Model-Assisted Probability of Detection for Ultrasonic Rail Flaw Detection
This report gives a roadmap for applying model-assisted probability of detection (MAPOD) to the ultrasonic detection of rail flaws. Researchers at ENSCO, Inc., selected this type of track inspection as the most suitable for the MAPOD approach.
MAPOD analysis involves identifying the factors that can affect the probability of detecting flaws using a defined technology. These factors are assessed either theoretically or experimentally. Experimental assessment involves physical tests with sample flaws. Analytical assessment involves computer modeling of flaw detection. MAPOD analysis combines the results from a relatively small number of physical tests with more extensive results from computer simulation. Two options exist for combining the theoretical and experimental results, both requiring analytical methods to be developed.
Ultrasonic rail flaw detection is suitable for MAPOD analysis because it can be modeled with commercially available software. The Federal Railroad Administration (FRA) has used this software to develop ultrasonic rail flaw detection models that could be used in this study.