Analysis of post-elastic damage measures in Nonlinear Dynamic Analyses of RC structures

The use of Nonlinear Dynamic Analyses, although commonly accepted as the most suitable method for determining the seismic demand on structures with non-linear behavior, provides a significant increase of the uncertainties on the structural demand. These uncertainties generally vary depending on the damage measures considered for the evaluation of the seismic demand. In order to evaluate the variability of the seismic demand, this study considers two damage indicators, including a local (the maximum section curvature) and a global measure (the maximum inter-story drift ratio). This analysis was performed by subjecting a reinforced concrete structure to different groups of spectrum-compatible real records (according to the Eurocode 8 provisions). The comparison of the different Engineering Demand Parameters permits to estimate the damage measure most representative of the seismic demand obtained from a Nonlinear Dynamic Analysis. The current seismic codes prescribe that if the response is obtained from at least 7 nonlinear time-history analyses the average of the response quantities should be used. However, when a structure is subjected to seismic records, the response quantity is characterized by a high variability and the average structural response could underestimate the seismic demand. For this reason a measure of the structural demand depending on its variability based on the statistical concept of upper tolerance limit Lu is proposed and compared with the arithmetical and geometrical mean of response quantities. The tolerance limit Lu is determined depending on two tabulated parameters obtained in function of the coefficients of variation of the structural response. The proposed method allows to consider a higher value of the response quantity when the output variability is high and the number of records is low.