基于梯度提升决策树的电力电子电路故障诊断

针对电力电子电路的常见故障类型,提出一种利用主元分析(PCA)提取电路状态的故障信息特征和基于梯度提升决策树(GBDT)分类的电力电子电路故障诊断方法.首先讨论利用PCA进行特征提取的步骤以及GBDT的分类原理;然后研究了基于PCA特征提取以及GBDT分类的电力电子电路故障诊断流程;最后利用三相桥式整流电路进行了建模、仿真、验证,实验结果表明,采用该方法进行电力电子电路故障诊断相比其他方法在低维空间具有更高的诊断准确率和更佳的样本泛化能力.

Fault Diagnosis of Power Electronics Circuit Based on Gradient Boosting Decision Tree

In order to eliminite the common faults of power electronics circuit,a fault diagnosis method of power electronic circuit is proposed,which is based on the principal component analysis(PCA) for extracting the faulty circuit information feature and the gradient boosting decision tree(GBDT) for classifying.Firstly,the steps of feature extraction using PCA and the classification principle of GBDT are discussed.Then the fault diagnosis process based on PCA and GBDT is studied.Finally,a three-phase full-bridge rectifier is used to model,simulate and verify the proposed method.The results show that this method can achieve a higher diagnosis accuracy and a better generalization capability than other methods.