LAUSR

Abstract Seismic performance evaluations of offshore platforms are of socio-economic importance, but could be more reliable in scope if probabilistic methods are employed. Probabilistic seismic demand models (PSDMs) for fixed pile-founded offshore platforms in the Persian Gulf have been developed that use probabilistic seismic demand analysis (PSDA). PSDA represents seismicity through a selection of ground motions. PSDM is based on a representative relation between ground-motion intensity measures (IMs) and engineering demand parameters (EDPs). In the present study, optimal pairs were selected from a large combination of IM-EDP pairs that relate to the criteria of practicality, effectiveness, efficiency, and sufficiency by employing a lognormal distribution and regression analysis. To generalize the results of PSDMs for geometric parameter uncertainty, five brace-configuration models that consider soil-pile-structure interactions have been generated. While PSDMs are being traditionally conditioned on a single IM and EDP, it has been proved that the uncertainty level in the probabilistic models depends on the IM and EDP used. The present study evaluated optimal PSDMs from 27 IMs against a range of EDPs at the local, intermediate and global levels. The results demonstrate the superiority of PSDMs conditioned by velocity-related IMs for most EDPs. Optimal pairs were provided for global-level Housner intensity-driftglobal (HI-DRglobal), for intermediate-level HI-DRjacket, and for local-level HI-DRcellardeck. Conversely, Sa(T1,5%), which is a widely used IM for probabilistic assessment of fixed pile-founded offshore platforms, performed unfavorably when compared with velocity-related IMs for predicting demand parameters.