Reliability analysis and risk assessment of pile-reinforced slopes considering spatial soil variability and site investigation

Site investigation data is of great significance for practical slope stability analyses and designs. Previous studies on the probabilistic analysis of pile-reinforced slopes, however, rarely considered the influence of site-specific information, which may lead to a biased judgment of the stability of the soil-pile system, thus resulting in unreasonable designs. This paper proposes a new integrated probabilistic framework to analyse the probabilistic characteristics and quantitative risk of the pile-reinforced slopes considering site investigation data. Particularly, with the aid of the framework, the responses of the stabilising piles under different schemes of investigation boreholes are systematically studied. The Random Finite Difference Method within the framework of Monte Carlo Simulation (MCS) and Subset simulation is employed to explore the uncertainty within the pile-reinforced slope and the associated pile responses. A soil slope is taken as an illustrative example to validate the effectiveness of the proposed framework through various parametric studies on the investigation scheme of boreholes (e.g., numbers, locations, and sampling intervals). It is found that the stability results obtained from the deterministic analysis with mean soil properties and the conventional unconditional random field MCS may be biased compared to the results by the proposed framework considering real-site situations.