The project was split into four main stages:
1. Airfoil Selection
2. Ideal Angle of Attack
3. Endplate Design
4. Placement on the Vehicle
1. AIRFOIL DESIGN
To easily identify Cl/Dl (coefficient of lift / coefficient of drag) ratios for various airfoils two databases were used. These were airfoil tools and the University of Illinois' Aerodynamic Institute's database. Both are linked above.
From the Airfoil Tools database, airfoils can be compared side by side to make performance decisions. The website uses XFoil software to estimate the performance of subsonic airfoils. XFoil has been found by Kallstrom(2022) to be accurate for testing at under 0.4 Mach but can over predict lift and under predict drag. These concerns fall outside of the scope of the project as our main concern is to compare multiple airfoils for performance relative to one another.
Comparison of NACA6412 and S1223-rtl Airfoils
The process used for comparing airfoils is demonstrated below using the two finalists, the S1223-rtl (top), and NACA6412 (bottom).
Several simulation values have to be considered prior to comparing airfoils. First is the Reynolds number of the system. The Reynolds number measures the ratio of inertial to viscous forces in a system, and is defined as:
Re = pvl/μ = vl/v
v = velocity of the fluid
p = density of the fluid
l = characteristics length ( chord length)
μ = dynamic viscosity of the fluid
v = kinematic viscosity of the fluid
For our purposes, we will look at simulations run with an Re of 500, 000. Our target velocity is approximately 36.6m/s, chord width is 0.21m and the SI value for the kinematic viscosity of air is 1.48E-5. This gives an Re of approximately 540,000 however the website runs simulations at 100K, 500K, and 1Mil, so 500,000 makes a reasonable approximation. The data below shows the NACA6412 (red) compared to the S1223 (green).
Based on the data above, the NACA wing makes less lift under all conditions when compared to the S1223 airfoil. However, it exhibits higher peak efficiency with a Cl/Cd of approximately 135 compared to 110. Based on this, we know that the S1223 airfoil makes more lift but substantially more drag.
The data was discussed with the vehicle owner and we decided to use the S1223 airfoil. The main focus is maximising downforce and drag on the car can be more effectively reduced in other areas and is less of a concern.
2. Finding Ideal Angle of Attack
Based on the data from Airfoil Tools, the ideal angle of attack (AOT) for the STL airfoil (in terms of L/D ratio) is just about 5°. This is for a single element airfoil, and doesn't suit our purposes. In order to find the ideal AOT, SolidWorks CFD was used to determine drag, lift, and flow over the airfoil. Tests were run with the main element (bottom) at 0°, and the top element from 5 - 40°. After finding the ideal AOT for the top element, the main element was adjusted and simulated from 0-15° .The same end plate was used for all tests (more on the importance of endplates later). The ideal AOT for each element is shown below.
When analysing different AOT's, three main factors were investigated.
1. Total Downforce
2. Total Drag
3. Flow over the airfoil
When looking at flow over the airfoil I was aiming to maximise angle of attack without inducing stall in the wing. Stall is a condition wherein the separation point of the airflow begins close to the wingtip. A large space of separated airflow results in a large turbulent zone, drastically increasing drag.
Flow simulations were run and visualised to investigate separation points, maximum lift, and maximum drag. The data for the final design is shown below.
From the flow simulations, areas of turbulence and separation can be identified. As shown in the simulation, areas of separated flow can be identified by the swirling lines. When testing the different AOTs, the goal keep the separation towards the trailing edge, as separated flows close to the leading edge indicate stall. Greater emphasis was put on the data shown on the left. This demonstrates average downforce (GG force Y in Newtons) and drag (GG Force Z in Newtons). This data is easier to interpret, especially to this inexperienced, but is equally limiting at the same time, providing only a partial picture.
3. Endplate Design
Endplates are incredibly important. On the topside of the wing, we have created a high pressure zone (denoted by red lines in the simulations above), and on the bottom we have a low pressure zone. Without endplates, air from the top side would spill over the sides of the airfoil to amend the pressure differential. To keep the pressure zones separate (and create downforce), we use endplates. This is especially important on a short and narrow airfoil like those found on cars.
For my design, simplicity was key. To keep costs downwind make manufacture simpler, I chose to minimise the size of the endplates. Had I had more time, money, and experience in manufacture, I would have gone with a much larger design as generally the larger the endplate the greater the downforce (within reason obviously). Instead, I had to be careful with my design.
As a general rule, most endplates have most of their surface area on the bottom side of the airfoil. This served to further isolate the wing's low pressure region, which is pertinent to downforce generation. It is usually easier to maintain a low pressure region on a wing versus a high pressure one due to the differences in energy. Higher energy regions are more susceptible to air flow separation. Because of this, putting more effort into maintaining a low pressure region will be more cost effective for our purposes.
The data on the left shows the drag and downforce relationships for the same wing with the final two endplate variations. Obviously, version 9 outperforms version 15 in all metrics.
Version 9 is shown on the right, and version 15 on the left. It might be a bit hard to see, but version 9 clearly shows distinctly higher pressure on the upper inside of the endplate. Concurrently, it held lower pressure on the bottom inside of the endplate. This is likely a major contributing factor to its superior downforce performance. Furthermore, the design seems to generate vortices further away from the trailing edge of the wing, potentially mitigating its influence on airflow leaving the wings surface.
Again, it may be a bit hard to see but the endplates on wing version 9 (left this time) seems to have a considerably reduced impact on airflow over the wing. Version 15's endplates seem to draw a greater volume of air into their vortices. This likely explains both the reduction in drag and increase in downforce.
As stated above, I recognise that I am a bit of a novice in aerodynamics. Because of this, greater emphasis was placed on quantitative data versus the flow analysis.
4. Placement on Vehicle
Due to technological limitations, this is likely the shortest section. Despite having access to a CAD model of the EK civic, I unfortunately don't have the computing power to run simulations. Nonetheless, some basic aerodynamic principles can still be used to identify a good mounting location and method prior to real-world testing.
1. The wing was designed to operate in clean air
2. Clean air is located away from the body of the car
3. It is easier to maintain low pressure than high pressure regions
With these rules in mind, it was decided that we would mount the wing with swan necks (mounts on the upper face of the wing) 50mm above the roof line and 100mm behind the rear wheels. The mounting location on the car will be directly in line with the rear wheels. If the model can be simulated prior to construction, I will update this page. A rough mock-up can be found below.
