Article Directory
1. Summary section
In actual industrial application scenarios, mechanical equipment is in a healthy state most of the time. Therefore, the number of healthy samples acquired by the monitoring system is often much more than the number of fault samples, which leads to the fact that in many diagnostic tasks, the number of fault samples is much less than the samples in a healthy state, that is, the design of the diagnostic model is faced with Data imbalance problem. Generally speaking, through artificial intervention design experiments, a sufficient amount of fault data can be obtained, and then a balanced data set can be constructed. However, limited by the cost of actual monitoring applications, it is difficult to obtain sufficient and balanced data sets for most monitoring tasks. Therefore, this study proposes a method based on data augmentation to solve