Relation Between Risk-Informed In-Service Inspection and Inspection Qualification

A growing need exists for a quantitative measure of inspection effectiveness as an input to quantitative risk-informed in-service inspection (RI-ISI). A Probability of Detection (POD) curve could provide a suitable metric, but there can be significant problems associated with generating realistic POD curves by conventional statistical methods employing practical trials. The ENIQ inspection qualification methodology is used to provide high confidence that an inspection system will achieve its objectives, but is not designed to provide a quantitative measure of the type that can be used in RI-ISI analysis. This paper describes the results of a project set up to investigate approaches to quantifying the confidence associated with inspection qualification. Three approaches were considered and applied in pilot studies. These approaches were 1) a direct judgement method; 2) a Bayesian approach that included weighting and scoring the various parts of the evidence presented in the technical justification; and 3) an approach based on the relationship between POD and margin of detection. The pilot studies resulted in a set of recommendations related to the quantification of inspection qualification. Particular attention was paid to the development of guidance to support the application of the Bayesian model that combines and quantifies the ¿soft¿ evidence from a technical justification with the "hard" evidence obtained from practical trials. Besides the application of the quantification approaches, the project also illustrated how, if defect growth can be modelled, it is possible to link inspection qualification results, risk reduction and inspection interval.

SIMOLA Kaisa;  GANDOSSI Luca;  SHEPHERD Barrie; 

2010-04-23

Office for Official Publications of the European Communities

JRC48676

https://publications.jrc.ec.europa.eu/repository/handle/JRC48676,