You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Describe the bug
Currently, when uniform bremsstrahlung splitting (UBS) is turned on with Russian Roulette, we do not split fluorescent photons resulting from atomic relaxations induced by electron impact ionization (EII) during Moller scattering of phat electrons. This results in the emission of phat photons.
To Reproduce
Use BEAMnrc to simulate a low energy (e.g., 225 keV) electron beam incident on a tungsten target using uniform brem splitting (UBS) with a splitting number (nbr_split) of, say, 100, and with Russian Roulette on. Ensure that atomic relaxations and EII are both ON.
Score phase space below the target
Run many histories (>= 10e6)
Observe photons with weight=1 in the phase space data
Expected behavior
Scored photons should all have weight 1/nbr_split. Scored charged particles should all have weight 1.
Operating system
[x ] Linux
Windows
[ x] macOS
Other:
EGSnrc version
develop branch
Other
This showed up when testing the egs++ DBS algorithm against BEAMnrc. Using BEAMnrc with UBS and nbr_split=100, I found that when I used a photon weight cutoff of 0.01 (i.e. eliminating phat photons), the energy spectrum with Russian Roulette on was up to ~40% lower than that with Russian Roulette off for energies < 40 keV.
The text was updated successfully, but these errors were encountered:
@blakewalters just to remind us that egs_brachy (egs_brachy.cpp: lines 1895 to 1902) and the egs_radiative_splitting AO (egs_radiative_splitting.h: lines 159 to 166) account for this. However, these implementations check wheter the particle is fat and hence it shouldn't creare any problem. Although if EGSnrc takes care of this, it might be less risky to have those external implementations removed.
Describe the bug
Currently, when uniform bremsstrahlung splitting (UBS) is turned on with Russian Roulette, we do not split fluorescent photons resulting from atomic relaxations induced by electron impact ionization (EII) during Moller scattering of phat electrons. This results in the emission of phat photons.
To Reproduce
Expected behavior
Scored photons should all have weight 1/nbr_split. Scored charged particles should all have weight 1.
Operating system
EGSnrc version
develop branch
Other
This showed up when testing the egs++ DBS algorithm against BEAMnrc. Using BEAMnrc with UBS and nbr_split=100, I found that when I used a photon weight cutoff of 0.01 (i.e. eliminating phat photons), the energy spectrum with Russian Roulette on was up to ~40% lower than that with Russian Roulette off for energies < 40 keV.
The text was updated successfully, but these errors were encountered: