@article {Morgan:2009:1354-2575:266,
title = "Automated Ultrasonic Testing (AUT) Qualifications and variations",
journal = "Insight - Non-Destructive Testing and Condition Monitoring",
parent_itemid = "infobike://bindt/insight",
publishercode ="bindt",
year = "2009",
volume = "51",
number = "5",
publication date ="2009-05-01T00:00:00",
pages = "266-270",
itemtype = "ARTICLE",
issn = "1354-2575",
url = "https://www.ingentaconnect.com/content/bindt/insight/2009/00000051/00000005/art00007",
doi = "doi:10.1784/insi.2009.51.5.266",
author = "Morgan, L L",
abstract = "The use of AUT systems is now becoming the norm in testing the integrity of girth welds during the construction of cross-country pipelines. It is good practice to qualify such systems before they are used in order to demonstrate that the contracting agency, with the procedure that is
to be used, can achieve the quality of performance that is needed. This is in terms of detecting the smallest defects that need to be found, with sufficient reliability and accuracy. This qualification is done by testing deliberately induced defects in test welds. Such tests are time consuming
and expensive. This study describes work to establish how these tests can be minimised using a statistical approach, while preserving the original objectives. This has been demonstrated using a large data-set from earlier work. The significance of data that lie well outside the statistical
norm is also considered and a criterion for exclusion is proposed.",
}