Seminar Schedule

Abstract: Actively changing the blade pitch angle of a vertical-axis wind turbine throughout the rotor rotation can increase the power production and reduce the structural vibrations, as demonstrated in previous work. Here, we experimentally explore the potential of pitch control to reduce the turbine’s drag, which is related to the wake deficit and recovery. The pitching kinematics are optimized with double objective Bayesian optimization, to reduce the turbine drag while increasing the power production. The optimized pitching yields a reduction of the turbine drag by 68% while maintaining a power production similar to the unactuated case. This decrease of turbine drag reduces the structural loads and improves the floater stability for offshore wind turbines. Wake flow measurements reveal a wake deficit reduction of 73%, as well as a reduction of the wake unsteadiness. The results confirm the potential of blade pitching to improve the wake recovery and generate smoother conditions for downstream turbines in a wind farm.

Abstract: Many previous works have studied fluid-structure interactions induced by thin flexible bodies. In most of these studies the body is nearly inextensible, with a moderate bending modulus. Here we consider softer materials extensible membranes that have zero bending modulus, and undergo significant stretching in a fluid flow that can lead to flutter. We develop a mathematical model and numerical method to study the large-amplitude flutter of rectangular membranes that shed a trailing vortex-sheet wake in a 3D inviscid flow. We consider 12 distinct boundary conditions at the membrane edges and compute the stability thresholds and the subsequent large-amplitude dynamics across the three-parameter space of membrane mass ratio, pretension, and stretching rigidity. We find that 3D dynamics in the 12 cases naturally form four groups based on the conditions at the leading and trailing edges. The conditions at the side edges are generally less important, but may have qualitative effects on the membrane dynamics e.g. steady versus unsteady, periodic versus chaotic, or the variety of spanwise curvature distributions depending on the group and the physical parameter values.

Abstract: In nature, flexible and porous structures are ubiquitous, aiding plants like dandelions in seed dispersion and reducing structural loads on leaves in strong winds. Natural fliers have compliant wings for thrust and maneuverability, while some bird feathers' porosity reduces noise emissions by suppressing coherent vortex shedding. Despite these natural advantages, poro-elastic structures are rarely used in engineered aerodynamic applications due to complex fluid-structure interactions and numerous design parameters involved.

This study explores the effects of poro-elasticity on thin, circular membrane disks by measuring deformation, unsteady flow fields, and drag forces at varying porosity levels in wind tunnel experiments. Our results show that porosity reduces the magnitude of vortex-induced membrane vibrations by stabilizing a large-scale vortex forming in the wake. The highly porous membranes stretch the vortex in the stream-wise direction, reducing the heavy load fluctuations and the high average drag experienced by the non-porous membranes. We present a poro-elastic scaling to model the aerodynamic loads and provide insights into the unsteady vortex shedding at various porosity levels. Finally, we relate our findings to poro-elastic phenomena observed in nature and offer guidance for the design of engineering devices.

Abstract: Autonomous ocean-exploring vehicles have begun to take advantage of onboard sensor measurements of water properties such as salinity and temperature to locate oceanic features in real time. Such targeted sampling strategies enable more rapid study of ocean environments by actively steering towards areas of high scientific value. Inspired by the ability of aquatic animals to navigate via flow sensing, this work investigates hydrodynamic cues for accomplishing targeted sampling using a palm-sized robotic swimmer. As proof of-concept analogy for tracking hydrothermal vent plumes in the ocean, the robot is tasked with locating the center of turbulent jet flows in a 13,000-liter water tank using data from onboard pressure sensors. To learn a navigation strategy, we first implemented Reinforcement Learning (RL) on a simulated version of the robot navigating in proximity to turbulent jets. After training, the RL algorithm discovered an effective strategy for locating the jets by following transverse velocity gradients sensed by pressure sensors located on opposite sides of the robot. When implemented on the physical robot, this gradient following strategy enabled the robot to successfully locate the turbu1ent plumes at more than twice the rate of random searching. Additionally, we found that navigation performance improved as the distance between the pressure sensors increased, which can inform the design of distributed flow sensors in ocean robots. Our results demonstrate the effectiveness and limits of flow-based navigation for autonomously locating hydrodynamic features of interest.

Abstract: Owing to continuously rising marine vessel traffic, shipping lanes around the world have become a prominent source of ocean noise emissions. One of the major sources of marine noise arises from unsteady cavitation around marine propellers. The current work aims to numerically investigate the dynamics of unsteady partial cavitation around rigid and flexible NACA66 hydrofoils. We elucidate the features of sheet-cavitating flow which enable the transition to cloud cavitation, evaluate the instabilities driving sheet-cavity breakdown and identify the vortex structures which drive cloud cavity collapse, and quantify the frequencies observed over the course of a cavitation cycle. For a rigid hydrofoil, our results demonstrate the emergence of a baroclinic torque instability arising at the vapor-liquid interface, which results in sheet cavity collapse and formation of vortex cavities. In a compliant hydrofoil, the deformation induced velocity fluctuations disturb the sheet cavity's formation, resulting in early cavity collapse. This results in significantly altered vortex dynamics as well as far-field pressure fluctuations. 

Abstract: Simulating fluid flows is a computationally demanding task that can be accelerated by modeling the interactions between point vortices. In this work, we model vortical interactions directly from data using graph neural networks (GNNs). We do this by constructing a hierarchy of GNNs to predict the dynamics of vortex clusters simulated using the Biot-Savart law. This hierarchy consists of a GNN that models the local interactions within a cluster and a GNN that models the global interactions between clusters. Additionally, we design the GNNs such that they enforce equivariance to rotations and translations. We show that this equivariant and hierarchical method is more accurate and faster than constructing a fully connected GNN. Furthermore, the method can predict the dynamics of test data containing a different number of vortices and clusters than exists in our training data, which can not be done with other data-driven methods like dense neural networks.

Abstract: The transient flow dynamics in the highly modulated near-wake region of a cantilevered square cylinder of height to width (h/d) ratio 4, protruding a thin laminar boundary layer, at a Reynolds number of 10600 is investigated using a novel 3D flow reconstruction technique. The technique builds on the multi-time-delay estimation technique of Hosseini et al. (2015) by implementing the finite-impulse-response spectral proper orthogonal decomposition (FIR-SPOD) of Sieber et al. (2016), to synchronize uncorrelated planar PIV measurements using surface pressure measurements. The FIR-SPOD successfully separates turbulent spatio-temporal scales within narrow spectral bandwidths, which in turn results in accurately retaining the correct phase relationships between pressure and velocity modes as is required for synchronizing coherent motions along the height of the obstacle. It is shown that the resultant low-dimensional 3D reconstruction of the flow field captures the cycle-to-cycle variations of the dominant vortex shedding process, which give rise to vortex dislocation events. Thus, the present methodology shows promise in 3D reconstruction of challenging turbulent flows, which exhibit non-periodic behavior or contain multi-scale phenomena.

Abstract: The propulsive characteristics of a foil undergoing a prescribed rotational pitch about its quarter cord in a uniform free stream flow of an incompressible Newtonian fluid are investigated numerically over a wide range of Reynolds and Strouhal numbers. At a fixed Reynolds number, the cycle-averaged thrust is shown to rise monotonically with the Strouhal number. Thus, for sufficiently low Strouhal numbers, the foil experiences a net drag while at a critical transition Strouhal number a self-propelling state in which the foil experiences no net force is attained. For Strouhal numbers larger than the critical Strouhal number, the imposed rotational pitch results in a net thrust. The cycle-averaged power coefficient rises sharply with the Strouhal number.  As a consequence, for a given Reynolds number, the propulsive efficiency for thrust generation attains a maximum at a specific Strouhal number. The dependence of the critical parameters namely the drag-to-thrust transition and maximum efficiency Strouhal numbers, and the peak propulsive efficiency on the Reynolds number will be discussed along with the implications of our findings for undulatory locomotion.

Abstract: The dynamics of wing-gust encounters are governed by the triadic interaction of the wing, its shed vorticity, and the ambient gust vorticity in the flow. Thus, the system state during the interaction is inherently tied to the vorticity field, which cannot be measured directly but only estimated through other measurements and the use of a system model. In this work, we make progress in model-based state estimation of unsteady flows by experimentally measuring the surface pressures, loads, and associated velocity fields of wings encountering transverse gusts, and by developing low-order vortex models that can predict these measured quantities. We find that accurate modelling of the Leading-Edge Vortex (LEV) is required to predict the suction side as well as the pressure side surface pressures and that modelling the deformation of the gust is important in predicting pressure during the gust exit. While these observations reinforce the need for global flow estimation, we also show that local surface pressure measurements aided by unsteady airfoil theory can provide useful information about the system state. 

Abstract: There are many examples in fluid dynamics in which we wish to use a limited amount of information about a flow (e.g., from sensors) to infer its larger behavior. Rather than treat this inverse problem deterministically, it is valuable to embrace its uncertainty, e.g., due to noisy sensor measurements, and work in a probabilistic setting. Even if we think we have non-noisy measurements, we can still learn a lot about the estimated flow by treating it in this setting: Is the estimate unique? What information is available (and not available) in the sensors? In this talk, I will review some basic tools from Bayesian inference and sequential estimation, and demonstrate their application in a vortex estimation problem. Since most of the challenge of applying these tools arises from the non-linearity of fluid dynamic problems, I will devote attention to the techniques we use to overcome those challenges.

Biography: Jeff Eldredge is Professor of Mechanical & Aerospace Engineering at the University of California, Los Angeles, where he has served on the faculty since 2003. Prior to this, he received his Ph.D. from Caltech, followed by post-doctoral research at Cambridge University. His research interests lie in computational and theoretical studies of fluid dynamics, including numerical simulation and low-order modeling of unsteady aerodynamics; investigations of aquatic and aerial locomotion in biological and bioinspired systems; and investigations of biomedical and biomedical device flows. He is the author of numerous papers, as well as the book Mathematical Modeling of Unsteady Inviscid Flows. He is a Fellow of the American Physical Society, an Associate Fellow of AIAA, and a recipient of the NSF CAREER award. He has served on the Editorial Board of Physical Review Fluids and as an Associate Editor for the journal Theoretical and Computational Fluid Dynamics.

Abstract: Metachronal motion is the most common and effective form of motion found in crustaceans to increase efficiency. This study examines the flow dynamics linked with krill's metachronal hovering. Krill, consisting of the five legs distributed along its body, can drive itself by properly coordinating their body and legs (pleopods), called metachronal motion. We will discuss the metachronal paddling motion of these legs and its role in producing two major types of forces: Added mass effects resulting from the inertia of the flow and vortex-induced forces generated from vortices. We will examine how the different kinematic parameters of the metachronal hovering mode affect the contribution of these forces and discuss the consequence of the coordinated opening/closure of the legs in shaping the wake and optimal hydrodynamic force generation at different Reynolds numbers.

Abstract: Wave-assisted propulsion (WAP) systems directly convert wave energy into thrust using elastically mounted hydrofoils. The wave conditions as well as the design of the hydrofoil drives the fluid-structure interaction of the hydrofoil and consequently, its performance. We employ simulations using a sharp-interface immersed boundary method to examine the effect of three key parameters on the flow physics, the fluid-structure interaction, as well as thrust performance of these systems - the stiffness of the torsional spring,  the location of the pitch axis and the Strouhal number. We demonstrate the utility of ‘maps’ of energy exchange between the flow and the hydrofoil system, as a way to understand and predict these characteristics. The force-partitioning method (FPM) is used to decompose the pressure forces into interpretable components and to quantify the mechanisms associated with thrust generation. Next, the multifoil system is investigated where the wake of the leading foil is utilized to enhance the thrust of the trailing foils.

Abstract: The successful deployment of oscillating hydrofoil turbine arrays depends on optimizing array configurations. This relies on understanding how the performance of different foils within an array is affected by the vortex-dominated wake behind each hydrofoil. In this work, an array of oscillating hydrofoil turbines in tandem configuration is experimentally studied to determine the optimal kinematics of the array. By characterizing wake-foil interactions between the leading foil-produced wake and the trailing foil, the kinematic configuration each foil in the array must have to maximize array power extraction is determined. This is done by prescribing leading foil kinematics that produce specific wake regimes and evaluating their effect on the performance of the trailing foil subject to different sets kinematics. Performance is evaluated through the power extracted by the foil over an oscillation cycle through force and torque measurements. Wake-foil interactions that lead to improvements in trailing foil performance are analyzed with time-resolved Particle Image Velocimetry. Constructive and destructive wake-foil interactions are compared, and it was determined that trailing foil performance could be improved either by avoiding interactions with wake vortices or by interacting directly with them. Additional insight is gained from applying the Force and Moment Partitioning Method to the experimental data.

Abstract: Turbulent flows are characterized by the dynamic coupling of widespread spatial and temporal scales. Non-intrusive diagnostic techniques such as Particle Image Velocimetry (PIV) can provide correlated spatiotemporal flow measurements but are often limited by temporal resolution. This work introduces a method for recovering spatiotemporal information from limited flow realizations through novel specification of a local convective velocity. Using the least squares minimization of the linear advection equation, the convective velocity of all scales is estimated. The definition is based on time signals and their local spatial derivatives, making it well-suited towards streamwise heterogeneity and spatially developing flows. Using a simplified semi-Lagrangian numerical technique, the convective velocity is used to estimate unknown velocity fields at intermediate times between successive flow measurements. The technique is assessed using three-dimensional direct numerical simulation (DNS) of a planar jet with Re = 10,000. Spectral analysis revealed that the energy content at frequencies several orders of magnitude beyond the Nyquist rate can be recovered. Furthermore, the results indicate that estimations based on local convective velocities are more accurate than those based on the local mean flow, particularly for recovering energy at the characteristic frequency of shear-dominated flows.

Abstract: Fish can significantly improve their swimming performance if their kinematics and paths are adequately adapted to the incoming flow. When the optimum path is unknown, the first step is to predict the performance that could be obtained for several route candidates. Next, we can select the trajectory that increases the lift or efficiency. We develop several force models with Koopman-based system identification tools to predict the CL evolution of a foil executing transitions in its motion inside an unsteady incoming wake. We test the optimum-path selection capabilities by predicting the CL evolution for a set of route candidates, and rank them based on CL production. We find that adding physically relevant information about the wake in the models helps to find the optimum path, achieving a correlation of 90% with numerical target data. Next, we optimize a more general form of the transition trajectory to minimize power consumption using the previously developed force models and evolution algorithms. We conduct CFD simulations and experiments to asses the performance of the optimiser in finding the most efficient transition inside the wake. 

Abstract: Bioinspired by flapping leaves, splitted-plates mounted on the windward side of a cylinder (inverted configuration) are potentially useful for energy harvesting applications. This work demonstrates the flow-induced rotational vibrations (FIRV) of an inverted rigid cylinder-plate, a simplified model of flexible inverted plates. Such a model is computationally less expensive, and we aim to address some open questions regarding the dynamics of such a class of FSI problems. For instance, whether the large amplitude flapping is VIV or flutter. We observe several regimes in the FIRV response: (i) Initial excitation, (ii) Intermediate symmetry-breaking, (iii) First chaotic regime, (iv) Lock-in, (v) Second chaotic regime, (vi) Final symmetry-breaking. We present a quasi-static aeroelasticity approximation-based mathematical model to understand better the dynamic response and the effects of various parameters, such as the mass ratio. We find that the cylinder-plate’s large-amplitude flapping motion is caused due to VIV, not flutter. A quantification of chaos is presented and the vortex-shedding patterns are discussed eventually.

Abstract: Modern small aircraft are asked to operate under severe atmospheric conditions in urban areas and turbulent wakes behind large structures. While understanding interactions between a strong gust and a wing is important, sweeping over the huge parameter space of extreme aerodynamic flows with expensive simulations and experiments is impractical, calling for data-driven approaches. We discuss how such complex aerodynamics under strong vortex gust-airfoil interactions can be analyzed in a low-order manner with nonlinear machine learning. For this study, we consider wakes over a NACA0012 airfoil at Re = 100 covering a range of angles of attack with a strong disturbance modeled by the Taylor vortex, producing a variety of complex wake patterns due to vortex-airfoil interaction. Such unsteady and violent vortical flows over the parameter space are compressed into merely three variables with a lift-augmented nonlinear autoencoder, revealing a low-rank manifold. Furthermore, the discovered manifold captures the fundamental physics of nonlinear interaction under extreme aerodynamics. Towards the end of the talk, we also show that the present approach can be used for real-time sparse sensor-based state estimation and fast control to significantly mitigate the impact of vortex gusts with local actuation.

