Aerofólio S809 - 1 - s2 0 - s0894177712000404 - naca2415

Aerofólio S809 - 1 - s2 0 - s0894177712000404 - naca2415

(Parte 1 de 3)

An experimental study on aerodynamics of NACA2415 aerofoil at low Re numbers

M. Serdar Genç a,⇑, _Ilyas Karasu a,b, H. Hakan Açıkel aWind Engineering and Aerodynamics Research Laboratory, Department of Energy Systems Engineering, Erciyes University, 38039 Kayseri, Turkey_Iskenderun Civil Aviation School, Mustafa Kemal University, 31200 _Iskenderun, Hatay, Turkey article i nfo

Article history: Received 13 September 2011 Received in revised form 20 January 2012 Accepted 23 January 2012 Available online 2 February 2012

Keywords: Low Reynolds number Transition Laminar separation bubble Hot-wire anemometry Oil flow visualization abstract

This study is a detailed experimental investigation on aerodynamics of a NACA2415 aerofoil by varying angle of attack from 12 to 20 at low Reynolds number flight regimes (0.5 105 to 3 105). For this investigation, pressure distributions over the aerofoil were measured using a system including a pitotstatic tube, a scanivalve unit and a pressure transducer. Moreover, time-dependant lift and drag forces and pitch moment of the aerofoil were obtained by using an external three-component load-cell system. Velocity measurements at different points over the aerofoil were carried out by using a hot-wire anemometer, and oil flow visualization method was used to photograph the surface flow patterns. The experimental results showed that as the angle of attack increased, the separation and the transition points moved towards the leading edge at all Reynolds numbers. Furthermore as the Reynolds number increased, stall characteristic changed and the mild stall occurred at higher Reynolds numbers whereas the abrupt stall occurred at lower Reynolds numbers. The stall angle varied with Re number due to the viscous effects and decreased with decreasing Re number. By the decreasing of the Re number, short bubble burst at higher angles of attack, which caused long bubble to occur. 2012 Elsevier Inc. All rights reserved.

1. Introduction

Low Reynolds (Re) number aerodynamics has gained more attention due to increasing applications of Unmanned Air Vehicle (UAV), Micro Air Vehicle (MAV) and wind turbine. At low Reynolds number flows, laminar separation bubble may cause negative effects, such as decreasing on lift, increasing on drag, reducing stability of the aircraft, vibration, and noise [1–3]. It is important to understand the behavior of laminar separation bubbles (LSBs) to design a control system of LSB and to compose a new aerofoil which is no affected from LSB. Gaster [4] performed an experimental study about LSB by using hot-wire anemometer and oscilloscope. This study was carried out over a wide range of Re numbers and in a variety of pressure distributions. The bursting circumstances of short bubbles were determined by a unique relationship between Re number and pressure rise. Consequently, LSB was classified as short and long bubble just as Tani [5] did. Tani showed that when Re number was decreased the short bubble burst at higher angles of attack, and that caused long bubble or unattached flow to occur. Rinioie and Takemura [6] carried out an experimental investigation on NACA0012 aerofoil at Re = 1.35 105. They obtained that the short bubbles formed at which angle of attack was less than 1.5 , the long bubbles occurred at which angle of attack was more than 1.5 . Tan and Auld [7] studied experimentally on Wortmann FX67-150K aerofoil at low Reynolds numbers. They concluded that as the Re number and turbulence intensity of the freestream increased, earlier transition occurred and this caused length of LSB to shorten. Sharma and Poddar [8] conducted on an experimental investigation about the formation of LSB and transition process on NACA0015 aerofoil by varying the angle of attack from 5 to 25 at low Re numbers. In this study, the flow visualizations were done by using oil flow technique for qualitative analysis of the transition zone. It was resulted that as the angle of attack was increased the separation bubble moved towards the leading edge of the aerofoil and then bursted at a particular angle of attack, and the bubble bursting caused abrupt stall to occur. Ricci et al. [1,2] studied on formation and controlling LSB which occurred on the upper surface of aerodynamic bodies at low Reynolds number. The experiments were carried out to get the locations of the laminar bubble characteristic points (separation, transition and reattachment points) with and without the acoustic disturbance. The frequency range of the acoustic force was between 200 and 800 Hz with a step of 100 Hz was inspected. They presented that a sinusoidal sound wave having determined frequency reduced the laminar bubble longitudinal dimensions retarding the separation and anticipating the reattachment.

