Worth it (at this point) to take an FEA class? 2

[3doorsdwn]

Hi. I’ m a structural engineer with a M.S. in structural engineering (and 12 years experience, including a great deal of FEA modeling experience), but I never completed a class that was solely about FEA. I certainly had classes where it was a topic (like my second structural analysis class, where the direct stiffness method was covered and my Advanced mechanics of materials class), but never one where it was covered by itself. At this point, is it really worth it (in anyone’s opinion) to try to take an on-line course in it somewhere? My concern is that what made me drop it in Grad school, is probably what would make me drop it again: impracticality. The class I was in (for a couple of months) in Grad school, we got so caught up in the mathematics of it, and it was disconnected from practical modeling concerns that it was useless. I’ve tried to study this subject, but I was considering a more structured approach (but I didn’t know how effective it could/would be). Thoughts?

 

[corus]

I've not known of any courses that teach the practicality of the subject. The best way I've found is to go on the courses that are often ran by the software companies themselves. These can teach you the fundamentals of element selection, the effects of mesh density, and then the more advanced aspects such as contact, buckling, heat flow etc. It depends on the software you have though and whether they offer courses in the use of their software. The best way I've learnt though is to have a mentor who can advise you through the experience they've gained, and of course to gain experience yourself, as every problem is unique.

corus

 

[Stringmaker]

I most definitely would take a class in it! I didn't do the MS route right after college. I waited a couple years and am about half way through now and I find I get so much more out of classes by being able to apply them to my job than I ever did in academia. FE would have been tough to take as an undergrad...but now that you know the power of it and it's uses why wouldn't it be a positive experience? I would definitely recommend taking a sit in class if possible! I speak for myself, but taking a class online wouldn't do anything to improve my knowledge of something I already had a background in.

 

[UcfSE]

It depends. Do you want to understand why you are doing what you're doing?

 

[3doorsdwn]

I appreciate the feedback everyone.

I would consider myself pretty proficient at modeling (in fact many people here approach me to fix their models (i.e. instabilities, etc.)). But I was just kicking around the idea. I’ve always felt that more fundamentals couldn’t hurt.

 

[Qshake]

I'm a structural engineer with a MS who took FEA and really enjoyed it. It is not for the weak of heart (or those weak in math skills). Our course was pretty much math intensive and we had laboratories on top of that to put what we went over in class to work on the computer. There is where we learned the practical side of modeling. So I agree with stringmaker and UcfSE.

I would note that you can't possibly get all the necessary work from one course though. So if your school doesn't have an advanced course with dynamics, consider seeking out a seminar offered by the software vendors.

Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.

 

[abusementpark]

I just took an undergrad class on FEA and I am not sure I would recommend it. I got a few good things from the class, but the vast majority of it was worthless. I am really sure I didn't gain any practical knowledge from learning about how stiffness matrices are assembled, how shape functions are derived, how displacement boundary conditions are applied through the penalty method, etc. The only practical things I learned was how to access error in modeling of 2D and 3D, which I may or may not use in the future. I think a software sponsored course would be way better.

 

[3doorsdwn]

Since we are on the topic of FEA instruction, does anyone here have A First Course in the Finite Element Method
by Daryl L. Logan ? Is it a good book (i.e. practical)?
Does it have practical applications in the Civil/structural area?

 

[Stringmaker]

I happen to think the best books on the FEM are those by O.C. Zienkiewicz. Cook also puts out a good text. Those are the only two I really have experience with. Check the reviews on Amazon or B&N and see if they align with what you are looking for.

 

[Qshake]

Agree with stringmaker on OCZ and Cook but wish to add K.J. Bathe to the bookshelf too.

Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.

 

[GBor]

I have the book by Logan. I used it in Undergrad many, many years ago. We concentrated on the math theory...something I appreciate now that I've been using different software packages and want to understand precisely what they use for element formulations and how those may affect the results that I get in my "pretty picture". As for "practical"...it has some, but they are very basic...structral frames, truss systems, some slightly more complicated.

 

[ThomasH]

Back to the OP.

If you think that after 12 years you know everything worth knowing om FEA you probably don't need the class (or classes).

As for practical courses, they are often good. But I would say that knowing the theoretical background (like shape functions) can be more rewarding in the long perspective. Do you know what limitations the shape functions give on the analysis results?

If you find a good course, go for it.

Good Luck

Thomas

 

[abusementpark]

I have to disagree with others and say that Cook's book is not good at all. It has no examples or step-by-step processes listed.

 

[GregLocock]

Why would someone with 12 years FEA experience need step by step instructions?



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[3doorsdwn]

GBor, thanks for your feedback on Logan's book. One more question on it though: does the text discuss things in a practical manner? By that I mean can/does he discuss a topic (such as shear locking for example) without spraying a bunch of equations all over the page without a comprehinsible explanation? Books (like Bathe's) tend to do that. Someone on this thread mentioned Adams (et al) book: 'Building Better Products With Finite Element Analysis'; that book is too far in the opposite direction.

 

[pja]

How can you explain shear locking without "spraying equations all over the page" ? You need to at least write the expression for strain energy or at the very least the expression for shear strain for the concept to make any sense.

Some of you sound as if you are afraid of mathematics. You remind me of some of my students expecting to go through a whole course learning only "practical" things whatever that means. I throw down some equations and they start to get a glazed look in their eyes.

 

[3doorsdwn]

"How can you explain shear locking without "spraying equations all over the page" ? You need to at least write the expression for strain energy or at the very least the expression for shear strain for the concept to make any sense."
--------------------
I don't deny that doing some "spraying" should/would be necessary. But most texts I have seen treat the subject have done a very poor job of it. (It is possible to treat this without a too esoteric explanation.) As we are all engineers here, I think we all recognize that some books in our field fall short (in some subject areas egregiously so); but in no area does this seem to be more so than in finite elements.

 

[GBor]

Logan uses a number of equations and does spend some time on theory. The book is geared more toward teaching you to write your own software rather than understand the limitations of a specific software package. It also has several basic programs written in fortran and published in the appendices.

Logan's equations are generally explained in good detail without being so indepth that you feel a need to get out your calculus book. I don't think he spends enough time addressing topics like convergence criteria, but he does give plenty of examples (even if they are more truss-related rather than plates, bricks, or other element types), and he provides flow charts for developing analytical code.

 

[jagad5]

I suggest you read the book by Vince Adams and Abe Askenazi (spelling?). The title is something like Designing better products with FEA. You should be able to find it through amazon.com or bookfinder.com very easily.

I also recommend that you consider at least being familiar with the theory. In my opinion, it's impossible to responsibly use FEA without knowing at least some of the theoretical foundations of the method. If the theory built into the software is incapable of modeling the problem you are attempting to solve, refining until you get a converged "answer" will have proved nothing. I've seen several analysts misuse and abuse fea tools because they didn't understand the theory.

Doug

 

[aliante]

Hi to all of yoy over there,

I completely agree with people who say : knowledge about theory of finite elements is essential for understanding how to model a structure a part of it , how to implement a material and so on. I also assume that you understand the behaviour of a structure under some kind of loads and so that you will be able to accept or reject the result of an analysis and then understand that you have to re-model the structure.
However, my advice is to study these two books that you can buy through amazon:
1) Fundamentals of finite element analysis by David V. Hutton
it's more practical then Zienkiewicz.
2) Finite Element design of concrete structures by G. A. Rombach .
Best wishes
Aliante