[KVRC] Khamsin Virtual Racecar Challenge 2016

[CAEdevice]

Thanks

Jeez man that must have taken a large chunk of your wallet

Jokes aside, so if I get it working i should be able to half the time roughly. So about 14 hours instead of 24...

It is the workstation I use for FEM simulations at work: based on second hand componenets (CPU, motherboard). The most expensive part were (in 2014) the RAM kit (8x8gb) and the Quadro 2000 graphic board.

Most of times, the multicore simulations with OCCFD take 6/7h (instead of 5h) because other simulations (ProMechanica) run at the same time on the machine.

[CAEdevice]

The bottle neck with the simulations of the KVRC is the RAM space: 8gb is not enough. 16 is ok.

[Alonso Fan]

Right, I see. Can it not be optimised to run at lower ram? I only have 8gb

[CAEdevice]

You could try including the front and rear wing into the body component (a finer mesh is used for parts including the word "wing"). I am not using the "front wing" component at the moment.

[Alonso Fan]

Yes, I am currently doing that. I currently have an analysis running so let's see how it goes

[CAEdevice]

What about the light mode of occfd? DR and DF are quite accurate, only the cooling performance is badly approximated.

[RicME85]

Worth mentioning what I found last year....when simulating just a wing profile for the thread Machin and I did, the Test option gave near identical results to the others.

[CAEdevice]

Worth mentioning what I found last year....when simulating just a wing profile for the thread Machin and I did, the Test option gave near identical results to the others.

"test" or "light" option?

[Alonso Fan]

How much does the light option cut time by?

[RicME85]

Worth mentioning what I found last year....when simulating just a wing profile for the thread Machin and I did, the Test option gave near identical results to the others.

"test" or "light" option?

I used the Test AWS option and compared with the full AWS option and it was less than 10N difference iirc

[Alonso Fan]

right, so I ran a successful occfd analysis in parallel. but post processing didn't complete, do I have no values for df/drag etc

how do I get the values from paraview?

[CAEdevice]

The dr/df values are written in the text files of the postprocessing folder (or in the log)

[Alonso Fan]

Thanks, and what about cop?

[CAEdevice]

The COP os not included into the logs. I often have that post processing issue, so I compute it from the force and moments (with an approximation about the vertical position of the drag).

We should ask Julien wich algorithm is used by OCVFD.

[Alonso Fan]

Ok, thanks. I'm running a light simulation to quickly check cop. Looks like I have a slow car. I'm not getting enough out of my diffuser, but I have not sacrificed cooling, so let's wait and see