Assessing Model Sensitivities in District Heating Planning

A Case Study Using a Python-Based Toolchain

Abstract

A python‑based toolchain was developed to generate optimal district heating network topologies from open‑source data, combining consumer demand modelling and network routing. Using a case study in the city of Meersburg, four optimization variants were analyzed regarding the influence of additional information: (1) baseline OpenStreetMap(OSM)‑only data, (2) peak loads assumptions enhanced by anonymized energy consumption data, (3) additional shortcut route connecting the lower and upper city through vineyards, and (4) both load and shortcut data. The four variants vary in main‑line routing and pipe diameter allocation, some of them including solutions that are feasible according to OSM but excluded by planners for accessibility reasons. Quantitatively, total built pipe lengths varied by 4–6 % and investment costs by 4 % relative to the manual plan. These results highlight the impact of incorporating realistic load peaks and spatial constraints in automated routing, while demonstrating that a fast, data‑driven approach can result in a first estimation that lies within a few percent of the final derived costs. The automatic connection of buildings to the streets represents a significant portion of the built pipes and should be further investigated.