Abstract: Natural undulatory swimmers are observed to adapt their waveform kinematics when migrating or when swimming against strong currents. To characterise the effects of waveform kinematics on the swimming performance of undulatory swimmers, we designed a bio-inspired anguilliform robot. We measured the robot’s swimming speed, efficiency, in terms of the cost of transport, and body kinematics in free swimming experiments, for a broad range of kinematic parameters, including joint amplitude, body wavelength, and frequency. We find that speed, in terms of stride length, increases for increasing maximum tail angle, described by the newly proposed specific tail amplitude. Maximum stride length is reached for specific tail amplitudes around unity. Minimum cost of transport requires a lower specific tail amplitude and body undulations close to pure travelling waves. Live anguilliform swimmers display a range of specific tail amplitudes that match our robot’s efficient regime, suggesting similar mechanisms of efficient locomotion. The results improve our understanding of anguilliform swimming and provide guidelines for improved design of undulatory swimming robots.

Abstract: Bubbly turbulent flows occur in many engineering and environmental applications. In these flows, an important aspect is the interaction of bubbles with vortical structures, where bubble deformability and size play a crucial role. To understand these complex interactions, we experimentally study an idealization, namely, the interaction of a single air bubble with a single water vortex ring, with the focus being on the effects of bubble deformability and bubble-to-vortex size ratio on both bubble dynamics and vortex ring dynamics. During these interactions, the bubble dynamics are captured using high-speed imaging, while the effects on the ring’s vortex core are captured using time-resolved PIV. 

In the talk, I will focus on the interaction of a deforming air bubble and a rigid buoyant particle (a rigid bubble), with a single water vortex ring, with deformability being the distinct difference between the two. Both the buoyant particle/ deforming bubble are captured by the vortex ring, due to the low pressure within its core, leading to a strongly coupled interaction between the bubble/buoyant particle and the vortex ring. The deforming bubble undergoes elongation both outside the ring during its capture, and within the ring after capture, with the latter stage leading to bubble break-up within the ring. In contrast, the rigid bubble remains spherical during capture, and stays more localized within the ring. These differences in deformability lead to distinct differences in the ring’s convection speed, azimuthal vorticity, and enstrophy. These results could have implications in bubbly turbulent flows such as in bubble drag reduction where many studies indicate that deforming bubbles are better for drag reduction than nondeforming ones.

Abstract: Cyclorotors are a propulsion system that use several rotating, periodically pitching blades to produce a net force in a single direction. Previous studies have found that they exhibit improved performance at low Reynolds Numbers compared to conventional rotors, making them of much interest for the field of Micro Unmanned Aerial Vehicles. Due to the curvilinear flow they are subjected to, the blades experience an effect known as virtual camber which has an impact upon cyclorotor efficiency. In theory, this effect could be countered by cambering the blades in the opposite direction, but previous studies have found this to have detrimental effects on the blade’s capability to generate a leading edge vortex (LEV), which can aid their force production. As an alternative, this study looks into using blades that only camber their trailing edge portion, leaving the leading edge portion intact for LEV production.  The effects of this camber on force production, efficiency and flow structures were investigated on a cyclorotor in the hover configuration, through a combination of dynamic force measurements coupled with particle image velocimetry.  The results show that the leading edge vortex is preserved when only the trailing-edge is cambered as opposed to the fully cambered blade. However, this still leads to an overall reduction in force production and efficiency compared to the symmetric blades.

Abstract: While several previous studies considered tandem cylinders coupled through flow between them, a hitherto unexplored elastic coupling with fluid flow between them significantly influences FIV. Therefore, we numerically study the transverse flow-induced vibration (FIV) of elastically coupled tandem cylinders at Reynolds number 100 using ViCar3D. A systematic comparison between the classic elastically mounted tandem cylinders and elastically coupled cylinders is presented. The latter configuration exhibits two vibration modes, in-phase and out-of-phase, with corresponding natural frequencies approaching the Strouhal frequency of the system. Using the obtained results, we answer the following questions. (a) What are the FIV regimes associated with elastically coupled tandem cylinders? (b) What is the mechanism for lock-in of classic and elastically coupled tandem cylinders? (c) What are the participating wake modes in a quasi-periodic FIV regime? (d) Does the gap vortex formation modify by the small amplitude FIV response? (e) What is the effect of the relative motion of the cylinders on galloping response? (f) Can the system be used for undamped FIV suppression/energy harvesting applications?

Abstract: The drag force on an accelerating object is usually described by a quasi-steady force that scales with the square of the instantaneous velocity and an added mass force due to the acceleration. This description could lead to a significant underestimation of the actual drag force. We aim to find a better description of the drag force on a flat plate in an unsteady flow by measuring both the drag force and velocity field for a large range of constant accelerations and velocities. Our experiments show that the force due to acceleration does not scale linearly with acceleration, contrary to what is expected from added mass. We associate this force to the generation and advection of vorticity at the plate surface and combine this into a single scaling law model, based on the history force for unsteady flow. This scaling avoids previous inconsistencies in using added mass forces in the description of forces on accelerating plates. This new scaling law has proved useful in predicting the drag force and total circulation in the wake of different plate geometries and for non-constant accelerations.

Abstract: We investigate theoretically and experimentally the stability of three interlaced helical vortices with respect to displacement disturbances whose wavelengths are large compared to the size of the vortex cores. A space-time analogy is used to present a model for the spatial evolution of semi-infinite helical vortices subjected to time-periodic disturbances. This model is used to investigate the spatially evolving wake of a three-bladed turbine, which is subject to the global or local pairing instability that represents different wake control strategies. Perturbations modify the minimal distance between neighboring pairs, but also their relative orientation, i.e., locally parallel, antiparallel, or perpendicular.  This is relevant for vortex merging, but also for the development of elliptical and Crow instabilities.

Abstract: Seals use their specialized whiskers to detect and follow prey underwater. Previous studies have revealed that the harbor seal whisker morphology suppresses lift oscillation and vortex introduced vibration through features including flattened cross section, undulatory thickness, and a unique 180° frontal-dorsal phase difference in the undulation. These features have inspired many flow sensor designs. However, to understand the broad range of flow and signal correlation experienced by the seal whisker and to optimize the design and expand the application of inspired sensors, systematic investigations into the vortex-introduced vibration (VIV) of seal whiskers are needed. In this study, the single degree-of-freedom (cross flow) VIV of a harbor seal whisker is solved using direct numerical simulation for parametrically varied reduced velocity and angle of attack (AOA) at a constant Reynolds number of 300. The whisker is modeled as a rigid body with a mass ratio of 1 and a damping ratio of 0.02. The results have revealed an abundance of modes of response in the parametric space. In this talk, we will elaborate on the vortex shedding modes and how they compare to those from basic shapes like circular and elliptical cylinders. Particularly, we will show how the dynamic mode decomposition (DMD) of the flow field helps to identify dominant modes from complex three dimensional wake structures.

Abstract: The influence of initial geometry on the formation, development, and decay of a vortex loop behind a polygonal disk at a Reynolds number Re = 20, 000 is investigated. Using particle image velocimetry in 5° azimuthal increments, we reconstruct a 360° visualization of the vortex loop generated behind square and circular disks starting impulsively from rest. The coherence of the loop is strongest behind a circular disk, remaining strong and maintaining its shape far downstream. Meanwhile, as the number of sides of the polygonal disk decrease, topological changes to the vortex loop are observed, and a faster transition to turbulence, as compared to the disk.

Abstract: Many engineering applications involve flow past a moving boundary or interface, which can be challenging to simulate with a body fitted mesh. Immersed methods avoid the need for mesh generation altogether, but these methods have so far been limited to second order accuracy when applied to 3D flows with moving boundaries. In this talk we present our progress towards an immersed interface method that achieves high order spatial and temporal accuracy in 3D simulations with complex geometries and moving boundaries. Our spatial discretizations combine dimension-split finite differences with high order weighted least squares interpolants near immersed boundaries, yielding third order spatial accuracy for advection terms and up to sixth order spatial accuracy for diffusion terms or elliptic PDEs. We also discuss the issue of "freshly-cleared cells" in sharp interface methods with moving boundaries, and present a treatment that maintains the temporal order of accuracy of arbitrary explicit Runge-Kutta schemes (as well as preliminary results for diagonally implicit schemes). We conclude by applying this approach to a 3D immersed interface discretization of the advection-diffusion equation on a multiresolution grid, which captures thin boundary layers on immersed surfaces through a combination of high order discretizations and grid adaptivity.

Abstract: Modern engineering devices such as wind/tidal turbines and micro air vehicles experience unsteady loads due to turbulence and gusts. This unsteady loading may fatigue turbine blades or the vehicle's wings over time and therefore the mitigation of unsteady load is critical for these devices. Natural flyers such as insects and birds acquire both active and passive systems to control their position and velocity, and to mitigate the impact of gusts and lulls.  Active control systems are established in engineering, whereas passive control systems are not fully understood. To address this, we investigate the unsteady load mitigation mechanism of an aerofoil made of a passive trailing-edge, which spans 2/3 of the chord. The aerofoil undergoes sinusoidal heaving motion at a mean angle of attack of 6 degrees and Reynolds number of the order of 10 thousand. This study reveals that the deflection and unsteady load mitigation is scaled with the product of two different Cauchy numbers, which are the non-dimensional number comparing the fluid force and elastic force. These results will pave the way for the development of low-order models to better predict passive deflection and unsteady loading.

Abstract: This study examined lift forces on two nonslender delta wings with sweep angles of Λ = 40° and 50°, both having flags attached to the leading edge. Notable increases in lift were observed in the post-stall regime for the clean wings, as well as delay in stall. When a flag was present, flow field measurements indicated that the shear layer reattached to the wing surface and the leading-edge vortex re-formation was observed. In contrast, wings without flags experienced a completely stalled flow under the same free stream conditions. Across various angles of attack, mass ratios, and flag lengths, maximum lift improvement was achieved when the flag oscillated at an optimal dimensionless frequency range, and the tip velocity of the flag had a sufficient amplitude. This optimal frequency range was consistent with the natural shear layer instabilities found for the clean wings. Excitation provided by the flag oscillations within this range significantly improved the coherency of the otherwise separated flow. The main mechanism in increasing lift appears to be the excitation of the shear layer, which exhibits a convective instability. This differs substantially from lift enhancements on airfoils, where flags lock-in to the wake instability, which is known to have an absolute instability at the post-stall angles of attack for the clean airfoil.

Abstract: The stomach is responsible for mixing, grinding, sieving, and chemically breaking down food. The peristaltic motion of the stomach walls combines with the secreted gastric enzymes to physically and chemically breakdown the ingested food. However, experimental investigation of these phenomena is challenging and cost-intensive. This study presents a computational model of the stomach based on imaging data. An enzyme is secreted from the stomach walls that mixes with the contents and hydrolyzes the protein in the liquid meal. We also use the model to study the dissolution of an orally ingested pill and the subsequent delivery of the dissolved drug into the intestines. The effects of weaker motility, caused by disorders such as gastroparesis, on the flow field inside the stomach and on the rate of protein hydrolysis and drug dissolution are analyzed. The findings demonstrate the potential of a computational approach in this field in quantifying the effects of disease and dysfunction.

Abstract: We investigate the hydrodynamic implications of organised surface textures on aquatic animals. This study considers a modified NACA foil with an egg-carton type roughness at Re=100,000. On the foil, we prescribe general undulatory kinematics that result in zero net thrust. In this presentation, we investigate a reduction in the required power for a roughness wavelength corresponding to λ=L/128. First, we use dynamic mode decomposition (DMD) to identify coherent flow structures; we stabilise the boundary layer by applying the body motion's inverse map to the flow field. The DMD analysis identifies distinct flow structures present in the smooth and power-reducing roughness topology that can be likened to Helmholtz rollers. Next, we identify sensitivities in the cases by using the DMD basis to perform resolvent analysis. Finally, we correlate resolvent modes with power frequency spectra, enabling targeted visualisation of flow structures at specific frequencies. The study concludes that optimally tuned rough surfaces do more than merely increase frictional drag; they also alter flow structures, occasionally increasing the performance of the swimmer.

Abstract: There are over 3000 snake species, each exhibiting various lifestyles, including terrestrial, amphibious, and marine behaviors. Throughout more than 100 million years of evolution, their elongated and limbless body structure has remained remarkably preserved. Despite their seemingly simple morphology, snakes have adapted and developed more than 10 distinct gaits to navigate diverse environments. For instance, some snakes like the dice snake (Natrix tesselatta) prefer swimming fully submerged to forage, while others like the garter snake (Thamnophis sauritus) are more inclined to swim on the water's surface.

We proposed to visualize the wake produced by these two swimming snakes using two different methods. We used Defocused Digital Particle Tracking Velocimetry (DDPTV) to observe the vortex structures created by Natrix tesselatta swimming underwater. The waves produced by Thamnophis sauritus swimming on the surface were measured using synthetic Schlieren imaging.

The PIV revealed the creation of multiple vortices along the body of the snake due to its undulation. The 3D structure of the vortices generally consisted of paired vortex tubes, some of which were linked together to form a hairpin structure. The observations match predictions from computational fluid dynamic studies of other anguilliform swimmers. The Schlieren imaging allowed us to observe that the snake was simultaneously producing dragging and thrusting waves, contributing to the general propulsion of the animal. 

Abstract: Passive flow control via fluid-structure interaction (FSI) is a promising control paradigm for unmanned aerial vehicles operating in vortex-dominated low Reynolds number regimes. The associated aerodynamic flows are unsteady, high dimensional and nonlinear. A flexible structure therefore has the potential to passively alter key unsteady vortex structures through its vibrations, if its intrinsic modal dynamics are carefully aligned with the driving flow processes. Towards this aim, we perform high-fidelity numerical FSI simulations to study the dynamic interplay between a separated aerodynamic flow with vortex shedding content (Re=500 and AOA 15) and a flexible flat plate modeled using linear Euler-Bernoulli beam theory. We focus on how the flow dynamics and associated lift are informed by the structural modal shape(s) and deformation timescales of the structural response, and how in turn this is informed by the structural parameters relative to the vortex shedding behavior of the baseline reference (rigid plate) case. The ending long-term dynamics are then found to be characterized by the natural frequencies of the beam, taking appropriate amount of fluid effects. These long-term limit cycle dynamics are described in detail by this characterization, drawing connections to the baseline rigid case as appropriate.

Abstract: Recent advancements in small size Unmanned Aerial Vehicles (UAVs) have led to their increased popularity due to their versatility, including vertical takeoff and landing (VTOL) capabilities and low operating costs. Although UAVs emit less noise than helicopters, their increased use, especially in densely populated areas at lower altitudes, will likely lead to stricter noise regulations. As the rotors still remain as one of the major noise source, understanding rotor noise generation at these low rotational speeds is crucial for developing certified systems that meet these standards. To gain insight on the underlying noise generation mechanisms, high-fidelity numerical simulation is employed to study the aeroacoustics of a rotor operating in the transitional regime. The investigation focuses on a two bladed NACA0012 rotor with a constant pitch angle. Result shows distinctive flow structures along its span. Inboard, attached laminar flow is observed, while midspan to the blade tip region experiences flow separation with 2D and 3D flow structures.

These distinct regions along the span leave characteristic signatures on the near and far field acoustic spectra. Spectral analysis of pressure fluctuations at the blade surface and in the near wake reveals tonal peaks at the blade passing frequency and harmonics, along with a broadband hump at higher frequencies. These findings are in line with the experimental observation. The broadband hump originates from separated shear layers on both the pressure and suction sides, while peaks within the hump also arise from the leading and trailing edges due to shear layer instabilities and their interaction with the following blade's edges, including the tip vortex and wake.

