skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Application of process monitoring to anomaly detection in nuclear material processing systems via system-centric event interpretation of data from multiple sensors of varying reliability

Abstract

In this paper, we apply an advanced safeguards approach and associated methods for process monitoring to a hypothetical nuclear material processing system. The assessment regarding the state of the processing facility is conducted at a systemcentric level formulated in a hybrid framework. This utilizes architecture for integrating both time- and event-driven data and analysis for decision making. While the time-driven layers of the proposed architecture encompass more traditional process monitoring methods based on time series data and analysis, the event-driven layers encompass operation monitoring methods based on discrete event data and analysis. By integrating process- and operation-related information and methodologies within a unified framework, the task of anomaly detection is greatly improved. This is because decision-making can benefit from not only known time-series relationships among measured signals but also from known event sequence relationships among generated events. This available knowledge at both time series and discrete event layers can then be effectively used to synthesize observation solutions that optimally balance sensor and data processing requirements. The application of the proposed approach is then implemented on an illustrative monitored system based on pyroprocessing and results are discussed.

Authors:
; ; ; ;
Publication Date:
Research Org.:
Idaho National Lab. (INL), Idaho Falls, ID (United States)
Sponsoring Org.:
USDOE National Nuclear Security Administration (NNSA)
OSTI Identifier:
1373093
Report Number(s):
INL/JOU-16-40316
Journal ID: ISSN 0306-4549; PII: S0306454916309380
DOE Contract Number:
DE-AC07-05ID14517
Resource Type:
Journal Article
Resource Relation:
Journal Name: Annals of Nuclear Energy (Oxford); Journal Volume: 103; Journal Issue: C
Country of Publication:
United States
Language:
English
Subject:
98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL PROTECTION; advanced safeguards; integrated time- and event-driven analysis; nuclear fuel cycles; process monitoring; system-centric anomaly detection

