PHMSA’s July 2026 pipeline-repair proposal leans on “proven, data-driven modern engineering concepts” and “Modern, Proven Engineering Modeling”, signaling a shift toward risk-based scheduling.
It starts with two words that stand out: “modern” and “proven.” The NPRM (Notice of Proposed Rulemaking) promises to apply “proven, data-driven modern engineering concepts” and specifically highlights “Modern, Proven Engineering Modeling” to determine where and when repairs are needed. Integrity teams have used advanced inspection data and modeling for years, so what’s changing is how much regulatory weight those models carry.
Instead of one-size-fits-all deadlines, PHMSA envisions scheduling fixes by actual threat and remaining safe life. But if an analytical model can set a repair date, its assumptions become crucial.
That one word “proven” is key. Does it mean validated by field performance?
Independently verified? An AMPP or industry consensus? Who decides… operators, PHMSA, or both?
For example, under the proposal, hazardous liquid pipeline operators using an Engineering Critical Assessment (ECA) for dents must now submit their ECA procedures to PHMSA (via a no-objection notification). PHMSA estimates ~710 operators would each spend about 400 hours preparing that documentation. That concrete detail shows models shifting from in-house tools to regulated requirements.
The technology isn’t new but giving it regulatory authority is. As we move from prescriptive to predictive safety approaches, defining what makes a model “proven” is the next frontier.
What do you think? How should “proven” be defined for integrity models influencing regulations?