Transportation: Pipeline integrity - 2 (Conclusion): Model choice governs accuracy of corrosion predictions

The Markov model works well over a wide range of pipe ages and defect depths and can predict the time evolution of the entire defect depth population and the evolution of the defects in subpopulations within different depth intervals. It is the only model that reproduces the time evolution of the corrosion defects from the depth data of the first inline inspection (ILI) and the value of corrosion rate - estimated from two or more repeated ILI. The model is also the only model that can capture the intrinsic, unavoidable stochastic nature of the corrosion process independent of the inspection tool's measurement errors or the initial defect depth distribution in the pipeline. Comparison of the models' performance before determining the Markov model as the model of choice for correct prediction of corrosion defect depth evolution and pipeline reliability is presented.