Effect of trailing edge serration on the lift and drag characteristics of NACA0012 airfoil wing

AIAA AVIATION Forum 5-9 June 2017, Denver, Colorado 35th AIAA Applied Aerodynamics Conference

10.2514/6.2017-4470

Effect of trailing edge serration on the lift and drag characteristics of NACA0012 airfoil wing Usama Hussain1, Saif Ul Malook1, Burhan Shabir1, and Ozaif Ali1

Downloaded by TU DELFT on March 20, 2019 | http://arc.aiaa.org | DOI: 10.2514/6.2017-4470

Ghulam Ishaq Khan Institute of Engineering Sciences and Technology, Topi, KPK, 23460, Pakistan

A Computational and experimental study of NACA0012 wing section with serrated trailing edge is presented. Two types of serrations triangular, and serrations with quadratic curvature added to the triangular edges were investigated for their influence primarily on Lift to Drag ratio of the wing. Both computations and experimentation were conducted at constant chord Reynolds number of 360,000. A two equation turbulence model, k-ω, was used to solve Reynolds Averaged-Navier-Stokes equations. Different dimensions of serrations were first tested in the simulated environment of ANSYS CFX, and then optimized serrated designs were manufactured to validate the results by Wind tunnel testing.

Nomenclature Aspect ratio Bird type serration CD CL D Deg Ρ Μ H W L Tri y

= = = = = = =

Length of serration parallel to chord / Length of serration parallel to trailing edge

= = = = = =

viscosity of air, kg/(ms)

Quadratic spline serration coefficient of drag coefficient of lift depth of serration perpendicular to the span and chord, inch Degree density of air, kg/m3 height of serration parallel to chord, inch width of serration perpendicular to chord and parallel to trailing edge, inch length, inch Triangular Chord wise distance from the tip of serration

I. Introduction Recently, modifications to the edges of the wings have been under scrutiny in order to improve the wing’s aerodynamic and acoustic parameters. Previous studies have shown that airfoils with truncated trailing edges could give higher coefficient of lift, at constant chord Reynolds number, as compared to conventional airfoils, but the added bluntness due to truncation will now increase the coefficient of drag as well [1]. Higher coefficient of drag can be attributed to the Von Karman type vortex shedding in the wake region of the wing. Reducing the intensity of these Von Karman vortices, by using serrated wing edges, have proved effective in keeping the coefficient of drag checked, and thus have improved lift to drag ratio. Tanner [2], for example, employed serrated trailing edges onto blunt trailing edge and achieved up to 65% drag reduction as compared to blunt trailing edge. Similarly Krentel & Nitsche [3], found drag reduction of up to 30% by employing square wave and stepped trailing edges on NACA0012 wing section. Further study on the subject by Nedic and Vassilicos [4], was focused on the fractal patterns at the trailing edge of wings. They were able to speculate and correlate the chevron angle of fractal patterns with the intensity of vortex shedding which is primarily responsible for higher coefficient of drag. Their research have concluded that an increase in iteration of fractal patterns could mollify the intensity of vortex shedding for small chevron angles, and hence would ameliorate the lift to drag ratio. 1

Student, Department of Mechanical Engineering, GIKI, KPK, Pakistan. 1 American Institute of Aeronautics and Astronautics

Copyright © 2017 by Usama Hussain, Saif Ul Malook, Burhan Shabir, Ozaif Ali. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.

