Physics-Preserving AI-Accelerated Simulations of Plasma Turbulence
This addresses the problem of computationally expensive plasma turbulence simulations for researchers in physics and engineering, representing a novel method rather than an incremental improvement.
The paper tackled the computational challenge of simulating plasma turbulence by combining Large Eddy Simulation with a machine learning-based sub-grid-scale model, reducing computational effort by about three orders of magnitude while preserving statistical physical properties.
Turbulence in fluids, gases, and plasmas remains an open problem of both practical and fundamental importance. Its irreducible complexity usually cannot be tackled computationally in a brute-force style. Here, we combine Large Eddy Simulation (LES) techniques with Machine Learning (ML) to retain only the largest dynamics explicitly, while small-scale dynamics are described by an ML-based sub-grid-scale model. Applying this novel approach to self-driven plasma turbulence allows us to remove large parts of the inertial range, reducing the computational effort by about three orders of magnitude, while retaining the statistical physical properties of the turbulent system.