Ultimate limit state assessment of dyke reliability using the random material point method

This paper investigates embankment reliability based on the ultimate limit state (ULS). The ULS is generally not clearly defined and, especially for flood defences, the ULS is currently under discussion. According to Dutch law [1], flooding which leads to either casualties or substantial financial damage is considered as the ultimate limit state of a flood defence structure. However, initial slope instability is regarded as flood defence failure according to guidelines for the assessment of macro-instability of dykes [2]. Allowing initial failure, but preventing a dyke breach, is not prohibited by the current Dutch regulations and can lead to more efficient design. Analysis of both large deformations as well as the influence of spatial variability of soil properties is important to assess the reliability of a dyke against breaching. This paper uses a new technique called the random material point method (RMPM) [3], which combines MPM [4] for modelling large deformations, with random fields [5] for modelling soil variability, in a Monte Carlo simulation.