F1technical.net

Could it work? Almost certainly.

Would it work right away, first time? No way.

All these different configurations have different strengths and weaknesses, but none work straight out of the box, even if you apply good aerodynamic principles to them, essentially because it is almost impossible for us to "calculate" airflow from multiple objects in our head. We've seen this with your car over the past few pages and we saw it with LVDH's little sidepod study: the first iteration is almost always a poor performer and the good performances only come with step by step iteration of the model following a review of the CFD results from each stage... And even then there will be dead-ends where you make a "mistake" and have to go back (hint: save every iteration of your model!)...

When Variante unleashed his first car, how many times did he run it through CFD before that race? I'd wager quite a few... (care to tell us Variante?). The other front runners all have very different concepts, and none have really strayed far from their original concept... They've just refined, refined, refined....

Your round 2 entry was a good example of this... In concept it was almost identical to Variante's class leading car (it is quite difficult to see any differences between the two). The difference in performance was simply down to refinement... And Variante went on and further refined that car for round 3...

I have no intention of going that route whilst we are working away on the current one, was just throwing the idea out there whilst everyone was coming up with solutions to the covers.

As for your iteration comment....you should see my Dropbox, it has every version of the car to date which is a lot

Got all excited thinking that I had results of the latest test but seems divergence was detected
I have no idea why this happens. I can use the exact same model with one area swapped out and it can be a simple shape that has been extruded along a path yet OCCFD throws a hissy fit

RicME85 wrote:Got all excited thinking that I had results of the latest test but seems divergence was detected
I have no idea why this happens. I can use the exact same model with one area swapped out and it can be a simple shape that has been extruded along a path yet OCCFD throws a hissy fit

Try the "very stable" option of OCCFD: it solves the divergence issue most of times, and the results are comparable to the official ones (less than 5% difference).

Cheers Matteo, have just sent the files to AWS so fingers crossed.
Whilst that is working away I will try and refine the mesh in some areas where Sketchup has messed them up through C+P and going in and out of Groups.

@machin and Ric
It took me about 30 simulations just to get to the first iteration of the car.
However, it took me just 2 simulations to reach half of the final downforce. I want to underline this aspect because, before trying with CFD anything that pops in mind, we have to think about what we're doing and about its consequences.
For example, it was not impossible to predict that the suspension cover we've discussed about would mess up the flow...
Obviously, the greater the experience with aerodynamics, the more reliable the predictions.
One last consideration: The car I designed for the first season of KVRC was designed exclusively by eye and intuition; also, I had no experience with aerodynamics not I had ever touched a CFD. Still, the car managed to reach 2nd place in the championship and grab some victories. I'm convinced that pretty much everybody here could do the same.

variante wrote: It took me about 30 simulations just to get to the first iteration of the car.
However, it took me just 2 simulations to reach half of the final downforce.....

As you go on to say, you had come off the back of a really good previous KVRC season, where no doubt you learnt a lot, so you are pretty good at producing a good design within a coupe of iterations, but even then you managed to double the downforce with further refinement! For someone not so good at getting it right first time you're probably looking at having the potential to achieve somewhere near 3 or 4 times the initial downforce after multiple refinements... the bottom line being that newcomers shouldn't be disheartened with an initial "disappointing" result... and not to expect a class leading car straight away....

About 20 simulations for me for the first realese and about 5 for every evolution step (one race = one step, most of times).

...and roughly what sort of aero coefficients did you get with your very first attempt?

machin wrote:...and roughly what sort of aero coefficients did you get with your very first attempt?

First attempt (only a studio, not rule compliant): df/dr = 2.9
First release (first race, after 20 simulations): df/dr = 3.5
Next releases (race after race): df/dr = 4.0 - 4.5 (it depends on the setup)

Last race (after about 20 + 5x5 = 45 simulations): df/dr = 4.75 (I hope so )

Results are in!
Keep in mind that I havent been able to use the same OCCFD setting as usual (KVRC AWS) but instead followed Matteo's advice and used KVRC Stable AWS,

That numbers speak for themselves! you're ready for the podium

Improvements have been massive.
Should make a good guide for any new folk next year