Estimating the Mass of UO3 Powder in a Container Using MCNP-PoliMi Simulations of NMIS Measurements

Grogan, Brandon R

Oak Ridge National Laboratory

关键词: Algorithms;    Geometry;    Fast Neutrons;    Neutrons;    Targets;   

DOI  :  10.2172/931145
RP-ID  :  ORNL/TM-2008/020
RP-ID  :  DE-AC05-00OR22725
RP-ID  :  931145

美国|英语

来源: UNT Digital Library

【 摘 要 】

The Nuclear Materials Identification System (NMIS) uses fast neutrons to conduct non-intrusive scans of both fissile and non-fissile materials. By time-tagging the source neutrons, the NMIS processor can measure the time of flight of these neutrons through the target of the scan and into detectors on the opposite side. With a monoenergetic neutron source, such as a DT generator, these time-of-flight measurements can approximately determine the number of neutrons that travel through the target uncollided. By comparing the results of a scan on an empty container and one made on a container with unknown materials inside, the NMIS processor can subtract the container and determine the configuration of materials inside. Using multiple small detectors, and many horizontal and vertical samples, the NMIS system can produce a picture of the object being scanned, known as a neutron radiograph. With this radiograph, it is thought that other characteristics of the scanned material can be inferred. In this thesis, the material being scanned consists of highly enriched uranium trioxide (UO{sub 3}) powder. Using the results of the time-of-flight measurements from two reference scans, a container with an unknown configuration was scanned and the mass of the UO{sub 3} powder inside was estimated. The MCNP-PoliMi computer code was used to simulate these scans. The process was applied to three different unknown geometries of powder, and the scan of each geometry was simulated using three different measurement times, resulting in a total of nine simulated scans. All of the simulated scans produced a mass estimate that is within 3% of the true mass inside of the container. An analysis of the uncertainty in these measurements determined that the error in the estimates is much larger than the uncertainty in most cases, indicating that the algorithm used to determine the mass produces some systematic error in the results. For powder configurations which are fairly close to those modeled here, this systematic error should be small enough to neglect, but for substantially different models, a correction to the algorithm may be required.