A significant proportion of European rail networks are built upon earthworks that are over one hundred years old. These earthworks are under increased pressure as they have to contend with heavier and more frequent traffic, far outside the scope of their design. To compound this problem further, recent years have seen unpredictable weather patterns develop with prolonged intense rainstorms commonplace. This has led to increased incidence of slope failures along rail networks, as many aged earthworks struggle to withstand such drastic changes in loading. Marginal engineered slopes fail depending on the triggering mechanism which presents itself first. Therefore the failure surface is intrinsically linked to the applied load i.e. surcharge loading will instigate a different type of landslide than prolonged rainfall. Therefore this paper proposes to analyse marginal slopes probabilistically as a system, where multiple slip circles are considered. A multi-modal optimisation algorithm LIPS (locally informed particle swarm optimisation) is used to locate all significant slip circles. In a slope with multiple potential failure surfaces the consequence of failure is not necessarily the same across the different slip surfaces. This paper addresses this deficit by examining the consequence of the different landslides should they occur. When combined with previously calculated probabilities of failure this will entail amount to a full geotechnical risk assessment of engineered slopes.

reliability analysis; risk assessment; slope stability; engineered slopes; multimodal