Abstract :
This work focuses on the determination of predicted neutron detector response accomplished using neutron importance derived from an adjoint discrete ordinates (SN) transport calculation. A hypothetical detector apparatus, intended to detect fast neutrons, was modeled using He-3 tubes with graphite moderation using the PENTRAN™ 3-D multi-group discrete ordinates parallel transport code system. The detector geometry was modeled using z-axis symmetry and discretized into 30,280 3-D Cartesian cells. The material spatial mesh was generated using the PENMSH™ code in the PENTRAN system. The 47-group BUGLE-96 neutron cross section library was used for construction of macroscopic neutron cross sections. Results from an S8 angular quadrature using P3 anisotropy are presented. An adjoint transport source was established in the model using group dependent He-3 response cross sections. Each He-3 tube contained an adjoint source aliased to group He-3 absorption cross sections to permit assessment of detector performance. The spectrally dependent detector response from neutron capture in He-3 tubes from an arbitrary source can, therefore, be readily determined. This response comes from the complete integral of the actual source strength weighted by the adjoint function at the source location for any source distribution scenario. For selected neutron energies, an equivalent forward MCNP Monte Carlo model was used to demonstrate good agreement with the detector response determined from the adjoint calculation. The graphite used in this design has a large impact on detector performance due to the increasing sensitivity inherent in He-3 gas as neutrons thermalize. Computational adjoint results presented here predict a fast neutron detector design that yields efficiencies between 30 and 50% for neutron energies below 3 keV, and up to 30% efficiencies for neutron energies between 3 keV and 1 MeV. Overall, the methodology applied here highlights the elegant nature of an adjoint transport calculation.
