Sliding time synchronous averaging based on independent extended autocorrelation function for feature extraction of bearing fault

Extracting fault feature is a key challenge during the early failure stage of bearings due to weak fault components and strong background noises. To solve this problem, an extension operation is introduced, and a novel algorithm called independent extended autocorrelation function is proposed. Subsequently, based on time synchronous averaging, the sliding time synchronous averaging method is developed to further enhance periodic components. The time dimension is added to generate a series of periodic impulses, which is advantageous to directly extract the fault characteristic. Therefore, the presented algorithm is called the sliding time synchronous averaging based on independent autocorrelation function (STSA-IEACF). Numerically simulated signals and two experimental datasets are employed to compare the performance of the STSA-IEACF with the mainstream blind deconvolution methods in fault feature extraction of bearing. The results show that the proposed algorithm performs better than the others in this terms of extracting fault characteristics.