[nmag-users] Re: List Error

[Fangohr H.]

Dear Josiah,

I think your email didn’t go through due to the attached zip file. It is true though that you will not get a copy of the email even if it goes through - that is a bit unusual. However, you can check the email list archive at https://groups.google.com/forum/#!forum/nmag-users whether any email has arrived.

Regarding your question: it is hard to say what is wrong. My first guess is that the length scale of the mesh is inaccurate (probably too large). The main factor is the “unit_length” argument in this call:


sim.load_mesh(“bar30_30_100.nmesh.h5”,
[(“Py”, mat_Py)],
unit_length=SI(1e-9,”m”))

The example above (taken from http://nmag.soton.ac.uk/nmag/current/manual/html/example2/doc.html#the-simulation) means that a spacing of nodes in the mesh of 1.0 is to be multiplied with 1e-9 m to convert to real units. In other words, two hypothetical nodes in the mesh at [1,0,0] and [4,0,0] are in real units 3nm separated from each other.

If you get this number wrong (say 3m instead of 3nm), it may well result in very odd physics.

If you have trouble converting meshes into the nmesh format, then have a look at this section:
http://nmag.soton.ac.uk/nmag/current/manual/html/finite_element_mesh_generation.html#nmesh-file-format

which explains the ascii (text-based) version of the file format. In the worst case, you can write code yourself to write whatever mesh you have into an nmesh mesh.

Finally, I am not aware of any Netgen tutorials, but if you load a geo file into netgen (from the GUI), and then explore the menus a little, you may be able to learn a lot about it.

I hope this is useful.

Best wishes,

Hans



On 13 Mar 2015, at 16:20, Mullins, Josiah <JMUL...@gru.edu<mailto:JMUL...@gru.edu>> wrote:


Good Afternoon Dr. Fangohr,


I apologize by contacting you directly like this instead of using nmag-...@lists.soton.ac.uk<mailto:nmag-...@lists.soton.ac.uk>, but I have tried to send the below email and attached file to them for three weeks and I have not gotten the email that I sent through the list. Since I am a member of this list, should I not receive a copy of everything that I send? This lead me to the conclusion that the email was not getting through. I hope that you know the reason for this error.


Thank you for your help,

Josiah M. Mullins


P.S. In your tutorials, you say that you use netgen, but I can not find any tutorials geared towards beginners. Do you know of any instructions for beginners on this software so that I can use to learn how to use it? Thanks.


Good Afternoon Sirs,


I am learning to create meshes for use with Nmag using Gmsh version 1.5 on windows. I would use a newer version but the 2.x series saves meshes with region id 0 and up, and this is not compatible with Nmag's load_mesh function. Anyway, I created the mesh (rectangular.msh) from the .geo file (rectangular.geo), then I copied it onto my fedora 20 installation. Thereafter I converted it into the .nmesh.h5 format as in the example and ran Nmag as the instructions say, but instead of having the z component converge to 9e05, it decays to zero as do the x and y components. I immediately suspected the nmag installation as being faulty, so I recompiled and reran both the original mesh and my mesh. The original mesh behaves properly, but mine is still showing the same decay.


I have tried to included all files pertinent to the simulation to aid you in your diagnostics.


I apologize that this is so wordy and I thank you for your assistance,

Josiah M. Mullins


P.S. I apologize if you have received this multiple times, but I never received this email. Also, the file had been compressed to .zip, but due to failed(?) email attempts, I compressed it using 7Zip.



<2.02_results.txt>