`hstar` not properly calculated for every confinement scaling

[chris-ashe]

This is a follow on from #3493.

The variable hstar which is meant to represent the $H$-factor value if the loss power has no radiation correction (i_rad_loss = 3) is only true if the IPB98(y,2) scaling is used.

The current implementation is:

if physics_variables.i_rad_loss == 1:
            po.ovarrf(
                self.outfile,
                "H* non-radiation corrected",
                "(hstar)",
                physics_variables.hfact
                * (
                    physics_variables.p_plasma_loss_mw
                    / (
                        physics_variables.p_plasma_loss_mw
                        + physics_variables.pden_plasma_sync_mw
                        + physics_variables.p_plasma_inner_rad_mw
                    )
                )
                ** 0.31,
                "OP",
            )
        elif physics_variables.i_rad_loss == 0:
            po.ovarrf(
                self.outfile,
                "H* non-radiation corrected",
                "(hstar)",
                physics_variables.hfact
                * (
                    physics_variables.p_plasma_loss_mw
                    / (
                        physics_variables.p_plasma_loss_mw
                        + physics_variables.pden_plasma_rad_mw
                        * physics_variables.vol_plasma
                    )
                )
                ** 0.31,
                "OP",
            )
        elif physics_variables.i_rad_loss == 2:
            po.ovarrf(
                self.outfile,
                "H* non-radiation corrected",
                "(hstar)",
                physics_variables.hfact,
                "OP",
            )

• The calculation is also done in outplas() which is not recommended as outplas() should only be for output and not a place for doing calculations.
• This also means that hstar is not a globally tracked physics variable which can be used in other parts of the code.