Abstract: During the COVID-19 pandemic, face masks were the first line of defense for the global population to reduce the spread of the virus. However, the effectiveness of face masks can be affected by a number of factors, including how the mask is worn, the type of mask, and the wearer's activities. One of the most common activities that can affect the effectiveness of a face mask is talking. When people talk, the morphology of their faces changes, which can cause the mask to move and create gaps that allow respiratory droplets to escape. This is especially true for people who wear cloth masks, which are less effective at filtering out respiratory droplets than surgical masks or N95 respirators. To explore this problem systematically, we have developed a new method for modeling the movement of facial features during talking. This method, called geometrically weighted principal component analysis (GWPCA), can be used to create a low-complexity model of the face features. The GWPCA model can then be used to create a dynamically moving facial shape for identified categories of talking syllabi. This information will be used to explore the fluid dynamics of facemasks.

Modeling the airflow between the face and a mask is very complex. This is because it depends on the shape of the face, the type of mask, the wearer's activities, and other factors. An analytical integral boundary layer model is pursued here to quantify the flow in the interface region between the face and mask using mask shape on a moving face from a detailed deployment mask mode. An analytical model is validated with a detailed flow simulation and then employed to find the relationship between the fitness of the mask during talking, mask porosity, and its level of leakage. The research outcome will be a fast  model that enables improved operational forecasting of respiratory disease spread. It will directly lead to better epidemic planning in the future and provide a technique to quantify the efficacy of mask recommendation strategies for the diverse population in real time during future pandemics. 

Abstract: Impulsive flows happen in various circumstances of interest to engineering applications. However, the direct measurement of the hydrodynamic load acting on a body is not always possible for different reasons, for example the wake sensitivity of sensors. Moreover, this force measurement does not allow to correlate the story of the force with the physics of the flow. As an alternative, the feasibility of applying a planar control-volume approach to time-resolved PIV (TR-PIV) experimental data is investigated. The focus is on reconstructing the unsteady load acting on an accelerated flat plate normal to the flow. Although the flow is in the turbulent regime (Re = 18×10^3), the assumption of two-dimensional flow is made at the mid-plane of a plate with aspect ratio AR = 5. Three different kinematic cases have been considered by varying the value of acceleration and keeping the same target velocity. The method adopted here consists of using the control volume approach to estimate the force, in combination with a Poisson solver for the pressure field. The method was tested on direct numerical simulation (DNS) data of an impulsively accelerated flat plate where different experimental uncertainties are taken into account to mimic real PIV data. This allows us to identify sources of error and how they contribute to the final result. In both the numerical and experimental data the pressure term appears to give the largest contribution to the noise in the total force. Then, the force estimation based on experimental PIV data is shown in comparison with the force measured by a force sensor. The results match the measured force for the first part of the motion, where the unsteady term and the pressure term dominate the drag force.

Abstract: Generally, natural fliers demonstrate greater flight control than manufactured aerial vehicles. Elastic joints allow natural wings to passively respond to gusts enabling steady flight in turbulence. To investigate the underlying physical principle, we consider the gust response of a fixed foil and a foil that can passively and elastically pitch. The incompressible Navier-Stokes equations are weakly coupled with a rotational harmonic oscillator within the open-source toolbox OpenFOAM. A 2D NACA0012 foil is modelled at an initial chord-based Reynolds number Re = 10^3, at an angle of attack of 5°. The initial spring moment is set to balance the fluid dynamic torque. The inflow velocity is doubled within one convective time following a hyperbolic tangent law. The results show that a passively-pitching foil experiences a total load fluctuation that is up to seven times lower than that of a fixed-pitch foil. Different gust types, including reducing inflow speed and changing inflow direction, have been examined, and results show similar effect on mitigating unsteady loads through passive pitch. We demonstrate through an analytical model that the quasi-steady variation of any force component, e.g. the lift, can be cancelled if the pitch axis lies along a semi-infinite line from the foil. We show how the pitching axis location influences the efficacy of the unsteady load mitigation for different gust types. We extend these results by testing selected cases at Re = 5*10^4 and Re = 10^6, where the same trends are observed. Furthermore, when the gust triggers vortex shedding on the fixed foil, this is suppressed if the foil is allowed to passively pitch. These results provide new insights into the design of a passive system to mitigate unsteady loads; a system that may enhance the controllability and resilience of aerial and underwater vehicles, and turbine blades.

Abstract: In this study, we introduce a cluster-based decomposition technique to create data-driven, coarse-grained representations of nonlinear dynamics in complex systems subjected to external forcing. We collect time-series measurements by simulating the system with added external forcing and apply k-means clustering for coarse-graining. We represent the measurement data as a weighted sum of the centroids associated with the clusters and train the corresponding weights using sparse regression. This results in a deterministic, nonlinear predictive model for the time-series measurements' evolution, contrasting with probabilistic linear Markov models found in previous studies. We integrate this approach with model-predictive control (MPC) to guide the system towards favorable states. We first demonstrate the method using the Lorenz system, then apply it to control the laminar flow over a flat plate at a high angle of attack (35 deg) by employing momentum injection. This approach effectively achieves high lift and low drag states.

Abstract: The hydrodynamic interactions between individual swimmers can lead to the formation of stable schools. These interactions, which occur between fish bodies and vortices shed by other fishes, are determined by the kinematics and spacing of neighboring swimmers. In this experimental study, we investigate how these interactions are affected by tailbeat frequency for a pair of pitching hydrofoils. We found that equilibrium constellations of the two-fish school can be manipulated by changing the pitching frequency. We analyzed these constellations using multi-layer stereo PIV to capture three-dimensional flow structures. Understanding schooling mechanisms at high frequencies can provide insights into the schooling of high-speed fish species such as tuna, as well as benefit the design and control of future high-speed multi-agent bio-inspired robotic platforms.

Abstract: We will explain the basic concepts of network theory and discuss the benefits of applying network-based analysis to fluid flows. We will review recent advances in the field, including the use of network measures to identify important flow structures, detect transitions in flow behavior, and analyze interactions in laminar and turbulent flows as well as fluid-structure interaction systems. We will also highlight some challenges in applying the network-based methodology, including the need for appropriate data acquisition and processing methods, and the need to develop more sophisticated network models that can capture the full complexity of fluid dynamics.

Biography:  Aditya G. Nair joined the Department of Mechanical Engineering at the University of Nevada, Reno in August 2020. His research interests are in the areas of computational fluid dynamics, unsteady aerodynamics, high-performance computing, data science, and control theory focused on modeling and control of high-dimensional fluid flow physics. Aditya G. Nair received his Ph.D. in mechanical engineering from Florida State University (Tallahassee) in 2018. Prior to this, he completed his M.S. in mechanical engineering from the University of Michigan (Ann Arbor) in 2013 and a B.E. in mechanical engineering from the University of Mumbai in 2011. Following his Ph.D., he served as a post-doctoral research associate in the mechanical engineering department at the University of Washington (Seattle) till July 2020. 

Abstract: The role of compressibility on airfoil flutter in the transonic regime is investigated with a series of two-dimensional direct numerical simulations based on an immersed boundary method. An energy map approach based on forced kinematics is used to identify limit cycle oscillations and quantify the influence of Mach number and spring stiffness on pitching amplitude fluctuations. A subcritical instability is observed for Mach numbers close to 0.7 which covers a range of frequencies. Three shock-induced mechanisms are identified that influence the extent of energy imparted from the flow to the airfoil. The primary mechanism is flow separation triggered by the generation of a lambda shock. The influence of the lambda-shock dynamics on airfoil flow separation is characterized for different Mach numbers providing insight into the "transonic dip" observed under the current flow conditions.  

Abstract: Unsteady separated flow is present on many physically large objects. Low-fidelity simulations cannot accurately estimate the influence of separated flow. The physical scale of some objects precludes full-scale testing in wind tunnels and complicates the acquisition of performance data in operational conditions. Small-scale wind tunnel tests can provide an estimate of the aerodynamic loading. Tufts can be used to qualitatively validate the surface flow between different investigations and the full-scale object in operational conditions. In the current investigation it is shown, with computational and experimental data, that tufts can provide a quantitative estimate of the unsteady wing loading. Unsteady 3D simulations of a NACA0012 wing, at different angles of attack and Reynolds numbers, are used to obtain unsteady surface flow and lift coefficient data beyond stall. The computational data provides a proof-of-concept by using a linear surrogate model based on pseudo tuft orientations, this model is then extended with a non-linear component. Experimental data of a NACA0012 wing, equipped with tufts and a force balance, has been used in combination with neural networks to infer quantitative information about the unsteady wing loading. This results in the ability to capture non-periodic lift and pitching moment fluctuations based on visual tuft observations.

Abstract: Vortex tubes with initial axial core-size variations develop twist waves, and collision of such twist can lead to a sudden radial expansion of the vortex core known as vortex bursting. For sufficiently large initial core-size variations and Reynolds numbers, bursting leads to significant increases in enstrophy and has the potential to destabilize the vortex core. Using direct numerical simulations of the 3D Navier-Stokes equations, we have simulated and analyzed vortex bursting on tubes with circulation-based Reynolds number of 5000. For a baseline analysis, we have considered bursting on rectilinear vortex tubes. Currently, we are probing the stability of the mechanisms involved when bursting occurs on vortices with curved centerline geometries, such as helical vortices and vortex rings. The results indicate that interaction between the bursting structure and the core dynamics affects the details of the flow evolution and further destabilizes the bursting structure and the core structure. In this talk I will present an overview of our results of vortex bursting on straight and curved vortices.

Abstract: We present a series of Direct Numerical Simulations of the planar flow past an impulsively started cylinder at Reynolds numbers up to 1,000,000. An intriguing portrait of unsteady separation is revealed; vorticity generation and vortex shedding entail a cascade of separation events on the cylinder surface that are reminiscent of Kelvin-Helmholtz instabilities. Primary vortices roll-up along the cylinder surface as a result of instabilities of the initially attached vortex sheets, followed by vortex eruptions, creation of secondary vorticity and formation of dipole structures that are subsequently ejected from the surface of the cylinder. The vortical structures and their relationship to the forces experienced by the cylinder are analyzed.  

Abstract: Characterizing the aerodynamic response of wings to oncoming gusts is critical to maintaining stability and efficiency of aircraft. In this study, surface pressure and particle image velocimetry measurements are used to analyze the unsteady flow physics of a NACA 0015 wing in a time-varying freestream flow. Unsteadiness exhibited in the wing's aerodynamic response to velocity acceleration and deceleration can be attributed to the dynamics of developing vortical structures. Whether the flow is accelerating or decelerating determines the temporal and spatial scales of the vortical structures, including the convective time, size, and location from which vortical structures develop and shed. These scales determine the degree to which vortical structures interact with each other and with the wing surface, thereby also influencing the unsteady loading on the wing.

Abstract: Mechanical Aortic Valves (MAVs) are routinely implanted as permanent replacements for malfunctioning or diseased human valves.Compared with bioprosthetic ones, MAVs are more durable and insusceptible to tearing and calcification, nevertheless they promote non-physiological hemodynamics which might lead to platelet activation and mechanical hemolysis. Previous research has correlated the blood damage to augmented levels of turbulent stress downstream of MAVs when compared to bioprosthetic devices. In this scenario, we  numerically investigate two emerging technologies proposed for mitigating such detrimental effects on hemodynamics: a tri-leaflet configuration and a bi-leaflet valve with vortex generators. Simulations are carried out by means of a finite difference flow solver with immersed boundary forcing. When compared to the baseline design, vortex generators are found to anticipate the break-up of the shear layers downstream of the leaflets, lowering the overall turbulent shear stress at peak flow rate. Conversely, the trileaflet design provides comparable haemodynamics at peak flow rate, but further reduced stress levels in the deceleration phase. The findings of this study could be used to improve the design of next-generation MAVs in order to reduce the risk of thromboembolic complications.

Abstract: This study is an extension of our previous work on surface actuation as a lift improvement strategy for flow past an airfoil at Re 1000. In the current work, we employ gradient based optimization to determine the spatio-temporal properties of actuation. With actuation introduced on the entire surface of the airfoil, we explore the configuration of actuation for lift and drag improvements separately. It is found that actuation has distinct behavior for the two performance imperatives. Further, actuation for lift improvement results in a drag penalty and vice versa. Apart from highlighting the differences in temporal variation of the aerodynamic coefficients for each case, our investigation explores the critical spatial locations on the airfoil surface, the temporal variation of actuation in relation to the underlying shedding process and key flow features for each performance goal. 

Abstract: Helical vortices, such as those generated in the wake of a rotor, are subject to various instabilities including displacement instabilities, which occur when the vortex core is shifted from its baseline position. After being perturbed, the vortices deform and begin to break down. The zero-wavenumber displacement instability mode can be triggered by introducing an asymmetry to the rotor producing the vortices. The vortex dynamics in this case are highly complex, so a simplified model based on an infinitely-repeating strip of point vortices is developed to reproduce the nonlinear instability evolution. The model is validated against a more sophisticated filament model and water channel experiments, showing remarkable agreement for a range of parameters relevant to industrial rotors. It is then used to investigate the effectiveness of different types of rotor asymmetries at accelerating vortex breakdown. Even small initial displacements around 5% of the vortex spacing substantially disturb the vortices, and the direction of the perturbation plays an important role in the speed of the instability development. These findings can then be used to design wind turbine rotors that minimize the detrimental effects of their wakes on downstream turbines within a wind farm.

Abstract: It is well established that a flat plate impulsively set into linear motion will result in the formation of a clockwise vortex, also known as the starting vortex. This study is aimed at characterizing the resulting starting vortex and shear layer in the wake of the airfoil based on three initial conditions, namely, the angle of attack, the surge speed and the surge distance. It will be shown that the circulation of such a starting vortex is a fraction of Wagner’s prediction. Further, it was observed that the shear layer breaks into secondary vortices beyond a critical surge speed. The primary vortex in this case is also characterized.

Abstract: Dynamic stall is known to be a ubiquitous phenomenon in rotary flows, leading initially to an overshoot in performance followed by a dramatic loss in such performance. Being able to prevent or delay this performance loss via flow control mechanisms is highly dependent on understanding the underlying flow physics at play, including identifying the dominant unsteady scales associated with the dynamic stall vortex shedding process.  Experimental particle image velocimetry data from a series of water tunnel experiments is used to study the flow evolution around a dynamically pitching NACA 0012 airfoil at low Reynolds number. The airfoil underent a linear pitch ramp maneuver at a fixed dimensionless pitch rate of 0.05 across three Reynolds number to reproduce a canonical light dynamic stall process. The primary objective of this study is to assess the scalability of the dominant scales associated with the flow perturbations across multiple Reynolds numbers. The wavenumber scales associated with this canonical dynamic stall process were extracted using a combination of two-dimensional Log Gabor filtering scheme and Riesz transform. From the transverse velocity spectra, the fluctuations in the flow were observed to reach an amplified state during the initial ejection of vorticity from the leading-edge region of the airfoil.

Abstract: Wing-gust encounters cause harmful lift transients that can be mitigated through maneuvering of the wing. This work presents a method to generate an open-loop (i.e., prescribed) maneuver that optimally regulates the lift on the wing during a transverse gust encounter. Obtaining an optimal maneuver is important for laboratory experiments on the physics of wing-gust interactions and may be useful for the future design of feedback controllers. Prior work of us has shown that an Iterative Maneuver Optimization (IMO) framework can generate an optimal maneuver by using a surrogate model to propose a control signal that is then tested in experiment or high-fidelity simulation. The input to the surrogate model is updated to account for differences between the test data and the expected output. The optimal maneuver is obtained through iteration of this process. This paper simplifies the IMO method by replacing the surrogate model with the classical lift model of Theodorsen, removing the process of optimization over the surrogate model, and removing the requirement to know the time-averaged profile of the gust. The proposed method, referred to as Simplified IMO (SIMO), only requires input and output data collected from simulations or experiments that interact with the gust. Numerical simulations using a Leading Edge Suction Parameter modulated Discrete Vortex Model (LDVM) are presented to generate the input and output data of the wing-gust encounters for this paper. The results show an optimal pitch maneuver and an optimal plunge maneuver that can each regulate lift during a transverse gust encounter.

