Where the wind tunnels are used to experimentally simulate conditions in flight, the measurement results have to be corrected to account for the difference between conditions in the wind tunnels and the free flight. This is not necessarily the case when phenomenological research is concerned. Here in the discussion of interferences we are not considering the scale effects that enter the non-dimensional similarity parameters, like the Reynolds number effect or any other similarity containing the length scale.
There are two principal categories of the differences to the free flight that have to be accounted for – the imperfections of the flow quality in the wind tunnel and the constraints to the air flow by the tunnel walls and the model suspension.
In order to quantify and correct for the first category, the empty test sections of the wind tunnels operated by DNW are all thoroughly calibrated. These calibrations typically yield, in addition to the information about the local flow quality, also information about non-uniformity of the flow along the test section. The latter can again be due to the flow constraints as stated above. Based on experience, the calibrations are repeated on regular intervals, possibly after configuration changes, depending on the sensitivity. The calibration data can be made available to the users of the wind tunnels in order to facilitate a joint evaluation of measurement results in addition to its being used by the operators.
The second category of differences to free flight stems from the fact that the wind tunnel is a bounded environment with walls or shear layers around the test section in the far field of the flow and the model has to be held in place by the model mount. The model mount can further be split into the sting attached to the model (belonging to the near field) and the sword supporting the sing at the other end (belonging to the far field). The influence of the far field components on the free stream flow in the empty test section is captured in the calibration process.
To correct for the wall and support interferences in the test, a number of methods are available. Most of these are described in the open literature, such as the AGARDograph 336 and conference proceedings. The most frequently used corrections are those based on the potential flow approximation and small perturbation theory. The last one means that the corrections can be linearly superposed on the measured values.
For correcting for the presence of boundaries of the flow in the wind tunnel, the potential flow approach, also known as the classical approach, works best in wind tunnel configurations with closed wall test sections. More care has to be exercised when the testing is performed in a free jet (also: open jet) test section, since the circulation caused by a lifting body affects the free jet and the quiescent air in the plenum. This transition from the free jet to the plenum air cannot be simply modeled in the classical approach.
Corrections for partially adapting slotted walls become even more challenging. DNW has refined its methodology over the years using a combination of numerical calculations and reference model tests providing increased confidence in the results.
A slightly different approach can be taken when the wind tunnel walls are solid but flexible. Then the wall shape can be adapted in such a way as to minimize the wall induced perturbations to the free stream flow at the tunnel centerline, the usual location of the model.
In addition to the wall constraints, the influence of the model mounting mechanism has to be quantified separately as well. In addition to the pressure distribution obtained in the empty test section, the interaction of the model support and the model has to be accounted for. in order to do this, a choice from a combination of different mounts, one of which is load carrying and another one not (called “dummy sting” in DNW) can be employed to quantify the contribution of any variant of the model suspension in the wind tunnels. The loads are measured by the standard wind tunnel balance. The quantification of the interference effect of a sting variant is achieved by simple subtraction of two load measurements, one with the “dummy sting” and another one without. The methodology is well documented and publications with a lot of convincing detail are available.