AERO / HX-02 / CUSTODY REVIEW
STOPPED BEFORE SCORED CFDA real benchmark candidate, stopped at the input gate.
HX-02 is Kim et al.'s open-access two-stream water plate-and-shell experiment. Aero audited what an independent analyst can legitimately know before a blind prediction, reconciled the visible dimensions, and preserved the experimental reference separately.
Disposition: INELIGIBLE_FOR_SCORED_REPRODUCTION
What the source defines
Two water streams, plate-side hot flow, shell-side cold flow, circular plate / chevron channels, operating envelopes, energy-balance and counter-current LMTD definitions, uncertainty specifications, and a T1-T20 plate-sensor layout.
Why the blind run is not authorized
| Open input | Engineering consequence |
|---|---|
| 3-D header/manifold topology | Core and port losses/flow distribution cannot be reconstructed. |
| Pressure-tap stations | A CFD pressure span could use the wrong measurement plane. |
| Thermal bulk-temperature definition | Heat duty and CHT comparison would be underdetermined. |
| Plate/shell material properties | Wall-temperature prediction would require an invented model. |
This is an input-eligibility stop, not a solver failure. No Luna analysis, analytical prediction freeze, CFD, experimental reveal, or validation claim exists for this stage.
Deterministic reconstruction check
The rounded source dimensions reproduce an effective area within about 2.10% and hydraulic diameter within about 2.56% of the printed values. Aero preserves the printed values and records the derived values as a self-check; it does not silently substitute one for the other.
Next legitimate action
A human custodian must supply or approve the missing topology, measurement planes, thermal definitions and material data. Once those are reconciled, a new additive stage may freeze Luna's plan and analytical predictions before any CFD. The sealed experimental reference remains private until that blind freeze.