|Evaluation of alternative damping approaches for nonlinear time history analysis, and their influence on the development of fragility curves used in seismic risk, for low-to-moderate seismicity regions.|
Jose Baena Urrea
One of the major sources of uncertainty in dynamic non-linear time history analysis (THA) of structures is modelling of damping as this parameter is not easy to measure and continues largely misunderstood. Damping simulates energy dissipation on structural and nonstructural elements that is not explicitly modelled. Although it is known to be frequency-independent and amplitude-dependent, these features are not typically considered by practicing engineers. The classical and most used model is Rayleigh damping, developed in the XIX century as a mathematic convenient representation of friction effects in acoustic waves transmission.
During the last decades, engineers have accepted this model in linear THA because the damping forces are relatively small compared with the other resisting forces in the structures and due to efficiency, simplicity and low computational cost. Contrary to this, when nonlinear THA are performed, several numerical pathologies appear such as spurious damping forces in the massless degrees of freedom and other issues resulting from stiffness decay during the analysis. Over the years researchers have developed damping models to overcome these challenges, but the scientific community still did not reach a generalised agreement. The most suitable model will depend, to a certain degree, on the specific structural problem. Unfortunately, one major concern with damping modelling is that the structural response can change drastically depending on the model used.
Nonlinear THA are commonly used to define fragility functions, which are statistical curves of performance relating demands (accelerations and displacements, typically) versus probability of exceedance of reaching a specific level of damage. They are used in the context of seismic risk analyses and will depend on the damping model employed.
This research focus on evaluating current models and proposing an alternative approach better matching experimental evidence of the response at the element level. Subsequently, the influence of these damping models on the development of fragility curves will be assessed. Finally, their impact on the final outcome of seismic risk assessment for low-to-moderate seismicity regions, such as Belgium, will be computed and interpreted.