Towards simple dynamic Urban Building Energy Models that capture annual space heating demand and peak load in residential buildings

Despite its momentum in building energy modelling, grey-box RC modelling is still faced with the challenge of scaling across the heterogenous building stock. Therefore, this paper presents a white-box model identification procedure that allows to develop simple white-box mathematical structures for the modelling of the dynamic heat balance of individual buildings within bottom-up Building-Stock Energy Models. The procedure uses artificial measurement data to identify simple grey-box RC structures. Simplified dynamic white-box RC structures are established through pattern searches between the grey-box parameter estimations and the aggregated artificial training data. The identified white-box models showed good performance for predicting the yearly heat demand in residential buildings at stock level.