Generally, crack detection in-line inspection (ILI) tools can be excellent at detecting planar crack-like reflectors. However, the performance specification outlines that there is no guarantee that reported crack-like reflectors will be confirmed cracking.

Upon receipt of a crack-detection ILI listing, a proportion of the reported crack-like reflectors could be actual cracks, e.g., stress corrosion cracking or weld lack of fusion, both of which are considered to be true positives in line with API 1163. However, a proportion may also be false positives (verified to be non-crack like).

Verifications provide an opportunity to evaluate ILI signal characteristics and make the link between reflections and actual integrity threats. However, the costs associated with repeat verifications of similar ‘False Positives’ can be frustrating and expensive for operators, particularly if there is a confirmed cracking threat. As per API 1176, it is therefore important to consider which reflectors are a ‘likely’ and ‘unlikely’ crack prior to executing a strategic targeted verification campaign.

This paper presents how a natural gas operator has utilised their extensive field verification database to develop an analytically led process to support the evaluation and prioritisation of newly reported crack-like indications to support the future management of SCC.

This process integrates:

• Extensive field data reviewing historic verification findings.
• Increased knowledge of the cracking threat for each system.
• An evaluation of signal characteristics 
• SCC susceptibility to identify the ‘likely’, ‘possible’ and ‘unlikely’ cracks as per API 1176.

The process output will enable the operator to perform a system wide or pipeline specific analysis upon receipt of a new crack detection dataset, supporting discussions with the ILI vendors to efficiently manage the threat of SCC and minimise future ‘unnecessary’ excavations. The process could also play a significant role to articulate and justify integrity related decisions to the key stakeholders.