<![CDATA[The 20 kV distribution network is crucial for ensuring the uninterrupted supply of power to users in the PT. PLN UP2D North Sumatra Region. Disruptions in this network, including short circuits, overloads, and transient disturbances, can diminish system reliability and prolong outage durations if not promptly identified and rectified. This study seeks to develop a disturbance detection and recovery system for a 20 kV distribution network utilizing real-time analysis through the Support Vector Machine (SVM) algorithm. The system is designed by utilizing real-time electrical parameter data, including current, voltage, and network operational conditions, sourced from monitoring devices. The data undergoes preprocessing, feature extraction, and classification stages utilizing SVM to differentiate between normal and fault circumstances. The classification outcomes serve as the foundation for decision-making in isolating the fault zone and restoring supply to the unaffected segments of the network. The system's performance is assessed according to detection accuracy, response speed, and its capacity to facilitate the disturbance recovery process both automatically and semi-automatically. This research aims to enhance the dependability of the 20 kV distribution network, expedite fault resolution, and facilitate the advancement of a more intelligent, efficient, and responsive electrical distribution system.]]>
Copyrights © 2026
