Development of machine learning model for automatic ELM-burst detection without hyperparameter adjustment in KSTAR tokamak
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
In this study, a neural network model inspired by a one-dimensional convolution U-net is developed to automatically accelerate edge localized mode (ELM) detection from big diagnostic data of fusion devices and increase the detection accuracy regardless of the hyperparameter setting. This model recognizes the input signal patterns and overcomes the problems of existing detection algorithms, such as the prominence algorithm and those of differential methods with high sensitivity for the threshold and signal intensity. To train the model, 10 sets of discharge radiation data from the KSTAR are used and sliced into 11091 inputs of length 12 ms, of which 20% are used for validation. According to the receiver operating characteristic curves, our model shows a positive prediction rate and a true prediction rate of approximately 90% each, which is comparable to the best detection performance afforded by other algorithms using their optimized hyperparameters. The accurate and automatic ELM-burst detection methodology used in our model can be beneficial for determining plasma properties, such as the ELM frequency from big data measured in multiple experiments using machines from the KSTAR device and ITER. Additionally, it is applicable to feature detection in the time-series data of other engineering fields
Abstract
In this study, a neural network model inspired by a one-dimensional convolution U-net is developed to automatically accelerate edge localized mode (ELM) detection from big diagnostic data of fusion devices and increase the detection accuracy regardless of the hyperparameter setting. This model recognizes the input signal patterns and overcomes the problems of existing detection algorithms, such as the prominence algorithm and those of differential methods with high sensitivity for the threshold and signal intensity. To train the model, 10 sets of discharge radiation data from the KSTAR are used and sliced into 11091 inputs of length 12 ms, of which 20% are used for validation. According to the receiver operating characteristic curves, our model shows a positive prediction rate and a true prediction rate of approximately 90% each, which is comparable to the best detection performance afforded by other algorithms using their optimized hyperparameters. The accurate and automatic ELM-burst detection methodology used in our model can be beneficial for determining plasma properties, such as the ELM frequency from big data measured in multiple experiments using machines from the KSTAR device and ITER. Additionally, it is applicable to feature detection in the time-series data of other engineering fields
참고문헌 (36)
- H. Zohm, "The physics of edge localized modes(ELMs)and their role in power and particle exhaust", Plasma Phys. Control Fusion, 38, pp.1213~1223 (1996)
- A. W. Leonard, "Edge-localized-modes in tokamaks", Phys. Plasmas, 21, pp.090501~ (2014)
- A. Loarte, "Characteristics of type I ELM energy and particle losses in existing devices and their extrapolation to ITER", Plasma Phys. Control Fusion, 45, pp.1549~ (2003)
- W. Suttrop, "The physics of large and small edge localized modes", Plasma Phys. Control Fusion, 42, pp.A1~ (2000)
- P. B. Snyder, "ELMs and constraints on the H-mode pedestal : peelingeballooning stability calculation and comparison with experiment", Nucl. Fusion, 44, pp.320~ (2004)
- C. Ham, "Filamentary plasma eruptions and their control on the route to fusion energy", Nat. Rev. Phys, 2, pp.159~167 (2020)
- S. W. Yoon, "Characteristics of the first H-mode discharges in KSTAR", Nucl. Fusion, 51, pp.113009~ (2011)
- J. Kim, "Nonlinear energy transfer from low frequency electromagnetic fluctuations to broadband turbulence during edge localized mode crashes", Nucl. Fusion, 60, pp.124002~ (2020)
