Fusion energy is making great strides towards realizing its potential as a clean energy source. For decades fusion scientists and engineers have relied not only on experiment and theory but simulation to design, operate, and understand fusion devices. Today a range of simulations are applied to the fusion field, from those run in control scenarios all the way to large-scale massively parallel turbulence codes running on exascale supercomputers. AI/ML offers the promise of enhancing and accelerating these codes. Here we show research into using AI to accelerate physics kernels of larger codebases, and the care needed to ensure these PDE (Partial Differential Equations) solver surrogates conserve physical conservation laws. We also show how AI can aid in comparing experiment to simulation, using simulation-based inference techniques powered by neural networks.
This live event includes a 30-minute networking event hosted on the AI for Good Neural Network. This is your opportunity to ask questions, interact with the panelists and participants and build connections with the AI for Good community.
Register here: https://aiforgood.itu.int/event/ai-for-advancing-fusion-energy-through-enhancing-simulation/