Experimental study of coherent flow structures past a wall-mounted square cylinder

Abstract

In high flow velocity areas like those suitable for tidal applications, turbulence intensity is high and flow variations may have a major impact on tidal turbines behaviour. Large boils that can be observed at the sea surface are emitted from the sea floor and may interact with the tidal turbines. These boils have then to be characterized. The Reynolds number, based on the rugosity height and the mean flow velocity, is rather high in this context: Re ¼ 2:5 107. For that purpose, experiments are carried out in a flume tank with Re as high as achievable in Froude similitude (in the tank: Re ¼ 2:5 105 and Fr ¼ 0:23) in order to study coherent flow structures emitted behind seabed obstacles. The obstacle is here a canonical square wall-mounted cylinder chosen to be representative of specific in-situ bathymetric variations. Using PIV and LDV measurements, the flow past the cylinder is investigated. Using a POD filter, large coherent structures are identified and their trajectories are analysed. By means of a Lamb-Oseen profile approximation, properties of these structures are determined. The formation mechanism of such structures is discussed in this paper and their behaviour is characterized. It is assumed that vortices periodically shed from the obstacle interact and generate hairpin structures. Highlights ► Experimental study of coherent flow structures past a wall-mounted square cylinder. ► Tests are carried out on a wall-mounted cylinder representative of seabed elements, in Froude similitude with high Reynolds number. ► PIV measurements are performed in vertical measurement planes and spatial analyses are performed. ► POD analysis and center detection allow to study the vortices behaviour.

Experimental investigation of the turbulent wake past real seabed elements for velocity variations characterization in the water column.

“…Particles displacement is calculated using a Cross-Correlation [Meinhart et al (1993)]. Outliers are then replaced with the Universal Outlier Detection [Westerweel and Scarano (2005)], example and precisions on that method can be found in [Ikhennicheu et al (2019b)]. PIV measurements are carried out at various transverse positions: y * = 0, y * = 1 and y * = 2 (denoted y0, y1 and y2 respectively) represented in figure 6.…”

“…A list of PIV planes performed with measurements spatial discretization are summed up in table 2 and PIV measurement planes are illustrated in figure 7. For the two measurement techniques, uncertainty is estimated to be around 2 % for LDV and 2.6 % for PIV, calculus are detailed in [Ikhennicheu et al (2019b)].…”

“…In [Ikhennicheu et al (2019b)], a comparison is made between two large aspect ratio cases: C 6 and C 7 . Results showed that streamwise velocity profiles are very similar.…”

“…For test case C L , large boils are observed at the tank surface, as illustrated in figure 9, while no boil is detected for every other cases. Turbulent structures emitted downstream of the cylinder are detected and tracked us-ing PIV measurements in [Ikhennicheu et al (2019b)] and an explanation for their creation is proposed. Kelvin-Helmholtz instabilities are generated in the wake of the obstacle.…”

“…As explained in section 2, LDV signals are re-sampled and a FFT is applied. More details on the LDV postprocessing method can be found in [Ikhennicheu et al (2019b)]. Results for all 3 velocity components are presented in figure 10 versus the Strouhal number S t = f H/U ∞ with f the frequency.…”