Prof. Gretar Tryggvason, Johns Hopkins University
Coauthored with Jiacai Lu and Lei Zeng
Numerical Studies of Bubbly Flows Containing Hydrophobic Particles
Many multiphase flow problems of practical interest consist of three or more phases that interact in complex ways. One example is froth flotation, where buoyant bubbles are used to separate hydrophobic particles from hydrophilic ones in a slurry and bring them to the top of a tank where they can be skimmed off. This is widely used in mineral processing, such as the separation of copper particles out of the ore. While many aspects of the process are reasonably well understood, the complex interactions make it difficult to predict the overall selectivity. Here, we present numerical simulations of the capture of hydrophobic particles by buoyant bubbles.
Prof. Steve Ceccio, University of Michigan
Cavitation Inception and its Noise Emissions during the Interaction of a Pair of Line Vortices
Cavitation inception can first occur in the cores of relatively weak vortices that are stretched when interacting with stronger, non-cavitating, vortices. We examine a canonical flow comprised of two initially parallel counter-rotating vortices that undergo the Crow instability, resulting in the wrapping and stretching of the weaker (secondary) vortex. This stretching process leads to cavitation inception. We use volumetric velocimetry to measure the evolving flow field and estimate the unsteady pressure within the incepting vortex. The pressures in the vortex cores are related to vortex stretching and vortex properties. Using this data, the inception event rate is compared for different nuclei populations of the flow with the aim of predicting the inception rate. Finally, the resulting acoustic emissions of the incipient bubbles is related to the nuclei population and the underlying vortical flow.
Prof. Andrea Prosperetti, University of Houston
Co-authored with Guangzhao Zhou, University of Chinese Academy of Sciences, Beijing, China
Wrong Bubbles for Joe
Taylor bubbles – gas volumes with an equivalent spherical diameter greater than the diameter of the vertical tube in which they rise – are not the type of bubbles that Joe cares about, but I hope that this talk, offered in the spirit of admiration and friendship, will nevertheless be of some interest to him.
It is obvious that, for a Taylor bubble to rise in a liquid-filled tube, the volume vacated by its tail must be replenished by fresh liquid falling in the film that separates the gas from the wall. Accordingly, if the volume flow rate in this film is increased or decreased, the bubble will rise with a greater or smaller speed. One way to thicken the film is to insert a porous cylinder, coaxial with the tube, and having a sufficiently small diameter. The same approach can be used to thin the film, but a somewhat surprising and less straightforward way is to cause the Taylor bubble to execute volume oscillations. Unexpectedly, the imposed oscillations have the effect of decreasing the liquid flow rate in the film with a marked reduction of the bubble rising velocity.
Prof. Charles Meneveau, Johns Hopkins University
Forward and Inverse Cascade in Turbulence
We revisit the issue of forward and inverse cascade in high Reynolds number turbulence. This is a topic to which Joe Katz contributed valuable experimental data based on his early developments and applications of high accuracy 2D and 3D Particle Image Velocimetry (Liu et al. JFM 1996, Tao et al. JFM 2002). Instead of the filtering method used in LES and in the above-mentioned experiments, we here use a description based on a generalized local Kolmogorov-Hill equation expressing the evolution of second-order velocity increments integrated over spheres in the inertial range. We propose a new definition of entropy of turbulence and its generation rate, and measure it from DNS data available at JHTDB. We present evidence that the fluctuation relation (FR) from non-equilibrium thermodynamics applies to turbulence in the inertial range. Specifically, the ratio of probability densities of forward and inverse cascade is shown to follow exponential behavior with the entropy generation rate if the latter is defined by including an appropriately defined notion of “temperature of turbulence” proportional to the kinetic energy. This work is done in collaboration with H. Yao and T. Zaki (JFM 2023, in press) and is supported by NSF.
