Statistical Assessment of Factor of Safety for Pile-Reinforced Slopes

Abstract

Statistical design of experiments was combined with coupled and uncoupled pile-reinforced slope stability analyses to draw conclusions about the factor of safety (FS) sensitivity to pile-related properties (pile location, spacing, embedment depth, and diameter), and to analyze differences between analytical techniques and space dimensionality. This evaluation performed two- and three-dimensional, limit equilibrium (LE), and finite-element (FE) analyses in a fully randomized block design, a definitive screening design, and a central composite design fashion. Pile location, embedment depth, and spacing had significant effects on FS. The optimum pile location is not fixed, and varies not only with changes in soil strength but also with variations in geometry properties because of the presence of interactions among the factors. The comparative analysis found no statistically significant difference between the studied methods. However, the confidence intervals indicated FS differences as high as 0.27, a variation of crucial importance for practical decision-making when FS is around the stability/failure threshold of 1. Recommendations for analysis and design of pile-reinforced slopes were given.