Fragility curves play a critical role in regional seismic risk assessment and are a key component of tools used to support emergency response and preparedness in California following an earthquake. To have an accurate assessment of regional damage, it is critical to provide fragility curves that best represent the bridge inventory. However, it is impractical to develop unique fragility curves for each structure across a regional portfolio. One strategy that has been adopted to address this challenge is to group bridges into classes with similar design or structural performance. Traditionally, this grouping has been performed based on a relatively subjective identification of sub-classes. However, such an identification leads to a number of bridge classes and unwarranted grouping. This work suggests a performance based grouping methodology to group the box-girder bridges in California, and is the first systematic approach in sub-binning bridge classes for the regional risk assessment. The proposed grouping and analytical fragility methodology is used to derive fragility relationships for single frame box girder bridges in California. The fragility curves generated as a part of this research will be implemented in ShakeCast, a web based post-earthquake situational awareness application that supports the emergency responders and managers in California. This work concludes with the application of machine learning techniques for the generation of bridge-specific fragility curves.
Dr. Reginald DesRoches and Dr. Jamie E Padgett (Rice University)
Dr. Brani Vidakovic (ISyE), Dr. Lauren Stewart, and Dr. Iris Tien