Discussion Forum

Hi

it looks like first your model is 2D-axi symmetric, and then you can plit it symetrically around some w=0 typically, so start simple with 2D-axi physics, and draw one coil "section" as a simple circular spot, with a current along phi

--
Good luck
Ivar

Ok, thanks. To apply the currents, I chooes the "electric currents" physics?

Mikhail,

a couple of comments to your model from a more physical point of view. Your drawing looks like a Helmholtz coil. So the two coils will not increase field gradients, they will decrease them. Helmholtz coils are known to produce particularly uniform fields. You will see that in the simulation when you compare a single coil to the two coil arrangement.
I don't see the purpose of the symmetry axis. Nothing in your setup is symmetrical regarding this axis. In an axisymmetric 2D model, as proposed by Ivar, the symmetry axis must extend through the centers of the coils.

If you need strong gradients you would rather define the currents in the two coils in opposite direction.

Cheers
Edgar

The "symmetry axis" I called its the axis, to define simply the position of the coils, so it would be clear, that the wires are positoned exactly above each other with a variable distance. I know that there is no such physical symmetry in this model.
The idea of the current direction came from a problem I tried to simulate some time ago with another tool (didn't have access to COMSOL at this moment). In the problem I have two permanent magnets, exactly above eacht other and with different assignments of poles. It was a 2D problem, they were positioned along y-axis. The first magnets hat the poles North in +y direction and South in -y direction. The second magnet had the South pole in +y, and the North pole in -y direction. The resulting field eliminated each other in between.

But anyway, the purpose of this try in COMSOL is to build one simply model, and then playing with parameters such as current direction to get different results. Later on I will try to add to the model more wires along the x-axis. So in the end there will be like four 1 turn wires in +xdirection, on same y-coordinate, and below them exactly the same number of wires on a different y-position, but on the same x-position as the above wires. And the interesting result is the field in between the wires.

Just be sure you get the quantities 'field' and 'field gradient' right. The gradient can be large even where the field itself is zero and vice versa.

Yes, I meant of course both parameters, field and field gradient.

Next question occured.
First I draw my cirlet, then I draw a huge Box around it, to describe the box as "region" and apply "air" as the material to the box. Then I apply "copper" to the wire. But what is with the space in the circlet? I should also apply "air" or "air1" to it, to make sure that the space in the circlet also belongs to the region, right?

Hi

as you said in 2D-axi cut view, you need some "air" domain to allow the field to "loop around", then a circe domain for the Cu, but if you do not want to look at what is going on inside the Cu you can limit yourself to the boundary. Check the model library you have several examples of coils, even Helm Holtz coils too

--
Good luck
Ivar

Hello again, thank to you all, I've managed to do exactly what I wanted.
I've attached a screenshot from the 3D view of the coils.

Now I am going to add more parallel coils along the z-axis. This is going to be a bit tougher, since I can't use the symmetry. So I guess I should drawing the coils in 3D.