Abstract: The optimal stiffness for soft swimming robots depends on swimming speed, which means no single stiffness can maximise efficiency in all swimming conditions. Tunable-stiffness would produce an increased range of high-efficiency swimming speeds for robots with flexible propulsors and enable soft control surfaces for steering underwater vehicles. We propose and demonstrate a method for tunable soft robotic stiffness using inflatable rubber tubes to stiffen a silicone foil through pressure and second moment of area change. We achieved double the effective stiffness of the system for an input pressure change from 0 to 0.8 bar.  We achieved a resonant amplitude gain of 5 to 7 times the input amplitude and tripled the high-gain frequency range compared to a foil with fixed stiffness. These results show that changing second moment of area is an energy effective approach to tunable-stiffness robots. We have carried out underwater force measurements and flow visualisation for pitching foils and static foils subject to disturbances with the goal of demonstrating the ability to use the foil as a propulsor and as a control surface. 

Abstract: The use of fixed and rear flaps (boat-tail) on heavy road vehicles is an established method to suppress vortex shedding and improve aerodynamic performance. In nature, jellyfish use moving and flexible flaps to swim with unmatched efficiency. Is it possible to use nature-inspired moving flaps to further improve the aerodynamic performance of bluff bodies? To answer this question, we have numerically studied the flapping dynamics of two rear pitching flaps in the presence of a laminar bluff body wake for the first time. The study has uncovered the fundamental physical mechanisms which influence the aerodynamic performance of bluff bodies and the strategies capable of producing net-energy-savings.

Abstract: Forced motion of membranes in a fluid offers various fascinating problems. A scarcely documented forcing consists in placing a submerged membrane in a water wave field, thus imposing the excitation frequency along the whole membrane. This type of interaction has first been studied for potential applications as a wave barrier. It also has been shown that a submerged membrane attached to the seafloor could be an efficient and robust wave energy harvester. So far, the interaction between waves and a submerged membrane has been studied mostly analytically and numerically, with a strong focus on applications, while only few experimental works have been performed to characterize it. In this study, we adress the problem of the interaction between water waves and a submerged membrane, by means of physical experiments, with a view to highlighting and quantifying the various physical phenomena that contribute to the interaction. A thin flexible membrane, clumped at one end, is placed horizontally in a wave field. The simultaneous measurement of the waves (using full reconstructions of the wave field based on top view visualizations of the experimental flume) and of the deformation of the elastic plate (using side view video recording) provides many key parameters for an in-depth understanding of the interaction, such as the energy reflected and transmitted by the structure, or the membrane deflection modes. It is found that the membrane reflects little energy but tends to withdraw energy from the waves, for wavelengths close to the membrane length. The mechanism for wave energy attenuation is further investigated using Particle Image Velocimetry. Observations suggest that a significant part of membrane momentum is transferred to the fluid, in the form of a local horizontal stream near the free edge of the membrane, instead of radiated waves.

Abstract: Many tasks in fluid mechanics, such as design optimization and control, are challenging because fluids are nonlinear and exhibit a large range of scales in both space and time. This range of scales necessitates exceedingly high-dimensional measurements and computational discretization to resolve all relevant features, resulting in vast data sets and time-intensive computations. Indeed, fluid dynamics is one of the original big data fields, and many high-performance computing architectures, experimental measurement techniques, and advanced data processing and visualization algorithms were driven by decades of research in fluid mechanics. Machine learning constitutes a growing set of powerful techniques to extract patterns and build models from this data, complementing the existing theoretical, numerical, and experimental efforts in fluid mechanics. In this talk, we will explore current goals and opportunities for machine learning in fluid mechanics, and we will highlight a number of recent technical advances. Because fluid dynamics is central to transportation, health, and defense systems, we will emphasize the importance of machine learning solutions that are interpretable, explainable, generalizable, and that respect known physics. 

Biography: Dr. Steven L. Brunton is a Professor of Mechanical Engineering at the University of Washington.  He is also Adjunct Professor of Applied Mathematics and Computer science, and a Data Science Fellow at the eScience Institute.  Steve received the B.S. in mathematics from Caltech in 2006 and the Ph.D. in mechanical and aerospace engineering from Princeton in 2012.  His research combines machine learning with dynamical systems to model and control systems in fluid dynamics, biolocomotion, optics, energy systems, and manufacturing.  He received the Army and Air Force Young Investigator Program (YIP) awards and the Presidential Early Career Award for Scientists and Engineers (PECASE). Steve is also passionate about teaching math to engineers as co-author of three textbooks and through his popular YouTube channel, under the moniker “eigensteve”. 

Abstract: This expository talk presents several neural network architectures that have been proposed for learning input-output maps that are governed by PDEs, including DeepONet, Fourier Neural Operator, and PCA-Net. We will define each network from a mathematical perspective, paying particular attention to the richness of the output space definition in each formulation. We will then present qualitative results that illustrate differences in the network performance across a range of test problems drawn from fluid and solid mechanics. Finally we will present results from a careful numerical study of the cost-vs-accuracy trade-off for the networks. 

Biography: Elizabeth Qian holds a joint appointment at Georgia Tech as Assistant Professor in the Schools of Aerospace Engineering and Computational Science and Engineering. Her interdisciplinary research develops new computational methods to enable engineering design and decision-making for complex systems. Her specialties are in developing efficient surrogate models through model reduction and scientific machine learning, and in developing multifidelity approaches to accelerate expensive computations in uncertainty quantification, optimization, and control. Elizabeth previously held a postdoctoral appointment as von Karman Instructor at Caltech in the Department of Computing + Mathematical Sciences. She earned her PhD, SM, and SB degrees from the MIT Department of Aeronautics & Astronautics. Highlights of her awards and honors include a Caltech-wide teaching award from the undergraduate student body, the 2020 SIAM Student Paper Prize, the Fannie and John Hertz Foundation Fellowship, and the NSF Graduate Research Fellowship. She is also an alumna of the U.S. Fulbright student program.

Abstract: For many animals, sensing the flow of water or air disturbance is vital to their survival: it helps them locate food, mates, and prey and to escape predators. Across species, many flow sensors take the form of long, flexible cantilevers. These cantilevers exhibit sustained oscillations when interacting with fluid flow. Although much is known about how flexible cantilevers oscillate under the influence of vortex shedding, little is known about the mechanisms governing these structures' oscillatory response without vortex shedding. Our present work employs a high-fidelity numerical solver to examine the dynamics of a long, flexible cylindrical cantilever at Reynolds numbers below the critical Reynolds number of vortex shedding, i.e., Re<45. Of particular interest is investigating the underlying mechanism of the cantilever's sustained oscillations in this Re regime. Results show that frequency lock-in/synchronization is the mechanism that facilitates the transmission of disturbances to the wake and leads to the cantilever's sustained oscillations. We show that the motion of the cantilever during synchronization results in periodic vortex shedding in water for Re>20. For the cantilever in the air, wavy patterns in the shear layer dominate the wake region during the vibrations, indicating that parallel shear layers synchronize with the cantilever's motion. The tip motion trajectory of the cantilever resembles a figure-eight shape in water and an oval shape in the air. The findings of this study lead to a view in which synchronization is regarded as an intrinsic characteristic of fluid-structure systems and suggest possible directions for designing artificial flow sensors.

Abstract: Hydrodynamic pressure is a physical quantity that is utilized by fish and many other aquatic animals to generate thrust and sense the surrounding environment. To advance our understanding of how fish react to unsteady flows, it is necessary to intercept the pressure signals sensed by their lateral line system. In this study, we propose a new, non-invasive technique for reconstructing the instantaneous pressure field around an undulating body from particle image velocimetry (PIV) data. The proposed method utilizes a physics-informed neural network (PINN) to predict an optimized solution for the velocity and pressure fields that simultaneously satisfies the governing equations (i.e., the Navier Stokes equations) and the constraints put forth by the measurements. The method was validated using a direct numerical simulation of a self-propelled fish. The results demonstrate that when compared to the Queen 2.0 algorithm by Dabiri et al. (2014), the PINN is less sensitive to the spatio-temporal resolution of the velocity field measurements and provides a more accurate pressure reconstruction, particularly on the surface of the body. The improved accuracy can be attributed to three main advantages of the PINN method: 1) the kinematics of the undulating body can be directly incorporated into the pressure reconstruction process, 2) the pressure at the fluid-body interface can be directly computed without the need for extrapolation, and 3) the user has the flexibility to only enforce boundary conditions that are physically relevant. These results demonstrate that PINNs are a useful data assimilation tool for studying biolocomotion or fluid-structure interaction.

Abstract: With a growing interest in low Reynolds numbers compressible flows, compressibility effects on the secondary instabilities developing on the circular cylinder periodic wake are investigated. The unsteady and time-averaged two-dimensional flows are characterized for Reynolds numbers from Re=200 to 350 and Mach numbers up to Ma=0.5, revealing different flow structures and characteristic length scales that correlate to the instability wavelengths. The two-dimensional time-periodic solution is used as base state for a global stability analysis performed by means of Floquet theory coupled with a time-stepping finite difference approach of the non-linear operator, identifying Mode A and Mode B secondary instabilities, which are responsible for the three-dimensionalisation of the two-dimensional periodic wake. A stabilizing or a destabilizing effect of compressibility is observed on Mode A, depending on the Reynolds number and the mode wavelength, while Mode B is found to be stabilized by the increase of Mach number.

Abstract: Fish modulates the caudal fin flexibility and curvature during cruising by activating the muscles connected to fin-rays. For this purpose fish uses the bilaminar structure of the fin-ray consisting of two bony layers (hemitrich) and one soft layer (intraray). However, the effect of fish muscle activation on propulsion performance is not fully known. Therefore, a two-dimensional realistic bilaminar sunfish caudal fin-ray model is developed to study the effect of fin-ray muscle control on propulsion performance. Specifically, first the material properties of the fin-ray are inversely optimized to match the experimental data. Then a parametric study is developed by considering a sinusoidal muscle activation function for the ray and changing the muscle activation magnitude and phase. The ray is implicitly coupled with a flow field to study the fluid structure interaction. The result shows that the muscle activation cannot both increase thrust force and propulsion efficiency. It also shows that the activation phase is a critical factor affecting propulsion performance in which the maximum propulsion efficiency occurs when the activation phase is 270°, accompanied by a weak leading edge vortex. It shows that increasing the thrust is accompanied by an increase in trailing edge vortex strength and trailing edge-leading edge amplitude ratio. The wake behavior of the flow field is also studied thoroughly. A new definition for the angle of attack is introduced based on the relative velocity of the center of mass of the ray and the chord line orientation. It is showed that the lower angle of attack leads to a more slicing-like motion, resulting in a weaker leading edge vortex and finally a higher the propulsion efficiency.

Abstract: Data-driven models that respect physical laws are robust to noise, require few training samples, and are highly generalisable. Although the dynamic mode decomposition (DMD) is a principal tool of data-driven fluid dynamics, it is rare for learned DMD models to obey physical laws such as symmetries, invariances, causalities, spatial locality and conservation laws. Thus, we present physics-informed dynamic mode decomposition (piDMD), a suite of tools that incorporate physical structures into linear system identification. Specifically, we develop efficient and accurate algorithms that produce DMD models that obey the matrix analogues of user-specified physical constraints. Through a range of examples from fluid dynamics, we demonstrate the improved diagnostic, predictive and interpretative abilities of piDMD. We consider examples from stability analysis, data-driven resolvent analysis, reduced-order modelling, control, and the low-data and high-noise regimes.

Biography: Peter earned an MMath from the University of Oxford and a PhD in Applied Mathematics at the University of Cambridge. He spent one year at Imperial College London as an EPSRC Doctoral Prize Fellow before moving to MIT as an Instructor in Applied Mathematics. His PhD thesis won the “Best Thesis Award” from the UK Fluids Network and was published in Springer Nature’s outstanding theses series.

Abstract: The ability of membrane wings to adapt their shape passively in unsteady flow conditions enables several aerodynamic advantages over rigid wings. In pursuit of a theoretical model for evaluating these benefits, a theoretical framework is developed to predict the two-dimensional membrane wing response to unsteady flow conditions in an inviscid flow. Our model assumes linear deformations of an extensible membrane under constant tension, which are coupled aeroelastically to external aerodynamic loads using the unsteady thin airfoil theory. The structural and aerodynamic membrane responses are investigated for harmonic heave oscillations, an instantaneous change in angle of attack, sinusoidal transverse gusts, and a sharp-edged gust. The unsteady lift responses for these scenarios produce aeroelastic extensions to the classic Theodorsen, Wagner, Sears, and Kussner functions, respectively, for a membrane airfoil. These extensions incorporate for the first time membrane fluid-structure interaction into the expressions for the unsteady lift response of a flexible airfoil. In this talk, we will explore how the membrane elasticity affects the unsteady lift and how these affects could be exploited for gust mitigation strategies in future applications.

Abstract: This work proposes a novel approach to characterize unsteady, undulating propulsion components through evaluating the momentum and energy equations using detailed control volume analyses. Specifically, it is difficult in such conditions to separate thrust from drag as pressure components are inseparable. In general, the goal is to link physical processes to equation-specific control volume assessments measured within Computational Fluid Dynamics (CFD) models. Findings indicate that the energy equation uniquely highlights lift-work in the energy budget and can separate propulsive and drag forces. The effort expands from previously validated CFD studies of thrust resulting from heaving and pitching foils; results indicate that the presented approach provides a novel method to separate the axial force into thrust and drag components. Overall, the results indicate promise to isolate loss mechanisms from propulsive ones along with novel metrics of efficiencies useful to measuring self-propelling vehicles. 

Abstract: An axisymmetric body at the sufficiently high angle of attack shows separated flow with a set of vortices that remain attached to the body.  A cone is a self-similar canonical representation of such asymmetric structures with a peculiar asymmetric wake and body pressure distribution. In this talk, we will examine the contributions of different flow features in the wake of a perfect cone by relating the localized surface pressure to coherent flow structures. Direct numerical simulation is used to solve the flow over the axisymmetric cone for a wide range of angles of attack. In order to ensure higher accuracy and to achieve the required resolution near the boundaries, the immersed boundary method with pseudo-body-conformal grids is employed. We will discuss how two near-wake stable primary vortices originate from the separated shear layer, and how they induce reverse flow in the wake and initiate other strong secondary vortical structures very close to the surface. At a higher angle of attack, the primary vortices become asymmetric inducing substantial side forces. The role of major vortical structures on the force generation and pressure distribution on the cone surface is investigated to find the origin of this force asymmetry.

Abstract: A propeller design method has been developed that passively mitigates the formation of coherent tip vortex structures in the near-field of the rotor wake. Using the blade root bending moment coefficient (C_B) as a surrogate variable, gradients in circulation in the radial direction are avoided in a constrained optimization problem. A series of wind tunnel experiments are utilized to verify design prediction and diagnose topological characteristics of the propeller vortex wakes for a conventional power-optimized design and a vortex-attenuated design. Each set of blades designs were optimized for the same specified thrust coefficient, freestream velocity, and design RPM. Phase-locked stereoscopic particle image velocimetry (stereo-PIV) was utilized to visualize the shape and behavior of both sets of rotor blades in the axial configuration. It can be shown through the stereo-PIV that the resulting equally loaded blades produce significantly different wake shapes and behaviors. The wake optimized propellers can be shown to be more advantageous in developing quite blade technologies.

Abstract: Cross-flow turbines, also known as vertical-axis turbines, convert the kinetic energy in moving fluid to mechanical energy using blades that rotate about an axis perpendicular to the incoming flow. In this work, the performance and wake of a two-turbine array in a fence configuration (side-by-side) were experimentally measured. The turbines were operated under coordinated control, characterized by synchronous rotation rates with a constant phase difference. The array was measured with turbines co-rotating, counter-rotating with the blades traveling upstream at the array midline, and counter-rotating with the blades traveling downstream at the array midline. From the performance data, we found significant dependence between phase difference and the array efficiency. From the wake data we hypothesize how phase influences interactions between turbines.

