Risk Assessment of the Transmission Pipeline Appurtenance Threat

Abstract The Enbridge liquids pipeline system contains more than 28,000 KM of transmission pipe. Systemwide assessments are conducted to evaluate the risks associated with integrity threats, whereby historically the focus has been on crack, corrosion, dents, and geohazards. However, as reported in the API 2022 Performance Report, equipment failure is the leading cause of liquid pipeline releases. Correspondingly while release volumes are typically low, there are other threats such as appurtenances, which can lead to failure and warrant risk assessment. Prior assessments used average historical failure rates for appurtenances which enabled high-level risk quantification; however, this approach was overly simplistic and did not differentiate between different characteristics of the same asset class. In 2021–2022 four appurtenance risk ranking models: valves, sleeves, flanges, and offtakes, were developed to enable comprehensive systemwide risk management. During model development, there was an objective to incorporate model inputs that are readily available from drawings and in-line inspection data. Correspondingly the model is easily implemental whereby the inputs included only: asset information (i.e. vintage, construction), site specific attributes (i.e. proximity to other appurtenances), and operational considerations (i.e. proximity to pump stations, pressure cycling, product types). The steps conducted to enable risk assessment of the appurtenance threat included: • Literature review of historical failures and failure modes based on operator specific databases and publicly available literature to identify key factors that lead to failures. • Application of model factors based on appurtenance characteristics and subject matter expert input. • Calibration of models so that the systemwide average and aggregated likelihood of failure align with those historically observed. • Validation of models based on the predicted likelihood of failure for historically failed assets. It was demonstrated that the models captured the key characteristics that impact possible failure modes for pipeline appurtenances.