Project Summary
This project aims to produce a damage detection system for railway bridges using an instrumented train that can inspect bridges on the network while travelling at operational speed.
Project Achievements
Activities • Building, training and testing data-driven damage detection and classification algorithm. • Conducting scaled testing using a model bridge and an instrumented model train. • Conducting a practice-based feasibility investigation using an over 100-year-old bridge using an operational instrumented train.
Conclusions
• Achieving an unprecedented accuracy of 100% in detecting and classifying damage using the developed data-driven damage system using an instrumented model train. • Demonstrating the viability of the concept for a case study bridge by demonstrating over 75% accuracy in detecting the change in the riverbed and 85% in age classification of train-borne signals. • Building the practice-based core and foundation for the follow-up activities that will lead to an automated data-driven, drive-by damage inspection system.
Next Steps
• Expanding the application of the system for other types of bridges using the additional train measurements for two different instrumentation systems and surveys covering approximately 200 mile of track, for 2-5 year period. • Automating bridge numerical modelling through 3D scanning and climbing robots.
