Comparison of MCNP5 results with experimental measurements for neutron-shielding materials - 016

In the 1980's and 1990's Ueki and his colleagues performed a series of experiments designed to systematically evaluate the effectiveness of various shielding materials. Those experiments measured the dose rate from a ²⁵²Cf source at the outer surface of various thicknesses of those shielding materials. The shielding materials included in the experiments discussed herein were Resin-F, NS-4-FR, Krafton-HB, and Type 304 Stainless Steel, in the form of slabs that were 5 or 5.3 cm thick. The source was placed in a paraffin howitzer, and the detector was placed 115 cm from the source. The slabs were positioned so that the back edge of the shielding material was 100 cm from the source, and the number of the slabs was increased in the direction of the source up to a maximum thickness of 30 or 31.8 cm. MCNP calculations were performed for each thickness of each shielding material, using 10,000 to 15,000,000 histories per case. Nuclear data for interactions with neutrons, photons, and electrons were taken from the ENDF66, MCPLIB04, and EL03 libraries, respectively. The dose rates were obtained by applying flux-to-dose conversion factors based on values from the ANSI standard given in the MCNP5 manual to a volumetric-flux tally. Initially, no variance-reduction techniques were applied, and calculations for all materials and shielding configurations required a few weeks of CPU time to reach 1% relative error]. Recent work applied both ring detectors to take advantage of the rotational symmetry and optimized weight-windows as variance reduction techniques reduced the required computational time required to achieve 1% statistical precision to approximately two hours. The results from the MCNP calculations are presented as the ratio of the calculated results to the corresponding measured values. The standard deviations from the MCNP5 calculations are less than 5% for each case, and so are the C/E ratios for almost all of them. Overall, the MCNP5 results match the measured values quite well. Furthermore, with the exception of Resin-F, there is no systematic trend in the ratio of calculated to measured results