Session R14: Aerodynamics: General

Experimental Study of Thin and Thick Airfoils at Low Reynolds Numbers

A recent surge in applications of unmanned air vehicles in various fields has led to increased interest in understanding the characteristics of airfoils at Reynolds number regime $\sim$10$^{4}$. At these low Re numbers, aerodynamics of an airfoil is influenced by laminar separation and its possible reattachment, which is in contrast to airfoil behavior at high Re numbers. This study focused on comparing the load characteristics of symmetric, thin (NACA-0009) and thick (NACA-0021) airfoils at low Re numbers $\sim$2 -- 4 $\times$ 10$^{4}$, and angles of attack between 2$^{\circ}$ to 12$^{\circ}$, along with simultaneous flow visualization. The experiments were performed in a low speed flow visualization water tunnel facility, and two-component Laser Doppler Velocimetry was used to quantify the inflow conditions and turbulence intensity. A high precision force/torque transducer was used for the load measurements, while hydrogen bubble technique was used for flow visualization on the suction side of the airfoils. The presentation will discuss the correlation between observed flow structures and instantaneous load on the airfoils, as well as the aerodynamic load characteristics of thin and thick airfoils at low Re numbers.   

Formation Flight: Upstream Influence of a Wing on a Streamwise Vortex

Aircraft flying together in formation can experience aerodynamic advantages. Impingement of the tip vortex of the leader wing on the trailer wing can increase the lift to drag ratio L/D and the unsteady loading on the trailer wing. These increases are sensitive to the impingement location of the vortex on the wing. Particle image velocimetry is employed to determine patterns of velocity and vorticity on successive crossflow planes along the vortex, which lead to volume representations and thereby characterization of the streamwise evolution of the vortex structure as it approaches the trailer wing. This evolution of the incident vortex is affected by the upstream influence of the trailer wing, and is highly dependent on the location of vortex impingement. As the spanwise impingement location of the vortex moves from outboard of the wing tip to inboard, the upstream influence on the development of the vortex increases. For spanwise locations close to or intersecting the vortex core, the effects of upstream influence of the wing on the vortex are to: increase the streamwise velocity deficit; decrease the streamwise vorticity; increase the in-plane vorticity; decrease the downwash; and increase the root-mean-square of both streamwise velocity and vorticity.   

Interaction of a perturbed trailing vortex with a wing

The interaction of a perturbed trailing vortex with a wing is explored in the context of formation flight. The incident vortex is generated from a leader wing at a fixed angle-of-attack, which is subjected to controlled oscillation in the vertical direction over a range of frequencies and at amplitudes an order of magnitude smaller than the chord of the wing. Particle image velocimetry is used to determine the flow structure of the perturbed vortex along the stationary follower wing. Two spanwise locations of vortex impingement are characterized: aligned with and inboard of the tip of the wing. Images of streamlines, vorticity and turbulent kinetic energy reveal the temporal and spatial evolution of the vortex structure. For the aligned interaction, the incident vortex induces a tip vortex of opposite signed vorticity on the wing thereby forming a dipole. For the inboard interaction, the incident vortex induces a vortex of same sign vorticity at the wing-tip and a shear layer of opposite sign across the wing surface. The upwash of the incident vortex promotes flow separation at the leading edge for both of these interactions; this separation gives rise to substantial levels of turbulent kinetic energy that persist downstream through the region of reattachment.   

Numerical Investigation of Capability of Self-Starting and Self-Rotating of a Vertical Axis Wind Turbine

The immersed boundary method is used to simulate the incompressible flow around two-dimensional airfoils at low Reynolds numbers in order to investigate the self-starting and self-rotating capability of a vertical axis wind turbine (VAWT) with NACA 0018 blades. By examining the torque generated by a three-bladed VAWT fixed at various orientations, a stable equilibrium and the optimal starting orientation that produces the largest torque have been observed. When Reynolds number is below a critical value, the VAWT oscillates around a stable equilibrium. However, the VAWT goes into continuous rotation from the optimal orientation when Reynolds number is above this critical value. It is also shown that VAWT with more blades is easier to self-start due to a wilder range of positive starting torques. Moreover, with a proper choice of load model, a VAWT is able to self-rotate and generate a designed averaged power.   

