[MVRC] CCE GP001

[CAEdevice]

On my car the front wing itself is not so much the problem, as the front wheel suspension members. They all have separation on the top and are the real cause of cptot loss of the cooling air.
if only one was allowed to angle these a couple of degrees...

It happens the same with my car. Also to lower the sospensions (maybe 50mm) would help.

[machin]

So, the last round is upon us.... and I never really felt like I got anywhere with the car... too much time trying to sort out that damn cooling! Indeed, the biggest change on this last version was also to the sidepods... the "top vents" are now removed again, and the internal "double-decker" ducts (maybe the eagle-eyed amongst you spotted them last time) in each duct are now adjusted, and visible through the cut outs in the side of the sidepods...



The only other change is to the diffuser profile... now featuring two "expansion zones" as spotted on some of the front runners...

The wings are clearly at least 30% down on downforce compared to the front runners... but I didn't change the front because at the moment the COP is too far forward, and I didn't change the rear because I am hoping the sidepod change will improve the rear wing downforce without requiring any changes to the wing itself.... time will tell...

[variante]

What's the purpose of the double decker ducts?

[machin]

What's the purpose of the double decker ducts?

Well it’s really just the result of adding a single long guide vane in the duct behind the heat exchanger... I think it looks like you had one there for COTA too (?), but mine is much longer, and inside a longer sidepod... so it effectively splits the duct into upper and lower portions... the reason for the longer sidepod is to try and limit the interaction with the rear wing...

[machin]

As many will have seen we have decided to keep the MVRC rules static for this year:-
viewtopic.php?f=14&t=28925

The final version of my 2019 car can be downloaded as a starting point by anyone who wants to use it for 2020.

Download CCE 2019 Round 5 here

So I thought I would post some more CFD images of the last version...

Even without the top vents the cooling performance was still way higher than needed ; flow 3.8M^3/sec (3.0m^3/sec required), and it no longer impinges on the rear wing. This should mean that you can put a really big front wing on and it will likely still make full power.



Performance of the 3 "primary performance areas" (Front Wing, rear wing, diffuser) was still way less than needed. So in the interest of community spirit, to help people who are just starting out, maybe we can get some tips from the championship front runners on how to read the CFD plots to improve these parts?

Let's start with the diffuser:-

The CFD images on the left are from the official MantiumFlow post processing pack (every competitor gets 588 CFD images after each race!), showing surface pressure and air velocity near the wall. The image on the right shows the profile of the diffuser. I've numbered the zones to make it easier to identify the different parts. Making reference to these images can anyone suggest what improvements to make and why?

[CAEdevice]

I think it is almost impossible to keep the flow attached in the central section, the transition is too sudden in my car too (almost identical).

[Koldskaal]

You certainly have to work around the sudden step created by the plank. I think the separation bubble here is unavoidable. One probably won’t get good results trying to apply a lot of suction directly on that bubble. A bit like increasing aoa on a wing that’s already stalling, doesn’t make a lot of sense. It is not impossible to make the center work though. You just need to keep the bubble “contained”.

I think I shared this U-near-wall picture after the final race last season (thinking it wouldn’t ever be relevant again). Notice that my car from last year only really had separation right after the plank.


Another thing: I think you could make the part attached to the stepped plane a bit more aggressive. Especially in zone 1 and 2. I would happily “sacrifice” zone 5 to make room for a more aggressive 1 and 2. (is the illustration on the right the actual cross-section?)

Also, it’s easy to feel like Adrian Newey if you only show cp down to -0.5 . Let’s just say, the surf_cprange_1 pictures (cp: -0.5 to 0.5) should be useless for evaluating the diffuser. See pic:


EDIT: Maybe have a look at this article by Willem Toet: https://www.racetechmag.com/2017/08/wil ... diffusers/

[machin]

Another thing: I think you could make the part attached to the stepped plane a bit more aggressive. Especially in zone 1 and 2. I would happily “sacrifice” zone 5 to make room for a more aggressive 1 and 2. (is the illustration on the right the actual cross-section?)

Thanks for the reply!

The image on the right is the actual profile in the side channel (i.e. looking from the side of the car).

So what feature(s) of the CFD plots are telling you that zones 1 and 2 aren’t aggressive enough? Or conversely, what features would be present if those areas were too aggressive?

By “more aggressive” do you mean a smaller radius in Zone 1, and a higher ramp angle in zone 2? Or the same ramp angle in zone 2 but a longer length (meaning zones 4 and 5 would be shorter in x-direction, and higher above the datum plane (Z-direction)?

[machin]

Maybe I can partially answer my own question by referring to the central section of the diffuser; here we see an area which only achieves CP -0.25 (green), whilst the speed is dark blue u= 0 to 5m/s), indicating a stalled area; showing the Ramp angle/curvature is too aggressive, right?

Compare that to the side channels of the diffuser, where there is no dark blue, indicating that the flow isn’t stalled anywhere, therefore it might benefit from being a bit more aggressive? And if I change it and I get some dark blue occurring that means I have gone too aggressive?

[variante]

Ideally, zone 1 should have a smaller radius, zone 2 slightly steeper, zone 3 looks ok, zone 4 higher from the ground, zone 5 less tall but still with a good upward kick.
In other words, you should promote the concave zones rather than the convex. That's because you want to maximize air expansion right away and exploit ground effect from the flat floor, therefore you have to work nearby the diffuser's throat.
Easier said than done, because i'm basically suggesting you a more aggressive diffuser which, however, requires big vortices to not stall.
Even in F1 we can still see different approaches, with much varying degrees of "aggression", meaning that beyond a certain geometry consistent gains become hard to achieve. I wouldn't be surprised if you'd get better performances from a less aggressive diffuser rather than from a more aggressive one during your first attempts. You'd have to improve 3D flows in order to make actual progress.

