Reliability Applied to Slope Stability Analysis

TLDR

Probabilistic slope‑stability methods have had limited practical uptake because engineers are unfamiliar with reliability concepts, there is confusion over reliability versus probability of failure, and absolute failure probabilities are difficult to estimate, making relative failure probabilities the most useful application. The study aims to show how field and laboratory data can be used to generate probabilistic soil‑parameter descriptions for slope‑stability analysis. The authors employ a first‑order, second‑moment method to model soil‑parameter uncertainties and apply it to the design of embankment dams. The example demonstrates how uncertainties in different parameters affect embankment reliability and shows that reliability analysis can inform consistent safety‑factor design values for various failure modes.

Abstract

Formally probabilistic methods for the analysis of slope stability have had relatively little impact on practice. Many engineers are not familiar with probabilistic concepts, and it has been difficult to incorporate concepts of reliability into practice. Also, there is confusion over what reliability and probability of failure mean. The most effective applications of probabilistic methods are those involving relative probabilities of failure or illuminating the effects of uncertainties in the parameters. Attempts to determine the absolute probability of failure are much less successful. The paper describes how probabilistic descriptions of soil parameters can be derived from field and laboratory data and applied in stability analysis. The first‐order, second‐moment approach is explored and applied to the design of embankment dams. The example illustrates the relative contributions of uncertainties about different parameters to the reliability of the embankment. Reliability analysis is especially useful in establishing design values of factors of safety representing consistent risks for different types of failure.