Bureau d'accueil : Bureau d'évaluation des risques sismiques pour la sûreté des installations (BERSSIN)
Date de début de thèse : octobre 2017
Nom du doctorant : Arnaud MONTABERT
Descriptif du sujet
The damage on buildings is often the only available datum on historical earthquake. Understanding the properties of seismic shaking on historical structure is a useful tool to improve knowledge of past seismicity. However, this is not a trivial task and its development requires a multidisciplinary approach including mechanical vibration, seismology, and archaeology. This innovative interaction draws the core of my PhD project.
On one side, structural models have to be both realistic and generic to represent an entire building typology. On the other side, the seismic motion has to be carefully selected and understood in particular looking at the 3 components (2 horizontal and 1 vertical) of the ground shaking. Indeed, the vertical component is often neglected in structural analysis. Thus there are no clear statements quantifying its impact on structural behavior, neither its peculiarities with respect to the horizontal components.
During these firsts months I am implementing a strategy aimed to describe the role of the vertical component on structural behavior. We know that the three components of the ground motion and their features depend on the rupture mechanism, on the wave propagation in the crustal and shallow layers and on the azimuth between the source and the station. The question is whether and how these characteristics are related with the structural behavior. Real data are affected by a huge complexity ; for this reason I am currently working with synthetic 3D waveforms, based on realistic rupture model, 1D layered crustal medium, and a full azimuth coverage.
The data are injected in a linear structural model based on finite element method analysis that I am currently refining: the impact of the ground motion on the results depends on the degree of complexity of structure model, but increasing the complexity also require longer computation times. Thus the lesson learned on this preliminary test will be applied to:
- optimize the strategy to analyze real data (what kind of signal treatment, rupture source, site-source geometry ?).
- set the sufficient level of structural complexity.
The final aim is to compare the numerical structural analysis with the observed damage, for this reason several reconnaissance missions on historical / archaeological sites are scheduled on 2018 to choose the case of study.