This paper is aiming at providing practical recommendations for effective and efficient integrity analyses of corroded onshore pipelines. The focus herein is on comparing integrity analysis approaches; namely deterministic safety factors calibrated safety factors and probabilistic analysis. The comparison study utilizes real life application of In-Line-Inspection (ILI) and field measurements of corroded onshore pipelines. The study addresses the advantages and disadvantages of each analysis approach emphasizing on the ability to capture measurements uncertainties and learnings from trending data properly. Such learnings can be addressed by a relative comparison of both ILI and field measurements (i.e. relative uncertainty measurements). Common practice showed that both measurements are contaminated with errors and/or biases. In this paper a new methodology for estimating ILI uncertainties based on minimizing the relative error between ILI and field measurements is presented. Deterministic codified safety factor approach is simply comparing the nominal burst pressure against the maximum operating pressure where some epistemic uncertainties may or may not be included. On the other hand safety factors can be calibrated to achieve a target reliability level while considering vendor claimed aleatory uncertainties. In reality the uncertainty of ILI measurements may be departure from vendors claim in terms of value and/or statistical properties. Therefore the calibrated safety factor may not be able to capture both aleatory and epistemic uncertainties under one umbrella. Recent advances in application of probabilistic approaches showed an added value for pipeline integrity programs along with it capability to address all types of uncertainties. The probabilistic or reliability-based approach has showed some superiority in this arena compared to deterministic integrity analysis approaches. In studying the three analysis approaches on onshore corroded pipelines operators would have the chance to assess the level of conservatism or non-conservatism in analyzing their corrosion integrity programs. This study presents how a full probabilistic analysis including the measurement uncertainty and calibrated corrosion ILI data can provide more effective and efficient estimation of the corrosion reliability measurements compared to traditional and recommended practice approaches.
Key words: uncertainty, measurement, calibration, In-line inspection (ILI), probabilistic, reliability, corrosion, probability of failure (PoF), failure pressure prediction (FFP), metal loss, integrity assessment, Ultra Sonic (UT) inspection, Magnetic Flux Leakage (MFL) inspection, Non-destructive examination (NDE) measurement