Enhancing the accuracy of measurement techniques in high Reynolds number turbulent boundary layers for more representative comparison to their canonical representations

Existing differences between experimental, computational and theoretical representations of a particular flow do not allow one-to-one comparisons, prevent us from identifying the absolute contributions of the various sources of uncertainty in each approach, and highlight the importance of developing suitable corrections for experimental techniques. In this study we utilize the latest Pitot tube correction schemes to develop a technique which improves on the outcome of hot-wire measurements of mean velocity profiles in ZPG turbulent boundary layers over the range 11 500 < Re-theta < 21 500. Measurements by Bailey et al. (2013), carried out with probes of diameters ranging from 0.2 to 1.89 mm, supplemented by other data with larger diameters up to 12.82 mm, are used first to develop a somewhat improved Pitot tube correction which is based on viscous, shear and near-wall schemes (which contribute with around 85% of the effect), together with a turbulence scheme which accounts for 15% of the whole correction. The correction proposed here leads to similar agreement with available high-quality datasets in the same Reynolds number range as the one proposed by Bailey et al. (2013), but this is the first time that the contribution of the turbulence scheme is quantified. In addition, four available algorithms to correct wall position in hot-wire measurements are tested, using as benchmark the corrected Pitot tube profiles with artificially simulated probe shifts and blockage effects. We find that the kappa B-Musker correction developed in this study produces the lowest deviations with respect to the introduced shifts. Unlike other schemes, which are based on a prescribed near-wall region profile description, the kappa B-Musker is focused on minimizing the deviation with respect to the (kappa) over tilde(B) over tilde relation, characteristic of wall-bounded turbulent flows. This general approach is able to locate the wall position in probe measurements of the wall-layer profiles with around one half the error of the other available methods. The difficulties encountered during the development of adequate corrections for high-Re boundary layer measurements highlight the existing gap between the conditions that can be reproduced and measured in the laboratory and the so-called canonical flows. (C) 2015 Elsevier Masson SAS. All rights reserved.