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

Post here information about your own engineering projects, including but not limited to building your own car or designing a virtual car through CAD.
User avatar
yinlad
1
Joined: Fri Nov 08, 2019 7:10 pm

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

Post

LVDH wrote:
Mon Apr 12, 2021 7:13 pm
CAEdevice wrote:
Mon Apr 12, 2021 7:38 am
Ok , I can set to flatted fapart of the deformed tyre 1mm below the ground l, but if you prefer not deformed tyre at all, no problem, it will be even easier to model.
A little bit is ok and actually very good as a completely round tire would be a bit unrealistic, but the CFD domain should define the plane that cuts the wheel off.


yinlad wrote:
Mon Apr 12, 2021 11:10 am
Will there be a 'standard' for creating meshes this year? I suspect there's a lot of performance in mesh optimisation and it doesn't really feel like that's what this competition is supposed to be about
For the official simulations a pre-defined template in MFlow is used. To develop your car you are welcome to use what ever settings you like. You should just make sure that the results of your own simulations correlate well with the official ones.
I mean for the STL mesh generation. If I recall correctly a lot of the top guys do a lot of work running scripts etc. to optimise the final surfaces being submitted for each round

User avatar
LVDH
31
Joined: Tue Mar 31, 2015 1:23 pm

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

Post

yinlad wrote:
Mon Apr 12, 2021 7:17 pm
LVDH wrote:
Mon Apr 12, 2021 7:13 pm
CAEdevice wrote:
Mon Apr 12, 2021 7:38 am
Ok , I can set to flatted fapart of the deformed tyre 1mm below the ground l, but if you prefer not deformed tyre at all, no problem, it will be even easier to model.
A little bit is ok and actually very good as a completely round tire would be a bit unrealistic, but the CFD domain should define the plane that cuts the wheel off.


yinlad wrote:
Mon Apr 12, 2021 11:10 am
Will there be a 'standard' for creating meshes this year? I suspect there's a lot of performance in mesh optimisation and it doesn't really feel like that's what this competition is supposed to be about
For the official simulations a pre-defined template in MFlow is used. To develop your car you are welcome to use what ever settings you like. You should just make sure that the results of your own simulations correlate well with the official ones.
I mean for the STL mesh generation. If I recall correctly a lot of the top guys do a lot of work running scripts etc. to optimise the final surfaces being submitted for each round
Ah, I see. I have experimented a lot with this in the past. It is possible to have stl files of various quality and this can affect the mesh created by snappyHexMesh. However the differences are more than minuscule and the differences in these meshes create even smaller differences in the end results of the cfd simulations. If people do invest time in this, it is for sure a big waste of time.
A more common issue that can affect the cfd results can happen when a mesh cell gets cut in a very unfortunate way. You can usually spot this because the simulation completely blows up at a specific point. Then you just have to move the geometry in that area by a half mm. This is more a problem of geometry and not stl quality and happens rather rarely and at random.

User avatar
jjn9128
644
Joined: Tue May 02, 2017 10:53 pm

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

Post

LVDH wrote:
Tue Apr 13, 2021 8:10 am
A more common issue that can affect the cfd results can happen when a mesh cell gets cut in a very unfortunate way. You can usually spot this because the simulation completely blows up at a specific point. Then you just have to move the geometry in that area by a half mm. This is more a problem of geometry and not stl quality and happens rather rarely and at random.
Is that with wings with zero thickness at the trailing edge?
#aerogandalf
"There is one big friend. It is downforce. And once you have this it’s a big mate and it’s helping a lot." Robert Kubica

User avatar
LVDH
31
Joined: Tue Mar 31, 2015 1:23 pm

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

Post

jjn9128 wrote:
Tue Apr 13, 2021 8:22 am
LVDH wrote:
Tue Apr 13, 2021 8:10 am
A more common issue that can affect the cfd results can happen when a mesh cell gets cut in a very unfortunate way. You can usually spot this because the simulation completely blows up at a specific point. Then you just have to move the geometry in that area by a half mm. This is more a problem of geometry and not stl quality and happens rather rarely and at random.
Is that with wings with zero thickness at the trailing edge?
Not sure what you mean with zero thickness at trailing edge. But if you mean you have a wing profile that gets so flat that over a certain length, you have the top and bottom surface touching each other at the trailing edge, then no. SnappyHexMesh is quite robust against this. Today this is mostly annoying during post-processing. Imagine you want to visualize the surface pressure on top of your wing. You will see both, the top and bottom, often resulting in a "checkerboard" pattern. You see the same thing when you are running non-closed surfaces. You will see the inside and outside pressure at the same time. Here in the challenge, as most people have rather small computers, this usually results in crashing simulations as the inside of the geometry gets filled with mesh, which then takes too much memory.

