FEA use nowadays 2

[321GO]

Hello to you all.

I've noticed a lot of engineers lack a decent understanding of fea (hold your horses i'm not claiming to be an expert) and base their engineering decisions on these results even without doing some form of crude verification.
For instance when asked "did the results converge?" they basically look at you clueless or say something like "this package is very advanced an does that for you..."

Even though i'm a Rookie i know my limitations using it and how easily it is to overlook stuff and mess up. Grabage -> Garbage out ->
So I'm weary of 'results' and basically don't trust them unless i get some form of confirmation (handcalc's whatever / don't care)

What strikes me even more is that management does not raise these questions. Come to think of it that's probably not so strange since everything must be done within the blink of en eye on minimal budget and quantity -> quality (till thing go sourer...) Even more striking, the knowledgeable engineer will run the risk of becoming outcasted because of his diligence.

Messed up IMHO.

Is this common? I would love to hear you guy's thoughts on this.

 

[corus]

Alas, it is common for management to expect anyone to use FEA, supposing they've been on a day's course on how to use it. If it makes a pretty picture then they're happy, as 'it's what the computer says'. In part it's due to software manufacturers who provide these 'click and go' packages, and thus eliminating any thought into the process and results. The result is that Design Engineers are given the tools to produce a FE model quickly, but not the brains to understand it.

 

[MikeHalloran]

I understand FEA well enough to not trust it without some kind of correlation to reality.
But I've run into two problems that may surprise you.


Maybe ten years ago, I got a nice NASTRAN package and FEMAP, and messed with them for a while on some simple problems. I got to talking with another senior engineer in the same outfit, and realized that we had both independently come to the same conclusion: If we showed any interest in working with FEA, we (each) would become 'the FEA guy', and we'd never be allowed to do anything else. ... and we found the process incredibly boring. ... so we each put our demo package on the shelf and busied ourselves with other things.


Not that long ago, I managed to get some usable FEA results out of the limited linear static package that comes with Solidworks. Simplified hand calcs made me suspicious of one portion of a competitor's skid that I had been ordered to duplicate. So I ran the SW package against it, and got some nice color plots showing the metal stressed past any achievable yield point, just from gravity loads. I forwarded the plots to my boss.

... who got yelled at because I was wasting my time with fancy expensive tools when I was just supposed to be copying someone else's work, which was assumed sufficient, because we were actually trying to steal their contract, and our first class salesman had managed to steal one of their blueprints.

... never mind that I had done the fancy plots on my own time for fun.

... never mind that the skids the competitor actually produced didn't look like the stolen blueprint, and wouldn't have the problem I had found.

In many ways I'm glad I don't work there anymore.









Mike Halloran
Pembroke Pines, FL, USA

 

[KevinDeSmet]

On the other hand it is elitist to say only people who have "x" years of experience, or spend a minimum of "x" percentage of their time performing simulation are the only ones who have the right to be doing so.

I agree SolidWorks with their Simulation tutorials do not do a great job of debunking this sticky topic. Just look at the first tutorial, tutor1, where there is plentiful use of the Fixed boundary condition to simply give the wrong idea to new users of Simulation. I get that it's the first tutorial and they should keep it simple, but they didn't just keep it simple with that one, they misguide new users.

Thankfully I think the VARs, at least the ones I've come in contact with, do their best to make things clear to users that it's not just going to be a case of *buy, go to training, install, run, profit*

Certified SolidWorks Professional

 

[TheTick]

The word "elitist" is most often used to describe people who understand the importance of what you don't know.

 

[fade2blue]

You definitely don't want to be the "FEA guy" in some companies. Running a "quick" FEA on an item to make sure it will handle 3x the load it was initially designed for because sales can say whatever they want to close the deal.

 

[321GO]

[tt]Food for thought..Thanks for the feedback![/tt]