openmc.plot_xs() should support NCrystal materials

[marquezj]

This is something we (@tkittel , @paulromano) forgot in PR #2222, but @zoeprieto brought to my attention. This function is really handy to plot the XS of whole materials, but as it is right now it ignores the contribution from NCrystal.

I will look into it, but I think it should be something easy to fix.

[shimwell]

I would be keen to see this feature and in general any attention the plot_xs function gets is good.

I've been tinkering with the plotter.py recently with #2413 and I think there is further scope to improve the functions in that file.

Looking forward to seeing NCrystal make it into the plotting function, would that allow us to plot things like parahydrogen at low temperatures?

[marquezj]

Hi @shimwell. Yes! NCrystal supports mixtures of scattering kernels, and each one is associated with the right nuclide, without need for pseudo isotopes. For instance, with the PR from @zoeprieto, it works like this:

cfg = """phases<0.95*para-H_T20.0K_ENDF8_massconvert_inelasticonly_dummydensity.ncmat&
                0.05*ortho-H_T20.0K_ENDF8_massconvert_inelasticonly_dummydensity.ncmat>"""
m1 = openmc.Material.from_ncrystal(cfg)
E1, xs1 = openmc.calculate_cexs(m1, 'material', ('total',))
plt.loglog(E1, xs1[0], label='95% para')

cfg = "para-H_T20.0K_ENDF8_massconvert_inelasticonly_dummydensity.ncmat"
m2 = openmc.Material.from_ncrystal(cfg)
E2, xs2 = openmc.calculate_cexs(m2, 'material', ('total',))
plt.loglog(E2, xs2[0], label='100% para')

plt.legend()
plt.ylim(0.2,100)
plt.xlabel('Energy [eV]')
plt.ylabel('Total Macro XS [cm^-1]')

(this is a typical example for cold neutron sources: check the effect of a given amount of liquid ortho-hydrogen in the emission of the source )

Edit: the ncmat files for these scattering kernels are in NCrystal-extra.

[marquezj]

Fixed by PR #2435. Thanks