Methods of predicting urban domestic energy demand with reduced datasets: a review and a new GIS-based approach

Abstract

This paper reviews existing bottom-up approaches to the urban scale prediction of domestic energy demand and introduces a new approach based on Geographical Information Systems (GIS). It describes how software tools can be used to derive predictors from the plan form of dwellings extracted from digital maps and, using a new default-data system, to satisfy the large data requirements of an embedded version of the BREDEM-8 domestic energy model. Energy consumption and CO₂emissions of urban dwellings can be predicted quickly and robustly without costly on-site measurements or statistically derived geometrical models. Urban areas of almost arbitrary size can be modelled and accuracy can be incrementally improved with additional data. The model is validated by comparison with results produced by the UK's National Home Energy Rating scheme software using complete site-surveyed datasets. Its ability to rapidly estimate energy demand is demonstrated for around 400 dwellings in Leicester, UK using only the dwelling age, the default data system and the GIS tools.