Abstract
The stability design of soil structures, such as retaining walls, has traditionally relied on deterministic analyses that use averaged soil parameters. However, modern computational advances, particularly Monte Carlo simulations (MCs), have highlighted the limitations of this approach. Engineers are now increasingly focused on understanding the probability of failure (PF) associated with specific safety factors. This study explores the impact of spatial variability in soil properties, specifically the friction angle and unit weight, on PF. By applying log-normal distributions and spatial correlation lengths, extensive MC simulations were combined with finite element limit analysis to establish relationships between safety factors and PF. This research makes several contributions to geotechnical engineering. It integrates MC simulations with finite element limit analysis to provide a statistically robust framework for evaluating PF in soil structures. The use of adaptive finite element meshes introduces new insights into failure mechanisms, addressing gaps in existing studies. Additionally, the study presents a set of parametric design charts that enable practitioners to estimate PF for specific safety factors effectively. These findings provide practical tools for design engineers, supporting better decision-making and increasing confidence in design outcomes.