A simple, scalable and low-cost method to generate thermal diagnostics of a domestic building

Traditional approaches to understand the problem of the energy performance in the domestic sector include on-site surveys by energy assessors and the installation of complex home energy monitoring systems. The time and money that needs to be invested by the occupants and the form of feedback generated by these approaches often makes them unattractive to householders. This paper demonstrates a simple, low cost method that generates thermal diagnostics for dwellings, measuring only one field dataset; internal temperature over a period of 1 week. A thermal model, which is essentially a learning algorithm, generates a set of thermal diagnostics about the primary heating system, the occupants' preferences and the impact of certain interventions, such as lowering the thermostat set-point. A simple clustering approach is also proposed to categorise homes according to their building fabric thermal performance and occupants' energy efficiency with respect to ventilation. The advantage of this clustering approach is that the occupants receive tailored advice on certain actions that if taken will improve the overall thermal performance of a dwelling. Due to the method's low cost and simplicity it could facilitate government initiatives, such as the 'Green Deal' in the UK.