【摘要】 Vienna整流器作为汽车直流充电桩充电模块最常用前级部分，前级系统的稳定运行直接影响着整个充电模块的运行状态，因此针对Vienna整流器故障诊断显得十分重要。针对Vienna整流器其功率开关和电解电容等核心器件的开路故障特点，文中提出了一种基于互补集合经验模态分解(Complementary Ensemble Empirical Mode Decomposition, CEEMD)和粒子群(Particle Swarm Optimization, PSO)优化随机森林(Random Forest, RF)算法的诊断方法。该方法以输入侧电流为原始信号，分析核心器件开路故障的波形特性，采用CEEMD方法对故障电流信号进行分解。在此基础上构造故障特征向量，并将提取的特征向量输入到粒子群优化的随机森林模型中进行故障状态识别。搭建了Vienna整流器仿真模型，验证所提方法的可行性和优越性。仿真结果表明该方法具有较好的诊断结果，诊断率达到了93.8%并且缩短了诊断时间，对汽车直流充电桩故障诊断具有现实指导意义。更多还原

【Abstract】 Vienna rectifier is the most common front part of the charging module of automobile DC charging pile, and the stable operation of the front system directly affects the running state of the whole charging module, so the fault diagnosis of Vienna rectifier is very important.In view of the open-circuit fault characteristics of Vienna rectifier′s power switch and electrolytic capacitor, in this paper, a Complementary ensemble empirical mode decomposition(CEEMD) and Particle swarm optimization are proposed.PSO optimizes the diagnostic method of Random forest(RF) algorithm.This method takes the input current as the original signal, analyzes the waveform characteristics of the open-circuit fault of the core device, and uses CEEMD method to decompose the fault current signal.On this basis, the fault feature vector is constructed, and the extracted feature vector is input into the particle swarm optimization random forest model to identify the fault state.The Vienna rectifier simulation model is built to verify the feasibility and superiority of the proposed method.The simulation results show that the method has a good diagnosis result, the diagnosis rate reaches 93.8% and the diagnosis time is shortened.It has practical significance for the fault diagnosis of automobile DC charging pile.更多还原