Khamsin Virtual Racecar Challenge 2014

[RicME85]

Have decided to use the Netfabb/Microsoft cloud model repairer.
To make life easier for me I have broken the car down into the various groups needed for KVRC race submission and exported to STL each group then uploaded to https://netfabb.azurewebsites.net/ and then put all the groups back in the same SKP.
Model looks better for it, will see what happens once run through OpenFOAM

[astracrazy]

just in case anyone wonders why i'm missing from any results, I won't be continuing with my car any further this year. That + running KVRC takes up a lot of spare time

[CAEdevice]

just in case anyone wonders why i'm missing from any results, I won't be continuing with my car any further this year. That + running KVRC takes up a lot of spare time

I'm sorry about that

[machin]

Now, a few people has made the observation that it seems that the KVRC cars are quicker when their wings are at a lower downforce setting than the real F1 cars.... I thought I would look into this, and sure enough, what I find is that real F1 cars are achieving Cl.A values in the region of 4.5m^2 and Cd.A's <1.4, i.e they are much more efficient than we're achieving so far....

If I apply these numbers into Virtual Stopwatch, what I find is that the cars are indeed quicker with higher downforce set-ups... exactly what we see on the track....



I guess that is the difference between an Amateur aerodynamicist and a team of professionals! But anyway, keep up the good work people, lets see how close you can get to the real F1 cars!

[cdsavage]

It may well be that the openfoam setup we are using overestimates drag / underestimates downforce fairly significantly - I would be hesitant to assume our results map well to reality.

[CAEdevice]

It may well be that the openfoam setup we are using overestimates drag / underestimates downforce fairly significantly

I can confirm that, comparing the results with the two solvers I used to test my car. I don't know wich is the more realistic simulation, but I have the same impression of cdsavage.

[CAEdevice]

Now, a few people has made the observation that it seems that the KVRC cars are quicker when their wings are at a lower downforce setting than the real F1 cars.... I thought I would look into this, and sure enough, what I find is that real F1 cars are achieving Cl.A values in the region of 4.5m^2 and Cd.A's <1.4, i.e they are much more efficient than we're achieving so far....

If I apply these numbers into Virtual Stopwatch, what I find is that the cars are indeed quicker with higher downforce set-ups... exactly what we see on the track....

http://www.competition-car-engineering. ... _14_v2.jpg

I guess that is the difference between an Amateur aerodynamicist and a team of professionals! But anyway, keep up the good work people, lets see how close you can get to the real F1 cars!

We should consider that the simulator (I think) works as a perfect traction control and that the answer of the wheels to vertical force is probably considered linear.
I would be curious to see what would happen if the wheels were simulated as harder than they are in the actual model.

Another difference: I'm testing a very comlicated diffuser, but with my mesh settings (similar to the OpenFoam settings) I just can't appreciate small differences and improvements that would increase drag and downforce.

[astracrazy]

just in case anyone wonders why i'm missing from any results, I won't be continuing with my car any further this year. That + running KVRC takes up a lot of spare time

I'm sorry about that

It's fine. I might do some sort of project after this season. Not sure what though. Maybe a previous f1 year or something. or just some research towards which ever direction we decide to take kvrc (rule book wise)

[machin]

We should consider that the simulator (I think) works as a perfect traction control and that the answer of the wheels to vertical force is probably considered linear.
I would be curious to see what would happen if the wheels were simulated as harder than they are in the actual model.

Virtual Stopwatch has a built in transfer function to account for the effect of road bumpiness on traction taking into account the set-up of the car, whether it has traction control or not, the tyre compound, weight distribution etc, although one area that is difficult to know for sure is the tyre data which changes from year to year... However I'm happy that the values I have used for all of these features are fairly close to reality since if I put in Cl.A and Cd.A figures representative of real F1 cars (or at least; figures that have been disclosed on the internet!) the lap times I get back are very indicative of actual F1 car performances at the respective tracks... the only variable then between real F1 cars and the KVRC cars is the aero coefficients.

I can't comment on the relative accuracy of one CFD solver compared to another.. but one thing I have been wondering about; the real F1 cars have rake, which I believe tends to improve underfloor performance (I'm no aerodynamicist so don't hold me to that one!), whereas we're running at zero rake?

The other thing is that because of the mesh size I believe you can't properly simulate very intricate features; now it is the same for everyone, so it isn't really a problem, but it does suggest why you might not be getting the same aerodynamic efficiency as a real F1 car.

[CAEdevice]

real F1 cars have rake

I agree, very good point

[JJR]

It may well be that the openfoam setup we are using overestimates drag / underestimates downforce fairly significantly - I would be hesitant to assume our results map well to reality.

I think that openfoam in current setup is quite accurate. I have done Perrinn P1 ( Le Mans prototype ) test and results ware around 5% off official numbers (L/D - openfoam 4,4, official 4,7 ).
I used openFoam Linux version with KVRC settings.

... and don ´t forget that for F1 teams only in CFD work 20+ people so they can try almost everything

[CAEdevice]

I'm trying to optimize the last winglet for Monza... then I'll decide to use the old MP001A or the new MP001B. In both cases I think that there will be two or three cars faster then mine...

[cdsavage]

The downforce/drag vs laptime chart with the added Cl.A lines last round was good - any chance we could see the same for the next 2 rounds?

[machin]

The downforce/drag vs laptime chart with the added Cl.A lines last round was good - any chance we could see the same for the next 2 rounds?

Sure thing.... coming very soon!

[astracrazy]

thought this might be of interest, courtesy of Kiril Varbanov posted in another thread http://learncax.com/index.php/en/blog/b ... h-examples