Gauge-invariant variational formulations of electromagnetic gyrokinetic theory
Ronald Remmerswaal, Roman Hatzky, Eric Sonnendrücker
Published: 2024/3/22
Abstract
The use of gyrokinetics, wherein phase-space coordinate transformations result in a phase-space dimensionality reduction as well as the removal of fast time scales, has enabled the simulation of microturbulence in fusion devices. The state-of-the-art gyrokinetic models used in practice are parallel-only models wherein the perpendicular part of the vector potential is neglected. Such models are inherently not gauge invariant. We generalise the work of [Burby, Brizard. Physics Letters A, 383(18):2172-2175] by deriving a sufficient condition on the gyrocentre coordinate transformation which ensures gauge invariance. This leads to a parametrized family of gyrokinetic models for which we motivate a specific choice of parameters that results in the smallest gyrocentre coordinate transformation for which the resulting gyrokinetic model is consistent, gyro-phase independent, gauge invariant and has an invariant magnetic moment. Due to gauge invariance this model can be expressed directly in terms of the electromagnetic fields, rather than the potentials, and the gyrokinetic model thereby results in the macroscopic Maxwell's equations. For the linearised model, it is demonstrated that the shear and compressional Alfv\'en waves are present with the correct frequencies. The fast compressional Alfv\'en wave can be removed by making use of a Darwin-like approximation. This approximation retains the gauge invariance of the proposed model.