Nuclear Engineering and Technology

Korean Nuclear Society (한국원자력학회)
권호 표지
DOI QR코드

DOI QR Code

Tokamak plasma disruption precursor onset time study based on semi-supervised anomaly detection

  • X.K. Ai (International Joint Research Laboratory of Magnetic Confinement Fusion and Plasma Physics, State Key Laboratory of Advanced Electromagnetic Engineering and Technology, School of Electrical and Electronic Engineering, Huazhong University of Science and Technology) ;
  • W. Zheng (International Joint Research Laboratory of Magnetic Confinement Fusion and Plasma Physics, State Key Laboratory of Advanced Electromagnetic Engineering and Technology, School of Electrical and Electronic Engineering, Huazhong University of Science and Technology) ;
  • M. Zhang (International Joint Research Laboratory of Magnetic Confinement Fusion and Plasma Physics, State Key Laboratory of Advanced Electromagnetic Engineering and Technology, School of Electrical and Electronic Engineering, Huazhong University of Science and Technology) ;
  • D.L. Chen (Institute of Plasma Physics, Chinese Academy of Sciences) ;
  • C.S. Shen (International Joint Research Laboratory of Magnetic Confinement Fusion and Plasma Physics, State Key Laboratory of Advanced Electromagnetic Engineering and Technology, School of Electrical and Electronic Engineering, Huazhong University of Science and Technology) ;
  • B.H. Guo (Institute of Plasma Physics, Chinese Academy of Sciences) ;
  • B.J. Xiao (Institute of Plasma Physics, Chinese Academy of Sciences) ;
  • Y. Zhong (International Joint Research Laboratory of Magnetic Confinement Fusion and Plasma Physics, State Key Laboratory of Advanced Electromagnetic Engineering and Technology, School of Electrical and Electronic Engineering, Huazhong University of Science and Technology) ;
  • N.C. Wang (International Joint Research Laboratory of Magnetic Confinement Fusion and Plasma Physics, State Key Laboratory of Advanced Electromagnetic Engineering and Technology, School of Electrical and Electronic Engineering, Huazhong University of Science and Technology) ;
  • Z.J. Yang (International Joint Research Laboratory of Magnetic Confinement Fusion and Plasma Physics, State Key Laboratory of Advanced Electromagnetic Engineering and Technology, School of Electrical and Electronic Engineering, Huazhong University of Science and Technology) ;
  • Z.P. Chen (International Joint Research Laboratory of Magnetic Confinement Fusion and Plasma Physics, State Key Laboratory of Advanced Electromagnetic Engineering and Technology, School of Electrical and Electronic Engineering, Huazhong University of Science and Technology) ;
  • Z.Y. Chen (International Joint Research Laboratory of Magnetic Confinement Fusion and Plasma Physics, State Key Laboratory of Advanced Electromagnetic Engineering and Technology, School of Electrical and Electronic Engineering, Huazhong University of Science and Technology) ;
  • Y.H. Ding (International Joint Research Laboratory of Magnetic Confinement Fusion and Plasma Physics, State Key Laboratory of Advanced Electromagnetic Engineering and Technology, School of Electrical and Electronic Engineering, Huazhong University of Science and Technology) ;
  • Y. Pan (International Joint Research Laboratory of Magnetic Confinement Fusion and Plasma Physics, State Key Laboratory of Advanced Electromagnetic Engineering and Technology, School of Electrical and Electronic Engineering, Huazhong University of Science and Technology)
  • Received : 2023.07.07
  • Accepted : 2023.12.01
  • Published : 2024.04.25
Download PDF
⟨ Previous Next ⟩

Abstract

Plasma disruption in tokamak experiments is a challenging issue that causes damage to the device. Reliable prediction methods are needed, but the lack of full understanding of plasma disruption limits the effectiveness of physics-driven methods. Data-driven methods based on supervised learning are commonly used, and they rely on labelled training data. However, manual labelling of disruption precursors is a time-consuming and challenging task, as some precursors are difficult to accurately identify. The mainstream labelling methods assume that the precursor onset occurs at a fixed time before disruption, which leads to mislabeled samples and suboptimal prediction performance. In this paper, we present disruption prediction methods based on anomaly detection to address these issues, demonstrating good prediction performance on J-TEXT and EAST. By evaluating precursor onset times using different anomaly detection algorithms, it is found that labelling methods can be improved since the onset times of different shots are not necessarily the same. The study optimizes precursor labelling using the onset times inferred by the anomaly detection predictor and test the optimized labels on supervised learning disruption predictors. The results on J-TEXT and EAST show that the models trained on the optimized labels outperform those trained on fixed onset time labels.

