Abstract submitted to PSHANI (by I. Zentner, F. de Martin, C. Cornou, P. Gehl, P. Sochala,  A. Imtiaz, K. Meza-Fajardo, E. Chaljub, G. Senfaute, B. Guillier, M. Wathelet, E. Maufroy)

The project EXAMIN (EXperimental Assessment and Modelling of ground motion spatial variability for performance based seismic risk assessment of industrial plants and Infrastructures) has the aim to develop a comprehensive approach for risk assessment of infrastructure networks based on physics-based numerical simulation accounting for spatial variability at different scales. The project has obtained funding from the French National Research Agency and runs until 2021.

EXAMIN develops an integrated approach where the physics-based wave propagation simulations allow for the generation of spatio-temporel waveforms to be used in the structural fragility analysis as well as the development of meta-models linking the site and source parameters to the surface ground motion intensity measures. The meta-models allow for the generation of ground-motion maps (Pseudo Spectral Acceleration PSA) required in the risk assessment chain while accounting for the epistemic uncertainties that have been previously introduced in the simulations. The impact of various assumptions (source event, hazard propagation) or damage configurations on loss metrics can be analyzed in this framework. The outcome of this project will allow for a more accurate evaluation the vulnerability and resilience of urban systems and infrastructures exposed to spatial seismic hazard.

The project puts a particular focus on spatial variability at different (local and regional) scales. Indeed, local variability is generally not accounted for in current seismic risk and loss assessment although it is now recognized that the spatial variation of seismic ground motion affects the dynamic response of extended and multi-supported structures and more generally on infrastructure networks or industrial plants comprising several adjacent buildings. The spatial variability at local scale (< 100m) is assessed experimentally and represented in the numerical models by random fields.  

This contribution proposes to present the methodology to be developed in the framework of EXAMIN and to show first results obtained by the partners (EDF, BRGM, ISterre).