A multi-terminal traveling wave fault location method for active distribution network based on residual clustering. (October 2021)

• Record Type: Journal Article
• Authors: Qiao, Jian; Yin, Xianggen; Wang, Yikai; Xu, Wen; Tan, Liming
• Is Part Of: International journal of electrical power & energy systems. Volume 131(2021)
• Publication Date: 2021-10
• Subjects: Fault location -- Traveling wave -- Residual clustering -- Active distribution network; 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.2021.107070
• Languages: English
• ISSNs: 0142-0615

Abstract

Highlights: DBSCAN algorithm is first utilized to identify the bad data which are affected by time errors. The remaining normal data are used to calculate the corrected solution of the fault distance. It can greatly improve the accuracy of fault location and has a strong tolerance to time errors. Simulation and field tests show that the proposed method has the advantages of high location accuracy and strong robustness to time errors. It can be effectively applied in the active distribution networks. Abstract: Since distribution networks have multiple branches, complex topologies and increasing penetration of the distributed energy resources (DERs), the accurate fault location is difficult to realize. The existing traveling wave fault location methods are strongly affected by the arrival time errors. To overcome the problems mentioned above, a multi-terminal traveling wave fault location method is proposed for active distribution networks based on residual clustering. Firstly, the traveling wave arrival times are utilized to construct a minimized optimization model for each section. The objective optimization function represents the minimization of the sum of squared errors (MSSE), and the global optimal solutions reflect the wave velocity and the fault distance. Subsequently, the particle swarm optimization algorithm (PSO) is used to solve the above optimization models, and the section with the minimum MSSE is judged as the faulty section. Finally, the density-based spatial clustering of applications with noise (DBSCAN) algorithm is utilized to group the residuals of the faulty section to identify the bad data, which are affected by the arrival time errors. And the normal data remained are applied to reconstruct the optimization model and calculate the optimal solution of the fault distance. Thus, the fault location results can be corrected. Simulation results and field tests indicate that the proposed method has high fault location accuracy, strong robustness to time errors and high adaptability for active distribution networks.