High impedance fault detection method based on improved complete ensemble empirical mode decomposition for DC distribution network. (May 2019)

• Record Type: Journal Article
• Authors: Wang, Xiaowei; Song, Guobing; Gao, Jie; Wei, Xiangxiang; Wei, Yanfang; Mostafa, Kheshti; Hu, Zhiguo; Zhang, Zhigang
• Is Part Of: International journal of electrical power & energy systems. Volume 107(2019)
• Page Start: 538
• Page End: 556
• Publication Date: 2019-05
• Subjects: Fault detection -- Intrinsic mode function -- Transient zero mode current -- Characteristic frequency component; Electrical engineering -- Periodicals; Electric power systems -- Periodicals; Électrotechnique -- Périodiques; Réseaux électriques (Énergie) -- Périodiques; Electric power systems; Electrical engineering; Periodicals; 621.3
• DOI: 10.1016/j.ijepes.2018.12.021
• Languages: English
• ISSNs: 0142-0615

Abstract

Highlights: Feature extract method : Used CEEMDAN algorithm to extract the CFC of transient zero mode current. Detection criterion : Constructed the starting criterion and discrimination criterion respectively, and realized the HIF accurate detection. Solved the HIF detection difficulty in DC distribution network, and the needed data window of the criteria is only 2 ms, and criteria is accurate, the calculation speed is faster. Abstract: Aiming at DC distribution network, we proposed a novel high impedance fault detection method in the paper, it main procedures are as follows: Firstly, used the algorithm of complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) to extract the first intrinsic mode function (IMF) of the characteristic mode. Secondly, calculated the singular point of mutation and the cumulative slope by the acquisition of the first order difference operation, and then, achieved to distinguish the fault state and the normal state by the comparison between the slope and the starting threshold. Thirdly, identified the first IMF with Prony algorithm to obtain the parameters of characteristic frequency components (CFC) and direct current components (DC), and calculated the energy ratio between them, and then, distinguished small impedance fault (SIF), medium impedance fault (MIF), high impedance fault (HIF) and load switching (LS) by different values of energy ratio. A large number of experiments show that the proposed method is accurate and effective. Compared with other methods, this method shows its merits in feature extraction accuracy, detection accuracy and calculation speed.