Superconductor critical current density parameterisations are not all kludged #4630

Description of issue / requirement to address

Several critical current density surface parameterisations are available in PROCESS, see https://github.com/ukaea/PROCESS/blob/main/process/models/superconductors.py

In odd parts of the space (e.g. T < 1.5 K, B -> B_crit) many of these functions can return negative results, and possibly imaginary results, and also possibly NaNs.

I think all these functions should return real, positive numbers (ideally tending towards, but never reaching 0.0 for really odd parts of the space). This is purely for optimisation purposes, below e.g. 1 A/m^2 the current density is effectively 0.0 but some gradients may be required to stop the optimiser getting stuck.

Proposed solution

Check which of these produces negative, or imaginary or NaNs for bad inputs and kludge appropriately.

I don't believe that treating negative input temperatures should be required after #4628 is implemented, but they are probably an issue at present.