中国物理B ›› 2023, Vol. 32 ›› Issue (7): 75202-075202.doi: 10.1088/1674-1056/accb44

所属专题： SPECIAL TOPIC — Plasma disruption

• • 上一篇    下一篇

Recent progress on deep learning-based disruption prediction algorithm in HL-2A tokamak

Zongyu Yang(杨宗谕)1,†, Yuhang Liu(刘宇航)1, Xiaobo Zhu(朱晓博)1, Zhengwei Chen(陈正威)1, Fan Xia(夏凡)1, Wulyu Zhong(钟武律)1,‡, Zhe Gao(高喆)2, Yipo Zhang(张轶泼)1, and Yi Liu(刘仪)1   

  1. 1 Southwestern Institute of Physics, Chengdu 610043, China;
    2 Tsinghua University, Beijing 100084, China
  • 收稿日期:2023-01-18 修回日期:2023-04-03 接受日期:2023-04-07 出版日期:2023-06-15 发布日期:2023-06-21
  • 通讯作者: Zongyu Yang, Wulyu Zhong E-mail:author:zy-yang@swip.ac.cn;zhongwl@swip.ac.cn
  • 基金资助:
    Project supported by the National MCF R&D Program of China (Grant Nos. 2018YFE0302100 and 2019YFE03010003). The authors wish to thank all the members at South Western Institute of Physics for providing data, technique assistance and co-operating during the experiment.

Recent progress on deep learning-based disruption prediction algorithm in HL-2A tokamak

Zongyu Yang(杨宗谕)1,†, Yuhang Liu(刘宇航)1, Xiaobo Zhu(朱晓博)1, Zhengwei Chen(陈正威)1, Fan Xia(夏凡)1, Wulyu Zhong(钟武律)1,‡, Zhe Gao(高喆)2, Yipo Zhang(张轶泼)1, and Yi Liu(刘仪)1   

  1. 1 Southwestern Institute of Physics, Chengdu 610043, China;
    2 Tsinghua University, Beijing 100084, China
  • Received:2023-01-18 Revised:2023-04-03 Accepted:2023-04-07 Online:2023-06-15 Published:2023-06-21
  • Contact: Zongyu Yang, Wulyu Zhong E-mail:author:zy-yang@swip.ac.cn;zhongwl@swip.ac.cn
  • Supported by:
    Project supported by the National MCF R&D Program of China (Grant Nos. 2018YFE0302100 and 2019YFE03010003). The authors wish to thank all the members at South Western Institute of Physics for providing data, technique assistance and co-operating during the experiment.

RichHTML

7

PDF (PC)

196

赞

摘要： Disruption prediction and mitigation is a crucial topic, especially for future large-scale tokamaks, due to disruption's concomitant harmful effects on the devices. On this topic, disruption prediction algorithm takes the responsibility to give accurate trigger signal in advance of disruptions, therefore the disruption mitigation system can effectively alleviate the harmful effects. In the past 5 years, a deep learning-based algorithm is developed in HL-2A tokamak. It reaches a true positive rate of 92.2%, a false positive rate of 2.5% and a total accuracy of 96.1%. Further research is implemented on the basis of this algorithm to solve three key problems, i.e., the algorithm's interpretability, real-time capability and transferability. For the interpretability, HL-2A's algorithm gives saliency maps indicating the correlation between the algorithm's input and output by perturbation analysis. The distribution of correlations shows good coherence with the disruption causes. For the transferability, a preliminary disruption predictor is successfully developed in HL-2M, a newly built tokamak in China. Although only 44 shots are used as the training set of this algorithm, it gives reasonable outputs with the help of data from HL-2A and J-TEXT. For the real-time capacity, the algorithm is accelerated to deal with an input slice within 0.3 ms with the help of some adjustments on it and TFLite framework. It is also implemented into the plasma control system and gets an accuracy of 89.0% during online test. This paper gives a global perspective on these results and discusses the possible pathways to make HL-2A's algorithm a more comprehensive solution for future tokamaks.

关键词: macroinstabilities, tokamaks, neural networks, magnetic confinement and equilibrium

Abstract: Disruption prediction and mitigation is a crucial topic, especially for future large-scale tokamaks, due to disruption's concomitant harmful effects on the devices. On this topic, disruption prediction algorithm takes the responsibility to give accurate trigger signal in advance of disruptions, therefore the disruption mitigation system can effectively alleviate the harmful effects. In the past 5 years, a deep learning-based algorithm is developed in HL-2A tokamak. It reaches a true positive rate of 92.2%, a false positive rate of 2.5% and a total accuracy of 96.1%. Further research is implemented on the basis of this algorithm to solve three key problems, i.e., the algorithm's interpretability, real-time capability and transferability. For the interpretability, HL-2A's algorithm gives saliency maps indicating the correlation between the algorithm's input and output by perturbation analysis. The distribution of correlations shows good coherence with the disruption causes. For the transferability, a preliminary disruption predictor is successfully developed in HL-2M, a newly built tokamak in China. Although only 44 shots are used as the training set of this algorithm, it gives reasonable outputs with the help of data from HL-2A and J-TEXT. For the real-time capacity, the algorithm is accelerated to deal with an input slice within 0.3 ms with the help of some adjustments on it and TFLite framework. It is also implemented into the plasma control system and gets an accuracy of 89.0% during online test. This paper gives a global perspective on these results and discusses the possible pathways to make HL-2A's algorithm a more comprehensive solution for future tokamaks.

Key words: macroinstabilities, tokamaks, neural networks, magnetic confinement and equilibrium

中图分类号:  (Macroinstabilities (hydromagnetic, e.g., kink, fire-hose, mirror, ballooning, tearing, trapped-particle, flute, Rayleigh-Taylor, etc.))

  • 52.35.Py
52.55.Fa (Tokamaks, spherical tokamaks) 07.05.Mh (Neural networks, fuzzy logic, artificial intelligence) 52.55.-s (Magnetic confinement and equilibrium)