论文部分内容阅读
为了提高含有噪声和野值的转子振动故障样本诊断精度,提出了基于WCFSE-FSVM的故障诊断方法。充分融合小波相关特征尺度熵(WCFSE)特征提取方法和FSVM故障诊断方法的优点,建立WCFSE-FSVM故障诊断模型。基于转子实验台模拟4种典型故障,获得原始故障数据;并利用WCFSE方法提取这些故障数据的WCFSE值,选取故障信号高频段中的尺度1和尺度2上的小波相关特征尺度熵W1和W2构造出振动信号的故障向量作为故障样本,建立FSVM诊断模型。实例分析显示:WCFSE-FSVM方法的转子故障诊断精度最高,即故障类别诊断精度为94.49%,故障严重程度的诊断精度为95.58%,二者都优于其它故障诊断方法。验证了WCFSEFSVM方法的可行性和有效性。
In order to improve the diagnostic accuracy of rotor vibration fault samples with noise and outliers, a fault diagnosis method based on WCFSE-FSVM is proposed. Fully combines the advantages of WCFSE feature extraction method and FSVM fault diagnosis method, and establishes WCFSE-FSVM fault diagnosis model. Based on the rotor test bed, four kinds of typical faults were simulated to obtain the original fault data. The WCFSE values of these fault data were extracted by WCFSE method. The wavelet feature scales entropy W1 and W2 of scale 1 and scale 2 in the high frequency band of the fault signal were chosen The fault vector of the vibration signal is used as the fault sample to establish the FSVM diagnosis model. The case analysis shows that the rotor fault diagnosis accuracy of the WCFSE-FSVM method is the highest, that is, the fault diagnosis accuracy is 94.49% and the fault diagnosis accuracy is 95.58%, both of which are superior to other fault diagnosis methods. The feasibility and effectiveness of the WCFSEFSVM method are verified.