Author Archives: paulwlucas

Updates

Hancha Group now operates under Velox Motorsports. We have released a few products and many more are close to being released. Come check it out!

Facebook Page: https://www.facebook.com/veloxmotorsports?ref=aymt_homepage_panel

Website: http://velox-motorsports.com/

Blog: http://velox-motorsports.com/blog/

Advertisements

Final Information on Team Pure Precision GTi

This work has been completed a while ago, apologies for just now sharing the work. The data will not be posted since it is proprietary to the team, however I can share some information and some of the post-processing work.

At first, we had to create a car model for the cfd work. We started off with just a plain GTi model and started work on making it resemble their current race car. I changed out the front bumper and added fender flares. I also modeled up the rear wing and located it on the car. The exact profile of the airfoil is proprietary so that cannot be shared, sorry. The wheels were then modeled up and placed on the car setting the ride height.

splitter_reference

 

The car model was then analysed using CFD. Multiple runs were completed to capture the aero balance (with no splitter) at different rear wing angles. Different splitter designs were then analysed to meet the requested aero balance at a certain rear wing angle. Multiple iterations of different splitter designs were made and some tweaking was done on the final design. Then the aero balance and total downforce could be changed by adjusting the rear wing angle. This is not optimum, but it was a good choice for this type of race car.

Below is one of the options tested.It is a splitter with side plates. You can see the pressure plot on the car (red is high pressure, blue is low pressure) and velocity streamlines (red is high velocity, blue is low velocity).

hancha_003_w_h streamline_004_h

hancha_010_w_h

 

The last picture is of the final design. This design was chosen since it met the criteria for front downforce to achieve the aero balance requested and had the least amount of drag. The drag was decreased on this design since the entire front wheels are covered. This design is not perfect (what is?) and could be improved. Basically time ran out for improvements since the car had to start track testing.

gti_current_002_w_hancha

 

Below is the car in its final form.

901821_580234605327550_1268998885_o

 

-Paul

Wind Tunnel Testing

Wind tunnel testing is another experiment, like CFD, to help validate aerodynamic changes.  The reason for writing this is to help clarify some of the misconceptions on wind tunnel testing. First, lets start with a little history of wind tunnels…because everybody loves history.

‘Whirling Arm’ apparatus was the first wind tunnel if it could be considered one. This was in the mid 1700’s to 1800’s. Francis Wenham was not happy with his experiences with the whirling arm, so he, with the British Aeronautical Society built the first wind tunnel in 1871. Osborne Reynolds (Reynolds number, ratio of inertial forces to viscous forces) showed a scale model could exhibit the same flow pattern as a full model in 1900. World War II led the rapid wind tunnel and aerodynamic advances. The United States had a 400 mph wind tunnel and Germany had three supersonic wind tunnels capable of Mach 4.4 by the end of the war. The rest is history…

One big misconceptions of wind tunnels are they reproduce what the car will see on the road or track. This is false because wind tunnels only simulates the conditions on the road or track. Simulations inherently deviate from reality and it is often are hard to quantify all the sources of the error.

What do wind tunnels idealize? 

-temperature change, wind boundary layer, wakes of other vehicle, ect.

Sources of wind tunnel inaccuracy:

1. Rolling road system

2. Wall Effects – tunnel sections

Different wall test sections

3. Tires – shape from deformation

4. Cornering Conditions – steer and yaw

Wind tunnel simulation vs. real world track situation

5. Sting interference – the sting holds the model in place (also can be on the wheels)

Sting located on center of car and on wheels

6. Wheel lift – wheel lift is not usually measured in a tunnel.

7. Model is not moving – aka relative motion isn’t the same

Wind tunnel data is precise but saddled with inaccuracy of simulation; road data is free from they inaccuracy but are not precisely measured. Anyone who will quote you extremely accurate numbers are most likely trying to sell you something. This is obvious by just reading above.

That is it for tonight on wind tunnels, but next will be information on scaled wind tunnel testing. That will get into Reynolds Number effects and Reynolds Scaling.

-Paul

 

Single Element Wing – Update

Single Element Rear Wing

The design is complete on the single element wing and we are currently getting quotes for the manufacturing.  The endplates have been optimized and all data collected. Below is the data collected from the 15 CFD runs at different angles of attack.

Type: Incompressible steady-state

Turbulence model: k-omega SST

Velocity: 100 mph

data

h-lse-001_alpha13_1

h-lse-001_alpha13_2

We should have a prototype soon! I am very excited to get this on a BRZ / FRS

-Paul

Splitter or Air Dam – Which Design is Best?

How do you choose which front end aerodynamic route you need to go? This is a tough question to answer so lets work through an example.  The example will be a 1990-1997 Mazda Miata. The Miata was chosen simply because 1. I have the model and 2. because different designs are commonly used.

