ETC2019 17th European Turbulence Conference Submission

10:45 Multiphase Flows 1

10:45 15 mins #483	Numerical and experimental investigation of regional deposition of glass fibres in the human respiratory airway Yan Cui, Miloslav Belka, Frantisek Lizal, Jure Ravnik, Matjaž Hriberšek, Paul Steinmann Abstract: Please read the attached pdf file.
11:00 15 mins #103	Drag reduction in turbulent channel flow of flexible fi Arash Alizad Banaei, Marco Edoardo Rosti, Luca Brandt Abstract: We perform numerical simulations of the turbulent channel flow of flexible fiber suspensions at Reb = 2800, based on the bulk velocity and channel half height. The fibers are considered as continuously flexible slender objects satisfying the Euler-Bernoulli equation. The fluid and solid motions are coupled with the immersed boundary method. We have considered different volume fractions, 1:5%, 3% and 6%, and two different fibers bending rigidities, one to study stiff fibers and the other for flexible fibers. The fiber aspect ratio is 20 and the length is 20 wall units for all cases presented. The results indicate that significant drag reduction can be achieved even for low volume fractions due to the suppression of the turbulent fluctuations induced by the fibers. The fibers remain almost straight and undeformed in the bulk of the channel, while they significantly deform in the near-wall regions. Indeed, their maximum deformation is observed in the region where the Reynolds stresses have their maximum value. Furthermore, significant increase in the fiber deformation is observed for increasing volume fraction. Statistics of the fluid and fiber motion will be presented, with interest in the relation between the turbulent time scales and the fiber oscillations.
11:15 15 mins #436	ORIENTATION DYNAMICS OF RIGID FIBRES IN A TURBULENT CHANNEL FLOW Subhani Shaik, Sofia Kuperman, Vladislav Rinsky, Rene van Hout Abstract: The orientation dynamics of fibres in turbulent flows is of great importance in various geophysical, biological, environmental and industrial processes. The motion of fibres in a flow is a combination of translation and rotation. Fibre orientation depends on the velocity gradients that the fibres experience along their trajectories. Most of the theoretical and experimental research is limited to isotropic turbulent flows and there is a need to extend this work to study fibre-flow interactions in non-homogeneous, near-wall turbulent flows. Although numerical studies on fibre dispersion in wall turbulence exist, the fibres are usually modelled as point particles, i.e. the fibre length is considered to be smaller than the Kolmogorov length scale. This assumption represents an unrealistic situation and there is a need to investigate length as well as inertial effects. The shape and inertial characteristics of fibres are quantified by their aspect ratio and Stokes numbers (St) respectively. The former is defined as λ=L/D where L is the fibre length and D is the fibre diameter, and the latter as the ratio between the fibre response time and the viscous time scale in the wall-bounded turbulence. In this study we seek to measure the effect of λ and St on the orientation dynamics of a dilute suspension of rigid fibres in a turbulent channel flow. Two sets of nylon fibres characterized by λ = 88.1, St = 0.24 and λ = 30.7, St = 2.16 were studied. Experiments were performed in a horizontal, closed-loop square water channel (50 mm x 50 mm internal cross-section) at a distance of 1.3 m from the channel entrance, where the turbulent flow was fully developed. The corresponding bulk and friction Reynolds numbers were 7353 and 435 respectively. The three-dimensional motion of the fibres in a volume of interest (VOI = 35 mmx 35 mm x 35 mm) was tracked using two orthogonal view, digital inline Fraunhofer holographic cinematography. The VOI was centered and extended from the bottom of the channel. Holograms were acquired at 800 Hz, i.e. 5 orders of magnitude greater than the fibre response frequency. Acquired holograms for each camera were numerically reconstructed and by combining the information from the two cameras, the fibre 3D orientations were determined. In addition, using the temporally resolved data, their tumbling rates were calculated. The reconstructed VOI was divided into several sub-volumes spanning the buffer and logarithmic regions of the turbulent boundary layer. The fibre orientation distributions and rotational dynamics were obtained in these sub-regions and compared with existing literature.
11:30 15 mins #335	THE LUMLEY TRIANGLE ─ A TOOL TO ANALYSE PARTICLE ROTATION ANISOTROPIES Helge Andersson, Kun Yang, Lihao Zhao Abstract: We explore how invariants frequently used to characterize correlations between turbulent velocity fluctuations, i.e. Reynolds stresses, can be used to assess the anisotropy of the particle angular velocity correlation tensor. The so-called anisotropy invariant map (AIM) and flatness parameter F introduced by Lumley and Newman [1] and Lumley [2] are routinely used to describe the anisotropy of the Reynolds-stress tensor [3]. Such correlations are essential in turbulence modelling, notably in second-moment closures. AIMs have only rarely been used to study the anisotropy of the vorticity correlation tensor. Knowledge of the complete correlation tensor is essential since both AIMs and F are based on the second (II) and third (III) invariants. In this study, however, usage of these invariants will be carried over to particle dynamics with the aim to diagnose the anisotropy of the particle rotation. To this end we consider inertialess spheroids embedded in an asymmetric turbulent Couette-Poiseuille (CP) flow simulated by Yang et al. [4]. The turbulence field near the shear-free moving wall is distinctly different from conventional wall turbulence prevailing near the opposite wall. The CP-flow is therefore particularly attractive for studies in which particle rotation measures are explored. We recently investigated different rotation modes, i.e. spinning versus tumbling, about the particle’s symmetry axis [5]. The particle rotation correlation tensor, on the other hand, represents particle rotations relative to the meanflow direction. The shape of the spheroids is parameterized by the aspect ratio λ. The AIMs in Figure 1 show the distinctly different rotational behaviors near the moving and stationary walls. Three-component (3C) almost-isotropic particle rotation (II ≈ III ≈ 0) is found in the sheared core region. Irrespective of shape, spheroids exhibit almost two-component (2C) rotation near the moving wall, reflecting a pancake-shaped [6] rotation correlation tensor. Near the stationary wall, however, the correlation tensor is cigar-like for oblate spheroids which follow the right boarder of the Lumley-triangle in Figure 1(a). Results for several other aspects ratios ranging from flat disks (λ = 0.01) to long rods (λ = 50) will be discussed in the conference presentation. Results for inertialess spheroids (St = 0) will also be compared with results for inertial spheroids (St = 30). The Lumley triangle may serve as an adequate diagnostic tool to assess rotation anisotropies.
