search

simulkade / FVTool

Finite volume toolbox for Matlab/Octave
http://fvt.simulkade.com
BSD 2-Clause "Simplified" License
97 stars 56 forks source link

Spherical coordinates not working #18

Closed avnarla closed 5 years ago

avnarla commented 5 years ago

I tried running the diffusiontutorial_spherical code, but I commented out line 20, and uncommented line 19 (switching BCs from left to right), and the computation no longer performs as expected. As I am just making a symmetric change in the BCs, is there a reason it doesn't work?

Thanks!

simulkade commented 5 years ago

Is the problem solved @avaneeshnarla ?

avnarla commented 5 years ago

Hi,

I guess the issue was more that I was confused as to what the boundary conditions mean in the Spherical, Cylindrical, Polar and Periodic cases. What does top/bottom mean when you are dealing with a circle? Is there any simple documentation I can go to? Right now, I have been trying to understand the definitions of the boundary conditions in each case.

avnarla commented 5 years ago

I think I figured it out. But also, I wanted to work in a radial geometry, but not have the annulus in the middle. Is there any way to avoid that?

simulkade commented 5 years ago

Sorry @avaneeshnarla for my late reply. Very busy these days with students. I don't think you can have that. In fact, what you want to do cannot be done with a structured grid, and FVTool uses only structured grids.

avnarla commented 5 years ago

Hi Prof. Eftekhari,

No problem at all with the delay! I really appreciate your help. I realised the same after poring through the data structures. I decided to adapt your code and use conformal mappings to achieve the same.

Thanks!

simulkade commented 5 years ago

Hi @avaneeshnarla . Feel free to submit the PR if you like to add you code to FVTool. Let me know if you have difficulties going through the code.