Diffuser / under body questions

[miqi23]

Well, all this argument for nothing lol, seriously they dont work at those speeds. My days from Uni, I remember students working on it just because they fancied aero and it gave them a chance of learning the theory behind it!

If some one can give out some dimensions of their FSAE underbody dimensions I can perform some CFD simulations and prove the changes that occur are useless!

[BreezyRacer]

Here's a quote from a 2006 technical paper by Xin Zhang on race cars in ground effect ..
___________
Aerodynamics, particularly ground effect aerodynamics, as applied
to open wheeled race cars is still mainly an experimental
science and will remain so for some time to come. This is
primarily due to the complex fluid flow physics involved. These
include
• separation as a normal feature
• surface character changes during an event lead to early transition
• suspension motion leading to unsteady flow
• highly complex physics: wall jet, shear layer instability, vortex
meandering and breakdown, etc.
• force enhancing vortices
• turbulent wake and ground boundary layer interaction
• compressibility
However, computational fluid dynamics CFD is becoming
much more important and its use complements model scale experiments.
This is particularly true in the case of flows around
geometries such as a front wing assembly, where the flow could stay attached over the majority of the aerodynamic surface, less so
for flows such as that associated with a diffuser, where the incoming
flow could be highly turbulent and distorted, and large vortex
flows are often coupled with flow separation.
The primary aim of race car aerodynamics is to generate a
desired level of downforce negative lift for the least possible
drag. However, the balance of the downforce under all conditions
of speed and acceleration is equally important. As such, the complex
flow features associated with individual components are often
interwoven and difficult to separate.
_______________

So he feels that CFD has a ways to go in fully explaining and emulating ground effects aero.

In any case if you would like an exercise to run, try a 55"x60" flat bottom with a 15"wide x 30" long diffuser at 10 degrees, and 2" of ground clearence. Velocities of 1.5ms, 3ms, and 5 ms.

[AeroGT3]

Well, all this argument for nothing lol, seriously they dont work at those speeds. My days from Uni, I remember students working on it just because they fancied aero and it gave them a chance of learning the theory behind it!

If some one can give out some dimensions of their FSAE underbody dimensions I can perform some CFD simulations and prove the changes that occur are useless!

Our undertray has been designed to produce more downforce than both wings combined. I cannot give out the design, but we have written our own codes and we have done CFD.

We will have track test data to back this up as soon as the motor is running!

[BreezyRacer]

We will have track test data to back this up as soon as the motor is running!

I appaud your enthusiasm for your design, but saying you'll have confirming results as soon as you test it is realy putting the cart before the horse, don't you think???? You may be too expectant of the results to be objective enough to evaluate them. Good luck though.

[AeroGT3]

We will have track test data to back this up as soon as the motor is running!

I appaud your enthusiasm for your design, but saying you'll have confirming results as soon as you test it is realy putting the cart before the horse, don't you think???? You may be too expectant of the results to be objective enough to evaluate them. Good luck though.

Our expectations are quite reasonable given the fidelity of the CFD we've done as well as the results of other teams with very similar designs.

Nonetheless, our tray will be worth its weight even if it produces half the downforce we've designed it to.

[BreezyRacer]

Just curious .. do you guys have a min weight in FSAE? Good luck on your design, just remember that the end result is what you're looking for, not the CFD. Though I guess the CFD helps your FSAE ranking doesn't it?

[AeroGT3]

Just curious .. do you guys have a min weight in FSAE? Good luck on your design, just remember that the end result is what you're looking for, not the CFD. Though I guess the CFD helps your FSAE ranking doesn't it?

We have no minimum weight. We are the lightest car by at least 30 lbs (which is huge), which makes designing Aero "easy" as we don't need much to make a difference!

[BreezyRacer]

We have no minimum weight. We are the lightest car by at least 30 lbs (which is huge), which makes designing Aero "easy" as we don't need much to make a difference!

Wow, that must be nice. 30 lbs is almost 10% for a good FSAE car. What kind of motor do you run?

[slimjim8201]

So he feels that CFD has a ways to go in fully explaining and emulating ground effects aero.

In any case if you would like an exercise to run, try a 55"x60" flat bottom with a 15"wide x 30" long diffuser at 10 degrees, and 2" of ground clearence. Velocities of 1.5ms, 3ms, and 5 ms.

He is referring to the inability of CFD to analyze the many transient phenomena that occur as the cars go through their very dynamic motions. Braking and acceleration pitch, cornering roll, etc. It's not so much that it can't be done, it just hasn't been done yet. Regardless, CFD remains FAR FAR FAR more efficient as a design tool than a wind tunnel could ever be. The cost of building numerous prototypes or implementing many design changes for wind tunnel testing is always going to be more expensive than analyzing the models in a CFD code. No, you won't get 100% accurate results, but no one expects that and frankly no one needs that. The correlations found in CFD codes (ie the percentange changes among many designs) ususally translate to real world data VERY well. Final testing should always be done in a wind tunnel, but CFD should be the foundation.

