A Bayesian model framework for calibrating ultrasonic in-line inspection data and estimating actual external corrosion depth in buried pipeline utilizing a clustering technique

In the petroleum industry, oil and gas pipeline operators routinely employ non-destructive in-line inspection (ILI) tools to perform integrity assessment, in which corrosion defects are detected, located and sized. However, the inspection technology is not perfect, and its accuracy is influenced by intrinsic measurement error of the ILI device as well as measurement noise. The quality of the inspection result should be assessed, and the inspection device should be calibrated before further assessment is performed. In the present work, a calibration model for an ultrasonic ILI device is proposed based on physical principles, and both systematic error and random error are able to be estimated. In addition to the errors from inspection devices, the soil environment introduces various uncertainties into the corrosion propagation. This paper presents a methodology to infer the actual corrosion defect depth based on detection theory and to account for the effect of soil property variation by combining cluster analysis with a Bayesian inferential framework. A numerical study on calibration uncertainty shows the influence of the number of field verifications. The proposed model framework is applied to a 110-km pipeline system to illustrate its application.