IDEAL and its partner project, BIGSMALL, were UK EPSRC-funded research projects to progress home energy monitoring and feedback methods.
The work aimed to give new insights into domestic energy demand and approaches to influencing it, contributing to the wider goal of facilitating the transition to a low carbon, secure and affordable energy system.
Phase one ran from April 2013 to June 2018.
Work centred around four core aims:
- Gathering a high-resolution household dataset: We developed and installed an ultra-low power wireless sensor network system in over 250 homes to collect high frequency data on whole-home energy use, boiler pipe temperatures and per-room temperature and humidity over extended periods of time (months to years).
- Inferring energy using behaviours: Appliance-level energy monitors in a subset of the homes and improved machine learning methods allow us to identify patterns of home appliance use and to infer which everyday behaviours account for a home’s energy usage.
- Understanding occupant behaviours: Our methodology of combining sensor data, machine learning inferences and survey data provides us with a long duration, high resolution dataset to investigate in great detail how everyday behaviours shape household energy use, and how these are influenced by socio-demographic, building, climatic and other factors.
- Enhanced energy feedback: Through a co-design approach with project participants we have developed a range of novel energy feedback tools and then evaluated their effects over extended periods of time in real homes, with real people, in a randomised controlled trial.
Impact and outputs
Our work aims to have impact on:
- policy insight into the underlying behavioural trends and ‘drivers’ of home energy use
- effective smart meter feedback design, policy and practice
- improved customer segmentation to better understand who engages in which high and low energy activities, and who uses energy feedback
- academic understandings of energy behaviours and practices and the factors influencing them, and Machine Learning “Non-Intrusive Load Monitoring” methods
- methodologies for evaluating effects of interventions and occupant effects on energy use in buildings
For a list of current publications and project outputs from these and other Sustainlab projects, see the Publications page.
|IDEAL and BigSmall were funded by the UK Engineering and Physical Sciences Research Council, grant references EP/K002732/1 and EP/M008223/1.
|A member of the TEDDINET network of energy demand projects.