Data and Knowledge-Driven Bridge Digital Twin Modeling for Smart Operation and Maintenance

The rapid expansion of civil infrastructure in China underscores the critical need for advanced solutions to ensure the structural health of aging bridges. This study introduces a novel data and knowledge-driven digital twin modeling (DK-DTM) framework designed to enhance the safe and efficient operation and maintenance (O&M) of bridges. Such a system should be capable of (1) monitoring structural dynamics in real time, (2) capturing spatiotemporal details and changes (e.g., defects and deformations), (3) analyzing structure deterioration patterns, (4) predicting structure failure risks, and (5) generating optimal maintenance and repair actions for ensuring structural safety. Previous studies have developed advanced sensing techniques and robust artificial intelligence algorithms for capturing and analyzing bridge health conditions. However, most existing techniques and algorithms heavily rely on high-quality data, which are difficult to obtain during bridge O&M. This raises the critical question of how to incorporate expert knowledge together with data-driven tools to establish a trustworthy DT for bridge O&M. This study presents the DK-DTM framework, which uniquely integrates multi-source data collection, spatiotemporal modeling, and expert knowledge reasoning. By combining these components, the framework supports smart structural health assessments of bridges, enabling comprehensive monitoring, prediction, and decision-making for efficient maintenance. The spatial and temporal models provide real-time data, while the expert knowledge model functions as an automated evaluation tool for structural health assessment. The results demonstrate that the proposed DK-DTM framework significantly enhances the accuracy and efficiency of O&M processes for aging bridges, addressing key gaps in existing digital twin methodologies.