Coupled Hydro-Geomechanical modelling of dike breaching

Dike breaching following overtopping event is considered as one of the most common failure mechanisms.  Understanding this process is critical, as breaches typically result in catastrophic flooding. While overtopping failures have been studied both experimentally and numerically, the coupled physical mechanisms remain complex. Erosion associated with high-velocity water flowing downstream has often been considered as the main leading cause of failure. Yet, suction pressure and water content fluctuations provide additional strength to the dike material. The effects of suction on the geomechanical strength of the dike material have often been disregarded.  In this work, we propose a proof-of-concept of a numerical model that encompasses what we consider as the main physical processes occurring during dike overtopping. First, we solve, in a traditional hydraulics approach, the Shallow-Water-Exner equations system to evaluate the water flow and the erosion potential. Second, we solve the Richards equation, for groundwater flow evaluation. This provides the information on the suction pressure evolution in the dike, spatially and in time, subject to overtopping.  Third, we propose a geomechanical approach that accounts for suction pressure effects on the mechanical strength of the soil. Large displacements of the geomaterial are computed by means of the Particle Finite Element Method (PFEM). It is a Lagrangian based method, that relies on a very efficient remeshing algorithm to simulate large displacements.  The resulting model is a proof-of-concept for advanced dike failure simulation. We compare the outcome of the model in a dike failure theoretical case with a purely hydraulic based model and with a sediment transport-based model. The analysis focuses on the differences between these models, as reflected in the output hydrographs. The aim is to underline the need for tightened coupling between hydrodynamic, sediment transport and geomechanical processes to accurately simulate dike breaching events and improve hydrograph prediction.