...unless, of course, you really like wasting time and money...

[nazo]

Here's a quote from a 2006 technical paper by Xin Zhang on race cars in ground effect ..
___________
Aerodynamics, particularly ground effect aerodynamics, as applied
to open wheeled race cars is still mainly an experimental
science and will remain so for some time to come. This is
primarily due to the complex fluid flow physics involved. These
include
• separation as a normal feature
• surface character changes during an event lead to early transition
• suspension motion leading to unsteady flow
• highly complex physics: wall jet, shear layer instability, vortex
meandering and breakdown, etc.
• force enhancing vortices
• turbulent wake and ground boundary layer interaction
• compressibility
However, computational fluid dynamics CFD is becoming
much more important and its use complements model scale experiments.
This is particularly true in the case of flows around
geometries such as a front wing assembly, where the flow could stay attached over the majority of the aerodynamic surface, less so
for flows such as that associated with a diffuser, where the incoming
flow could be highly turbulent and distorted, and large vortex
flows are often coupled with flow separation.
The primary aim of race car aerodynamics is to generate a
desired level of downforce negative lift for the least possible
drag. However, the balance of the downforce under all conditions
of speed and acceleration is equally important. As such, the complex
flow features associated with individual components are often
interwoven and difficult to separate.
_______________

So he feels that CFD has a ways to go in fully explaining and emulating ground effects aero.

In any case if you would like an exercise to run, try a 55"x60" flat bottom with a 15"wide x 30" long diffuser at 10 degrees, and 2" of ground clearence. Velocities of 1.5ms, 3ms, and 5 ms.

he was my lecturer.... lol.... zhang! got 88 for a cfd project with him

[BreezyRacer]

Here's a quote from a 2006 technical paper by Xin Zhang on race cars in ground effect ..

he was my lecturer.... lol.... zhang! got 88 for a cfd project with him

Sounds like you're a lucky and very smart guy!

[AeroGT3]

So he feels that CFD has a ways to go in fully explaining and emulating ground effects aero.

In any case if you would like an exercise to run, try a 55"x60" flat bottom with a 15"wide x 30" long diffuser at 10 degrees, and 2" of ground clearence. Velocities of 1.5ms, 3ms, and 5 ms.

He is referring to the inability of CFD to analyze the many transient phenomena that occur as the cars go through their very dynamic motions. Braking and acceleration pitch, cornering roll, etc. It's not so much that it can't be done, it just hasn't been done yet. Regardless, CFD remains FAR FAR FAR more efficient as a design tool than a wind tunnel could ever be. The cost of building numerous prototypes or implementing many design changes for wind tunnel testing is always going to be more expensive than analyzing the models in a CFD code. No, you won't get 100% accurate results, but no one expects that and frankly no one needs that. The correlations found in CFD codes (ie the percentange changes among many designs) ususally translate to real world data VERY well. Final testing should always be done in a wind tunnel, but CFD should be the foundation.

...unless, of course, you really like wasting time and money...

Well said, though I think you underestimate the importance of wind tunnel/testing. No matter how great CFD is, without experimentation you are still guessing as to accuracy. You NEED a heavy commercial CFD code like Fluent to get great accuracy and you NEED very skilled and specialized people and big computers. All of these are very expensive! Also, you do not need to build a tunnel, just rent time in one . . .

You say CFD is far more efficient, but that's not always true. What if you already have a tunnel and expertise to run it, but not the clusters, software, and CFD experts? What if you are partnered with another company or university that offers you tunnel time without the cost of building your own? How much does a well versed CFD engineer cost you? How much do good commercial packages (not floworks et al) cost you?

But otherwise, yes, CFD can model trays well. To address breezy's points from Xin Zaung:

separation as a normal feature

• suspension motion leading to unsteady flow
• highly complex physics: wall jet, shear layer instability, vortex
meandering and breakdown, etc.
• force enhancing vortices
• turbulent wake and ground boundary layer interaction
• compressibility

Maybe Mr. Zhang wrote his paper when computing power wasn't high enough to solve the reynold's stress equations for anything but course grids over simple bodies, but, other than using simple spallart-allmaras models, the Re stress eqn's and even the K-omega and K-epsilon models can predict stall quite well if implemented correctly.

The navier stokes and Re stress eqn's can handle the complex physics you described. It's whether or not the grid can!

Force enhancing vortices can be captured with the proper grid. Heck, using crocco's theorem (relates enthalpy to vorticity) you can even model vortical elements (like propellers) rather well using inviscid panel codes!

