Investigations to Introduce the Probability of Detection Method for Ultrasonic Inspection of Hollow Axles at Deutsche Bahn

Abstract The vast experience with the automated, ultrasonic system for the inspection of hollow railway axles used by Deutsche Bahn shows that much smaller flaws are detectable than required. This results in a number of false calls. False calls lead to unnecessary demounting and disassembling of wheelsets, which generates unnecessary additional costs. In order to adjust the sensitivity of the inspection system to reduce the number of false calls without compromising safety, the capability of the system to detect cracks needs to be comprehensively established. This capability can be quantified by using probability of detection (POD) curves for the system. The multi-parameter POD model makes it possible to include several factors that influence the crack detection in the analysis. The analysis presented in this paper shows that crack position, orientation, depth extension, and shape as well as the geometry of the axle all have influence on the ultrasonic response amplitude. For future work, calculation of the POD using multi-parameter POD model with these parameters is planned.