Gyroaveraging operations using adaptive matrix operators

TL;DR

This paper introduces an adaptive matrix-based gyroaveraging scheme for Particle-In-Cell codes that improves accuracy in complex geometries by adjusting to local plasma conditions, demonstrated in the XGC code.

Contribution

It presents a novel adaptive scheme for gyroaveraging in Particle-In-Cell simulations that enhances accuracy across varying magnetic and temperature profiles.

Findings

Improved accuracy in gyroaveraging with strong spatial variations.

Effective implementation in 2D and 3D geometries.

Successful integration into the XGC code.

Abstract

A new adaptive scheme to be used in Particle-In-Cell codes for carrying out gyroaveraging operations with matrices is presented. This new scheme uses an intermediate velocity grid whose resolution is adapted to the local thermal Larmor radius. The charge density is computed by projecting marker weights in a field-line following manner while preserving the adiabatic magnetic moment $\mu$. These choices permit to improve the accuracy of the gyroaveraging operations performed with matrices even when strong spatial variation of temperature and magnetic field is present. Accuracy of the scheme in different geometries from simple 2d slab geometry to realistic 3d toroidal equilibrium has been studied. A successful implementation in the grokinetic code XGC is presented in the delta-f limit.