Abstract
In rock masses, the presence of numerous randomly distributed joints introduces uncertainty, making the prediction of failure paths challenging. Among these, key joints significantly influence rock mass fracturing. This study proposes a peridynamics (PD) method based on Monte Carlo simulation analysis, discussing the impact of joints with different dips in complex joint networks on rock mass failure probabilities. Efficient parallel computing programs have been developed, markedly enhancing the computational efficiency of large-scale Monte Carlo simulations for PD analysis. The concept of material point failure probability (PFP) is presented, investigating the variation of PFP contour maps after excluding specific joint dips. Grid-based PFP contour maps and Grid-based JAIC (Joint Angle Impact Coefficient) contour maps are created, enabling a quantitative assessment of rock mass failure probabilities. The study reveals the influence of joint dip angles on the failure probabilities of rock masses with complex joint networks. Additionally, the concept of key and non-key joint dip angles based on the grid is introduced. Statistical methods for identifying key and non-key joints in rock mass grid regions are established, providing new perspectives and tools for understanding and predicting the failure of rock masses with complex joint networks. This research contributes to the reliability study of rock mechanics and provides new theoretical guidance for geotechnical engineering.