Deep.Meta has proven it can reduce steel production emissions by close to 10% at Spartan UK’s steel plant in Newcastle-upon-Tyne.
Deep.Optimiser-PhyX is an AI-powered Digital Twin — a smart digital replica of the steel production process that combines physics and machine learning to optimize furnace operations, simulating years of production in just a few hours.
By using real-time sensor data and material science, it more accurately predicts steel slab temperatures and improves scheduling, boosting energy efficiency, which in turn significantly cuts emissions. The technology will enter the live pilot stage at the Spartan UK plant.
Dr Osas Omoigiade, CEO and founder of Deep.Meta, and finalist of the second Manchester Prize, said: “We are developing Deep.Optimiser-PhyX to tackle inefficiencies that result in avoidable emissions — a crucial step in helping to decarbonize the industry. Through the Manchester Prize we have been able to integrate physics into our AI platform, which enhances its prediction capabilities further.
“Our ultimate ambition is to save 10 megatonnes of CO2 from entering the environment by 2030, creating a lasting impact here in the UK and across the steel industry. Our work with Spartan UK is a crucial step towards achieving that. If we are selected as the winner of the Manchester Prize, we want to scale our development work with furnace machine providers for integration across UK producers and continue expanding into other regions, including North America.”
Dr Kwangkyu Alex Yoo, Deep.Meta, senior machine learning scientist said: “Today’s machine learning models often operate as black boxes, lacking fundamental principles that clearly link inputs to outputs. This creates significant resistance when industries attempt to deploy AI technologies in real production environments. Our physics-based machine learning approach addresses these challenges by incorporating the underlying physical laws into both the training process and data generation. This leads to models that are more explainable and trustworthy, while enabling more reliable and robust decision-making.”
Michael Brierley, Spartan UK, CEO said: “Deep.Meta is a trusted partner, and we are piloting the Deep.Optimiser solution, because of the rising costs of energy and carbon. Increasing the efficiency of production is of high importance as energy costs form a significant part of our cost structure.”