High-impedance-to-ground fault detection method for radial distribution networks based on GWO–SVM

Le résumé fourni par la source

With the continuous increase in the cableization rates of medium-voltage distribution networks in Chinese cities, the low-resistance grounding system (LRGS) is being widely adopted in urban distribution networks, owing to its advantages such as fast fault-clearing speed and low overvoltage levels. To mitigate the difficulty of detecting high-impedance-to-ground faults (HIGFs) in LRGSs, this study proposes a HIGF-detection method for LRGSs, based on multifeature fusion. First, the equivalent models and fault characteristics of three different HIGF processes, namely, single-phase grounding, two-point in-phase cross-line grounding, and two-point out-of-phase cross-line grounding, are analyzed. Then, the value of the neutral-line zero-sequence current amplitude is selected as the startup criterion for the HIGF-detection model. Next, HIGF characteristic criteria suitable for the detection of the three types of HIGF processes are selected to construct the HIGF-detection model for LRGS. A simulation model is built using MATLAB/Simulink, and fault characteristic datasets are obtained by setting different HIGF conditions. Finally, these fault characteristic datasets are used to train a support vector machine optimized using the gray wolf optimizer, to construct the HIGF-detection model. The results show that the proposed HIGF-detection model for radial distribution networks based on multifault feature fusion can effectively detect ground faults with transition resistances up to 4800 Ω and accurately identify fault lines, which is significant in the field of HIGF detection for LRGSs.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.