Keywords

Acknowledgement

The authors are very grateful for the help of J-TEXT team. This work is supported by the National MCF Energy R&D Program of China under Grant No.2019YFE03010004 and by the National Natural Science Foundation of China under Grant No.51821005.

References

  1. J. Zheng, et al., Recent progress in Chinese fusion research based on superconducting tokamak configuration, Innovation 3 (4) (2022) 13. 
  2. F.C. Schuller, et al., Disruptions in tokamaks, Plasma Phys. Contr. Fusion 37 (11A) (1995) A135-A162. 
  3. V. Riccardo, et al., Disruptions and disruption mitigation, Plasma Phys. Contr. Fusion 45 (12A) (2003) A269-A284. 
  4. P.C. De Vries, et al., Survey of disruption causes at JET, Nucl. Fusion 51 (5) (2011) 12. 
  5. G. Pautasso, et al., The ITER disruption mitigation trigger: developing its preliminary design, Nucl. Fusion 58 (3) (2018), 036011. 
  6. G. Sias, et al., A locked mode indicator for disruption prediction on JET and ASDEX upgrade, Fusion Eng. Des. 138 (2018) 254-266. 
  7. B. Cannas, et al., Support vector machines for disruption prediction and novelty detection at JET, Fusion Eng. Des. 82 (5-14) (2006) 1124-1130. 
  8. J.M. Lopez, et al., Implementation of the disruption predictor APODIS in JET's real-time network using the MARTe framework, IEEE Trans. Nucl. Sci. 61 (2) (2014) 741-744. 
  9. R.A. Tinguely, et al., An application of survival analysis to disruption prediction via Random Forests, Plasma Phys. Contr. Fusion 61 (9) (2019), 095009. 
  10. W. Hu, et al., Real-time prediction of high density EAST disruptions using Random Forest, Nucl. Fusion 60 (6) (2021) 066034. 
  11. D. Wroblewski, et al., Tokamak disruption alarm based on a neural network model of the high- beta limit, Nucl. Fusion 37 (6) (1997) 725. 
  12. A. Sengupta, et al., Prediction of density limit disruption boundaries from diagnostic signals using neural networks, Nucl. Fusion 41 (5) (2001) 487-501. 
  13. C.G. Windsor, et al., A cross-tokamak neural network disruption predictor for the JET and ASDEX upgrade tokamaks, Nucl. Fusion 45 (5) (2005) 337-350. 
  14. Z. Yang, et al., A disruption predictor based on a 1.5-dimensional convolutional neural network in HL-2A, Nucl. Fusion 60 (1) (2019), 016017. 
  15. K.J. Montes, et al., Machine learning for disruption warnings on Alcator C-Mod, DIII-D, and EAST, Nucl. Fusion 59 (9) (2019), 096015. 
  16. B.H. Guo, et al., Disruption prediction on EAST tokamak using a deep learning algorithm, Plasma Phys. Contr. Fusion 63 (11) (2021) 115007. 
  17. Y. Zhong, et al., Disruption prediction and model analysis using LightGBM on JTEXT and HL-2A, Plasma Phys. Contr. Fusion 63 (7) (2021), 075008, 11pp. 
  18. Z. Yang, et al., In-depth research on the interpretable disruption predictor in HL2A, Nucl. Fusion 61 (12) (2021), 126042, 7pp. 
  19. B. Cannas, et al., A prediction tool for real-time application in the disruption protection system at JET, Nucl. Fusion 47 (11) (2007) 1559. 
  20. B. Cannas, et al., Disruption forecasting at JET using neural networks, Nucl. Fusion 44 (1) (2003) 68-76. 
  21. J. Vega, et al., Adaptive high learning rate probabilistic disruption predictors from scratch for the next generation of tokamaks, Nucl. Fusion 54 (12) (2014), 123001. 
  22. C. Rea, et al., Exploratory machine learning studies for disruption prediction using large databases on DIII-D, Fusion Sci. Technol. 74 (1-2) (2018) 89-100. 
  23. C. Rea, et al., Disruption prediction investigations using Machine Learning tools on DIII-D and Alcator C-Mod, Plasma Phys. Contr. Fusion 60 (8) (2018), 084004. 
  24. G. Pautasso, et al., On-line prediction and mitigation of disruptions in ASDEX Upgrade, Nucl. Fusion 42 (1) (2002) 100. 