Numerical investigation of the effect of sphere dimples on the drag crisis and the Magnus effect

The present study investigates the flow over a golf ball and a smooth sphere around the critical Reynolds numbers under both stationary and self-spinning conditions by conducting Large-eddy simulations (LES) based on high resolution unstructured grids. For the stationary cases, the present calculation results validate the promotion of the drag crisis at a relatively lower Reynolds number due to the golf ball dimples. It also shows that the golf ball dimples have a limited effect on the time-dependent lateral force development in the subcritical regime, whereas the dimples are beneficial in suppressing the lateral force oscillations in the supercritical regimes. With spin parameter $\Gamma =$0.1, the drag coefficients for the spinning smooth sphere increase slightly in all Reynolds number regimes when compared to the stationary cases, whereas for the spinning golf ball, the drag force decreases in the critical regime and increases in the supercritical regime. For both spinning models, the inverse Magnus effect was reproduced in the critical regime, whereas in the supercritical regime the ordinary Magnus force was generated. Relatively weaker lift forces were also observed in the cases of the spinning golf balls when compared to the spinning smooth spheres.   

Large-eddy simulation of a turbulent flow over a heavy vehicle with drag reduction devices

Aerodynamic drag contributes to a considerable amount of energy loss of heavy vehicles. To reduce the energy loss, drag reduction devices such as side skirts and boat tails, are often installed to the side and the rear of a heavy vehicle. In the present study, turbulent flow around a heavy vehicle with realistic geometric details is simulated using large-eddy simulation (LES), which is capable of providing unsteady flow physics responsible for aerodynamic in sufficient detail. Flow over a heavy vehicle with and without a boat tail and side skirts as drag reduction devices is simulated. The simulation results are validated against accompanying in-house experimental measurements. Effects of a boat tail and side skirts on drag reduction are discussed in detail.   

On the lift increments with the occurrence of airfoil tones at low Reynodls numbers

The aeroacoustic effects on the aerodynamics of an NACA 0006 airfoil are investigated experimentally at relatively low Reynolds numbers, $Re=30,000$ - $70,000$. By employing two wind-testing airfoil models at different chord lengths, $L = 40$ and $100$ [mm], the aerodynamic dependence on Mach number is examined at a given Reynolds number. In a particular range of Reynolds number, tonal peaks of trailing-edge noise are obtained from a shorter-chord airfoil, while no apparent tones are observed with longer chord length at a lower Mach number. Surprisingly, the occurrence of a tonal noise leads to a greater lift slope in the present wind-tunnel experiment, evaluated via a PIV approach. The lift curves obtained experimentally at higher Mach numbers agree well with two-dimensional numerical simulations, performed at $M = 0.2$. At the Mach number, the numerical results clearly indicate the occurrence of an acoustic feedback loop with discrete tones, within a range of angle of attack. A few three dimensional numerical results are also presented. In the simulation at $Re = 50,000$, the suppression of tonal noise corresponds to the development of a turbulent wedge in the suction-side boundary layer at the angle of attack $4.0$ [deg.], which agrees with the experiment.   

Reynolds number effects on flow over twisted cylinder with drag reduction and vortex suppression

We investigated the Reynolds number effects on the flow over a twisted cylinder in the range of 3×10$^{\mathrm{3}}\le $ Re$\le $1×10$^{\mathrm{4}}$. To analyze the effect of the twisted cylinder, a large eddy simulation (LES) with a dynamic subgrid model was employed. A simulation of the smooth cylinder was also carried out to compare the results with those of the twisted cylinder. As Re increased, the mean drag and lift coefficient of the twisted cylinder increased with the same tendency as those of the smooth cylinder. However, the increases in the mean drag and lift coefficient of the twisted cylinder were much smaller than those of the smooth. Furthermore, elongated shear layer and suppressed vortex shedding from the twisted cylinder occurred compared to those of the smooth cylinder, resulting in a drag reduction and suppression of the vortex-induced vibration (VIV). In particular, the twisted cylinder achieved a significant reduction of over 96{\%} in VIV compared with that of the smooth cylinder, regardless of increasing Re. As a result, we concluded that the twisted cylinder effectively controlled the flow structures with reductions in the drag and VIV compared with the smooth cylinder, irrespective of increasing Re.   

