Automated Generation of Synthetic Distribution Grids Based on open GIS Data

Abstract

This thesis discusses the generation of synthetic models for medium voltage (MV) networks from open-source data to fill the void left by the lack of publicly available real information in Germany. It aims to extend the state of the art by combining ideas, methods and tools that have been used in previous works. Concretely, it retrieves and leverages OpenStreetMap (OSM) data to locate MV infrastructure, validates it by comparing to other OSM and external data sources, extends it as necessary and creates a graph upon the connected street network which VRPy, a vehicle routing problem (VRP) solver, uses to generate routes of minimal total length from the capacitated VRP. Topological parameters of the generated network are vetted against similar and publicly known examples and its viability is checked via static power flow analysis with pandapower for diverse scenarios of network conditions. The grid model and other relevant data are visually presented in interactive html maps. The base case studied is the city of Darmstadt but the tool can be flexibly configured for many scenarios. The key indicators calculated for the generated Darmstadt model compared favorably with the reference data at hand. The network also responded appropriately to the scenarios presented. This is an initial endorsement of the viability of the tool and its approach for use in other projects requiring a synthetic MV grid model to stand in for unavailable official data. The approach used in this project can be replicated, adapted or expanded for other similar needs, such as creating synthetic low voltage grid models.