CFD Models

1. Stock 1990-1997 Mazda Miata
2. Stock 1990-1997 Mazda Miata at a 4in Ride Height
3. Small Front Air Dam at 4in Ride Height
4. Small Air Dam with Splitter at 4 in Ride Height
5. Large Air Dam at 4in Ride Height
6. Large Air Dam with Splitter at 4in Ride Height

Note: The air dam and/or splitter is 2 inches off the ground in study 3-6

The solver used for these analysis is a steady state incompressible solver with a k-omega SST turbulence model. Again OpenFOAM (as always) was used for pre-processing and solving and all post-processing was done using Paraview.

Cd = coefficient of drag
Cl = coefficient of lift
L/D = lift divided by drag / aerodynamic efficiency

Drag between all setups are all fairly close with the least drag being the setup with the large front splitter. The two splitter designs also makes significantly more downforce than the other setups. The two stock Miata setups makes lift instead of downforce which is expected since most road vehicles create lift from the factory.

Note: These are numbers and trends for common designs choices. Actual designs should be more refined after design goal is formulated.

The stock Miata simulation had a calculated coefficient of drag of 0.36. The 1990-1997 Mazda Miata had an indicated coefficient of drag of 0.38. The simulated Miata has a lower coefficient of drag; which was expected from the simplification of the vehicle. The simplification of the underside has an estimated coefficient of drag decrease of 0.015. That along with the simplified wheels and no internal flow, puts the coefficient of drag between the simulation and the indicated coefficient of drag within a reasonable error. This means the simulation passes the “sanity” check to ensure validity in the data.

XZ-Plane Pressure Cut Plot

XZ-Plane Pressure Cut Plot

Ground Pressure Cut Plot

Ground Pressure Cut Plot

XZ-Plane Velocity Cut Plot

Okay, now what does all this mean???

The main question is actually a trick question since there is not one “BEST” design.  Different designs are all about compromising.  First L/D and downforce goals should be known for best track times ***This is different with different tracks and cars*** The rear aero should be decided first and then the front design should balance the car back out.

-Paul

Mazda Miata Consultation

The first cfd consultation of Hancha Group. A Mazda Miata that races in SCCA Solo and track days wants to know whether aerodynamic changes actually helped performance.  He also wants to look into other aerodynamic performance changes.

actualPictureActual Picture of Car

s_2_cfd_modelCFD Model (No front splitter and no wing)

meshQuality_2Mesh Quality on Surface

The first analysis is of a stock Miata at stock ride height.   This was just for shear curiosity and to compare to the rest of the runs.

cd = 0.36

cl = 0.27

s_1_pressure_plot_1

Pressure Plot of Stock Miata

s_1_velocity_plot_1

Velocity Plot of Stock Miata

The next analysis was of a stock Miata lowered to have a 4 in ride height.  This was how the looked before the additional aero.

cd = 0.41

cl = 0.08

The results were as expected.  Lowering a vehicle will increase the downforce by increasing the velocity below the vehicle.  This does however increase drag.  The aero balance did not change between the runs.

s_2_pressure_plot_1

Pressure Plot of Lowered Miata

s_2_velocity_plot_1

Velocity Plot of Lowered Miata

This is all for now….currently running the current setup on the car to compare to it to the lowered Miata.  Stay tuned for more on this project.

Update on 01.14.2013

This project has been done for a little while. I am just now getting to the finished results.  The project was successful.

The next analysis was on the current cars setup.  The current car has a front splitter and a dual element rear wing. This setup did increase drag, as expected because of the rear wing. The big benefit that can be seen is the coefficient of lift.  The Miata is now making downforce!  Good news since drag currently isn’t the big concern.

cd = 0.51

cl = -0.93

s_3_pressure_plot_1

Pressure Plot of Current Setup

s_3_pressure_plot_4

Pressure Plot Under Rear Wing

s_3_velocity_plot_1

Velocity Plot of Current Setup

To improve this setup, the Hancha Single Element Wing was used instead of the dual element wing.  Single element wings are better to use if you can meet the downforce required because they will have less drag.  Now if you cannot meet the downforce required, duel element wings would have to be used. This setup improved the performance by lowering drag and increasing the downforce.  This is a double win by increasing the efficiency!

cd = 0.48

cl = -1.00

s_5_pressure_plot_1

Pressure Plot Improved Car

s_5_pressure_plot_6

Pressure Plot Under Rear Wing

s_4_velocity_plot_1

Velocity Plot of Improved Car

This setup still isn’t perfect.  The wing is stalled in the center, which can be seen in the pressure plots of the car and the wing.  The picture below shows the deadwater behind this section of the wing.  Improvements are still in the process.  This car will be getting one of the prototype single element wings!

s_5_stream_3

Deadwater in the Center of the Wing