Cando Naula, Diego JonnathanSanango Fernandez, Juan BautistaChuqui Cajamarca, Freddy HernanZambrano Asanza, Sergio PatricioFranco Baquero, John Fredy2022-03-032022-03-032021978-166544421-70000-0000http://dspace.ucuenca.edu.ec/handle/123456789/38277https://www.scopus.com/record/display.uri?eid=2-s2.0-85117610692&origin=resultslist&sort=plf-f&src=s&st1=Graph+partitioning-based+clustering+for+the+planning+of+distribution+network+topology+using+spatial-+temporal+load+forecasting&sid=501320cdffad2f8063c8a14f51d5c49d&sot=b&sdt=b&sl=141&s=TITLE-ABS-KEY%28Graph+partitioning-based+clustering+for+the+planning+of+distribution+network+topology+using+spatial-+temporal+load+forecasting%29&relpos=0&citeCnt=0&searchTerm=&featureToggles=FEATURE_NEW_DOC_DETAILS_EXPORT:1Planning the expansion and the new topology of distribution networks requires knowing the location and characterization of the load as well as its future growth. Spatial load forecasting is a key tool in this task, providing high spatial resolution and adequate temporal granularity. Nowadays, with the penetration of distributed energy resources, multiple microgrid connection strategies, and implementation of self-healing and protection schemes, it is necessary to identify load blocks to plan the new active network architecture. Based on spatial load forecasting information, this paper proposes a graph partitioning technique to create load clusters in the distribution feeders. A weighted graph is constructed by means of a minimum spanning tree that allows to consider adjacency relations. The results of the simulation, carried out in a real distribution network, have demonstrated the effectiveness of the proposed method.es-ESDistribution planningClusteringMicrogridsMinimal spanning treeSpatial load forecastingGraph partitioningARTÍCULO DE CONFERENCIA10.1109/ISGTLatinAmerica52371.2021.9543010