Early bearing fault diagnosis based on the improved singular value decomposition method

The traditional singular value decomposition (SVD) method is unable to diagnose the weak fault feature of bearings effectively, which means, it difficult retain effective components (SCs). Therefore, a new method, SVD based on FIC (fault information content), proposed, takes amplitude characteristics frequency as selection index components. Firstly, Hankel matrix original signal constructed, and applied in matrix. Secondly, proposed used evaluate decomposed SCs. Finally, SCs with are selected added obtain denoised signal. results bearing simulation signals experimental show that compared differential (DS-SVD), can select larger amount able under heavy noise interference. be for denoising extraction.