Discussion Forum

Hello All!

I am using the magnetic fields no current interface in Comsol 5.2 to study a stationary, 2D permanent magnet structure. The goal of my work is to strike a balance between increased magnetic field strength and decreased permanent magnet material ($) before I build the magnet.

First, take a look at images 1-3.

- I have created a domain probe to find the average value of the magnetic field strength over a circle.
- I want to move the domain which is being averaged from theta=0 degrees to theta = 45 degrees, in say increments of 2.5 or 5 degrees. The probe should be re-evaluated at each location. The result should be the average field in the circle as a function of the circle's location.

First, i tried to define parametric curves along the midline arc, however this was inadequate for my application (see figure 4).

Next, I defined a parameter (theta) and did a parametric sweep. This gave me exactly the output I was looking for - however this required re-solving the PDE for each parameter step. I am hoping to couple Matlab with the model to try some basic optimization, which requires increasing the efficiency of the Comsol simulation.

Is it possible to have a nested solver, where the PDE is solved for the field as a function of space in the outer loop, then my "theta" variable is sequenced in an inner loop, and the average field probe is updated for each degree?

Thanks so much in advance!

best,

Reed