Prof. Alexander Smits, Princeton University
Turbulence Control and Drag Reduction: Some Recent Examples
Reducing skin friction drag and achieving beneficial flow control continue to be important goals for research in turbulent flows. Three recent examples are presented. First, we discuss the use of SLIPS (Slippery Liquid-Infused Porous Surfaces) for underwater applications, where grooves in the surface hold a low viscosity fluid like heptane and provide a slip boundary condition for the overlying turbulent flow. Up to 30% drag reduction has been reported. Second, we consider modifications in pipe flow surface geometry inspired by Proper Orthogonal Decomposition. Short sections of pipe shaped according to POD Modes 3 and 15 are shown to have significant and long-lasting effects on the turbulence, and the interaction of the two modes is strongly nonlinear. Third, we review recent efforts in active control using transverse wall oscillation. This work reveals two pathways to drag reduction: inner-scaled actuation (ISA) on scales comparable to the near-wall motions, and outer-scaled actuation (OSA) on scales comparable to the outer layer. In contrast to ISA, OSA allows significant levels of drag reduction at high Reynolds number that are expected to increase with Reynolds number, it requires relatively low frequencies, and net power savings are possible.
Prof. Morteza Gharib, California Institute of Technology
Coauthored with Chris Willert and Jurgen Kompenhans
From Leonardo to Prandtl’s Particle Image Velocimetry
Leonardo da Vinci used small leaves and grass seeds to visualize flow patterns from bird flights to vortices generated by fast streams or in his glass models of heart valves. Justifiably, one can name him as one of the original particle trackers who used this technique to understand the flow phenomena. Four hundred years later, with the development of modern digital recording systems, PIV has become an essential tool for exploring new phenomena in fluid mechanics or testing and validation of CFD. In this talk, we will present some of the early attempts in producing time-resolved PIV image sequences during the late 1920s and early 1930s by Prandtl and his colleagues Tietjens and Müller using free surface water flumes. Recorded at 20 frames per second, the films visualize the process of unsteady flow separation and vortex generation on surface-piercing objects such as airfoils or cylinders. The visualization was achieved using small particles (aluminum powder, ferrous mica, or lycopodium powder) scattered on the water surface. Illumination from above resulted in high-contrast images of the random particle distribution that are very well suited for PIV analysis. Modern PIV software is used to process digitized versions of the films. In addition to the surface flow field, the time-evolving vorticity field and other quantities can now be visualized, which shows the importance of carefully documenting and archiving valuable data.
Prof. Michael Triantafyllou, Massachusetts Institute of Technology
An “Intelligent Towing Tank”: AI Strategies for Experimental Fluid Mechanics
AI strategies to expedite experimental mapping of the properties of complex flow-structure interaction problems that depend on multiple parameters offer unprecedented new capabilities to identify the parametric effect on the dominant physical mechanisms and optimize system configurations. First, the “Intelligent Towing Tank” used Gaussian Process Regression to generate extensive force databases for the combined in-line and cross-flow vortex induced vibrations of flexibly mounted cylinders as function of the Reynolds number. Then the tank was used to maximize the wave energy dissipation of architected artificial reefs with controlled porosity, providing drag coefficients at least an order of magnitude larger than existing reefs.
Prof. Fulvio Scarano, Delft University of Technology
Experiment Upscaling and Data Assimilation for 3D PIV
Three-dimensional velocity field measurements with Particle Image Velocimetry are nowadays frequently obtained for the investigation and visualization of complex three-dimensional flows. Most experiments still focus on the study of fundamental, turbulent flow phenomena, but some applications indicate that the technique is reaching the maturity needed to be deployed and advance aerodynamic technologies. The latter have become possible with the realization of seeding devices that produce neutrally buoyant helium filled soap bubbles in the sub-millimeter range for use in large wind tunnels.
Three-dimensional measurements, need to cope with intrinsic limitations of spatial resolution, due to the lower seeding density imposed by the occurrence of ghost particles. On the other hand, time-resolved measurements within a volumetric domain are suited for more advanced data assimilation treatments compared to the traditional signal-processing based (e.g. filtering in space and/or time, proper orthogonal decomposition). 3D and 4D data can embed governing laws in the post-processing algorithm that enforce their physical consistency.
The lecture discusses data assimilation techniques that invoke the Lagrangian transport principle to estimate spectra from PIV time series even when the Nyquist criterion is violated. Pouring space into time is demonstrated from advection-based Time Supersampling and later extended and generalized to turbulent shear flows using the vorticity equation (VIC, or vortex-in-cell).
