Lower Bound of Weight Window

User Support
1

Hello Everyone

Can someone explain following to me:

“We can also see that the weights are divided by a factor of two, which accounts for the typical 5× factor separating the lower and upper weight window bounds in OpenMC”.

wℓ(r)=ϕ(r)/2/max(ϕ(r))

Thanks in advance.

2

Hi @nrnaktr,

I suppose that it is based on the fact of 2 * 2 is almost 5, so the WW interval
[f(r) / 2, f(r) * (5 / 2)]
is almost log-symmetrical regarding to the central value of relative flux
f(r) = ϕ(r) / max_r(ϕ(r)).

The major principle of both MAGIC and FW-CADIS is based on the assumption of the optimal splitting rate for transition from point r_1 to point r_2 is approximately proportional to product of the statistical weight w and ratio of their flux densities [1]: w * ϕ(r_1) / ϕ(r_2). Sometimes, it is close to the truth.

References

1 . Cooper, M. A., & Larsen, E. W. (2001). Automated Weight Windows for Global Monte Carlo Particle Transport Calculations. Nuclear Science and Engineering, 137(1), 1–13. https://doi.org/10.13182/NSE00-34

3

I apologize for the mistake: FW-CADIS must be replaced with CADIS in my previous comment.