Validating ILI Accuracy Using API 1163

It is important to validate the accuracy of in-line inspection (ILI) tools to know how many excavations are needed to maintain the integrity of a pipeline segment. Performing sufficient excavations is important to ensure there are no defects left in the pipeline that have even a remote chance of failure. In some cases additional excavations may be necessary to ensure safety where in other cases no excavations may be necessary. This paper looks at using spatially recorded metal-loss data collected “in-the-ditch” to measure the accuracy of ILI tool results. Examples of spatial in-ditch data are laser scans for external corrosion and UT scans for internal corrosion. Spatially mapped metal loss, because all of the corrosion area is mapped, has the advantage of allowing more comparisons to be made for a given corrosion area and also allows the interaction among corrosion pits to be studied for examining burst pressure calculation accuracy. From our studies we find the depth error for shallow corrosion 10%–20% wt deep is often not representative of deeper corrosion in the same pipeline and the interaction criteria for ILI tools needs to be larger than the interaction criteria for in-ditch data. Examples are shown with these types of results, and by interpreting the results in conjunction with API 1163, certain ILI runs are shown that require no excavations where others may require additional excavations than suggested by normal +/−10% wt ILI data.