14:00
Fluid-Structure Interaction 2
14:00
15 mins
#12
|
On the dynamics of multiple elastically-bounded flapping plates for flow energy harvesting
Stefano Olivieri, Corrado Boragno, Roberto Verzicco, Andrea Mazzino
Abstract: Aeroelastic phenomena are gaining significant attention from the energy harvesting (EH) perspective with promising applications, e.g., as a way of supplying electrical power to low-power remote sensors~\cite{mccarthy2016review}.
Besides the development of the individual EH device,
further research efforts are required when considering multiple objects in order to create an array of devices. Due to the inherently nonlinear character of these systems, the resulting dynamics is generally different with respect to the single device and the optimisation of such array turns out to be nontrivial.
In the present work, this point is faced by considering a flutter-based EH system consisting of a rigid plate that is anchored by means of elastic elements and invested by laminar flow, experiencing a flapping motion with regular limit-cycle oscillations of finite amplitude~\cite{olivieri2017fluttering,boccalero2017power}.
In our investigation, we consider a simplified yet general physical model~\cite{olivieri2017fluttering} and employ three-dimensional numerical simulations based on a finite-difference Navier-Stokes solver combined with a moving-least-squares immersed boundary method~\cite{detullio2016}.
Focusing on the kinematic and performance-related quantities, we first report on the dynamics of the single device while varying the main governing parameters.
Hence, we focus on multiple devices and consider different array configurations, i.e. in-line and side-by-side. By varying systematically the mutual distance between the devices and tuning other structural parameters, qualitative insights are provided. Among our main findings, cooperative effects in the side-by-side arrangement are found to be substantially beneficial in terms of the resulting power coefficient, with increases of about 100\% with respect to the single-device case.
In order to confirm the numerical evidence, complementary results from wind-tunnel testing of experimental prototypes are presented.
Finally, we describe the behaviour when increasing further the number of devices, with the ultimate goal in developing a network of numerous aeroelastic energy harvesters.
|
14:15
15 mins
#239
|
MODELLING DYNAMIC STALL OF A PITCHING AIRFOIL IN LARGE-SCALE FREESTREAM TURBULENCE
ThankGod Boye, Zheng-Tong Xie
Abstract: The aspect of dynamic stall associated with large-scale freestream turbulence effect is not well understood particularly in flow over a pitching airfoil. Substantial understanding of the effect of large-scale freestream turbulence is crucial for pitching airfoil because of the need for innovative aerodynamic design of modern large wind turbine blades. Although considerable experimental and numerical work of laminar inflow have been done to investigate aerodynamic characteristics on the dynamic stall phenomena of an airfoil, Gandhi et al. [1] reported that there are a few experimental studies of the effect of the freestream turbulence on a pitching airfoil and even fewer numerical studies, perhaps due to the complexity of the methods involved to develop and implement the incoming turbulence inflow conditions in experiments and numerical simulations. Researchers who have studied turbulence effects on the dynamic stall of an airfoil focus on the effect of high turbulent intensity and small integral length-scales [2-3]. They found that an increase of the turbulent intensity resulted in an increase of the lift coefficient, but little research particularly on the effect of large integral length-scales has been conducted.
Turbulence integral length-scales in the atmospheric boundary layer are typically much greater than the airfoil chord length. The effects of the large-scales turbulence on the aerodynamic characteristics of airfoil are of greater importance than the effects of the high turbulence intensity [4]. Mahmoodilari et al. [5] identified that an increase in integral length-scale aversively affects the lift. Ravi et al. [6-7] revealed that the integral length-scale and turbulent intensity had opposing influence on the lift coefficient, and an increase in integral length-scale rendered an increase in lift. Harbst et al. [8] reported that an increase in integral length-scale enhances the probability of occurrence of short separation bubbles close to the leading edge of the airfoil. Therefore it is not clear whether the incoming large eddies improve performance of the aerodynamic forces of a pitching airfoil. The advanced numerical modelling methods such as large-eddy simulation (LES) are promising to simulate complex unsteady flow and to capture crucial flow physics of large-scale turbulence interactions with a pitching airfoil.
