Masters Theses

Date of Award

5-1997

Degree Type

Thesis

Degree Name

Master of Science

Major

Nuclear Engineering

Major Professor

Peter G. Groer

Committee Members

Laurence F. Miller, Gloria Mei

Abstract

A nuclear accident dosimetry system based solely on thermoluminescent dosimeters (TLDs) for calculating neutron doses is rare. Most nuclear accident dosimetry systems include activation foils and pellets for characterizing neutron fields from a criticality event. The following analysis provides the first quantitative description of the dose uncertainty for a nuclear accident dosimeter using radiation exposure data and the corresponding TLD response.

TLDs were exposed to three different possible neutron accident exposure fields. Two of the data sets fit linear regression models. The third provides an example that fits a non-linear "saturation" model. Bayesian linear regression and prediction techniques have been applied to the linear data sets to enable the prediction of TLD responses based on given exposures to neutron fields and a calibration data set of given neutron exposures and measured TLD responses. This method provides a quantification of the uncertainty of the predictions in the form of a probability density. A random value was chosen from an assumed distribution about each TLD result in the calibration data set and a representative TLD result for which prediction was desired. The regression and prediction procedures were then performed and repeated several times. Mathcad software was utilized to facilitate the calculations. Distributions were obtained which indicate the effects of observed TLD results on regression parameters and prediction.

The third data set was analyzed using a non-linear model to describe the saturation of the TLD response at higher exposure levels. Using Bayes' theorem and the non-linear model for the data, equations were derived for the joint posterior density function and predictive and calibrative density functions. The resultant functions were plotted to describe the uncertainty in the parameter set as well as the predictive and calibrative values for TLD response and radiation exposure, respectively.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS