[MVRC] Mantium Virtual Racecar Challenge 2020 (Grand Prix Cars)

[jjn9128]

Here you go. Definitely improvements could be made but rear wing is pretty free.

[variante]

Yeah, true. That looks pretty nice.
Even with my side exits, some streams manage to get close to the rear wing, so i know how difficult that management is.

[MaccaRacing]

Here you go. Definitely improvements could be made but rear wing is pretty free.

https://pt1kqw.db.files.1drv.com/y4mC-r ... pmode=none

Very nice pic! And really nice from you to share this pic!

I would like to ask you some question regarding Mantium Flow and openfoam.. is there any chance to get a "conditional" launch of a case? I explain better.. since I can launch three cases (fast) in two days more or less, and since I can't control my PC during the night and during work hours, I would like to launch the second or the third case according to the results of the previous.. maybe I am asking too much

[jjn9128]

I would like to ask you some question regarding Mantium Flow and openfoam.. is there any chance to get a "conditional" launch of a case? I explain better.. since I can launch three cases (fast) in two days more or less, and since I can't control my PC during the night and during work hours, I would like to launch the second or the third case according to the results of the previous.. maybe I am asking too much

Not sure if how it works with the mflow software but you should be able to write a batch/shell script to run 2/3 cases sequentially. You'd have to start it from the terminal rather than the mflow GUI though, so it depends how confident you are in that environment.

I think if you did the setup (build case) in MFlow you could write a really simple script like:

cd [case_file1]
bash runCase.sh
cd ..

cd [case_file2]
bash runCase.sh
cd ..

...etc

Which should run the cases one after the other.

How you call the program depends if you're running linux or windows, likewise I think the file extension is different... if windows you might be able to start through the bluecore terminal which is by default in the openfoam environment.

EDIT: Tried this with some empty folders (I don't have any cases to run at the moment) and it seems like it would work. Save the text file as a .bat and then run it from the bluecore terminal.

[MaccaRacing]

I think if you did the setup (build case) in MFlow you could write a really simple script like:

cd [case_file1]
bash runCase.sh
cd..

cd [case_file2]
bash runCase.sh
cd..

...etc

Which should run the cases one after the other.

Thank you for the reply, but maybe I was not clear enough in explaining it

Actually I am running the cases as you wrote, by launching a script through bluecfd in windows. However I would like to build a script that compares the results of the first simulation of the script with a reference case, and then, maybe with an "if" statement, bluecfd launches a case rather then another one..

For example, the first case is a new type of bargeboard (let's call it V2). At the end, Cz increased from 2 to 2.1, with no increase of Cx, with respect to the case with bargeboard V1 of the day before.

Then for example, I would like to test a rear wing. Assuming that in Mantium Flow I have already built two cases with the new rear wing in both and with bargeboard V1 in one and bargeboard V2 in the other, how can I manage to tell bluecfd to run the case with bargeboard V2?

I would imagine in my head something like this:

If cz > 2
cd '...\case2a'
Runcase.sh
else
cd '...\case2b'
Runcase.sh
end

(don't look at the condition, is just for example.. The ideal one would include going to the stopwatch website, but I admit this is far too complex then maybe reading the coefficiente of the last iteretion in the solver.log)

Sorry if I wrote a long post

[k.ko100v]

Hi Guys,

Something strange happens.

2 days ago, I recieved this kind of cooling flow shape, after designing a brand new diffuser, without any flaps.
https://imgur.com/Cq5TM8Z.jpg

I thought that the diffuser is stalling, so I managed to fix it, by slightly changing the shape and adding some flaps to guide the flow inside the diffuser. The flow shape was improved, BUT I lost arround 0.2 Lift and the CoP move to the front by 0.1m

https://imgur.com/PijJXdd.jpg

So I am a bit confused...

Has anybody seen something like this?

Regards,
Krasen

With any weird result like this check for holes in the mesh. The first case the streamlines look odd. Is there a second set of heat exchangers or some other accident?!?

Hm, No, the cooling setup is separate file and it is one and the same for both cases. I was wondering how a messy looking flow gives better results, compared to better looking flow and lower results.

I have the assumption that, if any aero element is stalled, it is not aero effective and I espected that after the "corrective actions" the case will be improved. Unfortunately no, but I hope that this is a sign of a good potential

I am testing the case on Mflow from last year. In the current version I miss the "U_nearWalls.vtk" file. I compared the both versions. What I've observed (based on my cases) the newer version is given equal or better results. I would recommend to use the older version for qually

EDIT:

Yeah... it looks like that the diff is stalling (or I am wrong), and somehow the flow is concentrated to the middle of it and makes it stronger

https://imgur.com/lll1FCz.jpg

Regards,
Krasen

Definitely stalling. Try a fence, a strake.

My saga continues with the rear diffuser problems. I tryed more than 20 different setups (diffuser shape, strakes shapes etc)... with very familiar results.
I am suspecting that there is something wrong with the rear tyre wake calculations.

Would you please check if this is normaw air behaviour arround a tyre?:


Compared with the front tyre wake:


Regards,
Krasen

[jjn9128]

Almost looks like it's rotating backwards somehow?!

