The Cryogenic Referent Registry
Ten referent records audited against six criteria: what each can and cannot support in cryogenic tank model validation.
Program 01 · Active · Cryogenic & Propulsion Modeling Systems
CPMS measures how well engineering models predict the behaviour of cryogenic propellant systems — and publishes the answer, including when the answer is that no one can tell.
Models that predict self-pressurization and boil-off in cryogenic tanks are used to size upper stages, plan orbital propellant transfer, and design liquid hydrogen storage for ships and aircraft. Many such models exist. Most are described as validated.
How much each one errs — and where it stops being supported by experimental evidence — is not systematically documented anywhere. Each study validates against its own subset of the public record, by its own method, against its own criteria. There is no common baseline, and no referee.
We evaluate models against the public experimental record and publish the comparison reproducibly. Every result is intended to be reconstructable by someone who does not trust us.
The public experimental evidence base — which experiments report measurement uncertainty, which publish usable data, and which validations can be reconstructed from the public record at all.
Published models, faithfully — stating exactly what was specified by the source, what we had to assume, and where our implementation deviates.
Error with the uncertainty that qualifies it: numerical, input, and the measurement uncertainty of the reference experiment itself. No number without its uncertainty.
A documentary audit and traceability study of the public experiment reports used to validate cryogenic tank self-pressurization models — which reports state measurement uncertainty, which publish usable data, and which can be reconstructed from the public record alone. Exploratory work on public documents; not conducted for or in association with NASA.
Ten referents were audited against six criteria fixed in advance, from the primary test reports rather than the papers citing them: K-Site, MHTB, SHIIVER, liquid nitrogen zero boil-off, Saturn AS-203, TPCE and ZBOT. 14 registry sources and 14 downstream-use documents — 28 artifacts in all — identified by SHA-256. The findings, the underlying records and an errata table of our own were published together.
Ten referents audited · Notes 000–003 and the supporting dataset published 12 August 2026 · five permanent identifiers, all resolving
Ten referent records audited against six criteria: what each can and cannot support in cryogenic tank model validation.
A 1992 liquid hydrogen dataset, five downstream validation documents through 2025, and the one piece of experimental context that still does not travel.
SHIIVER — the best-documented cryogenic tank test I have read — and what a modeler still cannot reconstruct from it.
Auditing the 161-page primary report beneath a liquid-hydrogen validation citation examined in this series.