Advancements in the Seismic Risk Assessment of Mid-rise Reinforced Concrete Buildings

Reinforced concrete (RC) buildings constitute a significant part of the existing stock in many areas exposed to high seismic risk, such as in urban centres in the Mediterranean region. Recent earthquakes in Italy confirmed the high vulnerability of RC buildings and highlighted the need for more reliable models to estimate the expected losses and quantify the possible risks. Such models are widely used in insurance industry and are based on the three main components of the risk equation, namely hazard-vulnerability-exposure. This Thesis investigates the hazard and fragility/vulnerability components of the risk equation and systematically assesses the effect of these on the economic losses estimated for populations of mid-rise RC buildings within a catastrophe risk framework. Regarding hazard, a variety of intensity measures (IM), including advanced and conventional IMs, are tested to determine the optimal IM to represent seismic hazard. Optimal IMs reduce uncertainty as they are better correlated with the buildings’ seismic response, and allow a smaller number of analyses to be performed. Regarding fragility and vulnerability, catastrophe models require the assessment of large populations of structures; however, simplifications are needed to this procedure due to computational limitations. To this aim, the simplified analysis approach FRACAS bridging capacity spectrum assessment and fragility curve generation is presented alongside with a new software interface. Additionally, a Bayesian emulator-based approach (BEA) to efficiently compute the probabilistic seismic response and develop fragility functions is also proposed. The BEA is based on fewer assumptions and is shown to achieve better uncertainty characterization compared to standard methods. Several combinations of the developed knowledge are applied to a probabilistic seismic risk assessment case-study to determine their influence on the resultant damage/loss prediction. It is shown that the choice of IM and fragility function have a significant impact on the losses, with the produced advanced IM and fragility models yielding the best results