Monte Carlo and deterministic calculation of the Bell and Glasstone spatial correction factor Original Research Article

The Bell and Glasstone correction factor is used in subcritical assembly pulsed neutron source experiments to correct the spatial dependency of the measured reactivity on the detector position. The correction factor is defined as the ratio between the reactivity obtained by computer codes in criticality mode and that obtained by computer codes in source mode. In the area method, the reactivity (in dollar units) of a subcritical assembly is given by the ratio between the prompt and the delayed areas; these areas are obtained by integrating the detector reaction rate over the pulse period. This work illustrates different methods to calculate the Bell and Glasstone spatial correction factor using both Monte Carlo (MCNPX) and deterministic (PARTISN) computer codes. The different calculation methods include: (1) the one-simulation dynamic method (which has been applied by MCNPX computer simulations); (2) the two-simulation static method (which has been applied by both MCNPX and PARTISN computer simulations); (3) the one-simulation static method (which has been applied by MCNPX computer simulations). In the one-simulation dynamic method: (1) the external neutron source is time dependent; (2) the detector reaction rate is obtained from a single pulse and it is superimposed until the delayed neutron contribution reaches the asymptotic value; (3) the prompt area is obtained as the difference between the total and delayed areas. In the two-simulation static method: (1) the external neutron source is time independent; (2) the total and prompt areas are obtained by two separate computer simulations (one with and the other without delayed neutrons); (3) the delayed area is obtained as the difference between the total and prompt areas. In the one-simulation static method, first introduced in this study, the prompt and delayed areas are tallied in the same MCNPX simulation, which halves the computing time and reduces the statistical error relative to the two-simulation static method.