Learning Forum

[Be Ru]

I am quite new to Ansys and I am trying to simulate a toroidal inductor. Since it is absolutely symmetric, I tried to only simulate a quarter of the toroid to reduce computation time. However, since I split the core I always get the error "Current leak to the air". I can't really understand where it is coming from. The current path is in the outer coil (see screenshot) and is still a contiuous path. Also, I added insulating boudaries to the coil so the iron core doesnt short it.

 

For the region, I added independent and dependent boundaries to the faces, where the rest of the core should contiue.

 

I also tried changing the mesh settings like it was recommended in another post. 

 

Nothing really helps currently and I am wondering what is wrong. Can someone please assist?

[Reshmi Raghavan]

Hello ,

Request you to please assign the excitation on the sheet on the outer part of the coil as shown in the below snapshot. Another point to note here is that the boundary condiction can be symmetric normal as the flux would be flowing normal to the cut planes.

[Be Ru]

Thanks, that helped as a first step. Nevertheless, I still get an error after about 10 passes. Additionally, the H-magnitude looks wrong to me. I set the boundary conditions to symmetry and normal flux.

[HDLI]

Hello Be Bu,

      Could you try to enable "Use dynamic surface resolution" in initial mesh setting for "leak" issue?
      Additionally, I think your matching boundaries should be good too.

HDLI 

 

[Be Ru]

Sadly, it still fails after a few passes. The B-vec in the core looks right, but the magnitude of H in the core cant be right. For mesh method, it doesnt improve anything if I change to "Classic" method or back to auto, also enabling/disabling "apply curvilinear meshing to all curved surfaces" or make the mesh finer does not help. Do you have any idea what could be the problem?

 

Thank you!

[HDLI]

Hello Be Ru,

1)     Please use matching boundaries.

2)     Could we use “current density terminal” and “current density” for this excitation of 3D model?

Please assign “current density terminal” in cross section and then “current density” in the object. We could use cylindrical coordinate in current density setting.

HDLI

[Reshmi Raghavan]

Hello Be Ru,

Please provide a small gap between the core and the coil and run the simulation.

[Be Ru]

Thanks for your replies!

@HDLI: 1.) so i should go back to dependent/independent boundaries? 2.) I assign the current density terminal to the outer radius of the coil.

And the current density to the complete coil with cylindrical coordinates:

Do I have to assign all 3 current parameters? What would be typical values? Since the current density on the inner radius is higher than the outside, how do I cope with that?

 

@Reshmi Raghavan: I added a small airgap between coil and core, but still the simulation failed.

[HDLI]

Hello Be Ru,

       1). Yes. 

       2). It is good that we just enter current density in Z because current density termial is XY plane and current is a loop. Others of X and Y could be zero. 

        Please enter current density = current / area of current density terminal.

HDLI

[Be Ru]

Thanks, it now seems to work, at least for a magnetostatic solution type. However, if I change it to eddy current solver and add e.g. a linear count for the current in the optimetrics, I get the "current leak to the air"-error again. I set the eddy-effect under excitations to coil and core.