Turbulence measurements in a compressible reattaching shear layer
Previously cited in issue 05, p. 591, Accession no. A83-17917
Engineering topics
Publications and source records attributed to Hayakawa, K..
Previously cited in issue 05, p. 591, Accession no. A83-17917
Previously cited in issue 15, p. 2345, Accession no. A82-31948
The performance of a constant-temperature normal hot-wire in a supersonic flow is critically examined. It is shown that this instrument is inherently unsuitable for measuring turbulent temperature correlations because of the highly nonlinear response to temperature fluctuations, particularly at low overheat ratios. The instrument is therefore limited to measurements of mean and fluctuating mass-flow rates. Suitable calibration procedures, as well as the limits on spatial and temporal resolution are discussed, and corrections for mean stagnation temperature changes are suggested.
Detailed hot-wire measurements of the longitudinal component of the mass-flow fluctuations have been made in an approximately self-preserving free shear layer reattaching on a 20 degree ramp at Mach number of 2.9. The experimental configuration is especially designed to provide a well-defined initial condition for the reattachment process. The absolute mass-flow turbulence intensity increases dramatically through the compression in the reattachment region, which is in sharp contrast to similar subsonic reattachments. It is clear that the mean dilatation contributes significantly to the turbulence amplification. In addition, the length scale is affected strongly by the presence of extra strain rates. Prediction of this flow will require some sophisticated modelling, and the challenge to the predictor is clear.
Measurements of the longitudinal component of the mass-flow fluctuations have been made in an 8-deg compression corner flow and a reattaching shear layer at a Mach number of approximately 2.9 and unit Reynolds number of about 7 x 10 to the 7th per m. The data include turbulence intensities and probability density distributions. Significant turbulence amplification occurs in both interactions. A qualitative explanation in terms of direct shock effects, extra strain rates, and their relative contribution, is suggested.