Aggregation of residential buildings for thermal building simulations on an urban district scale

Abstract

Knowing the energy demand at the scale of neighborhoods allows the conception of efficient energy administration systems that aid to reach sustainability in the built environment. When energy demand data is not available, simulation models can provide estimations and thus enable the analysis of a neighborhood. To simulate the space heating demand of a residential building stock, often an aggregation is carried out. Aggregation implies using one or a few representative models to replace a larger number of building models. This paper explores the effect of the aggregation method on model accuracy when applying a first-order building model for the space heating demand of an urban residential district. The results show that aggregation leads to inaccuracies when the district includes buildings with varying values for their properties: the errors reach 14% if the district is highly polarized and 8% for more diverse districts. Aggregating buildings with identical properties diminishes the error compared to a total aggregation − aggregating the building stock as a whole. Aggregation with respect to certain building properties yields better results than others, with window area fraction and U-value of opaque surfaces leading to the first and second lowest error values.