Pre-registration, Version 4: A Water-Trained π Look-Up Table for Critical Heat Flux, Blind-Tested on R-134a and R-12 Against the Scaled 2006 Look-Up Table

Pre-registration of a blind test, frozen on 9 October 2026 before either test article was opened. The 2006 water look-up table predicts critical heat flux (CHF) in another fluid only through a scaling map, and the Ahmad (1973) and Katto (1978) maps disagree; in the fluid-to-fluid test of version 3 (doi:10.5281/zenodo.23250748) they scored 45.1% and 34.5% relative RMS error on the same R-134a points, and version 3 failed at 48.2%. Version 4 is a smooth table of the water CHF data in four dimensionless groups of the fluid’s own properties (density ratio, Weber number, equilibrium quality, heated length over diameter): a Gaussian-weighted local linear fit of the logarithm of the boiling number, read at any fluid’s own groups, with no refrigerant data and no scaling map. The form and bandwidths were chosen in an exploratory development on R-134a data already seen, disclosed in full with its failed rounds; on those data version 4 scored 17.1% against 19.0% for the Katto-scaled table, which is not evidence. The test articles are Vijayarangan, Jayanti and Balakrishnan (2006), R-134a, and Katto and Yokoya (1987), R-12. H4: version 4’s relative RMS error on the pooled primary points is lower than that of the better scaled table. H4b: it is no more than 2.0 percentage points above it. Predictions at fixed conditions are published, and so is a defect found before freezing: at mid and high pressure version 4’s CHF rises with quality above about 0.55, where the table’s falls. The deposit holds the binding protocol, the test code with hash and once-only checks, the frozen model and its 24,156 water rows, the controls, two audits by a separate AI agent that had not written the package, the changes made in response, the development scripts and logs, and the R-12 development transcriptions. The test articles will be obtained through interlibrary loan after this record is public. Archive SHA-256: a0f1106db76b8d9372fb3b724eac9f508a61f2c95d8231ab3c38b9b9434e137a

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-10-09
DOI
https://doi.org/10.5281/zenodo.23254782
Primary Topic
Heat Transfer and Boiling Studies
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
preprint

Pre-registration, Version 4: A Water-Trained π Look-Up Table for Critical Heat Flux, Blind-Tested on R-134a and R-12 Against the Scaled 2006 Look-Up Table

Brian Guarino
Zenodo (CERN European Organization for Nuclear Research)
Heat Transfer and Boiling Studies
preprint

Pre-registration, Version 4: A Water-Trained π Look-Up Table for Critical Heat Flux, Blind-Tested on R-134a and R-12 Against the Scaled 2006 Look-Up Table

Brian Guarino
preprint en

Abstract

Pre-registration of a blind test, frozen on 9 October 2026 before either test article was opened. The 2006 water look-up table predicts critical heat flux (CHF) in another fluid only through a scaling map, and the Ahmad (1973) and Katto (1978) maps disagree; in the fluid-to-fluid test of version 3 (doi:10.5281/zenodo.23250748) they scored 45.1% and 34.5% relative RMS error on the same R-134a points, and version 3 failed at 48.2%. Version 4 is a smooth table of the water CHF data in four dimensionless groups of the fluid’s own properties (density ratio, Weber number, equilibrium quality, heated length over diameter): a Gaussian-weighted local linear fit of the logarithm of the boiling number, read at any fluid’s own groups, with no refrigerant data and no scaling map. The form and bandwidths were chosen in an exploratory development on R-134a data already seen, disclosed in full with its failed rounds; on those data version 4 scored 17.1% against 19.0% for the Katto-scaled table, which is not evidence. The test articles are Vijayarangan, Jayanti and Balakrishnan (2006), R-134a, and Katto and Yokoya (1987), R-12. H4: version 4’s relative RMS error on the pooled primary points is lower than that of the better scaled table. H4b: it is no more than 2.0 percentage points above it. Predictions at fixed conditions are published, and so is a defect found before freezing: at mid and high pressure version 4’s CHF rises with quality above about 0.55, where the table’s falls. The deposit holds the binding protocol, the test code with hash and once-only checks, the frozen model and its 24,156 water rows, the controls, two audits by a separate AI agent that had not written the package, the changes made in response, the development scripts and logs, and the R-12 development transcriptions. The test articles will be obtained through interlibrary loan after this record is public. Archive SHA-256: a0f1106db76b8d9372fb3b724eac9f508a61f2c95d8231ab3c38b9b9434e137a

Zenodo (CERN European Organization for Nuclear Research)
Heat Transfer and Boiling Studies
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.