[MVRC] Mantium Virtual Racecar Challenge 2018

[LVDH]

Yes, MFlow has a nice GUI but I prefer to launch from command line.

Good man, that is how the software was supposed to work. Setting the rake with a command is currently not possible but I will add this option. First it looks like I have to check out some bugs though, so this will not be available until next week.

I am sure Adrien Newey knows about MVRC and he would not be afraid to compete

Against you guys he would at least have a good challenge.

so I have a simulation running at the moment using blueCFD-Core-2017-2 and the latest MantiumFlow, the only geometry I have is wheels in the frt_wheels and frr_wheels folder and the rear wing in the high res folder.
It seems to be taking a lot longer...
Time = 19

smoothSolver: Solving for Ux, Initial residual = 0.18640888, Final residual = 0.00040734291, No Iterations 2
smoothSolver: Solving for Uy, Initial residual = 0.205356, Final residual = 0.00025854882, No Iterations 2
smoothSolver: Solving for Uz, Initial residual = 0.27623981, Final residual = 0.0003475833, No Iterations 2
GAMG: Solving for p, Initial residual = 0.44512728, Final residual = 0.004415026, No Iterations 16
GAMG: Solving for p, Initial residual = 0.0017387356, Final residual = 1.3195199e-005, No Iterations 17
time step continuity errors : sum local = 2.8274649e-013, global = 7.4193289e-016, cumulative = 9.599932e-016
smoothSolver: Solving for omega, Initial residual = 0.0039757704, Final residual = 3.0044545e-006, No Iterations 2
smoothSolver: Solving for k, Initial residual = 0.33430139, Final residual = 0.00036918317, No Iterations 2
ExecutionTime = 1354.347 s ClockTime = 1354 s

also cm,cd,cl etc are all 0

previously my execution times were mear seconds. anyone else got a simulation to run?

This does not sound good. I will send you a link to upload your case so I can have a look. First you should check the mesh and make sure that the car body is not meshed inside. That have a have created a huge mesh and would explain why it takes so long now.

Update: it looks like blueCFD-Core-2017 is required, but 2017-1 or 2017-2?

This is new. I did not know about this new version. I would assume that it will not make a differnce to just go for 2017-2. I will have to check though.

Hi!

Nice to get going again!
Will we use same OF version as last year? And what about fvSchemes, fvSolutions and DarcyForchheimerCoeffs? Are there any changes from last year ?

This year we are up to OF 5.0. This is because of the awesome blueCFD version for Windows which has been updated to the 2017(-2) version which corresponds to OF 5.0.

[Team_Bart]

Hi Team_Bart, the images seems to be correct. Have you tried with a sligthly smaller angle (59°)?

An image with the section of the car would help.

Yes, tried for 50 degrees as well but the same error is there. I used the case where the HEX is leaning bachwards btw. I can show an image with the car tomorrow (since I have no acces to the file atm) if it is not clear yet.

I tested the rotation last year (45°) and the software worked (even if the results were not good). Let's take a look to a local section tomorrow.

https://ibb.co/mdO7xT
Here a sketch of the housing I tried, I found out that for a small angle (if the angle which now is 60deg is 10deg) there is no error. The angle is with an exact horizontal plane (in part and assembly). The HEX is the grey part.

[CAEdevice]

Good man, that is how the software was supposed to work. Setting the rake with a command is currently not possible but I will add this option. First it looks like I have to check out some bugs though, so this will not be available until next week.

Ok, take your time, it is not a problem to use the GUI

[LVDH]

Yes, tried for 50 degrees as well but the same error is there. I used the case where the HEX is leaning bachwards btw. I can show an image with the car tomorrow (since I have no acces to the file atm) if it is not clear yet.

I tested the rotation last year (45°) and the software worked (even if the results were not good). Let's take a look to a local section tomorrow.

https://ibb.co/mdO7xT
Here a sketch of the housing I tried, I found out that for a small angle (if the angle which now is 60deg is 10deg) there is no error. The angle is with an exact horizontal plane (in part and assembly). The HEX is the grey part.

