본문 바로가기
KERD
논문KCI 등재

Robust transformer-based anomaly detection for nuclear power data using maximum correntropy criterion

Yi Shuang|Zheng Sheng|Yang Senquan|Zhou Guangrong|He Junjie
한국원자력학회 (Korean Nuclear Society)|2024.04|Nuclear Engineering and Technology|제56권 제4호|pp. 1284-1295|DOI 10.1016/j.net.2023.11.033

2

조회

0

다운로드

0

인용

0

참고문헌

키워드

Nuclear powerTransformerAnomaly detectionMCC
주제분류원자력공학

초록

Due to increasing operational security demands, digital and intelligent condition monitoring of nuclear power plants is becoming more significant. However, establishing an accurate and effective anomaly detection model is still challenging. This is mainly because of data characteristics of nuclear power data, including the lack of clear class labels combined with frequent interference from outliers and anomalies. In this paper, we introduce a Transformer-based unsupervised model for anomaly detection of nuclear power data, a modified loss function based on the maximum correntropy criterion (MCC) is applied in the model training to improve the robustness. Experimental results on simulation datasets demonstrate that the proposed Trans-MCC model achieves equivalent or superior detection performance to the baseline models, and the use of the MCC loss function is proven can obviously alleviate the negative effect of outliers and anomalies in the training procedure, the F1 score is improved by up to 0.31 compared to Trans-MSE on a specific dataset. Further studies on genuine nuclear power data have verified the model’s capability to detect anomalies at an earlier stage, which is significant to condition monitoring.

Abstract

Due to increasing operational security demands, digital and intelligent condition monitoring of nuclear power plants is becoming more significant. However, establishing an accurate and effective anomaly detection model is still challenging. This is mainly because of data characteristics of nuclear power data, including the lack of clear class labels combined with frequent interference from outliers and anomalies. In this paper, we introduce a Transformer-based unsupervised model for anomaly detection of nuclear power data, a modified loss function based on the maximum correntropy criterion (MCC) is applied in the model training to improve the robustness. Experimental results on simulation datasets demonstrate that the proposed Trans-MCC model achieves equivalent or superior detection performance to the baseline models, and the use of the MCC loss function is proven can obviously alleviate the negative effect of outliers and anomalies in the training procedure, the F1 score is improved by up to 0.31 compared to Trans-MSE on a specific dataset. Further studies on genuine nuclear power data have verified the model’s capability to detect anomalies at an earlier stage, which is significant to condition monitoring.

참고문헌 (46)

