Starting Particle Weight

User Support
1

I am reasonably new to using OpenMC. However, I have a question about how OpenMC handles starting particle weights. I have a custom source routine, which needs to bias the sampling of source particles to areas of low gamma emission. To do this biasing I am sampling the starting positions of the gamma particles uniformly and then adjusting the starting particles weight (particle.wgt) to take account this into account. This weight adjustment is done within the custom source routine. However, when doing a comparison to the same model in MCNP I get different results. Having looked at this for a while I am pretty sure this is down to the way OpenMC handles particle weights. In my case I have noticed that dividing the weights of the starting particles by an arbitrary factor (such as 2) does not appear to change the result of the calculation. In other transport codes dividing the weight of the source particles by 2 would half the results.

My questions are therefore: can OpenMC handle starting particles with weights > 1? and does OpenMC renormalise the weights of the starting particles at all?

Many thanks,

Tim

2

Hi Tim,

I assume you are running a fixed source calculation – in this case, there is no renormalization of the weight of the starting particles. However, it appears that at the time we accumulate tallies, there is a normalization based on the total weight of the starting particles. I think perhaps it should be normalized based on the number of particles rather than the total weight, but I will investigate further to make sure we don’t break other things in the code. I’ll go ahead an create an issue on github so we can track this; thanks for bringing it to my attention!

Best,
Paul

3

Speaking with Andy it looks like changing from current normalisation method to it being based on number of particles would help with this PR https://github.com/openmc-dev/openmc/pull/1251