Training of ideal magnetohydrodynamic models with physics informed neural networks

Date

2025-1-9

Author

Özdemir, Devrim

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

284
views

585
downloads

Cite This

Physics-Informed Neural Networks (PINNs) can be used to train a set of neural network models to achieve quick and cost-effective solvers for our physical models, including Magnetohydrodynamics (MHD), which governs the behavior of plasmas in nuclear fusions and astrophysical phenomena. By integrating the fundamental physics of MHD equations into neural networks, PINNs enable quick and cost effective predictions, reducing computational costs compared to traditional numerical methods. These models can approximate solutions for plasma states, creating an operator map from one state to another. The application of PINNs to MHD not only accelerates simulations but also holds a great potential for advancing nuclear fusion research by enabling real-time predictions and optimizations in plasma behavior.

Subject Keywords

magnetohydrodynamics, neural networks, pinn, computational physics

URI

https://hdl.handle.net/11511/113519

Collections

Graduate School of Natural and Applied Sciences, Thesis