Abstract: When two or more cylinder-like structures are placed in close proximity, the resulting interactions are known to alter the wake patterns significantly from those behind a single cylinder. One commonly investigated arrangement of two cylinders is the side-by-side configuration, wherein two identical cylinders are placed parallel to the incoming flow. In this work, the flow past two elliptic cylinders is studied experimentally in a water-tunnel. We consider two oval-shaped (elliptical) cylinders with large values of eccentricities (e = 0.89, 0.99), at varying gap ratios between the cylinders (0 < G < 2.0) and Reynolds numbers (200 < Re < 2000). The three well-known wake regimes (single-body wake, asymmetric wake, two parallel wakes) are quantified by dye visualization and planar particle image velocimetry (PIV). The formation length is observed to increase with decreasing gap ratio, and to decrease with increasing Reynolds number. In the case of the parallel coupled wake anti-phase regime, logarithmic spiral structures are observed in the outermost vortices. These outermost vortices are characterized by fitting logarithmic spirals to the flow visualization data, and evaluating the corresponding logarithmic spiral exponent, k. For an ellipse with e = 0.99 and G = 2.0, the value of k in the far-wake is shown to decrease with increasing Re.

Abstract: From air-sea gas exchange, oil pollution, to bioreactors, the ubiquitous fragmentation of bubbles/drops in turbulence has been modelled by relying on the classical Kolmogorov-Hinze paradigm since the 1950s. This framework hypothesizes that bubbles/drops are broken solely by eddies of the same size, even though turbulence is well known for its wide spectrum of scales. Here, by designing an experiment that can physically and cleanly disentangle eddies of various sizes, we report the experimental evidence to challenge this hypothesis and show bubbles are preferentially broken by the sub-bubble-scale eddies. Our work also highlights that fragmentation cannot be quantified solely by the stress criterion or the Weber number; The competition between different time scales is equally important. Instead of being elongated slowly and persistently by flows at their own scales, bubbles are fragmented in turbulence by small eddies via a burst of intense local deformation within a short time.

Abstract: In this talk, we numerically study the combined vortex and cavity synchronization/lock-in in a flexibly mounted hydrofoil at high Reynolds number. The coupled cavitation/fluid-structure system is solved using in-house body-fitted framework based on a variational finite element formulation. We identify a frequency lock-in phenomenon as the primary source of sustained large-amplitude vibrations of hydrofoil. We find that the unsteady lift forces lock into a sub-harmonic of the hydrofoil's natural frequency. During the cavity collapse and shedding, we observe a periodic generation of spanwise clockwise vorticity, leading to unsteady lift generation. We determine the origin of this flow unsteadiness near the trailing edge of the hydrofoil via the interplay between the growing cavity and adverse pressure gradient. In the frequency lock-in regime, large coherent cavitating structures are seen over the hydrofoil suction surface accompanied by a cavity growth-detachment-collapse cycle. For the post-lock-in regime, cavity shedding is primarily limited to the cavity trailing end and the attached cavity is observed to undergo high frequency spatially localized oscillations.

Abstract: This talk will report on the results obtained from applying resolvent analysis on two-dimensional incompressible laminar and turbulent flows through square and rectangular ducts. We identify the dominant linear energy-amplification mechanisms present in such flows, and in particular study the effects of both aspect ratio and the secondary mean flows that are present in turbulent cases (Prandtl's secondary flows of the second kind). We find that the presence of such secondary flows can either amplify or suppress amplification in different regimes, and we identify cases for which small amplitude secondary flows (approximately 1% of the streamwise velocity) can lead to entirely distinct energy amplification mechanisms, with resolvent gains that are approximately twice as large.

Abstract: In this talk, time-resolved particle image velocimetry (PIV) of the flow over stalled symmetric and cambered aerofoils at a chord-based Reynolds number Rec=71,000 will be used to elucidate low-order dynamics and determine noise characteristics. At this Reynolds number, the flow exhibits a combination of intense turbulent fluctuations and strong coherent 2D leading-edge vortex roll-up. The ability to capture the stall state from just 3 pressure probes using linear stochastic estimation (LSE) is explored. Further, the universality of the low-order dynamics between the symmetric and stalled aerofoils will be discussed, with important consequences for real-time sensing of aerofoil stall for any foil geometry. Finally, the LSE framework is demonstrated for determining the instantaneous flow structures correlated with noise generation.

Abstract: The high Reynolds number turbulent separated flow over a Gaussian speed-bump has presented turbulence modelling challenges for predicting flow separation, reattachment, turbulent transition, and relaminarization. This has motivated the computational fluid dynamics community to accelerate progress in this area. Unfortunately, the lack of time-resolved (TR) experimental data on the Bump has limited progress on understanding the link between the unsteady dynamics and energy transfer mechanisms that lead to flow separation, which would provide substantial insight for the advancement of turbulence models. The above challenges motivate the present work, which attempts to provide TR estimates of the velocity field from under-sampled particle image velocimetry (PIV) Bump data. We propose flow estimation techniques that utilize surface-mounted pressure sensors and recurrent neural network architectures to predict transient dynamics that are inherently missing from the under-sampled PIV time-series. Results after up-sampling the 15 Hz original PIV time-series to 3000 Hz reveal the complex unsteady dynamics characterizing the Bump flow. We observe a very low-frequency breathing mode that is likely associated with the contraction and expansion of the separation bubble, coupled with unsteadiness at higher frequencies linked with the flapping of the shear layer.

Abstract: The laminar-turbulent transition often involves the evolution of Tollmien-Schlichting (T-S) waves of small amplitude. Real surfaces are not perfectly smooth and even small imperfections can affect this process. Here, we investigate the two-dimensional scenario of T-S waves interacting with rectangular bumps using a compressible DNS code. The bumps used had sharp corners which represents the most critical bump shape. Effects of Mach and frequency were analyzed for bump heights varying from 5% to 40% of the displacement thickness. The effect of the bump varied linearly with bump height only up to 10%. In the linear regime, the effect amounts to less than a 5% increase in wave amplitude. It was found that the waves required a substantial extent of the plate to return to the smooth surface growth rate. The bump effect strongly increased with height, but, somewhat unexpectedly, the region extension was almost independent on the bump height. For heights below 10% the effect of the bump was almost independent on frequency, but for heights above it, the bump effect increased with frequency. Up to M = 0.6, compressibility effects were small. At M = 0.9 these effects increased sharply because the downstream regions affected by the bump increased substantially. Since the mesh required to simulate such small elements is demanding, an approximate method to include the bump effect in a coarser mesh was developed. This reduced the computational costs by an order of magnitude. It constitutes an improvement on previous methods that use a Taylor expansion to determine a boundary condition at the wall that causes the same effect of a surface irregularity. The improvement involves corrections based on the Stokes’ layer and on a second order approximation.

Abstract: The collective movement of fish in schools is driven by behavioral imperatives such as safety from predators, improved foraging and increased hydrodynamic efficiency. While the movement of each fish is powered by propulsive forces generated by its fin(s), patterns that emerge in fish schools are modulated not just by behavior but also by the complex flow fields encountered by fish swimming within a school. Hydrodynamics also plays a key role in enabling a fish to sense the position/velocity of neighbors through its lateral line system, and to control its own velocity and heading. A model of collective swimming that allows the investigation of behavior, as well as hydrodynamic sensing and energetics, could answer questions related to the behavioral ecology of fish as well as to inform the design of bioinspired swimming systems. In the current study, we present a new dynamical model of collective swimming of fish that has three key features:(a) the model is based on Newtonian dynamics; (b) the model includes vortex wakes as well as the interaction of fish with these vortex wakes; (c) the model is parameterized with data from direct numerical simulations (DNS) of swimming fish. The complex collective swimming patterns that emerge from this model are analyzed and recapitulated against observations. The model is also used to examine the effect of factors such as turbidity and swimming speed on school topology and stability.

Abstract: In this presentation, I will talk about closed-loop flow control studies applied to flows with transition. High-fidelity numerical simulators are leveraged for testing techniques applied to trailing edge noise attenuation and transition delay. The extremum seeking control technique is applied for online minimization of acoustic noise produced in airfoil flows at two differents Reynolds setups, where distinct noise generation mechanisms take place. We also present the application of deep learning for control design tested with the Kuramoto-Sivashinsky equation, which is able to model convective instabilities.

Abstract: Offshore wind turbines and wind farms are increasingly gaining popularity. When installed in deep waters, bottom mounted turbines are no longer possible and must be installed on platforms, whose design is inspired from naval architecture. For this reason, the aerodynamic analysis for Floating Offshore Wind Turbines (FOWTs) has to take into account factors such as platform motion, which triggers vortex instabilities, modifying the wake structure, influencing the flow reaching downstream wind turbines. We investigate this effect here by the aid of numerical simulation. The fluid flow is modeled by Large-Eddy Simulations (LES) and the turbine is modeled by Actuator Lines. The platform motion is imposed, at a given amplitude and frequency. We find that the frequency of the imposed motion strongly affects the structure of the wake. For example, when excited with half of the rotation frequency, those movements favour the coalescence of six tip vortices, generating large structures, possibly increasing alternate loads on downstream turbines. For larger excitation frequencies, for example 1.5 of the rotation frequency, the more classical out-of-phase vortex pairing mechanism comes into play. Those findings are compared with results from stability analyses.

Abstract: This work constructs three open-loop pitch maneuvers with the objective of mitigating the lift transient on a wing during a large-amplitude, transverse gust encounter. The open-loop maneuvers were experimentally implemented for different gust ratios, and force, pitch moment, and flowfield measurements were collected. The lift histories of all the cases are decomposed and compared to identify the flow mechanisms responsible for lift attenuation. The pitch maneuvers were found to mitigate a significant portion of the circulatory contribution to the lift force. Accurate modeling of added-mass was found to be critical for lift mitigation maneuver design. Pitch control input was found to double the range of the pitching moment coefficients during the gust encounters relative to the uncontrolled case. The source of the pitching moment fluctuations between the pitching and non-pitching cases differed. Quantification of the total momentum variation in the measured flowfield showed that pitch maneuvers reduced the disturbance to the gust's flowfield, and thereby reduced the momentum transfer between the gust and the passing wing. Leading-edge vortices shed from the wing during the encounters were found to have an initial linear growth region, followed by a plateau corresponding to vortex detachment. Comparison of vortex circulation strengths between experiments and unsteady discrete vortex model (DVM) simulations demonstrated the viability of using DVMs for maneuver design.

Abstract: Understanding the hydrodynamics of fish swimming is crucial to the study of interactions of fish with their surroundings. Hydrodynamic models based on potential flow theory, in particular the vortex dipole-based models, constitute a viable approach to analytically study fish swimming, offering a mathematically tractable representation of the flow physics. Despite their promise, the accuracy of the dipole-based modeling paradigms has never been validated. Here, we bridge this gap through a computational fluid dynamics campaign informed by experimental data. We demonstrate that dipole-based hydrodynamic models can capture the major characteristics of the flow around a swimming fish, while they cannot predict the geometric effect of the elongated fish body. To address this limitation, we propose an alternative model that assimilates the fish body by a pair of vortex sheets, which demonstrates an improved accuracy with a marginal increase in the computational cost.

Abstract: An overview of the suite of methodologies collectively known as global linear (modal and non-modal) flow instability will be presented and the relation of global linear theory to resolvent analysis and theoretical flow control will be highlighted. Aspects of the theoretical foundation and numerical implementation will be briefly discussed. Representative solutions of two-dimensional (BiGlobal) and three-dimensional (TriGlobal) eigenvalue and initial value problems will be presented at incompressible, supersonic and hypersonic flow conditions.

Biography: Vassilis Theofilis obtained his MSc in Applied Mathematics and Fluid Mechanics and PhD in Aerospace Engineering at the University of Manchester, UK. After a post-doc at the Department of Applied Mathematics of University of Twente, the Netherlands, he has been Alexander von Humboldt research fellow at DLR Goettingen, Germany and Ramon y Cajal Research Professor at the School of Aeronautics, Technical University of Madrid, Spain. He held visiting appointments at Caltech, Arizona, Maryland, USA and the Universidad Federal Fluminense (Rio de Janeiro), Brazil. In 2016 he was appointed at the Chair of Aerospace Engineering at the University of Liverpool, UK and since 2019 he is also Full Professor at the Escola Politecnica of Universidade São Paulo, Brazil. His research interests lie in fluid flow instability from the incompressible to the hypersonic regime, including development and application of accurate numerical methods for the solution of the pertinent large-scale eigenvalue and singular value problems.

Abstract: An overview of the suite of methodologies collectively known as global linear (modal and non-modal) flow instability will be presented and the relation of global linear theory to resolvent analysis and theoretical flow control will be highlighted. Aspects of the theoretical foundation and numerical implementation will be briefly discussed. Representative solutions of two-dimensional (BiGlobal) and three-dimensional (TriGlobal) eigenvalue and initial value problems will be presented at incompressible, supersonic and hypersonic flow conditions.

Biography: Vassilis Theofilis obtained his MSc in Applied Mathematics and Fluid Mechanics and PhD in Aerospace Engineering at the University of Manchester, UK. After a post-doc at the Department of Applied Mathematics of University of Twente, the Netherlands, he has been Alexander von Humboldt research fellow at DLR Goettingen, Germany and Ramon y Cajal Research Professor at the School of Aeronautics, Technical University of Madrid, Spain. He held visiting appointments at Caltech, Arizona, Maryland, USA and the Universidad Federal Fluminense (Rio de Janeiro), Brazil. In 2016 he was appointed at the Chair of Aerospace Engineering at the University of Liverpool, UK and since 2019 he is also Full Professor at the Escola Politecnica of Universidade São Paulo, Brazil. His research interests lie in fluid flow instability from the incompressible to the hypersonic regime, including development and application of accurate numerical methods for the solution of the pertinent large-scale eigenvalue and singular value problems.

Abstract: We study the stability and non-linear behavior of a compressible flow in an open cavity. First, global stability analysis was performed in a region of the parameter space where both Rossiter and centrifugal modes are unstable, with the Rossiter modes being substantially more unstable. Next, DNS was carried out investigating the interaction of these modes. Without the centrifugal modes, the non-linear regime of the Rossiter modes has sharp spectral peaks that diverge substantially from Linear Stability Theory (LST) predictions and approach empirical predictions of Rossiter modes. In the presence of centrifugal modes, the Rossiter modes are more broad band in spectra, but frequency and other aspects are substantially closer to LST predictions.

Abstract: We used the Force Partitioning Method (FPM, Menon and Mittal, JFM 907, A37, 2021) to decompose the hydrodynamic drag in a rough wall turbulent channel flow. The contributions of vortex and strain dominated regions on the pressure drag are quantified using an auxiliary surface-dependent potential field \phi. We have identified different sources of drag and quantified their relative importance. These sources are: Q-induced force (where Q is the second invariant of the velocity gradient tensor), viscous momentum diffusion induced pressure force, and viscous shear forces on the solid walls. Results have shown that the Q-induced force is responsible for about 50% of the pressure drag and is mainly generated by the strain-dominated (Q <0) regions before each roughness element. We also explored characterizing the equivalent sandgrain height k_s for 21 fully rough channel flows (DNS data provided by Aghaei-Jouybari et al., JFM, 912, A8, 2021) using \phi-dependent norms, and found an empirical correlation that can predict k_s with rms and maximum errors of 12 and 30 percent, respectively, satisfying the expected universality and accuracy of such predictions.

Abstract: Neural networks can be used to gain global situational awareness from local sensor measurements in fluid mechanics.  Although numerous models have been proposed to date for such application, the reconstruction is still challenging when sensors are sparsely populated.  Moreover, when sensors become on-/offline in time while moving their locations in practical applications, developing a robust technique is a major challenge.  In response, we propose a neural network-based flow field reconstruction technique from arbitrary number of sensors in motion.  We use Voronoi tessellation to obtain a structured-grid representation from low-dimensional sensor information.  This geometric data projection enables us to use a simple convolutional neural network for global field reconstruction without machine learning customizations.  We demonstrate the use of the present model for vortical wake flows, sea surface temperature, and wall-bounded turbulence.  The present technique achieves accurate flow field reconstruction from an arbitrary number of moving sensors while having robustness against noisy sensor inputs.  The flexibility of the present framework enables us to expect a broad range of extensions in both numerical and experimental studies.

Abstract: Fluid damping plays an important role in shaping damped oscillations of aeroelastic systems. In this study, we experimentally characterize the nonlinear fluid damping associated with vortices shed from the rounded leading edge and the sharp trailing edge of a rigid but elastically mounted pitching wing in the absence of a free-stream flow. We simulate the dynamics of the elastic mount using a cyber-physical system. We perturb the wing and measure the fluid damping coefficient from damped oscillations over a large range of pitching frequencies, pitching amplitudes, pivot locations and leading-/trailing-edge sweep angles. A universal fluid damping scaling based on the Morison equation is proposed and validated. Within the small-amplitude limit, the scaled non-dimensional fluid damping is found to increase linearly with the pitching amplitude, with a constant slope corresponding to the unsteady drag coefficient. This slope decreases as the pitching amplitude increases, presumably because the shed vortices no longer follow the rotating wing. Flow fields obtained using particle image velocimetry (PIV) are used to explain the nonlinear behavior of the fluid damping.