[machin]

Ideally, zone 1 should have a smaller radius, zone 2 slightly steeper...

Cool....So can you tell this from the two CFD images? What parts of the images show this? Is my post above more or less correct? Or should I be looking at different plots?

I’m trying to take the guess-work out of the refinement process!

[variante]

Ideally, zone 1 should have a smaller radius, zone 2 slightly steeper...

Cool....So can you tell this from the two CFD images? What parts of the images show this? Is my post above more or less correct? Or should I be looking at different plots?

I’m trying to take the guess-work out of the refinement process!

Nono. I can tell that because i know how F1 diffs look like (even though, again, they can vary a lot), and i'm using those shapes as archetypes. BTW you can find nice images here and there in the F1 cars threads.

Just looking at you images, i can only say that there is only one vortex along the inner strake, not so powerful. The diffuser is not stalling, so you can push it a little more. But, for bigger gains, you'd have to tweak the strakes (more outwash?) and the diffuser's concavity.

To be honest, i'll also need to work a lot on the diffuser if i want to keep being competitive during the next MVRC season...

[Koldskaal]

variante explained pretty much exactly what I meant by more aggressive. Smaller radius, steeper angle etc.

Cool....So can you tell this from the two CFD images? What parts of the images show this? Is my post above more or less correct? Or should I be looking at different plots?

Judging by the cp picture, I would estimate the peak suction to be around -1 to -1.5 in zone 1. This would make the design quite conservative. I don't think doubling the peak suction is unrealistic.

It is of course a bit uncertain judging only by two pictures, we don’t know how clean the air entering the diffuser is. If this is the front tire wake, then you may already be stressing the air as much as it can take.

One thing I have found useful is to look at curves like this: Concave produces low pressure, smaller radius -> lower pressure. Convex produces high pressure, smaller radius -> higher pressure. This is not true in every instance, but I think it is useful for eyeballing a starting point or how to improve.

It seems like you use a spline, or something similar, to define the shape. I simply use circular segments (in 1, 3 and 5) and straight lines (in 2 and 4). This gives me direct control of the radii and slope angles. It probably isn’t optimal, but I think it makes optimization easier.

[machin]

It seems like you use a spline, or something similar, to define the shape. I simply use circular segments (in 1, 3 and 5) and straight lines (in 2 and 4). This gives me direct control of the radii and slope angles. It probably isn’t optimal, but I think it makes optimization easier.

Yes, I agree, actually I have already changed to that approach with straights and arcs, and it is something I did in the past on previous years; honestly this car was thrown together as a starting point for others to use. I expected them to just delete my wings/diffusers, etc to start with their own.

It is of course a bit uncertain judging only by two pictures, we don’t know how clean the air entering the diffuser is. If this is the front tire wake, then you may already be stressing the air as much as it can take.

So is there another image I should be using to determine how clean the air is? What I was hoping we might be able to tease out of you guys , is something like this:-

"Looking at CFD Plot "X", there is a "Y" coloured patch [here], this indicates that "Z". You should change geometry "A" to rectify this".

Surely it is possible to read the CFD plots in this sort of way?

Isn't the central portion of the diffuser the perfect example? My attempt:

"Looking at the "u near wall" CFD plot below, there is a big dark blue patch (low velocity) centrally after the plank, this indicates that the flow has stalled, you should reduce the angle/radius of this portion in order to try and avoid flow separation here, or introduce a vortex upstream to re-energise the flow and keep it attached."

Or would you say that statement is wrong?



If there are CFD plots that would be more useful in the diagnosis of problems, that aren't already in the package Andre sends over, what would be more useful? If anyone can name them I'll see if Andre could add them.

I refuse to just accept the CFD plots as just pretty pictures!

[Koldskaal]

I think what we are seeing is a conflict of interest . I want to share my thoughts partly to give people a nudge in the right direction. Partly so that more knowledgeable people will correct me when I am wrong about something. But there are two reasons why I am holding back slightly:
1. I don’t think anyone wants the solutions to be “formulaic”. We all want to see creative solutions to these problems.
2. The obvious one, I want to win. I am going to keep some ideas private.

"Looking at CFD Plot "X", there is a "Y" coloured patch [here], this indicates that "Z". You should change geometry "A" to rectify this".

The example you give is sort of correct, but I think that an: if x then y approach is too simplistic. Yes, you could definitely decrease the slope in the central region to a point where there was no longer stall. But if you look at the cars from the last race: WB, variante, Mercury, JJR, CAEdevice and me have all seemingly arrived at the same conclusion: Inject fresh air into the diffuser. Work the fresh air instead of the separated bubble. This solution is hard to draw from a simple if x then y.

I am not saying this is the optimum solution, in fact I can think of a potentially better solution in this area that this basically “blocks”. I think in general it is more useful to clarify the actual problem, and make competitors come up with their now solutions.

My point is: pointing to the problem is more interesting than suggesting a solution. There are "universal truths" but they are basically: stalling is bad, don't let the rad exhaust hit the rear wing, etc.

Are you going to continue this “series” with front and rear wing? I think you may have started with the hardest one!

If there are CFD plots that would be more useful in the diagnosis of problems, that aren't already in the package Andre sends over, what would be more useful? If anyone can name them I'll see if Andre could add them.

LIC! Line integral convolution everywhere! I think everyone should be looking at LIC slices.