Hi all
I'm going to simulate an array of antennas around a cylindrical object. As can be seen, the structure is perfectly symmetric (Images 1 and 2).
But the meshing is different on aligned (image 3) and non-aligned (image 4) axial structures.
I think it is the reason of different results for identical antennas and configurations (image 5).
Please answer thus questions:
1. Is the different results due to the different meshing?
2. How can I fixed it ? (meshing enhancement is too time-consuming)
Thanks in advance
Amir
Hex mesh in the transient solver has to be structured, aligned across the Cartesian axes. Thats the probable reason for your discrepancy. You can try to manually increase the mesh, to level out these mesh errors, or use the Frequency solver, tet mesh like Giovanni suggested, but that will require a heck of RAM memory and simulation time.
You could try rotating the model by half the angle between antennas and see if that changes anything - to see if it is the mesh. I expect it is the mesh, and if you stay with the hexahedral mesh the problem will remain.
Have you set symmetry planes? If you can then you can increase the number of mesh cells.
Thanks to Giovanni, Marcelo and Malcolm, I try a new option in meshing: "Subgriding".
This technique reduced mesh cells significantly (image 1). But the problem in results still remains (Image 2). Inconsistency at high frequency confirms the Meshing problem idea.
But I realized something interesting, although the Sn,n results don't match each other, the similar channels match each other (Image 3)!! I think the problem arises only for (Sn,n)s.
The Impedance of ports are depicted in Image 4.
Can somebody explain the reason?
PS: Dear Malcom, I'll try your suggested method. Meanwhile, I can use symmetry planes just in main coordinates planes (Axial Aligned Antennas).
Thanks in Advance
- Badges
- Science topic