FEAforAll.com: another blog about finite element analysis
Good news, folks! Based on a thorough web search and literature review, my blog is the upmost and supreme authority on electrical machines, at least from the numerical analysis point of view! Wheeee!
Not that there’s much competition.
At least I haven’t been able to find any. Ansys and Comsol do understandably have their own content marketing blogs (quite good at that, at least if you’re using their products), but that’s about it.
However, if you are a mechanical engineer, you have a really nice resource called FEAforAll.com.
It’s is maintained by one Frenchman called Cyprien Rusu. He seems like a really cool fellow, at least based on the little online correspondence I’ve had with him. And that’s what counts, right?
He even has a really cool reading list on his site, which is definitely more than I can boast right now. Well, I could definitely make one, but for some reason people seem to kind of look down on Warhammer 30k fluff and fleet-on-fleet-action-packed space opera scifi. (By the way, I love hyphens, in case you haven’t noticed.)
Anyway, I highly recommend you check out the site. The link opens in new tab so you can easily come back to mine when you’re done.
Back already? Or finally? Matters not.
According to their mission statement, FEAforAll aims to make finite element analysis more accessible to engineers and designers alike, worldwide. Which is a cool and worthy goal indeed. More about that in a sec.
As you can see, FEAforAll approaches finite element analysis more from the structural mechanics’ point of view. Which is more than fine. Heck, if it were about machines I’d be ddos’ing the everloving crap out if it already, haha.
I’m kidding, by the way. I’m working in the academia, so I definitely can handle a little competition.
But, not having to is even better.
Where Things Differ
Besides, reading a little something about structural topics really shows what a hugely diverse field finite element analysis is after all. I’ve been elbow-deep in the guts of machine-FEM for years now, and I barely understood half of the topics on the site (before reading through the corresponding texts, of course. They are quite clear.).
Anyway, based on this brief scouting mission, these are probably the most obvious differences between structural and electromechanical FEM:
- Structural FEA utilizes far more diverse types of elements. Tetrahedral, brick, shell, you name it. For us machine people, it’s pretty much nodal (Lagrangian) only, or vector (Nedelec) in 3D.
- Talking about 3D, structural mechanics is very often that. Makes sense, since that’s what most really structures are like. Electrical machines are too, yeah, but their magnetomechanical behaviour is usually sufficiently governed by 2D-phenomena.
- They are pretty much always nonlinear, though. Solid mechanics quite often isn’t, unless you go into plastic deformation range. (I think…)
- Singularities. I really pity mechanical engineers for having to deal with them.
- Solid mechanics quite often deals with solid mechanics. Electromechanics pretty much always deals with motion (changes in mesh topology) and coupling to circuit equations.
- Time-domain analysis is quite different indeed. Mechanics are governed by second-order ordinary differential equations, whereas for machines it’s all about first-order DAEs (including the rotor motion might give us an extra second-order term, but that’s handled easily enough).
And that’s just a comparison between two applications of FEM. Two quite similar applications, mind you. Fluid dynamics, for example, is a whole another level of nightmare fuel.
Two Sides of the Same Huge Coin
So indeed, our topics – Cyprien’s and mine – are somewhat similar, yet still different. Moreover, we approach them from slightly different angles. Remember FEAforAll‘s goal of making finite elements more accessible to people?
I could say it’s mine too. (Well, that and blowing my own horn for the entire world to hear. But I digress.) Our ways of advancing this worthy ambition differ a bit, though. FEAforAll is full of practical tutorials, about all things FE. My blog is full of semi-coherent ramblings about mathematics, and the very (small part of the) cutting edge of elements-related research.
In my opinion, both are definitely needed. Many people – FEA users are that – are brutally lacking in an understanding of their software tools.
Indeed, it’s quite incredible what black-box magic-tool mentality FEA is usually regarded with. What I mean is that far too many people only concern themselves with the software itself, without much understanding about what goes on under the hood. And I’m not completely alone with this opinion, either (see here or here, to begin with), mind you. I hope some kind of renaissance movement is slowly gathering momentum.
When Simply Having a Tool is not Enough
Now, I’m not saying that not understanding a software is necessarily a bad thing. If you’re a graphic designer, you probably don’t need a very deep understanding of the inner workings of image processing algorithms. But that’s for a simple reason only – you can easily verify the quality of the outcome with your own two eyes.
With numerical analysis, that might not be the case. You can usually tell when the excrement interferes with the air-cooling in a bad way – your supposed-to-be-nice curves are all over the place. Fixing the problem might be a tad more complicated, however. Moreover, sometimes problems are not right away visible, unless you know how to look deeper. Finally, occasionally the generally-accepted way of doing stuff is simply a gross oversimplification, but doing anything about it requires coming up with a solution yourself.
In any of these situations, just knowing how to use a software won’t probably be enough. And even if it were, having a little more understanding surely wouldn’t hurt you.
Finite element analysis is an awesome, albeit under-blogged field. ’nuff said.
Need help with electric motor design or design software? Let's get in touch - satisfaction guaranteed!