  1. J. Ma, "Applications of fault detection and diagnosis methods in nuclear power plants : a review", Prog. Nucl. Energy, 53(3), pp.255~266 (2011)
  2. D. Bailey, "Practical SCADA for Industry" (2003), Newnes
  3. E. Gursel, "Using artificial intelligence to detect human errors in nuclear power plants : a case in operation and maintenance", Nucl. Eng. Technol, 55(2), pp.603~622 (2023)
  4. H. Kim, "and prognosis in thermal performance analysis for nuclear power plants", Nucl. Eng. Technol, 46(6), pp.737~752 (2014)
  5. M. G. Fernandez, "Status of research and development of learning-based approaches in nuclear science and engineering : a review", Nucl. Eng. Des, 359, pp.110479~ (2020)
  6. J. Contreras, "ARIMA models to predict nextday electricity prices", IEEE Trans. Power Apparatus Syst, 18(3), pp.1014~1020 (2003)
  7. B. N. Saha, "Snake validation : a PCA-based outlier detection method", IEEE Signal Process. Lett, 16(6), pp.549~552 (2009)
  8. L. -J. Cao, "KPCA and ICA for dimensionality reduction in support vector machine", Neurocomputing, 55(1–2), pp.321~336 (2003)
  9. T. Kanungo, "An efficient k-means clustering algorithm : analysis and implementation", IEEE Trans. Pattern Anal. Mach. Intell, 24(7), pp.881~892 (2002)
  10. T. Cover, "Nearest Neighbor Pattern Classification", IEEE T. Inform. Theory, IT, pp.21~27 (1967)
  11. M. M. Breunig, "LOF: identifying density-based local outliers", ACM SIGMOD Int. Conf. On Management of Data (2000)
  12. A. Banerjee, "A support vector method for anomaly detection in hyperspectral imagery", IEEE Trans. Geosci. Electron, 44(8), pp.2282~2291 (2006)
  13. K. Hundman, "Detecting spacecraft anomalies using LSTMs and nonparametric dynamic thresholding", KDD 2018 Deep Learning Day (2018)
  15. B. Zong, "Deep autoencoding Gaussian mixture model for unsupervised anomaly detection", 6th International Conference on Learning Representations (2018)
  16. J. Audibert, "USAD: unsupervised anomaly detection on multivariate time series", The 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (Virtual) (2020)
  17. Y. Su, "Robust anomaly detection for multivariate time series through stochastic recurrent neural network", 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (2019)
  18. D. Li, "MAD-GAN: multivariate anomaly detection for time series data with generative adversarial networks", 28th International Conference on Artificial Neural Networks (2019)
  19. A. Vaswani, "Attention is all you need", The 31st International Conference on Neural Information Processing Systems (2017)
  20. J.-H. Xu, "Anomaly Transformer: time series anomaly detection with association discrepancy", The 10th International Conference on Learning Representations (Virtual) (2022)
  21. X. -X. Wang, "Variational transformer-based anomaly detection approach for multivariate time series", Measurement, 191, pp.110791~ (2022)
  22. S. Tuli, "TranAD: deep transformer networks for anomaly detection in multivariate time series data", 48th International Conference on Very Large Data Bases (2022)
  23. W.-F. Liu, "Correntropy: a localized similarity measure", 2006 International Joint Conference on Neural Networks (2006)
  24. N. Khentout, "Fault supervision of nuclear research reactor systems using artificial neural networks : a review with results", Ann. Nucl. Energy, 185, pp.109684~ (2023)
  25. E. Y. Chow, "Analytical redundancy and the design of robust failure detection systems", IEEE Trans. Automat. Control, 29, pp.603~614 (1984)
  26. M. Basseville, "Detecting changes in signals and systems-A survey", Automatica, 24, pp.309~326 (1988)
  27. R. Isermann, "Process fault detection based on modeling and estimation methods-A survey", Automatica, 20, pp.387~404 (1984)
  28. R.V. Beard, "Failure Accommodation in Linear Systems through Self-Reconfiguration", pp.1971
  29. S. S. Godbole, "Applications of kalman filtering technique to nuclear reactors", IEEE Trans. Nucl. Sci, 20, pp.661~667 (1973)
  30. N. Khentout, "Fault monitoring and accommodation of the heat exchanger parameters of Triga-Mark II nuclear research reactor using model-based analytical redundancy", Prog. Nucl. Energy, 109, pp.97~112 (2018)
  31. J. W. Hines, "Lessons learned from the US nuclear power plant on-line monitoring programs", Prog. Nucl. Energy, 46(3–4), pp.176~189 (2005)
  32. B. M. Wise, "The process chemometrics approach to process monitoring and fault detection", J. Process Control, 6, pp.329~348 (1996)
  33. J. W. Hines, "Use of autoassociative neural networks for signal validation", J. Intell. Rob. Syst, 21, pp.143~154 (1998)
  34. N. Kaistha, "Incipient fault detection and isolation of field devices in nuclear power systems using principal component analysis", Nucl. Technol, 136, pp.221~230 (2001)
  35. N. Zavaljevskl, "Sensor fault detection in nuclear power plants using multivariate state estimation technique and support vector machines", The Third International Conference of the Yugoslav, Nuclear Society (2000)
  36. A. Papaoikonomou, "Deep learning techniques for incore perturbation identification and localization of time-series nuclear plant measurements", Ann. Nucl. Energy, 178, pp.109373~ (2022)
  37. S. Ryu, "Development of deep autoencoder-based anomaly detection system for HANARO", Nucl. Eng. Technol, 55(2), pp.475~483 (2023)
  38. J. C. Principe, "Information Theoretic Learning : Renyi’s Entropy and Kernel Perspectives" (2010), Springer
  39. P. P. Pokharel, "A low complexity robust detector in impulsive noise", Signal Process, 89(10), pp.1902~1909 (2009)
  40. X. -B. Chen, "Recursive robust least squares support vector regression based on maximum correntropy criterion", Neurocomputing, 97, pp.63~73 (2012)
  41. W. -C. Lu, "Short-term wind power forecasting using the hybrid model of improved variational mode decomposition and maximum mixture correntropy long short-term memory neural network", Int. J. Elec. Power, 144, pp.108552~ (2023)
  42. B.W. Silverman, "Density Estimation for Statistics and Data Analysis", pp.48~49 (1986), Chapman and Hall
  43. S. Kollias, "Machine learning for analysis of real nuclear plant data in the frequency domain", Ann. Nucl. Energy, 177, pp.109293~ (2022)
  44. X. Li, "Research on anomaly detection method of nuclear power plant operation state based on unsupervised deep generative model", Ann. Nucl. Energy, 167, pp.108785~ (2022)
  45. H.-W. Xu, "Unsupervised anomaly detection via variational auto-encoder for seasonal KPIs in web applications", The 2018World Wide Web Conference (2018)
  46. A. Siffer, "Anomaly detection in streams with extreme value theory", Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (2017)

