Diffuser Profile - Convex or Concave?

[BreezyRacer]

Drew
While we cannot really know what goes on behind this step design without some testing, it's fair to say that it doesn't seem be normal in the 2011 diffusers anyway. I wonder if it's there to stall out slow boundary layer airflow, thus filling the diffuser with high energy airflow?

[hardingfv32]

I wonder if it's there to stall out slow boundary layer airflow, thus filling the diffuser with high energy airflow?

Please expand. What is meant by stalling the boundary layer?

Brian

[BreezyRacer]

I wonder if it's there to stall out slow boundary layer airflow, thus filling the diffuser with high energy airflow?

Please expand. What is meant by stalling the boundary layer?

Brian

Once you get airflow over a surface you get a buildup of boundary airflow .. basically the frictioned airflow against the surface it's flowing over, which moves at speeds between the surface (no movement relative to airflow) and the high speed flow. You usually measure it in thickness. On something like an undertray it's very small so usually discounted or forgotten alltogether. *Perhaps* the rise stacks up the boundary airflow and allows a more highly charged airflow into the diffuser box. That's what I was throwing out there.

[ringo]

Boundary layer doesn't stall per se, it transitions from steady to unsteady flow.

When it gets unsteady it's height grows very rapidly. If the boundary layer height is the same as the ground clearance then you are in problems.

If the friction on the diffuser causes the flow to be unsteady, you are in problems again, due to the energy loss in the eddies and the turbulent flow.

viscosity of the fluid is a factor, the fluid speed, surface roughness, surface lenght. What is important is the shear stress on the surface mostly.

[BreezyRacer]

Boundary layer doesn't stall per se, it transitions from steady to unsteady flow.

When it gets unsteady it's height grows very rapidly. If the boundary layer height is the same as the ground clearance then you are in problems.

If the friction on the diffuser causes the flow to be unsteady, you are in problems again, due to the energy loss in the eddies and the turbulent flow.

viscosity of the fluid is a factor, the fluid speed, surface roughness, surface lenght. What is important is the shear stress on the surface mostly.

All understood and conventional aero theory .. BUT Ringo what do you think the step is about?

[ringo]

I don't know to be honest.
It may have to do with the range of speeds the diffuser operates in ?

Or transients as the diffuser moves with the body of the car relative to the flow under the diffuser.
We always imagine diffusers to be stationary relative to the expanding flow, but on an F1 car they move a lot.

If the air flow is moving through the diffuser at say 1cm per second for arguement's sake. The diffuser is 25cm long.
Everything will be honky dory if in 25 seconds the diffuser doesn't budge.

Now let's say the diffuser/car body has a simple harmonic motion and it oscillates before that stream of air can pass through from throat to exit. You're gonna have some funky effects under it with that expanding air. At times the diffuser will move away, increasing expansion rate, then it will move towards the ground doing the opposite.

It's these kinds of considerations why i can't really stick a pin on any reason for the step. There a lot to consider.

what's for sure though is that it is very critical, as it is near the throat.

[godlameroso]

You would also have to consider the philosophy driving the rear end of the cars during that era(Amazing how time flies). Before the goal was to converge the effects of the beam wing, diffuser and rear wing, so that they all influenced each other. The goal today is not too dissimilar, what is dissimilar is how air travels through each element. Nowadays it's become more challenging to combine their effects due to the regulations, whereas before, the upwash of the diffuser was feeding the upwash of the beam-wing, which was feeding the upwash of the main wing. The regulations actually encouraged this type of design.

The current regulations have teams scrambling for ways of combining these effects, EBD's, DDF's, diffuser gurneys, and on top of that there's the DRS to consider.

[gixxer_drew]

Yah I call it a "trip fence" (this is land speed nomenclature) its one of the first things I thought about, you trip the boundary layer helps keep the flow attached at higher angles of attack and sometimes to reduce drag. Entirely possible. I never tested this of value in CFD, however I deal with much lower cell counts than F1 and the tunnels are not nearly as accurate. Last but not least, I have not designed any ground up open wheelers, always had other types rules to contend with....

[shelly]

Boundary layer doesn't stall per se, it transitions from steady to unsteady flow.

When it gets unsteady it's height grows very rapidly.

ringo you are confusing transition with stall. They are two different things. You have stall when the boundary layer separates and wall shear stress goes to zero.

[BreezyRacer]

So it does seem like we've reached some kind of consensus on what this step is for .. it's there to strip away the boundary layer .. notice that does so AFTER the entry curve .. presumably the boundary airflow layer helps the flow to transition into the diffuser box, and after that it's not desired.

