[KVRC] Khamsin Virtual Racecar Challenge 2016

[CAEdevice]

Hi Andre. I'm sorry: in the first release of the article I included some words about your optimization technique, but I had to cut it (and some words about the teams that did not win a race but obtained good results). I hope to receive comments and visits so I can write more in the magazine.

About the heat exchanger dinensions on my car: wich car are yu referring to?

I published a lot of versions (including the article). Take as a reference the car with orange sidepods and black livery you can find on this thread and on the KVRC-CAEdevice thread.



That car is option2 compliant: heat exchangers have a 230000mm2 area but they have been rotated of about 10°.

They look big too (I used a perpective view so it may change your impression), especially considering the 100mm extruding distance.

Anyway: the 230000mm2 area is realistic in my opinion, but I would reduce the limits about rotation and outlet areas/positions for opt2.

[LVDH]

Then have another look when you actually make it rule complaint. You should have 250.000mm². And having yours positioned so far forward where the cockpit is quite wide I think they would have to be much higher, or extend outside the width limits. Or are the 250.000 for both sides together?

[CAEdevice]

I read somewhere that there was a proposal to reduce the area from 250000mm2 to 230000mm2. Maybe on this thread.

I confirm that area of each heat exchanger of the car in the above is 230000mm2. I don't have the impression that there is a big difference between my sidepods max section and your heat exchangers study.
The 10° rotation is quite important, but I can't reveal too much.

[CAEdevice]

By the way: we really need a definitive rulebook! Later I'll post a list of proposal after dedicating about 4h to the design in the last week.

[variante]

The area has been reduced to 230000 for each side. Still, given the 100mm extrusion and the fact that it's harder to model very complex radiators and ducts (in comparison with an easily achievable complex area template of option 1), the radiator area is still quite large.

Edit: also, a minimum distance of 10mm between the inner and outer flows must be kept. This is not required by option 1.

[LVDH]

When I take that huge block in my car and rotate it 30° it has almost the same height!
In the location I chose, it would help if it was allowed to rotate around Z.
In general I guess a somewhat smaller hx would look more realistic.
I will spend some time on Mulsanne's Corner.
Maybe I can find some nice pictures.

[AratzH]

So nice to see new cars. I think it is time to show some of my own.







I must agree that the 230000 mm^2 radiator area looks pretty huge and requires some clever packaging, but this also gives a greater sense or realism, specially when you have to model all the ducting.

I wonder if Julien could provide more info about the characteristics of the porous medium for the heat exchangers. Right now I am having a hard time to get to the 3m^3/s of flow.

If it's not too indiscreet, how large of an intake area are you using for the hear exchangers? Right now I am at aroun 60% of heat exchanger area....

[CAEdevice]

Nice car AratzH and nice landscape!!!

I have not a definitive opinion about opt2, since it is not possible to simulate it with occfd at the moment, but I have the impression that inlet/outlet/rotation limitations should be reduced.

The other (last) proposals (I have already wrote about it, but now I am more sure) is the reduction of the floor height (45mm >>> 25mm, 35mm would be a good compromise), in order to have realistic diffusers (maybe a bit lower, 50mm, if we prefer to keep similar df values).

At the moment, anyway, the most urgent things are a definitive rulebook and a working release of occfd including opt2 and without detecting convergence issues (I suspended the design in the meanwhile).

[LVDH]

Hi AratzH,
cool pictures. Looks like your car is in a nice place.
And thanks for the radiator image. On your car it looks somewhat OK.

After having a good look at some of the beauties at Mulsanne's Corner. I do think that our heat exchangers are a bit too big. But considering that we have a super modern engine which has quite some requirements when it comes to thermal management, they are not far too big. What I did notice is that cars with large radiators have them rotated to an almost flat position, maybe 30 deg missing to being horizontal. So maybe we could indeed rethink some of the rule limits.

The other (last) proposals (I have already wrote about it, but now I am more sure) is the reduction of the floor height (45mm >>> 25mm, 35mm would be a good compromise), in order to have realistic diffusers (maybe a bit lower, 50mm, if we prefer to keep similar df values).

That could lead to a CFD nightmare . Julien would not be very happy.

[LVDH]

If it's not too indiscreet, how large of an intake area are you using for the hear exchangers? Right now I am at aroun 60% of heat exchanger area....

