Learning Forum

[Shi Hao Ng]

https://github.com/Ice-Citron/FEA-Physics_IA/blob/main/First%20Runnable/final_read%20-%202.k

https://github.com/Ice-Citron/FEA-Physics_IA/tree/main/First%20Runnable/Second%20Runnable

 

I am trying to simulate a railgun of a particular topology and size for my college Physics Coursework. I am very new to LS-Dyna, but has managed to get used to LS-Dyna's structure and everything, etc. I am using the railgun exemplar as my guide. 

On very above of this post, these 2 links are my GitHub repo which contains the 2 variants of keyword file I created, which are based on the railgun exemplar on LS-DYNA examples: https://lsdyna.ansys.com/em-railgun/#animated-result

The issue that I am facing is this:

I am using the exact same settings as the exemplar, as far as I am concerned. The only difference between my .k file and the exemplar is that I am using a different solid mesh for my railgun, as I will be using a different topology soon after. My mesh is created from .STEP file exported from fusion 360, which I then selected one of the faces, and then mesh it to 1.2mm with auto-mesher, and then used element generation to generate the rest of the component using shell offset to solid shell. 

I checked the contact of both files, as in penetration and contact edges. The exemplar itself seems to also have penetration + contact edges natively too, and its simulation works despite the initial penetrations.

My version is the second one, and even when I try to remove the penetrations, etc. The same error output file is produced, which thus I in a way think penetration isn't the issue here with FEM-BEM solver being unable to converge?

I hope someone is willing to attempt to download my .k source files from my git repo, and then try and see if they can solve it, so that it runs like the one in the exemplar. 

 

 

[Shi Hao Ng]

{Version I am using: ls-dyna_smp_d_R13.1.1_27-g8731a0d8c5_winx64_ifort190.exe} (_d_, double precision, so I don't think there is an error source from precision.)

Furthermore, I had stumbled upon a post and tried to change the “NUMLS” in em_control, from 100 (of the exemplar), to 1000. Which, this solves the issue with the FEM-BEM failing to converge, which I don’t understand why. And as seen in the image below, it says that NUMLS should only be in effect when “EMSOL” == 2, which in my case “EMSOL” == 1, as I am using the Eddy Current Solver. Regardless, as seen in image below below, a different type of error is outputted. And when I try to view my calculated d3plot file, the result makes absolutely no sense, this applies for both the exemplar file and my personal .k file, in the sense that their result makes no sense once “NUMLS” is changed to 1000, which is weird it’s even taking effect in the first place.

This is likely irrelevant, but just an interesting detail that I noticed. But to be honest, I just don’t really understand why is the FEM/BEM failing to converge, because my mesh shouldn’t really be erroneous? You can also refer to the source of the .STEP file in my Github repository: https://github.com/Ice-Citron/FEA-Physics_IA/blob/main/src%20STEP%20file/Square%20Augmented%20-%20Further%20Splitted.step

I hope someone can help me out of this, as I am out of options now, both solution and time, as I had spent a lot of time on this and my deadlines are coming up for my Physics Coursework. You will have my deepest gratitude. Thank you.

 

[Shi Hao Ng]

new update: https://github.com/Ice-Citron/FEA-Physics_IA/tree/main/FP16%20archive%20-%20cords_flattened/Modified%20-%20fp16%20%2B%2075segment100mm

Update: I changed the coordinates manually, so that it fits the exemplar, as in all the nodes that are supposed to be in contact are now the very same. Yet the error still is as seen in the image below. Now assuming there is no more issue with the meshing, what can possibly be the issue with the armature not having contact with the rails? Maybe I should just try and do the conventional classic "em_contact" instead?