11:45 15 mins #402	SINGLE-DROP BREAKUP IN HOMOGENEOUS ISOTROPIC TURBULENCE Alberto Vela-Martín, Marc Avila Abstract: The emulsification of a disperse phase in a continuous phase is of key importance in chemical engineering. After initial transients, a statistically steady state, characterized by the probability density function (pdf) of drop sizes, is approached. This pdf is determined by the competition of drop breakup and coalescence processes, which are notoriously difficult to parametrize. We here focus on the breakup of a single drop in homogeneous isotropic turbulence, as studied recently by Changxiao et al. The Navier–Stokes equations are solved with the pseudo-spectral method of Cardesa et al., which was extended to simultaneously solve the Cahn–Hilliard equations governing the conservation of a phase-field variable. Our highly efficient GPU code enables a statistical study of the inherently stochastic breakup process. For simplicity, density- and viscosity-matched fluids are considered.
12:00 15 mins #449	Modeling of coalescence and breakup of fluid particles in turbulent flows Antonio Buffo, Marco Vanni, Daniele Marchisio Abstract: The prediction of the size distribution in turbulent liquid-liquid and gas-liquid systems is of a great importance in many industrial applications. Nowadays computational tools based on Computational Fluid Dynamics (CFD) and Population Balance Modeling (PBM) are available to estimate this property for different systems in large scale equipment. However these tools require the estimation of the outcome of coalescence and breakage events occurring at a smaller scale in the system. This information can be retrieved only if the interactions between the fluid particles and the surrounding turbulent fluid is properly accounted for. This led to the formulation of models (often called kernels), which describe the statistics of these events, such as the frequency of coalescence and breakage, as well as the number and the size of the daughter fluid particles after the break-up. Turbulence in fact plays a major role in these phenomena. As far as coalescence is concerned, the collisions between fluid particles are mainly caused by the chaotic relative motion induced by turbulent eddies of size similar of the colliding particles (i.e., the so-called turbulent fluctuation mechanism). This mechanism is well documented in the literature and it has brought to the formulation of many phenomenological models, each of them with different assumptions and sometimes predicting very different outcomes. Another important collisional mechanism is caused by the entrapment of very small fluid particles into small-scale turbulent vortices, close to Kolomogorov scale. In the formulation of the phenomenological models, particular attention has been paid also on the role of the molecular forces acting on the collided particles, to establish if the collision caused by turbulence lead to drainage of the fluid film between the fluid particles and therefore the formation of the new bubble. The breakage of fluid particles in turbulence is often described using an analogy with the kinetic theory of gases as a stochastic collisional event between the turbulent eddies and the fluid particle, where the breakage event occurs if the disruptive forces overcome the internal resisting forces, which are mainly interfacial and viscous forces. Although this phenomenon can be considered simpler compared to the coalescence since it involves only one fluid particle, in this case it is very difficult to predict the number of the daughter particle generated. This number depends on the strong interplay between the balance of the local stresses acting on the interface of the fluid particle and the shape of the particle during the latest steps of the breakage event. Due to the complex physics involved, the formulation of coalescence and breakage kernels is often a trade-off between an accurate description of the physical reality and the practical use of the final numerical expression, trying to avoid the solution of multi-dimensional integrals or similar mathematical objects that may slow down the numerical simulation. For this reason, the kernels have modeling constants, in which the uncertainty related to the modeling assumptions and simplifications are contained, usually fitted through experiments carried out on real systems. In this work, we reviewed most of the proposed approaches to quantify the coalescence and breakage of fluid particles, with a particular attention to the role that the turbulence and the modeling assumptions have on the description of such events. This work is a necessary step towards the formulation of more accurate models, where the modeling constants are substituted by universal parameters, for the prediction of the bubble/droplet size distribution in gas-liquid and liquid-liquid systems.
12:15 15 mins #625	Cloud-clear air interfaces: Population Balance Equation solutions by considering nucleation information from in-situ measurements, and by modeling the droplet growth on super-saturation fluctuation data from numerical simulation. Mina Golshan, Federico Fraternale, Marco Vanni, Daniela Tordella Abstract: In this study we will present a preliminary analysis of the droplet population as hosted by a turbulent shear-less mixing air flow which is mimicking a cloud/clear-air interface. The interface is subject to density stratification and vapor density fluctuation under super-saturation conditions. We use the Population Balance Equation (PBE) as a tool to represent a few aspects of the droplet size dynamics by taking into consideration both turbulence results coming from in-situ and laboratory measurements and from numerical simulations.
12:30 15 mins #34	Dynamics of small flexible fibers in turbulent channel flow Diego Dotto, Cristian Marchioli Abstract: We investigate the dynamics of flexible fibers in turbulent channel flow. We examine the effect of local shear and turbulence anisotropy on the translational and rotational behaviour of the fibers, considering different elongation (parameterized by the aspect ratio) and inertia (parameterized by the Stokes number). Results from 1-way and 2-way force-coupled simulations show that flexible fibers obey the same inertia-driven mechanisms that govern preferential concentration of spherical particles and rigid fibers.