ZedThree / s_alpha_gpo_poc

Proof-of-concept ideal ballooning stability boundary using Gaussian Processes

This is a demonstration of finding the ideal ballooning stability boundary with the gyrokinetic solver GS2 and Gaussian Process optimisation (GPO).

There are two Python scripts that each run 25 GS2 simulations across the same set of parameters to find the ideal ballooning stability boundary. scan_with_pk.py uses Pyrokinetics to do a 5x5 uniform scan over s and $\alpha$, while scan_with_gp.py uses GPO from Inference Tools to do the same scan in a more flexible way.

WARNINGS:

• this is just a demonstration of the techniques, and not necessarily the best way to achieve the results with these tools;
• the simulations are almost certainly massively underresolved in order for the whole scan to run quickly, and no physical interpretations should be made about the results;
• at the very least, you will need to change the path to gs2