Scientific software can produce a clean plot from the wrong interpretation.
That risk was central to research undertaken in an Airbus aerodynamic challenge context and recorded in the Haivian Research Suite. It is not evidence of Airbus deployment, conformity or endorsement. Before comparing solvers or proposing a quantum route, the programme had to make explicit an ambiguity in how a Taylor–Green vortex benchmark’s characteristic length was being used.
Two Reynolds-number profiles appeared to conflict. The research retained a more precise explanation: they could describe the same physical viscosity while using different characteristic-length conventions. The important work was preserving both defensible readings and designing a verification framework that could test them without pretending to know which convention the organizer intended. That organizer intent remained unresolved in the public evidence recorded for this candidate.
Same label and same physics are different questions
A Reynolds number is a ratio built from a velocity scale, length scale and viscosity:
Re = UL/ν
Here U is the velocity scale, L the characteristic length and ν the kinematic viscosity. If the declared length changes from L to 2L while U and ν stay fixed, the Reynolds-number label doubles even though the physical viscosity has not changed.
That creates two legitimate comparison questions:
- What happens when two methods use the same displayed Reynolds-number label?
- What happens when they use the same physical viscosity after reconciling the characteristic scales?
Those comparisons are not interchangeable. A benchmark report that mixes them can mistake a convention difference for a solver difference.
The public dual-profile foundation presents the programme’s two preserved profiles and their limitations. The public machine-readable projection provides the bounded milestone and claim state. Every result must remain tied to its declared profile; the public record does not establish which profile reflects organizer intent.
Verification before performance
The first accepted milestone was deliberately narrow: symbolic and dimensional consistency. It checked equations, dimensions, assumptions and deliberately altered cases within that declared scope.
That narrowness is valuable. It prevents a verified equation catalogue from becoming an implied claim of numerical accuracy, aerodynamic fidelity or industrial readiness.
The quantum question stayed downstream
The quantum question remained downstream of the verified benchmark foundation. No quantum target was admitted in the public programme state recorded for this candidate.
A future target would need evidence that it addressed a measured, decision-relevant bottleneck or representation opportunity. It would also need full accounting for loading, preparation, repetition, readout, reconstruction and the strongest relevant classical comparator.
That sequencing matters: profile semantics first, numerical evidence next, target admission later.
What the case teaches
The Airbus research illustrates a general method for scientific AI and hybrid-computing projects:
- Turn ambiguous conventions into named profiles rather than hidden assumptions.
- Separate same-label comparison from same-physics comparison.
- Verify equations, dimensions, assumptions and mutations before benchmarking performance.
- Bind every result to its profile, grid, time step, tolerance and claim ceiling.
- Admit an advanced-computing kernel only after classical evidence identifies a consequential bottleneck.
The most important contribution is not a new Reynolds-number identity. It is a benchmark discipline that stops a familiar ambiguity from contaminating every downstream comparison.
Evidence boundary
The public evidence was recorded for this candidate as at 2026-09-28. The reader-facing profile section and the machine-readable projection support the Airbus-specific state described here. The exploratory simulation download is not cited and cannot upgrade the accepted evidence. These sources do not establish Airbus endorsement, conformity, production readiness or quantum advantage.