The seismic wave propagation model SAM-4D (Seismic Advanced Model for Dams), a product of the previous three-year System Research activity, was developed to overcome the main limitations (in particular the excessive conservativeness of the results) highlighted by traditional and simplified methods of seismic analysis of dam-basin-foundation systems.

The model makes it possible to ideally reproduce the same behavior that would be obtained in a semi-infinite foundation (as it is in reality) by representing the propagation of seismic waves in a computational domain bounded by appropriate artificial contours (absorbing or non-reflecting contours) and realistically provided with mass.

The model, developed in last three years for use in the Abaqus finite-element code, has been implemented in the proprietary finite-element code CANT-SD (Code for Three-Dimensional Nonlinear Analysis – Static and Dynamic). This code, developed specifically for the analysis of dam-basin systems, is very reliable in modeling structural discontinuities of concrete dams, characterized by a remarkable practicality of the result extraction process, and is free of the need to acquire commercial licenses.

For the implementation of SAM-4D in the CANT-SD code, it was necessary to develop special elements, characterized by orthotropic (and massless) linear material, to be used as absorbing contours. In addition, a criterion was defined, implemented, and validated to enable the correct application of equivalent seismic forces (normal and tangent surface forces to be applied to the base and lateral contours of the foundation).

The preprocessing computer tools created for the Abaqus code with the purpose of automating the calculation and application of equivalent seismic forces were suitably modified in order to be able to use them as preprocessors for the CANT-SD code as well.
The proper functioning of the SAM-4D model in the CANT-SD code was verified by means of two case studies previously used for the validation of the model itself in the Abaqus code.

Finally, the case study of the Monticello arch dam was reused to further investigate the use of the deconvolution technique for defining seismic action in the context of problems for which experimental records are available.