Community scale dynamic building simulations using probabilistic behavioural patterns

The upscaling of energy models of buildings has entered mainstream discussion within the subject of energy system modelling (ESM), the primary role of such models being to generate policy; however, policy holds the potential to stimulate significant changes in energy demand, especially within the residential building sector. This cyclic system is not modelled fully at administrative levels in the UK. The mechanisms to infer useful insight into future demand caused by millions of new heat-pumps and electric vehicles, for example, must be implemented as an integrated part of the modelling process.

The move away from data-derived demand curves is exemplified in the present work. This focuses on a small community and explores methodologies to provide scalable solutions to characterise residential thermal demand. Despite being inherently deterministic, the tools employed have been configured here to run sequences of probabilistic inputs to deliver aggregate loads for a diverse building stock. The dwellings in this study have been classified into their respective archetypes based on building form and construction. Smart meter data have then been used to generate behavioural patterns which describe how the dwellings are used. Finally, probability distributions have been applied to the behavioural patterns to consider variability across the sub-groups within the stock.