Evaluation of a Recently Proposed Ground Motion Selection Method in Case of Vertically Irregular Frames

Rapid growth of performance-based earthquake engineering has caused increasing interest in Nonlinear Time History Analysis (NLTHA) as an effective tool for the estimation of dynamic structural demand and capacity. Considering the preparation of a set of ground motions as the key step for NLTHA; many ground motion selection and modification approaches have been proposed to ensure reliable analysis results by reducing possible bias due to the random selection of ground motions. Apart from the existing differences among these methods, there is a common aspect in almost all of them, which can be considered as a limitation: They are constructed on the basis of simplifying assumptions that are not necessarily valid for irregular or complex structural systems. This paper evaluates the efficiency of a recently proposed structure-specific record selection scheme in terms of collapse simulation of vertically irregular frames. Utilization of the method is assessed by case studies on which different strength, stiffness and combined irregularity patterns are applied. The influence of proposed reduction in the number of used records on the estimated collapse capacities is evaluated by statistical tools. The results confirm the ability of the method in estimating median collapse capacity with 82% reduction in computational cost and maximum observed error of 16%.