CALIBRATION

In all wind tunnels, the drag of the walls, floor and ceiling tends to slow the stream down at the edges and the walls have their own boundary layer characteristics, introducing errors into measurements taken near them. Before a tunnel is used, it has to be established by careful testing that the flow speed is even through and across the whole test section.

A measurement of particular importance for low speed work is the turbulence factor of the tunnel. Since so much depends on the boundary layer and its transition from laminar to turbulent flow, any small, microscopic turbulence in the tunnel will have a disproportionately large influence on the drag of the aerofoils under test In serious test work, the tunnel turbulence factor is reported and, to allow a very rough correction to be made, the Reynolds number of the test may be multiplied by this factor to yield an equivalent Re. It is also found that tunnels tend to have somewhat different turbulence factors at different flow speeds, so strictly a whole spectrum of turbulence measurements should be made. This is not often done.

Before F. W. Schmitz could carry out his classic tests on wing profiles at Cologne (see Chapter 8 and Appendix 2), he had to work for more than a year to improve the wind tunnel. He reduced the turbulence factor to 1.06. Modem tunnels should be better than this. Schmitz’s results are probably reliable to within 6% of the stated Reynolds numbers.

A few years earlier, the NAC A in America had carried out an extensive series of tests in the compressed air tunnel, down to Reynolds numbers of42,000. The results published in NAC A Report 586 covered all the most popular NACA four digit profiles, the 6409, 4409, 4412, etc. and are still quite often quoted by writers in modelling magazines and presumably are used in designing some models. Unfortunately, as the NACA authors reported at the time, the turbulence factor was 2.64, which implies that the stated Re of each test should be multiplied by this figure to arrive at a better but still very crude approximation to the truth. In other words, the Re of 42,000 of these tests (apparently well within the free flight modelling range), represents a true Re of 110,880, which takes these test results above the usual critical Re for most aerofoils. NACA Report No. 586 is not in general of much value to modellers. Other tunnel tests have suffered from the same difficulties, though when published, the turbulence factor is not always stated, so that not even the crudest correction can be attempted. Modellers should not take seriously any wind tunnel results which are quoted or published if the turbulence factor is not known.