Fault Feature Extraction and Degradation State Identification for Piezoelectric Ceramics Cracking in Ultrasonic Motor Based on Multi-Scale Morphological Gradient

Guoqing An,Hongru Li,Baiyan Chen
DOI: https://doi.org/10.20855/ijav.2019.24.41464
2019-12-31
The International Journal of Acoustics and Vibration
Abstract:Piezoelectric ceramics cracking is one of the main faults of the ultrasonic motor. According to the morphological mathematics and information entropy, a method based on multi-scale morphological gradient was proposed for ceramics fault feature extraction and degradation state identification. To solve the problem that traditional multi-scale morphology spectral (MMS) entropy cannot exactly describe the performance degradation of the piezoelectric ceramics, multi-scale morphology gradient difference (MMGD) entropy was proposed to improve the sensitivity to the fault. Furthermore, multi-scale morphology gradient singular (MMGS) entropy was presented to reduce the system noise interference to the useful fault information. The disturbance analysis of temperature, load, and noise for MMGD entropy and MMGS entropy was also given in this paper. Combining the advantages of the above two entropies, a standard degradation mode matrix was built to distinguish the degradation state via the grey correlation analysis. The analysis of actual test samples demonstrated that this method is feasible and effective to extract the fault feature and indicate the degradation of piezoelectric cracking in ultrasonic motor.
What problem does this paper attempt to address?