مشخصات پژوهش

Hydrogen, as a green gas, has recently garnered attention for its role in storage and its impact on wettability alteration and interfacial tension (IFT). In this study, the use of hydrogen as a cushion gas alongside carbon dioxide, methane, and nitrogen was investigated in an innovative approach within enhanced oil recovery systems. Its performance was compared with other gases using machine learning models. Laboratory data, after statistical preprocessing, were processed and validated to predict IFT and contact angle (CA) values. Random forest, random tree, and LSBoost models were implemented for this purpose, and hyperparameters were tuned using Bayesian optimization and random search methods. The LSBoost model demonstrated the best performance, with R2 values of 0.9986 for IFT and 0.9870 for CA. These results highlight the high accuracy and applicability of the proposed method.