论文部分内容阅读
针对永磁同步电机(PMSM)直接转矩控制(DTC)系统中常见的电力电子逆变器故障,提出了一种基于智能策略的诊断方法。根据PMSM DTC系统逆变器故障的特性,建立了一个自适应神经模糊网络的模型,选择PMSM的电流量作为故障检测和诊断的信号源,对系统正常及故障状态下的电流特性进行了分析,并利用训练好的模糊神经网络进行逆变器的故障诊断。仿真结果表明,该方法仅需检测电机的一相电流便可直接实现逆变器多种常见故障的诊断,摒弃了复杂的信号变换,同时节约了系统成本,保障了故障系统容错策略的实施。
Aiming at the common faults of power electronic inverter in PMSM direct torque control (DTC) system, a diagnostic method based on intelligent strategy is proposed. According to the characteristics of inverter fault in PMSM DTC system, a model of adaptive neural network is established. The current of PMSM is selected as the signal source of fault detection and diagnosis, and the current characteristics of the system under normal and fault conditions are analyzed. And using the trained fuzzy neural network for inverter fault diagnosis. The simulation results show that this method can directly diagnose many common faults of the inverter by detecting only one phase current of the motor, eliminate complex signal transformation, save system cost and guarantee fault tolerance implementation.