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

[yinlad]

Congrats to those in the podium.

A race to forget for me. Always knew it was going to be bad going in but didn't expect it to be quite THAT bad.

Don't think I'm going to find myself near the front runners any time soon but hoping to see a better finish next time out

[variante]

congratulations for the winner - Variante did a nice homework! La Grande Machina! (I really hope no mistakes happened)

Hahah thanks mate! (i read it with Vettel's voice)

Now some news, that doesn't come with ease: I decided to withdraw from this year competition

That's a pity... Hopefully, a break will make you change your mind. I also had moments where is seemed like any effort would just make the car worse (this happens especially when the starting point is a good car like yours), but eventually things always start rolling again.

[Stlawrks]

I noticed that the cooling flow of some cars (inlcuding mine) is 6.0 m^3/s: I am sure that I have a good cooling (above 3 m^3/s) but 6 m ^3/s is too optimistic

I am also wondering why some teams got a cooling flow of 6.0000 m^3/s? It differs a lot from other cars and it's exactly the same for all those teams I believe. Just curios.

Anyways, great job by all the teams and a lot closer racing than I was expecting. And I didn't finish last

[LVDH]

I noticed that the cooling flow of some cars (inlcuding mine) is 6.0 m^3/s: I am sure that I have a good cooling (above 3 m^3/s) but 6 m ^3/s is too optimistic

I am also wondering why some teams got a cooling flow of 6.0000 m^3/s? It differs a lot from other cars and it's exactly the same for all those teams I believe. Just curios.

MFlow automatically exports a line with all the values relevant to the Virtual StopWatch that can easily get copy-pasted. It seems nobody has noticed, because it also complains if it cannot detect all relevant values on its own. Most of the time this is caused by strange name for the HX monitor. Then MFlow just exports a 6 as default value (and gives advice on how to name the part). When submitting the numbers to Richard, I paid attention that on the cars with a 6 for cooling flow, the real value was above 3. I did not know that the numbers would wind up in the video.
Other people wondering about their numbers, should keep in mind that I remesh the monitoring surface and run the simulation for 5000 iterations. This can account for some differences.

[Stlawrks]

I noticed that the cooling flow of some cars (inlcuding mine) is 6.0 m^3/s: I am sure that I have a good cooling (above 3 m^3/s) but 6 m ^3/s is too optimistic

I am also wondering why some teams got a cooling flow of 6.0000 m^3/s? It differs a lot from other cars and it's exactly the same for all those teams I believe. Just curios.

MFlow automatically exports a line with all the values relevant to the Virtual StopWatch that can easily get copy-pasted. It seems nobody has noticed, because it also complains if it cannot detect all relevant values on its own. Most of the time this is caused by strange name for the HX monitor. Then MFlow just exports a 6 as default value (and gives advice on how to name the part). When submitting the numbers to Richard, I paid attention that on the cars with a 6 for cooling flow, the real value was above 3. I did not know that the numbers would wind up in the video.
Other people wondering about their numbers, should keep in mind that I remesh the monitoring surface and run the simulation for 5000 iterations. This can account for some differences.

Great! Thanks for clarifying that.

[k.ko100v]

Hi guys,

I want to ask something which is most probably already discussed in the previous topics. If it is, sorry for that.

On the residuals graphs I am observing some difference between the different simulations and most of the cases the difference is the different behaviour of the "k"

My question is: What it is the "k" coefficient on the residuals graphs and can I make a certain conclusion between the two behaviours of the "k" on the graphs bellow?


Regards,
Krasen

[Thomas2019]

k = turbulent kinetic energy
Wikipedia

I wouldn't worry, as long as "p" doesn't go ballistic.
And there is not really anything to draw from the difference in the two graphs.

Thomas

[Koldskaal]

I'm a bit late but congrats to variante and CAEdevice. I am very happy about my 3. place in such a strong field. It looks like there will be many podium contenders for the next race

Now for something completely different:
I have found a way to make pressure distribution plots in Paraview. I have done a lot of work on the front wing lately and found these essential for apples to apples comparisons. (no more judging the blueness of wings!) Here is an example, this is my Yas Marina spec front wing from last year:

A simple how to: Slice a patch. Plot the slice. Use Points_x for x-axis parameter.

