Researchers in Edinburgh are proposing the use of AI to help heavy industry model carbon capture and storage methods more quickly and cost effectively.
Experts at Heriot-Watt university have been working on a way to harness the technology to help carbon intensive industries such as steel, cement and chemicals reach net zero.
Called ECO-AI, the research team is developing specialist AI techniques for scientific computing, material discovery and financial forecasting.
This will enable CO2 capture and storage in deep geological formations whilst setting out the financial implications in deploying these techniques for businesses and for policymakers.
The university’s global research institute for net zero is spearheading the project – with £2.5 million funding from UK Research and Innovation – in partnership with Imperial College London.
It is expected to leave a significant framework for future researchers to build upon and will play an important role in helping the UK Government to reach its net zero target by 2050.
The project aims to develop energy-efficient solvents for CO2 capture followed by permanent storage of captured CO2 into geological storage sites, through various AI techniques.
Once captured, the CO2 is compressed into a fluid almost as dense as water and pumped down through a well into a porous geological formation.
The research is one of the projects being delivered by Heriot-Watt University’s global research institute which is focused on achieving net zero and beyond. Called iNetZ+, the team brings together a range of scientific expertise including chemical engineering, physics, geology, mathematics, computer science and economics.
By using specialist AI simulators, standard techniques can be replaced for modelling flow migrations, and simulations on a supercomputer that may have previously taken up to 100 days can now be achieved in just 24 hours.
Professor Ahmed H. Elsheikh, leader of the data and artificial intelligence research theme at iNetZ+ says: “Our efforts for the ECO-AI research are primarily focused on refining algorithms that can potentially be applied to CCS in the future in typically hard to decarbonise industries.
“Our research has the ability to really advance existing scientific research streams to source suitable options for safe storage of CO2 without consuming too much energy and without the need to deploy expensive and often time-consuming exploratory investigations.”
To strengthen its research, the university hosted a two-day workshop and three-day hackathon event, bringing together leading experts in AI, computational science, and CCS. The workshop event highlighted the vital role of interdisciplinary collaboration and discussed using digital twins for decision-making around reaching net-zero emissions under uncertainty, as well as incorporating simplified models into large-scale optimisation replicas for complex systems.
The three-day hackathon event provided an opportunity for teams to develop AI-based solutions for challenges related to CO2 capture, storage, and policy/economics. With access to tools, data, and expert support, they tackled tasks like discovering new materials, modelling subsurface fluid flow, and analysing patterns in carbon capture and storage patents.
Professor Clare McCabe, co-leader of the project’s carbon capture component and Bicentennial Professor at Heriot-Watt University adds: “The enthusiasm and ingenuity displayed during the hackathon was truly inspiring. Students and postdocs coded tirelessly, developing AI models that could reshape how we approach CCS technologies. I was very impressed with the workshop, but especially so with the optimism and energy in the hackathon. Students and postdocs from both within and outside the ECO-AI project coded for three days working on various AI models related to the ECO-AI project.”
