The seismic performance of a earth dam by different displacementbased methods

The performance-based design of earth dams and the rehabilitation of existing ones require the evaluation of seismic performance based on permanent displacements caused by expected the earthquake. The paper reports a comparison between different methods with increasing complexity for estimating seismic displacements: simplified rigid block method, based on empirical relationships (Bray and Rathje, 1998; Tropeano et al., 2009); simplified uncoupled method, again based on the sliding block analysis, but accounting for soil deformability; coupled ‘stick-slip’ approach, based on a 1D lumped mass model to calculate together dynamic response of the site and movement of sliding block (Tropeano et al., 2011); 2D finite differences analyses by the FLAC code, reproducing the heterogeneity of soil and topographic effects. The methods were applied to the case of the dam of Marello mountain across the Angitola river (Southern Italy). The parameters for static and dynamic geotechnical characterization of subsoil model have been taken from the results of the site investigation published in technical reports. The spectral shape and peak ground acceleration specified by the Italian Seismic Hazard Map, representative of input motion on outcropping bedrock, allowed to choose a set of spectrum-compatible acceleration time histories to simulate the seismic input. The sliding displacements predicted using simplified method resulted strongly dependent on topographic coefficient. Both uncoupled and coupled approaches have shown conservative permanent displacements compared to Newmark method. The average displacement of the sliding block by two-dimensional finite difference analysis, considering the stiffness variability related to depth, results comparable with values obtained by other methods.