基于PCA和SVM的柴油发动机冲击故障诊断方法研究

Research on Fault Diagnosis of Diesel Engine Impact Fault Based on PCA and SVM

针对柴油发动机振动信号进行故障诊断技术研究,提出了一种基于主成分分析和支持向量机的柴油发动机冲击故障诊断方法。首先利用小波包分解提取出冲击故障的特征;再利用主成分分析技术获得敏感特征参数,进而减小数据处理的复杂程度;最后利用支持向量机对敏感特征参数样本进行训练,获得分类模型,进而实现故障分类。将该方法用于柴油机实际故障分类,诊断准确率较高,结果证实了该方法对多种冲击故障诊断具有的有效性。

For the research on fault diagnosis technology of diesel engine vibration signal a fault diagnosis method of diesel engine based on PCA and SVM is proposed. First of all, the features of impulsion faults are extracted by wavelet packet decomposition. Then PCA is used to obtain the sensitive characteristics, which reduces the complexity of data processing. Finally, SVM can be used for training the sensitive feature subset to get the classification model, and then realize the fault classification. The method is applied to the actual faults of diesel engine, and it turns out to have high diagnostic accuracy, which confirms the validity of this method for multiple impulsion fault diagnosis.