Difference in neutron flux between OPENMC and MCNP6

Validation and Verification
1

Hi everyone,

I am encountering a 22% difference in neutron flux calculated between OPENMC and MCNP6 for a simple system with a fixed source. The flux was calculated in the volume of the system. From what I have read on the forum, it seems that the issue might be related to flux normalization, particularly when using the track-length estimator.

In MCNP6, I used the F4 tally with the SDEF source card, energy distribution and the total source weight. In OPENMC, I used a fixed source with the same energy distribuition and the total source weight, but the calculated flux was later divided by the cell volume.

**Source MCNP:** 

sdef pos=0 0 0 axs=0 0 1 vec=1 0 0 erg=d1 cel=1 rad=d2 ext=d3  wgt=1e20
si1 H 2.5119E-08 .... $ value in MeV
sp1 D 0.00 ....
si2 0 30.760 
si3  150 

nps 10000


**Source OPENMC:** 
 
settings = openmc.Settings()
settings.output = {'tallies': True}
settings.particles = 10000   
settings.batches = 50
settings.inactive = 0
settings.run_mode = 'fixed source'        
strength = 1E+20
space = openmc.stats.CylindricalIndependent(
      r=openmc.stats.Uniform(0, 30.760),
      phi=openmc.stats.Uniform(0, 2 * np.pi),
      z=openmc.stats.Uniform(150, -150),
      origin=(0.0, 0.0, 0.0)
)
angle = openmc.stats.Isotropic()
energy = openmc.stats.Tabular([ 2.5119E-01, .....], [0.00, .....]) #value in eV

source = openmc.IndependentSource(strength=strength,angle=angle, space=space,energy=energy) 
settings.source = source

settings.export_to_xml()

I’m not sure how to reduce this difference between the results from the two codes. Could someone provide guidance on how to proceed or if there is something specific I should adjust in the normalization process?

Thanks in advance for any help!

F4
F41119×41 7.26 KB

NETONFLUX
Track

2

@Criss The results of the tally can also come because the tally is different. Can you share the whole input? Also, it seems it is not a normalization difference, but something else (the MCNP result looks more moderated than the OpenMC result).

From what you shared, it looks like for the spectrum of the source you are using the default linear-linear interpolation in OpenMC, and histogram interpolation in MCNP. Maybe that could be a cause of difference.