In this paper, we aim to investigate the effect of large integral length-scales on a pitching NACA 0012 airfoil at a moderate Reynolds number Re = 1.35 x 105 based on the chord length and the freestream velocity. A turbulence inflow condition method for LES is employed for the study. At the inlet large integral length-scale greater than the chord length are used for modelling the energetic large-eddies impacting on the pitching airfoil at various reduced frequencies kred = 0.1, 0.15 and 0.2. Preliminary results of baseline simulations show good agreement with experimental and numerical data in [3]. The lift, drag and moment coefficients and corresponding flow-field of the large-scale turbulence structures will be presented. We are more focused on revealing the effects of the large integral length-scales on the dynamic stall of the pitching airfoil. The data and knowledge derived from this study will contribute to enhancing innovative aerodynamic design of modern large wind turbine blade that operates within the atmospheric boundary layer.
References
[1] A. Gandhi, B. Brandon and Y. Peet. Effect of reduced frequency on dynamic stall of a pitching airfoil in turbulent wake, AIAA SciForum,
AIAA Aerospace Sciences Meeting, 55: 1-15, 2017.
[2] X. Amandolese and E. Szechenyi. Experimental study of the effect of turbulence on a section model blade oscillating in stall. Wind Energy,
7(4):267-282, 2004.
[3] Y. Kim and ZT Xie. Modelling the effect of freestream turbulence on dynamic stall of wind turbine blade. Comp. fluids, 129:53-66, 2016.
[4] J.Stack. Test in the variable density wind tunnel to investigate the effect of scales and turbulence on airfoil characteristics. Technical notes
National advisory committee for aeronautics, 364:1-15 1931.
[5] M. Mahmoodilari. The effect of turbulence flow on wind turbine loading and performance. Ph.D. Thesis, University of Manchester, 2012.
[6] S. Ravi, S.Watkins, J. Watmuff, K. Massey, P. Peterson and M. Marino. Influence of large-scale freestream turbulence on the performance
of thin airfoi. AIAA Journal, 50:2448-2459, 2012.
[7] S. Ravi, S.Watkins, J. Watmuff and A. Fisher. Transient Loads occurring over a thin airfoil subjected to large-scale freestream turbulence.
AIAA journal, 51(6):1473-1485, 2013.
[8] S.L. Harbst, C.J. Kahler and R. Hain. Influence of large-scale freestream turbulence on an SD7003 airfoil at low Reynolds numbers. AIAA
Aviation Forum Applied aerodynamics conference, 1-13, 2018.
|
14:30
15 mins
#263
|
The effect of turbulence on the near-field of porous disks
Hauk-Morten Heimlund Lykke, Magnus Kyrkjebø Vinnes, R. Jason Hearst
Abstract: Wind tunnel experiments have been conducted with a solid disk (SD) and two different porous disks with identical blockage (57%) placed in flow with two different turbulence intensities (0.6% and 4.1%). The first porous disk was a wire mesh disk (WD) and the other was a machined acrylic disk (AD) with quasi-rectangular holes arranged in concentric circles. The experiments were performed at a Reynolds number of Re_D=110 000. Velocity measurements were performed in the wakes, and higher-order statistics were compared. The wakes were investigated over 0.5 <= x/D <= 30, and they remain disparate at the farthest downstream measurement position for all quantities measured. The objective was to quantify characteristics of the near wakes and to determine to what extent and how the shape of a porous disk affects the wake. The mean velocity U, turbulence intensity, skewness and kurtosis are reported. Few previous studies have investigated porous disks at this Re, and fewer still have reported skewness and kurtosis; one of the exceptions is Aubrun et al. In agreement with earlier studies, the decay rate of the wake half width was found to increase and centreline velocity deficit decreased for all disks when subjected to incoming turbulence. However, the onset of the wake recovery was found to be delayed for the WD, which is surprising. The WD and the AD create very different turbulence intensity profiles, where the WD forms a narrow and separated annular shear layer at the disk edge as opposed to the strongly turbulent region behind the entire AD, which more closely resembles the wake of the SD. The incoming turbulence appears to speed up the onset of self-similarity for the SD and the AD, which is very interesting considering earlier results stating that self-similarity was prevented by the incoming turbulence for x>60D. However, this might be explained by a faster transition to the wake's final state caused by the turbulence, including the appearence and then dissappearence of self-similarity in the region 30D
|
14:45
15 mins
#593
|
MODIFYING SPATIAL LARGE-SCALES USING BLOWING PERTURBATIONS
Venkatesh Pulletikurthi, Suranga Dharmarathne, Murat Tutkun, Luciano Castillo
Abstract: Study of wall-bounded flows to reduce drag is of major interest to researchers because of its economical impact on energy
consumption. Controlling the fluid flow in the near-wall region reduces the drag [3, 5], but most controlling techniques
entail altering whole length of the boundary, which incurs additional cost. Our study aims to control the flow using the
blowing perturbations near the inlet rather than entire boundary layer.
