The accuracy of CQC and Response Spectrum Analysis in the case of impulsive earthquakes

Comparison of structural outputs of Response SpectrumAnalyses (RSA) and linearTime History (TH) analyses are carried out for two existing RC frames. Four set of records, in which two of them were made of pulse-like records, were considered; aimed at checking the accuracy of CQC in response spectrum analyses. Results showed a statistically significant underestimation of RSA with respect to TH for maximum interstorey drift and drifts at upper storeys and for all record sets (up to ???20%). Pulse-like records with pulse periods (TP) lower than the fundamental period (T1) of the structures showed an increase of underestimation (up to ???40%). RSA should be carefully employed for both design and assessment when displacement or drifts have to be controlled. These trends are not only an issue in the case of pulse-like records with but in general.