Invigorated by the wings of an owl some researchers also examined the aerodynamic effects of serrations on the leading edge of a wing instead of that on the trailing edge. Winzen et el [5], for example, used Particle image velocimetry to study owl-based wing models with leading edge serrations. Their results showed a reduction in lift to drag ratio, owing to the higher coefficient of drag. However, they found a uniformity in the size of vortex structures shed, especially at low Reynolds numbers. This uniformity has a stabilizing effect on the flight allowing better maneuverability at the cost of adverse aerodynamic performance. Similarly, Hensen et el [6], used a planer sinusoidal leading edge serration on a two dimensional NACA65-021 and NACA0021 wing sections, and found an increase in the lift coefficient in the post stall region. This increase in the coefficient of lift in the post stall region also enhanced the maneuverability of the flight. So to sum up, serrations have an immense influence on the aerodynamic characteristics of a wing. The aim of this study is to examine the effects of not just triangular serrations, like in previous studies, but also of serrations with curvature added to their edges, named as bird type serrations for convenience. In this research efforts have been directed to prove the hypothesis that adding curvature to the triangular serration will enhance drag reduction, by virtue of shrinking the bluntness at the trailing edge, and hence will improve lift to drag ratio.


II. Methodology A. CAD Modeling The 3D wings of NACA0012 airfoils were modeled in SOLIDWORKS 2015 from the 2D x-y geometry coordinates, which in turn were extruded into the z axis. The serrations were made by cutting straight wedges (triangular prisms), and quadratic splines off from the original wing and applying a pattern of the cuts along the trailing edge. All the wings spanned 24” with a mean chord length of 6”. Mean chord lengths and hence the planform areas, for all the wings under consideration, were kept constant in order to isolate and study only the effect of serrations on the aerodynamic parameters. Two different types of serrations, namely triangular and bird type (quadratic spline), were tested both in simulated and experimental environments and the results were compared to that of a conventional NACA0012 wing section. Different dimensions of each type of serration were first modeled and tested in simulated environment and then optimized serrations were later manufactured for wind tunnel testing to corroborate the results. Table 1 shows the dimensions of serrations, of both types, which were modeled for computational analysis. Table 1. Dimensions of serrations modeled for Computational analysis Triangular

Width

Depth

Aspect

Bird-Feather

serrations

(inches)

(inches)

Ratio

Serrations

(W x H) Tri 0.5x0.5

Width (inches)

Depth

Aspect

(inches)

Ratio

(W x H) 0.5

0.5

1

Bird 0.5x0.5

0.5

0.5

1

Tri 0.5x0.75

0.5

0.75

1.5

Bird 0.5x0.75

0.5

0.75

1.5

Tri 0.5x1

0.5

1

2

Bird 0.5x1

0.5

1

2

Tri 0.5x1.25

0.5

1.25

2.5

Bird 0.5x1.25

0.5

1.25

2.5

Tri 1x0.5

1

0.5

0.5

Bird 1x0.5

1

0.5

0.5

Tri 1x0.75

1

0.75

0.75

Bird 1x0.75

1

0.75

0.75

Tri 1x1

1

1

1

Bird 1x1

1

1

1

Tri 1x1.25

1

1.25

1.25

Bird 1x1.125

1

1.25

1.25

Tri 2x0.5

2

0.5

0.25

2

0.5

0.25

Tri 2x0.75

2

0.75

0.375

Bird 2x0.75

2

0.75

0.375

Bird 2x1

2

1

0.5

Bird 2x1.25

2

1.25

0.625

4

0.5

0.125

Bird 4x0.75

4

0.75

0.1875

Bird 2x0.5

Tri 2x1

2

1

0.5

Tri 2x1.25

2

1.25

0.625

Tri 4x0.5

4

0.5

0.125

Tri 4x0.75

4

0.75

0.1875

Tri 4x1

4

1

0.25

Bird 4x1

4

1

0.25

Tri 4x1.25

4

1.25

0.3125

Bird 4x1.25

4

1.25

0.3125

Bird 4x0.5

2 American Institute of Aeronautics and Astronautics


Figure 1 & Fig.2 below shows triangular serrations whereas Fig.3 & Fig.4 shows bird type (Quadratic spline) serrations on NACA0012 wing section respectively. These serrations were made as per the dimensions given in Table 1.

