Abstract
Infrastructure systems are the backbone of modern societies and are critical for well-functioning communities. Natural and man-made hazards have the potential to disrupt the services provided by these systems, thus impacting normal community functions. For a community to assess risk from any hazard, it must first understand its dependency on the services provided by supporting infrastructure and, in turn, how the infrastructure is not only vulnerable to the hazard, but dependent on other infrastructure systems that may also be vulnerable. However, merely understanding the consequences and impacts of interdependent infrastructure failures on critical community services, let alone prioritizing capital investments to shore up aging, failing or otherwise vulnerable systems, is a daunting and often unachievable challenge for resource constrained communities.
This dissertation proposes a novel all-hazards analysis (i.e., AHA) methodology and knowledge management framework to enhance our understanding of risks to interconnected infrastructure, systems, and networks. The methodology advances risk analytic capabilities through the integration of concepts from graph theory, knowledge representation, and function-based engineering design. Infrastructure systems are modeled as multilayer networks and their behavior is simulated through the application of scalable function-failure logic designed to enhance risk mitigation guidance, while reducing the need for high-fidelity engineering data. In addition, the AHA Text Analytic System is proposed to enhance knowledge base population through the development and application of infrastructure specific named entity recognition.