关键词: 小波包特征熵 , SVM , 分类器 , 压缩机气阀 , 故障诊断

Abstract: The feature extraction and the failure identification model construction was extremely difficult due to the nonstationarity and nonlinearity of the compressor valve vibration signal and the little failure sample．According to the above
characteristic，the information with good fault －tolerance ability was extracted as the feature parameter，which outlines the overall statistical characteristic of the signal．The extracted wavelet packet entropy was used as the input vector to construct the decision function and establish the valve wear failure SVM classifier based on the available failure sample．The experimental result demonstrates the effectiveness of the model in the nonstationarity signal feature extraction and nonlinearity pattern classification with small sample，and this model provides a practical valve wear failure diagnostic method.

Key words: Wavelet packet entropy , SVM , Classifier, Compressor valve, Fault diagnosis