go_away

Author Topic: Computational Fluid Dynamics  (Read 23182 times)

0 Members and 1 Guest are viewing this topic.

gamefreak

• Supreme Robot
• Posts: 543
• Robo-Enthusiast
Computational Fluid Dynamics
« on: November 05, 2007, 11:48:47 AM »
Now as some of you know im currently working on an RC hovercraft, but im also a bit of a programmer and want to build a simulator, unfortunatly i dont know how to compute Fluid dynamics, any help on this?

All i know is that to make an accurate model then i need to do something with the pitch and speed of the propellor, then take that and do something with the duct size and gap, and then carry on with that, but what are the eqautions? I've found a few but they dont make much sense to me so a description would be nice.
All hail Rodney, the holy 555 timer
And Steve said: "Let there be lead!"

• Supreme Robot
• Posts: 11,680
Re: Computational Fluid Dynamics
« Reply #1 on: November 05, 2007, 05:20:11 PM »
Personally for what you are doing, the driving factor will be what you are capable building with your garage type tools . . . I would say you will probably only get a ~10% gain by using CFD optimization . . . and so its not worth your time . . .

but heck with my nay-saying, this might help:
http://www.societyofrobots.com/robotforum/index.php?topic=188.0

Basically, CAD your robot and save it as a .stl file. Then get some CFD software, applying your conditions (fluid flow field, borders, velocity, etc.), then map out a pressure gradient field. You want maximum pressure behind your vehicle and minimal pressure in front of your vehicle to maximize forward thrust . . . note that this doesn't correlate with controllability, just speed.

Oh and expect it to take 1 or 2 days to process on your computer

gamefreak

• Supreme Robot
• Posts: 543
• Robo-Enthusiast
Re: Computational Fluid Dynamics
« Reply #2 on: November 05, 2007, 05:44:19 PM »
is there a FREE CFD software? all i can find is stuff that seems for buisnesses, no free trials, just a number to call for quotes...

hmm student solidworks wont let me export to a .STL, it says it can but i cant find it..
« Last Edit: November 05, 2007, 06:25:02 PM by gamefreak »
All hail Rodney, the holy 555 timer
And Steve said: "Let there be lead!"

• Supreme Robot
• Posts: 11,680
Re: Computational Fluid Dynamics
« Reply #3 on: November 05, 2007, 07:24:42 PM »
Quote
is there a FREE CFD software?
thats a very good question . . . a very good one . . .

Quote
hmm student solidworks wont let me export to a .STL, it says it can but i cant find it..
im not sure about solidworks, but in most programs:
open up the part file (not the assembly file)
click save as...
you should then have the option to choose the file type.

oh, and CFD programs work much better the less complex/more smooth your part is . . . i actually have a secondary part file thats simplified just for CFD analysis . . .

ed1380

• Supreme Robot
• Posts: 1,478
Re: Computational Fluid Dynamics
« Reply #4 on: November 05, 2007, 08:48:02 PM »
is there a FREE CFD software? all i can find is stuff that seems for buisnesses, no free trials, just a number to call for quotes...

hmm student solidworks wont let me export to a .STL, it says it can but i cant find it..
legally free? I don't think so

but you have solidworks and IIRC solidworks can do a bunch of simulation stuff.

EDIT: cosmos floworks it's called
Problems making the \$50 robot circuit board?

gamefreak

• Supreme Robot
• Posts: 543
• Robo-Enthusiast
Re: Computational Fluid Dynamics
« Reply #5 on: November 06, 2007, 10:59:38 AM »
Yea i found some info about floworks this morning, but i cant find the option is solidworks...
All hail Rodney, the holy 555 timer
And Steve said: "Let there be lead!"

ed1380

• Supreme Robot
• Posts: 1,478
Re: Computational Fluid Dynamics
« Reply #6 on: November 06, 2007, 08:58:45 PM »
it's like a seperate thing. seperate licence.

I'm guessing you have it, but you need to activate it
Problems making the \$50 robot circuit board?

gamefreak

• Supreme Robot
• Posts: 543
• Robo-Enthusiast
Re: Computational Fluid Dynamics
« Reply #7 on: November 06, 2007, 09:57:14 PM »
nah, i got the solidworks from a BEST competition, i dont think they are kind enough to give me that thousand dollar software.

Does Autodesk Inventor have anything like this? I own the student version bu i haven't used it in awhile
All hail Rodney, the holy 555 timer
And Steve said: "Let there be lead!"

