Enhancements to Caltrans Bridge Fragility Relationships for Regional Seismic Risk Assessment in California

Click to enlarge


California Department of Transportation


Jazalyn Dukes, and Karthik Ramanathan

Bridge seismic fragility relationships are statistical functions that describe the probability of exceeding a specific damage threshold as a function of shaking intensity. They are a fundamental building block used in multiple (current and potential future) applications of interest to Caltrans including emergency response, design support using performance based earthquake engineering, planning and policy support. Bridge fragility relationships currently in use were mostly developed in the late 1990’s with bridge classes based on a limited number of parameters available in the National Bridge Inventory (NBI) and risk level based on limited sets of field damage observations. Attempts have subsequently been made to account for some differences in California bridge design, however the fundamental framework for describing bridge damage is not well aligned with Caltrans seismic design philosophy nor the California bridge inventory. The result is that bridge performance estimates have clearly not achieved their full potential. Exacerbating this situation is the lack of systematic organization of bridge design, retrofit, and maintenance data (beyond NBI parameters) required to make substantial improvements. The project holds specific endeavors for improving fragility relations for bridge classes in California including

  • Extending and subdividing existing bridge classes to better account for California bridge inventory
  • Calibrate and refine numerical procedures for generation of fragility curves to assure that they accurately reflect empirical damage observations, instrumental recordings of bridge response and extensive laboratory testing of bridge details.
  • Establish a refined fragility framework to be aligned with Caltrans bridge design and maintenance practices so that risk associated with specific failure mechanisms can be readily isolated and characterized.
  • Supplement the NBI information available about bridges with Departmental bridge design, retrofit, and maintenance data to create more and better bridge classes that have more consistent performance characteristics. For the newest bridges, this could include numerical values for specific seismic design-check parameters captured during the design phase.
  • Explore the potential of using multiple hazard parameters (say combinations of response spectral values, magnitude, distance bin, and duration) to create a more precise risk surface (rather than curve) with correspondingly lower uncertainty.