Fault Diagnosis Method for Distribution Terminals Based on Whole Process Integration and Federated Learning

With the development of distribution automation, the accuracy and efficiency of fault diagnosis for distribution terminals are of crucial importance. This paper proposes a fault diagnosis method for distribution terminals based on whole process integration and federated learning. Firstly, a whole process integration technology platform and a federated learning model are constructed. Subsequently, a data acquisition and processing system based on whole process integration is established. Then, a fault diagnosis model is constructed using federated learning, including model selection, training and optimization, and performance evaluation. Through actual case analysis and experimental verification, it is demonstrated that this method significantly improves the accuracy and efficiency of fault diagnosis compared with other methods, with an average accuracy reaching 97.23% and the time for fault location within 8.35 μs.