• Supreme Robot
• Posts: 11,680
Re: Computational Fluid Dynamics
« Reply #8 on: November 06, 2007, 11:30:29 PM »
Quote
Does Autodesk Inventor have anything like this?
nope

nottoooily

• Full Member
• Posts: 91
Re: Computational Fluid Dynamics
« Reply #9 on: November 07, 2007, 10:01:39 AM »
CFD can give you more accurate results. But for a rougher approximation in your case it's probably enough to use ordinary fluid flow finite element analysis software, possibly a 2D package if you have that kind of symmetry. There are many with free trial licences.

I know that for boat propellers there are simply tables of empirical data for all different diameters, pitch, speed, etc. And you can look up what the trust, efficiency, etc will be. Possibly the same exists for aeroplane propellers. You'd need to know what pressure/flow rate the propeller is producing to use FEA.

ed1380

• Supreme Robot
• Posts: 1,478
Re: Computational Fluid Dynamics
« Reply #10 on: November 07, 2007, 02:55:31 PM »
lookie what I accidently found
http://www.cfdesign.com/Products/inventor/default.asp

IIRC sollidworks student edition is about \$100, but has all features of regular solidworks, it just watermarks stuff
Problems making the \$50 robot circuit board?

• Supreme Robot
• Posts: 11,680
Re: Computational Fluid Dynamics
« Reply #11 on: November 07, 2007, 03:16:04 PM »
whoa a thermal fluids solver!!! i want!!!

Unfortunately they don't list a price (why does that not surprise me?) . . . so rather giving them my contact info so they can call me once a month, I browsed around and found a website dated nov 2006 that says CFdesign 9.0 is priced "From ?4,495"

and a dream ends.

ed1380

• Supreme Robot
• Posts: 1,478
Re: Computational Fluid Dynamics
« Reply #12 on: November 07, 2007, 03:57:34 PM »
*cough*torrents*cough*
Problems making the \$50 robot circuit board?

gamefreak

• Supreme Robot
• Posts: 543
• Robo-Enthusiast
Re: Computational Fluid Dynamics
« Reply #13 on: November 08, 2007, 06:44:20 AM »
so that is a add-on for inventor?

lol, why are all of these things so outrageously expensive.
All hail Rodney, the holy 555 timer
And Steve said: "Let there be lead!"

ed1380

• Supreme Robot
• Posts: 1,478
Re: Computational Fluid Dynamics
« Reply #14 on: November 08, 2007, 01:47:56 PM »
because it is a program for businesses and they can afford stuff like that, and they actually make a profit from it

PS- my computer pwns yours :p
Problems making the \$50 robot circuit board?

nottoooily

• Full Member
• Posts: 91
Re: Computational Fluid Dynamics
« Reply #15 on: November 08, 2007, 04:23:53 PM »
PS- my computer pwns yours :p

[/quote]

Haha, you have hypersonic blade tips. Bags not listening to those sonic booms!

• Supreme Robot
• Posts: 11,680
Re: Computational Fluid Dynamics
« Reply #16 on: December 13, 2007, 01:34:12 PM »
I got ahold of an unlimited use trial version of COMSOL

a few capabilities:
http://www.comsol.com/showroom/animations/

its normally worth \$10's of thousands . . . unfortunately my trial ends in 8 days . . .

if anyone wants me to simulate anything of yours, send me a CAD file with info on what you want simulated before then . . .

I might get it for work if it turns out to be useful enough to be worth spending all of your tax dollars on it

HDL_CinC_Dragon

• Supreme Robot
• Posts: 1,261
Re: Computational Fluid Dynamics
« Reply #17 on: December 13, 2007, 03:58:39 PM »
that is freakin cool, i just watched all the demo animated models... I want that program... but I have no real need for at the moment... but I still want it!
United States Marine Corps
Infantry
Returns to society: 2014JAN11

• Supreme Robot
• Posts: 11,680
Re: Computational Fluid Dynamics
« Reply #18 on: December 13, 2007, 05:47:47 PM »
I just started using it . . . its really buggy . . . wtf it just crashed while I typed 'buggy'!!!!

It was listening to me . . .

Anyway, I'm still learning how to use the software and figure out if its useful . . . Simulations take hours, mostly to find out I did it wrong and have to redo it . . . doh!

