M.R. Nurgaliev1, E.N. Khayrutdinov1, R. Clark2, D.M. Orlov2, G.F. Subbotin1
1 Next Step Fusion SARL, Mondorf-les-Bains, Luxembourg
2 University of California San Diego, La Jolla, CA 92093, United States of America
Plasma shape is one of the key parameters that could be used for optimization to achieve more stable and more efficient plasma confinement. Recent results [1] demonstrated the possibility of using reinforcement learning (RL) algorithms to develop efficient controllers that do not require equilibrium reconstruction and operate using magnetic diagnostics signals directly.
The goal of this work is to validate the simulation models for magnetic probes and flux loops for their future applications in the RL simulation environment.
A new Next Step Fusion Simulator (NSFsim) is used for modeling. It is a free-boundary equilibrium and transport solver based on the renowned simulation approach used in DINA code [2].
Download full poster here.
