Picture credit: MChe Lee on Unsplash

Classrooms remained empty back in September, when the government advised 104 schools to close all buildings made of reinforced autoclaved aerated concrete (RAAC), due to the potential for collapse. With RAAC used in building many public buildings in the 50s, 60s and 70s, it’s a fairly common substance. But the good news is that Loughborough academics have trained an AI (artificial intelligence) to detect RAAC deterioration.

While some people think the situation is not as dramatic as it was painted in the press, the reality is that tens of thousands of RAAC structural panels can be found across schools and public buildings, including hospitals, and some of them may be in danger of collapse.

The problems start when degradation occurs – due to water ingress, corrosion, spalling or overloaded structures. Essentially, it’s a maintenance problem. Compounded by the fact that RAAC was not necessarily intended to be used for such a long time. As Professor Christopher Gorse from Loughborough University puts it, there’s no evidence that RAAC was designed for a lifespan beyond 25 years.

The team at Loughborough’s School of Architecture, Building and Civil Engineering were called in by the NHS in 2021 to look at RAAC in hospitals. And as part of this work they started developing a machine learning tool that uses photographic archives to identify cracks in RAAC panels and predict how they might behave.

Dr Karen Blay, a Senior Lecturer in Digital Construction and Quantity Surveying at Loughborough, looked at survey methodology and worked with the maintenance team to develop a way to observe the behaviour of RAAC and come up with maintenance solutions.

At first the team fed the AI 85,225 images of cracks on normal concrete, followed by 1850 images of cracks taken within the NHS Trust. This training allows the AI to detect RAAC deterioration with 95.8% accuracy. A maintenance worker could run images through the software, and it will let them know which ones have cracks in.

A digital twin is then created so Loughborough’s scientists can understand how changes happen. Digital twins are a virtual representation of an object or system that spans its lifecycle, is updated from real-time data and uses simulation, reasoning and machine learning to aid decision making. Using the time stamp of the photographs along with the locations of the cracks, they’ll be able to plan maintenance and understand better how changes happen.

The project aligns well with the Building Safety Act and the golden thread of information. The act is part of the new safety regime which aims to prevent events like the Grenfell Tower disaster. It sets out a clear pathway on how residential buildings should be constructed, maintained and kept safe. The golden thread of information is a requirement to ensure safety is considered at every stage of the building’s life cycle.

Meanwhile, if you need assistance with the structural elements of an upcoming project, please do get in touch.