Global Sensitivity Analysis of Plasma Instabilities via Active Subspaces

TL;DR

This paper applies active subspace analysis to plasma models to identify key input parameter combinations affecting plasma instability growth rates, providing a new computational approach for global sensitivity analysis in plasma physics.

Contribution

It introduces novel computational methods to quantify uncertainty in plasma instability growth rates using active subspaces, tailored for the Vlasov-Poisson system.

Findings

Identified dominant input directions influencing plasma growth rates.

Quantified uncertainty propagation in plasma instability models.

Enhanced understanding of parameter sensitivities in collisionless plasma dynamics.

Abstract

Active subspace analysis is a useful computational tool to identify and exploit the most important linear combinations in the space of a model's input parameters. These directions depend inherently on a quantity of interest, which can be represented as a function from input parameters to model outputs. As the dynamics of many plasma models are driven by potentially uncertain parameter values, the utilization of active subspaces to perform global sensitivity analysis represents an important tool to understand how certain physical phenomena depend upon fluctuations in the values of these parameters. In the current paper, we construct and implement new computational methods to quantify the induced uncertainty within the growth rate generated by perturbations in a collisionless plasma modeled by the one-dimensional Vlasov-Poisson system near an unstable, spatially-homogeneous equilibrium in the linear regime.