[gixxer_drew]

So it does seem like we've reached some kind of consensus on what this step is for .. it's there to strip away the boundary layer .. notice that does so AFTER the entry curve .. presumably the boundary airflow layer helps the flow to transition into the diffuser box, and after that it's not desired.

I wouldn't say that, I think we laid out some possibilities. To really know this you need to have a full car model and an accurate way of testing it. Considering other teams likely had those photos and more, but did not copy the device tells me people at the level of F1 were still wondering if/how/why it worked.

Someone probably added it to the CFD test queue and it didn't work with all their other details.

[BreezyRacer]

So it does seem like we've reached some kind of consensus on what this step is for .. it's there to strip away the boundary layer .. notice that does so AFTER the entry curve .. presumably the boundary airflow layer helps the flow to transition into the diffuser box, and after that it's not desired.

I wouldn't say that, I think we laid out some possibilities. To really know this you need to have a full car model and an accurate way of testing it. Considering other teams likely had those photos and more, but did not copy the device tells me people at the level of F1 were still wondering if/how/why it worked.

Someone probably added it to the CFD test queue and it didn't work with all their other details.

Yes, I agree that it probably doesn't work .. when you think aero innovation the last team you would consider is the old Honda team. But someone had a theory and someone engineered it out and tried it .. IMO we're looking for the underlying theory here, not whether it worked. I haven't seen this step design in any other diffuser pic (I have quite a few .. I'm a diffuser junkie ). I think we've hit on the theory that spawned this design.

[godlameroso]

The thing is that theory no longer works under the scope of the current design regulations. It makes more sense to treat each element separately than it does to try to combine their effects. Perhaps the diffuser and beam wing could combine their effects given their proximity. The rear wing however is almost excluded from this "conversation". Of far more importance is how much flow can be fed into the rear element areas, and how can you best exploit the flow you do condition to go in said area.

Given that you've limited the amount of pressure drop in the floor of the car, you have to compensate by increasing it in the opposite side in order to get the same effect.

Whereas in the front, you can pretty much make the air do whatever you want, the rear is left in essence to scavenge what the front leaves in it's wake. So instead of just focusing on the diffuser, I think it would make more sense to see if you would work with the area of bodywork called diffuser, and make it function as a wing as well as a diffuser.

The combination of concave and convex shapes are the best compromise to both diffuse air travelling across the floor, and to increase the pressure on the upper side of the chassis, while also trying to feed the underside of the beam wing as much as possible.

[PNSD]

viewtopic.php?f=6&t=6223&start=30

[gixxer_drew]

I just had another thought on this issue. Its a good example of why you cant draw conclusions so simply. Something else to consider as a cause of a boundary trip fence on a diffuser.

It's really hard to define what is the "boundary layer" there is a dictionary definition but essentially we are merely humans observing the physics of the world around us and on top of that we cant measure everything without effecting it in the process. If low velocity is building up on a surface people tend to define it as "boundary layer". Now if the boundary layer buildup exceeds what you calculate it to be (and this happens ALL the time) you kind of scratch your head, but depending on budgets and time you pursue what you must to get a car ready and performing at the best you can with the tools and money you have. Especially when you have a rolling floor, everyone thinks that fixes everything.

So you go ahead and call it boundary layer, you make a device to trip that boundary layer and wham you get the RDF value you wanted or you think is competitive next year. Reality is that a wind tunnel has compromises of its own and inaccuracies of its own. Someone I really look up to once told me that you never really understand the car until you designed and tested the wind tunnel as well.

This could have been an issue with rolling belt geometry or sizing, effects from something else on the floor, nozzle geometry, straighteners or something else I haven't thought of. Pitot tubes are not really good at telling you a larger picture just a single momentary location's value. Nothing ever tells you why.

If your CFD and your tunnel data don't converge, what do you trust? A mathematical system designed around giving the same results as a wind tunnel or a wind tunnel?

Nobody else copied it, it only worked in one tunnel. Might explain their car's performance if it were true. DId that feature appear on the new car after they pulled out?

Just a theory, no real way to test it back in time, but imagining it is a good exercise for me. Keep me on my toes, always rethinking and rechecking my test data. What is 100% accurate on one car can be totally screwy on another. Just something to think about. I've seen this happen before in similar fashion. In my vivid imagination F1 tunnels are so amazing they always give perfect values but reality may be a bit different.