PMSM DTC系统逆变器故障的模糊神经网络诊断(英文)

Neural-fuzzy technique for inverter faults diagnosis in PMSM DTC system

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 frequently happened power electronics inverter faults in the well known permanent magnet synchronous motor （PMSM） direct torque control （DTC） system, an intelligent technique based on diagnosis method is proposed. An adaptive neural-fuzzy model was constructed considering the features of the inverter faults in the PMSM DTC system. The motor current was taken as the signature for the inverter faults diagnosis, and the adaptive neural-fuzzy network was used for the inverter fault diagnosis after the off-line training process. The elaborate simulation results show that only one phase current needs to be sensed for the diagnosis of different inverter faults, the complicated transformation is eliminated and the system cost is reduced. The proposed technique ensures the further fault tolerant implementation of the system.