CFD - Computational Fluid Dynamics, Motorsport, Formula 1

Here are our CFD links and discussions about aerodynamics, suspension, driver safety and tyres. Please stick to F1 on this forum.
User avatar
andylaurence
123
Joined: 19 Jul 2011, 15:35

Re: CFD - Computational Fluid Dynamics, Motorsport, Formula

Post

Just_a_fan wrote:I'd be hugely disappointed if we still have cars, or anything remotely like them, in 100 or 200 years. As for computing power, you're looking at having the equivalent of all of the world's supercomputers, squared, available for "shits and giggles". Sure, it might happen but only if there is nothing more interesting than studying an historic irrelevance - which is what F1 will be by then.
My cheap laptop right here has more processing power than the whole college where I studied (with 10000 other students). It has 4x the memory of the whole college and also has more than 100x the storage capacity of the college. That was 15 years ago. My phone has more memory and storage than the college and all the schools I attended put together. Given that the fastest supercomputers double in power every 14 months, it doesn't take long to realise that in 200 years, there will be a lot of power at people's fingertips. Whilst F1 might not exist at that point, I think it's a fair assumption that people will still be racing and that we'll still have an atmosphere, so CFD will be important. 200 years ago, people were still racing, but they raced horses rather than cars and that was the case for thousands of years before that.

Take a look at an Egyptian racing chariot hauled out of a tomb after 5000 years. They applied their technology to going faster with metal bushes to reduce friction and holes hacked out of the chassis rails to reduce weight.
Image

Just_a_fan
Just_a_fan
591
Joined: 31 Jan 2010, 20:37

Re: CFD - Computational Fluid Dynamics, Motorsport, Formula

Post

I'm aware that computing has moved in huge leaps over the years, but we're still talking about a system to track every molecule in a volume of the order of 22m^3 (to give a representative box around the vehicle). That's in the region of 6x10^26 discrete particles and all of their interactions. Those interactions will presumably be partly Newtonian and partly quantum.

People are using the current supercomputers (e.g. Titan) to carry out molecular interactions, some even at high speed, but they're looking at volumes on the molecular scale, not on the scale of a car.

And the argument about your phone being better than computers from 20 years ago is fine but that was driven by the huge advances in miniaturisation of computer components. There is a limit to how small we can make transistors which is why supercomputers just use lots of ordinary kit in a massively parallel way. I really struggle to see how we'll ever get something with the power of Oak Ridge's Titan next to our desks. Indeed, it's more likely that we'll just schedule time on a supercomputer somewhere else - much like Google runs a lot of its clever stuff on its own kit over the net and just dumps the result on to your phone. I suppose we might even use the world's workstations in a remotely parallel way; if everyone does that then you're no better off than you are using your own workstation though.

And if you can't have the equivalent of your own Titan by your desk then you've no chance of doing molecular CFD on a car.

Unless we really do crack quantum computing and can turn a cup of soup in to a hugely powerful quantum computer... :wtf: :lol:
If you are more fortunate than others, build a larger table not a taller fence.

User avatar
andylaurence
123
Joined: 19 Jul 2011, 15:35

Re: CFD - Computational Fluid Dynamics, Motorsport, Formula

Post

Just_a_fan wrote:And the argument about your phone being better than computers from 20 years ago is fine but that was driven by the huge advances in miniaturisation of computer components. There is a limit to how small we can make transistors which is why supercomputers just use lots of ordinary kit in a massively parallel way.
Yes, huge advances in miniaturisation of components made my phone more powerful in the same way as the PC from 20 years ago was hugely smaller than Colossus and the phone in 10 years will have more oomph than this laptop I'm typing on. Existing things improve and new things are invented all the time. Progress hasn't halted yet, so why should we assume it's going to slow down now?

