Document Type : Original Article

Authors

• Mohammad Reza Motahari ¹
• Pedram Asgari ²

¹ Associate Professor Department of Civil Engineering Faculty of Engineering, Arak University, Arak,, Iran

² Master's student Department of Civil Engineering Faculty of Engineering,, Arak University, Arak,, Iran.

Abstract

The failure behavior of rocks under true triaxial stress conditions is significantly influenced by the intermediate principal stress (σ₂). However, in many traditional rock failure criteria, the effect of this stress component is either neglected or treated in a simplified manner. The main objective of this study is to investigate the role of the intermediate principal stress in rock failure strength and to evaluate the capability of deep learning–based approaches for predicting the major principal stress at failure (σ₁) under true triaxial stress conditions. A comprehensive database comprising approximately 500 true triaxial experimental results obtained from five different rock types with distinct mechanical properties was employed. In the proposed model, the intermediate (σ₂) and minor (σ₃) principal stresses, together with the rock type, were considered as input variables, while the major principal stress at failure (σ₁) was defined as the model output. The experimental results indicate that an increase in the intermediate principal stress has a significant strengthening effect on rock failure resistance. Moreover, the deep learning model demonstrates high predictive accuracy and a strong capability in capturing the nonlinear relationships governing rock failure behavior under true triaxial stress conditions. These findings suggest that the proposed model can be regarded as a reliable tool for predicting rock failure strength in geotechnical and mining engineering applications.

Keywords

• Intermediate Stress
• True Triaxial Test
• Rock Failure Strength
• Deep Learning
• Artificial Neural Network

Main Subjects

• Geotechnics