Citation Formats

Garcia, Humberto E., Simpson, Michael F., Lin, Wen-Chiao, Carlson, Reed B., and Yoo, Tae-Sic. Application of process monitoring to anomaly detection in nuclear material processing systems via system-centric event interpretation of data from multiple sensors of varying reliability. United States: N. p., 2017. Web. doi:10.1016/j.anucene.2017.01.006.
Garcia, Humberto E., Simpson, Michael F., Lin, Wen-Chiao, Carlson, Reed B., & Yoo, Tae-Sic. Application of process monitoring to anomaly detection in nuclear material processing systems via system-centric event interpretation of data from multiple sensors of varying reliability. United States. doi:10.1016/j.anucene.2017.01.006.
Garcia, Humberto E., Simpson, Michael F., Lin, Wen-Chiao, Carlson, Reed B., and Yoo, Tae-Sic. Tue . "Application of process monitoring to anomaly detection in nuclear material processing systems via system-centric event interpretation of data from multiple sensors of varying reliability". United States. doi:10.1016/j.anucene.2017.01.006.
@article{osti_1373093,
title = {Application of process monitoring to anomaly detection in nuclear material processing systems via system-centric event interpretation of data from multiple sensors of varying reliability},
author = {Garcia, Humberto E. and Simpson, Michael F. and Lin, Wen-Chiao and Carlson, Reed B. and Yoo, Tae-Sic},
abstractNote = {In this paper, we apply an advanced safeguards approach and associated methods for process monitoring to a hypothetical nuclear material processing system. The assessment regarding the state of the processing facility is conducted at a systemcentric level formulated in a hybrid framework. This utilizes architecture for integrating both time- and event-driven data and analysis for decision making. While the time-driven layers of the proposed architecture encompass more traditional process monitoring methods based on time series data and analysis, the event-driven layers encompass operation monitoring methods based on discrete event data and analysis. By integrating process- and operation-related information and methodologies within a unified framework, the task of anomaly detection is greatly improved. This is because decision-making can benefit from not only known time-series relationships among measured signals but also from known event sequence relationships among generated events. This available knowledge at both time series and discrete event layers can then be effectively used to synthesize observation solutions that optimally balance sensor and data processing requirements. The application of the proposed approach is then implemented on an illustrative monitored system based on pyroprocessing and results are discussed.},
doi = {10.1016/j.anucene.2017.01.006},
journal = {Annals of Nuclear Energy (Oxford)},
number = C,
volume = 103,
place = {United States},
year = {Tue Jan 24 00:00:00 EST 2017},
month = {Tue Jan 24 00:00:00 EST 2017}
}
  • This work presents a novel modeling and analysis framework for graph sequences which addresses the challenge of detecting and contextualizing anomalies in labelled, streaming graph data. We introduce a generalization of the BTER model of Seshadhri et al. by adding flexibility to community structure, and use this model to perform multi-scale graph anomaly detection. Specifically, probability models describing coarse subgraphs are built by aggregating node probabilities, and these related hierarchical models simultaneously detect deviations from expectation. This technique provides insight into a graph's structure and internal context that may shed light on a detected event. Additionally, this multi-scale analysis facilitatesmore » intuitive visualizations by allowing users to narrow focus from an anomalous graph to particular subgraphs or nodes causing the anomaly. For evaluation, two hierarchical anomaly detectors are tested against a baseline Gaussian method on a series of sampled graphs. We demonstrate that our graph statistics-based approach outperforms both a distribution-based detector and the baseline in a labeled setting with community structure, and it accurately detects anomalies in synthetic and real-world datasets at the node, subgraph, and graph levels. Furthermore, to illustrate the accessibility of information made possible via this technique, the anomaly detector and an associated interactive visualization tool are tested on NCAA football data, where teams and conferences that moved within the league are identified with perfect recall, and precision greater than 0.786.« less
  • The aim of nuclear safeguards is to ensure that special nuclear material is used for peaceful purposes. Historically, nuclear material accounting (NMA) has provided the quantitative basis for monitoring for nuclear material loss or diversion, and process monitoring (PM) data is collected by the operator to monitor the process. PM data typically support NMA in various ways, often by providing a basis to estimate some of the in-process nuclear material inventory. We develop options for combining PM residuals and NMA residuals (residual = measurement - prediction), using a hybrid of period-driven and data-driven hypothesis testing. The modified statistical tests canmore » be used on time series of NMA residuals (the NMA residual is the familiar material balance), or on a combination of PM and NMA residuals. The PM residuals can be generated on a fixed time schedule or as events occur.« less
  • The development of in situ sensors for ultratrace detection applications in process control and environmental monitoring remains a significant challenge. Such sensors must meet difficult detection limit requirements while selectively detecting the analyte of interest in complex or otherwise challenging sample matrixes. Nowhere are these requirements more daunting than in the field of radionuclide sensing. The detection limit requirements can be extremely low. Nevertheless, a promising approach to radionuclide sensing based on preconcentrating minicolumn sensors has been developed. In addition, a method of operating such sensors, which we call equilibration-based sensing, has been developed that provides substantial preconcentration and amore » signal that is proportional to analyte concentration, while eliminating the need for reagents to regenerate the sorbent medium following each measurement. While this equilibration-based sensing method was developed for radionuclide sensing, it can be applied to nonradioactive species as well, given a suitable on-column detection system. By replacing costly sampling and laboratory analysis procedures, in situ sensors could have a significant impact on monitoring and long term stewardship applications. The aim of this review is to cover radionuclide sensors that combine some form of selective sorption with a radiometric detection method, and­as a primary aim­to comprehensively review preconcentrating minicolumn sensors for radionuclide detection. As a secondary aim, we will cover radionuclide sensors that combine sorption and scintillation in formats other than minicolumn sensors. We are particularly concerned with the detection of alpha- and beta-emitting radionuclides, which present particular challenges for measurements in liquid media.« less
  • An automated fluidic instrument is described that rapidly determines the total 99Tc content of aged nuclear waste samples, where the matrix is chemically and radiologically complex and the existing speciation of the 99Tc is variable. The monitor links microwave-assisted sample preparation with an automated anion exchange column separation and detection using a flow-through solid scintillator detector. The sample preparation steps acidify the sample, decompose organics, and convert all Tc species to the pertechnetate anion. The column-based anion exchange procedure separates the pertechnetate from the complex sample matrix, so that radiometric detection can provide accurate measurement of 99Tc. We developed amore » preprogrammed spike addition procedure to automatically determine matrix-matched calibration. The overall measurement efficiency that is determined simultaneously provides a self-diagnostic parameter for the radiochemical separation and overall instrument function. Continuous, automated operation was demonstrated over the course of 54 h, which resulted in the analysis of 215 samples plus 54 hly spike-addition samples, with consistent overall measurement efficiency for the operation of the monitor. A sample can be processed and measured automatically in just 12.5 min with a detection limit of 23.5 Bq/mL of 99Tc in low activity waste (0.495 mL sample volume), with better than 10% RSD precision at concentrations above the quantification limit. This rapid automated analysis method was developed to support nuclear waste processing operations planned for the Hanford nuclear site.« less