Should people without CFD/FEA backgrounds do CFD/FEA? 6

There is such a thing as self-taught engineering... while your particular experience may be GIGO, I see nothing wrong with non-background people doing that work... so long as they question what they don't understand and get proper input/feedback. Without that last part, it's a very steep hill to get anywhere useful.

Dan - Owner

MacGyverS2000,

I have had some scary conversations about using FEA for analysis. The FEA is a magic box that is wise and that can answer our questions. All sorts of CAD[ ]operators out there do not understand that CAD and FEA are tools, and that the job (in this case) is structural analysis, and that there may be catastrophic failure modes.

CFD is not as scary to me. That does mean that the job will be done properly.

--
JHG

I have worked with many engineers who had formal training in their discipline and are completely worthless or can't be trusted for results.

I have worked with many engineers with little to no formal training on their specific niche discipline/software/industry who are very capable and dependable.

In general I would say it's prudent to focus on trained individuals when evaluating candidates for a position, but years on the job and a track record of good results are perhaps more important than a certification.

At some point, everyone has been in the position of having no experience in something. That does not mean that they should not engage in that particular activity, or they are "garbage in, garbage out" engineers. It simply means that they do not have the expertise to perform that activity without aide and supervision. Nobody is born an expert. However, if someone is blindly using software, accepting the results without knowing if they make sense, and allowing widgets to be built based on the software output, that is not right.

"In my opinion, to do FEA you must have taken a class in FEA"

Hate to break it to you but when I started FEA it was teach yourself or don't do it. Never done a course on FEA in my life (and it would bore me to tears).

If we are flinging stupid assertions around, nobody should be doing FEA who has not worked in a stress lab, or run an experimental modal analysis.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

What separates analysts from users is deep knowledge--e.g. fluids, materials, failure modes. There's a lot to know in breadth and depth.

Recently I was working a support role for a large team of analysts. Believe me, most engineers are not cut out for that job. There's so much more than turning meshes into rainbow colors.

I can play piano, but I am not a pianist. I can run analyses, but I am far from being an analyst.

What Greg said! Same here.

I don't use FEA very often... it's a specialised tool. I've not taken any classes (classes may help, but not essential) in it, but have read several texts and have written my own FEA programs. Writing the programs really brings the chicken home to roost. With any tool, you have to know the limitations and with FEA, in particular, the effects of boundary conditions.

I was involved in a court case about 50 years back and one of the protagonists was one of my profs (and one of the referees who signed for me to join the professional association). The problem dealt with shear friction, prior to it being codified in Canada... he'd run FEA on the dowels and ended up with 13 ksi compression concrete stress at the dowel. In court he couldn't give a proper response to 'what happens to the stress distribution if the concrete yields'. An OK tool in its place.

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

Education teaches the basic theory and how to use the tools. Experience teaches when the theory and tools fail in the real world, and how to complete the work efficiently. Like anything else in life, many have become proficient entirely on their own. Many more have become fraudulent hacks bc fewer eyes make more mistakes, hence the many jokes about “youtube-certified” and the requirement to have both when selling to the public.

As alluded to above, running uncorrelated analysis for years doesn’t make someone a CAE expert regardless of education. Without correlating tests their analysis is still an unproven guess - who knows how accurate, prob not very. Personally I have spent enough time building and correlating corporate master models that I don’t waste much time on analyses that won’t be tested.

GregLocock said:

...nobody should be doing FEA who has not worked in a stress lab, or run an experimental modal analysis.

Click to expand...

There's a lot of truth in that!

Comparison of computed predictions to measurements provides a lot of perspective.

FEA brings out the best of Dunning-Krueger. If you are confident, you are almost certainly wrong. If you are self-aware enough to recognize the limits of your knowledge and abilities, and the limits of your tools, you might still be wrong, but at least you'll be aware.

I think a general rule of thumb is that people need more experience with the subject than simply using the program. Whether its FEA, pipe stress analysis, etc, you need to understand what you're doing not just be fluent in filling in the blanks the program requests.

I think a general rule of thumb is that people need more experience with the subject than simply using the program.

Click to expand...

Ha! Had an applicant whose resume claimed experience on a certain SW package, so I asked what sorts of systems they had modeled with it; "Weeelll, I actually didn't use, but I watched someone use it."

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert! faq731-376 forum1529 Entire Forum list

Import a CAD model of unknown fidelity.

Click - generate mesh.

Click - assign some random boundary conditions.

Click - generate colorful plots.

Easy. Any intern can do it.

What could possibly go wrong?

I've done CFD, CAD...and traditional hand calc's, and variations in between. They are tools, and like any other tool can be used...inappropriately.

Did a lot of work alongside some PhD's doing cutting-edge CFD with reactive, mixed super/subsonic flows. They would do a lot of whiteboard sketches in our shared cubical space, and I would of course kibbitz on the conversations. Funny, but we could usually (with the help of several colors of whiteboard markers) do a pretty good job of guessing the resultant pretty colored pictures from the CFD runs, using the Mk 1 CFD solver in our heads.