HDL_CinC_Dragon

• Supreme Robot
• Posts: 1,261
Re: Computational Fluid Dynamics
« Reply #19 on: December 13, 2007, 08:11:05 PM »
hahahaha well its good to hear feedback on it without wasting demo time
United States Marine Corps
Infantry
Returns to society: 2014JAN11

• Supreme Robot
• Posts: 11,680
Re: Computational Fluid Dynamics
« Reply #20 on: December 14, 2007, 01:36:29 PM »
hot off the presses!

after toying with COMSOL for 8 hours, I got it to mostly work . . .

here are a few images of my robot fish tail simulating fluid flow around it

• Supreme Robot
• Posts: 11,680
Re: Computational Fluid Dynamics
« Reply #21 on: January 24, 2008, 10:23:46 PM »
I decided to write a tutorial on it:
http://www.societyofrobots.com/mechanics_FEA.shtml

Enjoy!

ed1380

• Supreme Robot
• Posts: 1,478
Re: Computational Fluid Dynamics
« Reply #22 on: January 24, 2008, 10:41:26 PM »
great tutorial.

at the very end. isnt cosmos supposed to be cosmosm?
« Last Edit: January 24, 2008, 10:42:20 PM by ed1380 »
Problems making the \$50 robot circuit board?

• Supreme Robot
• Posts: 11,680
Re: Computational Fluid Dynamics
« Reply #23 on: January 24, 2008, 10:47:51 PM »
Nope, COSMOS:
http://www.cosmosm.com/

(the extra m stands for mo' money I think )

ed1380

• Supreme Robot
• Posts: 1,478
Re: Computational Fluid Dynamics
« Reply #24 on: January 24, 2008, 11:49:04 PM »
sorry.
i googled to be sure and cosmos didnt show anything
Problems making the \$50 robot circuit board?

mirali

• Beginner
• Posts: 1
Re: Computational Fluid Dynamics
« Reply #25 on: February 03, 2008, 04:11:32 PM »
Hey Guyz,

I am a new member to your group. I was reading the quest for a free CFD software. the good news is there is one try openfoam

gamefreak

• Supreme Robot
• Posts: 543
• Robo-Enthusiast
Re: Computational Fluid Dynamics
« Reply #26 on: February 03, 2008, 04:15:14 PM »
Admin, is it just me, or do I keep driving you over the edge and getting new tutorials made....
All hail Rodney, the holy 555 timer
And Steve said: "Let there be lead!"

• Supreme Robot
• Posts: 11,680
Re: Computational Fluid Dynamics
« Reply #27 on: February 03, 2008, 04:51:07 PM »
nah, I did the CFD tutorial mostly because I recently did a lot of CFD stuff, and I was in a mood to write about it.

kanzure

• Beginner
• Posts: 1
Re: Computational Fluid Dynamics
« Reply #28 on: March 05, 2008, 11:20:12 AM »
Hi all,

I found this website through the Wikipedia article on computational fluid dynamics. I thoroughly enjoyed the FEA tutorial that is posted up on the site, good job. It was interesting. However, there's one thing that made me register, and in particular it's that the majority of the users here think that CFD software costs tons of money. Where did this come from? Yeah, it's true that big businesses will shell out all of the money, but amateurs don't stand for that (I certainly don't).

So here's a list of free, open source CFD simulation packages. Have fun. I hear OpenFOAM is good. FreeFem++ is also interesting (has good web presentation). Currently I am trying to figure out how to get the preprocessing and postprocessing to work for me, for example how to define materials and the parameters for my simulation per each of the CFD simulation packages. It turns out that there's no standardization in the file formats. That's kind of odd. Sourceforge hosts a project for airplane simulators and there's an XML fileformat, but it's mostly for propellers and other specifications like that rather than abstract definitions for a variety of possible CFD simulations. So maybe we can change that?

Anyway, I have to get going, but if anybody wants to discuss this with me, hit me up over email or the IM protocols.

- Bryan

• Supreme Robot
• Posts: 11,680
Re: Computational Fluid Dynamics
« Reply #29 on: March 05, 2008, 11:39:39 AM »
Quote
However, there's one thing that made me register, and in particular it's that the majority of the users here think that CFD software costs tons of money. Where did this come from?
Mostly from my inability to find capable free software with documentation . . .

Which brings me to my next point, thanks kanzure for finding this!

Keep us updated on any of the packages you try out and your thoughts on them.