Federated Multi-Model Transfer Learning-Based Fault Diagnosis with Peer-to-Peer Network for Wind Turbine Cluster

Intelligent fault diagnosis for a single wind turbine is hindered by the lack of sufficient useful data, while multi-turbines have various faults, resulting in complex distributions. Collaborative intelligence can better solve these problems. Therefore, peer-to-peer network constructed with one node corresponding to cluster. Each equivalent and functional replicable new federated transfer learning method, including model based on multi-task fusion dynamic adaptive weight adjustment. Models convolutional neural networks are trained locally transmitted among nodes. A solution processes data management, information transmission, provided. Experiments conducted signal testing bed bearing dataset Case Western Reserve University. The results show excellent performance method gearbox