[MVRC] Mantium Virtual Racecar Challenge 2016

[CAEdevice]

I am evaluating this right now. I think it should be increased a lot. But I also do not want to make it impossible to design a car.

I have the suspect that to increase the required flow and, at the same time, the radiator resistence would lead to place the inlet in front of the car or in odd places, in order to have a direct air flow on it.

I am sure that the radiator resistance has to be increased, but not so sure about the cooling flow. How do you estimated it? The car are tested at a quite low speed (44.7 m/s)...

[LVDH]

I am sure that the radiator resistance has to be increased, but not so sure about the cooling flow. How do you estimated it? The car are tested at a quite low speed (44.7 m/s)...

Yes, from experience I know that what we were asking from you guys was too low. Even at the speeds we are using. Actually in MVCR we test at 40m/s. This is what is done in most wind tunnels and I know what to expect at those speeds.
What baffled me last year was how hard it was to get the cars to the cooling requirements. And I did tests on my own old car and had your issues as well. Now I took the time for deeper investigations and found out that I was right and the resistance and required cooling flow were too low. Still, why was it so hard to achieve? I am not sure. One thing that was quite obvious is that front down force and cooling flow are conflicting objectives. As you guys managed to create very good diffusers and rear wings you were all very focused on front downforce to balance the car. I think that you guys did not like giving up front down force as it meant that the nice diffusers and rear wings had to be scaled back too much and it was just too painful.
So probably to get a good car this year means to take care of cooling right from the first blank sheet of paper and take it as seriously as the other factors.

[CAEdevice]

One thing that was quite obvious is that front down force and cooling flow are conflicting objectives. As you guys managed to create very good diffusers and rear wings you were all very focused on front downforce to balance the car.

I agree, the rear diffusers are very efficient, maybe too much! A different rule, in order to have lower rear diffuser and without the double curvature would help. I t would make the car easier to balance (less rear df) and it would improve the flow from the cooling outlets (the central "hump" of the diffuser is an obstacle for the air).

[rjsa]

One thing to take into account is HX orientation. I guess no one actually uses flat X facing planes.

[CAEdevice]

One thing to take into account is HX orientation. I guess no one actually uses flat X facing planes.

https://paulsf1.files.wordpress.com/201 ... g_0536.jpg

Good point.

A great angle of inclination allows larger hx surfaces, and this (considering the same flow and porosity properties) reduces the internal resistence of the cooling ducts.

[LVDH]

You are right,
so far I have not included this into the MVRC edition of the software as it requires you guys to input a vector the describes the orientation. There are two things to consider here:
1. Not everybody will be able to do that.
2. The MVRC staff has to check if you guys give us the correct vector.

The solution is an oriented bounding box algorithm. It looks easy to implement. If I can get that done in the next days, you will get it. Otherwise you will have to wait to or three races but it will come.

BTW, these paper cars are fantastic.

[CAEdevice]

You are right,
so far I have not included this into the MVRC edition of the software as it requires you guys to input a vector the describes the orientation. There are two things to consider here:
1. Not everybody will be able to do that.
2. The MVRC staff has to check if you guys give us the correct vector.

The solution is an oriented bounding box algorithm. It looks easy to implement. If I can get that done in the next days, you will get it. Otherwise you will have to wait to or three races but it will come.

BTW, these paper cars are fantastic.

In my opinion, a good compromise would be:

1) To keep the present scheme
2) To increase the required flow (I have no empiric experience about hx, so if you have higer values in the wind tunnel it's ok for me)
3) To limit the "volume" (with minumum section dimensions and with mandatory rectangular or trapezoidal section) of the hx, and using omogenous porous properties

I know: this is a rough compromise, but quite simple to implement.

Note: the volume that we call "hx" now would not exactly represent the "hx", but the "volume including the hx"

[rjsa]

Can't you get it from the monitoring surface normal?

[CAEdevice]

You are right,
so far I have not included this into the MVRC edition of the software as it requires you guys to input a vector the describes the orientation. There are two things to consider here:
1. Not everybody will be able to do that.
2. The MVRC staff has to check if you guys give us the correct vector.

The solution is an oriented bounding box algorithm. It looks easy to implement. If I can get that done in the next days, you will get it. Otherwise you will have to wait to or three races but it will come.

BTW, these paper cars are fantastic.

What about only 3 alternatives for the hx orientation? 45° 0° -45°, with the vector written in a txt file.

[rjsa]

I was thinking something along these lines:



An external frame, 20mm thick, porous box and monitor plane bleeding into the frame.

May be it's too convoluted but I think it would make sure there are no leaks and makes it easier to move the HX around.

Being it always rectangular, I haven't seen anything different than that used.

[rjsa]

You are right,
so far I have not included this into the MVRC edition of the software as it requires you guys to input a vector the describes the orientation. There are two things to consider here:
1. Not everybody will be able to do that.
2. The MVRC staff has to check if you guys give us the correct vector.

The solution is an oriented bounding box algorithm. It looks easy to implement. If I can get that done in the next days, you will get it. Otherwise you will have to wait to or three races but it will come.

BTW, these paper cars are fantastic.

What about only 3 alternatives for the hx orientation? 45° 0° -45°, with the vector written in a txt file.

May be 15° increments in Y or Z, never both?

[CAEdevice]

May we use homogenous properites for the hx porous volume, in order to aboid any orientation issue? Would the difference be so big? I don't think so, especially if we would use a thinner volume (50mm), with mandatory 50mm of extruded (void) volume from the two faces.

[rjsa]

May we use homogenous properites for the hx porous volume, in order to aboid any orientation issue? Would the difference be so big? I don't think so, especially if we would use a thinner volume (50mm), with mandatory 50mm of extruded (void) volume from the two faces.

My understanding is that the orientation problem comes from integrating flow over the monitor surface. Isn't that the case? If it is, the normal is needed anyway. The projected area of the monitoring surface on YZ would also do the trick.

[LVDH]

Measuring the flow with the monitor surfaces works without any tricks. Using that surface to determine the normal vector of the HX is a good fall back idea though.
The vector is needed to specify the flow resistance through the HX as it is not isotropic and it would make a huge difference to neglect that. A real HX directs the flow and if you look in your old results in MVRC they do that as well. A while ago I checked the algorithm to implement an automatic detection and it does not look that difficult. Right now I need to fix the possibly new values for the HX and get an intro car rolling. Those two are more difficult.

[CAEdevice]

Ok. Do you have any proposal about the required flow (4÷5 m3/s ?) And porosity parameters? I would like to test my 2016 car with the new values.