Exhaust Blown Floor - Forward Exhaust Exit

[ringo]

This image shows a free stream, where the car does not interact with the flow

You know that, since you created it. Right?

This was from an equation. This image was just used to counter marekk's interpretation of the equations. He said they show that the flow goes under the car.

NOW, once we take into account that the diffuser generates low pressure, obviously that exhaust stream would be drawn back in, correct?

No it's not like that. The diffuser does not perform work like a pump. This is why it's not always safe to assume that the flow under the car will suck everything around it because the pressure is low. If you watch the F1 car move in the rain or dust, you wont see dust being drawn in from around the sides. The car does the work and it is always pushing the air. So there's no definite rule of thumb.

Since that blue line would theoretically crash into the rear tire in a free stream where the rear tire (and the rest of the car) had no effect on the air (which is what you just said), then if you take into account the effects of the diffuser and the tire, then that blue like would likely end up going inside the rear wheel and SURPRISE, SURPRISE, through the diffuser.

Tthe CFD takes into account all of these things you say and even more and iterates hundreds of times.

Therefore, if some of the exhaust stream ends up following that blue line, then SOME of it may end up going through the diffuser.

But at what quality?
Certainly not a higher quality of flow than a proper desinged redbull diffuser.

When it crashes into the tyre, it loses it benefits as the temperature is very low and the gases lose a lot of momentum and get very turbulent.
This was what made me doubt the diffuser blowing. The gas is at about 70 degrees C at the tyres and much of the energy is dispipated into the air by then as the gas has traveled about 1.5m.

It can be seen that this exhaust is not of a comparable quality going into the diffuser as other examples like the redbull and the mclaren.

The flow is slower, it is more turbulent; having crahsed into the tyre, the temperature is also lower.
So it can be reasoned that placing the exhuast so far upstream of the car to blow the diffuser downstream, cannot be as effective as blowing the sides of the diffuser point blank.

I do agree with you that the effect is probably largely to fill in the front tire wake (your CFD shows that quite well) and to heat the air at the leading edge of the floor, but your hard-headed stance that no exhaust will possible go into the diffuser is quite frankly unfounded (unless there are some vector or streamline plots that you are withholding... which is why I asked if you had them so I could better understand roughly where the air would go).

What was said previously is that in some cases a few puffs of gas will go between the tyres. It's somewhere in this thread.
But some people are acting like if a puff of smoke goes into a diffuser it's like Nirvana and the car turns into a downforce monster. That one puff is what they need to save face.

I am not withholding any thing at all. There is no sleeping giant.

[ringo]

Haha, ok.

Awaiting your less-hurried reply, then!

Diffuser is at 0m exhaust at 2.5m.

It would be wiser to place the higher energy nearer to the diffuser if the diffuser was the focus.

The high energy must be placed where it can be most effective.
For renault it's not at the diffuser. Its at the same place 2.5m upstream of the diffuser.

This is why rebull build the cannon exhausts. To keep in the energy and unleash it at the point it is most needed for their design; at the diffuser.




There is no reason to have the gas losing 80% of it's speed when it reaches the diffuser. It's a waste of energy and fuel.

This is very easy to understand. You can directly see why blowing the diffuser from a far is not effective, as the atmosphere soaks up the energy.

[malcolm]

No, I get all that... but you were inferring that the exhaust stream did not go back at all toward the diffuser and that it would just continue going outward, outside the tire or something.

I didn't say that was the intent of the design... just that some of the exhaust would end up there.

[PlatinumZealot]

I think he said with speed the direction of the exhaust changes. But you need the Downforce the most at medium speeds.

[shelly]

Two most important areas where I think ringo is wrong:
- gases do go under the diffuser. The pathlines are deflected inwards by low pressure on the diffuser kink
-to perform a decent cfd, you need a powerful computer and a well designed mesh. cad embedded cfd tools can not give significant answers for a car external simulation. So in this case marekk is closer than ringo about estimating needed resources, even if he is overestimating a bit.

[marekk]

I really appreciate ringo's cfd (i know how much work you have to put into this), but to get usable results one have to use terabyte-range datasets (with detailed real car model) and use one of moving meshes techniques to simulate at least the tarmac and rotating tyres (not to mention aeroelasticity effects).

I considered moving wheels and ground. Even the ground roughness is considered and tyre surface roughness.
You don't need terabyte range whatever you are saying. That's just techno babble, which you have no experience with.

