Discussion Forum

I am replying to my previous message. I used 7zip to zip the file! If someone could please take a look and see if they know how I can mesh correctly, I would appreicate it. Once again, I am using the CAD import module on COMSOL 6.1, and I imported a magnetic shield that I built in SolidWorks. I am new to this module and am having a hard time getting the mesh to work. Thank you!

Victoria,

the 7z file is empty. Compressing Comsol files is not efficient because they are zip files essentially.

You can ask Comsol support to help upload a bigger file.

Cheers Edgar

Hi,

Thank you for the reply! I have successfully reduced the size. The file is below.

Victoria,

this is quite a complex thing and I am afraid I cannot analyse it in detail. CAD imports can be difficult, because often the original CAD design wasn't done with FEM and the need of meshing in mind. So you may want to analyse the original CAD design for inaccuracies, intersections, missing conjunctions and the like. Or build the thing in Comsol which gives you more control of the geometry and also allows easy parameterization. You have thin structures in the geometry. This will likely require setting up a custom mesh. In most cases I start to set up meshes in the inner delicate parts and make my way to the outside boundaries of the geometry. It may also be possible to replace thin domains by suitable boundary conditions.

Good luck Edgar

The shield would be much better represented by zero thickness boundaries, which you can assign the magnetic shielding boundary condition. I would created the geometry within Comsol, it would not be very complex (see attached simplified model).

Note, I have not added the cut-outs in the shield (but would be simple to do). The magnetic shielding boundary condition allows you to add a thickness. In this model I have defined it with a relative permeability, but you could assign a BH curve (the one from the Silicon Steel material) too.

Obviously the attached is not to scale, and uses different current in the coil, but you get the idea... (note it is a version 6.1 model).

By doing something similar to the attached, you will avoid the horrors of meshing that imported geometry.

Mark

Thank you both so much! I will continue to work on it.

The shield would be much better represented by zero thickness boundaries, which you can assign the magnetic shielding boundary condition. I would created the geometry within Comsol, it would not be very complex (see attached simplified model).

Note, I have not added the cut-outs in the shield (but would be simple to do). The magnetic shielding boundary condition allows you to add a thickness. In this model I have defined it with a relative permeability, but you could assign a BH curve (the one from the Silicon Steel material) too.

Obviously the attached is not to scale, and uses different current in the coil, but you get the idea... (note it is a version 6.1 model).

By doing something similar to the attached, you will avoid the horrors of meshing that imported geometry.

Mark

Hi,

I want to follow up with this message. I followed this much simplier method of creating the magnetic shield and was able to mesh it successfully, thankfully. However, I am now running into an error when computing. It reads something similiar to "maximum number of Newton iterations reached" and I have been able to fix this error in the past when this error message appeared in different simulations; however, the methods that I have tried are not working now. I ignored many of the edges in the shield; however, I am not sure if that is helping me in this situation. The "ignore edges" can be disabled if need be. I also tried decreasing the relative tolerarnce from .001 to .0001, which also did not help. I cleared the mesh and the solutions and set the relative tolerance back to its original value of .001 and I have attached my file below if anyone could please take a look and let me know if they see what is wrong. Let me know, and thank you so much.

I think you can delete the "ignore edges"...

Also, I have changed the length factor in the Geometry Analysis to 4 (doesn't affect the solving but will affect the values of results).

In order to troubleshoot, I would start by making some further simplifications, maybe making your shield with a relative permeability magnetization model (maybe start with a value of 1000).

Also, start off with linear discretisation in the mf node, rather than quadratic.

I tried the above tweaks and it solves in a few minutes. You can then try adding in a BH curve definition to see if it is that causing the model to not solve... if it is, then there may be some further guidance you can seek with the knowledge that it is the BH curve defition that is causing the model to fall over.

As always, start simple and include more complexity if it is justified...

So, basically I have a working simulation involving two magnetic shields and coils that are actually going to be constructed very soon for the project that I am working on. I had to change the shape of the shield which is why I am now working with the file referenced above.

In the file above, I see that the model solves very quickly when using the relative permeability magnetization model; however, I have to use the BH curve model. When I use the BH curve model, the computation fails. In my working simulation, I use the BH curve model, use a quadratic discetisation, and I use the direct solver. I tried running my new simulation with the more simple settings that were suggested (linear discreisation, relative permebility magnetization model, and a different solver) but it did not work. I then enabled the "ignore edges" again (because in my working simulation, ignoring the edges was necessary) and I then tried computing with the same more complex settings that I have in my working simultion; however, that did not work either. I keep getting the same error message stating "Failed to find a solution. Maximum number of Newton iterations reached. Returned solution is not converged. Not all parameter steps returned".

I am not sure what is different between my new simulation and my simulation that works other than the fact that I changed the shape of the shield. I tried using a mesh with less degrees of freedom because I know the direct solver uses a lot of RAM; however, that also did not help. I attached both simulations below. I have not added the 2nd shield into the new simulation yet because I want to get it working first with just the one shield and then I will add in the second shield. If you / anyone could take a look I would greatly appreciate it!

The older simulation that works with the 2 shields is the one called "magshield_newcoils_fixed" The newer simulation is the one called "magshield...simplified"

OK, so it seems you have identified the BH curve as the source of the failure. The following Blog post may have the answer for you.

https://www.comsol.com/blogs/how-the-b-h-curve-affects-a-magnetic-analysis-and-how-to-improve-it/

The Blog looks at what can happen if the BH data is not smooth enough, and the error you have seen is mentioned.

It seems you may need to add some additional points and/or run your data through the B-H curve checker application (linked to in the Blog).

The material you use from the built-in library (Silicon Steel GO 3408 does not seem to have been optiomised according to the list on the blog).

I hope the above helps!

Mark

Hi Mark,

Thank you for the tip! I am trying this now.

The BH curve checker solved the problem!