About the project
The main research problem in this project will be to explore how AI tools can help domestic end users in their switch to renewable energy, and to do so in a trustworthy manner.
Investing in cleaner renewable energy generation in domestic settings is highly challenging for end users. Installing solar PV, battery storage and/or switching to EVs is expensive and requires reasoning about long-term costs, high uncertainty and behaviour will affect both the return on investment and the environmental sustainability of the installation.
This project will focus on several different aspects including:
- optimisation of a renewable energy installation given the properties of a user’s home and historical consumption data
- lifetime monitoring and optimisation of the system, including automatic energy management (through heating, car charging and import/export of energy)
- suggestion of behaviour interventions based on consumption data and limited interactions with the end user
Trust is a key aspect of this – the project will therefore explore techniques for quantifying and uncertainty, explaining calculations and assumptions clearly to users. To enable this, it will involve running focus groups, surveys and field trials with users.
Another aspect is to consider possible incentive schemes to encourage behaviour modifications or demand response.
In terms of methodology, the project will combine the use of optimisation, machine learning (for demand and behaviour predictions, including under incentives) and aspects of explainable AI.
