Loss-oriented hazard-consistent incremental dynamic analysis

Incremental dynamic analysis (IDA) is typically conducted using a generic record set, avoiding the arduous site-specific record selection process required for multiple stripe analysis (MSA). Traditional IDA is, however, limited in its ability to produce hazard-consistent estimates of engineering demand parameters (HC-EDPs). To address this limitation, this study presents loss-oriented hazard-consistent IDA (LOHC-IDA), a novel statistical framework for seismic loss assessment using IDA. LOHC-IDA comprises of three main steps: (i) performing IDA using a generic set of ground motions; (ii) post-processing the IDA results to develop simplified surrogate models at different intensity levels to predict engineering demand parameters (EDPs) using secondary ground motion intensity measures, and estimating the correlations among EDPs; and (iii) using the simplified EDP surrogate models and correlation matrices to adjust the original EDPs and simulate HC-EDPs that are consistent with the site- and intensity-specific ground motion characteristics. Unlike other methods in the literature, LOHC-IDA explicitly accounts for correlations among EDPs and their variation across intensity levels, and uses peak ground acceleration as a predictor for peak floor acceleration. Previously proposed methods also do not benchmark their loss estimates against MSA. Application of LOHC-IDA to two buckling-restrained braced frame buildings demonstrates that the simulated structural response and loss distributions across different intensity levels, as well as the expected annual losses, closely match the benchmark MSA results, with a 6 % average absolute error in EAL.