提出了一种基于神经网络故障诊断新方法。研究了基于波形直接分析和BP神经网络 的电力电子整流装置故障诊断方法。以三相桥式可控整流电路晶闸管断路故障为例,通过对一个电力电子电路进行仿真实验研究结果表明:该方法简单、直观,可以提高神经网络模型的学习效率和诊断的准确性,具有很好的故障识别能力,具有广阔的应用前景。 关键词:神经网络,电力电子,故障诊断 Abstract: A new method of fault diagnose distinguishing based on neural network is raised. Fault diagnose method of power electronic has been studied based on the wavy curve and neural network. Take the three-phase bridge-type controllable thyristor circuit disconnection fault as the example, The results of power electronic circuit experiment show that this method is simple, is direct-viewing, may enhance the neural network model the study efficiency and the diagnosis accuracy. The way has very good fault distinguishing ability and vast prospect. Keywords: Neural Network, Power Electronic, Fault Diagnose