Collision based k-eff estimate in Shift

[Kaspar Klenø]

Is it possible to calculate this with Shift?

The track length estimator in Shift gives me twice the sigma as I get with the collision estimate in Serpent, when running the same number of histories on the same model.

The absorption based estimator in KENOVI gives a sigma in between those two.

Best, Kaspar

[Rike Bostelmann]

Hi Kaspar,

The difference in statistical uncertainties of keff between Shift and KENO (and other Monte Carlo codes) is due to the use of different keff estimators. 

Shift in SCALE 6.3.3 only uses a path length estimator. For SCALE 7.0, we already added collision and absorption estimators, including simple and covariance estimated combinations of the 3 individual estimators. The results of these estimators compare well with, for example, KENO and MCNP. The keff results of the path length estimator usually yields larger statistical uncertainties compared to the other two estimators, so that's why the SCALE 6.3.3 Shift keff results show larger uncertainties.

This is the list of keff results reported with Shift in SCALE 7.0 beta:

• keff --> Soft Link {keff_cov-ptl+col+abs}
• keff-abs
• keff-col  
• keff-col+abs 
• keff-ptl  
• keff-ptl+abs  
• keff-ptl+col 
• keff-ptl+col+abs         
• keff_cov-col+abs         
• keff_cov-ptl+abs 
• keff_cov-ptl+col
• keff_cov-ptl+col+abs 
• keff_cycle-abs
• keff_cycle-col 
• keff_cycle-ptl

Best regards,

Rike

SCALE Team