In 1813, we couldn't drive as we'd not invented the car, we couldn't type as we'd not invented the computer and we couldn't fly because we'd not invented the plane. Getting around quickly was achieved on horseback but there were these new-fangled experimental devices called steam trains that people thought might be useful soon. If you wanted to send a message to someone far away, then you wrote the message down and paid someone to jump on a horse and ride to the other person with that message. That's how far we've come in 200 years. Analysing a car's aero at a molecular level looks like a small step in comparison, so I expect it to be possible in 2213.

stefan_
stefan_
696
Joined: 04 Feb 2012, 12:43
Location: Bucharest, Romania

Re: CFD - Computational Fluid Dynamics, Motorsport, Formula

Post

[youtube]http://www.youtube.com/watch?v=hTvWcpQl5TU[/youtube]


------------------------


"Pelican" vs."Non-Pelican" F1's car nose brief CFD analysis

The “non-pelican” nose:
Image

The “pelican” nose:
Image

As in the previous step-less vs. stepped F1's car nose brief CFD analysis:

- They have been used the real length between lines A-A and B-B (925 mm) besides the minimum external dimension (vertical) at the section A-A (275 mm) and the minimum external dimension (vertical) at the section B-B (400mm).

- It has been set, too, an airflow's speed of 50m/s because that's the maximum allowed by FIA (2013 Formula One Sporting Regulations) during wind tunnel testing [Rule 22.10].

This is the velocity vectors chart, for the “non-pelican” nose:
Image

Enlarged view of the above image:
Image

This is the velocity vectors chart, for the “pelican” nose:
Image

Enlarged view of the above image:
Image

Velocity vectors comparative chart for both noses:
Image

We can see, that the presence of the bulge under the nose increases airflow's speed.

Enlarged view of the above image:
Image

In addition to velocity vectors, we are able to analyse velocity and static pressure contours, too.

If we look at the contours of velocity, we can see how airflow's speed increases for the “pelican” nose case.
Image

Enlarged view of the above image:
Image

Static Pressure contours comparative chart for both noses:
Image

Enlarged view of the above image:
Image

For the “pelican” case, there will have a bigger static pressure gradient, in other words, the “pelican” nose (such as they[both noses] were drawn) will create “more downforce” than the “non-pelican” one.
We should note that, because we have not taken into account the front wing elements (main profile, flaps, etc.), after the iterative process we obtain “positive” values of the coefficient “cl” (downforce coefficient).
More than the values obtained for the downforce coefficient, what is truly important here, is its change as the “non-pelican” or the “pelican” one, may be the case.

“Non-pelican” nose results:
Image
“Pelican” nose results:
Image

We can see that clpelican<clnon-pelican, in other words, the presence of such a bulge underneath of the nose, induces a “more negative (smaller)” downforce coefficient (0.23<0.30).

Regards, MarleneKberg.
via Marlene K. Berg (SOURCE)
"...and there, very much in flames, is Jacques Laffite's Ligier. That's obviously a turbo blaze, and of course, Laffite will be able to see that conflagration in his mirrors... he is coolly parking the car somewhere safe." Murray Walker, San Marino 1985

shelly
shelly
136
Joined: 05 May 2009, 12:18

Re: CFD - Computational Fluid Dynamics, Motorsport, Formula

Post

Found this post on Gordon McCabe's blog:

http://mccabism.blogspot.it/2013/07/fro ... kdown.html

very interesting work by PhD student Jacques Heyder Bruckner comparing RANS vs DES in front wing-wheel aero.

Following the link on the post, http://eprints.soton.ac.uk/207263/the full file can be downloaded! Kudos to the author and to McCabe
twitter: @armchair_aero

User avatar
Alonso Fan
10
Joined: 06 Apr 2013, 18:21

Re: CFD - Computational Fluid Dynamics, Motorsport, Formula

Post

Does anyone know of any free cfd software?
SHR Modding
Youtube
Twitter
Discord

Sound Developer for Reiza Studios
Sound Modder for Assetto Corsa

User avatar
KeithYoung
24
Joined: 02 Jul 2003, 20:21
Location: USA

Re: CFD - Computational Fluid Dynamics, Motorsport, Formula

Post


User avatar
Alonso Fan
10
Joined: 06 Apr 2013, 18:21

Re: CFD - Computational Fluid Dynamics, Motorsport, Formula

Post

i mean free cfd visualisation software,

like stuff that can compete with Autodesk Project Falcon, or Autodesk Simulation etc

