Diffuser Profile - Convex or Concave?

[Tozza Mazza]

Very little info out there on the web, and pictures of Diffuser profiles are hard to come about.

Which shape is best for the aerodynamics of the diffuser, and why?

There was some info in the diffuser confusion thread, but it was so clogged up with babble and opinion I couldn't make sense of it.

[MIKEY_!]

It's the grey bit. Odd shape I know. Its to get the greatest expansion (hopefully) without stalling.

[ringo]

Chances are it's straight and the top surface is convex.
I haven't really got a good look at it. This question was on my mind as well.

[hardingfv32]

To get some idea of the complexity of the subject you propose, review this article on a F1 intake diffuser.

Design Optimization of a Two-Dimensional Subsonic
Engine Air Intake

http://www.soton.ac.uk/~nwb/lectures/Ae ... 23-118.pdf

Brian

[MIKEY_!]

Seems to match the scarbs pic. Hope that helps.

[ringo]

It's very interesting. Almost counter intuitive.
Wonder how they will look without the blown floor.

[raymondu999]

I haven't really got a good look at it. This question was on my mind as well.

How did you decide for your design then?

[hardingfv32]

Analysis the diffuser profile subject:

What are the design limiting issues that the profile introduces to the diffuser?

Flow separation for one? How does the present exhaust outlet location help prevent flow separation? The outlet is almost parallel to the front of the diffuser floor and angled away. Does the the low pressure created by the exhaust vortex help with flow separation at the roof of the diffuser?

I do not see how the exhaust is related to the roof profile.

Brian

[shelly]

@tozza: read the thread on the exhaust blown difuser and the thread diffuser function: there is a lot of info in that.

For downfoce you do not want a smooth concave diffuser as someone may think, but an abrupt change of direction on the kink line (as abrupt as separation allows).

[DRCorsa]

I would like to contribute -as i can- by presenting my own little research on the subject. I have designed a traditional carbon concave diffuser for a 7-type kitcar and i am in the process of constructing it.
I have "developed" this using CFD software and the final results was about 540N of downforce. I attach 3 pictures from the CFD results.





After looking at some pictures from the current F1 diffusers i decided to make some runs of a concave diffuser on CFD. I have not put a great effort on this yet, this is a basic and "fast" analysis. I have not even included any vertical strakes but i think that this run could shed some light on the subject.
Here are 3 pictures from the convex diffuser results.





CONVEX (TRADITIONAL) DIFFUSER
Downforce: 546N
Drag: 85N

CONCAVE DIFFUSER
Downforce: 394N
Drag: 47N

The air velocity was 50m/s (180kph)

First of all, we see that the drag with the concave diffuser is alost half!
Downforce is less but this could be improved with more development.

Looking at the static pressure images, i would say that for the traditional diffuser the pressure is more evenly distributed along the length of the diffuser, while on the concave one, the negative pressure is more "concentrated" on its front (horizontal) part.

Your thoughts?

[shelly]

My thoughts:

1) the comparison is biased by the fact that you have a lot of fins in the first diffuser. Especially the two central fins have a very strong effect, because they are protruding from the flat plane. You can see that the central part of the floor has a significantly lower pressure compared to the side channels because of this.

2) in order to better compare two cases, you should have the same color scale for both, instead of autoranging. So same color in the two pictures would mean same pressure. You could use cp instead of pressure, and set it from -2 to 1in both images for example

3)probably the concave floor is stalled (no significant pressure gradient along the diffuser visible)

4)even in this first attempt concave diffuser, suction peak on the kink line is clearly visible - lowest pressure in this case is significantly lower than minimum pressure in the other image

[DRCorsa]

My thoughts:

1) the comparison is biased by the fact that you have a lot of fins in the first diffuser. Especially the two central fins have a very strong effect, because they are protruding from the flat plane. You can see that the central part of the floor has a significantly lower pressure compared to the side channels because of this.

2) in order to better compare two cases, you should have the same color scale for both, instead of autoranging. So same color in the two pictures would mean same pressure. You could use cp instead of pressure, and set it from -2 to 1in both images for example

3)probably the concave floor is stalled (no significant pressure gradient along the diffuser visible)

4)even in this first attempt concave diffuser, suction peak on the kink line is clearly visible - lowest pressure in this case is significantly lower than minimum pressure in the other image

1) I know that the comparison is biased because i had no much free time to spend on the concave diffuser. But i will try to run it with vertical fins.
I have protruded the outer fins so forward because the flow tended to concentrate on the central part and leave the outer parts of the diffuser somewhat "ineffective". By placing those long fins there, i tried to "keep" the flow attached to the outer part of the diffuser.

2) You are absolutely right. Fixed now.

3) It actually seems that on the concave surface we have some lift force (red colour). But maybe the fact that this diffuser stalls is an explanation for the much lower drag? Or perhaps this is the F1 concept about? Just a thought.

4) I will try different radiuses on the kink to see the effect.

[DRCorsa]

I concluded a run with the same concave diffuser but with 4 vertical strakes added, similar to those of the first convex diffuser. The results were:
Downforce: 445N (394N without the strakes)
Drag: 58N (47N without the strakes)
But, the concave is still behind in terms of absolute downforce levels.

[shelly]

I suggest plotting pressure coefficient instead of pressure; also it could be useful to see velocity magnitude on the surface (if you can select it in the first cell, otherwise you will get 0 for no slip) or in a longitudinal section - to see it there is stall.

As for downforce, probably the simulation gives stall as a result: you should try with a less aggressive profile (slightly bigger radius on kink line and less angle for the initial part of the diffuser)

[Tozza Mazza]

Interesting stuff.

I reckon you could get the concave diffuser up to the DF levels of the convex one, with much lower drag!