Data-driven discovery of reduced plasma physics models from fully-kinetic simulations

TL;DR

This paper presents a data-driven method to discover reduced plasma physics models directly from particle-in-cell simulations, enabling efficient multi-scale modeling of complex plasma phenomena.

Contribution

It introduces an integral sparsity-based technique to identify PDEs governing plasma dynamics from noisy simulation data, recovering models from kinetic to fluid scales.

Findings

Successfully recovered plasma models from Vlasov to MHD

Robust identification despite particle noise

Accelerates development of reduced plasma models

Abstract

At the core of some of the most important problems in plasma physics -- from controlled nuclear fusion to the acceleration of cosmic rays -- is the challenge to describe nonlinear, multi-scale plasma dynamics. The development of reduced plasma models that balance between accuracy and complexity is critical to advancing theoretical comprehension and enabling holistic computational descriptions of these problems. Here, we report the data-driven discovery of accurate reduced plasma models, in the form of partial differential equations, directly from first-principles particle-in-cell simulations. We achieve this by using an integral formulation of sparsity-based model-discovery techniques and show that this is crucial to robustly identify the governing equations in the presence of discrete particle noise. We demonstrate the potential of this approach by recovering the fundamental hierarchy of plasma physics models -- from the Vlasov equation to magnetohydrodynamics. Our findings show that this data-driven methodology offers a promising new route to accelerate the development of reduced theoretical models of complex nonlinear plasma phenomena and to design computationally efficient algorithms for multi-scale plasma simulations.