Khamsin Virtual Racecar Challenge 2014

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
variante
138
Joined: 09 Apr 2012, 11:36
Location: Monza

Re: Khamsin Virtual Racecar Challenge 2014

Post

First of all: congratulation to Savage for the great win! You really know how to design great cars!

And, Astra, that race visual was a nice touch. Very enjoyable :D

BTW i've got a question concerning the displayed data: why have you chosen to show the Cl.A and Cd.A instead of the coefficients of lift and drag? And what do they mean exactly?

PS: you can use the pictures of my car that you find in the forum, obviously.

julien.decharentenay
julien.decharentenay
10
Joined: 02 Jun 2012, 12:31

Re: Khamsin Virtual Racecar Challenge 2014

Post

Hi. Thanks a lot for the feedback and congratulation to @cdsavage and all participants. Well done.

Race 1 Monaco:
Result page is at http://www.khamsinvirtualracecarchallen ... 14/results

And the race lap visualisation is at https://googledrive.com/host/0B9oqdnOAD ... onaco.html. The visualisation is provided for entertainment purposes: only final position and lap times are determined using Visual Stopwatch. Intermediate speed and position are provided for entertainment purposes only and are not indicative of actual performance...

Cl.A and Cd.A are downforce area and drag area. They are the product of the downforce coefficient by the frontal area and the drag coefficient by the frontal area. These coefficients are the one used in Virtual Stopwatch as they represents normalised drag and downforce.

The downforce and drag coefficient themselves may give an erroneous impression as a car may have a low drag coefficient but a large frontal area. Although I agree that in the KVRC challenge case, it is likely that the frontal area are reasonably close.

cdsavage
cdsavage
19
Joined: 25 Apr 2010, 13:28

Re: Khamsin Virtual Racecar Challenge 2014

Post

Well done everybody on a succesful round 1. Looks like I have some work to do on CoP. Is there any chance of seeing graphs for laptime /df /drag for the next round(s)?
variante wrote:BTW i've got a question concerning the displayed data: why have you chosen to show the Cl.A and Cd.A instead of the coefficients of lift and drag? And what do they mean exactly?
Should be Cd.A = drag/1225 and -Cl.A = downforce/1225, this way you dont need to worry about calculating the frontal area as you would need to do for Cd and Cl.
julien.decharentenay wrote:Also: I will be updating the back-end to CFD Apps next Monday - I tried very hard to not touch it while the race is on. The update should improve turn-around time for the generation of openFoam cases and add a couple of feature, but there should be no change in simulation process (i.e same parameters and the like). So the results should be the same. There will probably be some issues for a couple of days (let me know so I can fix them asap).
Will simulations run previously be directly comparable to any new simulations, ie will the mesh, boundary conditions etc be identical?

One minor thing I noticed with the portal-generated cases is that the refinement zone seems to be slightly off-center when viewed from above, my preference would be to keep it as it is though so results are still directly comparable.

User avatar
CAEdevice
49
Joined: 09 Jan 2014, 15:33
Location: Erba, Italy

Re: Khamsin Virtual Racecar Challenge 2014

Post

cdsavage wrote:Well done everybody on a succesful round 1. Looks like I have some work to do on CoP. Is there any chance of seeing graphs for laptime /df /drag for the next round(s)?
I have a similar question: would it be possibile to have (in addition to the graph with time lap lines depending on drag and lift with constant COP) a complementary graph showing what happens to timelap when Drag/Lift/FrontalSection are constant and only the pressure distribution changes (es. COP = 1.6/1.8/2.0). I've seen a similar graph in the 2013 thread.

Another request (sorry Machin I don't want you to waste too much time): I would be glad to know the total energy "consumed" by each car (in the hypotesis to spend all accumulated electric energy in the same time lap).

astracrazy
astracrazy
31
Joined: 04 Mar 2009, 16:04

Re: Khamsin Virtual Racecar Challenge 2014

Post

this is a strange question for me to be asking (I have asked Julien but no reply yet)

does anyone else use khamsin on there own? if so how far out are your results? my results are 3x better than the challenge. I'm not using rotating wheels and mesh size i put in is 0.02 when setting it up.

I thought it was the tunnel size too small so I done a bigger one but it was the same result.

It was out last year (because of rotating wheels i guess) but not by this much.

Any ideas what i'm doing wrong? As i've said before I'm not too good with the cfd side of it. I know it won't be bang on but i would expect it to be closer.
Last edited by astracrazy on 17 Mar 2014, 17:29, edited 1 time in total.

