Bearing Fault Diagnosis Based on Mel Frequency Cepstrum Coefficient and Deformable Space-Frequency Attention Network

The main bearing is the core component of gas-fired generator, and its reliability directly affects stability whole system. Therefore, it great significance to study fault diagnosis generator. In based on vibration signal, how extract signature features effectively key achieving accurate diagnosis. Based extracting faults, classify efficiently another this, we propose a method Mel frequency cepstrum coefficient (MFCC) deformable space-frequency attention network (DSFAN). view inconsistent feature distribution different types MFCC algorithm introduced preprocess original signals their features. Then, model DSFAN constructed mechanism (SFFAM). can global constraint distributed realize To make full use classification information, data processed by into three-dimensional cube as input DSFAN. Finally, validity proposed MFCC-DSFAN verified CWRU, XJTU, generator sets. experimental results show excellent performance for prove effectiveness module in extraction.