  25. B. Cannas, et al., An adaptive real-time disruption predictor for ASDEX upgrade, Nucl. Fusion 50 (7) (2010), 075004. 
  26. Y. Zhang, et al., Prediction of disruptions on ASDEX Upgrade using discriminant analysis, Nucl. Fusion 51 (6) (2011), 063039. 
  27. S.P. Gerhardt, et al., Detection of disruptions in the high-β spherical torus NSTX, Nucl. Fusion 53 (6) (2013), 063021. 
  28. R. Yoshino, Neural-net disruption predictor in JT-60U, Nucl. Fusion 43 (12) (2003) 1771-1786. 
  29. R. Yoshino, Neural-net predictor for beta limit disruptions in JT-60U, Nucl. Fusion 45 (11) (2005) 1232-1246. 
  30. J. Jang, et al., Feature concentration for supervised and semi-supervised learning with unbalanced datasets in visual inspection, IEEE Trans. Ind. Electron. 68 (8) (2020) 7620-7630. 
  31. M.H. Nguyen, Impacts of unbalanced test data on the evaluation of classification methods, Int. J. Adv. Comput. Sci. Appl. 10 (3) (2019) 497-502. 
  32. B. Cannas, et al., An adaptive real-time disruption predictor for ASDEX Upgrade, Nucl. Fusion 50 (7) (2010), 075004. 
  33. J. Zhu, et al., Hybrid deep-learning architecture for general disruption prediction across multiple tokamaks, Nucl. Fusion 61 (2) (2021), 026007. 
  34. A. Pau, et al., A machine learning approach based on generative topographic mapping for disruption prevention and avoidance at JET, Nucl. Fusion 59 (10) (2019), 106017. 
  35. Y. Xiong, et al., Recognizing multivariate geochemical anomalies for mineral exploration by combining deep learning and one-class support vector machine, Comput. Geosci. 140 (2020), 104484. 
  36. W. Zheng, et al., Disruption predictor based on neural network and anomaly detection on J-TEXT, Plasma Phys. Contr. Fusion 62 (4) (2020), 045012. 
  37. S. Esquembri, et al., Real-time Implementation in JET of the SPAD disruption predictor using MARTe, IEEE Trans. Nucl. Sci. 65 (2) (2018) 836-842. 
  38. R. Aledda, et al., Multivariate statistical models for disruption prediction at ASDEX upgrade, Fusion Eng. Des. 88 (6-8) (2013) 1297-1301. 
  39. I. Goodfellow, et al., Deep learning, Cambridge:MIT press 1 (2016) 499-507 (Chapter 14). 
  40. G.E. Hinton, Autoencoders, minimum description length and Helmholtz free energy, Advances in Neural Information Processing Systems San Mateo 6 (1994). 
  41. Scholkopf, et al., Estimating the Support of a High-Dimensional Distribution, Neural Computation 13 (7) (2001) 1443-1471. 
  42. M.Y. Su, Real-time anomaly detection systems for Denial-of-Service attacks by weighted k-nearest-neighbor classifiers, Expert Syst. Appl. 38 (4) (2011) 3492-3498. 
  43. S. Esquembri, et al., Real-time implementation in JET of the SPAD disruption predictor using MARTe, IEEE Trans. Nucl. Sci. 65 (2) (2018) 836-842. 
  44. H.P. Kriegel, et al., Angle-based Outlier Detection in High-Dimensional Data//Proceedings of the 14th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, ACM, Las Vegas, Nevada, USA, 2008. August 24-27. 
  45. C.S. Shen, et al., IDP-PGFE: an interpretable disruption predictor based on physics-guided feature extraction, Nucl. Fusion 63 (4) (2023) 046024. 
  46. Y.H. Ding, et al., Overview of the J-TEXT progress on RMP and disruption physics, Plasma Sci. Technol. 20 (12) (2018) 125101. 
  47. Y. Liang, et al., Overview of the recent experimental research on the J-TEXT tokamak, Nucl. Fusion 59 (11) (2019) 112016. 
  48. D.L. Chen, et al., Characterization of disruption halo current between 'w-like' graphite divertor and "ITER-like" divertor structure on east tokamak, Plasma Phys. Contr. Fusion 62 (9) (2020), 095019. 
  49. X. Gao, L. Zeng, M.Q. Wu, et al., Experimental progress of hybrid operational scenario on EAST tokamak, Nucl. Fusion 60 (10) (2020) 102001.