Accepting a Wind Tunnel: How to Verify It Actually Meets Spec
How flow quality is actually measured at handover, which tolerances are realistically checkable, and what an acceptance protocol needs to protect both sides.
A number in a contract — "turbulence intensity below 0.5%" — means nothing until both sides agree on exactly how that number gets measured: where in the test section, with what instrument, at what speed, averaged over what time. Acceptance testing is the moment that agreement either already exists on paper, or gets argued out in real time at handover, with money on the line and a facility standing idle either way.
1. What actually gets measured, and where
Flow quality is not a single number for the whole test section — it varies point to point, which is exactly why acceptance measurement uses a traverse : an instrument moved through a defined grid of positions across the working plane, rather than a single fixed probe. At each grid point, the measurement typically covers:
- Mean velocity , to confirm uniformity across the section — how much the flow speed varies from point to point at a given fan setting.
- Turbulence intensity , usually via hot-wire anemometry, since it resolves the fluctuating component of velocity that a simple Pitot-static probe cannot.
- Flow angularity — pitch and yaw relative to the tunnel axis — since a model sitting in flow that isn't parallel to the centreline sees an effective angle of attack it wasn't supposed to have.
None of this is measured once. It has to be repeated across the speed range the contract specifies, because flow quality at the top speed and at a low-speed coaching condition are not the same measurement, and a facility that only proves itself at one operating point hasn't been fully verified.
2. Which tolerances are actually checkable — and why the target has to match the tunnel's purpose
Not every tunnel is built to the same flow-quality target, and acceptance criteria have to reflect that rather than borrowing a number from a different class of facility. An anemometer calibration tunnel, for instance, is commonly specified around axial turbulence intensity below 2% and flow uniformity to within 0.2% across the instrument area — tight numbers, but not the same as the sub-0.1% turbulence figures that low-turbulence aerodynamic research tunnels are built around. Writing a calibration-grade tolerance into a research-tunnel contract, or vice versa, sets a target that either wastes money over-delivering or under-specifies what the facility actually needs to do its job.
The acceptance tolerance has to be derived from what the tunnel is actually for — the same needs-analysis question that should have shaped the concept design in the first place — not copied from a different project's contract.
3. What a protocol that protects both sides actually contains
- The methodology is agreed before commissioning starts , not improvised at handover. Instrument type, grid density, measurement plane locations, speed points and averaging time should all be contract exhibits, not on-site decisions.
- Raw data accompanies the summary report. A pass/fail statement without the underlying measurement grid gives the buyer nothing to independently verify and gives the builder nothing to defend against a dispute.
- A named reference standard for each instrument , with traceable calibration, so a turbulence or velocity reading isn't just a number from an uncalibrated sensor — this is the same traceability principle that governs accredited calibration tunnels, applied here to the acceptance instrumentation itself.
- A witnessed measurement, where the stakes justify it. For larger or higher-value facilities, having the buyer's own engineer or a third party present for the acceptance run removes any question about whether the numbers were generated honestly.
- A clear sign-off structure , specifying who has authority to accept on the buyer's side and what happens procedurally if a measurement falls outside tolerance — remeasurement, correction and retest, or a negotiated resolution — decided in advance rather than improvised in a tense moment.
4. Why this is worth getting right before commissioning, not during it
By the time a facility is powered up for acceptance testing, both sides have a strong incentive to see a pass: the buyer wants a working facility, the builder wants to close the project. That shared incentive is exactly why the methodology has to be locked down earlier, when neither side yet knows which way an ambiguity will cut. A protocol negotiated calmly during contracting protects both parties far better than one improvised under the pressure of a facility that's already built and a schedule that's already slipping.
5. What this means for your programme
Tell us what the facility needs to prove, and to whom , and we'll help you build an acceptance protocol — methodology, grid, tolerances and sign-off structure — agreed before a single fan blade turns, not negotiated at handover.
Sources and notes
- Flow-quality measurement methodology (traverse-based grid measurement across the working plane; mean velocity for uniformity; hot-wire anemometry for turbulence intensity; flow angularity relative to the tunnel axis; repetition across the full speed range rather than a single operating point) is TunnelTech engineering practice.
- Calibration-tunnel flow-quality figures used for comparison (axial turbulence intensity below 2% and flow uniformity to within 0.2% for accredited anemometer calibration tunnels, distinct from the sub-0.1% turbulence figures used in low-turbulence aerodynamic research facilities): consistent with our published article on
.
- Acceptance-criteria contract structure (the categories that belong in a specification, from speed range to instrumentation accuracy) is developed in our published article on
; this article extends that criteria table into the measurement methodology and protocol behind it.