DataSheet1_Proposed multi-signal input fault diagnosis of wind turbine bearing based on glow model.docx

In the fault diagnosis, the problem of insufficient fault samples and unbalanced number of fault category samples can occur. In this paper, we used Glow model to supplement the number of wind turbine bearing fault samples to enhance the diagnosis accuracy when fault samples are insufficient and the number of fault category samples is unbalanced. Meanwhile, we established a multi-input fault diagnosis model to achieve multi-location and multi-category fault diagnosis, and constructed some experimental under different scenarios. We took into account the noise interference in the actual operation and conducted comparison experiments under different scenarios, and the experimental results verified the new algorithm had good fault diagnosis effect.