Recently, I have accepted the position of a Visiting Professor at the Symbiosis Institute of Technology, Lavale, Pune, and have just begun teaching an introductory course on Finite Element Analysis to the students of MTech (Mechanical — CAD/CAM/CAE specialization) there, on a part-time basis.
The classes are held in the CAD/CAM/CAE lab., wherein they have everything one would have liked to see: an audio projector hooked to a PC + a networked PC for every student right in the class-room + no acoustic reverbaration problem in the class-room at all + timings convenient to me (OK to conduct classes on weekends and all). It’s a small class (< 20 students), and though it might be too early to form a judgment, the students do seem to be mature and responsive. (A couple of them have even worked in industry for a year or two.) I think I am going to enjoy teaching this course, even though my interest in FEM has been rapidly sliding down in the recent past—I am getting into the CFD field now.
FEM would, of course, continue to remain a matter of interest. Yet, the primary interest for the rest of my career, I have now firmed up my mind, now on is going to be: CFD. CFD using FVM and FDM and FEM, sure, but what I especially have in mind are some of the more recent particles-based approaches like LBM (the lattice-Boltzmann method) and SPH (the smoothed particle hydrodynamics). Especially, LBM.
Both LBM and SPH are the particles-based approaches that are essentially (i) local, (ii) transient, and (iii) spatially cascading and temporally propagating in nature—so close to the very core of how, I realized only recently, I have always had approached understanding nature. So much so that I have this “it’s me—it’s my way of doing things” feeling about it.
Indeed, my first memories of wanting a transient, local and propagating description goes as far back as at least the X/XI standard, right at the time when I was being taught Newton’s laws and calculus for the first time, and, unlike many other smart guys, had found the theoretical apparatus wanting when it came to exactly and directly capturing the physical reality the way I wished to. I didn’t have the words back then, though I would describe my differences animatedly. Now, I know the words, too. Things like: micro-dynamics, local, transient, cascading, propagating, feedbacks, emergent, etc. Back then, I didn’t have the words but still found the existing formalism wanting when it came to applying it to even simplest cases like the carrom-striker’s rebound, or the hitting of a ball by a bat. More on this all, again, later! (Do remind me!!)
Enough to say here, and for the time being, that the approach I developed during my PhD—an approach that is local and transient and propagating in nature—was a development that was very, very natural to me—my way of thinking. That way, I have always had fascination for fluids. Yet, comparatively, that’s secondary. The primary thing is the nature of this local and transient approach itself—something that can only become feasible when you are doing computational modeling, not otherwise.
And, so, it simply is marvelous for me to now find that there are these methods that are not only fairly well developed, but also capable of addressing questions of practically important engineering situations. Here, do realize that MSC Software has just last year entered into a partnership/marketing arrangement for the Next Limit’s LBM-based product: XFlow. A fortuitous circumstance, all in all. And, not only that, but also something more, further: given my approach to understanding physical reality in general, and my novel approach in QM in particular, LBM should also allow me to more easily span my interests from engineering to QM (or vice versa). That is a major attraction for me.
Anyway, that’s rather for the long-term future. Coming back to this course on FEM I am right now teaching.
I think I am going to make a departure from using C++. Instead, I am going to try using SciLab right from the beginning of this course. Since the course is the very first one on FEA for most students, I have little choice when it comes to the contents. For the most part, it’s going to be the usual linear and static FEA with simple isoparametric elements. However, I might try to introduce an exposure to diverse topics via assignment of student-delivered seminars. Such topics might include (all at a simple level): nonlinear FEA (as in fluids), plate and shell elements, FEA for fracture mechanics, FEA for simple coupled problems, etc. … It might also be a good idea to introduce a free (free of cost) FE software in this course. Though I have not yet finalized my mind, I am thinking of choosing Elmer.
Also, when it comes to evaluation, I might try out different approaches like open-book examinations, etc. Otherwise, for a topic like FEM, it can be excruciatingly boring to do all those calculations by hand. Not only that, the traditional examination format restricts you to only very simple, almost artificial sort of problems. People’s grades get determined to 50% extent or more by their skills in doing only simple truss and frame analysis using FEM, simply because no better questions can at all be posed. That, in turn, is because any better question would require more than 3 hours of examination to solve by hand. If you are going to conduct examination in 3 hours, then, probing for the things that really matter, goes out the window, and it becomes a testing for learning by the rote. Instead, I would like to see if I can allow them to use SciLab and refer to openly accessible codes, right during their formal written examinations. If things turn out OK, I might even consider assigning them more ambitious projects via open-book examination mode, for the final examination. Let’s see how things progress. … After all, it’s not just the teacher but also the students—they, too, have to grow comfortable with such ideas, including the idea that their grades may get determined, at least in part, via some approach with which they have had no prior experience… And, the university administration and management also has to be satisfied that sufficient quality was maintained during the evaluation procedures, too. That’s why, I will have to see how things progress…
And, yes, I am still very much on the look out for my primary job. Don’t forget that part.