EcoModder.com

EcoModder.com (https://ecomodder.com/forum/)
-   Aerodynamics (https://ecomodder.com/forum/aerodynamics.html)
-   -   Flow Illustrator - a simple windtunnel simulator (https://ecomodder.com/forum/showthread.php/flow-illustrator-simple-windtunnel-simulator-3709.html)

thirdman 07-11-2008 11:36 AM

Flow Illustrator - a simple windtunnel simulator
 
Howdy, all... just thought I would share this nifty FREE wind tunnel simulator that I sumbled across with y'all. This is not an ad, I am not affiliated with them in any way... it is just a cool tool for playing with aerodynamics.

Warning: it is VERY simple, 2-D and any results that it gives should be considered "rough estimates".

That being said, here is is:
Flow Illustrator
Flow Illustrator home page

Be sure to read the "Home" and "More Info" pages before diving in. There are some limitations to the movies that are not documented. It appears to die if you create a movie that is greater than 2mb. The best combination I have found is to use black-and-white BMP silhouettes that are 256x128 and the default settings with the exception of the time, which I cannot get to go longer than 25 seconds. YMMV.

Anywho, to see what I am talking about, here are some frame grabs I got from a simulation I did for my Ford Festiva, investigating different spoiler designs:

Stock Car:
http://farm4.static.flickr.com/3078/...24dda36e_m.jpg

Attached Spoiler:
http://farm4.static.flickr.com/3133/...47c5b150_o.jpg

Detached Spoiler:
http://farm4.static.flickr.com/3263/...344fa1fb_o.jpg

Again, the program itself kicks out movies, and the above are just frame grabs from said movies. Note that for the simulations that I did, it took 10-15 seconds for the simulation to "settle down" and for stable patterns to appear. So keep your BMPs small to allow for maximum run times.

Everything I know about this program I got from their website, or I have noted above. Please don't ask me any hard questions, as I promise I won't know the answers!! :)

thirdman 07-11-2008 11:45 AM

Also, my conclusions so far are that the following are good things that are easy to accomplish:
1) A small, flat, parallel to the ground, level with the roof spoiler.
http://farm4.static.flickr.com/3221/...1b71851d_o.gif
2) Reduced airflow under the car, either through lowering or air dam & skirts.

No big surprise there, right? ;) It is just really cool to SEE how this stuff can change air flow.

As always, YMMV. ;)

i_am_socket 07-11-2008 12:07 PM

now if I just knew what Re and dt are supposed to represent...

thirdman 07-11-2008 12:16 PM

dt has something to do with the speed of the "wind". Very small changes make a big difference... going from .01 to .05 made it show standing shockwaves all along the hood, so I am guessing that was REALLY fast. I am doing some runs now with dt's in the .02 and .025 range to see if I can tell what is going on...

texanidiot25 07-11-2008 12:20 PM

That's way cool. I'm gonna toss in my '69 CST/10 for kicks

i_am_socket 07-11-2008 12:48 PM

Doesn't seem to do as well with top-down views of my car as it does side views. Very nifty and can't beat the price :)

FastPlastic 07-11-2008 12:50 PM

Great find :thumbup: I've been lookin for something like this to better understand flow.

co_driver 07-11-2008 12:55 PM

Quote:

Originally Posted by i_am_socket (Post 43086)
now if I just knew what Re and dt are supposed to represent...

Reynolds Number = ( L * v ) / 'nu'

L = Effective length (or diameter of pipe) [m]
v = velocity [m/s]
'nu' = Greek letter 'n' = kinematic viscosity [m^2/s] (~1.6*10^-5 for air @ 20C)

Re is an indicator of effective flow when comparing scale models of real things (3/8 scale, or ?) - 2X speed for something half the size. So, in the modelling programs, a correct Re must be used.

For nearly all objects there is a typical Re number where transition occurs from laminar to turbulent flow. (a Moody diagram [pipe flow] shows this well)

texanidiot25 07-11-2008 01:11 PM

Quote:

Originally Posted by co_driver (Post 43106)
Reynolds Number = ( L * v ) / 'nu'

L = Effective length (or diameter of pipe) [m]
v = velocity [m/s]
'nu' = Greek letter 'n' = kinematic viscosity [m^2/s] (~1.6*10^-5 for air @ 20C)

Re is an indicator of effective flow when comparing scale models of real things (3/8 scale, or ?) - 2X speed for something half the size. So, in the modelling programs, a correct Re must be used.

For nearly all objects there is a typical Re number where transition occurs from laminar to turbulent flow. (a Moody diagram [pipe flow] shows this well)

:confused:

So what are the best numbers?