User avatar
CAEdevice
49
Joined: 09 Jan 2014, 15:33
Location: Erba, Italy

Re: Khamsin Virtual Racecar Challenge 2014

Post

Sorry, I'm not using OpenFoam/Khamsin yet.
If it could help, I (almost) solved my COP convergence problem: I was using a wrong C_SYS :(

Are you using an a bigger enough box? Rotating wheels can change te results but not so much (in my experience +/-50% not +/-300%)

astracrazy
astracrazy
31
Joined: 04 Mar 2009, 16:04

Re: Khamsin Virtual Racecar Challenge 2014

Post

its basically cl.a is 3... something instead of 1....

i tried the larger box posted in this thread and it was the same result (only a minor difference)

User avatar
CAEdevice
49
Joined: 09 Jan 2014, 15:33
Location: Erba, Italy

Re: Khamsin Virtual Racecar Challenge 2014

Post

What about turbulence parameters? Rotating wheels direction is right?
I don't know how rotating wheels are defined in Khamsin, but with my solver I noticed I used the wrong C_SYS after 2 weeks of simulations!

astracrazy
astracrazy
31
Joined: 04 Mar 2009, 16:04

Re: Khamsin Virtual Racecar Challenge 2014

Post

not using rotating wheels. see what julien comes back with

User avatar
variante
138
Joined: 09 Apr 2012, 11:36
Location: Monza

Re: Khamsin Virtual Racecar Challenge 2014

Post

@astra

i can't simulate rotating wheel too. To compensate i decreased the ride height to 35mm.
I use a mesh size of 0.005m for the car... This should be the most important factor.
Wind tunnel's dimensions are limited (way smaller than those of the official run), and i only test half the car.

The results are very similar to the official ones, even the COP.

User avatar
RicME85
52
Joined: 09 Feb 2012, 13:11
Location: Derby

Re: Khamsin Virtual Racecar Challenge 2014

Post

Not sure whats going on Luke, my results seem to be pretty spot on although I cannot say for definite as the version of my car you have is slightly different to the one I have results for.

User avatar
machin
162
Joined: 25 Nov 2008, 14:45

Re: Khamsin Virtual Racecar Challenge 2014

Post

The aero coefficients vs lap speed charts for the next two rounds are below (I have kept the convention of Cl being positive for downforce, so if you are using the other convention (positive for lift) remember to swap the sign!

Image

Image

I've extended the range to encompass the sort of results you were all getting in the last round. You can see that both of the above tracks reward a much lower drag set-up than Monaco.

Balance vs lap time is shown below (in this case for magny cours for a cars have Cd.A = 1.6 and Cl.A = 2.8, but the trend will be similar for other tracks):-

Image
Another request (sorry Machin I don't want you to waste too much time): I would be glad to know the total energy "consumed" by each car (in the hypotesis to spend all accumulated electric energy in the same time lap).
I'm afraid I would have to re-write the program to obtain this energy data. From a design point of view it wouldn't help you design your car; you should be aiming to increase your downforce, decrease your drag, and get your COP as close to 1.83m as possible.

If it were me doing the challenge my approach would be to design three basic cars; low downforce, medium downforce, and high downforce. I would then run some CFD to obtain basic aero cofficients for the three arrangements. I would then plot the three cars on the lap speed chart for the up-coming race, as I have done below:-

Image

In this example I find that my medium downforce set-up is the fastest. So I would use this set up as the basis for the Magny Cours race, and I would try and work to make small changes to improve the efficiency of that design....
COMPETITION CAR ENGINEERING -Home of VIRTUAL STOPWATCH

astracrazy
astracrazy
31
Joined: 04 Mar 2009, 16:04

Re: Khamsin Virtual Racecar Challenge 2014

Post

I had to make a partition on my mac to be able to run khamsin and i obviously didn't make it big enough because 0.02 is the lowest i can go. This is prob the problem I wonder? I should re-do the partition really but my internet it so slow it will take a long time to re-download everything

this is the force and moment i'm getting. I don't know if this gives away anything
(15.9046 4630.57 -5747.58) (-0.0275366 48.2703 3.14803)
(-11346 -16.0619 43.8073) (-9.01843 -0.0449242 0.169445)

CottrellGP
CottrellGP
4
Joined: 02 Sep 2011, 01:48

Re: Khamsin Virtual Racecar Challenge 2014

Post

Hi guys, im in this competition and i dont even know how to analyse my results and i dont know which figures mean what, can someone help and these graphs that keep getting posted i dont know how to read them? thanks Dan
Dan Cottrell

Master Of Innovation!

astracrazy
astracrazy
31
Joined: 04 Mar 2009, 16:04

Re: Khamsin Virtual Racecar Challenge 2014

Post

cl.A is the downforce (usually a negative number as its actually lift it refers too, but simplicity wise we keep it positive), cd.a is the drag (both based on area hence the a, but don't worry about that just take the number for what it is) and cop is the centre of pressure (in simple terms the balance of the car downforce wise). if its 1 its way forward, 3 its way back.
you always want the cl.a to be higher than cd.a. The bigger the gap the better - i.e then you'd be creating a lot of downforce and little drag (the goal)

the graphs are what you should aim for if you are doing your own testing. so lets take this graph

Image

left is obviously time and along the bottom is drag (cd.A)
machin has then drawn two lines based on cl.a = 1 and cl.a = 2.8
so for the perfect lap time your car would need a cl.A value of 2.8 and a cd.A value of 1, worst lap time cl.A = 1 and cd.A = 1.6

so if you take your results from monaco and you changed nothing the lap time you would get roughly is say 1:18.2. Why? Because you cd.A is 1 and you cl.a is 1.4

don't worry about the balance graph, just aim for roughly 1.8, you have 1.9 so don't worry to much.