In this two-part tutorial, we will introduce machine learning techniques to analyze and model fluid flows.  In recent years, machine learning has attracted attention with an explosion of ideas and applications.  We should however note that there are significant overlaps between machine learning methods and traditional fluid flow analysis techniques.  In the present tutorial sessions, we build upon the fluid-based techniques and take a journey toIn this two-part tutorial, we will introduce machine learning techniques to analyze and model fluid flows.  In recent years, machine learning has attracted attention with an explosion of ideas and applications.  We should however note that there are significant overlaps between machine learning methods and traditional fluid flow analysis techniques.  In the present tutorial sessions, we build upon the fluid-based techniques and take a journey to sample some of the latest developments in machine learning.  In the first tutorial, we will cover unsupervised machine learning, focusing on clustering and modal analysis.  The second tutorial will go over supervised machine learning, including regression and classification using neural networks.  We will consider examples from canonical fluid flows and discuss some of the open research questions. 

Biography:

Kunihiko Taira is a professor of mechanical and aerospace engineering at UCLA focusing on unsteady aerodynamics and flow control through computational and data-driven fluid dynamics.  

In this two-part tutorial, we will introduce machine learning techniques to analyze and model fluid flows.  In recent years, machine learning has attracted attention with an explosion of ideas and applications.  We should however note that there are significant overlaps between machine learning methods and traditional fluid flow analysis techniques.  In the present tutorial sessions, we build upon the fluid-based techniques and take a journey to sample some of the latest developments in machine learning.  In the first tutorial, we will cover unsupervised machine learning, focusing on clustering and modal analysis.  The second tutorial will go over supervised machine learning, including regression and classification using neural networks.  We will consider examples from canonical fluid flows and discuss some of the open research questions. 

Biography:

Kunihiko Taira is a professor of mechanical and aerospace engineering at UCLA focusing on unsteady aerodynamics and flow control through computational and data-driven fluid dynamics. 

Abstract: Linear stability analysis has been used for decades to study fluid flows. However, its use in fluid-structure interaction (FSI) problems has only been introduced in the last few years. The flow around an elastically-mounted bluff body has been commonly used as the model problem, and, in general, the investigations focus on the characterization of the least stable modes close to the first instability. In this presentation, a methodology to calculate the sensitivity of the least stable modes of fluid-structure interaction systems with respect to local forces is presented. We make use of the adjoint equations of the flow-structure coupled system to calculate the gradients. The methodology was applied to two-dimensional incompressible laminar steady flows around an elastically-mounted circular cylinder, and we obtained the gradients of the real and imaginary parts of the least stable eigenvalues with respect to forces located at arbitrary points in the flow domain. Selected values of mass ratio and reduced velocity were considered in the simulations, and the results were compared to those obtained for a fixed cylinder at the same Reynolds number. The sensitivity fields of the fluid-structure interaction system can be very different from its fixed structure counterpart, and amongst the cases, with an elastic structure, the fields vary greatly according to the reduced velocity. The sensitivity results were verified against linear and nonlinear simulations of flows with small control cylinders placed at locations selected according to the sensitivity fields. The agreement between the predictions made with the sensitivity analyses and the linear and nonlinear results of the forced flows was excellent. In some cases, it was possible to completely suppress the cylinder vibration.

Abstract: The aerodynamics of inclined flat plates in close ground proximity is experimentally investigated at a chord-based Reynolds number of 50,000 for aspect ratios (AR) of 1 and 2. The minimum ground height is varied between 0.1 to 1.0 chord lengths. Lift is measured directly using a force balance for angles of attack ranging from -90 to 90 degrees. Streamwise and cross plane velocity measurements are conducted using planar and stereo particle image velocimetry, respectively. In all cases, notable changes in loading are observed when the plate is below 0.75 chord lengths from the ground. The results show that the effect of ground proximity on the lift force is dependent on the combination of AR and angle of attack, but the most significant effects are observed for angles between 20 and 40 degrees for both AR. The analysis relates the observed changes in aerodynamic loading to the changes in flow field development, and pressure reconstructions are performed at the centreline plane to elucidate the associated changes in surface stresses.

Abstract: Oscillating foils have shown to be an effective way to extract hydrokinetic energy and offer benefits of shallow water operation, scalability, and low cut-in speeds. For arrays of oscillating foils, the role of wake structure is particularly important since the coherent structures can cause constructive and/or destructive interference with downstream foils. This presentation explores a wide range of oscillating foil kinematics within energy harvesting mode, with the goal of grouping and parameterizing the wake based on the input kinematic variables. A first approach based on the primary vortex strength shed from each oscillating foil kinematics is discussed and it is found three energy harvesting modes with respect to the foil’s relative angle of attack. To fully consider the wake between two foils, an autoencoder and clustering model is developed to capture the pertinent wake features from each kinematics and further cluster foil kinematics based on wake similarity. Four different wake patterns are obtained through this model and when compared with the primary vortex analysis approach, a connection between these wake patterns and foil relative angle of attack is determined, and hence provide insight for optimizing foil array configurations for energy harvesting.

Abstract: Pulsatile jet propulsion is a highly energy-efficient swimming mode used by various species of aquatic animals and which inspires engineers of underwater vehicles. Here, we present a bio-inspired jet propulsor that combines the flexible hull of a jellyfish with the compression motion of a scallop to create individual vortex rings for thrust generation. Similar to biological jetters, our propulsor generates a nonlinear time-varying exit velocity profile and has a finite volume capacity. The formation process of the vortices generated by this jet profile is analysed using time-resolved velocity field measurements. The transient development of the vortex properties is characterised based on the evolution of ridges in the finite-time Lyapunov exponent field and on local extrema in the pressure field, which is derived from the velocity data. Special attention is directed toward the vortex pairing observed in the trailing shear layer. During vortex pairing, the Lagrangian vortex boundaries first contract in the stream-wise direction before expanding in the normal direction to keep the non-dimensional energy at its minimum value, in agreement with the Kelvin-Benjamin variational principle. The circulation, diameter, and translation velocity of the vortex increase due to pairing. The vortex pairing takes place because the velocity of the trailing vortex is higher than the velocity of the main vortex ring prior to merging. The comparison of the temporal evolution of the Lagrangian vortex boundaries and the pressure based vortex delimiters confirms that features in the pressure field serve as accurate and robust observables for the vortex formation process.

Abstract: The pressure loads induced by vortices on immersed surfaces are central to numerous problems in fluid dynamics. While past studies in unsteady aerodynamics have focused mostly on the role of vortices in force production, we show using a force partitioning method (FPM; Menon & Mittal, J. Fluid Mech., 918, R3, 2021) that strain-dominated regions associated with vortices can in fact have a significant effect on aerodynamic loads in some situations. FPM allows us to quantify the loads induced on immersed surfaces by individual vortices as well as their associated regions of strain. By analyzing the forces on a pitching airfoil undergoing dynamic stall, we show that our current understanding of vortex-dominated phenomena could be incomplete without considering the substantial, and sometimes dominant, effect of strain-dominated regions that are associated with vortices.

Abstract: The onset and evolution of the dynamic stall vortex (DSV) are analyzed by means of large eddy simulations of an SD7003 airfoil undergoing periodic plunging motion in a transitional Reynolds number flow (Re = 6 × 10^4). Interactions between upstream propagating Kelvin-Helmholtz instabilities and a shear layer formed at the leading edge trigger flow separation. The former appear to be related to acoustic waves scattered at the trailing edge due to initial vortex shedding. Two freestream Mach numbers (M = 0.1 and 0.4) are employed to examine the flow differences due to compressibility variations. The existence of a common timing for the acoustic perturbations in both flows suggests a possible Mach number invariance for the birth of the Kelvin-Helmholtz instability. Increasing compressiblity, however, induces earlier spanwise fluctuations, higher flow three-dimensionality and a weaker and more diffuse DSV, which is formed further downstream of the leading edge and has lower residency time. Modal decomposition, performed with both the classical dynamic mode decomposition (DMD) and its multi-resolution variant (mrDMD), highlights key trends and demonstrates the capacity of the mrDMD to extract physically meaningful flow structures related to the stall onset. Such detailed characterization of the shear layer can be used for a systematic exploration of flow control strategies for unsteady airfoils.

Abstract: Fish schools are examples of active systems whose collective dynamics emerge from individual-level interactions. These systems are often modeled with self-propelled particles in unbounded domains subject to behavioral rules based on visual feedback that usually neglect hydrodynamic interactions. Little is known about how geometric confinement together with flow-mediated interactions affect the collective behavior of fish. Here, we combine vision-based rules with hydrodynamic interactions in a circular domain, and we map out the different collective phases that the group of fish can achieve. We show that new collective phases emerge where (1) the group follows the tank wall; (2) the group splits into two groups milling in opposite directions, in a double milling phase; (3) a new bistable regime emerges in which the school intermittently switches from schooling to milling and vice-versa. We analyze the bistable regime by constructing effective potentials on the coarse-grained translational and rotational order parameters. We find that the bistable regime is sensitive to the school size and the geometric confinement.

Abstract: Vertical-axis wind turbines are great candidates for wind energy diversification and could contribute to reaching a near-zero carbon emission electrical grid. The complex aerodynamics of vertical-axis wind turbines have challenged their development and integration into urban infrastructure. The blades of these turbine undergo periodic oscillations in inflow conditions, which continuously change the blade’s effective angle of attack and flow velocity. These oscillations often lead to the formation of large-scale vortices and the occurrence of dynamic stall. These flow structures allow an increase in torque production, but also cause heavy load transients jeopardising the turbine’s structural integrity and potentially leading to premature failure. This talk aims at presenting the dilemma vertical-axis wind turbines face over a wide operation envelope: generating torque while ensuring structural reliability.  Time-resolved particle image velocimetry and load measurements were performed to characterise the occurrence of dynamic stall on a scaled-down vertical-axis wind turbine for a wide range of operating condition.The timescales of vortex formation and the corresponding impact on the unsteady load response are analysed. Close attention is brought to the asymmetry in flow topology and load response found between the first and second half of the blade’s revolution. Future work on using these metrics for active flow control using blade pitching is also introduced.

Abstract: This talk presents new methods for modeling and simulation of the aortic and mitral heart valves and use of these methods to study congenital heart disease. To construct model heart valves, we specify that the heart valve supports a pressure and derive an associated system of partial differential equations for its loaded state. Using the solution to this system, we then derive reference geometry and material properties. By tuning the parameters in this process, we design the model valves. This process produces material properties that are consistent with known values, yet also includes material heterogeneity. When used in fluid-structure interaction simulations, these models are highly effective, producing realistic flow rates and robust closure under physiological driving pressures. Using these models, we study flows through the bicuspid aortic valve. Simulations show that a bicuspid valve, without alterations to the aorta anatomy, alters blood flow patterns dramatically. These flows suggest that hemodynamics may play a strong role in aortic dilation and aneurysm formation.

Abstract: We present spectral analysis modal methods (SAMMs) to perform POD in the frequency domain using non-time-resolved Particle Image Velocity (PIV) data combined with unsteady surface pressure measurements. In particular, time-resolved unsteady surface pressure measurements are synchronized with non-time-resolved planar PIV measurements acquired at 15 Hz in a Mach 0.6 cavity flow. Leveraging the spectral linear stochastic estimation (LSE) method of Tinney et al. (2006), we first estimate the cross correlations between the velocity field and the unsteady pressure sensors via sequential time shifts, followed by a Fast Fourier transform to obtain the pressure-velocity cross spectral density matrix. This leads to a linear multiple-input / multiple-output (MIMO) model that determines the optimal transfer functions between the input cavity wall pressure and the output velocity field. Two variants of SAMMs are developed and applied. The first, termed “SAMM-SPOD”, combines the MIMO model with the SPOD algorithm of Towne et al. (2018). The second, called “SAMM-RR”, adds independent sources and uses a sorted eigendecomposition of the input pressure cross-spectral matrix to enable an efficient reduced-rank eigendecomposition of the velocity cross-spectral matrix. In both cases, the resulting rank-1 POD eigenvalues associated with the Rossiter frequencies exhibit very good agreement with those obtained using independent time-resolved PIV measurements. The results demonstrate that SAMMs provide a methodology to perform space-time POD without requiring a high-speed PIV system, while avoiding potential pitfalls associated with traditional time-domain LSE.

Abstract: In this work, we present an approach to obtain a desired leading-edge vortex (LEV) shedding pattern from an unsteady airfoil through the execution of a suitable motion kinematics. Previous research revealed that LEV shedding is associated with the  leading-edge suction parameter (LESP) exceeding a maximum threshold. A low-order method called LESP-modulated discrete vortex method (LDVM) was also developed to predict the onset and termination of LEV shedding from an airfoil undergoing a prescribed motion kinematics. In the current work we present an inverse-aerodynamic formulation based on the LDVM to generate an appropriate motion kinematics to achieve a prescribed LESP variation and thus the desired LEV shedding characteristics from the airfoil. The algorithm identifies the kinematic state of the airfoil required to attain the target LESP value through an iterative procedure performed inside the LDVM simulation at each time step. Several case studies are presented to demonstrate events such as initiation and termination of LEV shedding at prescribed instants of time, inducing LEV shedding from the chosen surface of the airfoil, inducing or deterring LEV shedding during an unsteady motion on demand, and achieving similar LEV shedding patterns using different maneuvers. The kinematic profiles generated by the low-order formulation are also simulated using a high-fidelity unsteady RANS method. The flowfield results from CFD confirm  the occurrence of the desired  events at the prescribed  time instants.

Abstract: While hunting for prey, seals are able to use their whiskers for hydrodynamic trail following, a skill partly attributed to the unique shape of the whiskers themselves. The whisker’s undulated topography has thus been investigated for its hydrodynamic properties, demonstrating reduced drag and oscillating lift forces when compared to flow over a smooth cylinder however the exact mechanisms are not well understood. The current investigation parameterizes the seal whisker inspired undulated geometry into seven non-dimensional parameters and shows that individual modifications to the geometric parameters can strongly impact the overall flow response. In particular, simulations are performed for flow over five geometries with various undulation wavelengths (λ=1, 2, 3.4, 5, 6.9) and a comparable smooth elliptical cylinder. The effect on wake patterns and spanwise variation is investigated through comparison of Reynolds stresses and turbulent kinetic energy calculations. Changes in the flow response vary nonlinearly with respect to topography wavelength with minimal drag and oscillating lift occurring at a biologically relevant wavelength, λ=3.4. The analysis highlights the redirection of energy along the span and the resulting effect on wake structure and vorticity.

Abstract: Water channel experiments are presented for a pair of NACA 0012 pitching hydrofoils of aspect ratio 3. One foil is fixed, while the other is completely free to move in the horizontal plane. A side-by-side arrangement is found to be two-dimensionally stable to perturbations away from this equilibrium arrangement, and arises naturally from purely hydrodynamic forces. This provides the first experimental evidence of two-dimensional stability and supports the Lighthill conjecture, which is that hydrodynamics forces may be sculpting the structure of fish schools.  This minimal schooling arrangement also increases the swimming speed by 20% compared to an isolated swimmer.  These findings are supported by force measurements, trajectory measurements, and free-swimming simulations of two-dimensional pitching foils. Moreover, previously discovered one-dimensionally stable equilibria driven by wake vortex interactions are shown to be, in fact, two-dimensionally unstable, at least for an out-of-phase synchronization.  These newfound schooling performance and stability characteristics suggest that fluid-mediated equilibria may play a role in the control strategies of schooling fish and fish-inspired robots.

