Model Wind Tunnel Experiment (MWTE)

Updated 8 January 2026

MWTE is a rigorously scaled experiment that models aerodynamic, aeroacoustic, and scalar transport phenomena using physical or numerical methods in a wind tunnel.
MWTE adheres to non-dimensional scaling laws, such as Reynolds, Froude, and Mach numbers, ensuring dynamic similarity between model and full-scale systems for accurate flow quantification.
MWTE employs advanced diagnostics, active grid control, and real-time data acquisition to validate computational models and inform design in engineering and environmental studies.

A Model Wind Tunnel Experiment (MWTE) is a rigorously scaled, instrumented, and controlled laboratory test using physical or numerical models in a wind tunnel to investigate aerodynamic, aeroacoustic, or scalar (e.g., pollutant or heat) transport phenomena. MWTEs enable the quantification and visualization of flow fields, forces, moments, scalar distributions, and multi-physics interactions that underpin atmospheric, engineering, and geoscientific applications. Test procedures require strict adherence to similitude criteria, advanced measurement techniques, and comprehensive data analysis protocols to identify controlling mechanisms, validate computational models, and inform design or policy.

1. Physical and Numerical Scaling Principles

The validity of any MWTE rests fundamentally on the correct application of non-dimensional scaling laws that preserve the ratios of dominant physical forces and transport mechanisms between the model and the full-scale system. Essential dimensionless parameters include:

Reynolds number: $Re = UH/\nu$ or $Re = Ud/\nu$ (inertial vs. viscous forces; $U$ = velocity, $H$ = reference length, $d$ = model dimension, $\nu$ = kinematic viscosity).
Froude number: $Fr = U/\sqrt{gH}$ (inertial vs. gravitational bouyancy).
Richardson number: $Ri = g\beta \Delta T H / U^2$ (buoyancy vs. shear, $g$ = gravity, $\beta$ = expansion coefficient, $\Delta T$ = temperature difference).
Mach number: $M = U/a$ (compressibility, $a$ = speed of sound).
Strouhal number: $St = fD/U$ (shedding frequency, $f$ = characteristic frequency, $D$ = diameter/length).
Peclet and Schmidt numbers: $Pe = Re\,Pr$ , $Sc = \nu/D_m$ (advection-diffusion of heat and mass).
Other problem-specific groups: drag/permeability coefficient for vegetation ( $\lambda$ ), Damköhler number for combustion, among others (Zhao et al., 2023, Makowiecki et al., 2023).

Scaling is rarely possible for all nondimensional groups; for sharp-edged urban flows, $Re$ -independence can often be obtained above thresholds (e.g., $Re_{crit}\sim2\times10^4$ ) (Zhao et al., 2023), whereas buoyancy-driven or compressible regimes require matching of multiple numbers such as $Ri$ , $Fr$ , $M$ as dictated by the underlying physics (Ahlefeldt et al., 2023).

2. Experimental System Design and Instrumentation

An MWTE encompasses a precision wind-tunnel facility, appropriately scaled models, advanced actuation (e.g., gust-generating vanes, active grids), and multi-modal diagnostics:

Facility: Closed- or open-circuit tunnels, custom contraction ratios, variable test-section geometry, slotted/solid walls for boundary control, and—in advanced setups—pressure- or cryogenically-modulated conditions to simultaneously reach target $Re$ and $M$ (Ahlefeldt et al., 2023).
Model fabrication: Scale factor $M$ set by desired geometric and dynamic similarity; use of 3D printing, precision machining, or modular construction; attention to surface finish or roughness for boundary-layer transition control (Ellingsen et al., 2023, Varanwal et al., 17 Aug 2025).
Flow modulation: Active grids (16-axis in (Kröger et al., 2021)), oscillating vanes for gusts (Manolesos et al., 19 Dec 2025), or sloping test-sections for gravity effects (Makowiecki et al., 2023).
Sensors: Hot wire/cold wire anemometry, multi-axis force/moment balances, high-speed PIV, pressure taps (30+ for high-resolution wall distributions), chemiluminescence, tracer diagnostics, and synchronized multi-array microphones for aeroacoustic testing (Fellini et al., 2022 Ahlefeldt et al., 2023).
Automation: Real-time control and data acquisition (DAQ) systems with synchronized multi-channel sampling, servo/stepper actuation, and advanced simulation-DAQ integration (e.g., real-time hybrid simulation in (Du et al., 21 Apr 2025)).

A typical complex test may involve the integration of all of the above, e.g., measuring 3D pollutant fields with FID sensors at over 1000 points in an urban canyon array, under controlled inflow, for several tree densities (Fellini et al., 2022).

3. Examples of MWTE Workflows Across Domains

3.1 Urban Environmental Flows

Test sections up to 12 m in length, 3.5 m width, and 2 m height are employed to house arrays of urban blocks at $H$ = 0.1 m (1:200 scale), with tree rows (plastic, measured aerodynamic porosity $\alpha_p$ ) systematically varied to study their effect on pollutant dispersion and ventilation (Fellini et al., 2022).
Injection of passive scalar (e.g., C $_2$ H $_6$ line source, $Q_{et}=0.2$ L/min ethane in 4 L/min air), $C^* = C U_\infty L_s \delta / Q_{et}$ normalization for systematic comparison.
Metrics such as volume-averaged concentration, bulk exchange velocity $u_d$ , non-dimensional ventilation $u_d/U_\infty$ are derived from spatial mapping and mass balance.

