Multimodal convolutional neural networks for predicting evolution of gyrokinetic simulations

被引:0
作者
Honda, Mitsuru [1 ]
Narita, Emi [2 ]
Maeyama, Shinya [3 ]
Watanabe, Tomo-Hiko [3 ]
机构
[1] Kyoto Univ, Grad Sch Engn, Kyoto 6158530, Japan
[2] Natl Inst Quantum Sci & Technol, Naka Fus Inst, Ibaraki, Japan
[3] Nagoya Univ, Dept Phys, Nagoya, Aichi, Japan
基金
日本学术振兴会;
关键词
convolutional neural network; deep learning; GKV gyrokinetic simulation; multimodal model; turbulent heat flux;
D O I
10.1002/ctpp.202200137
中图分类号
O35 [流体力学]; O53 [等离子体物理学];
学科分类号
070204 ; 080103 ; 080704 ;
摘要
Gyrokinetic simulations are required for the quantitative calculation of fluxes due to turbulence, which dominates over other transport mechanisms in tokamaks. However, nonlinear gyrokinetic simulations are computationally expensive. A multimodal convolutional neural network model that reads images and values generated by nonlinear gyrokinetic simulations and predicts electrostatic turbulent heat fluxes was developed to support efficient runs. The model was extended to account for squared electrostatic potential fluctuations, which are proportional to the fluxes in the quasilinear model, as well as images containing fluctuating electron and ion distribution functions and fluctuating electrostatic potentials in wavenumber space. This multimodal model can predict the time and electron and ion turbulent heat fluxes corresponding to the input data. The model trained on the Cyclone base case data successfully predicted times and fluxes not only for its test data, but also for the completely different and unknown JT-60U case, with high accuracy. The predictive performance of the model depended on the similarity of the linear stability of the case used to train the model to the case being predicted.
引用
收藏
页数:12
相关论文
共 21 条
[1]   Direct multiscale coupling of a transport code to gyrokinetic turbulence codes [J].
Barnes, M. ;
Abel, I. G. ;
Dorland, W. ;
Goerler, T. ;
Hammett, G. W. ;
Jenko, F. .
PHYSICS OF PLASMAS, 2010, 17 (05)
[2]   A new gyrokinetic quasilinear transport model applied to particle transport in tokamak plasmas [J].
Bourdelle, C. ;
Garbet, X. ;
Imbeaux, F. ;
Casati, A. ;
Dubuit, N. ;
Guirlet, R. ;
Parisot, T. .
PHYSICS OF PLASMAS, 2007, 14 (11)
[3]   A high-accuracy Euleriangyrokinetic solver for collisional plasmas [J].
Candy, J. ;
Belli, E. A. ;
Bravenec, R. V. .
JOURNAL OF COMPUTATIONAL PHYSICS, 2016, 324 :73-93
[4]   An Eulerian gyrokinetic-Maxwell solver [J].
Candy, J ;
Waltz, RE .
JOURNAL OF COMPUTATIONAL PHYSICS, 2003, 186 (02) :545-581
[5]   Tokamak profile prediction using direct gyrokinetic and neoclassical simulation [J].
Candy, J. ;
Holland, C. ;
Waltz, R. E. ;
Fahey, M. R. ;
Belli, E. .
PHYSICS OF PLASMAS, 2009, 16 (06)
[6]   Integrated modelling and multiscale gyrokinetic validation study of ETG turbulence in a JET hybrid H-mode scenario [J].
Citrin, J. ;
Maeyama, S. ;
Angioni, C. ;
Bonanomi, N. ;
Bourdelle, C. ;
Casson, F. J. ;
Fable, E. ;
Goerler, T. ;
Mantica, P. ;
Mariani, A. ;
Sertoli, M. ;
Staebler, G. ;
Watanabe, T. .
NUCLEAR FUSION, 2022, 62 (08)
[7]   Electromagnetic turbulence suppression by energetic particle driven modes [J].
Di Siena, A. ;
Goerler, T. ;
Poli, E. ;
Navarro, A. Banon ;
Biancalani, A. ;
Jenko, F. .
NUCLEAR FUSION, 2019, 59 (12)
[8]   Comparisons and physics basis of tokamak transport models and turbulence simulations [J].
Dimits, AM ;
Bateman, G ;
Beer, MA ;
Cohen, BI ;
Dorland, W ;
Hammett, GW ;
Kim, C ;
Kinsey, JE ;
Kotschenreuther, M ;
Kritz, AH ;
Lao, LL ;
Mandrekas, J ;
Nevins, WM ;
Parker, SE ;
Redd, AJ ;
Shumaker, DE ;
Sydora, R ;
Weiland, J .
PHYSICS OF PLASMAS, 2000, 7 (03) :969-983
[9]   Persistence of Ion Temperature Gradient Turbulent Transport at Finite Normalized Pressure [J].
Ishizawa, A. ;
Urano, D. ;
Nakamura, Y. ;
Maeyama, S. ;
Watanabe, T-H .
PHYSICAL REVIEW LETTERS, 2019, 123 (02)
[10]  
Jenko, 1904, PHYS PLASMAS, V2000, P7