Diffuser Profile - Convex or Concave?

[shelly]

Here we see how quickly things get twisted. I this comparison vortex induce downforce is dominant for the convex diffuser

[miqi23]

@DRCorsa,you are making the mistake of running the setup in isolation, try including a region above the diffuser section that produces an upwash and the interaction should produce a much better representation.

[BreezyRacer]

@DRCorsa,you are making the mistake of running the setup in isolation, try including a region above the diffuser section that produces an upwash and the interaction should produce a much better representation.

hi buddy! How ya been? Good to see you here. Yes, I fully agree with what you've said. Something that resembles a diffuser top surface but is the same profile on both models would tell us more and should increase downforce levels for the diffuser as well as the upper surface itself.

[DRCorsa]

@DRCorsa,you are making the mistake of running the setup in isolation, try including a region above the diffuser section that produces an upwash and the interaction should produce a much better representation.

We haven't come yet to a conclusion as to which type of diffuser ie best for an F1 car. We are trying -through primitive analysis- to find the reason why F1 teams have chosen the concave (or bell) over to the convex one. The fact that still the convex is better for that simple analysis, proves that there is an interaction that makes it better.

[hardingfv32]

DRCorsa

Thank you for all your work so far......

To draw a conclusion as to which type of diffuser is best for an F1 car we are probably require you to expand your simulation to include more details about the shape and flow above the diffuser. Is it correct to say, keep it 2D so as not to bring vortices into the picture?

1) At this point, based on your simulation, a convex shape is better than a complex concave shape when restricting the testing with a simple bluff model. This correlates well with my research showing no use of concave shapes in venturis or diffusers in other industries.

2) It would seem that there is a reasonable to assume that the external flows above the F1 diffuser can save the concave design or....

3) There is some other reason other than to use the concave shape that we are not aware of.

Here is some data that I found. I am not sure how it correlates to our discussion. What is the harm to our F1 system of a low pressure recovery?

[DRCorsa]

I've a.lready designed this, which i think is way closer to an actual F1 floor but i am not sure how i should run it in CFD. Do you think that i should run it as it is with air flowing both on top and bottom surfaces, or it could be better to make its upper surface flat as in the previous experiment?

[hardingfv32]

I have no opinion on CFD, but I would be happier to see small steps in the simulation to help analyze each new modification to the simulation. In this case run first without the center convex section, endplates and strakes. Then make runs adding in these features one by one.

I would stay with a the std simple bluff car model at the front at this point.

I will work a some general layout of the shapes above the diffuser based on photo observation to provide some additional information for you.

Brian

[miqi23]

I've a.lready designed this, which i think is way closer to an actual F1 floor but i am not sure how i should run it in CFD. Do you think that i should run it as it is with air flowing both on top and bottom surfaces, or it could be better to make its upper surface flat as in the previous experiment?

The thing is that each design is unique and they work with the rest of the car in place. You will also need to consider the issue of ride height and rake angles and you will be able to get a better feel of what you have drawn.

For now, I would suggest that you run your model with air flowing over and under it.

[hardingfv32]

[
The thing is that each design is unique and they work with the rest of the car in place. You will also need to consider the issue of ride height and rake angles and you will be able to get a better feel of what you have drawn.

Do you propose considering all things at once? How do we learn from that? How do we know if mistakes have been made? If we test step by step, we gain knowledge at each step. We take that knowledge to the next step and if the results do not make sense with what has gone before, we examine for reasons or mistakes.

Ride height and rake angles in relation to the diffuser would be good.

These test work just fine using a simple fixed bluff test body. Trying to recreate the flows in front of the diffuser is way beyond what anyone on the forum can do in CFD (or demonstrated they can do).

Brian

[BreezyRacer]

[
The thing is that each design is unique and they work with the rest of the car in place. You will also need to consider the issue of ride height and rake angles and you will be able to get a better feel of what you have drawn.

Do you propose considering all things at once? How do we learn from that? How do we know if mistakes have been made? If we test step by step, we gain knowledge at each step. We take that knowledge to the next step and if the results do not make sense with what has gone before, we examine for reasons or mistakes.

Ride height and rake angles in relation to the diffuser would be good.

These test work just fine using a simple fixed bluff test body. Trying to recreate the flows in front of the diffuser is way beyond what anyone on the forum can do in CFD (or demonstrated they can do).

Brian

I would suggest 4 tires, the floor and a body to approximate an F1 car. It could be a bluff body with mods such as a cutout to emulate a raised nose with tea tray and tapered tail. You can learn a lot from that.

[marekk]

DRCorsa

...

Here is some data that I found. I am not sure how it correlates to our discussion. What is the harm to our F1 system of a low pressure recovery?

...
[/url]

Brian

I think you've already found the answer.
Concave shape is the best compromise between pressure drop (which means downforce in our case) and pressure recovery (less pressure recovery means more drag, less efficient beam wing and who knows what more due to backflow). Just draw the mean line between both pictured shapes.

[marekk]

Maybe this analogy helps (could be flawed as well, no warranty ):

In reference frame of every single air molecule entering the space under the floor's leading edge, the raked floor and diffuser form a moving surface of the sucking pump's piston. As the car moves, this surface goes up, volume increases, pressure drops. This cycle repeats for every next molecule, as long as the car moves.

Now, in time domain, we are discussing 2 scenarios of moving pump's piston in the same time over the same distance (from floor leading edge up to diffuser trailing edge):

1. raked floor + concave diffuser:
slow-slow-slow-slow-slow-slow-very quick-slower-slower-very slow
2. raked floor + convex diffuser:
slow-slow-slow-slow-slow-slow-slighlty quicker-quicker-much quicker


Work done in both cases is obviously the same, but in the first case, our impulse comes just in the right moment, in the second case it's to late, we are already drawing the air not only from under the car, but from all over the place. And the angle between the working surface and horizont quickly increases, which leaves us with less usable (downforce) and more harmful (drag) components of total (perpendicular to the surface) force.

[PlatinumZealot]

I REPOST THE RESULTS OF THE "BELL SHAPED" NEXT TO THE CONVEX DIFFUSER ADDING DOWNFORCE-DRAG FIGURES

"BELL" DIFFUSER
Downforce: 1381N
Drag: 220N

CONVEX DIFFUSER
Downforce: 1705N
Drag: 259N

Good work, but your convex diffuser could be much, much better. I suggest you try a straight line diffuser (within the FIA specs of course) as your "control."

[MIKEY_!]

Gurneys (slotted or otherwise) will be a key factor I suspect. Perhaps even that RBR flap. The flap will give a 'bell' profile to the concave design won't it?

[PlatinumZealot]

Gurney's change the environment pressure behind the diffuser but I think you can leave it out of the analysis.

Some say, that flap acts as a wing (but yeah same thing, best not to and an unnecessary variable to the analysis).