Fragility Assessment of Steel Frames under Post-Earthquake Hazards

Abstract

Post-earthquake hazards such as aftershocks and fire pose major threats to structural and life safety. Aftershocks often follow the mainshock within hours or days, increasing the risk of collapse. Similarly, post-earthquake fire (PEF) has historically caused additional losses and fatalities, sometimes exceeding the direct damage from the initial earthquake itself. However, most reliability-based approaches focus only on single-hazard evaluations. Therefore, this research explores a systematic fragility-based framework for assessing the performance of steel frames under post-earthquake hazards, including mainshock-aftershock (MS-AS) sequences and PEF. A fragility framework for corroded steel frames under MS-AS sequences was developed with several important considerations for probabilistic analysis. For example, as-recorded ground motion records were selected to capture realistic seismic input. Uncertain parameters affecting the lateral stiffness of the frame were included in the analysis. Spectral acceleration at the fundamental period was chosen as the representative intensity measure for mainshock and aftershock. Multiple Stripe Analysis (MSA) was used for time-history analysis due to its efficiency and ability to represent a wide range of intensities. Latin Hypercube Sampling (LHS) was adopted in the framework to generate unbiased samples of input random variables. Finally, three-dimensional fragility surfaces were developed, providing a more comprehensive assessment of structural vulnerability than two-dimensional fragility curves. Fire was also considered as a post-earthquake hazard. The concentrated plasticity model is an efficient approach to finite element modeling for seismic loading in OpenSees. However, the fire module relies exclusively on a distributed plasticity model to represent the temperature gradient across the section depth. Thus, the temperature-dependent distributed plasticity model has been used for PEF, although it could not capture degradation induced by earthquake loading. To overcome this shortcoming, low-cycle fatigue, which is available in OpenSees, was incorporated into “OpenSees for Fire” through modifications to the C++ source code. Using the modified version of “OpenSees for Fire”, the earthquake and fire analyses were performed sequentially, and the earthquake-induced damage was captured and carried over into the PEF stage. An efficient hybrid model was proposed, combining the modified distributed plasticity model for PEF scenarios with the concentrated plasticity model for floors subjected only to seismic loading. The hybrid model accurately predicted the structural response while offering improved computational efficiency. Following the enhancements made to the PEF analysis, a fragility framework was developed to evaluate the performance of steel frames subjected to PEF. The steps included generating samples, selecting and scaling earthquake records, defining fire curve parameters, selecting intensity measures, performing PEF analysis, checking performance criteria, and developing fragility contours. The samples were generated using LHS for important random variables in structural modeling. The earthquakes were selected and scaled within a defined spectral range to reflect higher-mode effects. Parametric fire curves were generated using LHS, with the opening factor, fire growth rate, and thermal inertia treated as random variables. Fire load density and average spectral acceleration were chosen as the intensity measures. MSA and Fire Stripe Analysis (FSA) were selected as efficient approaches for time history analysis. The framework was applied to a set of archetype frames (4-, 12-, and 20-story) evaluated at the collapse prevention performance level. Several engineering demand parameters were defined to account for various contributors to structural response. Fragility contours were developed for three earthquake hazard levels. A comprehensive comparison among the three archetypes revealed important findings in vulnerability and performance of steel moment-resisting frames under PEF. Together, these contributions establish a systematic framework for evaluating the performance of steel moment-resisting frames under post-earthquake hazards, including MS–AS and PEF. This framework can be extended to MS–PEF–AS, where the subsequent effects of seismic damage, fire-induced degradation, and aftershock may cause immediate collapse.