| Abstract Scope |
Nuclear and matrix graphite are key structural components in high temperature gas-cooled reactors because of their thermal conductivity and neutron-moderating capability. During air or water ingress accidents, however, accelerated oxidation can degrade component integrity and threaten reactor safety. Reliable prediction remains difficult as reported oxidation data vary with graphite grade, experimental configuration, gas environment, and specimen geometry, while additional testing is costly and time-consuming. This study presents a data-driven approach for predicting graphite oxidation under accidental air and water ingress conditions. A literature-based dataset was compiled from published papers and technical reports, including experimental parameters and oxidation responses such as mass loss, oxidation rate, and activation energy. Correlations among candidate variables were examined to identify the most influential features. Predictive models were then developed using different feature subsets, and their performance was compared to determine an accurate and efficient model for estimating oxidation behavior across multiple graphite materials. |