Figure 1. Triangular serrations. Top view

Figure 2. Triangular serrations. Front view

Figure 3. Quadratic spline serration. Top view

Figure 4. Quadratic spline serration. Front view

B. CFD Computational analysis were first conducted to establish an optimized configuration of serration. Both Triangular and Quadratic spline serrations were tested in the simulated environment of ANSYS CFX. All simulations were performed at fluid velocity of 33.33 ms-1 as that not only fell within the subsonic range, it also yielded a suitable chord Reynolds number of 360,000 and when tested in the wind-tunnel, gave the most stable and reliable results for experimental comparison. Lift and Drag values were calculated for all angles of attack between 0 degree and 20 degree with a step change of 1 degree. To model the wing a cuboid domain was established with the height and length approximately 100 times the chord length. The width was kept as 0.6096 m to assign a symmetry boundary layer which modeled the wing as one with an infinite span. The solid wing was removed from the fluid domain using a Boolean operation and was replaced by a no-slip wall. ANSYS meshing was used to generate unstructured meshes and various techniques like face sizing, inflation layers, virtual topologies and pinching were employed to achieve a mesh structure that was acceptable and allowed for the fastest possible computation without compromising the quality of the data acquired. Considering all the above mentioned issues, a mesh was created with the mesh statistics given in Table 2. The free stream temperature was set at 300 K, which is the same as the ambient temperature. The density of the air at the given temperature is ρ = 1.225kg/m3 and the viscosity is μ=1.7894×10-5 kg/(ms). To solve Reynolds Averaged-Navier-Stokes equations (RANS equations) a two equation turbulence model known as Wilcox k-ω model [7] was used in CFX. It is a density based solver which facilitates mimicking compressible flow over the body under experimentation very accurately. To simulate far field conditions, we have assigned opening boundary conditions above and below the wing by using the same velocity expressions as present within the inlet conditions, and have used a zero gradient condition to ensure uniformity of flow from inlet and openings. Moreover, our accuracy levels for residuals were set at 1 x 10 -4 to ensure both a decent level of accuracy and quick calculations, since we were dealing with approximately 640 distinct calculations so a time constraint was placed upon us.



Table 2. Mesh statistics Mesh Statistics Nodes (average) Elements (average) Orthogonal Quality Average Standard Deviation Max Min Skewness Average Standard Deviation Min Max Aspect Ratio Average Standard Deviation Min Max Element Quality Average Standard Deviation Min Max

404014 2190510 0.864 0.086 0.999 0.187 0.231 0.123 0.0003 0.962 1.852 0.474 1.165 12.74 0.845 0.097 0.145 0.999

C. Experimental Setup Wind tunnel available for experimentation was of ELD INC, Model 406 (B). It is a subsonic and open circuit wind tunnel with a test section of dimensions (Height x Width x Length) 24”x24”x48”. The manufactured NACA0012 wings, of chord length 6” and span 24”, were mounted horizontally in the test section by mounting plates in the vertical windows of the wind tunnel. The wings once fixed to the mounting plates could be rotated along with it to set angle of attack as desired. A dynamometer was calibrated with the available precise weights, as per user’s manual, and then was used to measure lift and drag on the wing sections. After thorough analysis of simulated results, one wing with three detachable serrations was manufactured for testing in the wind tunnel to serve the purpose of final verification of results. The two triangular serrations were of dimensions (Width x Depth) 1” x 0.5” and 1” x 1”. However the quadratic spline serration manufactured had dimensions (Width x Depth) 1” x 0.5”. These serrations can be fitted to the wing through screws embedded in the surface of the wing. The wing was manufactured with CNC machine under high precision to ensure a good structural integrity once assembled with detachable serrations. Figure 5 shows the CAD model and pictures of wing and its serrations.

a)

CAD of front part of the wing



b) CAD of Triangular trailing edge serration

c) CAD of Quadratic spline trailing edge serration

c) Actual Triangular trailing edge serration

d) Actual Quadratic spline trailing edge serration

e) Actual picture of assembly of manufactured wing and trailing edge serration

Figure 5. CAD model and pictures of wing and serrations


Coefficient of Lift

2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

CL-Simple CL Tri 0.5x1 CL Tri 1x0.5 CL Tri 1x0.75 CL Tri 1x1 CL Tri 1x1.25 CL Tri 2x1 CL Tri 2x1.25 0

