Self-consistent core-pedestal transport simulations with neural network accelerated models

Fusion whole device modeling simulations require comprehensive models that are simultaneously physically accurate, fast, robust, and predictive. In this paper we describe the development of two neural-network (NN) based models as a means to perform a snon-linear multivariate regression of theory-based models for the core turbulent transport fluxes, and the pedestal structure. Specifically, we find that a NN-based approach can be used to consistently reproduce the results of the TGLF and EPED1 theory-based models over a broad range of plasma regimes, and with a computational speedup of several orders of magnitudes. These models are then integrated into a predictive workflow that allows prediction with self-consistent core-pedestal coupling of the kinetic profiles within the last closed flux surface of the plasma. The NN paradigm is capable of breaking the speed-accuracy trade-off that is expected of traditional numerical physics models, and can provide the missing link towards self-consistent coupled core-pedestal whole device modeling simulations that are physically accurate and yet take only seconds to run.

[1]  David J. C. MacKay,et al.  Information Theory, Inference, and Learning Algorithms , 2004, IEEE Transactions on Information Theory.

[2]  A. D. Turnbull,et al.  Integrated modeling applications for tokamak experiments with OMFIT , 2015 .

[3]  R. E. Waltz,et al.  The first transport code simulations using the trapped gyro-Landau-fluid model , 2008 .

[4]  T. L. Rhodes,et al.  Validation studies of gyrofluid and gyrokinetic predictions of transport and turbulence stiffness using the DIII-D tokamak , 2013 .

[5]  L. Lao,et al.  Edge localized modes and the pedestal: A model based on coupled peeling–ballooning modes , 2002 .

[6]  Samantha S. Foley,et al.  Parameter Sweep and Optimization of Loosely Coupled Simulations Using the DAKOTA Toolkit , 2012, 2012 IEEE 15th International Conference on Computational Science and Engineering.

[7]  Dennis G. Whyte,et al.  Nonlinear gyrokinetic simulations of the I-mode high confinement regime and comparisons with experimenta) , 2015 .

[8]  P. B. Snyder,et al.  Exploration of the Super H-mode regime on DIII-D and potential advantages for burning plasma devices , 2016 .

[9]  R. Waltz,et al.  A gyro-Landau-fluid transport model , 1997 .

[10]  T. L. Rhodes,et al.  Progress in GYRO validation studies of DIII-D H-mode plasmas , 2012 .

[11]  H. R. Wilson,et al.  A first-principles predictive model of the pedestal height and width: development, testing and ITER optimization with the EPED model , 2011 .

[12]  R. E. Waltz,et al.  Gyro-Landau fluid equations for trapped and passing particles , 2005 .

[13]  J. Greene,et al.  Noncircular, finite aspect ratio, local equilibrium model , 1998 .

[14]  L. Lao,et al.  Variational moment solutions to the Grad–Shafranov equation , 1981 .

[15]  Arnold H. Kritz,et al.  Physics basis of Multi-Mode anomalous transport module , 2013 .

[16]  Anders Krogh,et al.  Neural Network Ensembles, Cross Validation, and Active Learning , 1994, NIPS.

[17]  C. Bourdelle,et al.  Ion temperature profile stiffness: non-linear gyrokinetic simulations and comparison with experiment , 2013, 1303.2217.

[18]  R. J. Groebner,et al.  Development and validation of a predictive model for the pedestal height , 2008 .

[19]  Samantha S. Foley,et al.  Multi-level concurrency in a framework for integrated loosely coupled plasma simulations , 2011, 2011 9th IEEE/ACS International Conference on Computer Systems and Applications (AICCSA).

[20]  Bernard Widrow,et al.  30 years of adaptive neural networks: perceptron, Madaline, and backpropagation , 1990, Proc. IEEE.

[21]  Jeff M. Candy,et al.  Multi-scale gyrokinetic simulations: Comparison with experiment and implications for predicting turbulence and transport , 2016 .

[22]  L. L. Lao,et al.  Integrated fusion simulation with self-consistent core-pedestal coupling , 2016 .

[23]  E. A. Belli,et al.  Gyrokinetic Eigenmode Analysis of High-Beta Shaped Plasmas , 2010 .

[24]  Jeff M. Candy,et al.  Tokamak profile prediction using direct gyrokinetic and neoclassical simulation , 2009 .

[25]  F. Hinton,et al.  Effect of finite aspect ratio on the neoclassical ion thermal conductivity in the banana regime , 1982 .

[26]  J. Stober,et al.  Transport properties of H-mode plasmas with dominant electron heating in comparison to dominant ion heating at ASDEX Upgrade , 2015 .

[27]  H. Wilson,et al.  Numerical studies of edge localized instabilities in tokamaks , 2002 .

[28]  Arnold H. Kritz,et al.  Predicting temperature and density profiles in tokamaks , 1998 .

[29]  J. Candy,et al.  Kinetic calculation of neoclassical transport including self-consistent electron and impurity dynamics , 2008 .

[30]  G. V. Pereverzev,et al.  Stable numeric scheme for diffusion equation with a stiff transport , 2008, Comput. Phys. Commun..

[31]  W. A. Peebles,et al.  L-mode validation studies of gyrokinetic turbulence simulations via multiscale and multifield turbulence measurements on the DIII-D tokamak , 2011 .

[32]  J. Kinsey,et al.  A theory-based transport model with comprehensive physicsa) , 2006 .

[33]  Jeff M. Candy,et al.  An Eulerian method for the solution of the multi-species drift-kinetic equation , 2009 .

[34]  Samantha S. Foley,et al.  The Design and Implementation of the SWIM Integrated Plasma Simulator , 2010, 2010 18th Euromicro Conference on Parallel, Distributed and Network-based Processing.

[35]  F. Felici,et al.  Real-time capable first principle based modelling of tokamak turbulent transport , 2015, 1502.07402.

[36]  L. L. Lao,et al.  ELMs and constraints on the H-mode pedestal: peeling–ballooning stability calculation and comparison with experiment , 2004 .

[37]  P. B. Snyder,et al.  The EPED pedestal model and edge localized mode-suppressed regimes: Studies of quiescent H-mode and development of a model for edge localized mode suppression via resonant magnetic perturbations , 2012 .

[38]  D. J. Campbell,et al.  Chapter 1: Overview and summary , 1999 .

[39]  H. Doerk,et al.  Gyrokinetic studies of core turbulence features in ASDEX Upgrade H-mode plasmas , 2015 .

[40]  F. Hinton,et al.  Theory of plasma transport in toroidal confinement systems , 1976 .

[41]  Maxim Umansky,et al.  Stability and dynamics of the edge pedestal in the low collisionality regime: physics mechanisms for steady-state ELM-free operation , 2007 .