SEDL / STP / STP798-EB / STP33213S



Statistical Methods for Estimating Crack Detection Probabilities

Berens, AP
Senior research statistician and associate research statistician, Service Life Management, Aerospace Mechanics, University of Dayton Research Institute, Dayton, Ohio

Hovey, PW
Senior research statistician and associate research statistician, Service Life Management, Aerospace Mechanics, University of Dayton Research Institute, Dayton, Ohio


Pages: 16    Published: Jan 1983


Download this paper for $25 PDF (232K)          View License Agreement
Abstract

To characterize the uncertainty in nondestructive evaluation (NDE), a probability of crack detection (POD) as a function of crack length is postulated where POD is defined as the proportion of cracks of a given length that would be detected by the NDE technique when applied by inspectors to structural elements in a defined environment. This paper presents a statistical framework for describing the uncertainty in the NDE determinations, and evaluates various characterizations of NDE reliability. The data from a recent Air Force study on NDE reliability are used to estimate the parameters of the NDE model. For these representative capabilities, NDE reliability experiments are simulated. Different NDE capability characterizations are computed for each simulated experiment and are statistically compared.


Keywords:
nondestructive evaluation, NDE reliability, NDE capability, probability of crack detection, crack detection reliability, inspection reliability, probabilistic fracture mechanics, fatigue (materials), fatigue crack growth, fracture mechanics

Paper ID: STP33213S
Committee/Subcommittee: E08.04
DOI: 10.1520/STP33213S
CrossRef ASTM International is a member of CrossRef.