Data-driven modeling to predict static contact angle in H2/brine/mineral/cushion gas systems: Implications for underground H2 storage


Mahmoudi Kouhi M. Piroozi G. Kalam S. Kammakakam I.
15 November 2025Elsevier Ltd

Journal of Energy Storage
2025#136

Underground hydrogen storage (UHS) is a viable solution for large-scale energy storage. Reservoir rock wettability, quantified by the contact angle, is a critical parameter that governs storage efficiency and hydrogen recovery. While experimental measurements are essential, they are often costly and time-consuming. This study develops machine learning (ML) models to provide a rapid and accurate alternative for predicting the contact angle in H2/brine/mineral/cushion gas systems. Using a comprehensive dataset of 1397 experimental data points, we evaluated several advanced ML algorithms based on rock composition, pressure, temperature, brine concentration, and cushion gas type. The CatBoost model demonstrated superior performance, achieving a coefficient of determination (R2) of 0.97 and 0.93, a root mean square error (RMSE) of 1.41 and 2.43, mean absolute error (MAE) of 0.85 and 1.41, and average absolute relative deviation (AARD) of 2.74 and 4.37 % for training and testing dataset, respectively. Furthermore, a novel and explicit mathematical correlation was developed using the Group Method of Data Handling (GMDH), providing a transparent equation with an R2 value of 0.83 and 0.82 for training and testing data, respectively. A key finding from SHAP analysis is that mineralogy, specifically the presence of calcite, is the most influential factor governing the contact angle. In addition, N2 as a cushion gas has the highest impact on contact angle, followed by CH4 and CO2. The novel CatBoost model significantly reduces the time needed to accurately predict contact angles in H2/brine/mineral/cushion gas systems compared to tedious laboratory experiments.

CatBoost , Contact angle , Cushion gas , Machine learning , Underground H2 storage

Text of the article Перейти на текст статьи

Department of Chemistry, Nazarbayev University, Astana, 010000, Kazakhstan
School of Mining and Geosciences, Nazarbayev University, Astana, 010000, Kazakhstan

Department of Chemistry
School of Mining and Geosciences

10 лет помогаем публиковать статьи Международный издатель

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