Abstract
Bridges are critical infrastructure assets that face a variety of stressors throughout their service life, requiring a life-cycle approach to assess their risk profile. Recent advancements in sensing and monitoring technologies provide a powerful data foundation to improve the accuracy of life-cycle risk assessment (LCRA). However, existing works that incorporate data for probabilistic risk assessment typically focus on individual bridges and rely on single-source data, limiting their scope and applicability. To this end, a system digital twin (SDT) framework based on Bayesian network (BN) is proposed to integrate multi-source data for LCRA of bridge networks. Specifically, the SDT can capture correlations and interdependencies across various scales, including within individual components (e.g., multiple failure modes), between components within a system (e.g., bridges along a route), and across interconnected systems (e.g., bridge and hydraulic systems). It integrates data from various sources including bridge inspections, traffic monitoring facilities, and water watch stations. A coastal bridge network in Miami-Dade County, FL, is used as an illustrative example to demonstrate how the SDT integrates multi-source data for risk assessment. Additionally, several future scenarios are hypothesized to showcase the applicability and flexibility of the proposed framework in supporting risk management for infrastructure systems.