The reverse process of estimating data at a higher spatial resolution corresponds to pouring time into space and is referred to as Super-Resolution. Techniques that achieve this goal are explained in their principles and by applications where quantities not measurable otherwise (e.g. 3D turbulent dissipation) can be estimated from experiments.
A snapshot of turbulent boundary layer measured with 3D-PTV and super-resolved with the VIC+ technique (Schneiders et al., 2017).
Prof. Oguz Uzol, Middle East Technical University
Experiments in the Aerodynamics Laboratory of METU Center for Wind Energy Research: From Porous Discs to Gust Generation to Airfoil Aerodynamics
This talk will present an overview of on-going research activities in the Aerodynamics Laboratory of METU Center for Wind Energy Research. Recent results from wake measurements downstream of porous discs and small wind turbines, which are frequently used for modeling wind turbine wakes in wind tunnels, will be presented. In addition, different experimental setups that we use to generate gusts in our wind tunnels will be introduced (2D and active grid based) and measurement results regarding wake response of porous discs and small wind turbines to incoming gusts will be discussed. Finally, examples of transition measurements over wind turbine airfoils using Infrared Thermography (IRT) will be presented. These experiments are conducted in the new large-scale subsonic wind tunnel at METU. The presentation will include details of the design and construction efforts for this wind tunnel, examples from various industrial tests and sample IRT measurement results for the DU00-W-212 and NLF(1)-0416 airfoils.
Prof. Kathleen Stebe, University of Pennsylvania
Active Surface Agents: Active Colloids at Fluid-Fluid Interfaces
Active colloids include bacteria which swim by action of flagella and biomimetic self-propelled colloids that move by consumption of chemical fuel. Active colloidal suspensions have generated tremendous excitement for their collective dynamics and non-equilibrium phenomena including enhanced diffusion, long-ranged correlations and active phase separations. We have been developing the concept of Active Surface Agents, active colloids trapped at fluid interfaces to promote interfacial transport. Fluid interfaces are highly non-ideal, complex domains that impose constraints that alter swimming behavior. We study the bacterium Pseudomonas Aeruginosa (PA01) at interfaces and characterize several distinct swimming behaviors. We measure the flow generated by these swimmers in pusher mode using a recently developed flow visualization method correlated displacement velocimetry. The flow field has unexpected asymmetries whose structure we describe fundamentally using hydrodynamic theory. We explore the implications of our results on mixing in the interface. By understanding how biological swimmers move at fluid interfaces, we can develop design rules for artificial biomimetic systems to promote transport at fluid interfaces with broad implications in chemical engineering processes. This work was initiated during the time that the Stebe group worked in collaboration with Joe Katz’s group and others under the support of the GOMRI initiative to address the Gulf of Mexico oil spill. We were inspired by Joe’s ability to advance fundamental science in real-world settings and his joy in tackling problems that impact humanity.
Prof. Kenneth Breuer, Brown University
The Rhythm of Life: Velocimetry of Cellular and Intracellular Flows
Advances in imaging technology, optical microscopy and micron-scale particle image velocimetry, measurements of motion at the cellular and intra-cellular scale have proliferated over the past fifty years. In this talk I will focus on three-dimensional velocimetry of synthetic intracellular flows biological systems comprised of microtubule bundles driven by kinesin motors. These active matter systems mimic the internal dynamics of the cell and generate complex chaotic flows at zero Reynolds number exhibiting both isotropic and self-organized non-isotropic characteristics.
Prof. Howard Stone, Princeton University
New Results for the Flow of Polymer Solutions in Non-Uniform Geometries
The flow of complex fluids, e.g., polymer solutions, is generally complicated by the nonlinear constitutive relation connecting stress to the rate of strain. In recent years we have analyzed pressure-driven flow of non-Newtonian or viscoelastic fluids in spatially varying, but narrow, geometries. These problems are normally treated numerically but we describe an approach that yields new analytical results. The analytical predictions are compared (favorably) with numerical results. The results also highlight where, and possibly why, there are differences between the numerical simulations and experimental results, which has been recognized in the literature, but is not often discussed.
In addition, if there is time, I will share an elementary way to think about the various time derivatives that appear in studies of viscoelastic fluid mechanics. (The work presented here is joint research with Evgeniy Boyko.)