Prediction of Fishbone Linear Instability in Tokamaks with Machine Learning Methods

Z.Y. Liu,H.R. Qiu,G.Y. Fu,Y. Xiao,Y.C. Chen,Z.J. Wang,Y.X. Wei

2024-02-23

Abstract:A machine learning based surrogate model for fishbone linear instability in tokamaks is constructed. Hybrid simulations with the kinetic-magnetohydrodynamic (MHD) code M3D-K is used to generate the database of fishbone linear instability, through scanning the four key parameters which are thought to determine the fishbone physics. The four key parameters include (1) central total beta of both thermal plasma and fast ions, (2) the fast ion pressure fraction, (3) central value of safety factor $q$ and (4) the radius of $q=1$ surface. Four machine learning methods including linear regression, support vector machines (SVM) with linear kernel, SVM with nonlinear kernel and multi-layer perceptron are used to predict the fishbone instability, growth rate and real frequency, mode structure respectively. Among the four methods, SVM with nonlinear kernel performs very well to predict the linear instability with accuracy $\approx$95%, growth rate and real frequency with $R^2\approx$98%, mode structure with $R^2\approx$98%.

Plasma Physics

What problem does this paper attempt to address?

This paper aims to address the issue of predicting fishbone linear instability in tokamaks. Specifically, the authors utilize machine learning methods to construct a surrogate model for predicting fishbone linear instability, growth rate, real frequency, and mode structure in tokamaks. By scanning four key parameters (including the central total beta value, fast ion pressure ratio, central safety factor $ q $ value, and the position of the $ q=1 $ surface), a database of fishbone linear instability was generated. The study employed four machine learning methods (linear regression, support vector machine with a linear kernel, support vector machine with a nonlinear kernel, and multilayer perceptron) for prediction. It was found that the support vector machine with a nonlinear kernel performed the best in predicting linear instability, growth rate, and real frequency, with accuracies of approximately 95%, 98%, and 98%, respectively. The main contribution of this paper is that, through machine learning techniques, it is possible to efficiently and accurately predict complex physical phenomena in tokamaks. This provides a foundation for future exploration of fishbone nonlinear evolution, saturation, and fast ion transport.

Prediction of Fishbone Linear Instability in Tokamaks with Machine Learning Methods

Prediction of fishbone linear instability in tokamaks with machine learning methods

Linear Hybrid Simulations of Low Frequency Fishbone Instability Driven by Energetic Passing Particles in Tokamak Plasmas

Nonlinear Hybrid Simulation of Internal Kink with Beam Ion Effects in DIII-D

Sensitivity of Kinetic Ballooning Mode Instability to Tokamak Equilibrium Implementations

Experimental Observation and Simulation Analysis of the Relationship Between the Fishbone and ITB Formation on EAST Tokamak

Hybrid simulation of fishbone instabilities with reversed safety factor profile

Random Forest Model Driven by CLT-Simulated Data for Predicting the Growth Rate of Internal Kink Mode in Tokamak

Thermal Ion Kinetic Effects and Landau Damping in Fishbone Modes

Fishbone Instability Driven by Trapped Fast Ions in a Toroidal Plasma with Reversed Magnetic Shear

Multiple interactions between fishbone instabilities and ITBs in EAST plasmas

Multiple Interactions Between Fishbone Instabilities and Internal Transport Barriers in EAST Plasmas

Gaussian process regression models for the properties of micro-tearing modes in spherical tokamaks

Hybrid Simulations of M/n = 1/1 Instability Driven by Energetic Counter-Passing Particles in Tokamak Plasmas

A machine-learning-based tool for last closed-flux surface reconstruction on tokamaks

Saturation of Fishbone Instability by Self-Generated Zonal Flows in Tokamak Plasmas

Development of a neural network model for peeling-ballooning stability analysis in the KSTAR tokamak pedestals

Numerical simulation of magnetohydrodynamic instability with variational principle in the tokamak

Saturation of fishbone instability through zonal flows driven by energetic particle transport in tokamak plasmas

Scientific Machine Learning Based Reduced-Order Models for Plasma Turbulence Simulations