Abstract:

Aiming at the problem that the fault arc characteristics in low-voltage DC systems are weak and difficult to accurately identify under different operating conditions,a DC arc fault diagnosis method based on the improved butterfly optimization algorithm (IBOA) optimizing random forest (RF) is proposed.Firstly,the whale optimization algorithm (WOA) is employed to optimize and improve the parameters of the adaptive noise complete ensemble empirical mode decomposition (ICEEMDAN),and the current signal is decomposed to obtain multiple intrinsic mode functions (IMF).Secondly,the effective components are selected and the hierarchical weighted permutation entropy (HWPE) is extracted to construct the feature vector.Finally,the reverse learning mechanism and dynamic boundary adjustment strategy are introduced to improve the butterfly optimization algorithm,and the IBOA-RF diagnostic model is established for fault identification.The results show that the average recognition accuracy of this method reaches 97.92% under various typical working conditions.The research verifies the effectiveness of this method in the field of DC arc fault diagnosis.

Key words: DC fault arc, feature extraction, hierarchical weighted permutation entropy, improved butterfly optimization algorithm, fault diagnosis