Implementing advanced data analysis techniques in near-real-time materials accounting

Materials accounting for special nuclear material in fuel cycle facilities is implemented more efficiently by applying decision analysis methods, based on estimation and detection theory, to analyze process data for missing material. These methods are incorporated in the computer program DECANAL, which calculates sufficient statistics containing all accounting information, sets decision thresholds, and compares these statistics to the thresholds in testing the hypothesis H/sub 0/ of no missing material against the alternative H/sub 1/ that material is missing. DECANAL output provides alarm charts indicating the likelihood of missing material and plots of statistics that estimate materials loss. This program is a useful tool for aggregating and testing materials accounting data for timely detection of missing material.