Glass Futures launches AI-driven digital twin to reinvent glass manufacturing

Glass Futures (GF) has installed a unique AI-driven ‘digital twin’ of its glass furnace capable of testing and predicting new and the best ways to make glass.

Located at its global centre of excellence in St Helens, Glass Futures has implemented a digital and immersive model of its first-in-the-world multi-fuel pilot furnace.

Combining data from the ‘real’ furnace and simulated information with the laws of physics, the digital twin can process over 11k calculations at once accurately predicting glass output.

It means industry will be able to test changes in temperature, pressure, density and other measures on the digital furnace revealing accurate predictions of how this would work in a real furnace.

The AI-GLASS project has been running for the last year with NVIDIA, a world leader in artificial intelligence and The University of Liverpool’s Virtual Engineering Centre, which has created a fully immersive model viewed through VR headsets.

Dr Jim Scotson, project lead for industrial digitalisation at Glass Futures, said: “The digital twin enables us to test things industry has never tried out before. It means we can drive innovation by allowing manufacturers of glass and other materials to try new fuels, electric heating, bubbling and other techniques and get an accurate prediction of how the end product will turn out.

"The AI is incredibly powerful. Thanks to "NVIDIA RTX Pro 6000 GPUs, we can combine real pilot-line data with 16 sensor points and NVIDIA PhysicsNeMo, meaning it is informed by the laws of physics.

"Rather than just learning from our real and simulated data it already has a physics informed neural network so when we put data in, it digests this with physics principals meaning it’s incredibly accurate and can calculate 11,530 calculations all at once. This allows us to remove the barriers to experimentation by making things possible, more quickly."

By creating an exact replica of a real furnace, it means industry can be more experimental, testing new ways to decarbonise, improve efficiencies and confidently adopt new low-carbon alternatives such as hydrogen and biofuels.

Anthony Hills, Director of UKI, said: "AI-GLASS shows how advanced AI can unlock entirely new ways of understanding and improving complex industrial processes. By combining NVIDIA’s accelerated computing with physics-informed modelling, Glass Futures is turning data into actionable insight at unprecedented scale. This kind of capability allows industries to move faster, experiment with confidence and make smarter decisions as they work toward more sustainable and efficient production."

Glass Futures will now work with its members and cross-industry partners in steel, ceramics and other sectors on projects using the digital twin and then the real pilot line and also use their expertise to support other teams to digitalise.

Dr Scotson added: “This project wasn’t without its challenges. When we talk about flame shape and know just by looking at it how the furnace is performing, how do you describe pointy, jagged or flat to a computer? We solved this via temperature measurement, but we’ve learnt so much that we now want to support others with their digital, AI-driven projects.

"It might not be about a complete digital twin but how we can support them to take analogue information, digitalise it and use AI and modelling to extend the life of existing equipment."

Dr Konstantin Vikhorev, Chief Technology Officer at the VEC, added: "AI Glass will revolutionise how industry approaches decarbonisation. By merging advanced modelling and AI, we can help manufacturers explore new fuels and materials in seconds, reducing risk and accelerating progress toward sustainable, efficient and cleaner glass production."

AI-GLASS is part of a wider £1.5m Innovate UK funding programme being delivered through Make UK.

For more information about the project or to discuss further please contact Jim at jim.scotson@glass-futures.org