5

10 15 20 Angle of Attack, Degree

25

CL Tri 4x1 CL Tri 4x1.25

Figure 6. Coefficient of Lift vs. Angle of Attack for various Triangular serrations in comparison with that of a simple wing 30 25

Lift/Drag Ratio


III. Results Results from the CFD simulations show that there is direct increase in Coefficient of Lift (C L) caused by introducing triangular trailing edge serrations. Secondly there is a direct correlation between increasing the depth of serration in a triangular shape and a depreciation of its Lift/Drag ratio for a constant width. Graphs in Fig.6 and Fig.7 has been developed from the Lift and Drag values acquired from the CFD-Post module in ANSYS CFX. The values were tabulated in Microsoft Excel and ratios calculated for Lift/Drag. Figure 6 below represents a comparison between the lift performance of a simple wing and Triangular serrated wings of various sizes. It is evident that any dimensions of triangular serration will at the very least have a higher maximum coefficient of Lift, due to added bluntness at the crests of serrations. Figure 7 however represents a comparison of Lift/Drag ratio between two triangular serrations, both with a serration width of 0.5 inch. The overall curve for the serration with the lower depth (1 inch) is greater than the curve for the serration with greater depth (1.25 inch). Both of these curves are held against the Lift/Drag curve for the simple un-serrated wing as a comparison which reflects the fact that addition of serrations can have both a negative and a positive impact on the performance of a wing, depending on the dimensions of serration. Furthermore, there is an overall forward tilt in the curve which represents the fact that the apex of performance is now achieved at a greater angle of attack than before. This can have a positive impact on the maneuverability of an aircraft as it can now perform sharper changes in direction without stalling.

20 15

CL/CD-simple

10

CL/CD Tri 0.5x1 CL/CD tri 0.5x1.25

5 0 0

5

10

15

20

25

Angle of Attack, Degree Figure 7. Lift/Drag comparison for triangular serrated wings of constant base width of 0.5 in and variable depth A similar comparison is shown in Fig.8 for a base width of 1 in, if the curve with an anomalous reading (it is suspected that it is due to a specific simulation not converging to a desirable degree) is ignored, our correlation 6 American Institute of Aeronautics and Astronautics

statement that the peak of, Lift/Drag vs. angle of attack, curve decreases with increasing depth of the triangular serration holds. Furthermore, the velocity contours in Fig.9, left to right, show how the region affected by vortex generation grows larger with increasing depth of serration. The contours are compared at the angle of attack of 10 degrees with Reynolds number of 360,000, to offer a comparison in the maximum performance range. Increase in bluntness due to increase in the depth of serration is the sole reason for the increase in drag, and hence the lower lift to drag peak in Fig.7 and Fig.8. This blunt region can be seen clearly in Fig.10. Adding triangular serration to a truncated wing have minimized blunt regions along the trailing edge. However the deeper the serration the greater the blunt region. So it has been established that serrations not only increase the lift coefficient of a conventional wing, but it also adds bluntness at the trailing edge which is responsible for lower Lift to Drag ratio. 35


Lift/Drag Ratio

30 25

CL/CD-simple

20

CD/CD Tri 1x0.5

15

CL/CD Tri 1x0.75

10

CL/CD Tri 1x1

5

CL/CD Tri 1x1.25

0 0

5

10

15

20

25

Angle of attack, Degree Figure 8. Lift/Drag comparison for triangular serrated wings of constant base width of 1 in and variable depth

Figure 9. Velocity contour for triangular serration with a constant width of 1 in and depth increasing from left to right

