Prediction of heat demand is of distinct importance for policy planning in social housing, where residents are in higher danger of falling into fuel poverty. Understanding the behavioural response of fuel vulnerable households against weather allows generating accurate baseline energy models and estimations of energy savings. This paper evaluates weekly heat demand profiles of 462 social housing dwellings in five tower blocks in the South of the UK, monitored over two years. Linear and segmented regressions are fitted through the 'segmented' package in R Studio to explore the relationship between heat demand (including Domestic Hot Water and space heating) and Outdoor Temperature, generating energy signature models for each flat. Three distinct heat demand profiles are found: (i) households that do not use space heating (11), (ii) irregular consumers, where the transition towards the heating season is not identifiable (33), and (iii) households with marked seasonal thresholds (56). Consumption trends as well as the effect of extremely weather events such as the 2018 storm 'The Beast from the East' on the heat demand are evaluated. Low consuming households show little to no variation in their demand patterns during extreme weather events, whereas higher consumers seem to reach a plateau in their demand even at extremely low temperatures. The variability of heat demand in dwellings which have identical physical properties, and are exposed to the same weather conditions, is attributed to occupant behaviour. This study highlights the heterogeneity of heat demand in social housing and the need to move away from national averages.
展开▼