Large-scale motions contribute to more than 60% Reynolds shear stresses and turbulent kinetic energy [1]. Castillo and
Johansson [2] observed that inlet conditions influence the flow evolution downstream. It was also found that blowing
perturbations lead to the proliferation of coherent structures [4]. In this study, we used Direct Numerical Simulations
(DNS) to simulate a channel flow (frictional Reynolds number, Re 394) with two—five equally spaced spanwise jets
with diameter D to create blowing perturbations near the inlet. We study the effect of the perturbations on the large-scale
motions and the influence of perturbations on the coherent vortical structures of large-scale motions.
Figure 1a shows the pre-multiplied streamwise velocity spectra at 3D downstream of the jets. It can be seen that outermost
contour stretches close to the top right corner of the graphs as the number of jets increases. The largest wavelength of the
structures increases with the number of jets and, as expected, the amount of energy supplied to the flow increases with the
number of jets. The inlet blowing perturbations effect on the large-scale coherent vortices can be seen from Figure 1b.
From 1D to 10D, there is a proliferation of large-scale vortical structures behind the jets, and it can be seen that heads
of hair-pin vortices are formed from the large-scale motions. These results show that upstream blowing perturbations
influence the large-scale motions downstream and can be used to modify the flow.
|
15:00
15 mins
#42
|
On turbulence and wave energy converters
Frederic Dias, Joao Bettencourt
Abstract: It is generally thought that viscous effects must be considered in the analysis of wave energy devices. Neglecting such effects might overestimate the power capture of the system. It is difficult to quantify viscous effects by tank experiments or real sea observations, whereas numerical modelling is a common way to estimate viscous effects (turbulence, vortex shedding). In order to assess the influence of turbulence on a special type of wave energy converter, namely the oscillating wave surge converter (OWSC), we performed numerical simulations using five built-in turbulence models: Standard k-ε Model (SKE), Re-Normalisation Group k-ε model (RNG), Realisable k-ε model (RKE) , Shear-Stress Transport k-ω model (SST) and Reynolds Stress Model (RSM). It was found that the torques of the flap due to the wave motion are not sensitive to the turbulence model and that vortex shedding from the flap is a short-lived, periodic phenomenon. For small Keulegan–Carpenter numbers, the shed vorticity layer rolls-up into vortices that do not travel far from the body. Consequently, under normal operating conditions, viscous effects and turbulence are not important issues for OWSCs.
However, in extreme seas, wave induced fluid velocities can be very high and extreme wall pressure fluctuations may occur. Moreover, the spanwise vortical fields oscillate violently in a wave cycle, inducing strong interactions between vortices and the device that can enhance the device motion. We studied the wall pressure and vorticity fields of the Stokes boundary layer in the intermittently turbulent regime through direct numerical simulation (DNS). The DNS results were compared to experimental measurements and a good agreement was found for the mean and fluctuating velocity fields. We observed maxima of the turbulent kinetic energy and wall shear stress in the early deceleration stage and minima in the late acceleration stage. The wall pressure field is characterized by large fluctuations with respect to the root mean square level, while the skewness and kurtosis of the wall pressure show significant deviations from their Gaussian values. The wall vorticity components show different behaviours during the cycle: for the streamwise component, positive and negative fluctuations have the same probability of occurrence throughout the cycle while the spanwise fluctuations favour negative extrema in the acceleration stage and positive extrema in the deceleration stage. The wall vorticity flux is a function of the wall pressure gradients. Vorticity creation at the wall reaches a maximum at the beginning of the deceleration stage due to the increase of uncorrelated wall pressure signals. The spanwise vorticity component is the most affected by the oscillations of the outer flow.
|
|