The impact of modelling uncertainties on the seismic performance assessment of reinforced concrete frame buildings

Simplified procedures have been used for the estimation of seismic response parameters by considering the epistemic uncertainties for an older reinforced concrete frame, and for two contemporary reinforced concrete structures. The simplifications in the procedure are associated with a simplified nonlinear method and models for the assessment of the seismic performance of the structure, whereas the effects of the epistemic uncertainty are treated by using the first-order-second-moment (FOSM) method, and the latin hypercube sampling (LHS) technique. The results of sensitivity analysis reveal that it is those parameters which affect the collapse mechanism and have a high coefficient of variation that have the greatest impact on the seismic response parameters for the near collapse limit state. The results of uncertainty analysis by using the LHS technique showed that epistemic uncertainties have an effect on the dispersion, and also on the median estimates of the response parameters. For all three example structures a reduction in the estimate for the median peak ground acceleration at the near-collapse limit state was observed. Thus, explicit consideration of epistemic uncertainties in the process of the assessment of structural performance can lead to more accurate results, and consequently also to more reliable assessment of seismic risk.