Through the method proposed in this paper, the feature extraction of gear uniform wear fault is carried out, and the fault gear is the intermediate shaft driving gear (G3). Firstly, the resampling signal is obtained by cot. The results are shown in Figure 1 and Figure 2.

The number of mode decomposition k = 5 and penalty factor α = 4 000 are selected for VMD processing. The IMF component and its corresponding kurtosis table are obtained. Imf2 is selected as the fault signal for fast spectral kurtosis processing and square envelope analysis. The results are shown in Figure 3 below.

In Fig. 3C, it can be observed that the speed order of the driving shaft OC, the speed order of the intermediate shaft oc2, the meshing order of the intermediate shaft driving wheel (G3) om, the double order ratio 2om and the triple order ratio 3om, and the meshing order of the input shaft driving wheel (G1) om1.

The position of each meshing order signal in the figure has not changed, but compared with the order spectrum under normal conditions, the harmonic amplitude value has changed, the amplitude of high-order harmonics has increased significantly, the gearbox has uniform wear, and it is known from the meshing order that the intermediate shaft driving gear (G3) has failed. The analysis results are consistent with the actual situation, which shows that the proposed method can extract the fault characteristics of gearbox gear under variable speed.