Winglet Performance

Design a winglet for a NACA 4412 airfoil and compare performance with and without it

* This page is a summary of work done in a lab group. A lot of credit is given to my lab members - my involvement is noted below.



  • Research
  • Design
  • CAD
  • 3D print
  • Wind Tunnel Testing
  • Analysis


The Winglet Challenge called for a wingtip device that increased the lift to drag ratio of a NACA 4412 wing section. The wing section provided has a chord length of approximately 15.6 cm, so the winglet design was scaled as such. Ease and simplicity of production was considered very important, so a low, gently sloping blended winglet was 3-D printed. At low speeds, wingtip vortices do not produce a large amount of drag. When the winglet was implemented, the L/D ratio was not significantly changed. This was because the winglet increased the wingspan of the wing increasing lift, but also increased the drag by increasing the surface area. The winglets designed were not effective on the wing designed in the scale presented.



  • Af = Frontal Area
  • At = Top Projected Area
  • α = Angle of Sting
  • AoA = Angle of Attack
  • Cd = Drag Coefficient
  • Cl = Lift Coefficient
  • Fd = Drag Force
  • Fl = Lift Force
  • L/D = Lift to Drag Ratio
  • Re = Reynolds number
  • ρ = Density of Fluid

I. Introduction

Winglets come in many shapes and sizes and can be found on many of today’s aircraft, from the small single-engine general aviation planes to the giant jetliners cruising across the sky. This simple yet sleek device serves one main purpose - reduce the amount of wingtip vortices to better flight performance and efficiency.

Wings produce lift through the differential of pressure across them, with a region of high pressure below and low pressure above. Due to Bernoulli’s principle, at the wingtips the air will move from the lower area of high pressure and around the tip to the above area of low pressure. This flow around the wingtip causes the air to rotate and form a vortex. This vortex causes a downward force onto the wing against the normal lifting force and can create significant amounts of induced drag. One way to deal with these negative effects is to increase the span of the wing, thereby reducing the amount of induced drag. However, constructing a higher aspect ratio wing will require more structural weight which may end up negating the effects of reducing drag. Also, and as is often the case with large airliners, the wingspan of the jet could not be made any larger due to the fact that it needs to fit in a gate at an airport. Instead, often the most practical way to reduce induced drag is by having a winglet.

By utilizing a winglet, the flow is obstructed from curling all the way around the wingtip. Instead of creating a large vortex at the wingtip, a smaller vortex is created at the tip of the winglet, causing less drag and less interference with the flow at the outer edge of the wing. In some cases, by moving the vortex to the tip of the winglet, this rotational flow can even provide a force component in the thrust direction, furthering the wing’s efficiency even more. Theoretically, while a winglet can minimize a vortex and move it away from the wingtip, it cannot completely eliminate it.

The Cal Poly low-speed wind tunnel is used to test the effects of fluid flow and various Reynolds numbers. In this lab, winglets were conceptualized and designed based on the NACA 4412 airfoil. They were constructed using additive manufacturing, and tested to determine if the design would improve the lift to drag ratio (L/D) of a given wing. Examining the lift and drag coefficients at different angles of attack and Reynold’s numbers help determine the effectiveness of the winglet. Smoke was added to further visualize the flow and easily see the vortex being created at the wingtip, with and without a winglet.

  Winglet Model

Winglet Model

  Early Winglet Designs

Early Winglet Designs

II. Methodology

A.  Winglet Design

The winglet was designed to maximize the efficiency of the wing, L/D. In order to do this, either lift needs to be increased, drag decreased, or a combination of both. Taking inspiration from the many winglets that have been carefully engineered and tested that are out flying around on aircraft today, many potential designs were examined. The split winglet design was considered, as seen on aircraft such as Boeing’s 737 MAX, however ultimately was not chosen to test due to the thought that there may be interference between the upward and downward tips that produces more drag. Another winglet that was considered was the more interesting looking “spiroid” winglet as seen on aircraft such as the Falcon 50. The spiroid was not chosen to be produced because of the difficulties in manufacturing that the complex design posed. Some initial design sketches (Fig.1) exhibit some of these features, but ultimately the design selected was as can be seen in Figure 2.

The design tested and produced was instead a single, upward-sloping blended winglet. This design was modeled to have this upward curve feature to reduce drag, as well as to mainly just simply increase the wing area, increasing lift. For the challenge, the winglet was limited to size in width (60mm) and height (150mm). The winglet produced made full use of the spanwise limit with the intention of increasing the wing area as much as possible, and subsequently increasing lift. The vertical dimension of the winglet was chosen purely by eye and was well under the limit, with the thought that more material extending vertically would simply cause more drag while not having an effect on the tip

  3-View Winglet Drawings

3-View Winglet Drawings

B. Manufacturing

Both winglets were manufactured with FDM (fused deposit modeling) 3D printing. With the advance in 3D printing technology coming into the consumer market, it was the ideal method for manufacturing winglets with a print time of only 4 hours and a tolerance of .003”. Due to the nature of 3D printing and the usage of 0.2mm layers, the surface of the winglets were not perfectly smooth. To overcome this potential source of error, we used a rough sandpaper and a fine sandpaper, 80 and 220 grit respectively. Though not critically important, we used a 25%  hexagonal infill for each winglet. This ensured a low print time while still maintaining proper layering and structure. While 3D printing allowed for very accurate fabrication of the winglets, we used aluminum foil tape to ensure connection to the wing when testing. The foil remained unintrusive in the testing, but increased security while in the wind tunnel.

