[KVRC] Khamsin Virtual Racecar Challenge 2016

[Alonso Fan]

I really don't mind race 2 being moved back a week or two. I have Uni exams and I don't even have time to make my car legal at the moment

I haven't touched it since submission, and won't be making any changes until my exams are up

[CAEdevice]

Hi Chris,

thank you for the update.

Sorry about the lack of input from us. Julien has had to wipe his PC and he won't be ready for round 2 for a bit longer - we'll need to postpone round 2 by at least a week, probably a bit more. I'll announce the new date as soon as possible.

I would have a problem in case Race 2 will be post-poned, beacuse I am leaving for at least a month and I can't do any simulation after next Thursday (the official/original submission date). Indipendentley from the new date for Race two, I would need to know as soon as possibile the new cooling rules, or I will have to submit the car completely guessing the cooling aerodynamics. Considering that there is a proposal to partially not consider the results of race 1, it would be a significative impairement.

For round 1, if we were going to award full points, we would need to find a fix, and then re-simulate every entry. I don't think this is likely to happen in any reasonable amount of time, so we're probably going to need to stick with the numbers we already have. The only decision is how to award points, I would lean towards awarding no more than 50% points, but if there's a clear majority opinion on what we should do, I have no problem following that.

I think that there is no need to re-compute the cars or to invalidate Race 1 ot to reduce points.
The cooling issue (=not physical behaviour of the inlet, caused by the bc) affects only the cars that have not enough pressure on the inlets, they could not use full power anyway. Re-computing the simulations with different bc would be the same as changiging the rule retrospectively. Please consider that the first 4 o 5 cars had acceptable cooling results because their teams simulated and checked it with OCCFD or equivalent OpenFOAM environements. In my case: I excluded a promising design (even superior to Mantium "monster" results about downforce), just because it showed the odd cooling issue.
The situation for teams that have geometry issue is different and not related to the proposed "solution".

The porous option is unlikely to be ready for round 2. Getting rid of the boundary conditions, while still measuring the pressure integrals, is a good possibility for the upcoming rounds if we can verify that it improves things. I'd also like to make a change to the submission process to make the inlet/outlet geometry less of an issue - this would probably involve asking for body.stl to be a single, solid body, and then producing the inlet and outlet surfaces a different way.

I completed a test to compare the two possibilities: "opt1" taken from the KVRC 2016 rules and "opt1b (?)" obtained suppressing the imposed flow bc. The results are encouraging:

1) The original car, has a pressure differential of 50 Pa*m2 without any issue (the average pressure is high enough to avoid local "suction" effect over the inlets)

2) The "opt1b" car cooling capabilities have been measured using the "force resultant" function in openFOAM: the same we can use to have separate results for the wings. I used a customized OpenFOAM configuration, but I think that it would be possible to obtain the measurement in OCCFD without changing anything of the code, just including all the cooling surfaces (inlets/outlets) into a file names as "*_wing_*" and looking into the "other wings" folder after the post processing.

3) The results of the "opt1b" car are comparable with the "opt1" car, considering that the absence of the imposed flow makes the pressure on the inlets higher (and lower on the outlets). I obtained 60N (pa*m2) that is a very reasonable value compared to 50 pa*m2 obtained by "Opt1" car.

4) I run a test modeling a (partially occluding) wall in front of the inlets: the "opt1" car showed the "not physical behaviour" (negative pressure, convergence difficulties, not repeateble results, geometry and meshing sensitivity), "opt1b" car showed a reasonable behaviour (pressure near to zero, no strange effects).

5) The pressure (force) estimation is written into the log files even during the preliminary simulation and not only in the final html report (that often fails to be generated). That is good thing because the preliminary simulation (500 quick iterations) often gives significant information in a short time and with small hardware resources.

The two CAD model are available for the KVRC staff. I can't share them because it is the car that I would have submitted next Thursday.

PS: To avoid problems, I would apply the new simplified simulation model to the engine inlets/outlets too

[HP-Racing]

I hope this isn't a dumb question, but if the BC on the inlet surfaces are suppressed for Opt 1, that would make them walls right? Hence we'd deflect any air reaching the inlet surfaces sideways and significantly disturb flow to the rear of the cars.

[CAEdevice]

I hope this isn't a dumb question, but if the BC on the inlet surfaces are suppressed for Opt 1, that would make them walls right? Hence we'd deflect any air reaching the inlet surfaces sideways and significantly disturb flow to the rear of the cars.

Yes, but this effect is already present: when the cooling ducts work in the right way (positive pressure on the inlets) a part of the flow runs along the sidepods.

[HP-Racing]

Okay. And what if we used a pressure outlet BC instead of velocity on the cooling intake? Wouldn't that be able to simulate the back pressure of the heat exchanger and we could measure the the flow exiting through the cooling intake.

I'm just just asking because I'm interested and would like to learn, not because I think I know better.

[Gridiot]

I hope this isn't a dumb question, but if the BC on the inlet surfaces are suppressed for Opt 1, that would make them walls right? Hence we'd deflect any air reaching the inlet surfaces sideways and significantly disturb flow to the rear of the cars.

Yes, but this effect is already present: when the cooling ducts work in the right way (positive pressure on the inlets) a part of the flow runs along the sidepods.

