Learning Forum

[Leon Allmannsdörfer]

Hello,

I´m simulating a special sensor by using Maxwell. It´s a magnetostatic simulation.

A gearwheel is imported as a step-file. Using a sweep oparation, the wheel turns for a certain angular. A magnet is placed in front of the gearwheel and between gearwheel and magnet is a tiny ferromagnetic box. With the fieldcalculator the fields inside the box are detected. If the objects are placed too close to each other, the signal changes drastically and is no longer usable.  

The mesh meanwhile changes every step of the sweep.

Is it possible, that the changing mesh is the source of the problem with the signal?

If it is, is there a way to use the same mesh for every step? If it is not, what could be the problem?

Thanks in advance

Leon

[Navya C]

Hi Leon,

What are you measuring as a signal here?

What do you mean by "the signal changes drastically and is no longer usable"?

If you are talking about the B field in the box, it will increase if you bring it close to the magnet.
Please share some screenshots of your setup to interpret it correctly.

Regards,

Navya

 

 

[Leon Allmannsdörfer]

Hi,

I´m using the field calculator to charge the magnetization of the box. Therefore I use the calculatet values of B, H and a constant. Following this, the angular between the calculated vector and a constant vector is charged. The cos() of this angular is the signal im measuring. An integration over the volume of the box, followed by a scaling is also needed.

Sometimes the calculation returns nan and sometimes the signal seems to be random, while it should be a cos()-function.

What could be the reason it returns nan?

In the latest attempts, i can see too high B-Vectors at the edges of the teeth of the gearwheel. I tried using a roung shape of the edges, but it only got worse.

Is there anything else i could try?

Regards,

Leon