Evaluation of Sampling Efficiency and Uncertainty in 3D Spatially Variable Slope Stability Assessment Using Conditional Simulation

Soil properties spatially vary due to the complex geological process in its formation. Recent studies have revealed that soil spatial variability largely influences geo-structure performance, i.e., slope stability. To facilitate the reliable slope design, revealing the spatially varied soil properties in-situ is necessary. However, they may require a large number of geotechnical investigations, i.e., the cone penetration test (CPT). This paper firstly explores the optimal sampling location of a 3D slope by considerng limited pseudo-CPT data. After the identification of the 3D slope failure mechanism, the optimal sampling location for each kind of failure mechanism is investigated by the calculated Euclidean distance using a designed simulation flow. Finally, the uncertainty in the estimation of the failure mechanism is discussed. In the analysis, soil properties (c, tan ϕ, and γ) are treated as random variables, and the limit equilibrium method is used to evaluate the slope stability. The conditional/unconditional simulation is used in tandem with a Monte Carlo simulation framework to evaluate the sampling efficiency together with the failure mechanism of the 3D slope.