In the present work we analyze three commonly used methods to determine the edge of pressure gradient turbulent boundary layers: two based on composite profiles, the one by Chauhan et al. (Fluid Dyn. Res. 41:021401, 2009) and the one by Nickels (J. Fluid Mech. 521:217–239, 2004), and the other onebased on the condition of vanishing mean velocity gradient. Additionally, a new method is introduced based on the diagnostic plot concept by Alfredsson et al. (Phys. Fluids 23:041702, 2011). The boundary layers developing over the suction and pressure sides of a NACA4412 wing section, extracted from a directnumerical simulation at chord Reynolds number Rec = 400, 000, is used as the test case, besides other numerical and experimental data from favorable, zero and adverse pressure-gradient flat-plate turbulent boundary layers. We find that all the methods produce robust results with mild or moderate pressure gradients, although the composite-profile techniques require data preparation, including initial estimations of fitting parameters and data truncation. Stronger pressure gradients (with a Rotta–Clauser pressure-gradient parameter β larger than around 7) lead to inconsistent results in all the techniques except the diagnosticplot. This method also has the advantage of providing an objective way of defining the point where the mean streamwise velocity is 99% of the edge velocity, and shows consistent results in a wide range of pressure gradient conditions, as well as flow histories. Collapse of intermittency factors obtained from a wide range of pressure-gradient and Re conditions on the wing further highlightsthe robustness of the diagnostic plot method to determine the boundary layert hickness (equivalent to δ99 ) and the edge velocity in pressure gradient turbulent boundary layers.
American Institute of Physics (AIP), 2016. Vol. 28