Since I announced Q&A column on my blog I got several great questions so far. I will try to answer as many of them as possible at the end of each month. If you have sent me a question and you didn’t find an answer here no worries! I store all of them and I will answer it later on!
A career in nonlinear FEA – what to do?
I like questions from LinkedIn simply because they look so nice and professional when inserted into the post 🙂
I think however that I would reframe the question a bit. Nonlinear analysis is not a career field… just as using Excel is not a career field. Sure this is a skill that gives you tremendous opportunities, but never “alone”. This is a tool you can do a lot with… as long as you know what you want to do with it.
In other words, don’t invest into the nonlinear FEA “alone”. This would mean you are like a “calculator operator”. Someone want to have something solved, so they came to you. They tell you more or less what is required and ask you to do it. A typical corporate job. There is nothing wrong with this of course, but I feel that the “true power” of nonlinear FEA lies elsewhere.
Before I started blogging I have spent 8 years (not counting studies) learning about structural stability in various problems. During that time I also learned FEA – but I started that few years along the way. I realized how big potential this method has. Since I already knew what I wanted to do I could appreciate the possibilities FEA gave me. I was aware of the problems in stability, and nonlinear FEA was an answer to those problems.
In other words, try to be an expert in some field other that nonlinear FEA and use FEA there. No worries it is not a choice for life! Simply don’t learn only about convergence, mesh refinements etc. Learn also about any industry you think nonlinear FEA will be useful. The more you know about this one industry, the more you will appreciate the possibilities nonlinear FEA has. But also you will be able to define problems and propose solutions on your own. This will give you a tremendous freedom of choice! After a while, you will know a lot in one industry. Thanks to this you will be able to “expand” into fields that have similar problems etc. From there it gets easier 🙂
My field was stability, which leads to shell stability and things related to different solver approaches, convergence issues etc. Then I expanded into thin-plate structures, simply because their issues are similar, and I learn this now. Later… who knows maybe dynamic stability or laminates? I will see where the life will take me 🙂
If I would focus only on nonlinear FEA I would most likely know more about it now. But I woul not be able to design anything myself. I would not know what to search for, and where the problems may be.
Nonlinear FEA is not a goal. It is a great tool that will help you to reach your goals. But it’s up to you to know what you want to use it for!
Where to start is a difficult question. Since you already use nonlinear analysis you have solved some problems already. If you are lucky they were from different fields. Just pick whatever you feel is interesting and go there. If you are unsure what to do, try to reach out to people on LinkedIn that do something similar. Ask them if nonlinear FEA makes sense in that field, what problems are there to solve etc. I’m sure you will find help from others, just as I’m helping you right now.
Stability of complex structure
It seems that with the last post I invoked a series of questions about stability. This one was special as I had a chance to talk with Grzegorz for half an hour on our latest Enterfea Hangout. If you missed this, no worries. I will do more of those in the future.
Grzegorz was curious about the stability of a structure he designed some years ago as a diploma thesis. I must admit that idea of the structure is pretty interesting – definitely worth taking a look:
There are 3 main components here. “Horizontal” girders aren’t of course “flat” but they are on the surface of the roof. They provide stiffness more or less in the horizontal direction. The second important elements are the vertical trusses. Those are rigid in the vertical plane. Finally longitudinal beams “tie” this lot together.
Grzegorz was curious how I would approach design of such a thing, so let’s give it a shot 🙂
First of all few observations.
- I would most likely assume that purlins need to carry the vertical load between vertical trusses. Sure, there is a “horizontal” girder between each of those. But I doubt it is rigid enough to take a significant load. At very max it would be an elastic support. Assuming that I would have a model its effectiveness is easy to check. I even have a lesson on how to do it in my starting with FEA online course.
- After that, I would like to check if the vertical arcs are strong enough to carry that load. Since on the first look they look slender I would also make a nonlinear analysis just to check if snap through may be an issue.
- Knowing that vertical direction is more or less ok I would check if the horizontal girders are ok. There must be the horizontal load on this “mouse” for sure. It would be great if it would not fall to the side 🙂
- Then I would tie this all in one big analysis checking if everything is stable together. Again a snap through check seems like a good idea.
I would approach this problem with following tools:
- The first choice is to make a decision. Do I want to analyze this structure in one “go” or do I want to analyze smaller parts first?
- If I would go with the whole structure LBA (Linear buckling) would be a nightmare – a LOT of forms will be connected with buckling of single elements, so before I would reach any global forms it would take significant time. If I would be designing this “for real” I would do it anyway of course. Still, it would be something that would make me think about making “isolated” models of each piece of the structure first. This way picking correct cross-sections would take much less time.
- With smaller isolated elements LBA would not be a great issue, and it could be easily done. Based on that I could calculate the critical force in elements and pick correct cross-sections (based i.e. on design according to EC).
- At the “isolated” models level I think I wouldn’t use nonlinear analysis yet. There are a lot of things such isolated models don’t take into account. As such, making an advance analysis there seems like a bad idea. Instead, I would create a nice 3D model of the entire thing. Since I did the isolated calculations I can input more-or-less proper cross sections now.
- At this stage making, LBA of the whole thing makes far more sense. However, it would take significant time in post-processing anyway. But knowing how this thing can deform (in a global manner) would greatly help me later.
- I could take imperfections from scaled shapes of global LBA failure and use them in nonlinear analysis. This way I could be sure, that I won’t miss anything important. Since elements can be designed according to a code (i.e. EC) in this step I’m not concerned about buckling of single elements. Instead, I want to check stability failure of the system as a whole (which is not exactly described in EC).
- And basically, that would be it I think 🙂
Also, let me know what you think about Q&A posts – is this a good idea? Do you find this useful, and should I continue them?