Embankment dam stability analysis using FEM

Abstract

This paper presents stability analysis of embankment dam with the surrounding heterogeneous rock mass using finite element method. In order to perform stability analysis of the dam and surrounding rock mass, several elastic-plastic material models for soil are customized and implemented in program package PAK. A 3D FE model of the embankment dam and the surrounding rock mass containing various material distributions according to their real distribution was made. The model includes a wider area around the dam in order to minimize the influence of the boundary conditions. The initial material parameters were determined using the identification of material parameters on the basis of the material from the dam body. Analysed dam is equipped with dam crest displacement transducers, as well as with transducers for the total and pore pressure in the clay core. Certain deviations have been noticed while comparing the measured values of these quantities with the results of the simulation. Since the analysed dam has been in operation for a long time, these deviations from the initial values of the parameters are caused due to the changes of the mechanical properties of materials during the dam operation. These changes are caused by several factors: The settlement of the dam and foundation, flushing of the material in the body of the dam and foundation, load changes, etc. In order to take into account the change of mechanical properties in materials and achieve the results of numerical simulation to describe the behaviour of the dam as close to the real behaviour, the calibration of the material parameters is carried out. Calibration of material parameters was performed using the measured displacements of the dam crest, as well as pore and total pressures in the clay core. Using the calibration we obtain new material parameters which give results of the numerical simulation that are closer to the behaviour of the real dam. In this manner we manage the dam safety and we can predict its future behaviour.