Go to Post Ok maybe you cant stop them, but you can utilize that pretty flashing LED pretty darn well! - Andy Grady [more]
Home
Go Back   Chief Delphi > Technical > CAD > Inventor
CD-Media   CD-Spy  
portal register members calendar search Today's Posts Mark Forums Read FAQ rules

 
Closed Thread
Thread Tools Rate Thread Display Modes
  #1   Spotlight this post!  
Unread 23-01-2004, 11:43
Neal Turett Neal Turett is offline
Registered User
#0201
 
Join Date: Nov 2002
Location: Rochester Hills
Posts: 34
Neal Turett is on a distinguished road
FEA Modeling

Does anyone know of a good piece of FEA modeling software, perferable that interacts with Inventor, and a team could get their hands on?

We just want to see if our bot will explode or not.
  #2   Spotlight this post!  
Unread 23-01-2004, 12:30
Kevin Sevcik's Avatar
Kevin Sevcik Kevin Sevcik is offline
(Insert witty comment here)
FRC #0057 (The Leopards)
Team Role: Mentor
 
Join Date: Jun 2001
Rookie Year: 1998
Location: Houston, Texas
Posts: 3,731
Kevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond repute
Send a message via AIM to Kevin Sevcik Send a message via Yahoo to Kevin Sevcik
Re: FEA Modeling

FEA software tends to be rather excitingly expensive. And requires hardware capable of ludicrous speed to analyze anything more complicated than a single part.

To top it off, it's rather easy to misapply it and misinterpret the results if you don't know what you're doing.

On the other hand, plain old hand calculations are cheap, and can probably get you enough accuracy to satisfy. If you don't know how to do it, find a MechE college student or a textbook on strength of materials. Half-priced books or Half.com should provide.

If all that fails, then someone here should probably write up a brief whitepaper on rudimentary strength calculations.

If I must recommend something, though, Ansys DesignSpace would probably work well. It appears that the educational edition, which has limited functionality, would run you $320.
__________________
The difficult we do today; the impossible we do tomorrow. Miracles by appointment only.

Lone Star Regional Troubleshooter

Last edited by Kevin Sevcik : 23-01-2004 at 12:47.
  #3   Spotlight this post!  
Unread 23-01-2004, 15:50
ChrisH's Avatar Unsung FIRST Hero
ChrisH ChrisH is offline
Generally Useless
FRC #0330 (Beach 'Bots)
Team Role: Engineer
 
Join Date: Jun 2001
Rookie Year: 1998
Location: Hermosa Beach, CA
Posts: 1,230
ChrisH has a reputation beyond reputeChrisH has a reputation beyond reputeChrisH has a reputation beyond reputeChrisH has a reputation beyond reputeChrisH has a reputation beyond reputeChrisH has a reputation beyond reputeChrisH has a reputation beyond reputeChrisH has a reputation beyond reputeChrisH has a reputation beyond reputeChrisH has a reputation beyond reputeChrisH has a reputation beyond repute
Re: FEA Modeling

Quote:
Originally Posted by Kevin Sevcik
FEA software tends to be rather excitingly expensive. And requires hardware capable of ludicrous speed to analyze anything more complicated than a single part.

To top it off, it's rather easy to misapply it and misinterpret the results if you don't know what you're doing..
The first part isn't necessarily true, but the second sure is. How long an FEA takes all depends on the level of detail in your model. For FIRST robots a coarse model of a 100 or so nodes or less should be adequate. Such a model should run in a couple of minutes on all but the most ancient PCs.

However a minor error like an incorrect end constraint or release can give you a very different answer. Depending on the program you are using, it can be hard to even find what you've assigned them to be. You kind of have to know what you're expecting before you look at the results to see if your model is behaving properly. That's how I find most of my mistakes.

Quote:
Originally Posted by Kevin Sevcik
On the other hand, plain old hand calculations are cheap, and can probably get you enough accuracy to satisfy. If you don't know how to do it, find a MechE college student or a textbook on strength of materials. Half-priced books or Half.com should provide.

