电气机车整流电路的故障诊断仿真分析

摘要

利用电力电子器件和仿真功能软件对我国电气机车整流器常用的控制电路(单相串联3重联结电路)进行了故障分析诊断；在Matlab/Simulink环境下对单相串联3重联结电路进行了电路分析和仿真,并在其基础上,提出了基于BP神经网络故障分析的方法；基于BP神经网络,利用电路分析产生的故障数据,采用归一化(或标准化)进行数据预处理,并提取故障特征及构造故障样本集,利用Matlab对神经网络进行学习训练,然后对所提出的故障特征进行了分析和辨识,最后通过仿真实验证明了该方法的正确性和可行性.%The control circuit (single -phase re -join the circuit in series 3) used in the electric locomotives adopts power electronic devices and simulation software for fault analysis and diagnosis. The single - phase circuits connected in series 3 are analyzed and simulated in Matlab/Simulink, based on which a theory is given for BP neural network fault analysis methods. By BP neural network circuit analysis, the normalized (or standardized) method for data preprocessing is used for fault data generation, Then Matlab is used for neural network to learn and train, the fault character is distinguished from proposed features, and finally the feasibility and validity of the method are proved by simulation.