R. B. Thompson, L. H. Brasche, Iowa State University, Ames, IA
Probability of Detection (POD) is the metric whereby the efficacy of an inspection is judged in probabilistic life management. Traditionally, this is determined empirically by first measuring the flaw response as a function of flaw size for a set of samples and conditions believed to be representative of the inspection situation of interest. The data is then analyzed by what is known as the â versus a technique. In the simplest implementation, a linear regression of log flaw response (â) versus log flaw size (a) is conducted. The regression line and the standard deviation of the data about this line, along with the inspection threshold, then determine the POD. This well accepted procedure suffers from the significant time and cost associated with the POD determination, which should be repeated for each change in the inspection problem. To overcome this limitation, there is significant interest in the use of models, derived from either careful laboratory experiments or physics-based computations, to generalize the results of one POD determination to another, related physical situation. The Model-Assisted POD (MAPOD) Working Group, led by the Center of NDE at Iowa State University, has been formed under the joint support of the Air Force, Federal Aviation Administration and National Aeronautics and Space Administration to explore this possibility. This paper will review recent progress of the MAPOD Working Group, including past demonstrations that have occurred, future demonstrations that are in the planning process, and protocols that are currently being developed.
Summary: This paper describes an emerging concept, Model-Assisted Probability of Detection (MAPOD), that is being developed to reduce the need for multiple, time consuming and expensive empirical determination of probabilty of detection (POD) in related situations. The papers is a summary of activities of the MAPOD Working Group and includes a discussion of the motivation, demonstrations that have been completed, demonstrations that are being planned, and emerging protocols
