Learning Forum

[mdibon]

I am currently trying to simulate the vertical motion of a fusion plasma and the resulting eddy currents and forces in the surrounding structures. The plasma is modelled as a toroidal conductor with circular cross section. The current in the plasma is 5.5 MA. Terminals for this current are on both ends of the plasma section.

The motion of the plasma works well. However, no current is flowing in the plasma. The currents in the surrounding control coils are correct. If I delete the motion setup, the current in the plasma is flowing correctly.

Is it possible to apply a current to a moving conductor?

[Navya C]

You can assign current to the moving conductor. If there is any issue with that, maxwell should generate an error wile validating the model before running the analysis.
With that said, I couldn't quite understand what your model is exactly.
Can you add some annotations to represent different parts in your model?
also, add some screenshots to show how you defined the band and excitations.
Regards Navya

[mdibon]

thank you very much for your quick reply.
The validation shows no errors and the simulation runs smoothly, which makes it surprising that there is no visibile current in the plasma.
The model is a simplified representation of a Tokamak fusion experiment. This model will be used to check EM forces on structures if the plasma uses vertical stability and moves in z-direction. The screenshot shows the components of the model.
Here are the plasma parameters.
And here are the band parameters.
There are no curved surfaces in the moving region. The vacuum region, which encloses the model, is shown in the next screenshot.
The experiment is axisymmetric. To conserve resources we only calculate a 40┬░ slice. The calculation of the full axisymmetric model led to the same problem, that there is no current in the plasma.
Again, thank you very much for your help. I hope you have some ideas, what we can do to solve our problem.
Best regards mdibon

[Navya C]

Hi @mdibon Since you have a motion set up I suggest you use matching(master/save till 2020R2) boundary condition.
When you define current, do not define it at the surface of the boundary but somewhere in the middle.

Regards Navya

[mdibon]

thank you very much for the suggestions.
With the master/slave boundaries at the edges and coil terminals in the XZ plane I get an error message that the band intersects with master/slave boundaries, which is correct. But I don't see how that can be resolved.
I also made a simplified full axisymmetric model consisting only of a segmented toroidal conducter in a polyhedral band in a region. The coil terminal is in the YZ plane.
The conductor moves perfectly, but there is zero current flowing in the conductor.
Somehow the winding excitation is completely ignored inside the band.
Is the solution type 'Magnetic transient' the correct solution type? 'Solve inside ' is active in all geometries.
Best regards mdibon

[Navya C]

Hi @mdibon Use two bands with more segments as shown below
It worked for me when I added an inner band also.

Regards Navya

[mdibon]

Dear thank you so much for your help. The inner band does the trick.
However, when I try to use a ring shaped inner band I get an error massage.
A ring shaped inner band and outer band would be necessary for my problem, because of the ring shaped vacuum vessel containing the conductor. Is this possible in Maxwell?
I thought of using a slice of the ring shaped geometry, but it produces the same error massage.
When the inner band is a non-model object I don't get this error, but the current in the conductor is zero again.
Master/Slave boundaries generate an error, that these boundaries cannot touch the band. This error persists even when the boundaries are well inside the band.
Thank you very much for your help.
Best regards mdibon

[HDLI]

There is a closed steel frame to fully cover the inner coil (plasma) in the geometry, so there is no entering for the current in both physics and geometry. We could not apply the current on the inner coil with motion in the real application if the geometry is correct.
Actually, Maxwell will follow the real physics, although it is just a FEA simulation. Thus, we could not apply the current on the inner coil for the existing model now.
Could you double check the geometry, what is the real excitation and how to enter the current for this rotating inner coil? Thanks.
HDLI

[HDLI]

This issue is that if there is a conducting object inside the band, then the limitation is that it must be a ÔÇ£simply-connected domainÔÇØ (no holes, and any circle that you draw in the volume can collapse to a point within the volume). We must be able to fit both a solid band and Inner Band so that there are no holes in the moving assembly.This is only critical when there is a conducting body within the band.
With the real structure, we could find an entering for the excitation in the geometry's center and could build a solid Band without the hole for it.
HDLI

[mdibon]

Dear thank you very much for your reply. You are correct. There is no current terminal for the moving conductor in reality.
In the real structure, the current in the moving conductor (plasma) is induced by ramping a current through the central solenoid. This ramp down usually lasts for 40 seconds. The vertical motion however takes place within a couple of milliseconds. Hence, the current in the central solenoid can be set constant over the time of the motion.
If we were to simulate the induction as well, how would the solid band and inner band look like?
This is the toroidal vacuum vessel with the control coils and the central solenoid (green)
Inside the vacuum vessel is the stabilizing plate.
and the toroidal plasma which carries the 5.5 MA current and moves vertically.
If you need more information, please let me know.
Best regards mdibon

[HDLI]

Both the central solenoid and toroidal plasma will move vertically, not rotating, is it correct? Could you clarify it? because rotating motion is used before.
Is the toroidal plasma a chunk solid or a coil with a few turns? Thanks.
HDLI

[mdibon]

thank you for the quick reply.
The central solenoid is stationary, while the plasma moves in z-direction. The plasma is one solid conductor with a single turn.
Best regards mdibon

[HDLI]

Hello @mdibon Thanks. For this, we will just set Plasma material with conductivity that allows Maxwell model generates eddy current, and then could build both Band and inner band for the hollow moving object, but this model could not be solved in 21R2 version for some reasons, while setting eddy effect on the plasma and using translation motion. Without the eddy effect, we can run the simulation with the translation motion.
It is no issue in the simulation with the rotation motion.
Thus, unfortunately, we may have to wait for 2022R1 version to fix or have enhancement for the translation motion. Thanks.
HDLI