Cambridge researchers from the Department of Architecture at the University of Cambridge have developed an AI model that can identify hard-to-decarbonize (HtD) houses, which are responsible for over a quarter of all direct housing emissions. These houses pose a significant obstacle to achieving net zero emissions but are often overlooked for improvement efforts. The deep learning model promises to make it easier, faster, and cheaper to identify these problem properties and develop strategies to improve their environmental credentials.
The AI model, trained using open-source data, can classify HtD houses with 90% precision, and the researchers expect this accuracy to increase as they add more data. This groundbreaking approach can save policymakers precious time and resources by directing them to high-priority houses that need decarbonization interventions.
The model also enables authorities to understand the geographical distribution of HtD houses, allowing for targeted and efficient deployment of interventions. The researchers fed the model data from Energy Performance Certificates (EPCs), street view images, aerial view images, land surface temperature, and building stock to identify HtD houses in Cambridge, UK. They identified 700 HtD houses and 635 non-HtD houses using this data.
The researchers are now working on an advanced framework that incorporates additional data layers related to factors such as energy use, poverty levels, and thermal images of building facades. This will further increase the accuracy and detail of the model’s predictions.
By empowering policymakers with this AI model, it becomes easier to identify and prioritize the decarbonization of homes. The researchers argue that making data more visible and accessible to the public will facilitate consensus and action towards achieving net zero emissions.
Cambridge is a particularly informative study site for this AI model because despite being relatively affluent, it faces challenges due to its old housing stock and building bylaws that restrict retrofitting and the use of modern materials in historically important properties.
The researchers will discuss their findings with Cambridge City Council and continue collaborating with Cambridge Zero and the University’s Decarbonization Network.
With its ability to pinpoint specific areas of buildings, such as roofs and windows, that contribute to heat loss, the AI model can guide residents and authorities in targeting retrofitting interventions to improve energy efficiency.
By utilizing AI and open-source data, policymakers can make informed decisions based on evidence, enabling effective climate change adaptation strategies.
This innovative research paves the way for more inclusive and accessible approaches to decarbonization and positions AI as a powerful tool in achieving net zero emissions.
