Operators of energy pipeline systems face a never-ending challenge to manage and balance integrity data and risk. An integrity threat of particular concern is the detection, classification, and sizing of mechanical damage anomalies, especially those caused by third parties. Mechanical damage, which often appears in the pipeline as a dent with coincidental gouging, is one of the leading causes of serious pipeline incidents globally each year (CER, PHMSA, CONCAWE, EGIG). Unfortunately, many in-line inspection (ILI) systems have difficulties inspecting the dent region for coincidental gouging and distinguishing it from corrosion.

Solving this issue began with the development of an innovative inspection platform that combines multiple ILI technologies into a single ILI tool for the most comprehensive inspection available, known as Multiple Dataset (MDS).  A recent enhancement to the MDS system performance was the development of the first gouge classifier backed by an industry compliant performance specification using advanced machine learning models. The specification outlines the classification of mechanical damage gouges, in addition to depth sizing of metal loss anomalies located coincident with a dent.

With a true dent/gouge classifier, operators will be able to confidently discriminate between corrosion versus gouging in a dent, the latter being a much greater integrity threat. Operators will then be able to prioritize repairs on those anomalies that pose the greatest threat and re-direct their budget to other activities that will enable better management of overall integrity.