What's up elsewhere in the internet
A couple of nice slash and/or interesting new from the last week or two. First of all, my FEA infographic was used (with my permission) by an engineering firm/group called Structural Design and Analysis, or SDA.
Indeed, they published a quite nice piece about what finite element analysis is, and how it’s performed. Their focus was more on mechanical engineering rather than magnetics, though.
But, it’s actually nicer that way. It’s always nice to learn about new things, plus it’s less competition for me.
Mechanical failures
Talking about mechanics – lots happening on that front too.
You may remember FEAforAll from some of my previous posts. In case you don’t – check it out. It’s a very nice resource, maintained by my buddy Cyprien.
Well, they have a really educating new post about the different ways mechanical structures can fail. So, if you can’t recite the differences between brittle and ductile materials in the middle of the night – I recommend you check it out!
The same goes for creeping (no, not that kind!) and fatigue.
And yielding.
And insufficient thickness.
Aand buckling.
Like I mentioned, the post is educational.
Plus it ties nicely with my ongoing post series about uncertainty quantification. Few things are certain with failure. Yielding might happen at a somewhat predictable less level alright, but even then some variations can be expected. Nobody knows what happens with fatigue and yielding, so those are very much uncertain, too. And as I wrote earlier, buckling is surprisingly stochastic despite its deceivingly-simple nature.
Buckling and stochastics
Talking about buckling, Łukasz from EnterFEA wrote a nice piece about nonlinear buckling analysis with imperfections. Note that the nonlinearity here extends beyond material parameters, to the geometry itself. Since this is such an unfamiliar concept for many electrical engineers, I recommend you check out my previous explanation of the topic.
And while you’re clicking links, check out my brief post on how linear buckling analysis relates to solving an eigenvalue problem. The corresponding eigenmodes can then be used in nonlinear buckling analysis.
You see, linear analysis is somewhat limited, as Łukasz has many times explained. Nonlinear analysis could better model the real phenomenon.
On the other hand, the nonlinear analysis has to be kick-started somehow. Buckling happens because a system is not geometrically perfect after all. So, starting nonlinear analysis from a perfect geometry doesn’t really make any sense, does it?
One option is to use one of the buckling modes as an imperfection, or initial state, in nonlinear analysis.
It’s an intuitive solution on many levels, but still not a perfect one. To see some of its limitations, see EnterFEAs post by clicking the picture above.
Actual stochastics
This approach doesn’t consider any stochastics, either. All you get is one value for the buckling load. No estimates on how it might be distributed, or anything.
For that purpose, uncertainty quantification can be applied. Somehow. Probably.
Indeed, Łukasz and I are working together on that topic right now. I’m not sure how it turns out, yet. I have a strong compute first – think second – do a literature review third approach. Great for learning new stuff and occasionally making some nice discoveries. Not so good for saving time.
In any case, below is a teaser sample for some random imperfections for starting the buckling analysis.
Until nex time!
-Antti
Check out EMDtool - Electric Motor Design toolbox for Matlab.
Need help with electric motor design or design software? Let's get in touch - satisfaction guaranteed!