Figure 10. Blunt region created by serration at the trailing edge shaded in yellow To increase lift to drag ratio a bird type serration, different from triangular shape, was also investigated for its aerodynamic parameters. This serration as shown in Fig.3 and Fig.4 has been obtained by adding quadratic curvature to the triangular serrated edges. As can be seen in Fig.2 and Fig.4, and explained through Eq.1 and Eq.2 that the blunt thickness for this new bird type serration has been reduced as compared to that of triangular serration. 7 American Institute of Aeronautics and Astronautics

The difference between Eq.1 and Eq.2 is positive for 0 < y < H, hence the blunt thickness for bird type serration will be smaller. 𝐷 𝐵𝑙𝑢𝑛𝑡 𝑇ℎ𝑖𝑐𝑘𝑛𝑒𝑠𝑠 (𝑇𝑟𝑖𝑎𝑛𝑔𝑢𝑙𝑎𝑟) = ( )𝑦 𝐻

(1)

𝐷 2 )𝑦 𝐻2

(2)

𝐵𝑙𝑢𝑛𝑡 𝑇ℎ𝑖𝑐𝑘𝑛𝑒𝑠𝑠 (𝐵𝑖𝑟𝑑 𝑡𝑦𝑝𝑒) = (

35

30

30

25

25

20

Lift/Drag Ratio

Lift/Drag Ratio

35

CL/CD Tri 0.5x1

15

CL/CD Bird 0.5x1

10 5

20

CD/CD Tri 1x0.5

15

CL/CD Bird 1x0.5

10 5

0

0

0

10

20

30

0

Angle of Attack, Degree

20

30

b)

35

30

30

25

25

20

Lift/Drag Ratio

35

CL/CD Tri 2x1

15

CL/CD Bird 2x1

10 5

20

CL/CD tri 4x0.5

15

CL/CD Bird 4x0.5

10 5

0

0 0

10

20

30

0


c)

10


a)

Lift/Drag Ratio


CFD results for these serrations are plotted in the same way as that of triangular serrations as mentioned previously. Figure 11 are comparative graphs of Triangular serrations and the bird type serrations on NACA0012 wing section. It can be concluded from these graphs that the bird type serration of the same scale as triangular has a higher peak of Lift to Drag curve, as per the previously established reasoning. Reduction in blunt region for bird type serration has reduced the wing area influenced by vortex shedding which can be seen in Fig.12. This reduction in the trailing edge vortices is responsible for increment in the maximum Lift/Drag ratio which is an immaculate improvement brought by adding quadratic curvature to triangular edges of serrations.

10

20


d)

Figure 11. Lift to Drag ratio vs. Angle of attack for Triangular and bird type serration


30

Results of CFD were finally validated by the wind tunnel testing of serrated wings of both types. Both experimental and computational results are in good agreement till a certain level of accuracy as can be seen in Fig.13. Small discrepancies are evident due to several experimental constraints, like boundary layer formation along the sides of wind tunnel, minute gap between the side wall of wind tunnel and the wing, small vibrations induced by the fan of wind tunnel, and fluctuations in the readings of dynamometer, etc. which were not taken into account in the computational models. But the experimental results, despite some small discrepancies, corroborates our computational solutions and confirm the correlations and trends established previously. 40 35

Lift/Drag Ratio


a) Velocity contours of triangular serration b) Velocity contours of bird type serration Figure 12. Velocity contours of Triangular, left, and Bird type, right, serrations. Significant reduction in the vortex generation can be seen for the bird type serration. This is due to the reduction in the blunt region at the trailing edge which has improved Lift to Drag ratio for this wing.

30 25 20

CL/CD Tri 1x0.5

15

CL/CD Tri 1x1

10

CL/CD Bird 1x0.5

5 0 0

5

10

15

20

25

Angle of attack, Degree Figure 13. Lift to Drag Ratio vs. angle of attack curve as obtained by wind tunnel experimentation.

