Refinement of Reliability-Based Safety Factors for Corrosion Assessment

Abstract Reliability-based safety factors for corrosion assessment have been developed for onshore natural gas and low vapor pressure (LVP) pipelines and published in 2018. These safety factors were developed using the limit states design (LSD) approach and are designed to achieve consistent safety levels for a broad range of pipeline designs and corrosion conditions. The factors were calibrated for two corrosion limit state categories: ultimate limit state, representing large leaks and ruptures; and leakage limit state, representing small leaks. They can be applied with a deterministic limit state equation (such as the corrosion burst pressure model) and a set of calibrated characteristic values of key input parameters for the equation (such as wall thickness and feature depth) to assess the acceptability of a corrosion feature. With this set of reliability-based safety factors, a corrosion feature assessment can be carried out deterministically while still achieving a consistent risk level. For the ultimate limit state, the safety factors were calibrated for a safety class system that is aligned with the six safety classes described in the Annex C of the 2023 edition of CSA Z662. Refinement of the previous formulation of the safety factors includes: 1) addition of more pipeline cases in the calibration to expand the applicability of the safety factors; 2) addition of the pipeline material strength in the safety factor calculation equations; and 3) further calibration of the safety factors to cover both the Annex O and Effective Area corrosion burst pressure models. These refinements were developed for natural gas pipelines, and LVP liquid pipelines are not considered in this paper. To ensure that these safety factors lead to practical solutions for real cases, the performance of the reliability-based safety factors was compared to that of the CSA Z662 Clause 10.10.2.5.1 and the full probabilistic approach of TC Energy (“TCE”). The comparison was based on the number of corrosion repairs required by each method and was conducted using a comprehensive set of TCE pipeline cases that covered a wide range of pipeline diameters (NPS 6 to 42), hoop stress factors (0.23 to 0.80), and corrosion severity (feature density from < 1 to > 5000 per km).