[LVDH]

Excellent work, that is how to climb to the top. Wasn't it you with the OF mesh morphing? Did you get somewhere with this? If I ever find the time, I would like to show how to parametrize geometries, hook it up with a Python optimizer and optimize the car. But if one of you guys can do it, even better.

[etsmc]

Glad you guys all understand what these plots mean.
I have been making progress on the car but its the old 1 step forward but 2 steps back I think I am heading in the right direction though, each time I increase the cooling I just have to rebalance the car some. So far the best progress has improved on my round 1 time by ~0.9s so can't be too disappointed at that.
Is anyone willing to share ideas/concepts as to how they are controlling rear tyre squirt?? I think i am getting a massive area of recirculating air causing issues with diffuser flow again but it seems to start at ground level and work its way up and into the diffuser.

[LVDH]

Short update: I worked on preparing all the reports yesterday evening but did not get through. I will probably send them out tomorrow, probably by using Wetransfer. I have one candidate who wants his report sent via ftp. Those with Dropbox can send me a file request.

[jjn9128]

Glad you guys all understand what these plots mean.
I have been making progress on the car but its the old 1 step forward but 2 steps back I think I am heading in the right direction though, each time I increase the cooling I just have to rebalance the car some. So far the best progress has improved on my round 1 time by ~0.9s so can't be too disappointed at that.
Is anyone willing to share ideas/concepts as to how they are controlling rear tyre squirt?? I think i am getting a massive area of recirculating air causing issues with diffuser flow again but it seems to start at ground level and work its way up and into the diffuser.

What are you doing already? A simple solution is to move the throat of your diffuser inboard away from the tyre contact patch and expanding it laterally, you lose downforce producing surface though. A vertical strake placed between the tyre and diffuser should create a vortex which rotates flow in the opposite direction to the squirt vortex. Other than that try to reduce the high pressure zone ahead of the tyre as much as possible. Outwashing the front tyre wake less will reduce rear wheel drag to the detriment of the floor performance.

[Koldskaal]

Excellent work, that is how to climb to the top. Wasn't it you with the OF mesh morphing? Did you get somewhere with this? If I ever find the time, I would like to show how to parametrize geometries, hook it up with a Python optimizer and optimize the car. But if one of you guys can do it, even better.

Yup! But I have only used it very little on the car. It is quite hassle to setup, and the results are not always great. I am rewriting the whole thing in python (univerisity licenses dont last forever ) with a new approach which should give higher quality morphing and a lot more control. I want to make it available to MVRC contestants when I am done.
Here is an example, original case (left) and morphed case (right):

And force convergens looks like this:

[machin]

I think i am getting a massive area of recirculating air causing issues with diffuser flow again but it seems to start at ground level and work its way up and into the diffuser.

Do you think you would be happy to share an image or two which leads you to the conclusion that the air is recirculating -(I’m not disagreeing with your conclusion, but I hope that posting an image might help others to give you some constructive feedback )

[LVDH]

Excellent work, that is how to climb to the top. Wasn't it you with the OF mesh morphing? Did you get somewhere with this? If I ever find the time, I would like to show how to parametrize geometries, hook it up with a Python optimizer and optimize the car. But if one of you guys can do it, even better.

Yup! But I have only used it very little on the car. It is quite hassle to setup, and the results are not always great. I am rewriting the whole thing in python (univerisity licenses dont last forever ) with a new approach which should give higher quality morphing and a lot more control. I want to make it available to MVRC contestants when I am done.
Here is an example, original case (left) and morphed case (right):
https://i.imgur.com/jP74WSa.png
And force convergens looks like this:
https://i.imgur.com/GeHQIq5.png

This is fantastic work. Have you implemented multi-objective optimizations already? Would help to keep the drag low when when asking for more downforce.