Modern methods of signal processing applied to gyrokinetic simulations

Abstract

Numerical simulations, like the ones necessary for e.g. electromagnetic gyrokinetic models in plasma physics, require large computational resources and long run times. Using tools from signal processing, it is possible to draw conclusions about frequencies, damping rates and mode structures using shorter runs. These tools can also be applied to analyse transient signals. We give a pedagogical review of two contemporary methods from signal processing: damped multiple signal classification and stochastic system identification. An application to simulations of Alfvén modes in a tokamak is presented.