Bayesian Techniques for Rail Reliability Modelling and Maintenance Decision Support

Host university




Infrastructure Health Monitoring and Predictive Maintenance

View Theme

Data collection, security, and integration

View Theme

Modelling, Simulations and Prognostics

View Theme


A key problem with rail infrastructure is the appearance of defects on the rail surface (e.g. squats, worn profiles, contact fatigue). If left unattended, these defects can develop and result in significant safety risk (e.g. derailment). To manage these risks, a number of inspection activities are typically undertaken, including: 1) inspection technologies (e.g. a track recording vehicle) that measure the rail profile; 2) work orders that briefly describe the execution and results of visual inspections; 3) detailed engineering reports with more comprehensive (but less structured) information on defects. The key challenge is to use the information from these multiple data sources to develop a complete statistical picture of the appearance and evolution of critical defects as well as their dependence on factors such as gross tonnage, traffic speed, curvature, etc. Moreover developing and communicating maintenance plans that respect regulatory constraints and resource budgets (e.g. staff, specialized equipment) would be greatly aided by the development of mathematical optimisation techniques — particularly in the case of multiple competing objectives (cost, risk, etc.) — to plan inspections and rail renewals.  


Michael E. Cholette

Chief Investigator
Queensland University of Technology
View Bio

Tommy Chan

RIIS Hub Lead Chief Investigator
Queensland University of Technology
View Bio

Joe Mathew

Partner Investigator
Asset Institute
View Bio

All projects