Diwan and Ramesh [9] investigated experimentally the length and height of the LSB on a flat plate at different Re numbers. It was obtained that both length and height of the LSB decreased,

0894-1777/$ - see front matter 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.expthermflusci.2012.01.029

E-mail addresses: musgenc@erciyes.edu.tr (M.S. Genç), ikarasu@mku.edu.tr (_I. Karasu), halilhakanacikel@gmail.com (H. Hakan Açıkel).

Experimental Thermal and Fluid Science 39 (2012) 252–264 Contents lists available at SciVerse ScienceDirect

Experimental Thermal and Fluid Science journal homepage: w.else vier.com/locate/etfs

and reducing ratio of the length is more than that of the height as Re number was risen. Yang et al. [10] carried out an experimental study about LSB over GA (W)-1 aerofoil at different low Reynolds number. It was resulted that while maximum length of the bubble was 20% of the chord length, the maximum height of the bubble was only 1% of the chord length. Furthermore, they showed that at more than angles of attack of 7 Kelvin-Helmholtz instabilities caused unsteady vortexes caused by LSB. Hain et al. [1] presented dynamics of LSBs at low-Reynolds-number aerofoils. It was found that Kelvin-Helmholtz instabilities had a weak coherence in the spanwise direction, and in a later stage of transition these vortices led to a three-dimensional breakdown to turbulence. Burgmann et al. [12] studied experimentally on the flow over SD7003 aerofoil used in wind turbine blades at low Re numbers. They concluded that vortex roll-up which was initialized by Kelvin-Helmholtz instability played effective role at transition process. Haggmark et al. [13] conducted experimental and numerical investigation on transitional separation bubble over elliptical leading edged flat plate. The flow was modelled numerically using engineering transition prediction methods and two-dimensional direct numerical simulations (DNS). They found that low-amplitude two-dimensional instability waves leaded to transition, and eN method gave a result in accordance with experiments and DNS. Lang et al. [14]

Nomenclature c chord length (m) b span length (m) a angle of attack ( ) D drag force (N) L lift force (N) M moment force (N)

U1 freestream velocity (m/s) u0 velocity fluctuation (m/s) q density of air (kg/m3) l dynamic viscosity of air (Ns/m2)

P1 freestream pressure (Pa) P local pressure (Pa)

Lb bubble length Xs separation point Xt transition point Xr reattachment point URe uncertainty of Reynolds number UC uncertainty of pressure coefficient UC uncertainty of lift coefficient UC uncertainty of drag coefficient UC uncertainty of moment coefficient

Fig. 1. Photograph of the experimental set-up.

Fig. 2. Schematic diagram of the experimental set-up.

M.S. Genç et al./Experimental Thermal and Fluid Science 39 (2012) 252–264 253

also carried out an experimental and numerical study about transition development in a separation bubble over elliptical leading edged flat plate. They showed that transition in LSB was driven by amplification of 2-D Tollmien/Schlichting waves and first stages of the 3-D disturbances played minor role in the transition. Moreover, the results showed that bidirectional vortexes lead to 3D breakdown. Brandt et al. [15–18] studied experimentally and numerically about effect of free-stream turbulence on the transition. The numerical results indicated important similarities with the flow structures in experimental studies on the secondary instability and breakdown of steady symmetric streaks. They concluded that the transition location moved to lower Reynolds numbers by increasing the integral length scale of the free-stream turbulence. Genç et al. [19–23] tested the behavior of the turbulence and transition models in the study of prediction of the LSB over the aerofoils at low and high Re numbers, and in the study of

Fig. 3. Manufacturing processes and aerofoil models manufactured with end-plate.

Reynolds number U U U U U

Fig. 4. Experimental set-up for pressure measurements over NACA2415 aerofoil. Fig. 5. Experimental set-up for the force measurements.

254 M.S. Genç et al./Experimental Thermal and Fluid Science 39 (2012) 252–264 controlling this LSB by using high lift, blowing and suction systems. The numerical results of the turbulence models indicated varying degrees of success in predicting the boundary layer flow field and the separation bubble, while the results of transition models accurately predicted the location and extent of the separation bubble in the single element aerofoil cases. In the control cases, it was predicted that the separation bubble was eliminated by using the slat, blowing and suction resulting in some marginal increase in the lift and decrease in drag.

The aim of this study is to evaluate aerodynamic performance of a NACA2415 aerofoil at the angles of attack from 12 to 20 at Re numbers of 0.5 105,1 105,2 105 and 3 105. Moreover, due to the fact that flow regime is low Re number flow, the LSB, transition and reattachment are important and their effects on aerodynamic performance is considered.

