F1technical.net

LVDH wrote: ↑

04 Nov 2018, 14:55

That is probably the switch from first to second order schemes. In the system folder there is a file called "fUpdate" or similar. You can check it out.

Oke thanks, I did not find that file but there was another file in the system folder were it was speaking about "Updating fvSchemes" at time is 560, so I think that it is the scheme switch indeed. For sure I have to learn more about OpenFoam soon…

LVDH wrote: ↑

04 Nov 2018, 14:55

Team_Bart wrote: ↑

03 Nov 2018, 21:34

For the MVRC team, I was wondering for what mistake I got a warning last race or is this a typo?

From reading the message in the post and memory I think you front suspension was visible. Somehow it was clear though that is was supposed to be in a different position. You have to send us the correct position you want the suspension to be in. Next time you mess that up will lead to a penalty. Anyone with warning should keep a list and always double-check the submissions.

Okay, strange but will check this for next race!

Hi, here is the car for the next race.

Finally, I had no time to redesign it, so I went back to the car I used in the first race and I refined it as much as possible: some small details and new cooling ducts. I also checked the rounds inside the diffuser, that were the same since 2017.

Cool... such a nice clean car... it’s going to be a tough challenge if Variante and JJR have made any gains...

machin wrote: ↑

10 Nov 2018, 12:11

Cool... such a nice clean car... it’s going to be a tough challenge if Variante and JJR have made any gains...

Variante's car is completely new, I fear he has a larger margin of growth

If anyone is looking for inspiration there are some interesting details on this Time Attack car...

well car is submitted and with out any penalties i am looking at a 1 minuet 20 second lap time so lets see if that can keep me in 4th or not...i have my doubts.

I'm gonna submit my car tuesday evening if its ok with our staff.

I submitted the car this morning. From the (bad) numbers I am seeing while a desperate simulation is running, it will be the final car for this race. I am happy: it is better than the car I used in the first race, I hope it will be enough.

Looking at the results of the last race, I think JJR is the favourite at Sao Paulo.

After a lot of trouble with running simulations and trying many different concepts that didn't really seem to work, I finally seem to have at least half a decent car .
Cooling should be sufficient, engine intake should be sufficient, exhaust should be fine.
Furthermore, the simulations point towards an improvement in downforce, however there's also a slight increase in drag.

Still, by looking at the numbers of last race, I'd assume that the finishing position should be around the same.

Just submitted my car. I've managed to run 20 simulations since the last race, mostly about optimization. Unfortunately, they've either provided unsatisfying results or improvements in areas that don't need to be improved right now (I'll save those for the future...). An incredibly unfortunate sequence, where it seemed like my car was already optimized before the optimization phase... Also, I'm sure CAEdevice and JJR will be super competitive, but despite all I might still win. I'm hoping for a close race, from everyone's side!

Some notes for Andre':
- Right radiator control surface is the one you gave me to avoid mesh problems. The left one is mine, with a problematic mesh. I've provided it just because the rules impose me so.
- Engine inlet caused quite some problems (simulation related) during the last race. You might want to doublecheck that area so you can run a smooth simulation.

variante wrote: ↑

12 Nov 2018, 00:44

Just submitted my car. I've managed to run 20 simulations since the last race, mostly about optimization. Unfortunately, they've either provided unsatisfying results or improvements in areas that don't need to be improved right now (I'll save those for the future...). An incredibly unfortunate sequence, where it seemed like my car was already optimized before the optimization phase... Also, I'm sure CAEdevice and JJR will be super competitive, but despite all I might still win. I'm hoping for a close race, from everyone's side!

I am convinced that MVRC is becoming a strategy challange about development.

After the first race, I noticed a significant issue inside my diffuser (the flow only passed through the central channel). I spent all my simulations (about 25) between race 1 and race 2 to solve it and it was a waste of time (the car was slower).

After race 2 I decided to renounce to solve that iusse (or to solve it changing the geometry around the diffuser and not the diffuser itself) and I could improve perfomance working on other details (20 simulations + 15 "fast" simulation with custom mesh, and light, setup): rearview mirrors supports where allowed (outside the visibility volume), cooling ducts (despite they are far from perfection), high nose setup, rear suspensions generated vortices reduction...

Unfortunately I had no time to make a nice show car for this round, but I can tell the new ButterFly has much better looks! More important, I found some extra (frontal) downforce without increasing the drag to much. Unfortunately I was not able to test the submitted iteration, but without penalties it should be close to etsmc. See it as the weather condition during the race what the exact time will be
Ps. After reading the posts above I think I need to buy an additional desktop to be able to compete with the front runners.

Team_Bart wrote: ↑

12 Nov 2018, 13:44

After reading the posts above I think I need to buy an additional desktop to be able to compete with the front runners.

Consider that when I started the project of my first LMPx car (maybe in 2015), I needed less than 5 simulations to reach the forth place, then around 50 to reach the podium and (after 2 years) more than 100 to win the challenge for the first time (with much luck). Cpu power and results are not exactly connected by a "linear" function.

I think 4 cores (i5 is ok) and and 16gb of RAM are the perfect solution. I use a virtual machine with 4cores and 32gb of ram and I am very happy. The other cores are used by some fem simulations I do for professional purposes.

Also consider that Machin two years ago developed a very decent car without cfd...

My general adivices are: try to design a basic car with very basic flow around it (you should be able to imagine the flow with mental cfd only). If you prefer try with a separate front wing (some cars used it in the first MVRC editions). Use inlets and outlets about 50% of the hx area for the first attempt. Then focus only on wings and diffuser optimization (a straight diffuser is not completely a bad idea to start with).

This is much more than you can do during first year if you have not previous experience with car aerodynamics and CAD.

But the most important rule is: during development try with only one change for each simulation run.

Yeah, I agree with CAEdevice. You can get fairly close to the front with just a bunch of simulations.

I've run a lot of them so far, beating my previous record by quite a margin. But that's only to gain those few tenths of a second necessary to win the race. Thats what you need to do when CAEdevice and JJR get each year more and more competitive...
(BTW I run an average of 1 simulation per day since the beginning of the season, with "fast" settings. Not so much after all.)

If you can't run many simulations, invest your time differently: study other cars, have a look at wing profiles leaked from F1, and always try to understand the reason behind such geometries. Then start thinking about a simple and linear concept, put sleek teardrop shapes around obstacles (wheels, cockpit,...), follow the natural airflow on top of the car to reduce drag, and force it into certain patterns underneath, to create downforce.
That's what I did during my first season(s), and it paid off quite well.

Once you get access to better hardware, you can run the necessary simulations to optimize your design, with the first place in mind, or to experiment with weird geometries, as I sometimes like to do.

One of the things I like about this championship is that our cars are more optimized (or, generally, better designed) than 99% of actual racecars.

I might write an article on my thread about this "behind the scenes" thing.

One thing I found to help with simulations was additional monitoring surfaces
1 at the floor entry and 1 at the diffuser entry. This way I could see if my changes were getting more or less flow through these areas and the images show me where it is coming from and going too.