[etsmc]

One issue i know from the past that affects the diffuser is turbulant flow from the front tyre wake getting under the floor. From your pic it dosent look as though you are trying to push the tyre wake out at all.
But wait for advice from one of the pros on here before you change anything.

[LVDH]

With any weird result like this check for holes in the mesh. The first case the streamlines look odd. Is there a second set of heat exchangers or some other accident?!?

Hm, No, the cooling setup is separate file and it is one and the same for both cases. I was wondering how a messy looking flow gives better results, compared to better looking flow and lower results.

I have the assumption that, if any aero element is stalled, it is not aero effective and I espected that after the "corrective actions" the case will be improved. Unfortunately no, but I hope that this is a sign of a good potential

I am testing the case on Mflow from last year. In the current version I miss the "U_nearWalls.vtk" file. I compared the both versions. What I've observed (based on my cases) the newer version is given equal or better results. I would recommend to use the older version for qually

EDIT:

Yeah... it looks like that the diff is stalling (or I am wrong), and somehow the flow is concentrated to the middle of it and makes it stronger

https://imgur.com/lll1FCz.jpg

Regards,
Krasen

Definitely stalling. Try a fence, a strake.

My saga continues with the rear diffuser problems. I tryed more than 20 different setups (diffuser shape, strakes shapes etc)... with very familiar results.
I am suspecting that there is something wrong with the rear tyre wake calculations.

Would you please check if this is normaw air behaviour arround a tyre?:
https://imgur.com/UPoTNJm.jpg

Compared with the front tyre wake:
https://imgur.com/d0Qszsi.jpg

Regards,
Krasen

That does look pretty strange. You can send me your case. From the image, I am wondering if you are using an extra coarse mesh? That could explain why the flow stays so attached.

[k.ko100v]

Hm, No, the cooling setup is separate file and it is one and the same for both cases. I was wondering how a messy looking flow gives better results, compared to better looking flow and lower results.

I have the assumption that, if any aero element is stalled, it is not aero effective and I espected that after the "corrective actions" the case will be improved. Unfortunately no, but I hope that this is a sign of a good potential

I am testing the case on Mflow from last year. In the current version I miss the "U_nearWalls.vtk" file. I compared the both versions. What I've observed (based on my cases) the newer version is given equal or better results. I would recommend to use the older version for qually

EDIT:

Yeah... it looks like that the diff is stalling (or I am wrong), and somehow the flow is concentrated to the middle of it and makes it stronger

https://imgur.com/lll1FCz.jpg

Regards,
Krasen

Definitely stalling. Try a fence, a strake.

My saga continues with the rear diffuser problems. I tryed more than 20 different setups (diffuser shape, strakes shapes etc)... with very familiar results.
I am suspecting that there is something wrong with the rear tyre wake calculations.

Would you please check if this is normaw air behaviour arround a tyre?:
https://imgur.com/UPoTNJm.jpg

Compared with the front tyre wake:
https://imgur.com/d0Qszsi.jpg

Regards,
Krasen

That does look pretty strange. You can send me your case. From the image, I am wondering if you are using an extra coarse mesh? That could explain why the flow stays so attached.

Yes, which exactly files do you need?
I suppose the "mesh.log" and "solver.log"

[LVDH]

The whole case would be good. Dropbox will not be able to deal with this. If you do not have any means to send it, I could setup an ftp account for you.

I just checked to be sure, the wheels are rotating correctly, which is good as otherwise I would have to hurt myself.

[LVDH]

Actually, from the images, it looks like, you are somehow post-processing the results from the potential solver, used to initialize the simulation. If you do not know, how to verify this, please send me an email, so we do not have to spam the forum full with this. I can later give you guys a summary.

[k.ko100v]

The whole case would be good. Dropbox will not be able to deal with this. If you do not have any means to send it, I could setup an ftp account for you.

I just checked to be sure, the wheels are rotating correctly, which is good as otherwise I would have to hurt myself.

Tell me about it. I want to shoot myself

Ok no problem. I will send you the case wherever you say.

[jjn9128]

hmmmmmm that's a little beyond me, but using the timestep data won't necessarily give you the desired result - depending how accurate you want to be the final timestep could be in the order of 0.1 out from the average. You can also find that just increasing downforce will not necessarily be better for cooling/ balance or efficiency so you really want to be checking 2/3 parameters before choosing how to progress.

But would it be something like

FOR "time = 5000" IN solver.log DO FIND "Cl"
IF Cl > x.xxx
...etc as before

??

[MaccaRacing]

hmmmmmm that's a little beyond me, but using the timestep data won't necessarily give you the desired result - depending how accurate you want to be the final timestep could be in the order of 0.1 out from the average. You can also find that just increasing downforce will not necessarily be better for cooling/ balance or efficiency so you really want to be checking 2/3 parameters before choosing how to progress.

This for sure, the example with the downforce was only to write something quickly. I did not know about the average, so this makes things much more difficult.. I think I will drop down this idea

In any case, thank you very much for answering me!