Research on early fault feature extraction technology of aviation bearing based on noise estimation ITD

Abstract

Abstract Early indications of faults in aircraft bearings are frequently accompanied by excessive noise. To enhance the accuracy of signal decomposition, this study presents the ensemble noise-reconstructed intrinsic time-scale decomposition (ENITD) technique. In addition, a highly sensitive mode component selection method is suggested to attain the goal of improving the precision of fault feature extraction. The findings demonstrate that the ENITD approach is successful in addressing the mode mixing issue and enhancing the precision of fault feature extraction. Unlike established decomposition methods, the estimated noise is applied for denoising instead of incorporating white noise. Furthermore, the estimated noise can introduce diverse frequency signals to their corresponding proper rotation component (PRCs), aiding in resolving the mode mixing problem. This paper examines the efficacy of the ENITD approach for extracting early fault features in aircraft bearings using both simulated and experimental signals.