Dude, this is exactly what I needed. Thank you so much. It's maddening that there seems to be no middle ground between 2 hour lectures explaining theory whose application isn't clear until 15 hours later in the course, and 15 minute videos breaking it down into pretty graphics that make the concepts more digestible but are simplified too much to assist when it comes to the level of granularity needed actually writing the code. It's so prohibitive of active learning and it is pervasive throughout formal education. Thank you for teaching this way.
@PolymerFEM
Жыл бұрын
Thanks for your kind words. I agree!
@manuelgarcia9368
7 ай бұрын
Thanks a lot! This is a great example! However, is there a better explanation on the code vars somewhere? I'm having a wee bit of trouble understanding the 'physical' meaning of some vars (like conn).
@mattgolman
Жыл бұрын
I really appreciate this look into some simple FE code, I think it's going to help a lot for a project of mine. I have two comments though: 1. It seemed like the video didn't add much that the comments in the code weren't already contributing. It would have been nice to have some more detail on exactly what the code is doing mechanically, especially in the global stiffness and forces/BCs section. 2. The output seems to vary dramatically based on the number of elements in the x dimension. With a 9x49 grid as shown, displacement is about 6.7mm, but doubling the resolution to 19x99 elements brings the displacement to 14.3mm. Around 49 x 249 the solve time starts to get inconvenient, but the displacement is up to 36.9mm. It seems like you get .75mm per element in the x dimension. Do you have any idea why that is? It's my understanding that increasing the mesh density should converge on more and more accurate figures.
@PolymerFEM
Жыл бұрын
1. Good point about the BCs. I have been considering adding a video explaining more about how that works. 2. The code is written such that each node with an applied force is given a force of 20 N. So if you increase the number of nodes, then the total force increases. You will need to scale the applied force based on the number of nodes in order to do a mesh refinement study.
@aravindaousl
2 жыл бұрын
Great job Dr Jorgen....
@waylandhui8340
Жыл бұрын
Hello, I have read your post and python code and now I have two questions. The first one is about strip problem. I think this is a plane stress problem but you say this is a plane strain problem. I mean in the strip problem the force 200 N is parallel to 50*10 plane rather than 10*1 plane. The second one is about approximation calculation. I have no idea to get result 8 . I just get the simplest result is 10. Could please give me some inspirations to this calculation?
@harsh_hybrid_thenx
2 жыл бұрын
Dear Sir, i am seeing that mesh_hx should be mesh_lx/mesh_nx not mesh_lx/mesh_ex from my reasoning . Please correct me if I am wrong. With best wishes, Harsh
@skippylegrandgourou777
Жыл бұрын
Hi, first of all, thank you for your work it is really helping me to understand the basics of FEM. I don't understand why we chose to assume plain strain and not plain stress, could you explain please ?
@vukasinsl1
2 жыл бұрын
what a great stuff! thank you jorgen!
@ghufranullahkhan7479
2 жыл бұрын
This is a great stuff, but I would suggest a little more focus on the code would be great because that is the difficult part. Hope we see some more great content.
@PolymerFEM
2 жыл бұрын
Ok thanks! That's good feedback.
@tarkesdora20
Жыл бұрын
It would be great is something similar example exists for Explicit FEM
@naghmezzz7177
2 жыл бұрын
That was very helpful. Can you please talk about hyperelastic Material model in the case of UMAT?
@rezaafshar3832
Жыл бұрын
Great videos! Thanks a lot for sharing!
@ghufranullahkhan7479
2 жыл бұрын
I am getting the follwing error when i try to run this code, kindly suggest me some solution for it. " Error in qhull Delaunay triangulation calculation: singular input data (exitcode=2); use python verbose option (-v) to see original qhull erro". The plt.tricontourf function has some issue. Waiting for your reply. Thanks
@PolymerFEM
2 жыл бұрын
I recommend that you download the python code by clicking "copy" next to the code on this webpage: polymerfem.com/full-finite-element-solver-in-100-lines-of-python/ I then recommend installing a full recent version of Python (like version 3.9). I have used the version from anaconda.com in the past. I just double checked it and the python code works.
@ghufranullahkhan7479
2 жыл бұрын
@@PolymerFEM Thank you so much. This work fine. I hope to learn more from you.
Пікірлер: 18