Predictive Maintenance of Civil Infrastructure with Multimodal Sensing, Machine Learning, and Numerical Modelling

This body of work consists of four major areas following the hierarchical framework of Data, Information, Knowledge, and Wisdom: (1) Data acquisition: applications of multimodal sensors for the digitalisation and condition assessment in multiple environments; (2) Material level defect detection: machine learning-assisted detection of damages in steel and concrete infrastructure; (3) Asset-level condition assessment: holistic condition assessment and risk rating of infrastructure with Convolutional Neural Networks (CNNs); (4) Scenario-based performance prediction: long-term prediction of infrastructure responses with multi-physics modelling.