Technical Reports

On-Line High-Speed Rail Defect Detection, Part II

  • 01
  • Mar
  • 2012
AUTHOR: Stefano Coccia, Robert Phillips, Ivan Bartoli, Salvatore Salamone, Piervincenzo Rizzo, Francesco Lanza di Scalea
SUBJECT: High-Speed Passenger Rail, Rail and Infrastructure Integrity, Track
KEYWORDS: Transverse cracks; ultrasonic guided waves; noncontact ultrasonic testing; laser ultrasonic; air-coupled sensors; finite element modeling; automatic defect classification
ABSTRACT: The objectives of this project were (1) to improve the defect detection reliability and (2) to improve the inspection speed of conventional rail defect detection methods. The prototype developed in this work uses noncontact transducers, ultrasonic guided waves, and real-time statistical pattern recognition to improve system reliability. The prototype was field tested seven times with the support of the Federal Railroad Administration (FRA) R-4 hy-railer vehicle. The latest tests at Herzog, Inc., in June 2010 showed excellent results. Two blind tests resulted in a defect detection reliability exceeding industry average and American Railway Engineering and Maintenance-of-Way Association recommendations. These tests also proved the potential for distinguishing good welds from defective welds, an excellent detestability of the vertical split head defect, and the potential for characterizing different rail surface conditions for grinding management purposes. Within this project, a 250-foot-long rail defect farm was also constructed at the University of California–San Diego, with in-kind donations from BNSF Railway, for the development of this and other rail inspection technologies of interest to FRA and to the railroad industry.
Pdf Download PDF Document [10.1 MB]