Limitations of Sa(T1) as an intensity measure when assessing non-ductile infilled RC frame structures

Abstract

An important decision in seismic risk assessment involves a good choice of intensity measure (IM) used to describe ground motion intensity. Depending on the structural typology and type of analysis, the choice of an optimal IM may differ. Ideally, it should have accurate response predictions and not be biased by parameters like scaling factors or other ground motion characteristics. This article investigates this topic in the context of non-ductile reinforced concrete (RC) frames with masonry infills. It examines several case study structures with IMs like peak ground acceleration (PGA), peak ground velocity (PGV), first mode spectral acceleration, Sa(T1), and average spectral acceleration, AvgSa. It reveals PGA and PGV to possess undesirable traits relating to inefficiency, sensitivity to scaling and dependence on other ground motion parameters. Sa(T1) is shown to be an optimal IM at lower intensities, but runs into difficulty once the infill panels collapse and a non-ductile mechanism forms in the structure. Further investigation shows that the response is strongly biased by the velocity-based characteristics of the record upon infill panel collapse when using Sa(T1). On the other hand, AvgSa is seen to possess many of the same benefits regarding efficient response prediction for both intensity-based assessment and fragility function development but is not biased by the ground motion velocity-based characteristics. In light of this, this study presents a case for the traditional IM Sa(T1) to be set aside and for a more promising alternative in AvgSa to be adopted when analysing existing infilled RC frames.