As you said, only part of the flow is doing that in normal condition, not all, and you can work to optimize that. If the inlet would be considered as a plain wall, I think teams would have to be allowed more time to tweak their design. I my case, I would clearly design the sidepod differently if that was the case (although I do not wish the challenge to go in that direction).

Regarding points distribution for the first round, being a bit selfish and looking at my own issue, I don't think it's fair to lose points because calculation of my COP is wrong. It clearly has a significant impact on my lap time for no "physical" reason. And I was hoping for a fix about that before continuing the development of my car, because, given the fact that the only simulation results I have are the races ones, how can I set the development route if I don't know the actual COP. If I look at the numbers, I should add loads of rear downforce, but it's clear from the analysis I can get from Paraview that the car is actually more balanced that results said. Therefore, apart from slight shape optimizations, I still don't know where to start developing.

[machin]

Has Chris or Julien explained the reason for the COP issue? I don’t think we've had that before... Is it linked to the Boundary Conditions issue?

[Gridiot]

No explanation yet. They investigate the issue, but I believe they also had a lot of work with this BC issue.
My problem is not related to it, nothing to do with inlets/outlets. If I remember correctly, Chris confirmed in his post that it was a totally separate issue.

[variante]

For round 1, if we were going to award full points, we would need to find a fix, and then re-simulate every entry. I don't think this is likely to happen in any reasonable amount of time, so we're probably going to need to stick with the numbers we already have. The only decision is how to award points, I would lean towards awarding no more than 50% points, but if there's a clear majority opinion on what we should do, I have no problem following that.

The porous option is unlikely to be ready for round 2. Getting rid of the boundary conditions, while still measuring the pressure integrals, is a good possibility for the upcoming rounds if we can verify that it improves things. I'd also like to make a change to the submission process to make the inlet/outlet geometry less of an issue - this would probably involve asking for body.stl to be a single, solid body, and then producing the inlet and outlet surfaces a different way.

If there's a chance that the cooling related problems will be found and solved, then we can just re-simulate the entries that had problems (or the whole lot). I would have no problems waiting for that moment. This would be the best solution.

If the solver's problems know no solution (which sounds unlikely to me), then I think that a new race should be run as the last race of the championship under new and effective cooling rules; this past race would be forgotten. Still on the Nurburgring.

I would exclude the half point decision, even though it's the simplest.

I understand that my point of view is pretty radical, but it would preserve the honor of the victory, and fully recognize the effort behind it.

No problems getting rid of the imposed flow for the next round.

[etsmc]

Been working on the car ready for the next race and one are that I was looking at was the diffuser at the front and rear.
for the development of the rear diffuser I have come up with a design that in my head should work but coming from someone with no aerodynamic knowledge what so ever I wanted to show some of the people that clearly know more than me and see if they can explain what's going on.

a slice is taken in the x-axis with the origin set to 0.5

this shows the concept I thought of where a cavity at the (not sure if this is the right term but) throat of the diffuser would cause an area for a large mass of air to go and help pull air out from under the floor.

a slice is taken in the z-axis with the origin set to 0.07

This image is just to show that there are no fences in the diffuser to direct the air in any way at the moment but the air flow does seem to be reasonably straight at the moment so some improvements look possible.

do you think with further development it might have legs??
I have tried running it through OCCFD with all the parts separated but I get divergence issues so the light test was run with all the parts in 1 STL file meaning I get no report at the end.

[machin]

Looks a bit like this "Vortex cell" (upside down of course):-

[LVDH]

I have seen stuff like that before. And allegedly it works. For me it is a bit implausible.

On the car diffuser we can now see if it does. All we need is the exact same simulation including images but with a closed vortex cell. That will give a first starting point in understanding what it does. And by the way. There is an unwritten rule in CFD: The flow goes from left to right.

Meanwhile I stumbled into this:
https://www.linkedin.com/pulse/aerodyna ... le-comment
Maybe those guys need some help from the KVRC people. 230kg of downforce do not sound that impressive.

[andylaurence]

If you're going to help a real racing team .... Ahem!

[machin]

Did you ever sort out your scanned model Andy?

[variante]

230kg of downforce do not sound that impressive.

It's 460kg for the full car. Anyway, it looks like they've got greater constraints than us...see the width of the gearbox/differential enclosing bodywork, for example. The wheelbase is also shorter: 2.65m VS 3.0m. Maybe the width too.

...On the other hand...

If you're going to help a real racing team .... Ahem!

...with these wings...


...I wouldn't dare to call myself "a real racing team", unless "real" is just referred to the fact that they build what they design.

I hope that front wing is just a temporary solution to have an idea of the influence of a front aerodynamic element (which would be bad anyway, as you don't want to design an entire car around a temporary solution).
The rear wing is designed by someone who knows about general laws of aerodynamics but has no experience on cutting edge racecar aerodynamics.
The airflow under the diffuser seem to be a bit messy, and the vortices under it not stable enough; I can't swear this, anyway...I would need better images.

I've been a bit harsh...I know. But I always hated unjustified deification. Sometimes, even a highschool student can do a better job than an old carapace (not saying that this is the case).

Still, I hope those guys will achieve good results, have fun, improve their performance and accept constructive criticism.