Shielding benchmarks analysis for transport/storage casks

The dose rate measurements of the TN 12/2, TN 28 VT and FS 65 has been used to evaluate the calculational procedures of Transnuclaire (TN). The three-dimensional (3D) Monte Carlo code TRIPOLI-3.4 which is used to optimize the shielding of TN casks, is applied to the analysis of a series of benchmarks. In the same cases the one-dimensional (1D) Sn code SN1D and the point kernel code MERCURE-V (3D) used for the more simplified calculations, are checked by the comparison with the measurements. The multi-group approximation used by the above codes, in order to reduces nuclear data, introduces errors due to the neutron cross-sections resonance treatment and the repartition of the gamma-ray spectrum (discrete) into an energy group structure. For a cask consisting of an iron shell of 250 mm of thickness, neutron dose rates can been underestimated of 50% if the resonances of the iron cross sections for high energy (above 1 MeV) are not taken into account. Also, depending on the energy group structure, gamma-ray dose rates can be over-estimated or under-estimated by the repartition of the gamma rays. The comparisons between measured and calculated dose rates are closer than 20% for the Monte Carlo calculations, 50% for the Sn calculations (1D) and a factor of 2 for the point kernel calculations. (author)