Reliability prediction for corroding natural gas pipelines

This paper aims to evaluate time-dependent reliability of a corroding onshore underground natural gas pipeline system over its lifetime. The reliability analysis is segment-based, as opposed to defect-based and the pipe segment is examined with respect to external metal loss corrosion. The non-homogeneous Poisson process (NHPP) and an empirical power law model are employed for generation of corrosion defects over time and for defect growth, respectively. The time-dependent probability of failure is evaluated by employing the limit state function for burst under internal pressure. Internal pressure is modelled using a Poisson square wave process-based method (PSWP). Thereinafter, the study uses the aforementioned analysis, in conjunction with a heuristic model, in order to investigate the influence of imperfect repairs on the reliability prediction of the pipeline system. The heuristic method is known as split system approach and it is used to describe the changes in reliability of the pipeline system over its lifetime with the consideration of multiple maintenance actions. A numerical example of the described methodology is presented based on a chosen maintenance policy. A parametric study is conducted to examine the impact of instantaneous generation rate of NHPP model, PSWP model and reliability control limit on the results of the numerical application. The proposed methodology aims to assist engineers in the decision making with regard to corrosion maintenance strategies.