88CRX 07-11-2008 01:25 PM

Awesome, Thanks!!!

texanidiot25 07-11-2008 01:34 PM

BTW, do use the dimensions he gave, and if your creating your own silhouettes, do not use any anti-aliasing or it screws up the uploaded image.

http://www.youtube.com/watch?v=LN1XPOsxPB0
^it's still processing, but shouldnt take long

cfg83 07-11-2008 02:00 PM

thirdman -

THANK YOU!

Can someone post good city/highway settings for the parameters?

CarloSW2

co_driver 07-11-2008 03:58 PM

Quote:

Originally Posted by texanidiot25 (Post 43111)
:confused:

So what are the best numbers?

Numbers to use:
L* = effective diameter of your car [1990 Firebird: (72.4" (width) * 49.8" (height) )^0.5 = ~60" = 1.53 m]
v = velocity [50 mph = 22.35 m/s]
'nu' = 1.6*10^-5 m^2/s

Thus: Re = 1.53 * 22.35 / 1.6*10^-5 = ~2,100,000

(*) Note:
A better reference would be the effective diameter rather than its length, as length is used for flat plates.

Now, on to making 2-D renditions of vehicles I have once owned...

texanidiot25 07-11-2008 04:05 PM

Quote:

Originally Posted by co_driver (Post 43180)
Numbers to use:
L* = effective diameter of your car [1990 Firebird: (72.4" (width) * 49.8" (height) )^0.5 = ~60" = 1.53 m]
v = velocity [50 mph = 22.35 m/s]
'nu' = 1.6*10^-5 m^2/s

Thus: Re = 1.53 * 22.35 / 1.6*10^-5 = ~2,100,000

(*) Note:
A better reference would be the effective diameter rather than its length, as length is used for flat plates.

Now, on to making 2-D renditions of vehicles I have once owned...

Alright, thanks!

Here's a few more I made out of boredom. One is my friend's 04 Avalanch, and for kicks I did my '69 CST10. Though these are with the default numbers, and the windspeed one at .1

YouTube - Chevy Avalanch:: Flow Illustrator
YouTube - 1969 CST10:: Flow Illustrator

Blue Bomber 07-11-2008 05:13 PM

Does this determine flow based on a flat 2d plane, or does it also take the width of the vehicle into account? Wondering if when making a truck shape, should the open bed be taken into account...

i_am_socket 07-11-2008 09:16 PM

Yes, it only does 2D calculations. It has no idea that the bed of your truck is empty or that there are dangly bits all over the place under the hood.

If you want that kind of data, though, you could do cross-sectional images (ie, exclude the sides of the bed)

cfg83 07-11-2008 09:18 PM

Hello -

Sooooooo, what should dt be for 35 MPH and 55 MPH?

CarloSW2

texanidiot25 07-11-2008 09:53 PM

Man this program has a total freakout when I set the RE to 2,100,000 and then the dt to .02.

Was testing the flow of air from a lower scoop (like it had for the radiator) toa heat extracter hood vent like some T/As had.

AndrewJ 07-11-2008 10:48 PM

Dammit, I got all the BMP's made but my mac won't play the video files back to me. :(

Time to do some detective work on finding the right codecs or whatnot.

adrive7 07-12-2008 12:07 AM

For a pickup truck I would leave off the side walls of the bed. Draw the Cab, the floor of the bed, and the tailgate. Otherwise you don't see the effect of the vortex in the bed. Leaving them on shows what it would look like with a tonneau cover.

texanidiot25 07-12-2008 01:39 AM

Quote:

Originally Posted by AndrewJ (Post 43281)
Dammit, I got all the BMP's made but my mac won't play the video files back to me. :(

Time to do some detective work on finding the right codecs or whatnot.

They are .avi videos. Try downloading DivX, if they have it for MAC then it's a player that supports .avi

Quote:

Originally Posted by adrive7 (Post 43304)
For a pickup truck I would leave off the side walls of the bed. Draw the Cab, the floor of the bed, and the tailgate. Otherwise you don't see the effect of the vortex in the bed. Leaving them on shows what it would look like with a tonneau cover.

Done:

OPEN:
YouTube - Truck bed test, open bed

TANO COVER:
YouTube - Truck bed test, Tano cover

FASTBACK:
YouTube - Truck bed test, Fast back

(give them time to process)

AndrewJ 07-12-2008 02:48 AM

Quote:

Originally Posted by texanidiot25 (Post 43321)
They are .avi videos. Try downloading DivX, if they have it for MAC then it's a player that supports .avi

No dice, I tried downloading DivX and XviD codecs, but still no luck...


