Reliability Assessment of Corroded Pipelines using Bayesian Updating Technique

In the oil and gas industry, metallic pipelines are widely used as the most efficient and safest way of hydrocarbon transportation. Pipelines are subjected to deterioration over time due to corrosion. Either internal or external corrosions, both can contribute to failure if left unmitigated. In-line Inspection (ILI) tools are used to inspect pipelines and are usually associated with a double-sided accuracy. The data from intelligent pigging however, does not give a 100% accurate result. Therefore, overestimation will lead to more frequent repairs and unnecessary inspection, which will be very costly. On the other hand, underestimation will cause critical damage which if left unmitigated can cause failure in the pipeline. Therefore, improvement of existing reliability functions based on Limit State Function (LSF) Design is needed to produce a reliable probabilistic model. A study of Bayesian updating theory incorporating both probability of failure and probability of survival is presented in this research in order to revise a calculated probability using additional data and prevent overestimation or underestimation of Maximum Allowable Operating Pressure (MAOP) in offshore pipelines. The case study involves a retired section of an offshore pipeline in Peninsular Malaysia. The updated probability of failure according to the different codes shows that the ASME B31G code predicted well with the burst test result, followed by the SHELL 92 code, and lastly the DNV RP-F101 code. The future prediction of the corrosion rate for the pipeline based on its historical corrosion data and also its respective updated probability of failure has also been projected and it is found out that the pipeline will run fine up until 2018.