In real life, airflow in a turn is curved, how does a wind tunnel simulate this condition?

If we just yaw the test object at some angle in wind tunnel, do we simulate the airflow in correct way?


  • $\begingroup$ Highly related What is airflow direction in turn? $\endgroup$ – Manu H Aug 1 '20 at 12:25
  • $\begingroup$ Definately not a duplicate question here! $\endgroup$ – Stuart Buckingham Aug 6 '20 at 16:44
  • $\begingroup$ The flowpaths that you have indicated are correct, and impossible to replicate in a stationary WT. According to the non-inertial Eularian frame of reference (attached to the body), there are two 'fictitious forces' that affect the flow: Centrifugal and Coriolis, that cannot be generated in a stationary WT with a stationary model. Dynamically rotating the model can work, but then you have other dynamics such as yaw acceleration, which will have an effect on the flow, let alone how difficult it would be to separate the dynamic forces from noisy aerodynamic forces $\endgroup$ – Stuart Buckingham Aug 6 '20 at 16:58
  • $\begingroup$ Because the flow is no longer steady with a rotating/oscillating model, the streamlines will be the same as the flowlines you drew, but the pathlines will be VERY different. Flow from the front of the vehicle will be nowhere near where it would be in a real yawing condition. The only example of replicating true cornering has been in a research setting - youtube.com/watch?v=9oD7Jk7mm8o . In automotive (which see much higher yaw rates than aircraft), true cornering is just estimated by having the model yawed to the incoming flow. $\endgroup$ – Stuart Buckingham Aug 6 '20 at 17:00

Wind tunnel is a tool to interrogate aerodynamics, not to investigate flight dynamics (which relies on aerodynamic results to predict vehicular dynamics). As explained in this answer, the additional aerodynamic effects from induced airflow due to rotation can be lumped into dynamic derivative coefficients. Some examples include:

  • Pitch damping: $C_{m_q}$
  • Lift due to pitch rate: $C_{L_q}$
  • Roll damping: $C_{l_p}$
  • Cross yaw derivative: $C_{n_p}$

etc. Note that the rates in these derivatives are expressed in dimensionless rates to maintain similitude, which is accomplished by dividing the body rates by $bV_\infty/2$, where $b/2$ is half span and $V_\infty$ is freestream speed.

To estimate dynamic derivatives in wind tunnel, the forced-oscillation method is typically applied. The model is excited in a single axis by a small amplitude sinusoid at representative range of dimensionless rates and reduced frequency. This paper provides a good introduction to the data collection and analysis.

Rotary balance tests can also performed in wind tunnel to measure the effect of sustained rotary motion whose angular velocity is coincident with the air velocity vector. These tests are mainly to explore spin characteristics, which occur at very high angles of attack and sideslip. Forced-oscillation method would not be adequate in these conditions.

  • $\begingroup$ cabrillo.edu/~jwelch/bernoulli/bernoulli.wmv ,Can you explain why flat plate in this test dont cause low pressure,you can post answer at my topic: "Why high air speed at flat surface dont allways produce low pressure?" $\endgroup$ – user50657 Aug 7 '20 at 16:17