Collapse Fragility Curves for RC Buildings Exhibiting Brittle Failure Modes

Defining the collapse probability of existing nonseismically designed RC buildings is a challenge mainly due to the fact that its determination relies on the simulation of the brittle failure modes controlling the collapse mechanism. Routine nonlinear dynamic analysis of ductile RC frames is not applicable for such buildings, which often fail prior to the attainment of their flexural capacity. An alternative approach to define collapse probability is proposed in this paper through the explicit definition of the failure mode hierarchy. A number of brittle failure modes are considered as credible and the corresponding capacity is calculated. The prevalent collapse mode is identified through capacity prioritization of the failure modes, and the corresponding peak ground acceleration at collapse is defined. This value, along with the associated variability, is substituted in a lognormal distribution to derive fragility curves for typical design and construction details. The main conclusion drawn from the processing of the results is that disregarding the brittle failure modes and accounting solely for flexural ductile behavior greatly underestimates the probability of collapse of such buildings.