Develop the ability to use plasmod/jetto profile directly in PROCESS

[ym1906]

When the plasma profile was refactored, it was done so with a view to using completely custom profiles. However, a custom case was never developed or tested in practice. Now we should test this ability.

1. Figure out the best way to import profile data into PROCESS.
2. Modify the code to handle this profile (add new argument, and create the profile object correctly)
3. Point generic profile object to data
4. Run, test.
5. Develop a tool for profile comparison between PROCESS and external profile.

Tools to read JETTO files:
https://git.ccfe.ac.uk/jintrac/jetto-pythontools

[mkovari]

See #766 (also #3254).

Most codes output profiles as a function of the normalised flux (usually poloidal flux), a quantity that PROCESS does not use. We would need:

• To be able to convert from the normalised minor radius to normalised flux, and
• To be able to calculate the volume element dV. In all JETTO runs, dV/drho is output in the signal profiles / DVEQ.

[mkovari]

Once the profiles are fixed, perhaps normalised by the central temperature and density, what variables remain for PROCESS to optimise? Is it still self-consistent to vary the major radius, aspect ratio, toroidal field, impurity concentration, plasma density and temperature?

Did you test an optimisation run using a custom profile? @ym1906

Opinions @stuartmuldrew @ajpearcey ?

[ym1906]

I have implemented this on the branch. I decided it was best to write the jetto data to json, and PROCESS just reads in the json with the profiles. The script to do this isn't very complicated but it uses a tool written by the jetto team and probably should stay outside of the PROCESS repo.

I have attempted an optimised run of something similar to STEP. This input file does eventually solve without a custom profile. When the custom profile is used, it cannot find a converged solution and this happens very quickly. Further investigation is necessary.