Rules-Based Approach for Unattended, Continuous Load Cell Monitoring of UF6 Feed and Withdrawal Stations.

摘要

This paper describes results from the installation of an unattended continuous load cell monitoring system designed to improve the detection of undeclared activities during feed and withdrawal (F&W) operations at an operating nuclear facility. The paper also describes laboratory work being conducted at Oak Ridge National Laboratory (ORNL) on layering unattended monitoring technologies to improve safeguards system effectiveness. The layered systems are being designed with rules-based algorithms for identifying off-normal operations that can be customized for site- and scenario-specific needs of the monitoring agreement. Continuous monitoring of load cell data at F and w stations can significantly improve the ability of inspectors to validate facility declarations of material transfers that occur between inspections. Rules-based event processing methods can use this operational data to indicate to operators when processes deviate from normal. This can facilitate the detection of problems such as human procedural errors and equipment failures. Furthermore, by integrating different monitoring technologies, advanced automated event processing can be used to screen innocuous events and facilitate nuisance event reconciliation for inspectors. ORNL has built a mock F and W system for controlled testing in a laboratory environment. This mock F and W system allows continuous material transfers from three feed stations to three product stations and two tails or waste stations via a single process tank. Each station is equipped with a dedicated load cell, from which data is collected nominally on a continuous basis. The stations also are equipped with RF antennas and readers for tracking feed, product, and waste containers fitted with radio frequency identification (RFID) tags as they enter and leave their respective stations or areas. Water is used as a surrogate for uranium fluids to provide an economical capability for testing and evaluating the integration of unattended continuous monitoring approaches (e.g., load cells, RF tracking, flow rate monitoring, and vido surveillance). T