Energetic scales in a bluff body shear layer

Abstract

A detailed experimental campaign into separated shear layers stemming from rectangular sections (having aspect ratios of 5 : 1, 3 : 1 and 1 : 1) was carried out at Reynolds numbers range between $1.34\times 10^{4}$ and $1.18\times 10^{5}$ based on the body thickness. Particle image velocimetry was used to locate the highest concentration of fluctuations in the velocity field and subsequent hot-wire measurements at those locations provided adequate spectral resolution to follow the evolution of various instabilities that are active within the separated shear layer. Similar to recent findings by this same group, the shear layer behaviour is observed to contain a combination of Reynolds invariant characteristics, including its time-averaged position, while other properties demonstrate clear Reynolds number dependency, including the spatial amplification of turbulent kinetic energy. Additional results here show that the ratio of side lengths of the body is a key parameter in revealing these effects. One reason for this is the level of coupling between modes of instability, which is evaluated using two-point correlation methods. These findings indicate that the separated shear layer on a bluff body is highly nonlinear. A specific set of scales responsible for these unique behaviours is identified and discussed, along with their relationship to other scales in the flow.

The Turbulent Flow over the BARC Rectangular Cylinder: A DNS Study

“…Their general features have been already described by Cimarelli et al (2018b). Once the flow becomes turbulent for x ≥ −2, the largest fluctuations are those of the streamwise component, in both the shear layer and the primary vortex, but the maxima of the three diagonal components are of comparable magnitude as also observed at larger Re (Moore et al 2019). The peak of uu is at (x, y) = (0.21, 0.97), i.e.…”

“…This suggests that at this Reynolds number the wake still influences the leadingedge region, as it happens at much lower Re (Hourigan et al 2001). The vortex-shedding frequency has been detected at these upstream sections also by Moore et al (2019) for Re = 13,400, but not by Rocchio et al (2020) at Re = 40,000; this suggests that the influence of the vortex shedding from the trailing edge on the first part of the shear layer gradually fades away as Re increases. Moving along the shear layer, a higher dominant frequency f ≈ 1.3 emerges, associated to the amplification of the velocity fluctuations in the shear layer (Cimarelli et al 2018b) owing to the Kelvin-Helmotz instability (Moore et al 2019).…”

“…For zero incidence they found St ≈ 0.115 which weakly increases with Re and with the intensity of the free stream turbulence. Moore et al (2019) experimentally found for Re = 1.34 × 10 4 St ≈ 0.1114 and L 1 ≈ 4.4; moreover, they found that St does not change with the Reynolds number in the range 1.34 × 10 4 ≤ Re ≤ 1.18 × 10 5 unlike L 1 which has a decreasing trend. Lastly, Cimarelli et al (2020) investigated via high-order implicit LES the influence of geometrical characteristics of the body, as the sharp leading-edge corners, the presence of separation at the trailing edge and the coupling between the two sides of the plate.…”

“…Table 1 shows that often experiments and simulations tend to underestimate the shedding frequency. This is attributed to the sensitivity of the BARC flow to slight deviations from the ideal setting, or to simulation inaccuracies; the finite Re should not affect the computed value, because according to Nakamura et al (1991), Schewe (2013) and Moore et al (2019) at Re ≥ 3000 the Strouhal number has already reached its asymptotic value.…”

“…Very recently, Cimarelli et al (2019), Moore et al (2019) and Rocchio et al (2020) presented and discussed the production term in the equation for the turbulent kinetic energy, with the latter two studies focusing on the region close to the upstream corners. Hence, before presenting the Reynolds stresses, we start with Fig.…”