- D. Eldon, "Controlling marginally detached divertor plasmas", Nucl. Fusion, 57, pp.066039~ (2017)
- M. Berta, "Automatic ELM detection using gSPRT on the COMPASS tokamak", Fusion Eng. Des, 123, pp.950~954 (2017)
- R. Panet, "Status of the COMPASS tokamak and characterization of the first H-mode", Plasma Phys. Control Fusion, 58, pp.014015~ (2015)
- A. Murari, "Extensive statistical analysis of ELMs on JET with a carbon wall", Plasma Phys. Control. Fusion, 56, pp.114007~ (2014)
- A. Krizhevsky, "Imagenet classification with deep convolutional neural networks", Adv. Neural Inf. Process. Syst, pp.1097~1105 (2012)
- Y. LeCun, "Deep learning", Nature, 521, pp.436~444 (2015)
- G. Shin, "Real-time classification of LeH transition and ELM in KSTAR", Fusion Eng. Des, 157, pp.111634~ (2020)
- Semin Joung, "Deep neural network Grad-Shafranov solver constrained with measured magnetic signals", Nucl. Fusion, 60, pp.016034~ (2020)
- P. Ronneberger, "U-net: convolutional networks for biomedical image segmentation, In medical image computing and computer-assisted intervention (MICCAI)", Lecture Notes Comput. Sci. Springer Cham, 9351, pp.234~241 (2015)
- A. D. Melnikov, "Deep learning for the precise peak detection in high-resolution LC-MS data", Anal. Chem, 92, pp.588~592 (2020)
- O. Çiçek, "3D U-net:learning dense volumetric segmentation from sparse annotation, In medical image computing and computer-assisted intervention (MICCAI)", Lecture Notes Comput. Sci. Springer Cham, 9901, pp.424~432 (2016)
- Q. Huang, "Robust liver vessel extraction using 3D U-Net with variant dice loss function", Comput. Biol. Med, 101, pp.153~162 (2018)
- J. H. Lee, "Diagnostics for first plasma and development plan on KSTAR", Rev. Sci. Instrum, 81, pp.063502~ (2010)
- J.G. Bak, "Initial measurements by using Mirnov coils in the KSTAR machine", Proceeding of the 37th EPS Conference on Plasma Physics, pp.5102~ (2010)
- T. Kass, "Characteristics of type I and type III ELM precursors in ASDEX upgrade", Nucl. Fusion, 38, pp.111~ (1998)
- V. Pascal, "Stacked denoising autoencoders : learning useful representations in a deep network with a local denoising criterion", J. Mach. Learn. Res, 11, pp.3371~3408 (2010)
- V. Nair, "Rectified linear units improve restricted Boltzmann machines", Proceedings of the 27th International Conference on Machine Learning, pp.807~814 (2010)
- P. K. Sahoo, "A survey of thresholding techniques", Comp. Vis. Gr., Image Process, 41, pp.233~260 (1988)
- M. Llobera, "Building past landscape perception with GIS: understanding topographic prominence", J. Archaeolog. Sci, 28, pp.1005~ (2001)
- M. Basu, "Gaussian-based edge-detection methods-a survey", IEEE Trans. Syst. Man. Cybern., Part C: App. Rev, 32, pp.252~260 (2002)
- Y. Kamada, "Disappearance of giant ELMs and appearance of minute grassy ELMs in JT-60U high-triangularity discharges", Plasma Phys. Control Fusion, 42, pp.A247~ (2000)
- J. -W. Ahn, "Confinement and ELM characteristics of H-mode plasmas in KSTAR", Nucl. Fusion, 52, pp.114001~ (2012)
- N. Oyama, "Effects of edge collisionality on ELM characteristics in the grassy ELM regime", Nucl. Fusion, 50, pp.064014~ (2010)
- N. Oyama, "ELM frequency dependence on toroidal rotation in the grassy ELM regime in JT-60U", Plasma Phys. Control Fusion, 49, pp.249~ (2007)
- J. Schirmer, "The radial electric field and its associated shear in the ASDEX Upgrade tokamak", Nucl. Fusion, 46, pp.S780~ (2006)
- A. Cathey, "Non-linear extended MHD simulations of type-I edge localised mode cycles in ASDEX Upgrade and their underlying triggering mechanism", Nucl. Fusion, 60, pp.124007~ (2020)
- J. Seol, "Study of type III ELMs in the KSTAR tokamak", 25th IAEA fusion energy conference (2014), IAEA
- J. Rapp, "Integrated scenario with type-III ELMy H-mode edge : extrapolation to ITER", Nucl. Fusion, 49, pp.095012~ (2009)