Session A38: Vortex Dynamics and Vortex Flows: Wakes I

The mechanism of unsteady wake transition behind large depth-ratio wall-mounted prisms

The onset of wake unsteadiness for small aspect-ratio (height-to-width) wall-mounted prisms depends on both flow and geometrical parameters, i.e., depth-ratio (length-to-width) and Reynolds number (Re). While past studies have characterized the onset of unsteady wake for infinitely-span and large aspect-ratio (finite span) wall-mounted prisms, the effect of depth-ratio remains unexplored. As such, we numerically investigate the onset unsteady flow in the wake of wall-mounted finite prisms with an aspect-ratio of 1 and varying depth-ratios between 0.016 and 4 at Re=50-2500. The minimum depth-ratio considered here represents the special case of a wall-mounted thin flat plate. Preliminary results indicate that the onset of unsteady wake for wall-mounted thin flat plate occurs at Re=190, while that of a wall-mounted cube occurs at Re=400. Transition to unsteady wake for prisms with large depth-ratio occurs at Re=600. The onset of unsteady wake is characterized by symmetric hairpin-like vortex shedding, which leads to an oscillatory drag force and separating side-edge shear-layers. Moreover, threshold Reynolds number of transition to unsteady wake for wall-mounted prisms is significantly larger than that of finite-suspended prisms and thin flat plates. It is known that transition to unsteady wake occurs due to Hopf-bifurcation for suspended bluff bodies, whereas the present results indicate a different mechanism that governs this phenomenon for wall-mounted prisms. Thus, we aim to identify and characterize the mechanism of transition to unsteady wake for wall-mounted prisms with changing depth-ratio. Furthermore, we aim to characterize the various regimes of unsteady wake to fully understand the transition-to-turbulence wake phenomenon induced by the increasing depth-ratio in wall-mounted prisms.   

On the relationship between wake and cylinder dynamics for quasi-periodic vortex-induced vibrations

The weakly nonlinear response of a forced self-sustained oscillator for vortex-induced vibrations (VIV) in the initial branch is investigated for a 2D cylinder near a plane boundary in a uniform flow (Re=200) perturbed sinusoidally at resonant, 2f_o, and near-resonant conditions, 2.2f_o (f_o is the natural shedding frequency). The cylinder exhibits a quasi-periodic response and its underlying physics are not captured using common VIV models. The total force acting on the cylinder is decomposed into a vortex-induced force, F_v(t), and a force induced by effective mass, F_s(t). Fv(t) is linearly coupled to the cylinder's motion, while F_s(t) serves as a nonlinear coupling mechanism. A semi-empirical model is proposed, showing that the time-varying nature of the effective mass in F_s(t) drives the nonlinear response. This model elucidates the physics underlying quasi-periodic VIV and explains observed frequency drift and crosstalk. This model represents a non-isochronous oscillator, exhibiting four branches of response and two lock-on regimes similar to what is observed for VIV. Hence, it shows promise as a predictive tool for VIV response under various flow conditions.   

Flow-induced vibrations of a pair of rigidly linked circular cylinders in cross-flow

Vortex-induced vibration (VIV) experiments with an elastically mounted low mass-damping system of two cylinders have been conducted. The system consists of a pair rigidly linked circular cylinders with the same mass m, length L and diameter D, that are submerged in a uniform flow with velocity U, generated by the free surface channel at the Laboratory for Fluid-Structure Interaction (LIFE) of the Universitat Rovira i Virgili. The cylinders are separated a distance S_x in the flow direction and S_y in the perpendicular to the flow or transverse direction. The set-up hangs from an air bearing system to minimise friction when oscillating. Restoring forces are provided by a set of springs with stiffness constant k.

In one of the con configurations studied, different S_x and S_y were used, ranging from the tandem or in-line positioning to the side-by side configuration. The dynamic response was fully characterized, including measurements of amplitude and frequency of oscillation, as well as fluid excitation. The experiments covered a large parametric space, consisting of variations in the separations and the free stream velocity, leading to a wide range of reduced velocities and Reynolds numbers. The rich fluid-structure interaction phenomena observed in the experiments will be discussed and analyzed in detail.   

Three-Dimensional Compressible Wakes of Cargo Aircraft Afterbodies

The swept afterbody of cargo aircraft promotes a highly three-dimensional wake that generates additional drag and can result in hazardous conditions for parajumpers exiting the aircraft. These challenges have prompted the exploration of cargo aircraft wakes using a sharp-edged simplified bluff body known as the slanted-cylinder. This study expands current knowledge of the slanted-cylinder and further relates the geometry to real-world cargo aircraft through the addition of an edge radius. The three-dimensional flow features of the sharp-edge and rounded-edge afterbodies with a 45deg slant angle are explored under mildly compressible conditions (M=0.30) using Scanning-Stereoscopic PIV (S-SPIV). S-SPIV is a novel technique that utilizes a continuously traversing SPIV system to conduct spatio-temporal averaging of the flowfield, thereby reconstructing the mean (3D/3C) velocity field in a volume. The flow morphology derived from these reconstructed whole-field measurements provides valuable insight into the afterbody wake dynamics and parameters that influence its principal features such as the separating shear layer and the separation bubble. We also explore the impact of vortex stretching and compressible effects through the vorticity transport equation.   

