[MVRC] Mantium Virtual Racecar Challenge 2016

[CAEdevice]

I guess I should have a look at what Chris did, maybe he made a mistake.

No, I checked the model: it is problem with one of my CAD tools (it costs 2000€ per year )

I don't know whym but when I export a component STL it "remebers" the old C_SYS...

[LVDH]

I checked and the files are correct.

Did you use rev2 of the guide parts?
If yes, maybe you imported them incorrectly into your CAD program. I know you are the expert but sometime even the smartest make mistakes.

[LVDH]

Ok, I see. Problem solved.

[RicME85]

Well I have installed OpenFOAM and am currently running the test case, looks like things are working fine at the moment.
How long should the test case take to run?

[LVDH]

It depends on your hardware, maybe an hour. If the solver is outputting a meaningful log you can abort it and get to your car.

[RicME85]

I have built a new PC for CFD.
Have the 8 core FX8320 and 16gb RAM with an SSD.
It has now finished so not bad.
Which are the log files you want?

[LVDH]

The log files from the test case are not so interesting for me at this point anymore. Those of your first car simulation would be good to see.

[RicME85]

Ok, just I saw you wanting more Windows logs.
Not sure if and when I can get a car simulation done however.

[Alonso Fan]

Ok, just I saw you wanting more Windows logs.
Not sure if and when I can get a car simulation done however.

Same here, I'm still designing my ducting

[LVDH]

Do not try to overengineer your first design. Bang out something simple and run your first simulation on it. When you see the results you will get a million more ideas.

[CAEdevice]

The simulation is running! About 600 iterations over 1500 (I think).

Meanwhile, I am trying to find the meaning of the coefficients:

Cm: ?
Cd: Drag coeff.
Cl: Lift Coeff. (negative means downforce)
Cl(f): Front lift coeff.
Cl(r): Rear lift coeff.

faceSource faceSource_engine_intake output:
areaIntegrate(engine_intake) for p = -0.97894

faceSource faceSource_engine_exhaust output:
areaIntegrate(engine_exhaust) for p = -0.124201

faceSource mSurf_cooling_inlet output:
areaNormalIntegrate(sampledSurface) for U = ?

faceSource mSurf_cooling_outlet output:
areaNormalIntegrate(sampledSurface) for U = ? (shuod be equal to the previous item?)




It looks that the results are comparable to the old solver (OCCFD, but not excactly the same (the difference is around 5-10%), as expected.

[CAEdevice]

I noticed that the virtual machine (virtual box) requires very few ram: I assigned 16gb, but the actual memory consumption is less than 1gb (with OCCFD more than 6gb were used).

There is something wrong with my settings?

[graham.reeds]

I noticed that the virtual machine (virtual box) requires very few ram: I assigned 16gb, but the actual memory consumption is less than 1gb (with OCCFD more than 6gb were used).

Wow, if it is that frugal on memory then it may be worth building a Raspberry Pi farm!

[CAEdevice]

I noticed that the virtual machine (virtual box) requires very few ram: I assigned 16gb, but the actual memory consumption is less than 1gb (with OCCFD more than 6gb were used).

Wow, if it is that frugal on memory then it may be worth building a Raspberry Pi farm!

It is not good news: I have the suspect that I choose the wrong settings, it is writing a lot of data on the hd...

[LVDH]

Send me your logs as soon as you are done. 1GB is very little. Are you using the fast option?

• Cm ->pitching moment
• the two first face sources are what they say: The pressure integrated over the specified surface
• The other two result from your cooling inlet and outlet surfaces. They should have the same value. However you will notice that they don't. This is mainly due to interpolation reasons (assuming your cooling duct is closed). The surfaces you use might only be triangulated with only two triangles. If you refine the stl better the results should match better. In the very beginning of the solver.log you can see out of how many triangles the surfaces consit.

On the last point I am planning a good post which will include on how to create a valid closed model. This will have to wait until the end of the week.