Khamsin Virtual Racecar Challenge 2015

Post here information about your own engineering projects, including but not limited to building your own car or designing a virtual car through CAD.
User avatar
andylaurence
123
Joined: 19 Jul 2011, 15:35

Re: Khamsin Virtual Racecar Challenge 2015

Post

I wonder if OCCFD might be easier if it didn't rely on a bunch of external tools and simply exported an OpenFOAM case that could be uploaded to Rescale. You don't have to worry about tool integration, nor about how much processing power you have - it just runs and doesn't cost very much at all.

User avatar
CAEdevice
49
Joined: 09 Jan 2014, 15:33
Location: Erba, Italy

Re: Khamsin Virtual Racecar Challenge 2015

Post

andylaurence wrote:I wonder if OCCFD might be easier if it didn't rely on a bunch of external tools and simply exported an OpenFOAM case that could be uploaded to Rescale. You don't have to worry about tool integration, nor about how much processing power you have - it just runs and doesn't cost very much at all.
I did not know "Rescale", very interesting!

Sabino
Sabino
1
Joined: 09 Feb 2015, 23:23

Re: Khamsin Virtual Racecar Challenge 2015

Post

cdsavage wrote:A copy of the test car is available here: http://www.mediafire.com/download/sar9z ... xample.zip. Again, this car doesn't conform to the current rulebook, I'm posting it mainly to make it clear how the model should be prepared before submitting.
MadMatt wrote:When saying 10° downward and 20° upward, I am not sure what is your reference point. I will use simple words here because I want to be sure everybody understands, so when looking down the X axis, is it 20° clockwise and 10° anti-clockwise?

In maths I would say +10° and -20° (just to be clear).
Rotation of 0 would mean that the inlet surface is facing directly forward. Rotation 'upward' would mean that the surface now faces forward and upward. I think the phrase 'looking down the x-axis' could be interpreted in different ways, but when the left-hand side of the car is facing the viewer, rotation upwards would be clockwise.
Out of curiousity i downloaded this file, what software did you design this with, im impressed with your blends,...

Sabino
Sabino
1
Joined: 09 Feb 2015, 23:23

Re: Khamsin Virtual Racecar Challenge 2015

Post

double,...
Last edited by Sabino on 16 Feb 2015, 23:45, edited 1 time in total.

cdsavage
cdsavage
19
Joined: 25 Apr 2010, 13:28

Re: Khamsin Virtual Racecar Challenge 2015

Post

CAEdevice wrote:I'm so curious about all your projects :)
To be honest the new rules would not change too much for me and the sidepods would look more realistic.
At the moment the harder rules to respect have been the 10mm thickness and the side impact profiles.
I agree the 10mm thickness rule is a difficult one - we'll be fairly lenient on this one as long as it's not an obvious breach. If you are exporting to STL, netfabb (free version) has a wall thickness measurement tool.

User avatar
CAEdevice
49
Joined: 09 Jan 2014, 15:33
Location: Erba, Italy

Re: Khamsin Virtual Racecar Challenge 2015

Post

Hi, is there any news about the rulebook? I'm going to start modeling the "definitive" car and it would be useful to have the final template (STEP) to assemble it and use as reference.

julien.decharentenay
julien.decharentenay
10
Joined: 02 Jun 2012, 12:31

Re: Khamsin Virtual Racecar Challenge 2015

Post

andylaurence wrote:I wonder if OCCFD might be easier if it didn't rely on a bunch of external tools and simply exported an OpenFOAM case that could be uploaded to Rescale. You don't have to worry about tool integration, nor about how much processing power you have - it just runs and doesn't cost very much at all.
I agree with you - OCCFD would be a lot easier if it was a full blow CFD software on its own. Most of the external tools are for pre and post-processing purposes rather than solver itself.

If you are using Amazon AWS (around $1 to $2 per run), then you need no installation of any of the external tools. I have the impression that people prefer to use their own resources prior to investigating cloud. So far I have focused on Amazon as it was - to my knowledge - the only one that was offering a pay-as-go approach. The other cloud providers requested an up-front payment...

User avatar
CAEdevice
49
Joined: 09 Jan 2014, 15:33
Location: Erba, Italy

Re: Khamsin Virtual Racecar Challenge 2015

Post

A tip about OCCFD: to avoid reaching the 10x10^6 cell limit in most cases it is enough to disable the wing refinement, just inlcuding the wings in the "body" component. Anyway, Julien will solve it in the next release.

