Effects of heterogeneous surface geometry on secondary flows in turbulent boundary layers

The effect of spanwise wavelength of surface heterogeneity on turbulent secondary flows

“…Recent studies showed that the size, spanwise location and direction of the secondary flows may be affected by the specific configuration of spanwise heterogeneity (Goldstein & Tuan 1998;Wang & Cheng 2006;Türk et al 2014;Vanderwel & Ganapathisubramani 2015;Stroh et al 2016;Awasthi & Anderson 2018;Chung et al 2018;Hwang & Lee 2018;Medjnoun et al 2018;Yang & Anderson 2018;Medjnoun et al 2020;Zampiron et al 2020). For strip-type roughness, the spanwise roughness wavelength Λ = 2S (figure 1a) seems to determine the behaviour of the secondary flows.…”

“…The recessed part of previously investigated ridge-type roughness is typically a flat surface, while the elevated part has many possible cross-sectional shapes: rectangular (Wang & Cheng 2006;Hwang & Lee 2018;Medjnoun, Vanderwel & Ganapathisubramani 2018, trapezoidal (Nezu & Nakagawa 1984), triangular (Goldstein & Tuan 1998;Medjnoun et al 2020;Zampiron, Cameron & Nikora 2020), semicircle (Medjnoun et al 2020), streamwise-aligned pyramids (Yang & Anderson 2018) and smoother, sinusoidal-like spanwise elevation variations (Colombini 1993;Wang & Cheng 2006;Awasthi & Anderson 2018). Studied surfaces often involve a combination of both ridge-type and strip-type roughnesses (Vanderwel & Ganapathisubramani 2015;Vanderwel et al 2019;Stroh et al 2020).…”

Channel flow with large longitudinal ribs

“…at z = h). This will also be explored here, but it is worth noting immediately that the profiles obtained at specific spanwise locations in laboratory experiments also lie below the regular log-law, as shown by Medjnoun et al (2020) and Zampiron et al (2020). It is much more difficult to obtain spanwise-averaged profiles below z = h in such experiments and that was not attempted in these latter works.…”

“…Medjnoun, Vanderwel & Ganapathisubramani 2018;Yang & Anderson 2018;Stroh et al 2020). For flows with longitudinal ribs, apart from the boundary layer DNS of Hwang & Lee (2018) and the channel large eddy simulation (LES) of Yang & Anderson (2018), there appears to be only the study containing data from (channel) DNS and a laboratory boundary layer (Vanderwel et al 2019) and, even more recently, wind tunnel boundary layer studies by Medjnoun, Vanderwel & Ganapathisubramani (2020) and a similar set of experiments, but in a water flume, by Zampiron et al (2020).…”

“…It is much more difficult to obtain spanwise-averaged profiles below z = h in such experiments and that was not attempted in these latter works. Medjnoun et al (2018Medjnoun et al ( , 2020 denote the profile offset by ΔU + in viscous units, in common with genuine rough-wall boundary layers; they call this the 'roughness function' but we will avoid this terminology, given that the boundary layer is not a genuinely rough-wall flow, as noted above. The present paper also discusses the spanwise-averaged turbulence stresses.…”