Quantitative and Localization Fault Diagnosis Method of Rolling Bearing Based on Quantitative Mapping Model

Abstract

Aiming to solve the problem of accurate diagnosis of the size and location of rolling bearing faults, a novel quantitative and localization fault diagnosis method of the rolling bearing is proposed based on the quantitative mapping model (QMM). The fault size and location of the rolling bearing affect the impulse type and the modulation degree of the vibration signal, which subsequently changes the complexity and randomness of the time-domain distribution of the vibration signal. According to the relationship between the multiscale permutation entropy (MPE) of the vibration signal and rolling bearing fault size, an average MPE (A-MPE) index is proposed to establish linear and nonlinear QMMs through the regression function. The proper QMM is selected through the error rate of fault size prediction to achieve a quantitative fault diagnosis of the rolling bearing. Due to the mathematical characteristics of the QMM, the localization fault diagnosis is realized. The multiscale morphological filtering (MMF) method is also introduced to extract the time-domain geometric feature of the fault bearing vibration signal and to improve the QMM accuracy of the fault size prediction. The results show that the QMM has a great effect on the quantitative fault size prediction and localization diagnosis of the rolling bearing.

Keywords

• rolling bearing
• quantitative and localization fault diagnosis
• multiscale permutation entropy
• multiscale morphological filtering
• regression function