Diverse occupant behaviour and urban building heterogeneity to enhance urban building energy modelling

• Introduced DOB-HUBS framework to address UBEM limitations by incorporating OB diversity and urban heterogeneity. • Assessed behavioural responses to Sweden’s 2027 tariff across heterogeneous building clusters. • Revealed strong seasonal impacts of window-opening on energy use and CO 2 emissions. • Planners and policymakers can design effective occupant engagement strategies. Upcoming energy transition initiatives, such as Sweden’s capacity-based electricity tariff in 2027, aims to incentivize changes in occupant behaviour (OB) influencing how occupants schedule and shift energy use. Urban building energy modelling (UBEM) is widely used for city-level energy assessment. However, most UBEM studies assume uniform occupant behaviour and homogeneous building properties due to modelling limitations in incorporating diversity and heterogeneity, which is computational infeasible at large scale. In practice, OB varies across demographics and seasonal routines, while heterogeneous building properties such as U-values, HVAC, archetypes directly affect energy impacts. Oversimplifying OB diversity and urban building heterogeneity can misrepresent energy demand, peak loads, and policy effectiveness of energy transition initiatives. This study introduces an enhanced framework, referred as DOB-HUBS, that explicitly incorporates OB diversity (e.g., seasonal, demographic) along with urban heterogeneity into UBEM. This enables ability to evaluate behavioural impacts across heterogeneous building clusters and occupants’ seasonal interactions. The approach employs unsupervised K-Means clustering and proportional stratified sampling to identify representative buildings, an automated bottom-up physics simulation workflow programmatically assigning diverse OB schedules and building properties, and an ensemble machine learning model that efficiently extrapolates predictions to the entire stock. This framework is demonstrated in Sweden, by evaluating seasonal impacts of excessive window opening, and behavioural responses to 2027 tariff policy. Results reveal strong seasonal dependence of OB, with deviations of up to 15% with 3.06% (standard deviation) energy use compared to traditional approach. Behavioural responses to 2027 policy could reduce peak loads by 6–17%, with impacts varying across small building clusters. By evaluating OB diversity and urban heterogeneity, DOB-HUBS enhances UBEM, supporting urban planners and policymakers in designing more effective occupant engagement strategies.