Internal corrosion is a complex phenomenon that includes under-deposit corrosion (UDC), microbially influenced corrosion (MIC), erosion, and localized and uniform corrosion mechanisms. For robust risk management of pipelines, there is a need to study the interactions of risk factors in internal corrosion. It is necessary to monitor the variations of corrosion risk factors and assess the pipeline's failure likelihood as a function of time. A Dynamic Object-Oriented Bayesian network (DOOBN) model is developed for this purpose. The DOOBN model has helped represent the probabilistic relationships among prevalent influencing risk factors and define their conditional dependencies. There are 94 risk factors considered for the different internal corrosion mechanisms. Results show that the probability of UDC occurrence for a given pipeline is 58, while MIC is 48. The study also confirms the increase in asset failure rate with the rise in internal corrosion. Results of the model are validated using filed data, and an accuracy of 93.22 is observed. The study serves as an early warning guide for the integrity management of pipelines against internal corrosion.
展开▼
