Affiliation of Author(s):自动化学院
Journal:IEEE Int. Conf. Progn. Heal. Manag., ICPHM
Abstract:Dc arc fault has no zero-crossing point and is difficult to identify, which poses a great challenge to the security and stability of the dc power supply system. Detrended fluctuation analysis (DFA) is suitable for analyzing non-stationary signal. This paper, based on the experimental data of current AC component from automatic fault diagnosis platform, utilizing the (DFA) algorithm to extract arc fault characteristic, the least squares support vector machine (LS-SVM) is as the classifier to identify the arc fault, the experiment show that the proposed method is with satisfactory fault classification performance which is better than that based on time-frequency analysis. © 2018 IEEE.
Translation or Not:no
Date of Publication:2018-08-27
Co-author:Yin, Zhendong,Zhang, Yaojia
Correspondence Author:王莉,Yin, Zhendong,Wang Li