Monitoring Aqueous Reprocessing Systems for Detection of Facility Misuse

Abstract

Traditional approaches to material control and accounting at used nuclear fuel reprocessing facilities rely in large part on laboratory-based destructive assay of periodic samples taken from key stages. These techniques are highly accurate, but limited inspection frequency may limit the size and operation of a reprocessing facility. Online monitoring techniques using in-line detectors can be applied to fuel reprocessing systems to improve the situational awareness as the facility is operating to provide additional indicators of the evolving state of the facility between normal inspection periods, potentially alerting regulators to the need for additional inspection. The multi-isotope process (MIP) monitor combines medium-resolution gamma detectors with statistical data analysis techniques to monitor reprocessing systems for anomalous conditions resulting from process control issues or malfeasance. The MIP monitor is attractive for process and safeguards monitoring because of its unsupervised approach to detection and its inherent data obfuscation. This paper presents the application of the MIP monitor to detect a notional diversion scenario in an operating aqueous reprocessing facility by analyzing gamma spectra with a combination of principal component analysis (PCA) and the sequential probability ratio test. Gamma spectra collected from the H-Canyon reprocessing facility were analyzed to demonstrate the efficacy of the MIP Monitor for detecting notional diversion scenarios. The proposed analysis successfully detected small, protracted diversions, which would alert regulators of the need for additional inspection.