A Novel Fault Diagnosis Scheme Based on Local Fault Currents for DC Microgrids

• Published in: IEEE transactions on power delivery Vol. 40; no. 1; pp. 570 - 583
• Main Authors: Li, Weiwei, Han, Hua, Sun, Yao, Chen, Shimiao, Liu, Hongyi, Zheng, Xinlong, Liu, Yonglu, Zhao, Jin
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.02.2025; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Capacitors; Circuit faults; Correlation; Correlation coefficients; DC microgrids; Direct current; Discharges (electric); Distributed generation; fault classification; Fault detection; Fault diagnosis; Faults; Feature extraction; Line current; Microgrids; Noise; Noise threshold; Real time; Resistance; weighted correlation coefficient
• ISSN: 0885-8977; 1937-4208
• DOI: 10.1109/TPWRD.2024.3510460

The diversity of line fault types, the uncertainty of fault resistances, the limitation of available fault information, and the similarity of positive pole currents under different work conditions challenge the fault diagnosis of DC microgrids. This paper proposes a double threshold fault diagnosis scheme for line faults with wide-range fault resistances only using local fault currents such as bus-side capacitor current and positive pole line current. One threshold is designed based on the bus-side capacitor current estimation through the bus-side capacitor voltage to detect low-resistance faults rapidly. The other threshold is designed based on a proposed weighted correlation coefficient between the bus-side capacitor current and positive pole line current to detect high-resistance faults under noise accurately. All types of line faults can be classified accurately in 1.5 ms based on the positive pole output current and the ratio of the bus-side capacitor current to the variation value of the positive pole output current. The proposed scheme improves the reliability and economy of fault diagnosis by only using positive pole information and non-real-time communication. Finally, the feasibility of the proposed fault diagnosis scheme is verified by simulations and experiments.