AratzH wrote: ↑06 Jun 2020, 14:58
How different are your results from the official ones? I am a bit confused. I ran the exact car I submitted with MVRC Long template (5000 iterations) and my results are notably different from the official ones:
Personal Pc:
Cd Cl Cl/Cd Cl(f) Cl(r) CoP Intake Exhaust Cooling
Total 1.305 -2.384 -1.827 -1.179 -1.205 1.718 5.56 -1.2979 3.12139
Official:
Cd Cl Cl/Cd Cl(f) Cl(r) CoP Intake Exhaust Cooling
Total 1.319 -2.431 1.715 4.62 -0.392 2.383
Is there any difference between our solver settings and the official one? This is almost a 1.5s difference in lap time and 4 positions...
Still 1.5 seconds off pace though haha
So, I checked your logs. You seem to have an interesting machine at your disposal. It looks like, we are not comparing the same case though. You have some different geometry from what is in the official run, resulting in different meshes. This explains why you get slightly different Ci values (I suppose the design is not very different in the two cases). The difference in cooling flow is pretty sure an incorrect tessellation, which I fix for everyone automatically.
k.ko100v wrote: ↑11 Jun 2020, 08:56
k.ko100v wrote: ↑10 Jun 2020, 13:52
LVDH wrote: ↑10 Jun 2020, 08:24
Yes, I sent out the reports. I had to create them with offscreen rendering which somehow deactivates AA, sorry for that.
I hope to get some other open issues tackled tomorrow. Currently I am crazy busy. That is why things like the homepage do not get updated. I hope to compile a small list of open issues where people can help. Some of it is coding scripts to automate the MVRC processes. I think, I will create some kind of Git account where people could contribute, if interested.
Not sure what this means. But if you have the exhaust exit surface in both folders they will be overlapping and the result will be arbitrary. This can mean some elements that are and exhaust and some that are a wall. As the BC is defined with a flow rate, this could cause crazy velocities.
I had only one exhaust out file and It stayed in the "vehicle body" folder. Now i moved it to the "special_bc" folder. Should I expect any performance drop, or increase?
So I added the exhaust_out file to the special_bc folder (only) and here are the results compared to the official ones.
It doesn't looks nice, but if it is not affecting the aero performance I am ok.
After loading the case in the GUI, you should be able see visually inspect the surface, I think it should be pink and not blue like the other car parts. You seem to have something else overlapping the part.
machin wrote: ↑11 Jun 2020, 09:44
Joe7218 wrote: ↑11 Jun 2020, 01:14
Hi, when looking at the official results for my car, I noticed the sum of the cl of each individual part was much greater then the total cl given at the top of the document (there was almost a 0.4 cl difference.) there is also a discrepancy between the summed cd values and the total cd value, however the discrepancy is much less (only about 0.08). Does this mean the race results for my car would be wrong? (unless there is supposed to be a discrepancy? thanks
If I remember correctly the list of components below the totals doesn’t include the forces on the heat exchangers... so it doesn’t add up...
I’m not sure why they’re excluded... Andre may be along to comment why soon, but definitely the top line is the correct one to use for the lap time prediction.
Yes, the single parts do not contain the forces of the porous parts. You should also check if really every single part is listed in the report. In earlier versions, I did not list very many parts on purpose, after more and more complaints and questions about this, I included more. I am not sure if every part is now part of the force per part list.
Also GCBB sent me stl files of the cooling monitoring surface. The mistake made there is that there is one stl file for the monitoring surfaces on the right and left hand sides of the car. This is not good, as we only run a half model. This means there is monitoring surface that is not part of the flow domain. OpenFOAM measures strange stuff, when you do that. Newer versions are even crazier, as they do not measure ridiculous numbers anymore so you might actually trust them, which seems to have happened here. You should either only create a surface for the right hand side of the car or have a separate surface for each side.