Driving the optimization of nuclear fusion reactor design through machine learning techniques

As the world seeks cleaner and more sustainable energy solutions, nuclear fusion has emerged as a promising contender. This thesis proposal aims to adopt machine learning techniques to make fusion reactors even better.

Objectives:

• Improve the design optimization process by means of machine learning techniques, both aimed at exploring currently adopted design patterns and generating new ones accordingly
• Develop a framework that integrates Electro-Magnetic, Monte-Carlo and Magneto-Hydro-Dynamics simulations, to provide a more accurate and complete picture of the reactor’s behavior, hopefully at a fraction of the computational cost!

Details:

🤝 Collaboration with the Department of Applied Science and Technology (DISAT), with Prof. F. Laviano, Dr. D. Torsello

🏫 Possibility of a period abroad (MIT)

💰 Possibility for a monthly grant