Density limit disruption prediction using a long short-term memory network on EAST

Kai ZHANG (张凱); Dalong CHEN (陈大龙); Bihao GUO (郭笔豪); Junjie CHEN (陈俊杰); Bingjia XIAO (肖炳甲)

doi:10.1088/2058-6272/abb28f

Plasma Science and Technology > 2020 > 22(11): 115602. > DOI: 10.1088/2058-6272/abb28f

Kai ZHANG (张凱), Dalong CHEN (陈大龙), Bihao GUO (郭笔豪), Junjie CHEN (陈俊杰), Bingjia XIAO (肖炳甲). Density limit disruption prediction using a long short-term memory network on EAST[J]. Plasma Science and Technology, 2020, 22(11): 115602. DOI: 10.1088/2058-6272/abb28f

Citation:

PDF (763 KB)

Density limit disruption prediction using a long short-term memory network on EAST

¹ Department of Engineering and Applied Physics, School of Physical Sciences, University of Science and Technology of China, Hefei 230026, People's Republic of China
² Institute of Plasma Physics, Chinese Academy of Sciences, Hefei 230031, People's Republic of China

Funds: The work is supported by the National Magnetic Confinement Fusion Energy R&D Program of China (2018YFE0304100 and 2018YFE0302100), Anhui Provincial Natural Science Foundation (1808085MA25).

More Information

Received Date: April 07, 2020
Revised Date: August 22, 2020
Accepted Date: August 24, 2020

Graphical Abstract

Abstract

Abstract

Disruption prediction using a long short-term memory (LSTM) algorithm has been developed on EAST, due to its inherent advantages in time series data processing. In the present work, LSTM is used as the model and the AUC (area under receiver operation characteristic curve) is used as the evaluation index. When the model is trained on data from the plasma current flattop phase and tested on data from the same period multiple times, the highest AUC is 0.8646 and the training time is about 6900 s per epoch. For comparison, the last 1000 ms of the flattop phases are intercepted as short time sequences. When the model is trained on data from short time sequences and tested on data from the same period, the highest AUC is increased to 0.9379 and the training time is restricted to 36 s per epoch. When the best model trained on the short time sequences is applied to the flattop phase for testing, the AUC is up to 0.9189. The experiment results show that it is possible for LSTM to train the model on data from short time sequences and migrate the model to the entire flattop phase, with a shorter training time and higher AUC value.
- disruption prediction,
- tokamak,
- neural networks,
- flattop phase

FullText(HTML)

References (38)

References

[1]	Cannas B et al 2007 Nucl. Fusion 47 1559
[2]	Hender T C et al 2007 Nucl. Fusion 47 S128
[3]	Chen D L et al 2018 Nucl. Fusion 58 036003
[4]	Taylor P L et al 1999 Phys. Plasmas 6 1872
[5]	Zheng W et al 2018 Nucl. Fusion 58 056016
[6]	Yokoyama T et al 2019 Fusion Eng. Des. 140 67
[7]	Zongyu Yang et al 2020 Nucl. Fusion 60 016017
[8]	Cannas B et al 2004 Neural approaches to disruption prediction at JET Proc. 31st EPS Conf. on Plasma Physics (London) 1.167
[9]	Cannas B et al 2011 Fusion Eng. Des. 86 1039
[10]	Yoshino R 2003 Nucl. Fusion 43 1771
[11]	Rattá G A et al 2010 Nucl. Fusion 50 025005
[12]	Rattá G A, Vega J and Murari A 2012 Fusion Eng. Des.87 1670
[13]	Vega J et al 2013 Fusion Eng. Des. 88 1228
[14]	Yoshino R 2005 Nucl. Fusion 45 1232
[15]	Rea C and Granetz R S 2018 Fusion Sci. Technol. 74 89
[16]	Rea C et al 2018 Plasma Phys. Control. Fusion 60 084004
[17]	Rea C et al 2019 Nucl. Fusion 59 096016
[18]	Montes K J et al 2019 Nucl. Fusion 59 096015
[19]	Cannas B A et al 2013 Plasma Phys. Control. Fusion 55 045006
[20]	Cannas B et al 2014 Plasma Phys. Control. Fusion 56 114005
[21]	Cannas B et al 2013 Nucl. Fusion 53 093023
[22]	Kates-Harbeck J, Svyatkovskiy A and Tang W 2019 Nature 568 526
[23]	Ferreira D R 2018 arXiv:1811.00333
[24]	Lei T et al 2018 Simple recurrent units for highly parallelizable recurrence Proc. 2018 Conf. on Empirical Methods in Natural Language Processing (Brussels, Belgium) (Association for Computational Linguistics) 4470 (arXiv:1709.02755)
[25]	Kohonen T 1990 Proc. IEEE 78 1464
[26]	Hochreiter S and Schmidhuber J 1997 Neural Comput. 9 1735
[27]	Bengio Y, Simard P and Frasconi P 1994 IEEE Trans. Neural Netw. 5 157
[28]	Eck D and Schmidhuber J 2002 Learning the long-term structure of the blues Proc. Int. Conf. on Artificial Neural Networks (Madrid, Spain) (Berlin: Springer) p 284
[29]	Liu Y D et al 2017 Short-term travel time prediction by deep learning: a comparison of different LSTM-DNN models Proc.2017 IEEE 20th Int. Conf. on Intelligent Transportation Systems (Yokohama, Japan) (Piscataway, NJ: IEEE) 1
[30]	Wang B et al 2016 Plasma Sci. Technol. 18 1162
[31]	Pau A et al 2018 IEEE Trans. Plasma Sci. 46 2691
[32]	Macrander A T 2008 Rev. Sci. Instrum. 79 10F701
[33]	Cannas B et al 2004 Nucl. Fusion 44 68
[34]	Kingma D P and Ba J 2014 arXiv:1412.6980 Kingma D P and Ba J 2014 Adam: A Method for Stochastic Optimization (arXiv:1412.6980)
[35]	Fluss R, Faraggi D and Reiser B 2005 Biom. J. 47 458
[36]	Bradley A P 1997 Pattern Recognit. 30 1145
[37]	Pereira A et al 2015 Fusion Eng. Des. 96–97 907
[38]	Vega J et al 2020 Nucl. Fusion 60 026001

