Fault diagnosis of high-voltage circuit breaker based on open-set theory fusion model

Abstract

Fault diagnosis of high voltage circuit breaker is an important aspect of electrical equipment intelligence. To effectively identify unknown faults, this paper proposes a high-voltage circuit breaker fault diagnosis method based on open set fusion model (OSFM). Firstly, the current data and vibration data are processed using sequential variational mode decomposition and Fourier transform, respectively, to extract data features, thereby constructing the original feature set of the current-vibration signal, which is then input into the Transformer model for further feature extraction. Secondly, the open-set discriminant model based on the extreme value theory is proposed, and the data output by transformer is input into classifier to realise open-set fault diagnosis. Finally, the tree-structured parzen estimator is used to optimise the selection of transformer model parameters and discriminator acceptance probability. The efficacy of the OSFM was evaluated through experimentation on experimental platform. The results demonstrated that the OSFM method can effectively recognise previously unidentified class faults while maintaining accurate recognition of known classes. Compared with other open-set classification techniques, OSFM can improve the recognition accuracy by up to 38.36%.