A priori analysis of the performance of cross hot-wire probes in a rough wall boundary layer based on stereoscopic PIV

Error in measuring Reynolds shear-stress in turbulent boundary layer flows over a rough surface with a cross hot-wire has been reported in the literature and attributed to the existence of strong ejection and sweep motions that cause rectification, e.g. deviation of the velocity vector angle outside the acceptance cone of these probes. Using stereoscopic particle image velocimetry measurements and the concept of effective cooling velocities, the objective of the present study is to perform an a priori analysis of the cause of errors occurring when employing cross hot-wire anemometers. Besides the above-mentioned rectification effect, the role of the non-measured component is investigated. It is shown to be responsible for a non-negligible underestimation of the measured velocity variances and Reynolds shear-stress. This often overlooked source of error is intrinsic to turbulent flows and not limited to flow over rough walls.

Error analysis on the instantaneous shear-stress in the roughness sublayer of a rough boundary-layer measured by a (a) generic XHWA modeled using SPIV data : (c) influence of the non-measured component v, and relationships with Q-events (b) without and (d) with the non-measured component influence.