MadMatt
MadMatt
125
Joined: 08 Jan 2011, 16:04

Re: Khamsin Virtual Racecar Challenge 2015

Post

For me the issue is airfoil proximity, and moving them further apart gave horrible results (worse than having the gap closed because of the mesh)! :)

User avatar
CAEdevice
49
Joined: 09 Jan 2014, 15:33
Location: Erba, Italy

Re: Khamsin Virtual Racecar Challenge 2015

Post

MadMatt wrote:For me the issue is airfoil proximity, and moving them further apart gave horrible results (worse than having the gap closed because of the mesh)! :)
I use gaps no smaller than 15mm, and I'm not having problems.

MadMatt
MadMatt
125
Joined: 08 Jan 2011, 16:04

Re: Khamsin Virtual Racecar Challenge 2015

Post

CAEdevice wrote:
MadMatt wrote:For me the issue is airfoil proximity, and moving them further apart gave horrible results (worse than having the gap closed because of the mesh)! :)
I use gaps no smaller than 15mm, and I'm not having problems.
I use gaps smaller than this hence the problems, but I hope it can be solved otherwise I will increase the gap but my airfoil profile is such that if I increase the gap I lose a lot, and then I would have to find a profile that is less sensitive to gap.

User avatar
RicME85
52
Joined: 09 Feb 2012, 13:11
Location: Derby

Re: Khamsin Virtual Racecar Challenge 2015

Post

CAEdevice wrote:Hi, is there any news about the rulebook? I'm going to start modeling the "definitive" car and it would be useful to have the final template (STEP) to assemble it and use as reference.
Going by the PM I have received changes are going ahead.
cdsavage wrote:Hi,

Any suggestions on the rule changes I mentioned in the last PM? I'm happy to work with you if they would necessitate a big redesign on your end. If not, I'd like to get these changes published as soon as possible.

Thanks,
Chris
cdsavage wrote:Hi,

After having spoken to a couple of the other entrants, this is a slightly different rule modification that we have come up with:
- lower side impact structure may be moved downwards by upto 40mm
- cooling inlet inner template is 500mm long
- cooling outlet surface must be at least 800mm rearward of the inlet
- cooling inlet outer template 1/2 lengths of 575/325mm

Let me know if this would require you to do a major re-design, if so we can make further changes. If time is a major concern to you, I will ask Julien about pushing the first round back by a couple of weeks at our next planning meeting.

Thanks,
Chris

cdsavage
cdsavage
19
Joined: 25 Apr 2010, 13:28

Re: Khamsin Virtual Racecar Challenge 2015

Post

We will go ahead with these changes:
- lower side impact structure may be moved downwards by upto 40mm
- cooling inlet inner template is 500mm long
- cooling outlet surface must be at least 800mm rearward of the inlet
- cooling inlet outer template 1 length of 575, outer template 2 length of 325mm
I'll update the rulebook ASAP. The guide files don't require any changes, other than the "example" cooling inlet templates.

julien.decharentenay
julien.decharentenay
10
Joined: 02 Jun 2012, 12:31

Re: Khamsin Virtual Racecar Challenge 2015

Post

OCCFD v1.0: I finally got it together and tested. The mesh size (maximum) is increased to 20 millions cells. The new version is the final one - to the exception of bug fixes - no further changes in mesh size, boundary conditions, etc. Except a reasonably big step in term of mesh size - my experience is around an increase from 2.7-3 millions to 4.7-5.4 millions cells... Just letting you know that it has increased the computational cost by around 100 to 150%.

In the right-click menu, there is an option for KVRC (Light) - under the More Options entry. This is for a smaller mesh size. The KVRC is based on the regular KVRC option.

The URL is http://www.khamsinvirtualracecarchallen ... .0-000.zip

Edit: Let me know if you discover any bug/issue...

Edit 2: Thanks Matteo. The link is corrected...
Last edited by julien.decharentenay on 19 Feb 2015, 22:25, edited 1 time in total.

User avatar
CAEdevice
49
Joined: 09 Jan 2014, 15:33
Location: Erba, Italy

Re: Khamsin Virtual Racecar Challenge 2015

Post

Well done Julien! Wich reduction of the total cells number do you estimate switching from strandard mode to light mode?

EDIT: After editing the link (the correct version is: http://www.khamsinvirtualracecarchallen ... .0-000.zip) I tested the "light" run and it works perfectly.

During the night I'll run the standard solver with the same model.