Abstract: Water tunnel experiments were conducted to study the interaction of gusts with airfoils as well as swept and unswept finite wings that are loaded or unloaded. A single vortex filament was generated by plunging an upstream airfoil at zero angle of attack in transient motion. The details of the interactions were documented by means of phase-locked force measurements and particle image velocimetry (PIV) measurements in multiple spanwise planes. The vortex filament, which was initially quasi-two-dimensional, deformed, diffused, and lost coherence as it interacted with the wing. The results show that when the counter-clockwise vortex passes the wing in proximity, the lift coefficient exhibits a sharp rise followed by a sharp decrease before recovering to the static value. The first peak is observed before the vortex arrives near the leading-edge, and is quasi-steady in nature. For the unloaded swept wing, the local structure of the vortex only depends on the relative location of the wing cross-section at that particular spanwise plane. For the loaded swept wing, the wing sweep has the largest effect at a post-stall angle of attack. Close interactions may produce vortex shedding parallel to the leading-edge of the swept wing, increasing the magnitude of the lift peak relative to the unswept wing.

Abstract: Multi-vessel coordination and controlled maneuvering through upstream wakes is important to a wide range of applications; from surface ships to autonomous underwater vehicles. In this work we study the predictive performance of physics-based and machine-learning (ML) models for unsteady inflow maneuvering forces using tandem flapping foils as a model system, in an effort to replicate the ability of fish to manoeuvre and obtain performance gains from schooling formations. Two physics-based approaches, one following simple quasi-steady assumptions and another using the classical Theodorsen have been modified to account for the flow unsteadiness, and are found to perform fairly well when there are only mild interactions with the upstream wake, with minimum error levels of around 6%. However, this error increases to 40% when there is strong wake interaction. Three ML models were trained and tested; a Long Short-Term Memory (LSTM) model, a Neural Ordinary Differential Equations (NODE) model, and a Sparse Identification of Nonlinear Dynamics (SINDy) approach. We find that all three models can match the low error of the physics-based for mild inflow unsteadiness and are capable of improving the predictions in the case of strong interactions, reducing the error levels below 20%. While these ML models require substantial training data and care in choosing their meta-parameters, their predictions do prove to be more reliable for a wider range of unsteadiness conditions as well as potentially still producing human-interpretable models (in the case of SINDy), making them an interesting research direction for further study.

Abstract: The stability characteristics of a trailing vortex pair interacting with its image in the ground plane is formulated analytically and studied numerically. The theoretical framework models the inviscid interaction between a counter-rotating pair of perturbed finite-core vortices near a planar surface. The stability equations are derived by matching the Biot-Savart integrals of the vortices with their temporally-varying position vectors, which are subject to constraints that represent a ground-image system. The stability problem is then cast into an optimal perturbation analysis to deduce the maximum growth rate for a prescribed disturbance as a function of time and the induced vortex trajectories.

Abstract: We numerically investigate linear modal instabilities of flow over finite-span untapered NACA 0015 wings at Reynolds number of 400 and a range of angles of attack and sweep on two wings having aspect ratios 4 and 8. Base flows were generated by direct numerical simulation, marching the unsteady incompressible three-dimensional Navier-Stokes equations to a steady state, or using selective frequency damping to obtain stationary linearly unstable flows. Global stability analysis is performed to identify the instabilities that lead to complex wake structures seen behind the wing. Unstable three-dimensional linear global modes of swept wings are identified for the first time using spectral-element time-stepping solvers. On unswept wings, stability analysis revealed that the most unstable global mode peaks in the midspan region of the wake with the peak of the mode structure moving towards the tip as sweep is increased. We show that the growth of these modes leads to the formation of vortical structures observed in the flow. In addition, data-driven modal analysis is employed to identify the most energetic structures of the nonlinear wake. On unswept wings, the dominant mode at low angles of attack is a Kelvin-Helmholtz-like instability, qualitatively analogous with global modes of infinite-span wings under same conditions. At higher angles of attack and moderate sweep angles, the dominant mode is a structure denominated the interaction mode. At high sweep angles, this mode evolves into elongated streamwise vortices on higher aspect ratio wings, while on shorter wings it is similar to a tip-vortex instability.

Abstract: An experimental investigation is conducted on a three-dimensional laminar separation bubble forming on the suction surface of a finite wing at a chord Reynolds number of 125 000. The rectangular semispan wing has a NACA 0018 airfoil section and an aspect ratio of 2.5. Measurements are performed using surface pressure taps and particle image velocimetry. Over a majority of the wingspan, the separated shear layer of the three-dimensional bubble rolls up into spanwise uniform vortices which behave similarly to the vortices in the two-dimensional separation bubble. Near the wing tip, where the spanwise pressure and convection velocity gradients are strongest, the vortices display a higher degree of three-dimensionality. The wing tip vortex inhibits boundary layer separation and transition in proximity to the wing tip, and the amplification of disturbances in the shear layer is reduced. Unlike a canonical two-dimensional separation bubble, the three-dimensional separation bubble does not form a closed region of recirculating flow. Fluid enters the recirculation region of the three-dimensional separation bubble at the wing tip, where it is drawn towards the wing root. The three-dimensional separation bubble is classified as a crossflow separation. The results show that away from the wing tip, the three-dimensional separation bubble is essentially spanwise uniform, whereas near the wing tip, three-dimensional effects substantially modify the bubble's mean structure and dynamics.

Abstract: Flow interactions are thought to be beneficial for animals flying and swimming in groups, and fish located behind but laterally displaced from upstream neighbours seem to exploit neighbour-induced vortices to save energy. However, it is not clear if fish actively seek these favorable positions or whether flow interactions create conditions in which flapping swimmers lock into stable formations that lead to energy savings. Recent experiments with pairs of self-propelled robotic flappers indicate that followers in any position relative to the neighbor ahead obtain hydrodynamic benefits if they match their tailbeat phase with the flow velocity of the leader's wake, a strategy that freely swimming fish also seem to exhibit even in the absence of visual and lateral line sensing. Meanwhile, pairs of hydrofoils, positioned in tandem with no means of adjusting their flapping motions, were shown to swim together cohesively, even when flapping at dissimilar kinematics, due to the interaction of the follower with the leader's wake. Are the hydrodynamic mechanisms at play in these distinct arrangements, in tandem hydrofoils and laterally-displaced fish, the same? To address this question, we employed a minimal vortex sheet model that captures salient features of the flow interactions among flapping swimmers, and we analyzed the free swimming of a pair of in-line and laterally-displaced flapping swimmers. We found that flow interactions stabilize formations of flapping swimmers at lateral positions that favor vortex phase matching for any flapping phase of the follower. Our results are consistent with the hypothesis that fish, although may choose to, do not need to actively match their tailbeat phase with the local vorticity to save energy and that hydrodynamics alone could create these energetically favorable formations.

Abstract: In the unsteady flow fields generated by pitching airfoils and plates, vortex wakes and coherent vortex interactions have been a constant focus because vortex behaviors such as formation, shedding, merging, and breaking, have been shown to be relevant to the performance dynamics such as lift, drag, or propulsion. In order to identify the important vortex structures and investigate their dynamics, the topological data analysis tool Persistent Homology (PH) was applied to both numerical and experimental flow fields generated by pitching flat plates. Several possible ways of using PH in this context were compared. The first comparison investigated point vortices shed by a pitching plate, as generated by the Discrete Vortex Method (DVM). The results show how the choice between the “standard” Euclidean distance and a vorticity-adapted Euclidean distance affects the resulting Vietoris–Rips (VR) complexes and their Betti numbers (number of connected components and topological holes). In both cases the PH was computed snapshot-by-snapshot. The second comparison investigated a set of phased-averaged experimental PIV vorticity fields of a pitching trapezoid with strouhal number St = 0.24, recorded on an Eulerian rectangular grid. The results show the difference between PH of the grid points treated as point vortices, and PH of the vorticity field itself computed using the cubical complex. In both cases, the changes in persistence diagrams and other representative diagnostics were correlated to the lift and drag felt by the pitching trapezoid, and interpreted in terms of the physical mechanism of vortex formation and evolution.

Abstract: When an object is accelerated in a fluid, a primary vortex is formed through the roll-up of a shear layer. This primary vortex does not grow indefinitely and will reach a limiting size and strength. Additional vorticity beyond the critical limit will end up in a trailing shear layer and accumulate into secondary vortices. The secondary vortices are typically considerably smaller than the primary vortex. In this paper, we focus on the formation, shedding, and trajectory of secondary vortices generated by a rotating rectangular plate in a quiescent fluid using time-resolved particle image velocimetry. The Reynolds number Re based on the maximum rotational velocity of the plate and the distance between the centre of rotation and the tip of the plate is varied from 840-11150. At low Re, the shear layer is a continuous uninterrupted layer of vorticity that rolls up into a single coherent primary vortex. At Re = 1955, the shear layer becomes unstable and secondary vortices emerge and subsequently move away from the tip of the plate. For Re > 4000, secondary vortices are discretely released from the plate tip and are not generated from the stretching of an unstable shear layer. First, we demonstrate that the roll-up of the shear layer, the trajectory of the primary vortex, and the path of secondary vortices can be predicted by a modified Kaden spiral for the entire Re range considered. Second, the timing of the secondary vortex shedding is analysed using the swirling strength criterion. The separation time of each secondary vortex is identified as a local maximum in the temporal evolution of the average swirling strength close to the plate tip. The time interval between the release of successive secondary vortices is not constant during the rotation but increases the more vortices have been shed. The shedding time interval also increases with decreasing Reynolds number. The increased time interval under both conditions is due to a reduced circulation feeding rate.

Abstract: Active flow control is a technology that is becoming increasingly more compelling for a multitude of aerodynamic applications. One of the implementation problems faced by the community is related to finding good locations to place the microjet actuators such that their effect is maximized with minimal input. Due to the vastness and complexity of the parameter space in the highly non-linear Navier-Stokes plant, this problem has only been explored computationally to date, with several restrictions such as the linearization of the plant. In this work, we will discuss a fully experimental implementation of the actuator placement problem in the fully non-linear context of the experiment, which provides a powerful platform for practical optimization to find effective actuator patterns in complex geometries, given a goal function of engineering relevance. The solutions encountered will then be briefly discussed for physical insight to understand the mechanisms leveraged by the optimization algorithm.

Abstract: This research adopts a bioinspired approach to improve how unmanned aerial vehicles (UAVs) and small aircraft fly by studying the aerodynamic responses of a red–tailed hawk (Buteo jamaicensis) when flying through a vertical gust. Red–tailed hawks remain stable and mitigate strong wind gusts that small aircraft struggle to fly through. However, the specific maneuvers that hawks perform to stabilize within natural or artificial wind gusts are not fully understood. To study these aerodynamic responses in a controlled environment, a flight–testing arena was developed. Four industrial fans placed perpendicular to the hawk’s flight path were used to produce an average vertical gust velocity of 6.5 – 8.5 m/s at 0.8 m above the fans, the location the hawk flies at. This uniform gust region was introduced in the flight path, encompassing the entire wingspan of the hawk. Hawks normally fly at around 10 m/s at level flight, making the gust magnitude between 65% and 85% of the hawk’s flight speed. The gust responses were recorded using calibrated multi–camera videography from two GoPro Hero 6 Black cameras at 240 frames per second. The hawk’s beak, tail, wings, and wrist were tracked in 3D to study the pitching response of the hawk when flying through the vertical gust. Tracking these specific points on the hawk will provide knowledge about how red–tailed hawks can morph their wings and tail to mitigate strong wind gusts while their body remains stable. This presentation reviews the experiment methodology and discusses some preliminary tracking data of the hawk’s responses to vertical gust events.

Abstract: Neutrally-buoyant near ground swimmers experience alterations in their added mass, quasi-steady, and wake-induced forces compared to swimming far from a ground plane.  In fact, using a simple freely-swimming pitching hydrofoil as a model near-ground swimmer it has been shown that a hydrofoil will be attracted to a stable equilibrium altitude due to competing hydrodynamic forces.  Here, a potential flow decomposition method using the unsteady Bernoulli equation is presented and applied to understand the competing forces that give rise to equilibria.  It is shown that many previous hypotheses do not hold and that equilibria are a balance between negative time-averaged quasi-steady lift and positive time-averaged wake-induced lift, while the time-averaged added mass lift is nearly zero across all ground proximities.  Results that run counter to previous hypotheses are examined in detail. 

Abstract: An aspect ratio 9.5 rectangular wing is revolved in a cylindrical domain at 45-degree angle of incidence and a Reynolds number Re = 300, based on the wing velocity two chord lengths from the axis of rotation. Four cases are considered. Case A represents the physical problem in which the approach velocity varies linearly with the distance from the axis of rotation and Coriolis and centripetal accelerations are active in the non-inertial reference frame attached to the wing. Case B implements the same reference frame attached to the wing, but without rotational accelerations. In cases C and D, the rotational accelerations are the same as A and B, respectively; however, the inflow is uniform along the span. Each case exhibits a strikingly different behavior of the leading-edge vortex (LEV), demonstrating that inflow shear is an important factor governing LEV behavior, in addition to the rotational accelerations. However, the mechanisms are active at different times in the vortex evolution. Vorticity transport analyses were conducted in chordwise planar control regions, at z/C = 2.0 (measured from the axis of rotation), revealing distinct differences in the contributions of the vorticity sources governing leading-edge vortex development for the four cases investigated.

Abstract: The Betz efficiency, since its derivation over one hundred years ago, has served as an upper bound on the power-conversion efficiency of wind-energy systems in theory and in practice. Recently, however, Dabiri (Phys. Rev. Fluids, 2020) suggested that relaxing the steady-flow assumption of the Betz derivation can lead to efficiencies that exceed the Betz limit. We thus seek to determine the effect of unsteady streamwise motion on the efficiency of wind-energy devices both analytically and experimentally. We first model the influence of periodic streamwise motion of an actuator disc using time-varying velocity potentials, which allow us to quantify the effects of various motion profiles on the theoretical efficiency. We find that certain classes of velocity-potential models and streamwise-motion waveforms yield time-averaged efficiencies above the Betz limit. These analytical results then motivate the construction of an experimental apparatus to investigate the applicability of the theory to a horizontal-axis wind turbine actuated in surge motions. The planned experiments will clarify the relationship between unsteady streamwise motion and turbine efficiency, and will inform the design and control of both existing and novel energy-harvesting systems.

Abstract: The encounter of a vortex gust with an aerodynamic body is a canonical fluid–structure interaction with implications for the prediction of transient loads on fliers and swimmers and their generation of vortex sound. In this talk, we will present an analytical model that is used to investigate the dynamically-coupled interactions of vortex gusts encountering a symmetric Joukowski airfoil on linear elastic supports. The model is solved as a potential flow problem using a time-dependent conformal mapping. The Brown and Michael framework models the unsteady shedding of vorticity from the airfoil into the wake, and the aeroelastic motion of the airfoil is analyzed using unsteady airfoil theory. The proposed model may be used to investigate the combined effects of airfoil thickness, airfoil motions and shedding of vorticity on gust-airfoil interactions problems. Special attention is paid to the effects of gust position and strength alongside structural parameters on direct vortex impingement.

Abstract: A stationary wing placed in a wake generated upstream experiences lift enhancement and stall delay. This effect is the most predominant in the post-stall conditions and when the wing is placed at an optimal offset distance from the wake centerline. The von Karman vortex street in the wake causes excitation of the separated flow in the post-stall conditions. The increase in the time-averaged lift force is associated with the flow separation, leading-edge vortex formation, and subsequent reattachment in a process similar to the dynamic stall of oscillating wings. At the optimal location, direct impingement of large vortices is avoided, and the velocity fluctuations in the wake are much smaller than those at the wake centerline. The small-amplitude excitation of the shear layer separated from the leading-edge, coupled with the appropriate wake frequency, lead to large vortices and separation bubbles, providing increased lift in the time-averaged sense. In contrast, the large-amplitude velocity fluctuations at the wake centerline lead to alternating partially attached and totally separated flow (deep stall), which is not as effective for the lift enhancement. The mechanism of delayed flow separation is similar to that of active flow control separation, with similar optimal frequencies. The degree of the lift enhancement is remarkable, given that the wake at these Reynolds numbers is turbulent, with relatively small spanwise correlation length.

