search

ukaea / PROCESS

PROCESS is a systems code at UKAEA that calculates in a self-consistent manner the parameters of a fusion power plant with a specified performance, ensuring that its operating limits are not violated, and with the option to optimise to a given function of these parameters.
https://ukaea.github.io/PROCESS/
MIT License
34 stars 11 forks source link

PF coil current allocation including elongation, triangularity, X-points #1420

Open jonmaddock opened 2 years ago

jonmaddock commented 2 years ago

In GitLab by @pc6783 on Oct 11, 2021, 15:40

Currently, the PF coil current allocation algorithm for Group 3 (equilibrium and shaping coils) uses a SVD formulation to target the vertical field at the plasma major radius while minimizing the sum of the squares of the PF coil currents. This is good, but it seems to me that this is a very solid foundation on which to include more considerations in the PF coil current allocation.

For a double-null divertor, one could use the existing SVD solver to target two additional field points: Zero (r and z) field at the X-points. You'd need to add the plasma current, which could be approximated as a filament at the major radius. Since the elongation and triangularity determine where the X-points are, this self-consistently folds elongation and triangularity into the current allocation.

Has anyone been looking at the PF coils? What do we think about this plan? I don't see any showstoppers as of yet.

jonmaddock commented 2 years ago

In GitLab by @pc6783 on Oct 11, 2021, 15:41

@stuartmuldrew You expressed interest in this at one point

jonmaddock commented 2 years ago

In GitLab by @mkovari on Oct 13, 2021, 13:38

Brilliant idea.

I am not sure that it would be sufficient to approximate the plasma current as a single filament, since this would underestimate the field at the X-point substantially.

Wouldn't it be possible to use this method for a single null?

I am concerned that to form an X-point you may need to adjust the current distribution in the CS as well as using a divertor coil. At the moment the CS in PROCESS is not broken into separate segments.

jonmaddock commented 2 years ago

In GitLab by @mkovari on Oct 13, 2021, 13:43

One more point - no-one currently working here understands the SVD algorithm - all we know is that it doesn't complain.