Browsing by Author "Bellman, Christopher"
Now showing 1 - 2 of 2
- Results Per Page
- Sort Options
Item A comparison of three computational phantoms for calibration of a radiation portal monitor used for measurement of internal contamination.(2014-12-01) Bellman, Christopher; Waller, EdThe goal of this research was to investigate if portal monitors could measure internally deposited radioisotopes over a short count time (1 second) with a detection limit sufficient to measure 1/10 of an annual limit on intake (ALI) of specific radionuclides. The goal was to determine if portal monitors in nuclear facilities were able to effectively screen workers for internal contamination knowing the counting efficiencies for measuring internally deposited radioisotopes is poorer than measuring the same activity of external contamination and that intakes of radioisotopes result in a committed lifetime dose that is greater than the dose expected from external contamination. This research investigated the counting efficiency of the Thermo PM12 personnel portal monitor for the measurement of internal contamination localized to the lungs. A counting efficiency curve was calculated by Monte Carlo analysis using Monte Carlo N Particle (MCNP) software. The counting efficiencies were used to calculate the minimum detectable activity for 241Am, 57Co, 60Co, 137Cs and 40K as a function of sample (personnel) measurement time. Three different computational phantoms were considered for this work: the adult male Bottle Mannequin Absorber (BOMAB) phantom, the University of Florida – Oak Ridge National Laboratories (UF-ORNL) stylized phantom, and the International Commission on Radiological Protection (ICRP) adult male voxel phantom. A percent difference ranging from 10-15% and 25-30% was observed at high energies (100-2000 keV) for the UF-ORNL and voxel computational Page iii phantoms respectively using the counting efficiency measurements calculated for the BOMAB as a baseline. At low energies (< 100 keV) the percent difference dropped 20% (e.g., 10% to -10%) within a span of 60 keV for both the UF-ORNL and ICRP computational phantoms. The ICRP voxel and the UF-ORNL computational phantoms allowed for greater accuracy for the source distribution. The BOMAB computational phantom was limited in that the source distribution was limited to one or more bottles. The BOMAB phantom was necessary to compare physical measurements to simulated measurements to assess the validity of the PM12 computational model across a range of energies. This work shows that the PM12, and other portal monitors of similar build, are able to achieve detection limits of 1/100 ALI for the beta/gamma radiation emitting radioisotopes assessed. In general, the PM12 was well suited to measure internally deposited radionuclides with gamma emissions greater than 100 keV and with an ALI greater than 105 Bq with a short measurement time. Applying this work to radioisotopes with gamma emissions below 100 keV or with low ALI values (less than 105 Bq), such as 241Am should be done with caution.Item Using consumer-grade brain-computer interface devices to capture and detect unaware facial recognitions(2017-08-01) Bellman, Christopher; Vargas Martin, MiguelThe brain's natural reaction to viewing and processing faces in an aware manner is an area of research that has been explored for previously, however the brain's unaware reactions to these stimuli prove to be fairly less explored. An experiment was performed where recruited participants viewed images of individuals' faces while their brains' electroencephalography signals were recorded using a consumer-grade BCI device. The chosen images were assigned one of three classes of recognition, corresponding with what we expect the images to be recognized as: No Recognition, Possible Unaware Recognition, and Possible Aware Recognition. Using modern filtering and analysis techniques, it was found that, in effect, using consumer-grade brain-computer interface devices, the three previously-defined classes of recognition are easily identified, both with the human eye and machine learning tools, and previous efforts to detect unaware/subconscious facial recognition have been improved on using a variety of methods for data manipulation.