Board logo

Chassis Design - CAD and Analysis
onenastyviper - 7/7/11 at 12:29 PM

If I use CAD and FEA to perform analysis of a chassis, would it be a good idea to only use it as a comparitive tool?
I am thinking of starting with a very rough chassis design then start adding and refining the members using the FEA results as a "better-worse" analysis.
A sort of virtual balsa model?

What do you guys think?

Doctor Derek Doctors - 7/7/11 at 12:53 PM

Depends how good you are with FEA.

"put crap in and you get crap out" was the phrase we were taught.

Probablky best to research existing designs and tools (hand calcs, etc etc) first and then do FEA alongside as a learning process, you should never 'rely' on FEA reults, especially without a feedback loop.

Have fun with the project though

coyoteboy - 7/7/11 at 02:00 PM

Check the FSAE solidworks frame design tutorials online, they should help you get started. I'd not be afraid of working the way you describe, however bearing in mind they're not perfect and you're not perfect and common sense must prevail. I'd not be trying to shave things to a bare minimum weight based solely on the results of FEA without significant testing, but as a comparative tool they're fairly powerful. GIGO does definitely apply though, I've had a few people telling me that torque varies along the length of a plain shaft in pure torque because the FEA proves it...when you've mis-applied your loads.

v8kid - 7/7/11 at 04:14 PM

Alan B - 7/7/11 at 05:26 PM

quote:
Originally posted by Doctor Derek Doctors
Depends how good you are with FEA.........


Got to agree with this bit.....and with no disrepect of the OP intended....it's a bit like the old "can I use composites/aluminium for my chassis"....if you have to ask, then most likely the answer is "no, it's probably not a good idea....if you could do it, you'd already know"

Just IMO

Alan

Doug68 - 8/7/11 at 02:51 AM

Look at the FRAMEWORK 2D+3D program on this page, its freeware.
The software is a structural analysis package rather than a FEA type program that requires you to mesh a model therfore IMHO it much more suited to doing a tube chassis.

As said above these things are just sausage machines and you need to be able to spot when its telling you things that cannot possibly be true. The more complex the software then the bigger the chance of you stuffing up the inputs and getting bad data as a result, this is another reason why i like the simpler approach of this particular software.

Here's a shot of my model in the software...

britishtrident - 8/7/11 at 07:38 AM

The biggest difficulties with FEA on a chassis are figuring out the boundary conditions ie; loadings and restraints ----- Get them wrong and any numbers that come out are valueless and it takes insight and experience to get them right.

onenastyviper - 8/7/11 at 12:05 PM

quote:
Originally posted by britishtrident
The biggest difficulties with FEA on a chassis are figuring out the boundary conditions ie; loadings and restraints ----- Get them wrong and any numbers that come out are valueless and it takes insight and experience to get them right.


I agree. I want to learn how to use the packages first. I know this isn't going to be a week to production and I used to be good at structures - argh: revision again

D Beddows - 8/7/11 at 12:58 PM

If this is for a '7' chassis you do realise you're completely wasting your time/re-inventing the wheel as it's been done lots and lots of times before and no one ever comes up with anything significantly better?

If not carry on!

onenastyviper - 11/7/11 at 09:09 AM

Definately not a se7en

quote:
Originally posted by D Beddows
If this is for a '7' chassis you do realise you're completely wasting your time/re-inventing the wheel as it's been done lots and lots of times before and no one ever comes up with anything significantly better?

If not carry on!