ILI Performance--Validating Rupture Pressure Prediction Performance of In-Line Inspection Tools

Successful application of in-line inspection (ILI) data for assessing the integrity of pipelines depends on understanding the performance of the specific technology employed. Actual performance of these technologies can vary from that claimed by the inspection tool vendor depending on a number of pipeline design, construction and operational variables. Performance claims for metal loss tools are generally limited to accuracy of metal loss depth and positional measurements but depth is only one limit state generally to be considered, the other is a prediction of burst pressure for corroded pipe. Metal loss depth performance is an industry standard and 80% confidence of +/-10% wall thickness is an often stated performance for ILI technologies. There are no performance claims for accuracy in burst pressure performance which is due in part to the effect of interacted defect length and burst pressure criteria, in addition to defect depth used to predict a burst pressure. However, an understanding of actual in-line inspection tool performance can help pipeline operators gauge the relative level of conservatism associated with decisions to accept or reject metal loss features based on an ILI log prediction. Accurate and reliable correlation of burst pressure predictions from ILI with direct examination predictions depends on matching of appropriate areas of corrosion as well as the accuracy of the in-ditch methods used for validation. Complex areas of corrosion can be difficult to match with ILI predictions and introduce possible error in validation correlations. This paper examines the practical technical issues involved in making validation comparisons between in-line inspection predictions and in-ditch validation and presents new data analysis tools and techniques, particularly applicable to high resolution laser and ultrasonic direct examination technologies that can be employed to increase accuracy and reliability of burst pressure validation.