Multimodal convolutional neural networks for predicting evolution of gyrokinetic simulations

Multimodal convolutional neural networks for predicting evolution of gyrokinetic simulations
复制标题

用于预测回旋模拟演化的多模态卷积神经网络

DOI:
10.1002/ctpp.202200137
复制
发表时间:
2023
影响因子:
1.6
通讯作者:
Watanabe Tomo‐Hiko
Watanabe Tomo‐Hiko
中科院分区:
物理与天体物理3区
文献类型:
--
作者:
Honda Mitsuru;Narita Emi;Maeyama Shinya;Watanabe Tomo‐Hiko

文献摘要

参考文献

被引文献

相似文献

由于湍流在托卡马克中比其他输运机制更重要,因此需要对湍流通量进行定量计算。然而,非线性陀螺动力学模拟在计算上是昂贵的。开发了一个多模态卷积神经网络模型,该模型可以读取非线性陀螺动力学模拟产生的图像和值,并预测静电湍流热通量,以支持高效运行。将该模型扩展为考虑与准线性模型中通量成正比的静电势波动的平方,以及包含波动电子和离子分布函数以及波数空间中波动静电势的图像。该多模态模型可以预测输入数据对应的时间和电子、离子湍流热通量。在Cyclone基本情况数据上训练的模型不仅成功地预测了其测试数据的时间和通量,而且还以高精度预测了完全不同的未知JT - 60U情况。模型的预测性能取决于用于训练模型的情况与被预测情况的线性稳定性的相似性。
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.
DOI: 10.1088/1741-4326/ac4557
发表时间: 2021
期刊: Nuclear Fusion
影响因子: 3.3
作者:
Shaokang Xu;S. Maeyama;T. Watanabe
通讯作者: T. Watanabe
ASDEX Upgrade H 模式等离子体中核心湍流特征的回旋运动研究
DOI: 10.1063/1.4919022
发表时间: 2015
期刊: Physics of Plasmas
影响因子: 2.2
作者:
A. Navarro;T. Happel;T. Goerler;F. Jenko;J. Abiteboul;A. Bustos;H. Doerk;D. Told;The Asdex Upgrade Team
通讯作者: The Asdex Upgrade Team
探索高电子温度燃烧等离子体中的多尺度湍流相互作用
DOI: --
发表时间: 2021
期刊:
影响因子: --
作者:
Zalo?nik An?e;Doerner Russell P.;Schwarz-Selinger Thomas;Miyamoto Mitsutaka;Brezinsek Sebastijan;S. Maeyama
通讯作者: S. Maeyama
在识别波数空间图像的卷积神经网络模型的辅助下,实现非线性回旋运动模拟的高效运行
DOI: 10.1088/1741-4326/ac70e8
发表时间: 2022
期刊: Nuclear Fusion
影响因子: 3.3
作者:
E. Narita;M. Honda;S. Maeyama;T.-H. Watanabe
通讯作者: T.-H. Watanabe
JET 混合 H 模式场景中 ETG 湍流的集成建模和多尺度回旋验证研究
DOI: 10.1088/1741-4326/ac7535
发表时间: 2022
期刊: Nuclear Fusion
影响因子: 3.3
作者:
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.
通讯作者: Watanabe T.