Numerical Methods for Propulsive Plasma Simulation

Peter Norgaard, Princeton University

Space thrusters which utilize the electromagnetic acceleration of plasma are widely recognized as an effective solution for space flight applications ranging from station keeping to interplanetary travel. The motivations for numerical simulation of these thrusters are clear - to achieve predictive capability of performance characteristics such as thrust and power, and to gain insight into the dominant physical processes. However, the effective numerical simulation of plasma thrusters remains a difficult task. A resistive magnetohydrodynamic (MHD) treatment of the plasma is the most practical method, given the length and time scales of interest. Unfortunately, this completely ignores such relevant physics as plasma-wall interaction, plasma sheaths, ionization physics, and the micro-instabilities that lead to anomalous transport. Since a full kinetic or particle based simulation is computationally impractical, I instead focus on ways in which the resistive fluid model can be augmented or hybridized to include the necessary physics.

Several candidate methods are being considered to improve the physics capabilities of our parallel finite volume resistive MHD code. Wherever possible, it is desirable from a numerical standpoint to use an empirical law which calculates coefficients using the fluid quantities. That failing, it may be possible to intermittently conduct either particle or kinetic micro-simulations in order to calculate certain coefficients. An interesting alternative to that is the ‘equation-free’ method. Another difficulty is that in some regions of the thruster the density becomes so low that a fluid model is invalid. Here, we might consider generating an interface between fluid and particle models. In each of these cases, fundamental issues of compatible time scales, model accuracy, and numerical cost leave numerous possibilities for substantial future research.

Abstract Author(s): P.C. Norgaard, E.Y. Choueiri, and S.C. Jardin