In the aerospace industry, the probability of detection (POD) has been accepted as an effective figure of merit for reliability of NDE performance, and is increasingly used as a condition-based component life management tool. However, the broadest possible use of POD estimates is yet to be achieved because of cost constraints, both in terms of sample preparation and time required. A strong desire in this particular industry sector to break the economical barrier has led a formation of the MAPOD Working Group where the use of NDE models as a part of the POD analysis procedure has been investigated. The objective is to endow POD analyses with a predictive power afforded by physicsbased NDE models. A predictive POD methodology can, for example, reliably extrapolate known POD results for existing inspection problems to other, newly encountered inspection conditions. In what follows, we will describe the proposed MAPOD methodologies in two classes, namely the full MAPOD framework briefly, and the transfer function MAPOD approach in some detail. The MAPOD methodologies have potential impact on nuclear system applications. This paper presents considerations on potential use of the predictive POD methods, specifically in two areas: First, inspection qualification is a possible area of the applications. Once an inspection is qualified in terms of an empirical POD for a given component condition with particular instrumentation, similar related inspections with modified component conditions and/or instrumentation may be qualified in terms of predicted POD estimates transferred from the empirical POD via a MAPOD methodology. The second area of the possible applications is the extended use of existing POD analysis results, obtained from components that were taken out of service, thus permitting destructive degradation characterizations, or otherwise well-characterized samples. The MAPOD methodologies will be able to predict POD from the empirical POD for related inspections, thus leading to full utilization of these extremely valuable data to estimate reliability of the related inspections.
Model-assisted probability of detection. Use of NDE models and statistical qualification of model predictions
Modell unterstützte Nachweiswahrscheinlichkeit. Anwendung von ZfP-Modellen und statistische Qualifikation zur Modellvorhersage
2009
10 Seiten, 4 Bilder, 2 Tabellen, 18 Quellen
(nicht paginiert)
Conference paper
Storage medium
English
Herschel Sorption Cooler Qualification Models
British Library Conference Proceedings | 2005
|HERSCHEL Sorption Cooler Qualification Models
Springer Verlag | 2005
|Accelerometers, Bendix Model Ttg-5 Qualification Tests
NTIS | 1955
|MADEQ: A model for Avionics Device Qualification
IEEE | 2013
|