Particle Merging Algorithm for PIC Codes

TL;DR

This paper introduces a particle merging algorithm for PIC codes that conserves charge, momentum, and energy, enabling efficient simulations of complex plasma and quantum electrodynamics scenarios with significant speedup and memory savings.

Contribution

It presents a novel merging algorithm based on particle coalescence that fully conserves physical quantities and improves simulation efficiency.

Findings

Conserves charge, momentum, and energy during merging.

Reproduces particle distributions accurately in various scenarios.

Enables significant speedup and memory savings in large-scale simulations.

Abstract

Particle-in-cell merging algorithms aim to resample dynamically the six-dimensional phase space occupied by particles without distorting substantially the physical description of the system. Whereas various approaches have been proposed in previous works, none of them seemed to be able to conserve fully charge, momentum, energy and their associated distributions. We describe here an alternative algorithm based on the coalescence of N massive or massless particles, considered to be close enough in phase space, into two new macro-particles. The local conservation of charge, momentum and energy are ensured by the resolution of a system of scalar equations. Various simulation comparisons have been carried out with and without the merging algorithm, from classical plasma physics problems to extreme scenarios where quantum electrodynamics is taken into account, showing in addition to the conservation of local quantities, the good reproducibility of the particle distributions. In case where the number of particles ought to increase exponentially in the simulation box, the dynamical merging permits a considerable speedup, and significant memory savings that otherwise would make the simulations impossible to perform.