You would be surpised. After 5 years of designing and building compute clusters (that's my daily job) i'm quite sure i have some experience, and there are reasons other then techno babble people spend +$1M on those systems.

On "normal" PC (even strong one) solver run will take ages to finish. Even on 10TFlop, 1TB RAM cluster much simpler models run for hours/days.

I know how to use the PC resources well. Day long calculations are overkill. The differences are very minute, and the effort is not worth it if it will only be rejected because some guy says his cousin knows someone working in F1.

As said, i really appreciate your work, lacking any real car's CFD in public domain, your pictures/videos provide at least some basic understanding of what's going on (you know - one picture says more than million words), so i think it's still worth the effort.
On the other hand i don't think differences to real cars are very minute, your simulation lacks any visible bending of the flows due to diffuser work (which counts for 25-40% of total car's downforce) clearly visible on posted Toyota's CFD, bargeboard vortexes are not visible to.

PS: I hardly know anyone working in F1, AFAIK Kubica is/was the only guy in F1 from my country - and even if i hear/read some "insider" info i always try to remember that real information in F1 is worth big chunk of money ($100M for McLaren for example) and it's not given to anyone for free

[marekk]

...I do agree with you that the effect is probably largely to fill in the front tire wake (your CFD shows that quite well)...

And i doubt it.

Frontal area of F1 front wheel (660mm x 355mm) is 0,213m^2 and at 100m/s (top speed, when drag reduction really matters) this wheel displaces some 26,4kg of air every second.
We agreed back in the thread, that exhaust flow at 18,000RPM is about 0,216 kg/s pro pipe - less then 1%. Even taking into account that due to front wing's work external flow doesn't "see" the whole front wheel, i don't think it's significant amount of gases to 'fill" the wake.

[marekk]

There is no reason to have the gas losing 80% of it's speed when it reaches the diffuser. It's a waste of energy and fuel.

This is very easy to understand. You can directly see why blowing the diffuser from a far is not effective, as the atmosphere soaks up the energy.

It's just as easy to understand, that hot exhaust gases loose some part of thermal energy by radiation (to the tarmac, floor, outside world), and this is responsible for total volume contraction, which in turn lowers the pressure in constrained space under the floor and produces downforce, so it's not wasting of energy, it's using the energy in another way.
Even just making air, floor and diff under the car hotter increases car's efficiency.
Getting hot, thin air in front of rear wheels reduces drag big time (even if it's only 80 degree above ambient, density reduces 25%, and drag is proportional to the cube of velocity).
Kinetic energy (momentum) of exhaust is used to increase mass flow (jet pump effect), seal outer parts of floor's leading edge and push flow lines outwards, so it's not wasted either.

RBR's solution is more about sealing diffuser sides and "jet pumping" additional massflow, they use mostly the kinetic part of exhaust energy.

Renault's solution is more efficient IMO.

[Ferraripilot]

There are reports that W02 will show up in Spain with a FEE. Brawn and Newey seemed to be fond of the design during testing, and now that W02 is adopting it perhaps it has a decent amount of merit per MB's wind tunnel testing?

[ringo]

Two most important areas where I think ringo is wrong:
- gases do go under the diffuser. The pathlines are deflected inwards by low pressure on the diffuser kink

I'll show you something later. Accompanied by some pictures.

What i will say now is that air doesn't migrate under the floor as easily as you think. You are not considering the front tyre wake interaction, rear tyre interaction and the sidepods.
If air migrated under the floor easily high nose designs wouldn't be necessary; neither would tapered floor bottoms.
Little on that later.

edit: if air migrated under the car as well, the car would lose downforce.
The car floor is like a refrigerator. How so?

A refrigerator opposes the nature of hot air flowing to colder regions. Reason is work is being done to make that happen.
In the case of the diffuser, the car body can perform work to prevent high pressure from flowing to low. I stress can, i don't know it for sure.
The car's floor can squeeze low pressure to high to create much lower pressure.

I'll post the shots later to show that.

-to perform a decent cfd, you need a powerful computer and a well designed mesh. cad embedded cfd tools can not give significant answers for a car external simulation. So in this case marekk is closer than ringo about estimating needed resources, even if he is overestimating a bit.

