A two-stage intelligent condition monitoring method for electromagnets in high-voltage circuit breakers

Abstract

Intelligent condition monitoring of electromagnets in high-voltage circuit breakers is crucial for the safe operation of power systems. In this study, a hybrid two-phase approach is proposed, aiming at intelligently identifying key feature points in electromagnet current signals by a modified one-dimensional U-Net model, as well as determining the operating state of electromagnets based on the location and magnitude of these key feature points. In the first data-driven phase, an improved U-Net model applicable to time series is introduced for the intelligent identification of key features in the electromagnet current signal. In the second knowledge-based phase, the operating state of the electromagnet is accurately identified based on the position and size of the key feature points by exploring the motion mechanism of the electromagnet. The experimental results show that the proposed method can effectively identify the key points, and the recognition success rate is close to 100%. The proposed method can realize the adaptive identification of various electromagnet faults with only a few fault samples. Therefore, the proposed method paves the way for robust state identification of electromagnets and enjoys the merit of strong anti-interference.