基于EWT-WOA-SVM的三电平逆变器早期故障诊断

Incipient Fault Diagnosis of Three-level Inverter Based on EWT-WOA-SVM

针对NPC型三电平逆变器中绝缘栅双极型集体管(IGBT)早期故障信号特征不明显,导致故障特征提取难度大,准确率低等问题,提出了一种基于经验小波变换(EWT)与支持向量机(SVM)相结合的故障诊断方法.该方法对逆变器早期故障进行分析,提取不同故障情况下的三相线电压作为原始特征信号,利用EWT对信号进行故障特征提取;针对EWT方法中需要人为设置分割层数的问题,采用自适应频谱分割方法根据采集到的原始信号自适应确定分割层数;利用鲸鱼优化算法(WOA)优化SVM的诊断方法对故障进行分类.通过仿真实验,并与其他方法进行对比,进一步验证了所提方法的有效性和准确性.

In order to solve the problem that the incipient fault signal characteristics of insulated gate bipolar collective tube(IGBT) in the NPC three-level inverter are not obvious, which leads to the difficulty of fault feature extraction and low accuracy, a fault diagnosis method based on the combination of empirical wavelet transform(EWT) and support vector machine(SVM) is proposed. Firstly, the early faults of the inverter are analyzed, and the three-phase line voltage under the different faults is extracted as the original characteristic signals. Meanwhile, the fault features are extracted by EWT. Furthermore, in order to solve the problem that the number of segmentation layers needs to be set manually in the method EWT,an adaptive spectrum segmentation method is adopted to determine the number of segmentation layers adaptively according to the collected original signals. The fault classification is carried out by utilizing the SVM diagnosis method optimized by whale optimization algorithm(WOA). Finally, the simulations are conducted and comparison is done with other methods. The effectiveness and accuracy in the proposed method are verified.