IV. Conclusion The results acquired from the CFD simulations are dependent upon the quality of mesh and the accuracy of our setup. Given that limited time and computational power was at the authors’ disposal, the best possible mesh could not be used. However, certain iterations were repeated with higher quality mesh and the results were either confirmed or improved upon, providing enough reason to claim that these results can hold up to scrutiny. However, further scope for investigation is obviously present as all these results are based on a single Reynold’s Number and limited by only a single kind of airfoil design. Repeating these tests over a broad range of Reynold’s numbers and employing a broad spectrum of airfoils could lead to a better understanding of the effects of serration on aerodynamics of a wing. Results from CFD and wind tunnel experimentation of subsonic flows at low Reynolds number can be categorized into the following major research outputs:  There is direct increase in the Coefficient of Lift, CL, caused by introducing trailing edge serrations 9 American Institute of Aeronautics and Astronautics




Introducing serrations also increases the Coefficient of Drag, CD, due to vortex formation in the wake of blunt regions formed by serrated trailing edges  There is a direct correlation between increasing the depth of serration in a triangular shape and a depreciation of its Lift/Drag ratio for a constant width  For the same width and depth, serrations formed by adding quadratic curvature to triangular serrations have an inherent advantage of giving higher Lift/Drag ratio as compared to both conventional and triangular serrated wings As a concluding remark, it is important to discuss what the results of this research could yield. The results of this research could make a difference both in the aeronautical industry and in the wind energy sector, because serrations tend to enhance both aerodynamic and aero acoustic parameters. Improvement in these parameters could be revolutionary for reconnaissance purposes, both in the air and in water. Reduction in aero acoustic noise, if achieved by subsequent researchers, from wind turbines would allow the wind farms to be located substantially closer to residential areas hence, minimize the cost of electricity transmission. The “Bird Feather” serration could mean quieter, faster helicopters, UAVs, turbine blades that have a higher efficiency and an almost universal advantage, especially in the world of additive manufacturing where the complexity of shapes is no longer a major concern in the market. 3D printing can allow for such small features like serrations to be manufactured accurately and be stronger than a product made from traditional methods. Even in areas where manufacturing constraints are present, researchers around the world are working on developing methods of producing serrated wings using the same methods that developed conventional wings. Therefore, extensive study of the effects of serrations on aerodynamic and acoustic parameters is the need of this era which our research intends to serve.

References 1Smith,

H. A. and Schaefer, R. F., “Aerodynamic Characteristics at Reynolds Numbers of 3.0 x 10 (6) and 6.0 x 10 (6) of Three Airfoil Sections Formed by Cutting Off Various Amounts From the Rear Portion of the NACA 0012 Airfoil Section,” Tech. rep., NACA TN-2074, 1950. 2Tanner, M., “A method for reducing the base drag of wings with blunt trailing edge,” Aeronaut Quart, Vol. 23, 1972, pp. 15–23. 3Krentel, D. and Nitsche, W., “Investigation of the near and far wake of a bluff airfoil model with trailing edge modifications using time-resolved particle image velocimetry,” Experiments in Fluids, Vol. 54, No. 7, 2013. 4J. Nedić and J. C. Vassilicos, "Vortex Shedding and Aerodynamic Performance of Airfoil with Multiscale Trailing-Edge Modifications", AIAA Journal, Vol. 53, No. 11, 2015, pp. 3240-3250. 5Andrea Winzen, Benedikt Roidl, Stephan Klän, Michael Klaas, Wolfgang Schröder, “Particle-Image Velocimetry and Force Measurements of Leading-Edge Serrations on Owl-Based Wing Models”, Journal of Bionic Engineering, Vol. 11, No. 3, 2014, pp. 423-438 6Hansen, K. L., Kelso, R. M., and Dally, B. B., “Performance Variations of Leading-Edge Tubercles for Distinct Airfoil Profiles,” AIAA Journal, Vol. 49, No. 1, 2011, pp. 185–194. 7Wilcox, D. C., “Formulation of the k–ω Turbulence Model Revisited”, AIAA Journal, Vol. 46, No. 11, 2008, pp. 2823– 2838.


Effect of trailing edge serration on the lift and drag characteristics of NACA0012 airfoil wing

Recommend Documents