Title: A series arc fault diagnosis method based on random forest model

Authors: Qianhong Hou; Yongxin Chou; Jicheng Liu; Haifeng Mao; Mingda Lou

Addresses: School of Mechanical Engineering, Yancheng Institute of Technology, Yancheng, China; School of Electrical and Automatic Engineering, Changshu Institute of Technology, Suzhou, China ' School of Electrical and Automatic Engineering, Changshu Institute of Technology, Suzhou, China ' School of Electrical and Automatic Engineering, Changshu Institute of Technology, Suzhou, China ' Suzhou Future Electrical Co., Ltd., Suzhou, China ' Suzhou Future Electrical Co., Ltd., Suzhou, China

Abstract: The current of series arc fault is too weak to be detected by the circuit breaker, which is one of the causes of electrical fire. Therefore, an intelligent diagnosis method of series arc fault based on random forest (RF) is proposed in this study. Firstly, the high-frequency current signals of six kinds of loads are collected as experimental data. Then, 13 features are extracted from time domain and frequency domain, and the feature is reduced to four dimensions by principal component analysis (PCA). Finally, a classifier for series arc fault diagnosis is designed using RF. The experimental data in this study are collected by the low-voltage AC series arc fault data acquisition device developed by ourselves. The identification accuracy of series arc fault is 99.95 ± 0.03%. Compared with the existing series arc fault diagnosis methods, it has higher recognition performance.

Keywords: arc fault; intelligent diagnosis; random forest; feature extraction; principal component analysis; PCA; high accuracy.

DOI: 10.1504/IJMIC.2024.135539

International Journal of Modelling, Identification and Control, 2024 Vol.44 No.1, pp.23 - 31

Received: 16 Aug 2022
Received in revised form: 26 Sep 2022
Accepted: 27 Sep 2022
Published online: 18 Dec 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article