are there any available free?
SHR Modding
Youtube
Twitter
Discord

Sound Developer for Reiza Studios
Sound Modder for Assetto Corsa

User avatar
andylaurence
123
Joined: 19 Jul 2011, 15:35

Re: CFD - Computational Fluid Dynamics, Motorsport, Formula

Post

You can do CFD for free by modelling in Sketchup, processing with the Khamsin plugin and viewing the results in Paraview. All are freely downloadable from the vendors' websites.

User avatar
Alonso Fan
10
Joined: 06 Apr 2013, 18:21

Re: CFD - Computational Fluid Dynamics, Motorsport, Formula

Post

Thanks!
SHR Modding
Youtube
Twitter
Discord

Sound Developer for Reiza Studios
Sound Modder for Assetto Corsa

User avatar
Alonso Fan
10
Joined: 06 Apr 2013, 18:21

Re: CFD - Computational Fluid Dynamics, Motorsport, Formula

Post

ok, i'm not a master at sketchup.

but, i have the khamsin plugin, and i also have paraview, i have selected boundry conditions and defined all volumes necessary, but when i click on analysis, there is no Go(local) option in the drop down menu... there is Go(remote) but i can't remote into another pc, because i don't have one...

please tell me what to do... i'm pretty stuck...
SHR Modding
Youtube
Twitter
Discord

Sound Developer for Reiza Studios
Sound Modder for Assetto Corsa

User avatar
andylaurence
123
Joined: 19 Jul 2011, 15:35

Re: CFD - Computational Fluid Dynamics, Motorsport, Formula

Post

Alonso Fan wrote:ok, i'm not a master at sketchup.

but, i have the khamsin plugin, and i also have paraview, i have selected boundry conditions and defined all volumes necessary, but when i click on analysis, there is no Go(local) option in the drop down menu... there is Go(remote) but i can't remote into another pc, because i don't have one...

please tell me what to do... i'm pretty stuck...
Have you configured the local solvers? Select the Khamsin tool, then right-click > Configure > Step 3 (I think - it's called local solvers). Go (remote) is probably because you've selected a remote machine or Amazon AWS for processing.

User avatar
Alonso Fan
10
Joined: 06 Apr 2013, 18:21

Re: CFD - Computational Fluid Dynamics, Motorsport, Formula

Post

ok, there's no local solver

there's

Dolfyn solver
LTsinterfoam
buoyanbusinesssimplefoam
laplacianfoam
potentialfoam
simplefoam

it was preselected on simplefoam.

do i need anything else except paraview?

paraview is there in the tools section...
SHR Modding
Youtube
Twitter
Discord

Sound Developer for Reiza Studios
Sound Modder for Assetto Corsa

User avatar
KeithYoung
24
Joined: 02 Jul 2003, 20:21
Location: USA

Re: CFD - Computational Fluid Dynamics, Motorsport, Formula

Post

I use Paraview. In fact I'll be releasing an article with Post Processing from Paraview soon.

Image

Letting you guys know I uploaded a calculator I made to simulate the underfloor of a flat bottom race car. Let me know if you like it and if it works.

Couette Flow Race Car Ground Effect Velocity Profile Calculator

User avatar
andylaurence
123
Joined: 19 Jul 2011, 15:35

Re: CFD - Computational Fluid Dynamics, Motorsport, Formula

Post

Alonso Fan wrote:ok, there's no local solver

there's

Dolfyn solver
LTsinterfoam
buoyanbusinesssimplefoam
laplacianfoam
potentialfoam
simplefoam

it was preselected on simplefoam.

do i need anything else except paraview?

paraview is there in the tools section...
Paraview is for after you've done the solving. You've not got that far yet. When you have simplefoam selected and right-click > Configure, do all the tools have locations in the "Configure local third party solvers and tools" menu or do some of them have "*** Not Found ***"?