Assessing damaged pipelines transporting hydrogen

• Published in: Journal of Pipeline Science and Engineering Vol. 2; no. 3; p. 100066
• Main Authors: Andrews, R.M., Gallon, N., Huising, O.J.C.
• Format: Journal Article
• Language: English
• Published: KeAi Communications Co. Ltd 01.09.2022
• Subjects: Cracks; Damage assessment; Dents; Fitness for service; Hydrogen transport; Volumetric corrosion
• ISSN: 2667-1433; 2667-1433
• DOI: 10.1016/j.jpse.2022.100066

There is worldwide interest in transporting hydrogen using both new pipelines and pipelines converted from natural gas service. Laboratory tests investigating the effect of hydrogen on the mechanical properties of pipeline steels have shown that even low partial pressures of hydrogen can substantially reduce properties such as reduction in area and fracture toughness, and increase fatigue crack growth rates. However, qualitative arguments suggest that the effects on pipelines may not be as severe as predicted from the small scale tests. If the trends seen in laboratory tests do occur in service, there are implications for the assessment of damage such as volumetric corrosion, dents and mechanical interference. Most pipeline damage assessment methods are semi-empirical and have been calibrated with data from full scale tests that did not involve hydrogen. Hence the European Pipeline Research Group (EPRG) commissioned a study to investigate damage assessment methods in the presence of hydrogen. Two example pipeline designs were considered, both were assessed assuming a modern, high performance material and an older material. From these analyses, the numerical results show that the high toughness material will tolerate damage even if the properties are degraded by hydrogen exposure. However, low toughness materials may not be able to tolerate some types of severe damage. If the predictions are realistic, operators may have to repair more damage or reduce operating pressures. Furthermore, damage involving cracking may not satisfy the ASME B31.12 requirements for preventing time dependent crack growth. Further work is required to determine if the effects predicted using small scale laboratory test data will occur in practice.