Reynolds Number Effects on Helicopter Rotor Hub Flow

The 12 inch diameter water tunnel at the Pennsylvania State University Applied Research Laboratory was used with the objective of quantifying effects of Reynolds number scaling on drag and shed wake of model helicopter rotor hub flows. Hub diameter-based Reynolds numbers ranged from 1.06 million to 2.62 million. Measurements included steady and unsteady hub drag, as well as Particle Image Velocimetry. Results include time-averaged, phase-averaged, and spectral analysis of the drag and wake flow-field. A strong dependence of steady and unsteady drag on Reynolds number was noted, alluding to the importance of adequate Reynolds scaling for model helicopter rotor hubs that exhibit interaction between various bluff bodies.   

Numerical Investigations of an Optimized Airfoil with a Rotary Cylinder

Numerical Investigations of an optimized thin airfoil with a rotary cylinder as a control device for reducing separation and improving lift to drag ratio have been performed. Our previous investigations have used geometrical optimization for development of an optimized airfoil with increased torque for applications in a vertical axis wind turbine. The improved performance was due to contributions of lift to torque at low angles of attack. The current investigations have been focused on using the optimized airfoil for micro-uav applications with an active flow control device, a rotary cylinder, to further control flow separation, especially during wind gust conditions. The airfoil has a chord length of 19.66 cm and a width of 25 cm with 0.254 cm thickness. Previous investigations have shown flow separation at approximately 85{\%} chord length at moderate angles of attack. Thus the rotary cylinder with a 0.254 cm diameter was placed slightly downstream of the location of flow separation. The free stream mean velocity was 10 m/sec. and investigations have been performed at different cylinder's rotations with corresponding tangential velocities higher than, equal to and less than the free stream velocity. Results have shown more than 10{\%} improvement in lift to drag ratio when the tangential velocity is near the free stream mean velocity.   

Comparative Study of Airfoil Flow Separation Criteria

Airfoil flow separation impacts a multitude of applications including turbomachinery, wind turbines, and bio-inspired micro-aerial vehicles. In order to achieve maximum performance, some devices operate near the edge of flow separation, and others use dynamic flow separation advantageously. Numerous criteria exist for predicting the onset of airfoil flow separation. This talk presents a comparative study of a number of such criteria, with emphasis paid to speed and accuracy of the calculations. We evaluate the criteria using a two-dimensional unsteady vortex lattice method, which allows for rapid analysis (on the order of seconds instead of days for a full Navier-Stokes solution) and design of optimal airfoil geometry and kinematics. Furthermore, dynamic analyses permit evaluation of dynamic stall conditions for enhanced lift via leading edge vortex shedding, commonly present in small flapping-wing flyers such as the bumblebee and hummingbird.   

A Computational Modeling Mystery Involving Airfoil Trailing Edge Treatments

In a curious result, Fairman (2002) observed that steady RANS calculations predicted larger lift than the experimentally-measured data for six different airfoils with non-traditional trailing edge treatments, whereas the time average of unsteady RANS calculations matched the experiments almost exactly. Are these results reproducible? If so, is the difference between steady and unsteady RANS calculations a numerical artifact, or is there a physical explanation? The goals of this project are to solve this thirteen year old mystery and further to model viscous/load coupling for airfoils with non-traditional trailing edges. These include cupped, beveled, and blunt trailing edges, which are common anti-singing treatments for marine propeller sections. In this talk, we present steady and unsteady RANS calculations (ANSYS Fluent) with careful attention paid to the possible effects of asymmetric unsteady vortex shedding and the modeling of turbulence anisotropy. The effects of non-traditional trailing edge treatments are visualized and explained.