From MVRC Staff asked 4 years ago
Hi I’ve some questions about the input files and the post rpocessing results. For my simulations now, my input files are the whole (mirrored) car. I was wondering if it is possible to decrease the simulation time by only using half of the car? Secondly I now create multiple .stl files for my vehicle body, resulting in cd/cl values for all parts seperately. Is it quicker/beter to merge all parts to 1? After the simulation I noticed the cd/cl values are not always the same for the same parts left and right, I think this is quite strange, since those parts are exactly the same and the difference is in some cases significant. Also there is a couple of parts which all have exactly the same values, while I think this is very unlikely. I used the long template for this simulation. At last I noticed some strange behaviour in the residual and flux plots, probably this clarifies more about this values. During the first 1400 – 2000 iterations it is quite a smoot line and it is asymptotically reaching a certain value. However at a sertain moment they start fluctuating a lot and show some strange behaviour in my eyes. At last I have a problem which is mentioned in a earlier question as well, postprocessing does not show CoP and Cl(f) and Cd(f).
MCAE Support answered 4 years ago
Ok, this is more than one question, at least for the most part these are easy to answer:
I was wondering if it is possible to decrease the simulation time by only using half of the car?
No, simulations in MantiumFlow MVRC Edition always use the symmetry plane with Y=0.
Secondly I now create multiple .stl files for my vehicle body, resulting in cd/cl values for all parts seperately. Is it quicker/beter to merge all parts to 1?
The additional values are only for your convenience and add additional information. Merging all into one file would speed up the simulation by maybe 0.5%. I do not think it is worth that considering the info that gets lost. You could also manually modify the forceCoeffs file in the system folder if you do not like this behavior. But again, it is probably not worth it.
After the simulation I noticed the cd/cl values are not always the same for the same parts left and right, I think this is quite strange, since those parts are exactly the same and the difference is in some cases significant. Also there is a couple of parts which all have exactly the same values, while I think this is very unlikely. I used the long template for this simulation.
That is indeed odd, as parts on the left hand side of the car never make it into the simulation. Are you sure that the parts are separated properly? Also parts with the exact same value sound strange. Possibly these are also not really separated like you intended.
During the first 1400 – 2000 iterations it is quite a smoot line and it is asymptotically reaching a certain value. However at a sertain moment they start fluctuating a lot and show some strange behaviour in my eyes.
That can happen in CFD and is bad luck and annoying. You can send me your case so I can check if I have to make the numerical schemes more robust.
At last I have a problem which is mentioned in a earlier question as well, postprocessing does not show CoP and Cl(f) and Cd(f).
Again a bit odd, assuming you are using the 2018_7c version.
You are welcome to send your case, so I can check it myself.
From MVRC Staff answered 4 years ago
Hi thanks for the quick answers! I figured out the strange values, this was my own wrong interpetation (I opened it on a bit to small screen probably…). All ”mirrored/left” parts have the same values and due to that indeed they are different from the right hand side parts…. Anyway still a bit strange probably. For the other things I can send you my case if it is useful.