Prediction of mixed-mode I/II fracture toughness of rock-concrete bi-material disc with interface crack: Interpretable NRBO-XGBoost-SHAP model and experimental validation

• An interpretable NRBO-XGBoost-SHAP model is developed to predict the effective fracture toughness ( K eff ) of R-C bi-material. • Data from laboratory CSTBD testing and published literature were used to train and test NRBO-XGBoost-SHAP model. • Seven machine learning models were compared, and the NRBO-XGBoost model performed best in predicting K eff . • To enhance the model’s interpretability, SHAP analysis was integrated with classical fracture mechanics principles. • The NRBO-XGBoost model’s reliability was further substantiated through experimental validation utilizing a new dataset.