Implementation and Validation of Reliability-Based Crack Assessment for Natural Gas Pipelines

Abstract

Abstract Traditional in-line inspection (ILI)-based crack management programs use deterministic methods, where the calculated failure pressure ratio (FPR) and ILI-reported crack depth are compared with their respective thresholds. In recent years, TC Energy has developed a probabilistic crack assessment method, where annual probability of small leak (POSL) and probability of failure (POF, i.e., probability of burst) are evaluated. The mitigation plan is then made by comparing the annual POSL and POF with their respective thresholds. The advantage of the probabilistic method over deterministic method is that the former portrays reality better by explicitly accounting for the uncertainties associated with pipeline geometric and material properties, ILI-reported crack sizes, crack growth and burst pressure models. This study demonstrates the safe implementation of the probabilistic assessment method for stress corrosion cracking (SCC) based on EMAT-reported ILI data and correlated dig data associated with three natural gas pipelines in Canada. Comprehensive validation was conducted by comparing the EMAT-based mitigation plan with the in-ditch assessment of a large set of dig data. Three key questions were addressed in the validation: (1) Does the probabilistic method capture all critical features identified in the field? (2) Whether the features avoided by the probabilistic method were unnecessary to excavate based on in-ditch assessment? (3) What is the benefit of the probabilistic method in comparison with the traditional deterministic method? The examination indicates that the developed probabilistic assessment process captures all the critical SCC features identified in the field, and the digs avoided by the probabilistic method are confirmed to be unnecessary in the field. The result demonstrates that the reliability-based method can reduce a significant number of unnecessary digs without compromising safety.