Diffuser Confusion

[vonk]

I think there’s general agreement that the air entering the diffuser will do so at speeds/pressures that vary with the speed of the car. But the exit condition is always the ambient static pressure in the car’s wake.

So, regardless of the philosophy by which one arrives at the inflow conditions, the diffuser must always have a matching area ratio to avoid back flow. A fixed wall diffuser can’t do that.

The following drawings depict three expansion conditions that might be expected in a concave diffuser (if it works ).



But it’s late and I’m going to bed. Good night, I’ll talk to you tomorrow.

Continuing from above, the grey bands represent the interface between the turbulent region denoted by the black vortex and the inflow. What goes on in that interface, and how broad it would be, would depend on the inflow conditions.

As I mentioned before, all of this is highly conjectural on my part. To what degree is the inflow at B laminar? Will an interface actually develop, or will the inflow burst into turbulence the moment it enters the diffuser. Is “diffuser” a misnomer for this device? Since it is the equivalent of “backflow”, the turbulent region would be at ambient pressure, though.

Somebody else must have asked the same question and the found it worth a try.



vonk

[PlatinumZealot]

Hey it will work.. but it might not work as well as a rigid wall... i do not know. I was setting up an ultrasonic flowmeter to measuere some drainage flows the other day.. and the closer the sensor got to the abrubt opening at the end of the pipe the more turbulence was experienced (seen as an inaccuaracy in the reading). The end of this pipe was bounded by the same fluid itself. So a non-rigid boundary in most cases cause energy losses because the fluid in the smooth part of the stream can turn into it. I think you have to play some tricks to get that concave diffuser to work as well if not better than a rigid wall diffuser. I think the main advantage of it would be having a "variable wall" almost like the f-duct of diffusers.

I think there’s general agreement that the air entering the diffuser will do so at speeds/pressures that vary with the speed of the car. But the exit condition is always the ambient static pressure in the car’s wake.

So, regardless of the philosophy by which one arrives at the inflow conditions, the diffuser must always have a matching area ratio to avoid back flow. A fixed wall diffuser can’t do that.

The following drawings depict three expansion conditions that might be expected in a concave diffuser (if it works ).



But it’s late and I’m going to bed. Good night, I’ll talk to you tomorrow.

How do you know that the air stream doesn't curl back into the turbulent zone... The car is moving froward and something has to fill the space, and the air in that pocket won't stick on like glue. There will be some interaction of the air in the pocket and the air stream. I think the only way to see if it works well is to do some sort of experiment.

You see, what you are essentially saying with the concave diffuser is that you don't need a diffuser on the car, just a horizontal plate that juts out the back. So this is a big claim. I think some tests have to be done.

[vonk]

n smikle, (my comments in blue)

How do you know that the air stream doesn't curl back into the turbulent zone... The car is moving froward and something has to fill the space, and the air in that pocket won't stick on like glue. There will be some interaction of the air in the pocket and the air stream. I think the only way to see if it works well is to do some sort of experiment.
I agree, as I wondered in my post: “Will an interface actually develop, or will the inflow burst into turbulence the moment it enters the diffuser. Is “diffuser” a misnomer for this device?” I wish I had access to a water table, like the one I used for flow visualization in college.

You see, what you are essentially saying with the concave diffuser is that you don't need a diffuser on the car, just a horizontal plate that juts out the back. So this is a big claim. I think some tests have to be done.
I am definitely NOT saying that, and I make no claims. Instead, I am talking about a boxed compartment in which, I suspect, the curved upper surface would be vital to keep the (probably turbulent) air mass in that volume rotating. I seem to remember, that the Kamm Effect relied on a series of large, shedding vortices that were generated by the car’s tail cut off and the slipstream. And yes, tests would be nice. A water table would be an inexpensive way to begin.

vonk

[Just_a_fan]

On the cars, the "vertical" bit of your diffuser is only a few millimetres. The scale is important in making this work on the cars.

[vonk]

On the cars, the "vertical" bit of your diffuser is only a few millimetres. The scale is important in making this work on the cars.

In this, I am trying to understand only the device shown in the picture. It may be a bit early for Reynolds Numbers.

[Just_a_fan]

Picture not showing here so don't know what you're referring to...

And what has Reynolds number have to do with the size of the diffuser fitted to the car - this is constrained by the rules.

[shelly]

vonk, I think the basic issue that has to be ironed out is that there is no 90° degree surface turn in the picture of the rbr diffuser you posted. Once this misunderstanding is removed, discussion could maybe proceed.

Also note this: a section like the one you sketched would not have upward deflecting streamlines, but rather an horizontal wake boundary line with no pressure recovery (and no downforce on the floor). Remember that in a subsonic field information travels both upstream and downstream.

[hardingfv32]

The angle of the turn is about 30 deg with a 2" radius at the junction with the floor based on the photo from the RB7 thread.

Brian

[PlatinumZealot]

Yes.

After looking at a few under car photos. the dirt on the diffusers give the effect that it is concave.

Vonk, I think your device can work, but I suspect some other effects. That will work against it.


For the restm when I say might as well stick a horizontal plank on the back of car and run without a diffuser I will draw a diagram and show you what I mean.

[Just_a_fan]

The concave nature of the McLaren diffuser is shown here too:


Indeed this diffuser is mixed with concave outer channels and a convex inner section with a complex 3d shape joining these together.

This nicely shows how complex this area is and how armchair experts are always going to struggle to do more than describe it in the most basic way.

[hardingfv32]

Could the flow at the throat/kink be that slow as to allow the steep rake (say 30 deg) at the front of the diffuser without stalling?

Brian

[vonk]

Picture not showing here so don't know what you're referring to...

Here is a repeat:

In this, I am trying to understand only the device shown in the picture. It may be a bit early for Reynolds Numbers.

And what has Reynolds number have to do with the size of the diffuser fitted to the car - this is constrained by the rules.

The Reynolds number is important in the (wished for) water table test. I was just joking.

BTW what do you mean by "vertical" bit?

[hardingfv32]

Could this situation be thought of as a Backward-Facing Step flow shape?

Here is a paper on the subject. Does it offer any insight into our concave diffuser discussion? It is beyond my skill set. At least I might have found a description to use for searches.

http://www.flow3d.com/pdfs/tn/FloSci-TN81.pdf

Brian

[mep]

Diffusors like this are nothing new. They are called Carnot diffusor and are used in industry for example to mix fluids. It’s really surprising but they do work. (Xpensive for sure also knows about them.)



Uploaded with ImageShack.us

They have a sudden pressure change which is not wanted. For race cars they possibly create to much drag and are therefore not that efficient as a Bernouli diffusor in regards to downforce vs drag.

When teams really go that route now then maybe to get bag some of the lost downforce by the reduced diffusor exit cross area. They have anyway enough engine power and want to get as much downforce as they can get so they just don't care about increased drag when they can improve over all (car) efficiency.

This doesn’t mean you can freak out now because some of the theories here are, hm lets say, scary. Well at least the topic title fits: Diffuser Confusion.

[n_anirudh]

Well, if it was a 90deg bed there, it would act like some sort of a giant gurney, wouldnt it?

Combining that with the backward facing step, it may actually be there to create an artificial curvature to the flow..