Anybody feel like helping a fella out?

http://i239.photobucket.com/albums/f...Kammbitmap.jpg

http://i239.photobucket.com/albums/f...r/K-Normal.jpg

http://i239.photobucket.com/albums/f...K-boattail.jpg

cfg83 07-12-2008 02:51 AM

AndrewJ -

Quote:

Originally Posted by AndrewJ (Post 43334)
No dice, I tried downloading DivX and XviD codecs, but still no luck...

Why not upload it to YouTube? You can delete it afterwards if you don't like it.

CarloSW2

cfg83 07-12-2008 03:11 AM

EDIT: SUPERCEDED BY POST #65 :
http://ecomodder.com/forum/45192-post65.html


Hello -

I didn't change any of the default numbers except for duration (40 seconds), so maybe this is all wrong. First, I did a stock analysis on my SW2 :


1999 Saturn SW2 Aero-Analysis
...removed to save download time...

After around 6 seconds, a "Virtual Kamm-Back" appears (at least that's what I call it). Then I did an analysis with a Kamm-Back based on MetroMPG's rain-drop example, ending at the bumper :

1999 Saturn SW2 + Kamm-Back Aero-Analysis
...removed to save download time...

Finally, I went back to the first analysis and added the "Virtual Kamm-Back" :

1999 Saturn SW2 + Kamm-Back II Aero-Analysis
...removed to save download time...

I like the Virtual Kamm-Back because it would mean not having to worry about moving the 3rd brakelight and maintaining almost full rear visibility. It also seemed to produce the most stable swirly-wake during the run. A nice compromise.

Question: Could my Virtual Kamm-Back be "tailored to the program"? What I mean is, could I be making a shape that the program likes?

CarloSW2

Chris D. 07-12-2008 04:07 AM

I had no luck getting the player to work for me.. :(

I uploaded it, it loaded and let me download the video, but it wouldnt play on anything I have..

I think I have the sizes right.. donno..
http://i4.photobucket.com/albums/y10...s/areosilo.jpg

http://i4.photobucket.com/albums/y10...eosiloMINI.jpg

i_am_socket 07-12-2008 12:05 PM

The issue with the videos is that they are AVI's encoded with WMV3 which is the format for Windows Media Player 9. Took me the better part of an hour just to be able to grab a frame to do side by side comparisons.

From all the googling I did on that, it doesn't look like there's much to help out the Mac peoples. Since the codec's a bit old, there might be something you guys can use. HERE's a thread from 2005 on the subject.

trikkonceptz 07-12-2008 12:31 PM

For us Mac guys, how about using a program like popcorn, which converts file types to just about anything ...

adrive7 07-12-2008 12:43 PM

If you are having trouble viewing the files, download VLC Media Player. It plays pretty much anything, including corrupted files.

VLC media player - Overview

Good for windows or Mac.

cfg83 07-12-2008 12:56 PM

Chris -

I processed yours and put it on YouTube :

YouTube - Chris Ds Taco

I made the Taco smaller and put a road underneath.

I'll delete it from YouTube if you want me to.

CarloSW2

Blue Bomber 07-12-2008 07:20 PM

Did a bunch of vids of my car in different scenarios:

YouTube - 93 MR2 - Windows Closed - Flow Illustrator
Windows Closed

YouTube - 93 MR2 - Windows Open - Flow Illustrator
Windows Open

YouTube - 93 MR2 - Windows Closed + Rear Diffuser - Flow Illustrator
Windows Closed + Rear Diffuser



YouTube - 93 MR2 - Windows Closed + Vented Hood - Flow Illustrator
Vented Hood

YouTube - 93 MR2 - Windows Closed + Hood + Diffuser - Flow Illustrator
Vented Hood + Rear Diffuser



Only problem I had was that I didn't know what numbers to enter into the program besides the movie time. The only thing I'd want to mess around with them for is aerodynamic difference between around 50 and 80mph, since those are the speeds I travel most.


Does this program use gravity in any of these movies? If not, then maybe we could do top-down movies, testing things like mirrors and front/rear shapes, as well.

adrive7 07-12-2008 08:00 PM

For Reynolds number use 2,500,000. At car speeds, and since this is air, the Reynolds number is always going to be around this for what we are interested in. dt is the time step of the movie, so keep it small. Gravity doesn't play into this simulation, so your top down view will work fine (you just wont see the effect of the air moving under and over the vehicle.

