Method for determining degradation point of rolling bearing based on maximum value of spectral kurtosis feature

The invention discloses a method for determining a degeneration point of a rolling bearing based on a maximum value of a spectral kurtosis characteristic, and the method comprises the steps: carrying out the sampling of a full-life-cycle vibration signal of a tested rolling bearing for T times, thereby forming a time sequence W = [S1, S2,..., ST], and enabling the sampling signal of each time unit to be S; spectral kurtosis calculation is carried out based on the sampling signal S of each time unit, M time domain feature values are extracted according to the spectral kurtosis of each time unit, a feature matrix XT * M is formed, M is the number of extracted features, and the number of M is at least five; smoothing the feature time sequence of each column of the feature matrix; 5% of feature data of each column of feature time sequence in the beginning and end stages are subjected to zero assignment processing; and selecting the moment t'corresponding to the highest occurrence frequency as the bearing degradat