Research on the Learning Method of Wiener Process Degradation Trajectory of Electromagnetic Relay Contact Arc Energy Based on Variational Autoencoder

The contacts of electromagnetic relays, as the key components responsible for current interruption, inevitably suffer from performance degradation due to electrical arc erosion during operation. Most existing research on relay contacts has focused on the performance degradation of contact resistance while neglecting the variations in arc energy during the breaking and closing processes. The magnitude of arc energy directly reflects the intensity of arc erosion on the contacts and serves as a critical parameter. Therefore, this paper models the drift Wiener process of arc energy sequences during contact breaking and closing using a variational autoencoder model based on contact arc energy. This approach captures the degradation process of contact arc energy and enables the lifetime prediction of electromagnetic relays.