A Dynamic Resilience Framework for Urban Bridge Networks under Extreme Hazards

Thesis Abstract

In the face of increasing frequency and severity of extreme hazards—such as seismic events, flooding, and high-wind storms—urban bridge networks exhibit nonlinear and dynamic degradation patterns that threaten structural integrity and functional connectivity within metropolitan transportation systems. This study addresses the need for a dynamic resilience framework capable of guiding adaptive decision-making for mitigation, response, and recovery across interdependent bridge networks. The aim is to develop and validate a dynamic resilience framework that integrates network science, structural reliability, and adaptive management to quantify and enhance urban bridge network resilience under multi-hazard scenarios. Specific objectives are (1) to conceptualize a dynamic resilience metric that captures structural performance, functional accessibility, and recovery trajectories; (2) to identify hazard–network interaction mechanisms through a combined theoretical lens of the Network Theory of Resilience and the Theory of Critical Infrastructure Interdependencies; (3) to calibrate and validate the framework on a representative urban bridge network consisting of 120 bridges categorized by structural type, age, and maintenance status; (4) to assess the framework’s decision-support capacity under three hazard scenarios—magnitude-adjusted earthquakes, flood-induced scour, and wind-induced fatigue—over a 20-year planning horizon; and (5) to formulate evidence-based guidelines for strategic interventions and prioritization of retrofit, repair, and redundancy measures. The methodology employs a mixed-methods research design anchored in systems engineering and urban resilience theory. The population comprises urban bridge networks within a major metropolitan area and corresponding transport demand data. A purposive sample of 120 bridges is analyzed, with data collected from structural health monitoring records, inspection reports, traffic load inventories, and emergency response logs. Instruments include a standardized bridge performance questionnaire, sensor-based data streams (displacement, strain, seismic accelerations), GIS-based network topology datasets, and hazard scenario templates developed from regional hazard models. Validity and reliability are ensured through triangulation of structural measurements with inspection classifications and historical outage records, and by piloting data collection on a sub-network of 15 bridges prior to full-scale deployment. The analytical framework integrates (i) a dynamic resilience metric operationalized via a coupled agent-based and network-analytic model, (ii) probabilistic structural reliability analyses using Monte Carlo simulation, (iii) scenario-based optimization for recovery prioritization, and (iv) regression analysis to identify drivers of resilience degradation. Theoretical grounding draws on the Network Theory of Resilience, the Theory of Critical Infrastructure Interdependencies, and the Dynamic Systems Theory to capture nonlinearity, feedbacks, and time-dependent recovery processes. Data analysis proceeds in four stages. First, event-driven simulations generate hazard-triggered damage states for each bridge, producing time-series data on functionality and connectivity loss. Second, a dynamic resilience index is computed for each time step, integrating structural performance, service level, and network criticality, with sensitivity analyses across hazard intensities. Third, multiple regression models quantify determinants of resilience trajectories, while a random-effects panel approach accounts for intra-network heterogeneity. Fourth, a scenario-optimization module identifies cost-effective intervention portfolios under budget constraints, prioritizing risk reduction, redundancy, and rapid recovery. Expected findings indicate that incorporating interdependencies and time-varying recovery improves predictive accuracy of resilience scores by 25–40% compared with static approaches. The framework is anticipated to reveal that targeted fortification of critical bridges with high betweenness centrality and redundancy, coupled with adaptive maintenance scheduling, yields sustained serviceability under multi-hazard conditions. The study contributes to knowledge by presenting a novel dynamic resilience framework that operationalizes resilience in urban bridge networks through an integrated, time-sensitive metric and decision-support tools. It advances methodology by combining agent-based network modeling with probabilistic structural analysis and scenario optimization, offering transferable templates for other critical infrastructure clusters. Practically, it provides planners and engineers with a reproducible workflow for hazard-informed retrofit prioritization, emergency response planning, and long-term resilience investments. The main conclusion envisaged is that dynamic, interdependency-aware resilience frameworks yield superior risk reduction and service continuity relative to static or siloed approaches. Recommendations include embedding the framework into metropolitan transport agencies’ resilience plans, expanding SHM data integration, and conducting periodic re-calibration using real hazard events to sustain model validity and applicability.