관련 논문AI추천

Research on unsupervised condition monitoring method of pump-type machinery in nuclear power plantFu Yin, Wang Zhichao, Xia Hong, Zhang Jiyu, Zhu Yihu한국원자력학회 · Nuclear Engineering and Technology2024 · 제56권 제6호Imbalanced data fault diagnosis method for nuclear power plants based on convolutional variational autoencoding Wasserstein generative adversarial network and random forestDu Baigang, Guo Jun, Liu Shiqiao, Sun Xiang, Wang Yulong한국원자력학회 · Nuclear Engineering and Technology2024 · 제56권 제12호An intelligent fault detection method for PWR-type nuclear power plants using neuro-encoder binary cellsAli Wasiq, Arshad Furqan, Haseeb Fazle, Peng Minjun한국원자력학회 · Nuclear Engineering and Technology2025 · 제57권 제11호Event diagnosis method for a nuclear power plant using meta-learning김종현, 이대일, 이희재한국원자력학회 · Nuclear Engineering and Technology2024 · 제56권 제6호Long-term prediction of safety parameters with uncertainty estimation in emergency situations at nuclear power plants김종현, 김효진한국원자력학회 · Nuclear Engineering and Technology2023 · 제55권 제5호Using artificial intelligence to detect human errors in nuclear power plants: A case in operation and maintenanceAgarwal Vivek, Boring Ronald L., Coble Jamie Baalis, Gursel Ezgi, Khojandi Anahita, Madadi Mahboubeh, Reddy Bhavya, Yadav Vaibhav한국원자력학회 · Nuclear Engineering and Technology2023 · 제55권 제2호

이 간행물의 다른 논문 (39건)

A validation study of the SLTHEN code for hexagonal assemblies of wire-wrapped pins using liquid metal heating experiments양원식, 최선락, 예휘열, 홍종간, 한준규pp. 1125-1134Design and dynamic simulation of a molten salt THS coupled to SFRGuo Zhongxiao, Liu Yizhe, Wang Jin, Yang Jun, Nuerlan Areaipp. 1135-1144Simulations for the cesium dynamics of the RF-driven prototype ion source for CRAFT N-NBILiang Lizhen, Yang Yalong, Zhang Rui, Wu Yong, Liu Wei, Xie Yahong, Wei Jianglong, Hu Chundongpp. 1145-1152Autonomous exploration for radioactive sources localization based on radiation field reconstructionHu Li, Zhou Ying, Hu Xulin, Wang Junling, Huo Jianwen, Guo Yunleipp. 1153-1164Evaluation of reactor pulse experimentsPungerčič A., Snoj L., Trkov A., Švajger I., Čalič D.pp. 1165-1203
원문 보기