System reliability analysis and risk assessment of a layered slope in spatially variable soils considering stratigraphic boundary uncertainty

Due to complex geological deposition, post-deposition processes and limited site investigation data in engineering practice, soil properties are often spatially variable and the stratigraphic boundary in layered soils is generally characterized by uncertainty. The spatial variability of soil properties has been well investigated in past years, whereas the question of how the stratigraphic boundary uncertainty affects the slope stability remains unanswered. This paper attempts to investigate the effects of the stratigraphic boundary uncertainty on the system reliability and risk of a layered slope in spatially variable soils. In this paper, the spatial variability of soil properties is simulated by non-stationary random fields that are generated by an extended Cholesky decomposition technique, while the stochastic nature of the stratigraphic boundary location is simulated by a discrete random variable. Monte Carlo simulation (MCS) is suggested for evaluating the system failure probability and risk. Various comparisons between probabilistic analysis results obtained from considering and neglecting the stratigraphic boundary uncertainty have been made for different statistics of soil properties. Results show that the stratigraphic boundary uncertainty plays an important role in identifying the slope failure mechanism. Moreover, neglecting the stratigraphic boundary uncertainty would generally overestimate the slope failure risk for different statistics, except at small coefficients of variation of the friction angle, where the results are underestimated.