Deep learning models to predict CO solubility in imidazolium-based ionic liquids.

Journal: Scientific reports
Published Date: Jul 21, 2025

Abstract

This study focuses on predicting CO solubility in imidazolium-based ionic liquids using deep learning models with input parameters of critical pressure, critical temperature, molecular weight, and acentric factor. The models evaluated include Bayesian Neural Networks (BNN), Deep Neural Networks (DNN), Gradient Boosting Neural Networks (GrowNet), Tabular Neural Networks (TabNet), Random Forest (RF), and Support Vector Regression (SVR). The results were compared with two PC-SAFT models, namely cQC-PC-SAFT-MSA (1) and cQC-PC-SAFT-MSA (2), where deep learning models performed better than SAFT models. Graphical and statistical analyses revealed that the GrowNet model, with a root mean square error of 0.0073 and a coefficient of determination of 0.9962, exhibited the lowest error compared to other models. In addition, Pearson correlation coefficient (PCC) and Shapley additive description (SHAP) analyses highlighted pressure (P) as a key parameter determining CO solubility in imidazolium-based ionic liquids, significantly contributing to model performance.

Authors

  • Amir Hossein Sheikhshoaei
    Petroleum and Petrochemical Engineering School, Hakim Sabzevari University, Sabzevar, Iran.
  • Ali Sanati
    Petroleum and Petrochemical Engineering School, Hakim Sabzevari University, Sabzevar, Iran. a.sanati@hsu.ac.ir.
  • Ali Khoshsima
    School of Petroleum and Chemical Engineering, Hakim Sabzevari University, Sabzevar, Iran. a.khoshsima@hsu.ac.ir.

Keywords

No keywords available for this article.

External Resources

Popular Topics
Recent Journals