In 2010 the housing stock was responsible for 30.5% of all energy consumed in the UK. The UK government has set a transition target to reduce the energy used from space heating in dwellings by 29% by 2020 as part of their drive to lower CO2 emissions and mitigate the risks of global climate change. Housing stock energy models have been developed as research tools to identify pathways to a low energy future. These tools use assumptions about how homes are heated that may reduce their effectiveness at making accurate energy predictions.
This thesis describes the collection and analysis of temperature data from over 300 homes in Leicester to develop better understanding of how dwellings are heated. The temperature measurements were assessed for error and a final sample of 249 dwellings was established. Mean winter temperatures (December February) were found to be 18.5°C and 17.4°C for living rooms and bedrooms which are comparable with temperatures reported in previous studies. Statistically significant relationships were established between seven descriptors; three technical (house type, house age and wall type) and four social (household size, employment status, age of oldest occupants and tenure). Only 24% of the variation in mean winter temperature could be explained by these descriptors.
Ten heating practice metrics were developed to give insight into how homes are heated; these included the duration of the heating period and the average temperature when heated. Statistically significant relationships were found between the heating practices and a number of technical and social household descriptors. It is concluded that the variation in heating practices which relates to social household descriptors will result in models being unable to make accurate predictions at the regional of city scale. Furthermore, this work has shown flaws in the idealised temperature profile as used in BREDEM. It is suggested that the findings of this work are considered in the development of future stock models.