I read this whole thing twice now I do not quite understand if there is an issue or open question related to MantiumFlow. If so I can send you a link to where you can upload your case, so I can have a look at it.


I found a bad bug which affects Windows users. It is now fixed and I will send out the new version of MantiumFlow tomorrow morning. Right now I do not think that any Windows user can use the software. So you all have time working on nice renders for the team section on the homepage.

[Team_Bart]

I tested the rotation last year (45°) and the software worked (even if the results were not good). Let's take a look to a local section tomorrow.

https://ibb.co/mdO7xT
Here a sketch of the housing I tried, I found out that for a small angle (if the angle which now is 60deg is 10deg) there is no error. The angle is with an exact horizontal plane (in part and assembly). The HEX is the grey part.

I read this whole thing twice now I do not quite understand if there is an issue or open question related to MantiumFlow. If so I can send you a link to where you can upload your case, so I can have a look at it.


I found a bad bug which affects Windows users. It is now fixed and I will send out the new version of MantiumFlow tomorrow morning. Right now I do not think that any Windows user can use the software. So you all have time working on nice renders for the team section on the homepage.

My question was if the showed HEX orientation is correct/legal. According to the rulebook I think it is, but when loading the design in MantiumFlow it says the HEX is rotated incorrect/rotated more than 60 degrees.

[LVDH]

Now I understand. I checked the code and it looks good but it is in theory possible that the detected rotation is something like 60+n*180 deg. I almost assume you run into this unlikely issue. Please send me your HX so I can see why you seem to be the first to run into this issue.

[LVDH]

Ok, I fixed the issues reported so far. The software should now work under Windows again and hopefully calculate the HX rotation a bit better. I have also fixed the gearbox issue in the template files.

[Team_Bart]

Ok, I fixed the issues reported so far. The software should now work under Windows again and hopefully calculate the HX rotation a bit better. I have also fixed the gearbox issue in the template files.

Nice, the HEX error is fixed now! Thanks for the support

[LVDH]

Cool good to hear. It turned out that instead of measuring 60deg he could come up with -120 and complain because of that.
Also keep in mind that warnings are just for your information. If you are sure about what you are doing just proceed.

[CAEdevice]

I tested MFlow 2018 with BlueCFD 2017-2 (very very fast template, only 200000 cells) and it works perfectly on Win10. Welcome (back) automated post processing!

[RicME85]

I have just signed up. Gone back to the old team name of RicME Motorsport. Have included a picture of last years car (from the MVRC website) with the main sponsors 'shopped onto the car.
The team is registered for Intro but I dont know whether to go back to Full or stay and try retain the Intro title.
In the off season I change my PC from an FX8320 to a newer Ryzen 2200G with faster RAM plus NVMe boot drive with standard SSD for applications, Im hoping that when I get the software installed tomorrow that a test run will show the outlay was worth it

[CAEdevice]

I tested MFlow 2018 with BlueCFD 2017-2 (very very fast template, only 200000 cells) and it works perfectly on Win10. Welcome (back) automated post processing!

Almost perfectly: in the post rpcessing report (both txt and html) I can't find the total cd/cl

[CAEdevice]

I completed the first simulation dith the new solver and new case folder.

I noticed that the parameters of the SnappyHexMeshDict are the same as in 2017 but fvSchemes parameters are slightly different.
I also noticed that the "mesh refinement proces" take about 3000s instead of 1800s (same geometry).

The total downforce is about the same (not matter the absolute value, since I am using an old 2017 car to test to CFD environement) but the drag is 15% higher: do you think it could depend on the paramters of the fvSchemes file?

[schwepes]

Are there some "leaks" from already registered teams? Spyshots?

[CAEdevice]

Are there some "leaks" from already registered teams? Spyshots?

I used a 2017 car with 2018 livery for the registration, but I might have something ready in a few hours...