User avatar
jjn9128
644
Joined: Tue May 02, 2017 10:53 pm

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

Post

LVDH wrote:
Tue Apr 13, 2021 8:32 am
Not sure what you mean with zero thickness at trailing edge. But if you mean you have a wing profile that gets so flat that over a certain length, you have the top and bottom surface touching each other at the trailing edge, then no. SnappyHexMesh is quite robust against this. Today this is mostly annoying during post-processing. Imagine you want to visualize the surface pressure on top of your wing. You will see both, the top and bottom, often resulting in a "checkerboard" pattern. You see the same thing when you are running non-closed surfaces. You will see the inside and outside pressure at the same time. Here in the challenge, as most people have rather small computers, this usually results in crashing simulations as the inside of the geometry gets filled with mesh, which then takes too much memory.
My understanding is it's always best practice to have a 1.5-2mm thickness on trailing edges rather than going to a point.

I'm well aware of non-closed surface issues :lol: onshape seems to have a proclivity towards exporting at least 1 hole per geometry. I've taken to exporting as .step and using gmsh or meshlab to produce a passable surface mesh.
#aerogandalf
"There is one big friend. It is downforce. And once you have this it’s a big mate and it’s helping a lot." Robert Kubica

LegendaryM
LegendaryM
0
Joined: Mon May 11, 2009 8:56 pm

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

Post

I found this very helpful script for Onshape: https://forum.onshape.com/discussion/95 ... ure-script which makes a nice aerofoil from a chord line and a few parameters, with I think a TE thickness of 2mm. It's very similar to what's used in industry and I'm quite happy I found it because drawing aerofoils is quite a tedious process.

User avatar
jjn9128
644
Joined: Tue May 02, 2017 10:53 pm

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

Post

LegendaryM wrote:
Tue Apr 13, 2021 12:55 pm
I found this very helpful script for Onshape: https://forum.onshape.com/discussion/95 ... ure-script which makes a nice aerofoil from a chord line and a few parameters, with I think a TE thickness of 2mm. It's very similar to what's used in industry and I'm quite happy I found it because drawing aerofoils is quite a tedious process.
Not far off how I define aerofoils. Essentially an ellipsis and a pair of splines (a spline where you can control magnitude of the handles is preferable to giving a dimension but OnShape doesn't have that option within a sketch). Designing wings is one of my favourite things :lol:

Image
#aerogandalf
"There is one big friend. It is downforce. And once you have this it’s a big mate and it’s helping a lot." Robert Kubica

User avatar
LVDH
31
Joined: Tue Mar 31, 2015 1:23 pm

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

Post

jjn9128 wrote:
Tue Apr 13, 2021 10:01 am
LVDH wrote:
Tue Apr 13, 2021 8:32 am
Not sure what you mean with zero thickness at trailing edge. But if you mean you have a wing profile that gets so flat that over a certain length, you have the top and bottom surface touching each other at the trailing edge, then no. SnappyHexMesh is quite robust against this. Today this is mostly annoying during post-processing. Imagine you want to visualize the surface pressure on top of your wing. You will see both, the top and bottom, often resulting in a "checkerboard" pattern. You see the same thing when you are running non-closed surfaces. You will see the inside and outside pressure at the same time. Here in the challenge, as most people have rather small computers, this usually results in crashing simulations as the inside of the geometry gets filled with mesh, which then takes too much memory.
My understanding is it's always best practice to have a 1.5-2mm thickness on trailing edges rather than going to a point.
It does not hurt, but nowadays I just do not think it is worth having a headache about this anymore. Keep in mind though, that the MVRC rules do have requirements about this. I am not sure what will happen to them.


jjn9128 wrote:
Tue Apr 13, 2021 10:01 am
I'm well aware of non-closed surface issues :lol: onshape seems to have a proclivity towards exporting at least 1 hole per geometry. I've taken to exporting as .step and using gmsh or meshlab to produce a passable surface mesh.
Having these kind of scripts can indeed give a competitive advantage over the other teams. Most CAD software is still surprisingly bad a t exporting good (enough) stl files.
One of the best software tools I know, to create crazy shapes, while not really a CAD software, that does export perfect stl files is Blender. While not used so far by other teams, I can highly recommend it for this competition.

DP_CFD
DP_CFD
5
Joined: Tue Apr 13, 2021 3:08 am
Location: Toronto, Canada

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

Post

jjn9128 wrote:
Tue Apr 13, 2021 2:20 pm
LegendaryM wrote:
Tue Apr 13, 2021 12:55 pm
I found this very helpful script for Onshape: https://forum.onshape.com/discussion/95 ... ure-script which makes a nice aerofoil from a chord line and a few parameters, with I think a TE thickness of 2mm. It's very similar to what's used in industry and I'm quite happy I found it because drawing aerofoils is quite a tedious process.
Not far off how I define aerofoils. Essentially an ellipsis and a pair of splines (a spline where you can control magnitude of the handles is preferable to giving a dimension but OnShape doesn't have that option within a sketch). Designing wings is one of my favourite things :lol:

https://db3pap006files.storage.live.com ... pmode=none
Quite similar to my my method too!

