Hydrogen is a promising alternative energy source due to its high energy density and clean characteristics. However, its safe and efficient storage remains a challenge. The wide range of potential storage materials presents significant time, energy, and cost barriers for experimental screening. This study investigates the potential of machine learning based data-driven models in predicting hydrogen storage capacity in solid-state materials, focusing on metal hydrides, complex hydrides, metal-organic frameworks (MOFs), and carbon materials. A dataset comprising 239 sourced materials was compiled and analysed using three machine learning models: Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Support Vector Regression (SVR). The RF model demonstrated the best performance with an R² score of 0.8657, Mean Absolute Error (MAE) of 0.7102, and Root Mean Squared Error (RMSE) of 1.0187. Sensitivity and Shapley Analysis indicated that material type and adsorption pressure significantly impact hydrogen uptake, with complex hydrides exhibiting the highest storage capacity (up to 18.5 wt%). The findings confirm that artificial intelligence (AI)-driven modelling accelerates material selection and optimization for hydrogen storage applications, reducing the reliance on costly and time-consuming experimental methods.
Moses et al. (Thu,) studied this question.