Unsupervised Discovery of Inertial-Fusion Plasma Physics using Differentiable Kinetic Simulations and a Maximum Entropy Loss Function

Abstract

Plasma supports collective modes and particle-wave interactions that leads to complex behavior in inertial fusion energy applications. While plasma can sometimes be modeled as a charged fluid, a kinetic description is useful towards the study of nonlinear effects in the higher dimensional momentum-position phase-space that describes the full complexity of plasma dynamics. We create a differentiable solver for the plasma kinetics 3D partial-differential-equation and introduce a domain-specific objective function. Using this framework, we perform gradient-based optimization of neural networks that provide forcing function parameters to the differentiable solver given a set of initial conditions. We apply this to an inertial-fusion relevant configuration and find that the optimization process exploits a novel physical effect that has previously remained undiscovered.

arXiv comment2nd AI4Science Workshop at the 39th International Conference on Machine Learning (ICML), 2022

References in corpus (17)

• SciPy 1.0--Fundamental Algorithms for Scientific Computing in Python
• Learning data driven discretizations for partial differential equations
• Evidence for Electron Landau Damping in Space Plasma Turbulence
• Automation and control of laser wakefield accelerators using Bayesian optimisation
• New method to design stellarator coils without the winding surface
• Coherent control of plasma dynamics
• Learned discretizations for passive scalar advection in a 2-D turbulent flow
• Kinetic Enhancement of Raman Backscatter, and Electron Acoustic Thomson Scatter
• The DESC Stellarator Code Suite Part III: Quasi-symmetry optimization
• ViDA: a Vlasov-DArwin solver for plasma physics at electron scales
• Optimized finite-build stellarator coils using automatic differentiation
• Simulations of Kinetic Electrostatic Electron Nonlinear (KEEN) Waves with Variable Velocity Resolution Grids and High-Order Time-Splitting
• Convective Raman Amplification of Light Pulses Causing Kinetic Inflation in Inertial Fusion Plasmas
• Adjoint approach to calculating shape gradients for 3D magnetic confinement equilibria
• Optimizing Laser-Plasma Interactions for Ion Acceleration using Particle-in-Cell Simulations and Evolutionary Algorithms
• Gradient-based optimization of 3D MHD equilibria
• Adjoint approach to calculating shape gradients for three-dimensional magnetic confinement equilibria. Part II: Applications

Cited by in corpus (2)

• Learning the dynamics of a one-dimensional plasma model with graph neural networks
• Learning collision operators from plasma phase space data using differentiable simulators