| Abstract: | Assessment of the seismic performance of existing structures requires due consideration of both aleatory and epistemic sources of uncertainty; the former being typically associated with the randomness in ground motion records and the latter with the uncertainty in numerical modelling. Using a numerical modelling approach calibrated to available experimental test data collected from the literature, the uncertainty associated with different modelling parameters for existing reinforced concrete frames in Italy was quantified via an extensive numerical study. This was done to quantify the propagation of modelling parameter type uncertainty to the overall dispersion of the demand parameters typically used in seismic assessment, namely peak storey drift and peak floor accelerations. In addition, the impact of such modelling uncertainty on the median intensity and dispersion of the collapse fragility function was also examined. From the results of this study, empirical values of modelling parameter uncertainty were quantified with a view to being used in the assessment of existing reinforced concrete frames with masonry infill designed prior to the introduction of seismic design provisions in Italy during the 1970s. Comparing these empirical values to those available in the literature, it is seen how the fundamental behaviour of the frames differs from more modern frames with ductile detailing to the extent that values available in guidelines such as FEMA P58 cannot be reasonably adopted for these structural typologies. |
