Performance of urban building energy models and its implications for district heating network design optimization

District heating networks can supply densely populated areas with renewable heating, but need precise planning to be economically efficient. Urban building energy models and district heating network design optimization can be leveraged to design thermal networks. This work proposes a method that combines building heating demand estimation with a highly spatially and temporally resolved district thermal energy grid design optimization model, and validates the results using building-specific hourly measurements. A district-wide coefficient of variance of root mean square error (CVRMSE) of 27.6% and a normalized mean bias error (NMBE) of 5.9% were achieved in the investigated case study of 130 substations with an hourly resolution, without model calibration. Cost-optimal multi-step designs lead to diversification of energy sources and longer, interconnected networks with limited increases in total annuitized costs of under 3% compared to single design point solutions. Different heating demand profiles can increase total annual district heating costs by up to 16% relative to the measurement-based scenario. Non-residential thermostat setpoint schedules had the largest impact on both heating demand accuracy and district heating network design outcomes, highlighting the need to further develop archetypes and user occupancy models for heterogeneous building types. The method enables energy planners and engineers to assess the potential effects of heating demand models on district heating network design outcomes, providing a flexible framework that can be applied to other case studies.