We ran a gas pipeline report against DNV-RP-F123

A design report checked against DNV-ST-F101 for gas, then DNV-RP-F123 for hydrogen. What hydrogen adds: 26 findings, 18 of them critical.

By Leo Ruspini — Nuclear engineer, PhD · Co-founder & CEO, normify.ai

In March 2026 DNV published DNV-RP-F123, its recommended practice for hydrogen pipeline systems. It covers new pipelines and, for most operators the bigger question, existing gas pipelines being requalified for hydrogen.

Requalification starts with the paperwork. A gas pipeline comes with a design report, material certificates, weld procedures and years of inspection records, all written for natural gas. Before it carries hydrogen, someone has to work out what those documents can still prove.

So we tested it.

The test

We wrote a design report for a fictional 18-inch subsea gas export pipeline, Cormorán, off Tierra del Fuego: X65 seam-welded pipe, walls of 14.3 and 15.9 mm, 130 barg maximum operating pressure, 32 km long and in service since 2013, with inspection history, a dent from trawl gear and two girth-weld signals from the last in-line inspection. Revision 4 of the report states that it's the basis for converting the line to up to 100 % hydrogen.

We checked it twice:

Keeping it fair

A 10-page summary report never carries every piece of evidence a standard asks for. The F101 check already found 46 things to complete for natural gas: transient hydraulics, overpressure protection, flaw sizing, the dent assessment, seal qualification, integrity management and more.

Those gaps aren't caused by hydrogen, so we don't count them here. Of the 43 findings in the F123 check, we removed the 17 that the F101 check had already raised. What's left is what hydrogen adds.

What hydrogen adds

26 findings, 18 of them critical.

Undocumented doesn't mean the pipeline is wrong. It means nobody has shown it's right for hydrogen yet.

Where the gas-era report conflicts with the RP

Fluid category and safety class (§2.4). The report takes its safety classes from F101 for natural gas, fluid category D. Hydrogen is category E. And for walls under 20 mm in location class 2, which both of Cormorán's landfall and platform sections are, the RP points to safety class very high unless a hydrogen risk assessment shows the risk is equivalent to natural gas. The report says high.

Every safety factor that follows (§2.4). Pressure containment, buckling and fatigue checks all use the factors of the gas safety classes. Change the class and every utilisation in the report has to be recalculated.

Fatigue (§6.2, §5.4.2). The report checks the seam welds with an S-N curve in air and Miner summation, and passes comfortably for gas. The RP has no S-N approach for hydrogen: it asks for crack growth by fracture mechanics, with hydrogen data. Same pressure cycles, different method, different answer.

Gas detection (§2.3.9, §8.5.3). The detectors at the landfall station are calibrated for methane.

Requalification itself (§9.1). The report presents itself as the basis for hydrogen service, but its integrity conclusion covers natural gas to 2043. The requalification steps the RP lays out (safety assessment, structural integrity with hydrogen as the new fluid) aren't done.

What the report doesn't document

The 21 undocumented findings read like a requalification work list. The main ones:

What this means

The pipeline didn't change. The steel didn't change. The rulebook did, and a report written for the old one can't answer most of the new questions.

The RP itself says its hydrogen safety factors are not yet fully calibrated. So read these findings as a work list, not a verdict on the pipe: what has to be tested, calculated or documented before it carries hydrogen.

That's the check normify runs: a document against the norm it now has to meet, clause by clause. Each finding comes with the clause, the passage in the document, and a recommendation. Where it holds, where it conflicts, where it's undocumented.

One run per norm, on a fictional document we wrote. Counts vary between runs; the findings above are from the reports of 13 October (DNV-RP-F123) and 19 October 2026 (DNV-ST-F101).

Sources