Application of SPA-HRDE in Acoustic Signal Fault Diagnosis of Mechanical Equipment
Aiming at the problems of contact acquisition and low accuracy of existing fault diagnosis methods,a mechanical fault diagnosis method combining smooth prior analysis and hierarchical reverse dispersion entropy is proposed.First,the acoustic signal is decomposed into trending and de-trending items by SPA.Then,the HDE is used to extract the hierarchical entropy values of trend item and de trend item signals,and fault feature samples are constructed;Finally,the honey badger algorithm is used to search the key parameters of support vector machine,a fault recognition model with optimal parameters is established,and the fault features are input into the HBA-SVM classifier for fault recognition.The experimental evaluation based on centrifugal pump and rolling bearing proves the effectiveness of the proposed method.The test results show that the method achieves 100％and 97％identification accuracy respectively in the fault identification of centrifugal pump and rolling bearing.Compared with other fault diagnosis methods,the method can fully extract the fault information in the acoustic signal,achieve higher precision fault diagnosis,and has strong robustness.
万方数据
维普
