁链误差和磁链位置角的模糊分级来实现逆变器开关的分级控制。仿真及实验结果表明,与常规DTC相比,该控制方法转矩、转速、磁链响应速度快、响应脉动小,具有良好的稳态跟踪性能和优异的动态响应。

In view of the questions of greater torque for IM′s direct torque control,the DTC strategy for IM based on neural network and fuzzy logic is presented. With the trained neural network, the stator flux observer is established,and the fuzzy control algorithm is applicated with state switch selector. The inverter switch is grade controlled through torque error,error of stator flux and flux position angle of fuzzy classfication. The simulation and experimental results show that the control torque, speed and flux response is fast, and response fluctuating is small compared with conventional DTC,it has a good steady-state tracking performance and excellent dynamic response, for the practical application.