Resilient Infrastructure Systems: From Seismic Damage Evaluation to AI-Enabled Quality Assurance and Digital Design

Abstract

Intelligent technologies are playing an important role in the development of resilient infrastructure that builds the ability to identify structural weaknesses, provide quick and accurate assessments, and enhance decisions across the infrastructure life cycle. This review explores the use of AI and machine learning in three related fields: seismic damage assessment, construction quality assurance, and digital design. After an earthquake, computer vision and deep learning based. AI can help identify cracks, spalling, corrosion, deformation, etc. and help process the information rapidly; earthquake regional scale models can help with portfolio level damage assessment and recovery planning. Automated imaging and sensor systems can provide continuous quality monitoring during construction, thus improving the ability to detect defects early in reinforcement, concrete and welding and minimize potential rework, all of which can be expensive. In digital design, generative AI can analyze several design options, and Building Information Modeling and digital twins can synchronize design data with real-time operational data to enable predictive maintenance and life cycle performance management. The applications are connected to each other by vibration, strain, imaging and other sensors that are used for structural health monitoring, giving a crucial link between these applications, allowing continuous condition assessment. While there has been significant progress, there are still practical limitations in using the models, such as the availability of limited datasets for rare and high consequence events, poor generalization across structure types and environmental conditions, interoperability issues, uncertainty, and the need for robust validation.