There’s More to NDT Reliability than POD – and Statistics Plays a Role

Abstract

Quantifying Nondestructive Testing (NDT) detection capability using Probability of Detection (POD) methods has become a more routine and often required part of NDT, resulting in a strong partnership between NDT and statistics. However, there is more to NDT reliability than POD. Components of NDT reliability can be identified at each step of an NDT process, from the selection of the NDT method to sizing identified, relevant indications. By integrating statistical methods throughout the entire NDT process, a more comprehensive quantitative assessment of NDT reliability can be achieved. In addition to an overview of the statistical methods underlying the industry standard POD methods, this paper provides an overview of additional statistical methods that may be used to evaluate other aspects of NDT reliability such accuracy and precision in flaw sizing, stability of an NDT process, and robustness of a POD specimen set design.