2. Experimental apparatus and methods 2.1. Wind tunnel and models

The experiments were carried out in a low-speed, suction-type wind tunnel with a square working section of 500 m 500 m located at the Department of Energy Systems Engineering, University of Erciyes. The ratio of cross sectional area of contraction cone was 9:1 and the side walls of the working section were expanded with a divergence angle of 0.3 on each side to minimize boundary layer effects on the working section walls, and to give a constant static pressure. Maximum speed near center of the working section was about 40 m/s, free-stream turbulence intensity at maximum speed was about 0.3%; about 0.7% at lowest speed (5 m/s) [24].

Fig. 6. Sample of oil-flow visualization experiment over NACA2415 aerofoil on which oil laid. Fig. 7. Constant temperature anemometry system and the probe over the NACA2415 aerofoil.

x/c

-C p

3.0 Experiment-Erciyes Un Experiment-Bath Un.[23, 29]

Suction surface Pressure surface α =8 º Re=200 0

Fig. 8. Comparison of C distributions over the NACA2415 aerofoil in present study (Erciyes University) and the study in University of Bath [23,29] (Re = 2 10 , a =8 ).

M.S. Genç et al./Experimental Thermal and Fluid Science 39 (2012) 252–264 255

The experiments were carried out at Reynolds (Re) numbers of 0.5 105,1 105,2 105 and 3 105 based on chord length of aerofoil (c) and free-stream velocity (U1). The experimental setup and schematic diagram are shown in Figs. 1 and 2.

The aerofoil models were manufactured out of glass fiber and two component epoxy resin covering up NACA2415 foam aerofoil using a steel mold (Fig. 3). Epoxy resin was mixed carefully with epoxy hardener at a certain rate (5:1) according to the manufacturers directions. After a thin mixed epoxy layer was applied on the NACA2415 foam which was prepared using a hot-wire and two wood aerofoils, glass fiber covered up NACA2415 foam. The remaining layers of glass fiber were laid onto the previous layers successively in a similar way. Later, NACA2415 foam covered up with glass fiber and epoxy resin located in the NACA2415 steel mold. The aerofoil in the mold was left to solidify for 8–12 h at room temperature. The edges of the solidified composite aerofoil was shaved to obtain clean plate edges and approximate the NACA2415 profile definition within an accuracy of 0.1 m. The composite plates were left at room temperature for three days to complete the curing period and to reach the maximum strength. Later, plexi-glass end plates were assembled to two side of aerofoil. The manufactured aerofoils have a span length of 290 m, and a chord length of c = 180 m.

The uncertainty of the measurement depends on the uncertainties of the calibration device, linearization, repeatability, the accuracy of reference calibration device, position of the probe, air density, ambient pressure, etc. The uncertainty of the velocity measurement using hot-wire anemometry and the pressure measurement using pressure transducer were determined as 4.0% and 4.6% for Re = 200,0, respectively. The uncertainty values in the pressure coefficient (UC ), lift force coefficient (UC ), drag force

Blockage ratio of NACA2415 at the angle of attack of 0 in our wind tunnel was 0.18 0.15 0.29/(0.5 0.5) = 0.031 ( 3%). Blockage ratio of NACA2415 at the angle of attack of 20 was 0.071( 7%). The blockage corrections are not made on the experimental results, because the blockage effects on the experimental results are negligible when the blockage ratio is less than 10% [25,26].

2.2. Pressure measurements

For the measurement of pressure distributions of suction and pressure surfaces on the NACA2415 aerofoil, a system including a pitot-static tube, a scanivalve unit, a pressure transducer and 24 pressure tappings of 0.8 m in diameter, which are flush along the mid-span of the upper and lower surfaces of the wing was used (Fig. 4). Pressure measurements were carried out by using a computer-controlled data acquisition system. The pressure was measured by using Honeywell 163PC01D75 model differential pressure transducers with a pressure range of 623 Pa. Calibration of the pressure transducer was made by using CEM DT-8920 manometer and Kimo TPL-03–300 pitot tube with an accuracy of x/c

-C p

Pressure surface Suction surface x/c

-C p

Suction surface Pressure surface

x/c

-C p

Pressure surface Suction surface x/c

-C p

Pressure surface Suction surface

(a) (b)

256 M.S. Genç et al./Experimental Thermal and Fluid Science 39 (2012) 252–264

1 Pa. The maximum response time of the pressure transducer was about 1 ms. The pressure signals were acquired at a sampling rate of 1000 samples per second and each of the mean pressures was obtained by means of averaging 16,384 data points for over 16.4 s. A software was written in C programming language to acquire signals in conjunction with a 16-bit A/D converter, and final post-processing was complemented in MATLAB to calculate the mean pressure distributions. Experiments were conducted over a range of angles of attack in order to calculate the pressure coefficient distribution.

(Parte 1 de 3)

Comentários