3.2 Aeroelasticity, Aeroacoustics and Gust Simulation

Free-rotation models (“MiRo”) use Cardan joints to enable full 3-axis rotation; stereo high-speed imaging yields sub-mm/0.1° precision in attitude estimation (Muller et al., 2023).
Aeroacoustic tests: 96-microphone arrays, high-dynamic-range A/D (16–24bit, 250 kHz), slotted/closed-wall comparison, CLEAN-SC algorithm for sparse 3D source mapping; separation of $Re$ and $M$ dependencies by pressure/temperature variation for full-scale validity (Ahlefeldt et al., 2023).
Gust generators: Four NACA 0015 vanes, servo-actuated ±20°, frequencies up to 20 Hz; customized motion law to minimize negative-peak-factor while sustaining gust ratio (waveform $u_g(t)$ , metrics $G,\,\mathrm{NPF}$ ) (Manolesos et al., 19 Dec 2025).

3.3 Turbulence and Replication of Realistic Atmospheric Fields

Reproducible turbulence realized via 16-axis active grids, time-series downscaling from atmospheric LiDAR, high-frequency actuation, full look-up table (LUT) calibration for each angle, DAQ at 20 kHz (Kröger et al., 2021).
Filtering and cross-covariance analysis ( $\rho_{ij}$ ) used to define reproducible time/length scales, guide selection of $f_\mathrm{act}$ for target structure sizes.

3.4 Stochastic Load and Uncertainty Quantification

Building wind-load simulation through MWTE-calibrated Proper Orthogonal Decomposition (POD) of pressure tap data ( $\sim$ 500 taps), cross-spectrum estimation, spectral representation method (SRM) for stochastic realization, explicit quantification of variance/correlation error, and optimal mode truncation for computational efficiency (Duarte et al., 2023).
Uncertainty propagation via polynomial chaos expansion or compressed sensing in inflow-driven variability scenarios (Detomaso et al., 2020).

4. Key Data Analysis, Validation, and Computational Integration

Robust MWTEs tightly couple experimental measurements with computational surrogates and in situ/in silico uncertainty analysis:

Decomposition methods: Bi-orthogonal/POD decomposition for pressure and velocity fields; separation of mean, primary, and higher modes (e.g., identification of dominant vortex-shedding frequencies or global load contributions) (Ellingsen et al., 2023).
Surrogate modeling: PCE or machine learning regressors (regression trees/neural networks) embedded into control/dynamics simulations (e.g., injector LWC/MVD modeling (Hernández-Hernández et al., 2024)), stochastic load generation (Duarte et al., 2023).
Hybrid control frameworks: Real-time adaptive control/estimation via extended/unscented Kalman filters, bidirectional simulation-physical interaction via UDP, time synchronization at ~1 ms (Du et al., 21 Apr 2025).
Validation metrics: Direct comparison of modeled vs. measured variables (drag, force, aerodynamic moments) within $<10\%$ in global coefficients, sub-degree resolution in angles, sub-percent error in wind-load SRM statistics, or $<5$ % error in RMS velocities (flow/gust field matching) (Muller et al., 2023 Manolesos et al., 19 Dec 2025 Duarte et al., 2023).

5. Best Practices, Limitations, and Application-Specific Considerations

MWTEs require discipline in similarity criteria, instrumentation, and analysis:

Maintain model blockage $<5\,\%$ ; verify $Re$ independence for flow regime of interest.
Systematically calibrate all diagnostics (anemometers, pressure transducers, microphones) across the actual parameter range; apply advanced background subtraction where possible.
Use multi-point (array) measurements for spatial coherence, BOD/POD for modal decomposition, and ensemble statistics for reproducibility (cross-covariance, filtering).
Adapt model geometry to match not only first-order statistics but also boundary-layer properties (e.g., roughness/artificial tripping to match supercritical regimes), though only global parameters—not local modal structure—may be reproducible at low $Re$ with artificial roughness (Ellingsen et al., 2023).
Hybrid and real-time control/identification approaches enable dynamic exploration of parameter space (variable mass, stiffness, damping), but demand synchronization, with explicit handling of delays ("covariance matching," predictor-corrector structures) (Du et al., 21 Apr 2025).

6. Representative Case: Urban Tree Effects on Street-Canyon Ventilation

A detailed realization (Fellini et al., 2022):

Large wind-tunnel (12 m×3.5 m×2 m); H/W = 0.5 square canyons, 2D block array, and synthetic model trees at variable densities.
Boundary-layer, turbulence characterization: $U_\infty=5$ m/s, $\delta=1.1$ m, $I=5$ –10%, $Re_H=1.2$ –3.3 $\times10^4$ .
FID detection grid, 1000+ points per config, sampling 2 min/point.
Ventilation defined by volume-integrated $C_\mathrm{vol}$ and $u_d/Q_{et}$ exchange; result: tree presence reorganizes scalar field 2D→3D, but bulk $u_d/U_\infty$ changes remain within 20% range, no monotonic trend with $S_T/H$ .
Implication: local tree-induced recirculation peaks do not straightforwardly predict changes in urban pollutant exposure at street level.

7. Future Directions and Methodological Trends

Increasing use of high-reproducibility turbulence via active grids, programmable gusts, or controlled inflow in advanced MWTE facilities.
Enhanced coupling with high-fidelity CFD/LES/URANS codes (including mesh deformation, transient gust, and combustion/ignition modeling).
Multi-physics integration: simultaneous measurement of flow, temperature, concentration, heat flux, emissions, and dynamic response—enabling validation and calibration of comprehensive city-, building-, vehicle-, or device-scale models.
Application to planetary/low-ambient $g$ and $P$ conditions (in situ simulation of Martian or exoplanetary surface flows) (Kruss et al., 2019).
Focus on rigorous uncertainty quantification, surrogate modeling (PCE, ML), and end-to-end workflow reproducibility.