Nonlinear coupled hydro-mechanical dynamic finite element analysis of the seismic response of earth dams

It is known that saturated soil exhibits a different mechanical response than dry soil, especially during transient dynamic loading, such as earthquakes. Relative motion of the soil skeleton and the confined pore water may result in excess pore water pressure development and some additional damping that may affect significantly the seismic acceleration amplifications. This is pronounced in the case of geotechnical structures that can vibrate significantly, in their fundamental mode in particular, such as high earth dams.
More practically, one of the critical considerations when analysing or designing an earth dam is whether a coupled hydro-mechanical analysis that considers pore water-soil skeleton interaction is necessary. Such coupled dynamic formulations exist and are being investigated by a number of researchers but they are not usually accessible by practising engineers, as they have not been implemented in many commercial engineering analysis software. Consequently, in many cases, practising engineers perform only uncoupled mechanical dynamic analyses, thus ignoring the dynamic interaction between the pore fluid the soil skeleton.
It is therefore, important to examine the effects of coupled and uncoupled dynamic finite element analyses that are usually used in the seismic analysis of earth dams. The aim is to understand the influence of pore water-soil skeleton interaction on the seismic response of earth dams and investigate the cases when a simple uncoupled analysis can be considered as appropriate.
This paper considers a well-documented earth dam case study for which relevant material properties, geometry and deformation data are available, along with the associated earthquake acceleration motions. A nonlinear finite element model is developed, and two cases are considered: (a) uncoupled mechanical and (b) coupled hydro-mechanical formulations. The predicted seismic response of the dam is compared to the available field records as a means of model validation. Subsequently, the two modelling cases are compared revealing the beneficial effects of the coupled formulation. Finally, through a series of parametric studies, the range of applicability of the uncoupled formulation is established.