Characterizing anomalous transport in pulsed power with continuum kinetic simulations and reduced transport models
Genia Vogman
Lawrence Livermore National Laboratory
Center for Applied Scientific Computing
Pulsed power inertial confinement fusion experiments rely on multi-megaamp currents to compress centimeter-scale Z-pinch loads on 100-nanosecond timescales. While high-beta load dynamics are effectively modeled using magnetohydrodynamic (MHD) simulations, the surrounding low-beta collisionless plasma is governed by nonlinear kinetic physics that leads to difficult-to-predict parasitic currents. Predictive modeling and experimental optimization thus require a multi-fidelity approach: high-fidelity kinetic simulations that resolve the underlying physics and reduced models that capture its net transport effects in global MHD simulations. To investigate instability-driven anomalous mass, momentum, and energy transport in these ExB systems, we employ kinetic theory and a fourth-order accurate, conservative finite-volume Vlasov solver. Refactoring this solver for GPUs yields 340-fold increase in throughput and 50-fold speedup, enabling realistic mass ratio multi-species simulations that were previously impractical. Using these simulations, we show how acceleration-driven lower hybrid drift instabilities generate anomalous resistivity and heating, and we assess the degree to which this physics is captured by reduced models. We derive and test a complete quasilinear diffusion model and a further-reduced parametrized model that relates linear growth rates to an effective collisionality. We discuss how the bulk effect of kinetic physics can be incorporated into MHD simulations and we outline the remaining challenges for robust predictive modeling. Prepared by LLNL under Contract DE-AC52-07NA27344.