Answers are grounded in this paper and cite their sources.
Modeling study demonstrates accurate prediction of bentonite hydraulic conductivity using XGBoost, indicating reliable barrier assessment for geological repositories with minimal inputs.
Hydraulic conductivity of bentonite-based materials is a key parameter for evaluating the long-term performance of engineered barrier systems in deep geological repositories. Predicting hydraulic conductivity remains challenging because it is affected by multiple interacting factors. This study developed a machine learning framework based on a database of 780 experimental records and evaluated four machine learning algorithms for hydraulic conductivity prediction. Model performance was assessed using five-fold cross-validation after hyperparameter optimization. Among the evaluated models, tuned XGBoost achieved the best overall performance, with a mean cross-validation R2 of 0.91 and an RMSE of 0.29. SHAP analysis identified bentonite content and dry density as the dominant controlling factors. External validation using independent experimental datasets demonstrated that the tuned models accurately reproduced hydraulic conductivity under varying dry density, permeant solution chemistry, and temperature conditions. The results indicate that reliable predictions can be achieved using a limited number of readily obtainable input variables.
No takes yet. Share an insight, caveat, or question.
Zhao et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: