Elsevier

Annals of Nuclear Energy

Volume 151, February 2021, 107947
Annals of Nuclear Energy

A radionuclide outlier identification and correlation analysis on low- and intermediate-level radioactive waste

https://doi.org/10.1016/j.anucene.2020.107947Get rights and content

Highlights

  • An outlier and correlation analysis were carried out for 32 generation facilities.

  • A outlier concentrations were determined in radioactive waste.

  • There was a significant interaction in Cs-137 and Tc-99 according to the facility.

  • Analyzing radionuclide data is an important factor for radwaste management.

Abstract

In this paper, we collected the analysis data of radionuclide concentrations and performed verification analysis on the data integrity and outliers in radioactive waste. These analysis results investigated the correlation between the radionuclide concentration data. As a result of analyzing the characteristics and outliers of the radioactive waste radionuclide data by using the collected original and repackaged drum information, the several generation facilities with high average radionuclide concentrations were identified. As a result of discriminating radioactive concentration outliers by standardizing the values for each nuclide, it was confirmed that the I-129(2.45), Mn-54(2.31) and Nb-94(2.31) nuclides had outliers. Also, it was found that there was a significant interaction of concentration in the middle class of Cs-137 and Tc-99 according to the waste generation facility and middle-class classification. We expect to be able to efficiently manage and dispose the radioactive waste inventory through analyzing the radionuclide related data.

Introduction

Recently, as the utilization of radioactive materials and radiation has increased, the amount of radioactive waste has increased. Radioactive waste is not only produced from nuclear power, but is also generated from the extensive use of isotopes in medicine, military and research. All type of radioactive waste must be managed in accordance with the regulatory requirements of the Nuclear Safety Act, and the disposal method is determined according to the concentration standards of each nuclide. In South Korea, there has been increasing concerns regarding data errors that occurred during the analysis of low- and intermediate-level radioactive waste nuclides. The cause of the data errors was found to be a human error, such as an error in analyzing a drum that is not representative in the group due to procedural problems, or a data management system operation error. To solve the problem, South Korean regulators and several civic groups are demanding the establishment of a system with a high level of reliability to safely manage data related to nuclide concentrations. The Korea Atomic Energy Research Institute (KAERI) has been conducting research into the patterns of radionuclides contained in radioactive waste to address these demands.

Radioactive waste is a material that emits radiation and must be properly stored, managed, and disposed for safety. Radioactive waste is expensive to be dispose, and it is essential to reduce the amount of waste by classifying it through middle-class classification and the concentration of nuclides to prevent damage to humans (Corkhill and Hyatt, 2018). In order to manage and dispose radioactive waste, it is currently generated and managed in a non-uniform and discontinuous manner in waste generation and treatment facilities according to business procedures. In particular, the nuclide concentration data must be managed for a long time from the generation of the radioactive waste to its delivery to the disposal site. When appropriate management and information are required, it is very important to present accurate and consistent data (Carter et al., 2018).

Radionuclide data analysis is used in the nuclear field to predict and evaluate and secure stability. Statistical analysis is frequently used to analyze nuclide data, for example, observing changes in the concentration pattern of nuclides through measuring the long-lived nuclides (Jenkins et al., 2009, Schrader, 2010), analyzing nuclide concentration through gamma-ray analysis, and using scale factor analysis (Povinec, 2019, Tompkins et al., 2004). This statistical analysis method is suitable for clustering data and classifying patterns (Okuno et al., 2002), and it is a method for determining the meaning of data and deriving unexpected patterns (Chandala et al., 2009, Hodge and Austin, 2004). In this analysis method, data mining plays the most important role, and outliers are determined to secure data reliability and usefulness (Gogoi et al., 2011, Gupta et al., 2013, Zhang et al., 2010). ‘Outlier’ is defined as the one that appears to deviate markedly from other members of the sample in which it occurs (Mokhtar et al., 2018). Specifically, outlier analysis is of great interest to the data mining field, due to the fact that it can reveal rare but important phenomenon, and find interesting or unexpected patterns (Xu et al., 2018). Researchers can obtain vital knowledge which assists in making better decisions about data by measurement error can reduce the statistical tests (Buckley, 2006), the removal and detection of anomalies or outliers (Mommaert et al., 2017, Dulanská et al., 2009) in a nuclear industry.

In this paper, we provide a brief overview of the outlier detection methods to efficiently manage data of radioactive waste by nuclide correlation analysis, which can reduce the cost of disposal and minimize radioactive waste. We examine the process level of low- and intermediate-level radioactive waste and the status of nuclide data contained in waste. We also describe the characteristics and outliers of radioactive waste nuclide data based on the data set to confirm the consistency of the nuclide data. The confidence intervals of radionuclide concentrations were calculated for each radioactive waste generating facility and for each middle-class classification. As a result, outliers were identified. It was investigated whether an exceptional event occurred through the presence or absence of an outlier. When including an outlier, the data should be reviewed for possible transcription errors, laboratory errors and possible reasons for the variation. Especially, The IAEA explains why the outlier assessment is important in the radiochemistry analysis stage during the assessment of waste characteristics in relation to the calculation of scale factors and data verification (IAEA, 2008). In order to understand the characteristics of radionuclides in radioactive waste, the correlation between factors was analyzed for factors affecting the radionuclide concentration. Through the results of this study, it was possible to confirm the justification and validity of why it is necessary to analyze the radionuclide concentration pattern contained in radioactive waste. The results could be used as an important basic data for establishing procedures and policy decisions, and this analysis is expected to be a useful tool to reduce radioactive waste reduction and disposal costs.

Section snippets

Radioactive waste life-cycle

From the generating department to the disposal site, a work process of radioactive waste was established with reference to relevant laws and guidelines. Fig. 1 shows the flow of documents based on the Business Process and shows the life cycle of radioactive waste. The work ledger, processing instructions, and work documents of the relevant business procedures were investigated, and the messages and data flow between the departments were identified and visualized.

In radioactive waste generation

Characteristics of radioactive waste radionuclide data

Through the analysis of radionuclide data, the characteristics of nuclide data were analyzed for inventory management and efficient disposal of radioactive waste. Based on the collected information of 3087 collected original drums and 2520 repackaged drums, it was confirmed that the main factors in the radioactive waste life cycle are generating facilities and middle-level classification. To test the hypothesis that two factors will affect nuclide concentrations, the characteristic values are

Conclusion

We conducted preliminary research steps to establish a radionuclide pattern and identified outliers through the characteristic values of radionuclide data. In order to eliminate outliers, confidence intervals of nuclide concentrations for each facility and waste classification were estimated based on the nuclide master data set. After selecting the 6 objects and 5 events, the correlation between the major factors was analyzed to verify the factors affecting the specific radioactive

CRediT authorship contribution statement

Sung-Chan Jang: Writing - original draft, Methodology, Visualization. Hyunjong Woo: Writing - original draft, Conceptualization, Methodology, Software. Dong-Ju Lee: Resources, Methodology, Investigation. Jeong-Guk Kim, Jongjin Kim: Writing – review & editing, Validation. Il-Sik Kang: Data curation. Jin-Woo Lee, Won Jong Park: Supervision, Software. Hee-Seoung Park: Supervision, Writing – review & editing, Project administration.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement

This research was supported by Nuclear Energy R&D Program through the National Research Foundation of Korea (NRF) funded by Ministry of Science and ICT (2019M2C9A1059067).

References (18)

  • J.H. Jenkins et al.

    Evidence of correlations between nuclear decay rates and Earth-Sun distance

    Astropart. Phys.

    (2009)
  • H. Schrader

    Half-life measurements of long-lived radionuclides—new data analysis and systematic effects

    Appl. Radiat. Isot.

    (2010)
  • Buckley, N., 2006. Clearance and Exemption Principles, Process and Practices for Use by the Nuclear Industry, A Nuclear...
  • Carter, A., Fekete, J.I., Ciambrella, M., Bienvenu, H., McMahon, K., Lambert, S., 2018. Metadata for Radioactive Waste...
  • V. Chandala et al.

    Anomaly Detection: A Survey, ACM Computing Surveys

    (2009)
  • C. Corkhill et al.

    Nuclear Waste Management

    (2018)
  • S. Dulanská et al.

    Scaling model for prediction of radionuclide activity in contaminated soils using a regression triplet technique

    J. Radioanal. Nucl. Chem.

    (2009)
  • P. Gogoi et al.

    A survey of outlier detection methods in network anomaly identification

    Comput. J.

    (2011)
  • M. Gupta et al.

    Outlier detection for temporal data: a survey

    IEEE Trans. Knowl. Data Eng.

    (2013)
There are more references available in the full text version of this article.

Cited by (0)

View full text