AndrewJ 07-13-2008 02:02 AM

adrive7: Thanks, the VLC program worked like a charm.

here's a few vids of old Harold in action.

Chris D. 07-14-2008 12:02 AM

Quote:

Originally Posted by cfg83 (Post 43392)
Chris -

I processed yours and put it on YouTube :

YouTube - Chris Ds Taco

I made the Taco smaller and put a road underneath.

I'll delete it from YouTube if you want me to.

CarloSW2

right on, thanks..

how did you get it to work? I wasn't clear on the settings.. :confused:

can you send it to me? :D PentastarTurbo@aol.com

for the record, VLC didn't play it for me :( already had it..

cfg83 07-14-2008 12:36 AM

Chris D -

The only value I changed was the time. I changed it to 40 seconds. Now I think the YouTube is wrong because, according to adrive7, the Re needs to be 2,500,000.

Quote:

Originally Posted by Chris D. (Post 43726)
right on, thanks..

how did you get it to work? I wasn't clear on the settings.. :confused:

can you send it to me? :D PentastarTurbo@aol.com

for the record, VLC didn't play it for me :( already had it..

I will do the same analysis with an Re of 2500000 and send you an e-mail.

CarloSW2

Peakster 07-14-2008 12:44 AM

This is just too much fun!!! Totally found my new favorite toy :D.

http://www.youtube.com/watch?v=dNEDdJhnnKQ

cfg83 07-14-2008 02:06 AM

1 Attachment(s)
co_driver -

Quote:

Originally Posted by co_driver (Post 43180)
Numbers to use:
L* = effective diameter of your car [1990 Firebird: (72.4" (width) * 49.8" (height) )^0.5 = ~60" = 1.53 m]
v = velocity [50 mph = 22.35 m/s]
'nu' = 1.6*10^-5 m^2/s

Thus: Re = 1.53 * 22.35 / 1.6*10^-5 = ~2,100,000

(*) Note:
A better reference would be the effective diameter rather than its length, as length is used for flat plates.

Now, on to making 2-D renditions of vehicles I have once owned...

I'll plug my SW2 numbers in :

L* = effective diameter of my car
[1999 Saturn SW2: (66.7" (width) * 54.5" (height) )^0.5 = ~60.29" => 1.53 m]
======> Same as Firebird Example, so the rest of my numbers will be the same.

v50 = velocity [50 mph = 22.35 m/s]
v55 = velocity [55 mph = 24.59 m/s]
v60 = velocity [60 mph = 26.82 m/s]
v65 = velocity [65 mph = 29.06 m/s]

'nu' = 1.6*10^-5 m^2/s

Thus: Re50 = 1.53 * 22.35 / 1.6*10^-5 = ~2,140,000
Thus: Re55 = 1.53 * 24.59 / 1.6*10^-5 = ~2,350,000
Thus: Re60 = 1.53 * 26.82 / 1.6*10^-5 = ~2,560,000
Thus: Re65 = 1.53 * 29.06 / 1.6*10^-5 = ~2,780,000

I've been using 2,500,000 for my second run, so that means I've been calculating for a little under 60 MPH.

Based on the above, I have attached a super-simple Reynolds Number calculator using Microsoft Excel. Enjoy!

CarloSW2

Peakster 07-14-2008 04:08 AM

You know I was just curious :p:

YouTube - Example of high wind resistance

i_am_socket 07-14-2008 11:09 AM

I'm getting some interesting movies using the high Re numbers.

I get a lot of this:
http://spidermouse.us/modding/car/stangMovie.gif

using 2,500,000. Is it supposed to look like that?

Chris D. 07-14-2008 11:10 AM

Quote:

Originally Posted by i_am_socket (Post 43818)
I'm getting some interesting movies using the high Re numbers.

I get a lot of this:
http://spidermouse.us/modding/car/stangMovie.gif

using 2,500,000. Is it supposed to look like that?

2000 era mustang, cool :thumbup:

mite have some slight color differences on the background causing the turbulence..

mite have to go over the whole background in one color, try that and repost it..

i_am_socket 07-14-2008 11:15 AM

Here's the source image:
http://spidermouse.us/modding/Car/stang_flowshape.bmp

2 bit, no antialiasing. I'll try bumping it up to 8 bit.

*EDIT*

Here's the video on 8 bit. (haven't figured out the embedding thing yet)

http://www.youtube.com/v/7w_nCKP-co4


All times are GMT -4. The time now is 02:27 AM.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.
Content Relevant URLs by vBSEO 3.5.2
All content copyright EcoModder.com