This paper presents a methodology to evaluate the reliability of corroding pipelines by simultaneously considering the growth and generation of corrosion defects. The non-homogeneous Poisson process is employed to model the generation of corrosion defects, whereas the non-homogeneous gamma process is used to characterize the growth of corrosion defects once generated. The parameters included in the non-homogeneous Poisson process and non-homogeneous gamma process are evaluated from the inline inspection data using a hierarchical Bayesian model. The measurement errors associated with the inline inspection tools are taken into account in the Bayesian updating. The time-dependent failure probability of the corroding pipeline is evaluated using the Monte Carlo simulation technique. The methodology is illustrated using a natural gas pipeline that has been subjected to multiple inline inspections over a period of time. The results illustrate the necessity to incorporate the generation of new corrosion defects in the reliability analysis of corroding pipelines.

This content is only available via PDF.
You do not currently have access to this content.