CONDITION MONITORING ASSET INTEGRITY MANAGAMENT: A DATA-DRIVEN APPROACH

Abstract

Pipelines are considered as an essential asset within oil and gas industry as it is the safest way of transporting it. Therefore, it is vital for plants to maintain integrity of their pipelines. This asset’s weakness is that they are prone to deteriorate due to various factors: corrosion development, crack growth, etc. To tackle these issues, the real-time condition monitoring asset integrity management models are of high demand. That type of model will enable operators to observe the operational risk online and decide on an optimum inspection period at any time. This paper proposes a methodology where the integrity of pipeline is dependent on two defects at a time: corrosion depth growth and crack depth aggregation. One of the defects, corrosion depth, is based on dynamic Bayesian network model, whereas another defect is associated with estimations in stress corrosion cracking. The mentioned is employed to estimate probability of failures of the pipeline. The estimated overall probability of failure started rising at month 77, while increasing until 0.27 in the following 31 months. Further, estimation of the consequence of corrosion based on API RP 581 standards is provided within this study. Incorporating probability of failure along with consequence of failure, the real time operational risk is estimated. The risk started increasing abruptly at month 77, and within 2.6 years of period the risk of failure was about $475’000. Next, the optimal inspection interval is assessed via introducing utility function and maximum acceptable risk line. The optimum period for performing maintenance operations was at month 78, after which risk was considerably reduced. The proposed methodology is manifested through hypothetical data generated within MS Excel software.