Fault feature extraction method of aircraft engine rolling bearings based on comprehensive dynamic screening

In view of the problem that the fault signal of aero-engine rolling bearing is often affected by high frequency and wide background noise, and the difficulty to extract and characterize the features, a new fault feature extraction method for rolling bearing was proposed. This method firstly performed wavelet packet decomposition on the bearing vibration signal and calculated the kurtosis value, correlation coefficient value, and energy ratio of each node component. Then using variance as an indicator, the most suitable weight for different vibration signals was found and integrated into comprehensive dynamic screening indicators. The contribution of each component was calculated. The first i node components with contribution reaching the threshold for reconstruction was selected to obtain a denoised new signal. Finally, the weak fault features of bearings were extracted through envelope spectrum analysis. Through verification by simulation signals, the signal-to-noise ratio of the denoised signal increased by 8.82 dB compared with that before denoising. Simulated tests on the intermediate shaft bearing failure of a certain type of aircraft engine and the main shaft bearing failure of a certain type of turbofan aircraft engine were conducted, effectively validating the methods proposed. Results indicated that the method presented accurately extracted the fault feature frequencies and their harmonics of the rolling bearings in aircraft engines, thereby achieving fault diagnosis. Through theoretical analysis and test validation, it can be considered as one of the effective methods for complex signal processing and diagnosis of rolling bearings in aircraft engines.