This study explores the effectiveness of machine learning approaches for estimating mode-I fracture toughness (KIC) of rocks, a crucial factor in both geotechnical engineering and materials science. Six sophisticated machine learning algorithms were constructed and thoroughly assessed using a compilation of 500 experimentally acquired samples. Vital input variables such as grain size, porosity, density, water saturation, and rock classification were gathered from varied geological environments to encompass inherent diversity. The outcomes demonstrated that the Random Forest Regressor (RFR) delivered the highest average forecasting accuracy among the evaluated models. Nonetheless, the Friedman–Nemenyi statistical test showed that models like SVR and ANN exhibited similar performance levels, with no statistically significant differences within the critical difference interval. SHAP analysis additionally improved model transparency by pinpointing porosity and water saturation as the primary factors influencing fracture toughness estimates. The results also underscore the existence of complex nonlinear relationships among input features, demonstrating the capability of machine learning models to capture meaningful physical relationships in heterogeneous rock systems. By merging robust experimental datasets with cutting-edge modeling and interpretability methods, this research advances data-driven techniques in rock mechanics and offers a dependable framework for anticipating fracture behavior in intricate geological materials.
Alaa R. Al-Shamasneh
— Department of Computer Science, College of Computer & Information Sciences, Prince Sultan University, Rafha Street, Riyadh 11586, Saudi Arabia
Arsalan Mahmoodzadeh — Center of Research and Strategic Studies, Lebanese French University, Erbil, Iraq
Yahia Said — Center for Scientific Research and Entrepreneurship, Northern Border University, 73213, Arar, Saudi Arabia
Abdulaziz Alghamdi — Department of civil engineering, University of Tabuk, Tabuk, Saudi Arabia
Ibrahim Albaijan — Mechanical Engineering Department, College of Engineering at Al-Kharj, Prince Sattam Bin Abdulaziz University, Al Kharj 16273, Saudi Arabia
Amal Alshardan — Department of Information Systems, College of Computer and Information Sciences, Princess Nourah bint Abdulrahman University, P.O. Box 84428, Riyadh 11671, Saudi Arabia
Ahmed Babeker Elhag — Center of Engineering and Technology Innovation, King Khalid University, Saudi Arabia
Parisa Khoshvaght — Department of Biosciences, Saveetha School of Engineering, Saveetha Institute of Medical and Technical Sciences, Chennai, 602105, India; International Center for Materials Sciences and Technology, Western Caspian University, Baku, Azerbaijan
PDF Viewer
Preview is limited to the first 3 pages. Sign in to access the full PDF.
