Multi-Hazard Risk Assessment of the Interconnected Infrastructure Systems

The escalating frequency and intensity of natural hazards necessitate robust methodologies for assessing and enhancing the resilience of interconnected infrastructure systems. This thesis introduces a comprehensive framework for multi-hazard risk and resilience assessment within these systems, with a particular focus on the impacts of climate change. The framework underscores the potential for cascading failures, where the malfunction of one component can initiate a chain of cascading failures throughout the network, resulting in widespread economic disruption and significant societal consequences. The first study proposes an integrated framework, BN-SPM, based on Bayesian Network (BN) and Strongest Path Method (SPM) to assess natural hazard risks to interconnected infrastructure systems. By leveraging Bayesian inference, the BN transforms prior failure probabilities into posterior ones, while SPM prioritizes network entities based on impact, vulnerability, and risk indices. This model evaluates various scenarios, including non-hazard conditions, flooding hazards, and mitigation strategies. The results highlight and emphasize the need for robust risk assessment in interconnected systems. The second study addresses the limitations of static risk assessments by introducing a Dynamic Bayesian Network (DBN) framework for evaluating the dynamic behavior of interconnected infrastructure systems under multi-hazard scenarios. This study investigates the complex network of critical infrastructure, focusing on the probabilistic conditions and functional dynamics of disruptions and restoration processes. The results reveal and underscore the necessity of dynamic risk assessments for multi-hazard scenarios to develop robust risk management strategies. The third study advances the assessment methodology by integrating Fuzzy Dynamic Bayesian Networks (FDBNs) with auxiliary nodes, aiming to reduce computational complexity while modeling uncertainties and temporal dynamics in hazard impacts and recovery processes. Two scenarios are investigated: multi-hazard events involving flooding and landslides, and climate change-induced flooding. The findings highlight differential impacts on infrastructure resilience, according to the crucial role of some entities rather than others. The integration of FDBNs with auxiliary nodes improves model accuracy and reduces computational complexity, providing strategic insights for resilient infrastructure planning. The implications of the abovementioned phases of this study are evaluated on Saint Lucia’s vulnerable southern coast, including Hewanorra International Airport (HIA). Overall, this thesis advances methodologies for resilient infrastructure planning and management, addressing critical gaps in current approaches. The findings emphasize strategic planning and adaptive resilience-building measures to enhance infrastructure robustness and minimize socio-economic disruptions in hazard-prone regions.