Discussion Forum

Here is my attached model

A quick look at your model shows several severe problems with your boundary conditions.
(1) On one of these rods you have placed lumped coaxial ports. ?? But there is nothing coaxial in your problem.
(2) On the other rod, you have set perfect magnetic conductor (PMC) boundary conditions. That makes no sense either. PMC conditions force the magnetic field lines to be perpendicular to the surface. Why would your cylinder have magnetic field lines perpendicular to its surface?
(3) Even if you fix those two major problems, your external sphere is too small. But then, that's probably the least of the issues here.
After all, it is easy to see that, even after you set the boundary conditions correctly, these will be very-very-electrically-short (i.e., infinitesimal fractions of a wavelength) rods. There is almost no place for any currents launched on them (whether by connection or induction) to actually go (except to induce miniscule levels of surface charges), even if you were to set the boundary conditions properly. This suggests to me that the actual physical problem in which you are interested is very probably something quite different than the geometry you created, but perhaps you thought your geometry was equivalent to it? Real-world systems of currents at such extremely low frequencies (e.g., the 10, 50, or 100 Hz in your model) are normally in the geometry of *closed paths*, not short stubby cylinders. Did you actually mean to have a closed current loop somewhere? Alternatively, did you mean for your cylinders to be *infinitely long*, but decided to make them short just to keep your problem computationally manageable? If so, you should instead be modeling this as two solid circles next to each other in a 2D model, since that would represent infinitely-long cylinders.

I hope that helps. Good luck.

Thank you for your reply, Robert.

Actually, I want to study magnetic induced current of a metal rod inside a resonator with a RF input. I think it would be easier to start with the magnetic induction between two rods.

I don't know how to setup an AC current on a rod in 3D. I have studied some cases of AC currents from the web but there are no similar cases to mine. Would you mind teaching me how?

To develop the required understanding of both the physics involved and the effective use of Comsol Multiphysics, I recommend that you work studiously and methodically through at least some of the examples in the AC/DC and RF Model Libraries. (You can update the model library by selecting File-->Help-->Update COMSOL Model Library.) If you are having trouble creating proper 3D models, practice building and executing several 2D models first, and be sure to model simple textbook examples (like the magnetic field around a single long, straight wire) and then be sure to compare your Comsol results to the analytic results in your EM textbooks -- I'm presuming you own at least a couple of good EM textbooks. Anyway, working through your EM textbooks and building your own Comsol models of the examples in those books, starting with the simplest ones, can be a good approach. And again, supplement that with exploring the library models specifically made available to you by Comsol (some of which are expressly tagged as "tutorial") and using the Comsol help system (which is actually sometimes useful), and (when you have specific questions) going directly to Comsol support or asking for advice from others here in the discussion forum.