Data-driven modeling and pseudospectral analysis for MHD systems in two and three dimensions

First principles models of plasmas lead to high-dimensional nonlinear systems of equations requiring complex MHD or kinetic simulations. Projection-based and data-driven modeling algorithms, such as dynamic mode decomposition (DMD), offer a powerful approach to building low-dimensional reduced models of plasma systems. These reduced models can permit improved active control of plasmas without highly resolved, expensive simulations of the dynamics. This poster will present optimized DMD [1] analysis of a set of astrophysical accretion flow simulations [2] with varied toroidal guide field and a transition from magnetorotational to magneto-curvature instability. The visco-resistive MHD equations are also known to be non-modal [3], permitting finite-time amplification even in stable systems. To understand these dynamics, pseudospectral analysis will be presented using a new 2D MHD code within the OpenFUSIONToolkit and the corresponding linearized MHD operator as discretized for 2D tearing systems.

[1] - Ashkam and Kutz, SIAM J. Applied Dynamical Systems (2018)

[2] - Ebrahimi and Pharr. The Astrophysical Journal, 936:145 (2022)

[3] - D. MacTaggart. J. Plasma Phys. (2018)