Plasma radiation distribution tomography based on deconvolutional neural networks

doi:10.16568/j.0254-6086.202103003

Abstract

Abstract: A deconvolutional neural network (deCNN) which allows to implement a pixel-wise emission distribution reconstruction is developed for plasma radiation distribution tomographic reconstruction. By adopting the structural similarity (SSIM) as the loss function, the method reveals better reconstruction accuracy in phantom data experiments. The simulated experiments results show that this method could still achieve an admissible reconstruction precision and is noise robustness when the line-integrated noise magnitude is 10%, 15% and 20%.

Key words: Plasma radiation distribution tomography, Deconvolutional neural networks, Line-integrated diagnostic system, Image reconstruction

摘要： 发展了一种基于逆卷积神经网络的图像级重建方法用于聚变等离子体辐射分布的断层反演。通过引入结构相似度(SSIM)作为损失函数，该方法在模拟数据实验中表现出了较好的重建效果。模拟实验结果表明，在弦积分信号噪声强度为10%、15%及20%时，该方法的重建结果依然具有良好的精确度和鲁棒性。

关键词: 等离子体辐射分布断层重建, 逆卷积神经网络, 弦积分诊断系统, 图像重建

CLC Number:

TL65+2

WANG Zhao-hui, XIAO Jian-yuan, QIN Hong. Plasma radiation distribution tomography based on deconvolutional neural networks[J]. Nuclear Fusion and Plasma Physics, 2021, 41(3): 207-214.

王朝辉, 肖建元, 秦宏. 基于逆卷积神经网络的等离子体辐射分布断层重建方法[J]. 核聚变与等离子体物理, 2021, 41(3): 207-214.

References

[1] Ingesson L, Alper B, Peterson B, et al. Chapter 7:Tomography diagnostics: bolometry and soft-X-ray detection, fusion science and technology [M]. Taylor & Francis, 2008. 528–576.
[2] Imrisek M, Weinzettl V, Mlynar J, et al. Use of soft x-ray diagnostic on the compass tokamak for investigations of sawteeth crash neighborhood and of plasma position using fast inversion methods [J]. Rev. Sci. Instrum., 2014, 85(11): 11E433.
[3] Weinzettl V, Naydenkova D, Sestak D, et al. Design of multi-range tomographic system for transport studies in tokamak plasmas [J]. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2010, 623(2): 806–808.
[4] Craciunescu T, Bonheure G, Kiptily V, et al. A comparison of four reconstruction methods for jet neutron and gamma tomography [J]. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2009, 605(3): 374–383.
[5] Mlynar J, Craciunescu T, Ferreira D R, et al. Current research into applications of tomography for fusion diagnostics [J]. Journal of Fusion Energy, 2018, 38(3‒4): 458‒466.
[6] Mlynar J, Weinzettl V, Bonheure G, et al. Inversion techniques in the soft-X-ray tomography of fusion plasmas: toward realtime applications [J]. Fusion Science and Technology, 2010, 58(3): 733–741.
[7] Yoshino R. Neural-net disruption predictor in JT-60U [J].Nucl. Fusion, 2003, 43(12): 1771–1786.
[8] Farias G, Dormido-Canto S, Vega J, et al. Automatic feature extraction in large fusion databases by using deep learning approach [J]. Fusion Engineering and Design, 2016, 112: 979–983.
[9] Rastovic D. Simulation and control of turbulence at tokamaks with artificial intelligence methods [J]. Journal of Modern Physics, 2012, 3: 1858‒1869.
[10] Demeter G. Tomography using neural networks [J]. Rev. Sci. Instrum., 1997, 68: 1438.
[11] Clayton D J, Tritz K, Stutman D, et al. Electron temperature profile reconstructions from multi-energy sxr measurements using neural networks [J]. Plasma Physics and Controlled Fusion, 2013, 55: 095015.
[12] Ronchi E, Conroy S, Sundén E A, et al. Neural networks based neutron emissivity tomography at JET with real-time capabilities [J]. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2010, 613 (2): 295–303.
[13] Krizhevsky A, Sutskever I, Hinton G E. Imagenet classification with deep convolutional neural networks [M]. Advances in Neural Information Processing Systems 25, Curran Associates, Inc., 2012. 1097–1105.
[14] Ferreira D R, Carvalho P J, Fernandes H. Full-pulse tomographic reconstruction with deep neural networks [J].Fusion Science And Technology, 2018, 74(1‒2): 47–56. [15] Matos F A, Ferreira D R, Carvalho P J. Deep learning for plasma tomography using the bolometer system at JET [J]. Fusion Engineering and Design, 2017, 114: 18–25.
[16] Wang Z, Bovik A C, Sheikh H R, et al. Image quality assessment: from error visibility to structural similarity [J]. IEEE Transactions on Image Processing, 2004, 13(4): 600–612.
[17] Dong C, Loy C C, He K, et al. Image super-resolution using deep convolutional networks [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, 38 (2): 295–307.
[18] Jain V, Seung S. Natural image denoising with convolutional networks [M]. Advances in Neural Information Processing Systems 21, 2009. 769–776.
[19] Wang G, Ye J C, Mueller K, et al. Image reconstruction is a new frontier of machine learning [J]. IEEE Transactions on Medical Imaging, 2018, 37(6): 1289– 1296.
[20] Chen K, Xu L, Hu L, et al. 2-D soft X-ray arrays in the EAST [J]. Rev. Sci. Instrum., 2016, 87(6): 063504.
[21] Ronchi E, Conroy S, Sundén E A, et al. A parametric model for fusion neutron emissivity tomography for the KN3 neutron camera at JET [J]. Nucl. Fusion, 2010, 50(3): 035008.
[22] Kingma D, Ba J. Adam: A method for stochastic optimization [R]. International Conference on Learning Representations, 2013.