If all that fails, then someone here should probably write up a brief whitepaper on rudimentary strength calculations.
There is some info on this in my whitepaper on Lifts and Arms. If you have a very specific question about a general problem maybe we can start a thread to discuss how to solve it. For example, if you're trying to build a telescoping lift that will reach 10ft without breaking. That would be a specific problem of general interest, but the answer would be very different than how to build an arm to pick up a big ball at a distance of 5 ft. (another poblem of general interest)

.
Quote:
Originally Posted by Kevin Sevcik
If I must recommend something, though, Ansys DesignSpace would probably work well. It appears that the educational edition, which has limited functionality, would run you $320.
Norton Engineering used to offer a very cheap student version of their NASTRAN (the grandaddy of all FEA programs) program. The NE student version would run a 300 node model and has the advantage of using the NASTRAN standard formats, which ALL FEA programs will read. I got mine several years ago and I'm not sure they are even still in business. I haven't used Ansys so I can't comment on that.

FEA programs in general are by nature pretty user-unfriendly. They are used to produce mathematical represenations of how things might behave in the real world, if certain assumptions are in fact correct. Picking the right math and assumptions is an art unto itself. Go ahead and explore FEA, it is a useful skill and you might be one of those rare people with a talent for it. But don't bet the robot on the answers you get unless you have proffessional guidance. The textbook approach Kevin suggested is much cheaper and safer.

I have dabbled in FEA for our robot designs. I might run a model or two after our robot is designed in an effort to catch any "Oh My God!s". But even so I will get guidance from one or more of the stress analysts I know at work. I don't do it enough to be proficient enough to get useful results in a hurry.
__________________
Christopher H Husmann, PE

"Who is John Galt?"
  #4   Spotlight this post!  
Unread 23-01-2004, 18:05
Kevin Sevcik's Avatar
Kevin Sevcik Kevin Sevcik is offline
(Insert witty comment here)
FRC #0057 (The Leopards)
Team Role: Mentor
 
Join Date: Jun 2001
Rookie Year: 1998
Location: Houston, Texas
Posts: 3,731
Kevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond reputeKevin Sevcik has a reputation beyond repute
Send a message via AIM to Kevin Sevcik Send a message via Yahoo to Kevin Sevcik
Re: FEA Modeling

Quote:
Originally Posted by ChrisH
The first part isn't necessarily true, but the second sure is. How long an FEA takes all depends on the level of detail in your model. For FIRST robots a coarse model of a 100 or so nodes or less should be adequate. Such a model should run in a couple of minutes on all but the most ancient PCs.
Ummm... heh. I'm all with you that a robot COULD be modelled with 100 nodes if you're using nifty beam and plate elements and such, which I'm sure is what NASTRAN loves to use. In fact, I agree that ideally the robot should be modelled this way since it saves so much on computing power.

Neal was asking for an FEA package that would integrate with Inventor, though. This kind of FEA package rather stubbornly uses brick elements to represent the solid model and just throws a ton of computing power at the problem, unless FEA packages have advanced greatly in the last few months.

But yes, the practical upshot of all this is that a good textbook or junior MechE is a rather lot cheaper and safer than FEA. Heck, most college students will probably work for food. I know I do.
__________________
The difficult we do today; the impossible we do tomorrow. Miracles by appointment only.

Lone Star Regional Troubleshooter
  #5   Spotlight this post!  
Unread 27-01-2004, 09:05
Neal Turett Neal Turett is offline
Registered User
#0201
 
Join Date: Nov 2002
Location: Rochester Hills
Posts: 34
Neal Turett is on a distinguished road
Re: FEA Modeling

The problem i was working on (a 2-speed tranny using overrunning clutches to shift) became a moot point once we actually got our hands on the clutches....

Thanks for all the advice anyway.
  #6   Spotlight this post!  
Unread 16-02-2004, 20:33
patTeam241 patTeam241 is offline
College Student
AKA: Pat Cleary
no team
 
Join Date: Jan 2004
Rookie Year: 2001
Location: Boston, MA
Posts: 54
patTeam241 will become famous soon enoughpatTeam241 will become famous soon enough
Send a message via AIM to patTeam241
Re: FEA Modeling

Just as a note for teams that want FEA programs, JL Analyzer 8.0 has a free version that isn't bad, I'm having trouble getting Inventor files into it, but its free. Go to www.autofea.com/Download-FEA-software.htm.
Closed Thread


Thread Tools
Display Modes Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
What modeling program did you use? archiver 2000 5 23-06-2002 23:24
Computer modeling of drivetrain and devices archiver 2000 1 23-06-2002 23:02


All times are GMT -5. The time now is 01:20.

The Chief Delphi Forums are sponsored by Innovation First International, Inc.


Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
Copyright © Chief Delphi