A Probabilistic Framework for Multi-Hazard Evaluations of Buildings and Communities Subject to Fire and Earthquake Scenarios

Abstract

Community resilience to extreme events is an issue of increasing concern in our interconnected and urbanized societies. Meanwhile, cascading multi-hazard events, such as fire following an earthquake, can cause major social and economical losses in a community. Fire and fire following earthquake are low-probability but high-consequence events. The evaluation procedure under such extreme loading scenarios involves many uncertainties, such as the intensity, location, and characteristics of primary and secondary hazards, and properties and response of structural elements. However, current design methodologies for fire are mainly based on prescriptive approaches. Therefore, a probabilistic framework for evaluation of structures at elevated temperatures and risk of fire ignition should be developed.

The research presented in this dissertation investigates the hazards of fire and fire following earthquake, within a probabilistic framework and at three levels: (1) Probabilistic models for fire load and material properties at elevated temperatures are developed and applied to perform reliability analysis of a structural element at elevated temperatures; (2) performance of a prototype building is evaluated under fire and fire following earthquake within a probabilistic framework using OpenSees; and (3) historical data is used to develop a post-earthquake fire ignition model at the community level.

The probabilistic models developed at the first level are used as an input to measure reliability of a building at the second level under post-earthquake fires. Performance of the prototype building is evaluated for different engineering design parameters. As part of the procedure, the thermal module in OpenSees is enhanced for cascading fire following earthquake analysis.

Finally, the ignition model at the third level calculates the probability of ignition for individual buildings and provides an estimate for the total number of ignitions in a community following an earthquake. The tools and models developed in this dissertation can be used to identify vulnerable parts of a community to fire ignitions after an earthquake, and perform reliability analysis of structures under fire-only or post-earthquake fire scenarios.

The primary impact of this research work is that the models, tools, and frameworks developed in this dissertation contribute to creating a risk management platform to propose recommendations, produce new policies, and re-conceptualize existing ones for resiliency planning of cities under fire and fire following earthquake.