engleski

Hybrid Shielding Methods Validation Using Graphite Shielding Measurements

This paper presents validation of the several computational hybrid shielding methods by means of Ueki experimental measurements of a graphite slab shield. The objective was to estimate neutron dose attenuation factors for different shield thicknesses at point detector located at fixed distance of 20 cm behind the graphite slab shield. Several state-of-the art shielding codes were used for that purpose. The ADVANTG3.0.1 code was used to generate variance reduction (VR) parameters for continuous energy Monte Carlo (MC) code MCNP6.1.1b. The deterministic module of ADVANTG code is Denovo, recently developed multigroup, discrete ordinates (SN), Boltzmann transport solver which approximates adjoint neutron flux in phase-space. From deterministic, adjoint-based solution, the VR parameters are automatically prepared in the form of weight-windows and biased source distribution for final MC simulation. Similar methodology can be found in SCALE6.1.3 code package, within MAVRIC shielding sequence with functional modules KENO-VI (MC geometry, material, cross sections) and Monaco (multigroup MC code). The same Denovo solver, but in a lighter form, is implemented inside the MAVRIC sequence, providing VR parameters (importance map and biased source) to Monaco MC module. With both hybrid shielding codes we have demonstrated efficient way of obtaining accelerated MC neutron dose rates and attenuation factors, based on Denovo adjoint SN solver and CADIS methodology. Accelerated MC results were compared to analog MC simulation, showing necessity for VR parameters as the shield thickness increases. Overall performance of final adjoint-driven MC process was done by analyzing so called figure-of-merit (FOM) factor. Excellent agreement between calculated and reported attenuation factors (C/E ratios) was obtained.

Slab shield, Monte Carlo, variance reduction, discrete ordinates, CADIS

nije evidentirano