Izvestiya vuzov. Yadernaya Energetika

The peer-reviewed scientific and technology journal. ISSN: 0204-3327

Detection of anomalies in reactor equipment operation using neural network algorithms

9/16/2020 2020 - #03 Modelling processes at nuclear facilities

Trykov E.L. Trykova I.V. Kotsoev K.I.

DOI: https://doi.org/10.26583/npe.2020.3.14

UDC: 621.039

The problem of detecting anomalies by algorithms based on machine learning methods, in particular, on neural network methods, has recently been very relevant in many industries, including nuclear energy.

The existing system of technical diagnostics of NPP equipment consists in the additional use of rapid diagnostics systems during operation and non-destructive testing tools during planned preventive repairs (PPR).

In this regard, the introduction of predictive analytical systems for the in-depth processing of process control data for the early detection of equipment malfunctions, as well as for the analysis of its resource characteristics, becomes extremely urgent.

The main task of predictive analytics is to build and optimize a digital model to search for anomalies in equipment operation, determine the time interval for equipment trouble-free operation and adjust the scope of maintenance and repair.

The paper presents an algorithm for detecting anomalies in equipment operation, based on the use of neural networks. The efficiency of the developed algorithm was confirmed by the operation of the MCPs at Unit 6 of the Novovoronezh NPP.

The developed algorithm demonstrates high sensitivity to changes in MCP operation modes and makes it possible to control their operation both online and offline.

Implementation of the developed methodology is possible within the framework of the complex diagnostics system (CDS) supplied by JSC «STC Diaprom» to various nuclear power plants.

References

  1. Predictive Analytics and Diagnostics of Nuclear Power Plants. Library of Technical Diagnostics of Nuclear Power Plants. Ed. by V. I. Pavelko. Moscow. JSC «NTCD» Publ., 2019, 69 p. (in Russian)
  2. Kingma D.P., Welling M. Auto-Encoding Variational Bayes. arXiv preprint. 2014, arXiv:1312.6114v10, 14 p.
  3. Goodfellow I., Bengio Y., Courville A. Deep Learning. MIT Press, 2016. 787 p.
  4. Yao K., Cohn T., Vylomova K., Duh K., Dyer C. Depth-Gated LSTM. arXiv preprint. 2015, arXiv:1508.03790v4, 6 p.
  5. Hochreiter S., Schmidhuber J. Long Short-Term Memory. Neural Computation. 1997, v. 9, pp. 1735-1780; DOI: 10.1162/neco.1997.9.8.1735 .
  6. Graves A., Schmidhuber J. Framewise Phoneme Classification with Bidirectional LSTM and Other Neural Network Architectures. Neural Networks. 2005, v. 18, pp. 602-610; DOI: 10.1016/j.neunet.2005.06.042 .
  7. Luong M., Pham H., Manning C. Effective approaches to attention-based neural machine translation. arXiv preprint. 2015, arXiv:1508.04025, 11 p.
  8. Bahdanau D., Cho K. Bengio Y. Neural machine translation by jointly learning to align and translate. arXiv preprint. 2014, arXiv:1508.04025v7, 15 p.
  9. Pereira J., Silveira M. Unsupervised anomaly detection in energy time series data using variational recurrent autoencoders with attention. 2018 XVII-th IEEE International Conference on Machine Learning and Applications (ICMLA). 2018, 8 p.; DOI: 10.1109/ ICMLA.2018.00207 .
  10. Bahuleyan H., Mou L., Vechtomova O., Poupart P. Variational attention for sequence-to-sequence models. arXiv preprint. 2018, arXiv:1712.08207v3, 11 p.
  11. Polykovsky D.A. Attention Mechanisms in Neural Networks. Final Qualifying Work. Moscow. MGU im. M.V. Lomonosova Publ., 2017, 22 p. (in Russian).
  12. Vaswani A., Shazeer N., Parmar N., Uszkoreit J., Jones L., Gomez A.N., Kaiser L. Attention is all you need. arXiv preprint. 2017, arXiv:1706.03762v5, 11 p.
  13. Luong M. Neural Machine Translation. Thesis PHD. December 2016. 156 p.
  14. Pereira J. Unsupervised Anomaly Detection in Time Series Data using Deep Learning. Thesis to obtain the Master of Science Degree in Electrical and Computer Engineering. Tecnico Lisboa Publ., 2018, 67 p.
  15. Rumelhar D.E. Hinton G.E., Williams R.J. Learning Representations by Backpropagating Errors. Nature. 1986, v. 323, pp. 533-536; DOI: 10.1038/323533a0.
  16. An J., Cho S. Variational Autoencoder based Anomaly Detection using Reconstruction Probability. arXiv preprint. 2015, arXiv:1802.03903v1, 12 p.
  17. Zhang C., Chen Y. Time Series Anomaly Detection with Variational Autoencoders. arXiv preprint. 2019, arXiv:1907.01702v1, 7 p.
  18. Mujica L.E., Rodellar J., Fernandez A., Guemes A. Q-statistic and T2-statistic PCA-based measures for damage assessment in structures. Structural Health Monitoring. 2010, v. 10, no. 5, pp. 539-553.
  19. Jackson J., Mudholkar S. Control Procedures for Residuals Associated With Principal Component Analysis. Technometrics. 1979, v. 21, no. 3, pp. 341-349.
  20. Runger G. Alt, F., Montgomery D. Contributors to a Multivariate Statistical Process Control Chart Signal. Communications in Statistics. Theory and Methods. 1996, v. 25, iss. 10, pp. 2203-2213.

neural networks time series anomalies in the operation of reactor equipment the system of technical diagnostics of NPP equipment