Comparison of GEANT4 and MCNPX/CINDER for use in Active Interrogation Simulation

摘要

Active interrogation works on the concept of inducing fission in Special Nuclear Materials (SNM). The signal from fission and the fission products is more difficult to shield than the passive SNM decay signal. The beam pulse makes evaluating the prompt signal difficult so the delayed signal is of considerable interest. The delayed signal out is very involved and dependent on materials, radiation type, energy, and shielding. The spectrum and time evolution of that delayed signal must be well understood to design a usable interrogation system. MCNP is a typical workhorse simulation code for fission, but to fully model the beam transport in, SNM fission, daughter product distribution, product decay, and delayed signal out requires interfacing MCNP with other codes such as CINDER2008. A standalone platform allows for calculations without having to modify parameters based on required input or output formats, greatly increasing flexibility of parameters tracked. In this work we compare MCNPX 2.6 coupled with CINDER2008 beta 5 with GEANT4.9.4 for active interrogation simulation. We have modified GEANT4 to allow for the full time evolution of daughter products from neutron-induced fission. Test calculations were run using a neutron surface source normal to a small, highly enriched uranium sphere, for experimental comparison. Fission outputs compare favorably between GEANT, MCNPX/CINDER, and experiment. Thus, GEANT appears to provide a valid, single platform environment for Monte Carlo calculations fission and delayed signals, allowing greater flexibility in modeling and simulation of active interrogation.