On the need to improve the seismic risk analysis for nuclear plants safety in France preliminary lessons and recommandations from the research sinaps project.

International audienceSINAPS (Earthquake and Nuclear Plants Improving and Sustaining Safety) is a five-year research project starting in 2013 and financially supported by the French government. This project was motivated by the need of characterizing the potential seismic margins of existing nuclear facilities. SINAPS conducts critical analysis at each step of the seismic risk assessment, from seismic hazard, to site effects, interactions between the wave field the soil and structures, structural and nuclear components vulnerabilities current French practices are challenged against site data and innovative methods more adapted to model complex nonlinear processes that occurred during the seismic loadings transfer. The seismic margins are appreciated through all assumptions and the epistemic/aleatory uncertainties treatment during the analysis. The present contribution expose some lessons learned from SINAPS in the area of the seismic hazard assessment. Among them, we present recent published strong motions databases and associated ground motion prediction equations that should be considered as references for SHA in France; SINAPS also investigated Bayesian approach and clearly showed their betterment in the GMPE's selection and weighting process with respect to the classical and less objective expert's advices practice. We also discuss recent progress in the site-specific SHA as the host-to-target and kappa corrections