Cross-scenario fault diagnosis of solar insecticidal lamp IoTs using multi-model feature fusion and dynamic transfer learning

This dataset was collected from Chuzhou, Anhui Province, China, which shares a similar fault label space with the source domain. However, significant distribution discrepancies exist between the two datasets due to device heterogeneity and varying environmental conditions in the deployment sites. Consequently, it is of practical significance to investigate transfer learning-based fault diagnosis across diverse devices and environments to ensure system reliability.