Develop synthetic benchmark models

[paulromano]

A performance benchmarking repository has been set up and now needs to be populated with models covering different areas:

Geometry

• Nested cylinders
• Nested spheres
• Nested torii
• Complex geometry with 1000s of cells
• Regular lattices
• Hexagonal lattices
• BEAVRS

CAD Geometry

• Paramak models
• Stellarator model @paulromano

Physics / data / cross sections

• Assembly model with ~250 nuclides
• Problem stressing unresolved resonance region
• Problem stressing thermal scattering (for various moderators)
• Coupled neutron-photon transport
• Photon transport only
• Temperature interpolation

Sources

• Mesh sources (various meshes)
• Mesh source w/ domain rejection
• Point cloud source

Variance reduction

• Deep penetration problem (requires FOM metric) @jtramm

Tallies

• Depletion-like reaction rate tallies
• Mesh tallies (structured meshes)
• Unstructured mesh tallies @pshriwise

Depletion/activation

• Single timestep activation, single material, full chain
• Single timestep activation, many materials, full chain
• Decay source generation, many materials

[pshriwise]

@paulromano I believe you generated a nice tetrahedral model for the mesh volume calculation feature. Would that be a good candidate as well?

[paulromano]

Yup, @amandalund and I just synced up on this task and I mentioned that model, which would work well for the "complex geometry with 1000s of cells" after some modifications

[paulromano]

Needs for photon transport:

• Stress different energy regimes dominated by photoelectric, Compton scattering, pair production
• Testing thick-target bremsstrahlung
• Testing with and without atomic relaxation (need a setting for this in OpenMC)

[paulromano]

For thermal scattering:

• Point source at say 1 eV in a large sphere of a single material
• Materials:
  • H2O (incoherent inelastic)
  • Graphite (coherent elastic)
  • ZrH (coherent inelastic)

[paulromano]

For coupled neutron-photon, we want to create a problem that is heavy on photon production from neutron reactions. To do this, we can have a neutron source with a few discrete energies (using openmc.stats.Discrete) that correspond to strong resonances in Gd:

• 2.568 eV = strong (n,γ) resonance in Gd155
• 33.23 eV = strong (n,γ) resonance in Gd156
• 16.77 eV = strong (n.γ) resonance in Gd157
• 22.3 eV = strong (n,γ) resonance in Gd158
• 222 eV = strong (n,γ) resonance in Gd160