Subgrid modeling for relativistic magnetohydrodynamics with machine learning
Open Access
- 4 September 2026
- journal article
- research article
- Published by American Physical Society (APS) in Physical Review D
- Vol. 114 (6) , 063010
Abstract
Resolving the impact of magnetic field instabilities in triggering small scale turbulent flow and the associated rearrangement of the field is of critical importance in understanding multimessenger observables in binary neutron star mergers and angular momentum transport in neutron stars and accretion disks. Direct simulation of these instabilities are unfeasible, however large-eddy simulations can incorporate the impact of this turbulence with a subgrid model. We present the first machine-learning-based subgrid model for special relativistic magnetohydrodynamics, trained using a neural network. We demonstrate its performance in online simulations of the 3D Kelvin-Helmholtz instability through both and tests. Evaluated in a low resolution simulation, our model captures magnetic field amplification of a simulation at 4 times the resolution with a speedup factor of 44. This demonstrates the applicability of such methods in general relativistic simulations of neutron star mergers and other scenarios.Keywords
All Related Versions
Funding Information
- European Commission
- European Research Council (InspiReM-101043372)
- European Research Council (BinGraSp-714626)
- Deutsche Forschungsgemeinschaft (524726453)
- Deutsche Forschungsgemeinschaft (INST 275/334-1 FUGG)
- Deutsche Forschungsgemeinschaft (INST 275/363-1 FUGG)
- Gauss Centre for Supercomputing (pn67xo)
- Gauss Centre for Supercomputing (pn76li)
- HLRS (INTRHYGUE/44215)
- HLRS (MAGNETIST/44288)
This publication has 65 references indexed in Scilit:
- Turbulence modelling in neutron star merger simulationsLiving Reviews in Computational Astrophysics, 2024
- Deep neural networks for data-driven LES closure modelsJournal of Computational Physics, 2019
- Deep Learning for real-time gravitational wave detection and parameter estimation: Results with Advanced LIGO dataPhysics Letters B, 2018
- Reynolds averaged turbulence modelling using deep neural networks with embedded invarianceJournal of Fluid Mechanics, 2016
- Large Eddy Simulations in AstrophysicsLiving Reviews in Computational Astrophysics, 2015
- Large-Eddy Simulations of Magnetohydrodynamic Turbulence in Heliophysics and AstrophysicsSpace Science Reviews, 2015
- An unsplit Godunov method for ideal MHD via constrained transport in three dimensionsJournal of Computational Physics, 2008
- Nucleosynthesis, neutrino bursts and γ-rays from coalescing neutron starsNature, 1989
- Evaluation of subgrid-scale models using an accurately simulated turbulent flowJournal of Fluid Mechanics, 1979
- Energy Cascade in Large-Eddy Simulations of Turbulent Fluid Flows1975