Abstract: Ring vortices are efficient at transporting fluid across long distances. They are observed in nature in various ways: they propel squids, inject blood in the heart, and entertain dolphins. These vortices are generally produced by ejecting a volume of fluid through a circular orifice and have been widely studied and characterised. After four convective times, three events happen simultaneously: the vortex moves faster than the shear layer it originates from, it separates from the shear layer, and the circulation and non-dimensional energy of the vortex converge. The simultaneity of these three events obfuscates the causality between them. To analyse the temporal evolution of the vortex independently of the separation, we analyse the development of vortices generated in the wake of cones. The vortex rings that form behind the cones have a self-induced velocity that causes them to follow the cone. They continue to grow as the cone travels well beyond the limiting vortex formation times observed for vortices generated by pistons. The non-dimensional circulation, based on the vortex diameter, and the non-dimensional energy of the vortex rings converge after three convective times. This result proves that the convergence of non-dimensional quantities is not just a consequence of the separation. In addition, the evolution of the vortex is modelled with an axisymmetric discrete vortex method. The model predicts accurately the evolution of the vortex.

Abstract: This talk will report on the development of a (semi) analytic method to approximate resolvent (pseudospectral) mode shapes and their amplification levels. These modes seek to reproduce wall-normal vorticity modes of maximal amplification of the linearized Navier-Stokes equations in wall-bounded parallel shear flows. The method assumes that mode shapes can be accurately approximated by a sum of suitably-defined wavepackets. The small number of parameters prescribing the shape of these wavepackets may be found by solving a low-dimensional optimization problem, thus eliminating the need to form and decompose discretized linear operators. We demonstrate the applicability and assess the performance of this method on the simplified scalar Squire operator in laminar plane Couette and plane Poiseuille flow. In particular, we show that the method can be applied to cases where leading resolvent modes are affected by boundary conditions and/or multiple critical layers. We next introduce a modified Laplacian inner product in order to apply this method to the full Navier-Stokes system for parallel shear flows. We additionally explore the capabilities of this method to compute suboptimal mode shapes and amplification levels, and discuss prospects for applying this method to more complex systems.

Abstract: Peripheral pulmonary artery stenosis (PPAS) is the most common cardiovascular abnormality associated with Williams and Alagille syndromes, two congenital disorders for which structural cardiovascular abnormalities comprise the leading cause of morbidity and mortality. Narrowing of the central and peripheral pulmonary arteries (PAs) in PPAS results in lung perfusion disparity, right ventricular (RV) hypertension, RV hypertrophy, and ultimately RV failure. No clinical consensus exists regarding the optimal treatment strategy for Williams and Alagille patients with PPAS, as transcatheter interventions limited to the central PAs have been associated with unfavorable outcomes, and extensive surgical repair requires long operating hours and specialized expertise. In this work, we lay the foundations for a virtual treatment planning platform capable of identifying lesions most critical for normalizing pulmonary vascular resistance (PVR) and thus PA pressures. We demonstrate our ability to accurately predict post-stent PA hemodynamics by incorporating the relevant autoregulatory physiology into our CFD simulations with fluid-structure interaction. Based on controlled comparisons of PA hemodynamics following different stenting plans within the same patient cohort, we further develop preliminary clinical recommendations. We note that our methods are broadly applicable to CFD investigations of other cardiovascular interventions.

Abstract: Propulsive flapping foils are widely studied in the development of animal-like autonomous systems and other engineering applications.  In this work, we carry out extensive numerical simulations to address various aspects of flapping propulsion. First, we explore the validity and applicability of strip theory to model flapping foils. We show that there exists an intermediate range of Strouhal numbers where the strip theory can be applied and that 3D effects dominate outside of this range. Second, we examine the possibility of predicting cycle-averaged peak forces of 3D foils based on 2D simulations. We show that an aspect-ratio based-correction, analogous to Prandtl finite wing theory, enables the use of strip theory in finite flapping-foil design. Finally, we study the importance of the leading-edge sweep angle to the performance of flapping foils. We carefully control the foil parameters of tail-like and flipper-like kinematics for a range of sweep angles. We observe no significant change in mean force, power, moment and efficiency for tail-like and flipper-like motions as the sweep angle increase, leading to a conclusion that fishtails, flukes and flippers can have a large range of potential sweep angles without a negative impact on hydrodynamic performance.

Abstract: Sensor placement and feature selection are critical steps in engineering, modeling, and data science that share a common mathematical theme: the selected measurements should enable solution of an inverse problem. Most real-world systems of interest are nonlinear, yet the majority of available techniques for feature selection and sensor placement rely on assumptions of linearity or simple statistical models. We show that when these assumptions are violated, standard techniques can lead to costly over-sensing without guaranteeing that the desired information can be recovered from the measurements. In order to remedy these problems, we introduce a novel data-driven approach for sensor placement and feature selection for a general type of nonlinear inverse problem based on the information contained in secant vectors between data points. Using the secant-based approach, we develop three efficient greedy algorithms that each provide different types of robust, near-minimal reconstruction guarantees. We demonstrate them on a problem where linear techniques consistently fail: sensor placement to reconstruct a fluid flow formed by a complicated shock-mixing layer interaction.

Abstract: Investigations into the biomechanics of cervical cerclage. Premature cervical remodeling, or cervical insufficiency, is a medical condition during pregnancy in which the uterine cervix softens and dilates before full term, usually between 18 and 22 weeks gestation, such that a preterm birth occurs. It is a common cause of second trimester pregnancy loss. Part of the clinical treatment of this condition is to perform a cervical cerclage (a purse string suture to close the cervix). Studies on the efficacy of this procedure are conflicting and mostly rely on statistical investigations. The importance of biomechanical factors such as cervical length, canal geometry and tissue softness are not as well documented. The purpose of this investigation is to examine the mechanical limitations of the cerclage. Working with physicians from The George Washington University Hospital, we create generalized, synthetic models of the cervix from ultrasound images and fabricate them with silicone to mimic physiological, softening cervical tissue. Aspects of the cervical geometry (length of cervix, shape and width of dilatation) and suture material used in the cerclage are varied. The synthetic cervices are stitched by physicians according to clinical techniques. Each synthetic cervix is placed into a capsule designed to contain the material while it is being compressed. Using a custom steel insert attached to a 5kN load cell, force is applied directly to the suture and measured as a function of time until failure (designated as when the cervical tissue begins to rip). The results of this study will provide insight into the most effective clinical interventions and the mechanism of their success.

Abstract: Zebrafish is used by biologists as an animal model to study the effects of neurotoxicants and drugs on locomotion and develop pharmacological treatments. Very few fish swimming simulations have been derived from real body deformations to investigate the complex and stereotyped escape response of zebrafish and support animal experimentation. This experiment-driven numerical approach combined experimental imaging for modeling body deformations and three-dimensional (3D) numerical simulations to compute the actual energetic performances. To this end, a novel 3D reconstruction of the zebrafish shape was generated and deformed according to experimental data. As a first application, three escape locomotion were recorded and simulated across six high-viscosity fluids. In addition to kinematic data such as traveled distance, velocity, and bending amplitude, the expended energy and cost of transport were computed based on the power output. Mean power and expended energy seemed to remain steady over viscosity, while the cost of transport was found highly correlated to fluid viscosity despite considerably altered escape phenotypes. Eventually, fictitious simulations were performed by combining body deformations and fluid viscosity, especially for challenging the experimental escape motions. Such simulations have revealed zebrafish was particularly efficient to escape and energetic expenditure could be emphasized by increasing fluid viscosity. These results provided preliminary insights for the implementation of an effort test, involving zebrafish experiments, viscous fluids, and energetic performances from numerical simulations.

Abstract: We study the kinematics and dynamics of a highly compliant membrane disk placed head-on in a uniform flow. With increasing flow velocity, the membrane disks deform nonlinearly into increasingly parachute-like profiles. The experiments were carried out in a closed-loop low-speed wind tunnel with Reynolds number in the range of 10^4 -10^5. Remarkably, these aerodynamically sustained membrane disks show a higher flow resistance (drag) than similarly shaped rigid concave bodies. We model the steady structural response of the membranes using a nonlinear aeroelastic model. The predictions of the model agree well with the mean deformations of the membrane disks for the full range of experimental parameters studied. Through a simultaneous quantification of the unsteady membrane kinematics and forces, we detect the onset of large amplitude membrane fluctuations, match with the observed drag modulation and have their origins in the resonance between the flow structures and the membrane’s natural frequency. A drum model with anisotropic spring-stiffness is proposed, which quantitatively captures the observed resonant response. Further, PIV experiments are being conducted to yield deeper insights into the steady and transient fluid-structure interactions. 

Abstract: The current understanding of biologically inspired aquatic propulsion is extensive for simple two-dimensional models but is lacking for more complex three-dimensional models. This experimental work used a novel two degree-of-freedom full fish platform to investigate the relationships between model kinematics and system performance. This platform was used to understand which conclusions from simple models can be extended to full fish models. Phase-averaged torque input, thrust output, and kinematics were acquired for 567 cases that spanned a motion parameter space that included the maximum trailing edge excursion of the caudal fin, the phase offset between the tail and caudal fin motion, and the heave-to-pitch ratio. We have shown that thrust and efficiency are maximized when the phase offset is between 90 and 115 degrees. The optimal heave-to-pitch ratio was highly dependent on the trailing edge excursion and ranged between 0.43 and 0.78. The optimal parameters shown here are different than those for simplified two-dimensional models and warrant further investigation. We have shown that efficiency is maximized when the peak in power input is temporally aligned with the peak in power output. Together, these results hint at strategies for anticipating or designing model kinematics for specific performance goals.

Abstract: We experimentally optimise the pitch angle kinematics of a flapping wing system in hover to maximise the stroke average lift and hovering efficiency with the help of an evolutionary algorithm and in-situ force and torque measurements at the wing root. Additional flow field measurements are conducted to link the vortical flow structures to the aerodynamic performance. The pitch angle profiles yielding maximum average lift have trapezoidal shapes and high average angles of attack. These kinematics create a strong leading edge vortex early in the cycle which enhances the force production. The most efficient pitch angle kinematics resemble sinusoidal evolutions and have lower average angles of attack. The leading edge vortex grows slower and stays close-bound to the wing for the majority of the stroke-cycle. This increases the efficiency by 93% but sacrifices 43% of the lift in the process. We estimate the shear-layer velocity at the leading edge solely from the input kinematics and use it to scale the average and the time-resolved evolution of the circulation and the aerodynamic forces. The experimental data agrees well with the shear-layer velocity prediction, making it a promising metric to quantify and predict the aerodynamic performance of the flapping wing hovering motion.

Abstract: Potential flow plays an important role in many applications, including flow estimation in aerodynamics. For the models in these applications to work efficiently, it is best to avoid Biot-Savart interactions between the potential flow elements, particularly for 3D models. This work addresses grid-based computations for planar potential flows and their implementation in a low-order vortex model for fast modeling of separated aerodynamic flows and gust interactions. The model uses the immersed boundary projection method to solve for the vector potential field subject to the constraints introduced by the presence of a body, any edge conditions, and Kelvin’s circulation theorem, with each constraint adding a Lagrange multiplier to the overall saddle point system. Sharp edges are treated by decomposing the body forcing Lagrange multiplier into a singular and non-singular part. To enforce the Kutta condition, the non-singular part can then be tuned to remove the singularity introduced by the sharp edge. The equations are discretized on a staggered Cartesian grid and solved using the lattice Green’s function. The accuracy of these computations is demonstrated for a flat plate shedding singular vortex elements in 2D and the extension to 3D flows will be discussed.

Abstract: Visual observations of fluid-structure interactions (e.g. the swaying of trees in the wind) encode information about the surrounding flow conditions. We present a physics-based model that leverages the relationship between the force on a structure due to incident flow and the resulting structural deflection to measure normalized flow speeds. Visually-measured wind speeds are compared to ground truth, anemometer-measured values for wind tunnel experiments on flexible cantilevered cylinders and trees. Results suggest the potential for flow-speed inference using measurements of structural deflections, which can reduce the need to further instrument or visualize the flow in order to measure its speed.

Abstract: The fluid-structure interaction of an inverted flag in a uniform flow exhibits three different vibration modes: the stationary, flapping and deflected modes. The vortex-induced vibration of such a flag is studied both numerically and experimentally to identify its flapping behavior. Experiments of flags made of spring steel were conducted in a wind tunnel, where the wind speed was swept up and down through the various oscillatory modes of the inverted flags, while the numerical simulation of the same problem was thoroughly studied. It was noticed that during its mode transition, a difference in critical velocity occurs during upsweep and downsweep, creating a hysteresis that was shown to be controlled by the combination of the leading edge and the trailing edge vortices. To better understand this, we adapted a recently proposed force partitioning method technique and extended the method for flexible structures using integral boundary formulation of thin structures to identify the modal force contributions from the major flow features. We will discuss the connection between hysteretic behavior and the leading coherent vortices from the fluttering flags and explain how the flag aspect ratio affects its bistable vibration.

Abstract: Transcatheter heart valves (THV) suffer from clinically silent complications like subclinical leaflet thrombosis which may result in fatal outcomes for the patient. Such malfunction is detected incidentally during post-implant follow-up, and common imaging techniques are either invasive or expose the patient to radiation and are cost prohibitive. This informs a critical need for a novel, non-invasive and non-toxic continuous monitoring modality of THVs which can provide persistent and longitudinal monitoring of prosthesis function. We conduct a data-driven, in-silico investigation into the viability of wireless, remote monitoring of prosthetic aortic valve health using pressure microsensors. To do this, we developed a versatile reduced-order valve model capable of simulating a wide range of valve conditions. The strong coupling between leaflet mobility and downstream hemodynamics facilitates correlating pressure measurements at strategic locations in the vicinity of the THV with leaflet status. High-fidelity simulations of transvalvular flow in a canonical aorta model with various valve conditions allow us to construct hemodynamic “signatures” of healthy and dysfunctional leaflets. These signatures can be analyzed using supervised learning methods to determine optimal sensor configuration, detect the presence of reduced leaflet motion (RLM) and quantify its severity. Preliminary results demonstrate pressure measurements at as few as two discrete locations per valve leaflet can be used for accurate retrospective as well as prospective prediction of leaflet status (“Healthy”/ “RLM”) and its range-of-motion.

Abstract: The choice and placement of sensors and actuators is an essential factor determining the performance that can be realized using feedback control. This determination is especially important, but difficult, in the context of controlling transitional flows.  The highly non-normal nature of the linearized Navier-Stokes equations makes the flow sensitive to small perturbations, with potentially drastic performance consequences on closed-loop flow control performance. Full-information controllers, such as the linear quadratic regulator~(LQR), have demonstrated some success in reducing transient energy growth and suppressing transition; however,  sensor-based output feedback controllers with comparable performance have been difficult to realize. In this study, we propose two methods for sensor selection that enable sensor-based output feedback controllers to recover full-information control performance: one based on a sparse controller synthesis approach, and one based on a balanced truncation procedure for model reduction. Both approaches are investigated within linear and nonlinear simulations of a sub-critical channel flow with blowing and suction actuation at the walls. We find that sensor configurations identified by both approaches allow sensor-based static output feedback LQR controllers to recover full-information LQR control performance, both in reducing transient energy growth and suppressing transition. Further, our results indicate that both the sensor selection methods and the resulting controllers exhibit robustness to Reynolds number variations.

Abstract: We develop a cluster-based feedback control strategy to characterize the complex dynamics of cavity flow. This probabilistic model is utilized to modify the evolution of flow states to a desired dynamics using limited sensor measurements for capturing the attractor physics. The flow state is encoded into a low-dimensional feature space, which is partitioned into a set of clusters, where the occurrence of each flow state is represented by a cluster probability vector. The evolution of the flow-state probability distribution is given based on the transition probability matrix, which encapsulates the intra-cluster and inter-cluster transitions. We perform community detection on the transition probability to identify and characterize the flow physics among the most probable inter-cluster transitions. Further, we exploit the linearity of the probability transitions and formulate an LQR problem to control the cluster probability vector to a desired probability distribution. We first demonstrate this probabilistic control approach on the canonical Lorenz-63 system with bistability. We then employ the cluster based characterization of dynamics exhibited by the 2D laminar cavity flow. The feature space is chosen as the limited sensor measurements along the cavity wall. The clusters are further divided into communities and we identify a probable transition path from the low rms pressure fluctuations to large rms pressure fluctuations. We discuss our efforts to reduce the pressure fluctuations in the cavity using this cluster based approach.