Abstract:In order to realize the fault diagnosis of bearing fast and accurately, a fast fault diagnosis method for bearing was constructed based on cross correlation and mutual information. In this method, each single fault vibration signal (including inner ring, outer ring, ball and cage) was decomposed by Finite Impulse Response (FIR) which reduce aliasing caused by interference in the process of signal decomposition. Taking the fault vibration model established by mechanics analysis as the reference and calculated the correlation coefficient, the subsignal that characterizes the fault feature was selected. Calculated the mutual information of subsignal and constructed the fault feature matrix. Finally, the recognition results of K-Nearest Neighbor (KNN) is used to validate that the algorithm has the advantage in rapid identification of bearing failure.