Fig 1.
The flowchart of the proposed method.
Step 1: Load the raw signal and define the parameter ranges for L and n to optimize FMD. Based on reference [22], the range for L is [2, 50], and the range for n is [2, 7]. Step 2: Set the minimum envelope entropy as the fitness function for SABO-optimized FMD. Step 3: Use SABO to identify the optimal combination of parameters [L, n]. Step 4: Calculate the kurtosis values of the n IMFs and retain the IMF with the highest kurtosis value. Step 5: Analyze the retained IMFs using envelope spectrum analysis.
Table 1.
The basic parameters of the drive-end bearing.
Fig 2.
The experimental apparatus used to extract the experimental data and its schematics.
Table 2.
The fault characteristic frequencies (multiple of running speed in Hz).
Table 3.
The data description of 170.mat.
Fig 3.
The iterative process of the SABO.
Fig 4.
The envelope spectrum after processing by the proposed method of 170.mat.
Fig 5.
The time-domain waveform after processing by the proposed method of 170.mat.
Fig 6.
The envelope spectrum after processing by the comparison method 1 of 170.mat.
Fig 7.
The time-domain waveform after processing by the comparison method 1 of 170.mat.
Fig 8.
The envelope spectrum after processing by the comparison method 2 of 170.mat.
Fig 9.
The time-domain waveform after processing by the comparison method 2 of 170.mat.
Table 4.
The experimental results of 170.mat.
Table 5.
The data description of 210.mat.
Fig 10.
The iterative process of the SABO.
Fig 11.
The envelope spectrum after processing by the proposed method of 210.mat.
Fig 12.
The time-domain waveform after processing by the proposed method of 210.mat.
Fig 13.
The envelope spectrum after processing by the comparison method 1 of 210.mat.
Fig 14.
The time-domain waveform after processing by the comparison method 1 of 210.mat.
Fig 15.
The envelope spectrum after processing by the comparison method 2 of 210.mat.
Fig 16.
The time-domain waveform after processing by the comparison method 2 of 210.mat.
Table 6.
The experimental results of 210.mat.
Table 7.
The data description of 130.mat.
Fig 17.
The iterative process of the SABO.
Fig 18.
The envelope spectrum after processing by the proposed method of 130.mat.
Fig 19.
The time-domain waveform after processing by the proposed method of 130.mat.
Fig 20.
The envelope spectrum after processing by the comparison method 1 of 130.mat.
Fig 21.
The time-domain waveform after processing by the comparison method 1 of 130.mat.
Fig 22.
The envelope spectrum after processing by the comparison method 2 of 130.mat.
Fig 23.
The time-domain waveform after processing by the comparison method 2 of 130.mat.
Table 8.
The experimental results of 130.mat.
Table 9.
The data description of 236.mat.
Fig 24.
The iterative process of the SABO.
Fig 25.
The envelope spectrum after processing by the proposed method of 236.mat.
Fig 26.
The time-domain waveform after processing by the proposed method of 236.mat.
Fig 27.
The envelope spectrum after processing by the comparison method 1 of 236.mat.
Fig 28.
The time-domain waveform after processing by the comparison method 1 of 236.mat.
Fig 29.
The envelope spectrum after processing by the comparison method 2 of 236.mat.
Fig 30.
The time-domain waveform after processing by the comparison method 2 of 236.mat.
Table 10.
The experimental results of 236.mat.