AI-based models for predicting rock/mineral-hydrogen-brine contact angles using experimental data


Hajibolouri E. Shafiei A.
7 November 2025Elsevier Ltd

International Journal of Hydrogen Energy
2025#186

In this paper, six intelligent models were developed and tested using various artificial intelligence algorithms for accurate prediction of contact angle (CA) for 10 rock/mineral types in rock/mineral-H2-brine system. A comprehensive database with 1085 experimental data that covers a wide range of geological conditions was used. Both static and dynamic CAs were considered in the modeling to tackle the uncertainty associated with common CA measurement methods. Monte-Carlo SHAP analysis showed that organic acid concentration, pressure, and mineral type are the most sensitive variables in this process. The proposed models can accurately predict wettability characteristics under various geological conditions. The CatBoost and XGB models achieved a great performance with coefficients of determination of 0.979 and 0.978 and root mean square errors of 3.186 and 3.199, respectively. This research work provides a reliable tool for CA prediction which is required for adequate assessment of hydrogen storage capacity.

Contact angle , H2 geo-storage , Machine learning , Rock/mineral-H2-brine , Wettability

Text of the article Перейти на текст статьи

Petroleum Engineering Program, School of Mining & Geosciences, Nazarbayev University, Astana, 010000, Kazakhstan

Petroleum Engineering Program

10 лет помогаем публиковать статьи Международный издатель

Книга Публикация научной статьи Волощук 2026 Book Publication of a scientific article 2026