  3D Printer and Winglet

3D Printer and Winglet

C. Experimental Setup

The NACA 4412 wing section was tested with and without the winglet in the wind tunnel at Cal Poly. In the test section, the wing was mounted to the Sting equipped with a load cell measuring the forces and moments in the x, y, and z direction. The load cell also measured torque in the x,y and z directions, but that data was unused in this experiment. The Sting is mounted on the ceiling of the wind tunnel, and the wing attached to it upside down in order to keep the wake of the wing from interfering with the sting (Fig.5). So, the measurements for lift are actually a downforce. The tunnel was run at one speed of 25 meters per second. This can be achieved with an RPM in the low speed wind tunnel of approximately 625. This was considered to be the maximum safe speed of the wind tunnel with the Sting and the wing in the tunnel. Even at this speed, an experimenter must always observe the wing, sting, and winglets too make sure that all remain secure.

  Wind Tunnel Sting Setup

Wind Tunnel Sting Setup

D. Test Matrix

In this lab, the test matrix was open ended. Two main variables were at play for being tested: angle of attack and wind speed/velocity. In the time given, only one variable was able to be tested with sufficient data points - the test matrix chosen (Table.1) was created in order to attain information about lift and drag for varying angles of attack. On top of comparing lift and drag for the wing with and without winglets, the matrix was created in order to find stall points. Both setups, wing without winglets and wing with winglets, were tested at the same wind speed each with the same set of angle of attacks. Anticipating a stall somewhere between 10 and 20 degrees, the step size of the test matrix was lowered to attain more data points near this separation point. 

E. Data Analysis

The data that the Sting produces that was used to calculate lift and drag. One hundred data points of force in the x- and z-direction were averaged at each angle of attack. As the AoA changes, the axes presented in Figure 5 undergo a frame rotation about the y-axis equal to the AoA. At 𝛼 = 0, the lift is equal to the force in the positive z-direction, and the drag is equal to the force in the negative x-direction. However, at varying angles of attack, the lift and drag follow Equations 1 and 2 respectively.

L     =      X sin(α)      +   Z cos(α)

D     =     -X cos(α)    +   Z sin(α) 


III. Results


IV. Discussion

The winglet did not perform its goal of increasing L/D at a significant value. At higher angles of attack, the lift was greater, and wingtip vortices increase as a function of lift. The vortices are created from the high pressure fluid below the wing flowing towards the low pressure area above the wing. Therefore, the change in lift to drag ratio by adding the winglets should be more visible at higher angles of attack where more lift is produced. Figure 8 however, shows no significant or consistent change in L/D with or without the Winglets.

Wingtip devices being more effective when more lift is produced is why they are usually only implemented on heavier aircraft that cruise at higher velocities. Heavy aircraft need to generate more lift, so wingtip vortices are more prominent on those aircraft. Our wingtip devices cannot be written off at all speeds of flow. At higher speeds, the winglets may be much more effective.

As can be seen in Figure 10, the wingtip devices were far from perfect. By utilizing a smoke machine, it was easy to see how the flow still curled around the winglet device. It is hard to see, but the some of the flow seemed to wrap around the leading edge of the winglet, and was not fully obstructed as designed. This is a decent visual representation of how wingtip vortices operate on an wing.

Objects in an external flow can affect fluid upstream. In concern to the experiment done in this lab, the Sting apparatus holding the winglet has a slight effect on the fluid flow over the winglet. Though this effect may be minimal, it may be a source of error in the lift and drag measurements. With the Sting disturbing the flow on the wing, performance characteristics such as the critical angle of attack may be affected. Due to the interference, turbulence may be induced, leading to an earlier stall. Fortunately, as can be seen in Figure 9, the effect of the Sting apparatus remains constant between the tests, and can be easily subtracted off of the raw force data.

  Flow Visualization over winglet

Flow Visualization over winglet

  Vortex Generated at Wingtip Viewed from Wind Tunnel Intake

Vortex Generated at Wingtip Viewed from Wind Tunnel Intake

V. Conclusion


Winglets are a very effective and simple device to minimize wingtip vortices and reduce induced drag. Although the winglet tested in this lab did not seem to indicate this, there were many other factors in play that could have affected the results. In most cases, winglets are most effective in high speed cruise conditions, which were not present during these tests. Although winglets are minimally important at low speeds, in high-lift producing scenarios, their inclusion can noticeably increase the L/D ratio of a wing. A simple and relatively easy to produce winglet, such as the one that was designed and tested in the lab could significantly increase the performance of a wing section, increasing the overall efficiency.



My lab group

Dr. Doig

Mr. Baldovin