[1]	Jianqing CAI, Yunfeng LIANG, Zhongyong CHEN, Wei ZHENG, Di HU, Lei XUE, Zhifang LIN, Xiang GU, the EHL-2 Team. Disruption prediction and mitigation strategies in the EHL-2 spherical torus[J]. Plasma Science and Technology, 2025, 27(2): 024013. DOI: 10.1088/2058-6272/adb40b
[2]	Yumin WANG, Qifeng XIE, Renyi TAO, Hui ZHANG, Xiaokun BO, Tiantian SUN, Xiuchun LUN, Lin CHEN, Weiqiang TAN, Dong GUO, Bihe DENG, Minsheng LIU, the EXL-50 Team. Design of the electron cyclotron emission diagnostic on EXL-50 spherical torus[J]. Plasma Science and Technology, 2024, 26(3): 034008. DOI: 10.1088/2058-6272/ad0d54
[3]	Mingxiang HUANG, Zhengkang REN, Feiyue MAO, Zhoujun YANG, Yuan GAO, Zhichao ZHANG, Shunfan HE, Guoliang LI, Jinrong FAN, Wei TIAN, Nengchao WANG, Zhipeng CHEN, Yonghua DING, Yuan PAN, Zhongyong CHEN. Behavior of multiple modes before and during minor disruption with the external resonant magnetic perturbations on J-TEXT tokamak[J]. Plasma Science and Technology, 2022, 24(6): 064013. DOI: 10.1088/2058-6272/ac6da9
[4]	Bin CHEN, Yubao ZHU, Qing ZHOU, Jiangbo DING, Xianming SONG, Shaodong SONG, Yuanming YANG, Xin ZHAO, Enwu YANG, Minsheng LIU, the EXL-50 Team. Microwave preionization and electron cyclotron resonance plasma current startup in the EXL-50 spherical tokamak[J]. Plasma Science and Technology, 2022, 24(1): 015104. DOI: 10.1088/2058-6272/ac3640
[5]	Kai ZHANG (张凱), Dalong CHEN (陈大龙), Bihao GUO (郭笔豪), Junjie CHEN (陈俊杰), Bingjia XIAO (肖炳甲). Density limit disruption prediction using a long short-term memory network on EAST[J]. Plasma Science and Technology, 2020, 22(11): 115602. DOI: 10.1088/2058-6272/abb28f
[6]	Yonghua DING (丁永华), Zhongyong CHEN (陈忠勇), Zhipeng CHEN (陈志鹏), Zhoujun YANG (杨州军), Nengchao WANG (王能超), Qiming HU (胡启明), Bo RAO (饶波), Jie CHEN (陈杰), Zhifeng CHENG (程芝峰), Li GAO (高丽), Zhonghe JIANG (江中和), Lu WANG (王璐), Zhijiang WANG (王之江), Xiaoqing ZHANG (张晓卿), Wei ZHENG (郑玮), Ming ZHANG (张明), Ge ZHUANG (庄革), Qingquan YU (虞清泉), Yunfeng LIANG (梁云峰), Kexun YU (于克训), Xiwei HU (胡希伟), Yuan PAN (潘垣), Kenneth William GENTLE, the J-TEXT Team. Overview of the J-TEXT progress on RMP and disruption physics[J]. Plasma Science and Technology, 2018, 20(12): 125101. DOI: 10.1088/2058-6272/aadcfd
[7]	Jinxia ZHU (竹锦霞), Yipo ZHANG (张轶泼), Yunbo DONG (董云波), HL-A Team. Characterization of plasma current quench during disruptions at HL-2A[J]. Plasma Science and Technology, 2017, 19(5): 55101-055101. DOI: 10.1088/2058-6272/aa5ff2
[8]	WANG Bo (王勃), Robert GRANETZ, XIAO Bingjia (肖炳甲), LI Jiangang (李建刚), YANG Fei (杨飞), LI Junjun (李君君), CHEN Dalong (陈大龙). Establishment and Assessment of Plasma Disruption and Warning Databases from EAST[J]. Plasma Science and Technology, 2016, 18(12): 1162-1168. DOI: 10.1088/1009-0630/18/12/04
[9]	DING Yonghua (丁永华), JIN Xuesong (金雪松), CHEN Zhenzhen (陈真真), ZHUANG Ge (庄革). Neural Network Prediction of Disruptions Caused by Locked Modes on J-TEXT Tokamak[J]. Plasma Science and Technology, 2013, 15(11): 1154-1159. DOI: 10.1088/1009-0630/15/11/14
[10]	ZHUANG Huidong (庄会东), ZHANG Xiaodong (张晓东). Development of a Fast Valve for Disruption Mitigation and its Preliminary Application to EAST and HT-7[J]. Plasma Science and Technology, 2013, 15(8): 745-749. DOI: 10.1088/1009-0630/15/8/05

