Automated Generation of Urban Medium-voltage Grids using OpenStreetMap Data

Realistic geo-referenced electrical distribution grid (DG) models are of great importance for power system analysis and resilience studies. However, DG data are usually not publicly available. In this study, we develop a new process for the automated generation of medium-voltage (MV) grid topologies, specifically for urban areas, based on openly available data and open-source software. OpenStreetMap (OSM) data on power infrastructure, street layouts, and land use, are used as the only input source. In contrast to previous works on DG reconstruction, we use available OSM data on substation locations. Different existing methods are combined in a new, hybrid approach by considering the incompleteness of OSM data, taking the street network into account, and applying the Capacitated vehicle routing problem (CVRP) to find cost-optimal routes for power lines. Our method is tested with a German city as a case study. Furthermore, we verify the result using land use data and evaluate the quality of power-related OSM data. The results demonstrate that our approach can yield realistic geo-referenced MV grid topologies, even with incomplete OSM power data.