Study on preferential concentration of inertial particles in homogeneous isotropic turbulence via big-data techniques

Abstract

We present an experimental study of the preferential concentration of sub-Kolmogorov inertial particles in active-grid-generated homogeneous and isotropic turbulence, characterized via Voronoï tessellations.We show that the detection and quantification of clusters and voids is influenced by the intensity of the laser and high values of particles volume fraction φ v . Different biases on the statistics of Voronoï cells are analyzed to improve the reliability of the detection and the robustness in the characterization of clusters and voids. We do this by adapting Big-Data techniques that allow to process the particle images up to 10 times faster than standard algorithms.Finally, as preferential concentration is known to depend on multiple parameters, we performed experiments where one parameter was varied and all others were kept constant (φ v , Reynolds number based on the Taylor length scale Re λ , and residence time of the particles interacting with the turbulence). Our results confirm, in agreement with published work, that clustering increases with both φ v and Re λ . On the other hand, we find new evidence that the mean size of clusters increases with Re λ but decreases with φ v and that the cluster settling velocity is strongly affected by Re λ up to the maximum value studied here, Re

Dynamic effects of inertial particles on the wake recovery of a model wind turbine

“…The resulting Voronoi diagrams contain clearly defined regions suggestive of clustering and voids typically found in turbulent flows. The intersection points observed between the probability distribution of cell areas and that of a random process show similarity to those obtained in more exhaustive studies [27,30]. Although a more focused discussion on cluster and void areas and distributions is necessary for full Voronoi analysis, the validated results presented suggest this method to be useful for defining particle behavior in turbine wakes and inspires further examination.…”

“…But the question remains as to how the presence of such particles affect turbine wake behavior and how wake turbulence may alter their behavior and trajectories. As previously noted, the vast difference in density between wind and denser debris such as rain and granular object suggests the interaction is a multi-phase phenomenon, gaining insight here from research describing particle behavior and effects within turbulent flows[28,18,22,30]. Studies toward particle behavior in a variety of turbulent flows show that in homogenous turbulence, particles of smaller size tend to become entrained in vortices, while larger particles are propelled away by large scale features[13].…”

“…Voronoi analysis was performed within the far wake region of the turbine model to observe clustering behavior. Rather this intersection speaks to variations in experimental conditions and lack of convergence due to limited data which should be examined for future study [27,30]. This study does not fully investigate Voronoi analysis, as it would also allow to characterise the size and distribution of clusters and voids.…”

“…Voronoi cell distribution example with randomly distributed points[34] Data used for this purpose consists of 2-D images. The cell creation method used here is referred to as the 'perpendicular bisector method', where the size and shape of the cells are dependent on proximity of individual particles with their surrounding neighbors[15,27,30,34]. The walls of the polygon are defined as lines (or surfaces in 3-D framework) which are orthogonal to, and located at the midpoint of a straight line from one particle to the next.…”

Effect of Reλ and Rouse numbers on the settling of inertial droplets in homogeneous isotropic turbulence

“…In the same facility, previous research has found, utilizing 2D images, evidence of preferential concentration under similar experimental conditions as those studied here [3]. Based on the approach of [55], we decided to normalize ∆V by the cluster velocity V cl ∼ C cl A cl ρ p g/(ρ air ν), where ρ p is the particle density, C cl , and A cl are the clusters concentration, and area, respectively. We estimate the latter quantities as C cl ≈ 4φ v from 2D correlations in the same facility [3,22,32], and A cl = 2.1 × 10 −5 St −0.25 max Re 4.7 λ φ 1.2 v [3].…”