
Scientific Reports, Год журнала: 2025, Номер 15(1)
Опубликована: Апрель 21, 2025
Hydrogen storage is a crucial technology for ensuring sustainable energy transition. Underground Storage (UHS) in depleted hydrocarbon reservoirs, aquifers, and salt caverns provides viable large-scale solution. However, hydrogen dispersion cushion gases such as nitrogen (N2), methane (CH4), carbon dioxide (CO2) lead to contamination, reduced purity, increased purification costs. Existing experimental numerical methods predicting coefficients (KL) are often limited by high costs, lengthy processing times, insufficient accuracy dynamic reservoir conditions. This study addresses these challenges integrating data with advanced machine learning (ML) techniques model dispersion. Various ML models-including Random Forest (RF), Least Squares Boosting (LSBoost), Bayesian Regression, Linear Regression (LR), Artificial Neural Networks (ANNs), Support Vector Machines (SVMs)-were employed quantify KL function of pressure (P) displacement velocity (Um). Among methods, RF outperformed the others, achieving an R2 0.9965 test 0.9999 training data, RMSE values 0.023 0.001, respectively. The findings highlight potential ML-driven approaches optimizing UHS operations enhancing predictive accuracy, reducing computational mitigating contamination risks.
Язык: Английский