Voltage Sag Homology Detection Based on DBSCAN Algorithm

Meijing Jiang

Abstract: Inverters, AC contactors and other equipment used in high-tech manufacturing are very sensitive to voltage sags. Voltage sags may cause equipment failure, production interruption, data loss, damage to sensitive equipment and unstable energy supply. A short circuit fault may trigger multiple power quality monitoring devices to record voltage sag waveforms. The problem of voltage sag data redundancy seriously affects data application. Therefore, identifying the source of voltage sags is of great significance for scientifically and rationally evaluating the severity of regional power grid voltage sags. Therefore, this paper proposes a voltage sag source identification algorithm based on the DBSCAN algorithm. By adopting appropriate feature engineering, three-dimensional clustering features are selected, and then appropriate clustering algorithm parameters are selected through iterative method for clustering. Finally, the algorithm effect is evaluated through six clustering evaluation indicators. Experiments were conducted on the jupyter notebook programming platform using data provided by a provincial power company. The final results prove the effectiveness of the proposed algorithm.

Keywords: voltage sag, clustering, DBSCAN, voltage sag homology detection