Corrosion threat assessment of crude oil flow lines using Bayesian network model

Abstract

Abu Dhabi Company for Onshore Oil Operations operates multiple carbon steel oil flow lines, which are emplaced on the desert surface of Abu Dhabi. The company's main oil pipelines are buried with coating and cathodic protection and internally protected by chemical inhibition, but the flow lines are without coating and cathodic protection. Over the years, this approach has been successful for flow lines, but the frequency of corrosion related leaks has increased recently due to changing operating and external conditions. This paper describes the use of a Bayesian network model that combines physics based models and expert knowledge of the flow lines to predict corrosion flaws depth and leak probability. It is shown that the Bayesian network approach can be useful in estimating location specific probability of failure and thus providing input to the prioritisation of inspections and corrosion mitigation. The approach was validated for five selected flow lines, where detailed field examination was available for comparison to model predictions.