Virtual Plasma Control System (FreeGSNKE Pulse Design Tool)

[kpentland]

Hi all,

In this PR, we introduce a virtual plasma control system (PCS) class that couples with FreeGSNKE's evolutive solver. This is useful for plasma scenario and control design enabling users to build plasma, shape, and position controllers that regulate the poloidal field coil voltages applied to the evolutive solver at each time step. It also enables one to encode machine safety limits such as coil voltage and current limits and coil activation times. It is inspired by the MAST-U PCS and more details can be found in the pre-print on arXiv.

The easiest way to see how the virtual PCS works is to check out the above paper and the new Example11 notebook. The different controllers can be used all together in the PCS framework or individually (for example, you may only care about the vertical controller, not the plasma current controller).

As can be seen this is quite a large PR and includes many small fixes to the FreeGSNKE ecosystem. The main changes related to the virtual PCS are the:

• "freegsnke/control_loop" directory, which contains the PCS class ("pcs.py"), the individual modules that handle the different aspects of plasma control in this class (e.g. the "_category.py" files), and some "useful_functions.py" that enable all of the above.
• Example11 notebook which demonstrates how to use the new virtual PCS with the evolutive solver in the MAST-U-like machine.

Other changes:

• Example5 has been renamed to Example5a (to make way for two new notebooks Example 5b and 5c).
• Example5b demonstrates how to use the (linear) evolutive solver without solving the full Grad-Shafranov equation at every time step (for rapid simulations to check scenario feasibility).
• Example5c demonstrates how to use the (linear) evolutive solver (with or without solving GS at each step) but with automatic relinearisation throughout the simulation (i.e. updating the internal Jacobian matrices).

Here we can see a few screenshots from Example11 demonstrating the control of plasma current, shape parameters, and vertical position in the MAST-U-like tokamak:

Plasma current control:

Vertical position control:

Shape parameter control:

[kpentland]

Some comments are questions or suggestions; others require a proper fix before merging.

Apart from all the comments throughout the code, just wanted to add the code would benefit of a systematic suite of tests (I can help with this). Also, there's virtually no type hint coverage in the arguments of the methods: this would help greatly when actually using the code (some methods have +20 arguments) and for debugging purposes.

Thanks for the notes @capstayn! I've addressed them all (except for the unit tests) and will wait now for @georgeholt1 to do a quick Codex review. Maybe we can add some tests before that.

[georgeholt1]

@codex review

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: c2fffd9f47

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[kpentland]

Thank you all for the review comments and help building this PR (it's a big one)!

Hopefully should enable some more control-orientated simulation studies!