A rolling bearing fault diagnosis approach based on LCD and fuzzy entropy

When the fault happens, the complexity of the intrinsic oscillation from the mechanical system will change. Fuzzy entropy (FuzzyEn), which is defined to measure the complexity and self-similarity of the time series, can be utilized to measure the complexity of vibration signal and reflect changes of complexity of the intrinsic oscillation. Since the changes distribute in different scales, a new non-stationary signal analysis method, local characteristic-scale decomposition (LCD), is proposed and used to decompose the vibration signal adaptively into series of intrinsic scale components (ISC) in different scales. And then a new rolling bearing fault diagnosis approach based on LCD and FuzzyEn is proposed. Namely, firstly, by using the LCD rolling bearing vibration signal is decomposed into numbers of ISCs; then the FuzzyEns of the first few ISCs that contain main failure information are extracted. Thirdly, the FuzzyEns obtained are taken as the inputs to the adaptive neuro-fuzzy inference systems (ANFIS) classifier. Finally, the proposed method is applied to experimental data, and the analysis results show that the proposed method performs effectively for the rolling bearing fault diagnosis.