A Resampling Strategy Based Adaptive Diagnosis Method for Industrial Machinery Faults

Currently, the fault diagnosis with balanced data and single-point characteristics has received mass concern, and the related research achievements are remarkable. However, because of the scarcity of fault signals and the correlation of faults characteristics, diagnosing mechanical faults ubiquitous in industrial systems remain a thorny task. Therefore, for finding a solution to the issue, a resampling scheme is developed, further, a diagnostic plan to industrial machinery faults is formulated in this paper. Firstly, a resampling plan is formed to get the balance diagnostic sample sets. Then, an improved cost-function is developed for enhancing feature learning richness of deep sparse autoencoder (DSAE). Next, a de-correlation condition is formed to weaken adverse effects of this correlation. Finally, necessary tests are carried out to show the excellent performance of the designed method.