That is a very good starting point. Just have a look at some slices through the domain showing velocity, preferably with LIC enabled. If the stream lines all look good and attached you could try to reduce the inlet opening a bit. Do not forget the exits though. The nice thing here is that you only have to deal with one velocity you have to design the ducts for. So you can make them perfect.
Next step for you would be to make the duct tangential to the radiator. Like you have on the lower left corner in the image. The top left and lower right are examples that need some more work.

[CAEdevice]

What I did notice is that cars with large radiators have them rotated to an almost flat position, maybe 30 deg missing to being horizontal. So maybe we could indeed rethink some of the rule limits.

Finally someone else noticed that

I agree with that: 45° would be ok. I would suggest a wide angle, but the limits on the outlet ducts (number of openings, total area, rotations) would be cancealed.

I was thinking about the required flow: I like the idea of a smooth transition from 2015 rules to 2017, but what about considering the real flow (2017), without any limitation for opt2, except obviously the heat exchanger mimimum area, but with a less dramatic impact on engine performance for both opt1 and opt2? At the moment 10% less of cooling flow/pressure_area = 10% less power, that is realisti,c but it could be modified to be less severe in 2016, es. computing the power loss only on the 50% of the total power.

The other (last) proposals (I have already wrote about it, but now I am more sure) is the reduction of the floor height (45mm >>> 25mm, 35mm would be a good compromise), in order to have realistic diffusers (maybe a bit lower, 50mm, if we prefer to keep similar df values).

That could lead to a CFD nightmare . Julien would not be very happy.

I don't know: I would use the same number of cells between the floor and the ground, only about 20% thinner (I consider 35mm, 25mm would be a problem actually).

[CAEdevice]

If it's not too indiscreet, how large of an intake area are you using for the hear exchangers? Right now I am at aroun 60% of heat exchanger area....

I started with a typical ratio: inlet area = 25% of the heat exchanger area.

Honestly I could not find a way to make it work, but consider the I'm having a pause in my design since I published the picture of the car.

[CAEdevice]

Hi everybody, I have a proposal to try to obtain a compromise between the idea of having more design freedom and the difficulties that Julien has having about the heat exchangers simulation (Julien is making a good but heavy job, considering that he is working alone on the simulation).

I would keep:

- the heat exchanger template (200000-250000 mm2) and 30° rotation freedom (or increased to 45°)
- the normal extrusion (100mm): the volume between front and rear heat exchanger surfaces will be part of the body
- the required flow through the heat exchanger (2016 or 2017 value)

... but I would apply the "flow boundary condition" directly to the virtual heat exchanger surfaces and then I would compute the pressure differential between the heat exchanger surfaces.

In a few words: a compromise between Opt1 and Opt2, taking the computation simplicity from Opt1 and the design freedom from Opt2.

[LVDH]

The other (last) proposals (I have already wrote about it, but now I am more sure) is the reduction of the floor height (45mm >>> 25mm, 35mm would be a good compromise), in order to have realistic diffusers (maybe a bit lower, 50mm, if we prefer to keep similar df values).

That could lead to a CFD nightmare . Julien would not be very happy.

I don't know: I would use the same number of cells between the floor and the ground, only about 20% thinner (I consider 35mm, 25mm would be a problem actually).

So the CAD guy does not believe the CFD guy. I ran into that situation once somewhere...
Maybe you could listen to aerodynamic reasons. 45-50mm is a good height. If you get too low you run into other trouble.
Have a look at this:
http://www.consultkeithyoung.com/conten ... ide-height

[CAEdevice]

That could lead to a CFD nightmare . Julien would not be very happy.

I don't know: I would use the same number of cells between the floor and the ground, only about 20% thinner (I consider 35mm, 25mm would be a problem actually).

So the CAD guy does not believe the CFD guy. I ran into that situation once somewhere...
Maybe you could listen to aerodynamic reasons. 45-50mm is a good height. If you get too low you run into other trouble.
Have a look at this:
http://www.consultkeithyoung.com/conten ... ide-height

The CAE guy thinks that the image below (thanks for the link to the article) demonstrates that a slightly reduced ride height would require more realistic diffuser angles (I mean the angle of the first part of the KVRC diffusers, that are so high at the and to allow some other strange things).



But I agree that the could be some meshing problems lowering too much (and I don't want to cause problems to Julien): 35mm would be ok, 40mm would be better anyway.