Turbulent wakes and boundary layer interactions have been captured by CFD for a while. shock boundary layer interactions (shock induced separation sticks out here) can be captured with rather course grids in fluent for example. I can show you some examples if you wish.

Compressibility is easily modeled in CFD.

Perhaps Mr. Zhang's evaluation of CFD for the undertray is dated?[/b]

[BreezyRacer]

Here's a quote from a 2006 technical paper by Xin Zhang on race cars in ground effect ..

AeroGT3 .. That was his latest publication on undertrays and diffusers, published in 2006. I think it's fair to say that Mr Zhang has access to whatever resources he wants and if you've read any of his papers you'll see that he has extensively studied "generic" ground effect design for at least 6 years, amongst other issues as well.

Given that I'm really just a racer/enthusiast with an applicable project car I feel somewhat out of my element in speaking to CFD. Can you tell me where all the CFD is derived from .. meaning is it's assumptions and variables based on actual wind tunnel results .. or did CFD rewrite aero theories?

I get the impression that CFD coding is based on a long history of actual results in aero studies, thus CFD still has problems in perfecting undertray design because of the lack of actual test results upon which coding can be written to collaborate the actual results. Having said that, I do not doubt, in any way that CFD will be there, given that the coding is generated, and I'm not here to dis CFD.

IMO there are still many details in reference to undertray/diffuser design that are largely unanswered/unpublished. I'll list a few that still puzzle me ..

1. Strakes and their effect on generating votexes. For instance does adding strakes (usually used to "smooth" airflow) create add'l vortexes, or do they have the potential to weaken existing vortexes?

2. Optimal diffuser width and length ratio. There appears to be an effective range depending on the undertray deminsions, so are strakes used really to create a series of optimal ratio diffusers?

3. Given that lots of air is pulled in from the sides of the vehicle do we need/want any air inlets at the high velocity leading edge of the undertray? I would think we would because the high velocity is more valuable than the weaker air along the sides of the body. Also front inlets would guarantee a supply of air during braking and other low ride height situations for the undertray to work. F1 designs don't directly relate to this issue due to the stepped floor requirement, but you can still see all the VGs forward of the sidepods that are designed to control airflow under the sidepods. Toyota and Ferrari especially seem to be hard at work in these areas.

These questions are not thrown out to be a challenge, they are just things that are still unanswered to me after reading all the relevent publications I have been able to find on diffuser/undertray designs. No doubt lots of people have done a lot of work on these issues and for good reason that info is kept within that small community. But I'm all ears and anxious to hear some feedback.

Thanks to everyone that contributes to this great site.

[Carlos]

Strakes -- Do the weaken or strengthen vortexes -- probably both -- I first noticed the concept used in the design of turbine blade design -- and their effect in turbine engines -- a aero- engine turine enclosure is an extremely controlled situation -- but to me a strake is something like an elongated turbine blade - seving a similar purpose. I first noticed strakes applied to the Allard and Mazda Lemans cars upper body sufaces -- then the application to underbody aero -- I wonder if he CDF databases of GE would be of value to us -- Ther remains a lot of work to be done-- another idea -- touched on in our threads -- blown flaps -- could they have any ubderbody aero applications -- performing the function of strakes in under body aero -- fortunately -- there a lot of exploration to be done in this field -- a lot of our members , who are students have the chance to enter an interesting field -- a lot of opportunity to build careers in this field.
These are just the observations of a racing enthusiast who has watched aero developement over 4 decades.

[BreezyRacer]

Ther remains a lot of work to be done-- another idea -- touched on in our threads -- blown flaps -- could they have any ubderbody aero applications --

Are you refering to recessed air charged VGs? If so I don't believe they are legal in any racing series I know of. In the Japanese GT series the Nissan Skyline Team skirted the rules on this and arranged their radiator exits into the diffuser with a big radiator fan, of course. They were, amazingly, allowed to use it for the season but rules were subsequently tightened to prevent this again. Still that makes me curious as to how they best did this.

If anyone has any pics/diagrams on how this was achieved other than the obvious basics I would love to see them. Not because I could use it but because it might help us to understand better the dynamics in diffuser design.

As for strakes, from the studies I've read I understand that to get the absolute most from a diffuser you have to work to develop the vortexes that form at the upper corners of the diffuser throat. It takes a certain amount of energy to form these vortexes. I imagine that adding a full strake to an already effecient design would weaken these vortexes, but maybe *magically* it might add another set of vortexes that combined with the original vortexes would be stronger .. I doubt it but I don't know and I haven't found any studies that address the issue of strakes, and the related issue of diffuser width/length ratio. I've found studies on strakes in general diffuser design but they do not focus on the aspect of vortex strength or generation, mostly cavitation. I know the aero experts amonst us will snicker at that since both are roughly the same, but the goal of those studies was not to enhance vortexes, rather to prevent cavitation. As you can see sometimes I think too much.