Detecting external failures in coated, buried pipelines:: Theoretical model and experimental verification

An experimental prototype, comprising a buried pipeline, was built with the purpose of calibrating a transmission line model and an artificial neural network (ANN) algorithm for locating and assessing the severity of external corrosion damage of defects in the pipeline's coating when the pipe was subjected to different levels of cathodic protection (CP) and to different environmental scenarios. The environmental scenarios were created by changing the value of different variables in the physical model, such as soil resistivity, defect (holiday) location, positions of reference and counter electrodes, and level of CP. The theoretical models were able to establish trends in the impedance signatures collected on the pipe system and corresponding to different damage locations and extents without further experimental work. The proposed theoretical-experimental methodology accurately could assess the level of CP, locate the position of holidays without the need to excavate, and estimate the severity (qualitative scale 1 to 100) of the corrosion damage at the defect sites.