Simulation-based plasma scenario development, optimization and control are crucial elements towards the successful deployment of next-generation experimental tokamaks and Fusion power plants. Current simulation codes require extremely intensive use of HPC resources that make them unsuitable for iterative or real time applications. Neural network based surrogate models of expensive simulators have been proposed to speed up such costly workflows. Current efforts in this direction in the Fusion community are mostly limited to point estimates of quantities of interest or simple 1D PDE models, with a few notable exceptions. While the AI literature on methods for neural PDE surrogate models is rich, performance benchmarks for Fusion-relevant 2D fields has so far remained flimited. In this work neural PDE surrogates are trained for the JOREK MHD code and the STORM scrape-off layer code using the PDEArena library (https://github.com/microsoft/pdearena). The performance of these surrogate models is investigated as a function of training set size as well as for long-term predictions. The performance of surrogate models that are trained on either one variable or multiple variables at once is also considered. It is found that surrogates that are trained on more data perform best for both long- and short-term predictions. Additionally, surrogate models trained on multiple variables achieve higher accuracy and more stable performance. Downsampling the training set in time may provide stability in the long term at the expense of the short term predictive capability, but visual inspection of the resulting fields suggests that multiple metrics should be used to evaluate performance.
Accelerating integrated modeling with surrogate-based optimization: the MAESTRO workflow
MIT Plasma Science and Fusion Center（麻省理工学院等离子体科学与聚变中心）；Universidad Politecnica de Madrid（马德里理工大学）；University of California San Diego（加州大学圣地亚哥分校）；Commonwealth Fusion Systems（联邦聚变系统）
Fast Dynamic 1D Simulation of Divertor Plasmas with Neural PDE Surrogates
Eindhoven University of Technology（埃因霍温理工大学）；École Polytechnique Fédérale de Lausanne（洛桑联邦理工学院）；Dutch Institute for Fundamental Energy Research（荷兰基础能源研究所）；Forschungszentrum Jülich GmbH（于利希研究中心）
5D Neural Surrogates for Nonlinear Gyrokinetic Simulations of Plasma Turbulence
ELLIS Unit, LIT AI Lab, Institute for Machine Learning, JKU Linz（ELLIS单元、LIT AI实验室、林茨约翰内斯·开普勒大学机器学习研究所）；UKAEA (United Kingdom Atomic Energy Authority)（英国原子能管理局）；Emmi AI GmbH（Emmi AI有限公司）
Machine-learning surrogate models for nonlinear energetic-particle transport predictions in ITER
Oak Ridge National Laboratory（橡树岭国家实验室）；Institute for Fusion Studies, The University of Texas at Austin（德克萨斯大学奥斯汀分校聚变研究所）；Universidad Carlos III de Madrid（马德里卡洛斯三世大学）
Autoregressive long-horizon prediction of plasma edge dynamics
Computational Sciences and Engineering Division, Oak Ridge National Laboratory（橡树岭国家实验室计算科学与工程部）；Fusion Energy Division, Oak Ridge National Laboratory（橡树岭国家实验室等离子体能源部）；Computer Science and Mathematics Division, Oak Ridge National Laboratory（橡树岭国家实验室计算机科学与数学部）
GyroSwin: 5D Surrogates for Gyrokinetic Plasma Turbulence Simulations
ELLIS Unit, Institute for Machine Learning, JKU Linz（JKU林茨机器学习研究所ELLIS单元）；United Kingdom Atomic Energy Authority, Culham campus（英国原子能局库勒姆校区）；EMMI AI, Linz（林茨EMMI人工智能）
LEAP: A Rapid Neural Surrogate of Multi-Fluid MHD at Europa
Jet Propulsion Laboratory, California Institute of Technology（加州理工学院喷气推进实验室）；University of Alberta（阿尔伯塔大学）；Alberta Machine Intelligence Institute（阿尔伯塔机器智能研究所）；University of Michigan（密歇根大学）；Johns Hopkins Applied Physics Laboratory（约翰斯·霍普金斯大学应用物理实验室）；University College London（伦敦大学学院）
AI Surrogate Modeling for Real-Time Tokamak Equilibrium Prediction: Benchmarking Neural Architectures and Validation on EXL-50U
Beijing ENN Fusion Energy Science and Technology Co., Ltd.（北京新奥聚变能源科技有限公司）；Xi’an University of Science and Technology（西安科技大学）；Xi’an Jiaotong University（西安交通大学）；Nankai University（南开大学）
AI-Accelerated Gyrokinetic Predictions of Turbulent Transport for Stellarator Design Optimization and Experimental Planning
Princeton Plasma Physics Laboratory（普林斯顿等离子体物理实验室）；University of Maryland, College Park（马里兰大学帕克分校）；Oak Ridge National Laboratory（橡树岭国家实验室）；Type One Energy；California Institute of Technology（加州理工学院）；Argonne National Laboratory（阿贡国家实验室）；Princeton University（普林斯顿大学）；University of Wisconsin, Madison（威斯康星大学麦迪逊分校）
DIFFER – Dutch Institute for Fundamental Energy Research（DIFFER——荷兰基础能源研究所）；Eindhoven University of Technology（埃因霍温理工大学）；MIT Plasma Science and Fusion Center（麻省理工学院等离子体科学与聚变中心）；VTT Technical Research Centre of Finland（芬兰VTT技术研究中心）；United Kingdom Atomic Energy Authority（英国原子能管理局）；Forschungszentrum Jülich GmbH（于利希研究中心有限公司）
Fast and Differentiable MHD Surrogate for Solar Wind using Neural Operator
University Corporation for Atmospheric Research（大气研究大学联盟）；Space Weather TREC, University of Colorado（科罗拉多大学空间天气TREC）；Queen Mary University of London（伦敦大学玛丽女王学院）；National Center for Atmospheric Research（国家大气研究中心）；University of Michigan（密歇根大学）
The Machine Learning Approach to Moment Closure Relations for Plasma: A Review
School of Physical and Chemical Sciences, Queen Mary University of London（伦敦大学女王学院物理与化学科学学院）；Space Weather TREC, University of Colorado（科罗拉多大学空间天气TREC）
Flux pumping and bifurcated relaxations of helical core in 3D magnetohydrodynamic modelling of ASDEX Upgrade plasmas
Max Planck Institute for Plasma Physics（马克斯·普朗克等离子体物理研究所）；Eindhoven University of Technology（埃因霍温理工大学）；Dutch Institute for Fundamental Energy Research（荷兰基础能源研究所）；Consorzio RFX (CNR, ENEA, INFN, Università di Padova, Acciaierie Venete SpA)（RFX联盟（意大利国家研究委员会、国家能源与新技术机构、国家核物理研究所、帕多瓦大学、威尼斯钢铁公司））；Istituto per la Scienza e la Tecnologia dei Plasmi, CNR（意大利国家研究委员会等离子体科学与技术研究所）；Proxima Fusion GmbH（Proxima Fusion有限公司）
