Fraser Gray, Susannah Turner, Jack Davies
Highgrade Associates Limited, Newcastle upon Tyne, UK
Abstract
In-line inspection (ILI) is the primary technique used for detecting and sizing corrosion metal loss in pipelines. The Pipeline Operators Forum (POF) requirements state that ILI vendors should provide Probability of Detection (POD) and sizing specifications for metal loss anomalies of different sizes (classified by the detected anomaly length and width). However, these sizing classifications do not represent the often complex morphology of corrosion. For example, deep pits within shallower general corrosion may be simplistically categorized within the general sizing classification.
This paper explores the impact of corrosion morphology on ILI tool detection and sizing capabilities, using real life examples. The review considers corrosion validation data for three anonymized pipelines, with ILI data from three different vendors using two different tool technologies, Ultrasonic Wall Thickness Measurements (UTWM) and axial Magnetic Flux Leakage (MFL). Methodologies given in API 1163 (2021) have been used for the analysis of ILI tool performance and examples of validated corrosion and associated ILI sizing are discussed. The potential impacts on integrity assessment are also considered.
The review presented in this paper provides insight on the performance of both MFL and UTWM ILI detection and sizing capabilities for complex corrosion anomalies (such as deep pits within shallow general corrosion). The potential for integrity assessment of pipelines with this type of corrosion morphology to be non-conservative due to treatment of ILI sizing tolerance is also discussed, considering both through-wall leak and pressure-driven burst failures. The findings of this paper further support previous research stressing the importance of validating more than just the deepest ILI reported anomalies, to understand the corrosion morphology, the actual sizing uncertainty of the ILI, and the associated impact on integrity assessment.



