Enhancing Power Distribution Feeders Restoration with a Probabilistic Crew Dispatch Method: Case Studies using Historical Data from a Brazilian Power Distribution Company

Abstract

Abstract This paper proposes a methodology to repair crews dispatch during distribution feeders restoration immediately after remotely controlled actions have been taken. Our method determines the sectors patrol sequence using the expected number of faults for each sector considering a previously calculated fault probability. Next, a road graph mapping of the buses is obtained by associating each bus with its real-world position. The crews’ route inside each sector is determined by the application of the nearest neighbor algorithm on that mapping. The case studies analyzed 24 real faults available in the database of a Brazilian power distribution company, and, a comparison was made between the time spent to locate the failures by simulating both the proposed methodology and the usual greedy strategy. The average time spent in localizing all the failures was from 9.49 % to 41.81 % shorter than in the usual method. Moreover, the speed and distance of repair crews showed weak influence in the crew dispatch methodology efficiency. Such results indicate an enhancement of the power distribution QoS indicators by using the proposed methodology to deal with faults.