Coupling spectral and Finite Element methods for 3D physic-based seismic analysis from fault to structure: Application to the Cadarache site in France

• The domain reduction model is implemented and verified for SEM-FEM coupling. • 3D physics-based simulations for dynamic analysis from fault to nuclear structure. • Investigation on input motion for SSI studies, bedrock signal is best accurate. • Amplification in structure response due to complex wave field input and surface wave. • The SEM-FEM coupling allows a reduction up to 50% of the modeled domain in FEM. This paper presents the application of a coupling strategy between the Spectral (SEM) and Finite Element (FEM) Methods to solve the soil-structure interaction (SSI) problem. The SEM-FEM coupling benefits from the mesh refinement capabilities of the FEM in modeling the structure and the near soil, with the realism of the SEM for regional scale earthquake simulations from the fault to the site. To this end, the Domain Reduction Method (DRM) introduced by ( Bielak et al., 2003 ) is herein employed. The DRM formulation solves the problem of multi-scale earthquake simulations by subdividing the fault-to-structure problem into two simpler ones, namely: problem I- containing the structure and (hereafter solved using the FEM) and problem II- containing the fault, the regional geological and surface topography, the sedimentary basin (hereafter solved using the SEM). In this work, the coupling between the FEM code CAST3M and the SEM code SEM3D is presented, verified and compared to conventional one-dimensional (1D) deconvolution methods for SSI analysis. A seismic simulation is performed, considering a simplified virtual nuclear reactor building, artificially positioned on the Cadarache site in South-Eastern France. The applicability of the SEM3D-CAST3M coupling is demonstrated, for SSI analysis in linear elastic regime and with a 50 % reduction of the model dimensions. Moreover, the coupling captures the amplified response for the nuclear structure due to the 3D spatially variable input field and to the resulting surface wave propagation in the SEM-FEM SSI analysis.