A Comparison of Numerical and Analytical Methods in Dynamic Analysis of Embankment Dams (Case Study Balarud Embankment dam)

Due to accuray of numerical methods and constitutive equations, a comparison of these methods with the classical methods of dynamic analysis of earthfill dams seems necessary. The results of Seed-Makdisi, Sarma and Ambrasys-Sarma analysis of a homogenous dam are compared with those of Quake/W and Slope/W softwares. A code is developed for calculating permanent displacements based on Newmark double integration method. The results of linear and equivalent linear models show that, due to unrealistic linear assumption of the behaviour of the material, the calculated accelerations are high and the analytical methods are overdesign. This is due to realistic constitutive models. Also, in homogenous dam, as the assumed critical wedge in analytical methods is different from that calculated in numerical analyses, the critical acceleration estimated is different too (0.45g in Ambrasys-Sarma method compared to 0.26g in equivalent linear model for a wedge of 0.6h depth). This shows that realistic models must be utilized. In the developed code first the average aceleration of assumed wedge is calculated using acceleration calculated with finite element codes utilizing modern constitutive models and consequently the permanent displacements are calculated. The results show that the current methods are overestimating. Also, the assumed critical wedges are unrealestic and pass through the dam core. While the critical wedges pass through downstream shell. The permanent displacement of a wedge with 0.6h depth is estimated to be 3 cm compared to zero in Sarma method. The developed code was used for dynamic analysis of Balarud earthfill dams.