No you don't need a powerful computer. You can't tell me that becuase i've tried it.
Making the mesh is the easy part. Generating it is another thing, and running the calculation.
I am telling you that the disgustingly high levels of CFD are only more accurate, but a normal level will still use the same basic mathematical and thermodynamical etc funtions. Because the package uses the same calculations regardless of computer power, one simulation will not be drastically different than another.

See what i am saying?
If i run an exhaust manifold CFD on a lesser computing system then run it on one 100 times better, the principles are the same and the results will generally be the same. You will only see hyper details on the better system.
In this case of the R31 there will not be a massive departure using a better sytsem, as the mathematical rules are the same. Apreciate what I am saying?

I had a slower system before, and what takes me hours now, would take a day on that. Were the results any different? Yep, but mostly numerical but not much in terms of visual.
A better system will get more accurate numbers but the principles will be the same.
So in my case, the flow wont go somewhere else on a better system with an infinitely better mesh. You may see more vortices and more track and tyre interaction, but the general result wont be different.

[ringo]

...I do agree with you that the effect is probably largely to fill in the front tire wake (your CFD shows that quite well)...

And i doubt it.

Frontal area of F1 front wheel (660mm x 355mm) is 0,213m^2 and at 100m/s (top speed, when drag reduction really matters) this wheel displaces some 26,4kg of air every second.
We agreed back in the thread, that exhaust flow at 18,000RPM is about 0,216 kg/s pro pipe - less then 1%. Even taking into account that due to front wing's work external flow doesn't "see" the whole front wheel, i don't think it's significant amount of gases to 'fill" the wake.

Oh boy. It's not as simple as you make it. You have no idea.

[strad]

No it's not like that. The diffuser does not perform work like a pump. This is why it's not always safe to assume that the flow under the car will suck everything around it because the pressure is low. If you watch the F1 car move in the rain or dust, you wont see dust being drawn in from around the sides. The car does the work and it is always pushing the air

.
Ringo,,For myself I kinda just assumed this. The diffuser to my mind just does a very efficient job of letting the air out,,,so much so it creates a low pressure area...so to speak...
Wouldn't the air flow squeezing in from under the front splitter sort of push the exhaust gases out of the way,,,that being to the outside and that would be the seal. Yes..no?

[shelly]

No you don't need a powerful computer. You can't tell me that becuase i've tried it.
Making the mesh is the easy part. Generating it is another thing, and running the calculation.
I am telling you that the disgustingly high levels of CFD are only more accurate, but a normal level will still use the same basic mathematical and thermodynamical etc funtions. Because the package uses the same calculations regardless of computer power, one simulation will not be drastically different than another.

See what i am saying?
If i run an exhaust manifold CFD on a lesser computing system then run it on one 100 times better, the principles are the same and the results will generally be the same. You will only see hyper details on the better system.
In this case of the R31 there will not be a massive departure using a better sytsem, as the mathematical rules are the same. Apreciate what I am saying?

I see clearly what you are saying, and I think you are wrong. If your mesh is not good enough (and in this case, good enough means also big enough) you can not really see that much. Especially pathline tracking is very dependendent on your volume mesh resolution (and pathline reconstruction method also: you can see different result for the same cfd case just tweaking the tolerance on cell-to cell path reconstruction).
Everybody knows that the equations are the same, but solution methods and resolution play much a bigger part than what you are saying. So using a much better system you will see different things (n-s are non linear!). Maybe you have not seen the transition in solution from your first pc to your second pc; but if you go to a much bigger pc with suitable software and suitable input mesh you would see it.

Advice: stop being that aggressive and that stubborn: try to discuss normally as others do.

[Just_a_fan]

I've done a bit of CFD stuff (for fire simulations) and the results can vary markedly with the same software on the same PC but with different mesh dimensions. That's with an LES method rather than a RANS method.

[ringo]

Shelly i hear you, but what I am saying is that the exact same mesh can run on a slower system.
Users only limit their mesh because of sanity. Time and your electrical bill is the enemey with a slower computer. Worse if the CFD is being used commercially.

If i a run a 1 million cell mesh on a super computer, I can run 1 million cells on a slower one. One will take an hour the other may take a week.
Results will be the same if settings are exactly the same.

But if i try to economize on time, i may cut down cell count on a slow computer and also other refinements to try to make the time equal. That is where the slow computer fails, when you turn down the calculation details to compensate time wise.
That's the compromise, especially in F1 where time is critical with resource restrictions ect.