Image

I take it just a step further in terms of simplicity as I like to minimize my design space. One important thing with this is that it's accompanied by a spreadsheet that calculates the y_1 and y_2 locations on the suction and pressure sides in order to set the LE radius, and also such that the minimum radius occurs at the LE. Assuming the angles are set, this only leaves the TE tangency points (x_4) and the mid-cord (x_2) locations to play with for each side.
Last edited by DP_CFD on Wed Apr 14, 2021 3:48 am, edited 2 times in total.
aka David Penner

lynch
lynch
0
Joined: Sun Mar 15, 2009 2:28 pm

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

Post

Hello,
I just found this competition and I a really interested in participate.
Will it run this year?

User avatar
jjn9128
644
Joined: Tue May 02, 2017 10:53 pm

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

Post

DP_CFD wrote:
Wed Apr 14, 2021 1:49 am
jjn9128 wrote:
Tue Apr 13, 2021 2:20 pm
LegendaryM wrote:
Tue Apr 13, 2021 12:55 pm
I found this very helpful script for Onshape: https://forum.onshape.com/discussion/95 ... ure-script which makes a nice aerofoil from a chord line and a few parameters, with I think a TE thickness of 2mm. It's very similar to what's used in industry and I'm quite happy I found it because drawing aerofoils is quite a tedious process.
Not far off how I define aerofoils. Essentially an ellipsis and a pair of splines (a spline where you can control magnitude of the handles is preferable to giving a dimension but OnShape doesn't have that option within a sketch). Designing wings is one of my favourite things :lol:

https://db3pap006files.storage.live.com ... pmode=none
Quite similar to my my method too!

https://i.imgur.com/itEwtUZ.png

I take it just a step further in terms of simplicity as I like to minimize my design space. One important thing with this is that it's accompanied by a spreadsheet that calculates the y_1 and y_2 locations on the suction and pressure sides in order to set the LE radius, and also such that the minimum radius occurs at the LE. Assuming the angles are set, this only leaves the TE tangency points (x_4) and the mid-cord (x_2) locations to play with for each side.
I find having the ellipsis section beneficial for a couple of reasons, 1) it gives extra points to fix splines to for wing lofts - I like at least 4 guide splines per wing section but 5 is optimal, 2) it makes it easier to split the leading edge for meshing reasons, having the maximum curvature region separate allows the mesh to be tailored, leaving fewer cells where the curvature is more benign.
#aerogandalf
"There is one big friend. It is downforce. And once you have this it’s a big mate and it’s helping a lot." Robert Kubica

User avatar
CAEdevice
40
Joined: Thu Jan 09, 2014 2:33 pm
Location: Erba, Italy

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

Post

Wouldn't be better to use a 1mm round (or bigger) instead of the 2mm segment?

It should be easy to check and would avoid cuspide shape.

User avatar
LVDH
31
Joined: Tue Mar 31, 2015 1:23 pm

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

Post

lynch wrote:
Sun Apr 18, 2021 3:44 pm
Hello,
I just found this competition and I a really interested in participate.
Will it run this year?
Yes, the idea was to be close to an announcement this weekend, but I am afraid we will have to hope for next weekend. I am getting very close though. Running this competition takes a lot of time and while we always somehow just bang out a new season without being done with everything, getting as much done as possible makes running the competition easier and is nicer for the competitors, so I hope you can wait that extra time.

lynch
lynch
0
Joined: Sun Mar 15, 2009 2:28 pm

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

Post

LVDH wrote:
Sun Apr 18, 2021 7:59 pm
lynch wrote:
Sun Apr 18, 2021 3:44 pm
Hello,
I just found this competition and I a really interested in participate.
Will it run this year?
Yes, the idea was to be close to an announcement this weekend, but I am afraid we will have to hope for next weekend. I am getting very close though. Running this competition takes a lot of time and while we always somehow just bang out a new season without being done with everything, getting as much done as possible makes running the competition easier and is nicer for the competitors, so I hope you can wait that extra time.
Of course I can wait, I hope I will be able to design a competitive car.

User avatar
CAEdevice
40
Joined: Thu Jan 09, 2014 2:33 pm
Location: Erba, Italy

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

Post

Hi, the front wheels are ready, rear wheels will be ready in 24h.

I was wondering if we are going to change the wheelbase.

In 2022 wheelbase will not be free in F1 but limited to 3600mm (2021 F1 cars have a wheelbase of about 3700mm-3750mm and 2020 MVRC cars used a wheelbase of 3500mm).