Radiation correct of H factor

[jmorris-uk]

Original issue (#652)

I have just looked at our calculation for Hstar (non-radiation corrected H-factor) in plot_proc.py and I believe it is wrong. In our radiation correction we calculate P_loss = P_heat - P_core. If iradloss = 1

The current implementation reads

hstar = hfact * (powerht / (powerht + psync + pbrem)) ** 0.31

I think this is only valid for our old radiation model, where pcore = psync + pbrem but I think it probably should be (if iradloss = 1)

hstar = hfact * (powerht / (powerht + pcore)) ** 0.31

or if iradloss = 0

hstar = hfact * (powerht / (powerht + prad)) ** 0.31

Papercut review

Agree that quantities should not be calculated in plot_proc.py, all calculated quantities should exist inside of process. The hstar calculation should be removed from plot_proc.py.

Rich points out on the original issue that rather than correct for one scaling post calculation elsewhere in the code, each confinement time scaling should be reviewed and any radiation correction should exist in the confinement time function itself. Producing both a corrected and uncorrected value.

The part of the code which calculates the energy confinement time could be rewritten to calculate it for corrected and uncorrected
powers, and PROCESS could output a proper H-factor for the confinement time scaling selected. This would be a bit of work but >
not, I think, excessively so. We shouldn't (in PROCESS) fix the calculation for only one confinement time scaling.

For the ITER98(y,2) scaling (isc=34) the proposed correction was

The radiation correction (with iradloss == 1) has only been matched to the ITER98(y,2) scaling. So for all other confinement scaling there should be a warning, that this has not been evaluated yet.

pinjmw = pnbitot - nbshinemw - porbitlossmw

and this should also be taken into account by writing

if (iradloss == 1)
    hstar = hfact * (powerht / (powerht + pcore + nbshinemw + porbitlossmw)) ** 0.31
elif (iradloss == 0 )
    hstar = hfact * (powerht / (powerht + prad + nbshinemw + porbitlossmw)) ** 0.31
else
    hstar = hfact * (powerht / (powerht + nbshinemw + porbitlossmw)) ** 0.31

The original issue has been closed and discussion on this model moved here.