CO₂ corrosion of carbon steel is a major degradation mechanism in industries including oil and gas, geothermal energy, and carbon capture and storage (CCS). Accurate prediction of corrosion rates is essential for materials selection, asset integrity management, and safe operation. As a result, a wide range of prediction tools has been developed, ranging from empirical correlations to sophisticated mechanistic models. Over recent decades, advances in understanding of the physicochemical processes governing CO₂ corrosion have enabled the development of increasingly more sophisticated models. Mechanistic approaches seek to represent the interactions between electrochemical reactions, mass transport, fluid flow, and evolving surface chemistry, providing both corrosion rate predictions and insight into the processes controlling corrosion behaviour.
This seminar reviews the evolution of CO₂ corrosion prediction models, with particular emphasis on the mechanistic models. Model strengths, limitations, and practical challenges are discussed, including the requirement to balance physical realism with robustness, computational efficiency, and industrial applicability.
Following the historical review, research conducted at the University of Leeds over the past five years will be presented, focusing on the interrogation, evaluation, and advancement of mechanistic CO₂ corrosion models through multiphysics modelling. By independently recreating and assessing published models, the work demonstrates how interactions between the underlying physical and chemical processes can be examined, how model sensitivities can be evaluated, and how opportunities for further development can be identified.
The seminar will then highlight ongoing research at Leeds aimed at addressing key limitations of current mechanistic models. These activities include the application of machine learning techniques to improve computational efficiency, robustness and accessibility. In addition, the extension of modelling approaches to complex geometries and more representative flow conditions, and the development of improved methods for incorporating the formation, evolution, and protective effects of corrosion products.
Finally, the broader potential of combining multiphysics modelling with machine learning will be considered. Together, these approaches offer a pathway towards corrosion prediction tools that are more accurate, computationally accessible, interpretable, and relevant to the complex operating environments encountered in industry.
Professor Richard Barker is Professor of Corrosion Science and Engineering at the University of Leeds and a Fellow of the Institute of Corrosion, with over 15 years of academic experience. In 2021, he was awarded the Kurt Schwabe Prize by the European Federation of Corrosion, recognising scientific and technical contributions of early-career researchers under the age of 35 in the field of corrosion science.
Professor Barker’s research focuses on understanding corrosion mechanisms in complex and demanding energy-sector environments, including oil and gas production, geothermal energy systems, and carbon capture, utilisation and storage (CCUS). A particular strength of his group is the development of novel experimental systems, apparatus and methodologies to investigate corrosion processes under representative service conditions. The group also works closely with industry to optimise corrosion management strategies and develop innovative approaches and models for corrosion assessment and control.
