Fast modeling of turbulent transport in fusion plasmas using neural networks
Top Cited Papers
Open Access
- 1 February 2020
- journal article
- research article
- Published by AIP Publishing in Physics of Plasmas
- Vol. 27 (2), 022310
- https://doi.org/10.1063/1.5134126
Abstract
We present an ultrafast neural network model, QLKNN, which predicts core tokamak transport heat and particle fluxes. QLKNN is a surrogate model based on a database of 3x10(8) flux calculations of the quasilinear gyrokinetic transport model, QuaLiKiz. The database covers a wide range of realistic tokamak core parameters. Physical features such as the existence of a critical gradient for the onset of turbulent transport were integrated into the neural network training methodology. We have coupled QLKNN to the tokamak modeling framework JINTRAC and rapid control-oriented tokamak transport solver RAPTOR. The coupled frameworks are demonstrated and validated through application to three JET shots covering a representative spread of H-mode operating space, predicting the turbulent transport of energy and particles in the plasma core. JINTRAC-QLKNN and RAPTOR-QLKNN are able to accurately reproduce JINTRAC-QuaLiKiz Ti,e and n(e) profiles, but 3-5 orders of magnitude faster. Simulations which take hours are reduced down to only a few tens of seconds. The discrepancy in the final source-driven predicted profiles between QLKNN and QuaLiKiz is on the order of 1%-15%. Also the dynamic behavior was well captured by QLKNN, with differences of only 4%-10% compared to JINTRAC-QuaLiKiz observed at mid-radius, for a study of density buildup following the L-H transition. Deployment of neural network surrogate models in multi-physics integrated tokamak modeling is a promising route toward enabling accurate and fast tokamak scenario optimization, uncertainty quantification, and control applications.Funding Information
- H2020 Euratom (633053)
This publication has 27 references indexed in Scilit:
- Angular momentum transport modeling: achievements of a gyrokinetic quasi-linear approachPlasma Physics and Controlled Fusion, 2013
- Quasilinear transport modelling at low magnetic shearPhysics of Plasmas, 2012
- Validating a quasi-linear transport model versus nonlinear simulationsNuclear Fusion, 2009
- Chapter 2: Plasma confinement and transportNuclear Fusion, 2007
- A theory-based transport model with comprehensive physicsPhysics of Plasmas, 2007
- The tokamak Monte Carlo fast ion module NUBEAM in the National Transport Code Collaboration libraryComputer Physics Communications, 2004
- Turbulent electron thermal transport in tokamaksNew Journal of Physics, 2003
- Massively parallel Vlasov simulation of electromagnetic drift-wave turbulenceComputer Physics Communications, 2000
- Approximation by superpositions of a sigmoidal functionMathematics of Control, Signals, and Systems, 1989
- New techniques for calculating heat and particle source rates due to neutral beam injection in axisymmetric tokamaksJournal of Computational Physics, 1981