Cited By

Cited by

Periodical cited type(7)

1.	Farahani, F.A., Depla, D. Phase composition of sputter deposited tungsten thin films. Surface and Coatings Technology, 2024. DOI:10.1016/j.surfcoat.2024.131447
2.	Shi, Q., Kajita, S., Ohno, N. Tungsten fuzz growth at low temperatures (<900 K) on a surface with nanocones. Nuclear Materials and Energy, 2024. DOI:10.1016/j.nme.2024.101668
3.	Kajita, S., Yagi, T., Ohno, N. Thermal Conductivity Measurement of Fuzzy W Using a Thermoreflectance Method. IEEE Transactions on Plasma Science, 2024, 52(9): 4365-4370. DOI:10.1109/TPS.2023.3348465
4.	Feng, S., Kajita, S., Yasuhara, R. et al. Formation of corrugated nano-fuzz tungsten thin film on silicon via helium plasma irradiation. Japanese Journal of Applied Physics, 2024, 63(1): 010904. DOI:10.35848/1347-4065/ad12ef
5.	Feng, S., Natsume, H., Kajita, S. et al. Fabrication of tungsten-based optical diffuser using fiberform nanostructure via efficient plasma route. Optics Express, 2023, 31(16): 25438-25445. DOI:10.1364/OE.493993
6.	Zhou, J., Lu, F., Cong, J. et al. Effect of plasma jet on electrochemical properties of silk fibroin hydrogel doped with PEDOT:PSS. Energy Science and Engineering, 2023, 11(7): 2521-2534. DOI:10.1002/ese3.1471
7.	Kajita, S., Yagi, T., Ohno, N. Thermal conductivity measurement of fuzzy W using a thermoreflectance method. Proceedings - International Symposium on Discharges and Electrical Insulation in Vacuum, ISDEIV, 2023. DOI:10.23919/ISDEIV55268.2023.10200254

Other cited types(0)

Get Citation

PDF

XML

Article views (170) PDF downloads (100) Cited by(7)

Turn off MathJax

Article Contents

Abstract

References

Density limit disruption prediction using a long short-term memory network on EAST