Reinforcement Learning for Plasma Control: A Proof of Concept for NTM Suppression

Neoclassical tearing modes (NTMs) are magnetohydrodynamic instabilities that generate magnetic islands in tokamak plasmas, degrading confinement and potentially limiting high-performance operation. Their stabilization typically requires precise alignment and appropriate injection of electron cyclotron (EC) power beams, making real-time control a challenging task. In this work, we present a proof-of-principle study aimed at investigating the potential role of neural networks in the control of plasma instabilities. The objective is not to develop a device-specific controller, but rather to explore, within a synthetic environment, how a learning-based agent can autonomously discover effective stabilization strategies. To this end, a neural network controller is trained using reinforcement learning techniques, resulting in an intelligent and agnostic control system. The controller is defined as intelligent in the sense that it learns the optimal strategy directly from interaction with the environment, without being explicitly programmed or guided by a predefined control law. It is agnostic because it does not rely on equilibrium reconstruction or explicit knowledge of the deposition location relative to the island. Instead, it operates solely on feedback derived from a representation of the magnetic island width, using this information to adapt its actions. Two control tasks are considered: pure angular alignment and combined angular alignment with power control. This exploratory study establishes a framework for assessing the potential advantages of data-driven approaches in magnetic island control and provides a basis for future investigations aimed at improving alignment and suppression strategies in fusion plasmas.