TF coil cross-section

[jonmaddock]

In GitLab by @mkovari on Sep 15, 2014, 12:28

Modify so as to eliminate problems with negative space available for conductor.
It might also be useful to simplify the winding pack as a trapezium.

Maybe the solution is to redo the the calculation starting with the coil width and going outwards.

@mkovari needs to redo the max B calculation, with improved parameter ranges - such as a coil that is very thin or very thick in the radial direction

• say 0.05 < z < 0.9

[jonmaddock]

In GitLab by @jmorris-uk on Oct 30, 2014, 15:53

Note just to add that we should swap to the rectangular winding pack shape to be in line with the EUROfusion projects.

[jonmaddock]

In GitLab by @mkovari on Nov 3, 2014, 09:35

A rectangular winding pack would create problems for some models that require a large radial depth - the consequence would be a winding pack that is narrower than it should be, giving a higher peak field. A trapezoidal winding pack makes better use of the space, and is also closer to that used in the stress model.

It might be good to have the option of a rectangular winding pack to allow us to replicate the EUROfusion model, but I am not convinced of this. Probably not worth a lot of effort.

[jonmaddock]

In GitLab by @ajpearcey on Aug 5, 2022, 13:36

Is this issue still relevant? Or have it be superseded by newer models?

[ajpearcey]

Is this now captured by the options in i_tf_wp_geom documented here.

Do you have any comments @mkovari @jmorris-uk

[mkovari]

Yes, the rectangular winding pack is now an option(i_tf_wp_geom = 0).
The remaining suggestion is still valid: We need to redo the max B surrogate model, with improved parameter ranges. Perhaps it might be possible to use Bluemira to do this rather than Matlab as before. @je-cook ?
@ajpearcey

[je-cook]

We have some work ongoing for TF coil winding pack design and optimisation Fusion-Power-Plant-Framework/bluemira#3043. Once this is in a B surrogate model could be created for PROCESS.

Current wip image for a taster of whats coming:

[ajpearcey]

Which max B field calculation need to be redone? Is the ripple correction?

What data was used for the surrogate before? Would we to some xy-plane magnetostatics for machines with 16, 18, 20 etc coils of various wp geometries.

[mkovari]

The non-axisymetric part of the toroidal field is important on the outboard side because it affects the plasma. In Process this is called ripple. It is important on the inboard side because it affects the peak field in the superconductor. A while ago I did some MATLAB calcs and created surrogate models for both of these, but as always the parameter range was not quite enough. The geometry I used was trapezoidal, I think. Ideally we would have either a physical model or a surrogate model for all the different geometries available. I did mention this to James C as he might be able to do something using Bluemira. It is still possible to redo the original Matlab calcs for a wider parameter range and choice of geometries.

[ajpearcey]

I propose that we wait till we can use Bluemira, then we can build new surrogates with a wider range of width to depth ratios.

[mkovari]

OK with me. I suggest we keep the issue open but mark the Papercut as complete.