Evaluating 2D and 3D VIV Wake Modes and Forces through Forced Motion Simulations

Simulation based approaches to studying and modeling vortex induced vibrations (VIV) can offer additional understanding to fluid structure interactions that may be difficult to study with experimental approaches, however simplified numerical approaches may lack the proper physics to capture all real effects. 2D simulations of a cylinder in a free stream forced to oscillate perpendicular to the free stream cannot capture spanwise effects that occur on finite span cylinders, resulting in less accurate solutions when compared to 3D, however they are considerably faster computationally. Using the verified boundary data immersion method (BDIM) fluid model ``WaterLily", 2D and 3D forced cylinder motion simulations at a Reynolds number of 4000 are performed and compared with measured experiments with equivalent parameters. WaterLily is written in ``Julia" and solves the unsteady incompressible Navier-Stokes equations on a Cartesian grid. Several combinations of reduced velocity (Vr) and non-dimensional transverse amplitude (Ay/D) from expected wake regimes (e.g., 2 single vortices shed per cycle, 2 pairs of vortices shed per cycle) are selected and both the wake formation and magnitude and phasing of the lift forces are compared. Results show that 3D simulations agree well with measured data whereas 2D simulations show better accuracy in low amplitude regimes and are in many cases unable to produce widely observed wake mode regimes at higher reduced velocity. This comparison provides insight as to where 2D simulations are appropriate in modeling VIV and where 3D effects play an important role in properly modeling the physics. This study provides a first step towards developing 2D forced motion models that can be corrected for 3D effects to yield a model with 3D accuracy and 2D computational speed.   

New insights into complex vortex-induced vibrations through 3D modeling and wake response analysis

In this work we focus on simulating vortex-induced vibrations (VIV) using real-time numerical simulations and validating their accuracy. We employed the boundary data immersion method (BDIM) to conduct simulations for one-degree-of-freedom (1DOF) VIV. The simulations demonstrated good agreement with experimental data at low reduced velocities and non-dimensional amplitudes. However, at higher reduced velocities and amplitudes, three-dimensional (3D) effects become more significant, resulting in poor comparisons of 2D simulations with physical experiments. Simulations were extended to 3D, showing that 2D simulations were inadequate in reproducing certain wake modes, such as 2P mode, whereas 3D simulations successfully reproduce these modes. Additionally, 2D simulations were used to recreate the wake response map of 1DOF VIV while applying Proper Orthogonal Decomposition to identify the most coherent modes in the boundary regions. These findings provide valuable insights into wake modes associated with VIV and emphasize the significance of considering 3D effects for more accurate simulations and understanding.   

Numerical Simulations of Flow Patterns in Crossflow Rotary Devices

Crossflow rotary devices, including both Darrius-type turbines and cyclorotors used for propulsion, have been widely proposed for application in aquatic contexts. Increasing understanding of the negative effects of anthropogenic acoustic noise on aquatic ecosystems has driven interest in numerically predicting radiated noise from marine devices. To date, most of the regulatory and research efforts have focussed on the effects of conventional marine propellers, while crossflow devices remain relatively unexplored. In turbomachinery applications, the primary sources of noise are shed vorticity and cavitation, when it occurs. Predicting radiated noise from vortices and cavitation requires an accurate solution of the shed vorticity in the wake and the resulting fluctuating pressure. In this study, we show that the flow patterns inside the crossflow rotor is a primary determinant of fluctuating pressure. We present a set of CFD simulations spanning advance ratios from 0.6 to 3.6. The solutions of the unsteady Reynolds-averaged Navier-Stokes (URANS) equations suggest that the production of noise cannot reliably be predicted by the operational speed or flow-through rate of the device and is instead dependant on vortex-blade interactions.   

A comparison between the near wake behind a cantilevered square cylinder and a periodic array of staggered vortex ring pairs

A comparative analysis is presented between the evolution of half-loop shedding patterns, identified in the wake of a cantilevered square cylinder, and that of a periodic street of staggered, titled vortex ring pairs. The near-wake of the cylinder with a height-to-width ratio h/d = 4, protruding a thin laminar boundary layer of thickness δ/d = 0.21, is investigated experimentally using time-resolved stereoscopic particle image velocimetry at a Reynolds number (Re) of 10600, based on the obstacle diameter. This wake region is characterized with half-loop shedding patterns, which connect successive Kármán vortices. As a simplified model of the half-loop patterns, the evolution of a periodic street of staggered, opposite-signed vortex ring pairs is simulated at a circulation-based Re of 5000. Expressed in terms of the ring radius R, the vortex rings are initialized with streamwise and spanwise distances of R and 0.4R, respectively. Further, the rings on either side of the symmetry plane are tilted with equal but opposite angles of π/6 radians. We discuss the qualitative similarity between the evolution of the model flow and the experimental measurements in terms of the global vorticity field. More importantly, we point out key differences between experiments and model due to the effect of the incoming boundary layer over the ground plate.   

Effect of elastic coupling on flow-induced vibration (FIV) of elastically mounted tandem cylinders.

The flow-induced vibration (FIV) of elastically coupled tandem cylinders is useful in modeling many engineering applications such as overbridges between buildings, structural connections between parallel tubes and pipelines, connected offshore wind turbines, twin bridge connections, etc. While there are several studies on FIV of uncoupled tandem cylinders, there is no study on the coupled system, to our knowledge. Using numerical simulations at low Reynolds number (Re=100), we present the FIV response of the cylinders for a wide range of gap ratios (1.1-5) and reduced velocities (4-18). The stationary tandem cylinder configuration can result in six possible flow structures, while the elastically coupled cylinders exhibit two possible mode shapes (in-phase and out-of-phase). Consequently, ten FIV regimes result from a combined flow and elastic coupling between the cylinders.

Our results suggest that the same system with an identical mass ratio can be tuned for FIV suppression or energy harvesting. The in-phase mode limits the upstream cylinder amplitude below 0.3 for the considered range of reduced velocity and gap ratio, without adding any damping. Similarly, the out-of-phase mode shows a higher energy harvesting potential for a single energy harvesting unit than two on the classic uncoupled tandem cylinder configuration.