Decision making for a new pipeline’s design and provision of the most effective maintenance or repair measures for a pipeline in operation can be a long and costly process. The final decision made, whether during design or operation, may not always reduce the risk or remediate the threat. This is mainly due to the uncertainty and missing information regarding the field chemistry for current and future pipeline operating conditions, that were not considered and quantified during the assessment.
In this paper, two case studies of pipeline internal corrosion risk are presented, one for pipeline in design and the latter for pipeline in operation. Both cases were assessed using Bayesian Networks. Bayesian Networks (BN) have been used to quantify the value of information of uncertain and missing data. BN displays the cause-effect relationships of these data in the form of conditional probabilities to describe how one’s data is influencing internal corrosion rates probability. Thus, predicting the pipeline’s conditions over the design life. Operators can visualize the development of internal corrosion within a pipeline over time and gain clearer understanding of the causal relationships that could lead to pipeline failure. The results allowed operators to confirm the effects of the parameter and followed by a sensitivity analysis to find out which data to prioritize in acquisition and validation before proceeding to decide on how the pipeline should be designed and maintained/inspected in future.