For pipelines with successive ILI runs the detected population of corrosion defects can be compared to identify both internal and external corrosion growth. Depending on the number of defects to be compared, the assessment can demand significant effort and expertise to ensure accurate and meaningful correlations between often very large ILI data sets. Specialist ILI comparison software facilitates efficient and accurate signal-to-signal matching and the determination of defect specific growth rates across very high defect populations. However, since ILI as a measuring technique is subject to inherent uncertainties, the prediction of where corrosion is active and the rate of growth from consecutive ILI runs also has a degree of uncertainty. The level of uncertainty is influenced by several sources of error:
The ability to accurately match the metal loss sites between the two ILI data sets
Identification of measurement bias associated with the ILI tools
Understanding the repeatability errors between the two ILI tools
There are various approaches that are used to compensate for these inherent errors. For instance, there are different ILI data matching methods that can be used and depending on the level of precision employed and the input data available these will result in varying levels of accuracy. We state signal-to-signal matching is the most precise and accurate approach that can be used over other methods such as box matching, but is there a common understanding of what “signal-to-signal matching” means, what information is required to perform it, what are the ways it can be done and the relative merits? This paper focuses on these questions in relation to comparing magnetic ILI tool data and looks at the challenges for signal matching across magnetic ILI tools with differing resolutions and even from different vendors. In addition, we discuss the importance of understanding tool bias and repeatability and minimizing the impact of these errors.