Physics of Magnetic Perturbations for Fusion Plasma Control

Magnetic confinement fusion requires robust control of plasma stability and transport to ensure reliable and sustained operation. Among the available actuators, externally applied non-axisymmetric magnetic perturbations, and in particular Resonant Magnetic Perturbations (RMPs), have emerged as a versatile tool for multiple control tasks, including the correction of intrinsic error fields, the mitigation or suppression of Edge Localized Modes (ELMs), and the regulation of plasma rotation. Understanding and optimizing the plasma response to these perturbations is therefore essential for both present experiments and the design of future reactors. This Thesis investigates the role of magnetic perturbations as a control actuator in fusion plasmas. A consistent numerical framework, based on equilibrium reconstruction with CHEASE and linear and quasi-linear plasma response modeling with the MARS-* family of codes, was developed and extensively applied to interpret experiments and to perform predictive optimization studies across three devices of increasing complexity and scale: RFX-mod, MAST-U, and JT-60SA. On RFX-mod, Resonant Field Amplification (RFA) experiments and MARS-F modeling showed that externally imposed perturbations preferentially excite intrinsic eigenstructures. Counter-rotating perturbations sustain marginally stable equilibria, while static ones reveal contributions from intrinsic error fields. These results consolidate RFA as a selective probe of stability and lay the foundation for the development of a robust MHD spectroscopy technique. On MAST-U, optimization of ELM control with RMPs was addressed through a hierarchy of linear and quasi-linear metrics. Dedicated pedestal density scans highlighted the sensitivity of metric optima to edge conditions, underscoring the importance of accurate kinetic reconstructions. Modeling identified a dominant intrinsic n=2 error field originating from the poloidal field system, capable of degrading confinement and reducing control performance. Complementary experiments revealed direct evidence of edge-localized toroidal acceleration correlated with the RMP waveform, interpreted as neoclassical toroidal viscosity torque in MARS-K simulations, with core braking attributed to MHD torques and amplified by intrinsic asymmetries. For JT-60SA, a predictive framework for ELM control optimization was established by systematically scanning coil phasing and current ratios. Linear and quasi-linear indicators identified intermediate phasing as optimal for maximizing edge response while minimizing core impact. Time-dependent quasi-linear simulations confirmed that optimized perturbations flatten the edge ExB rotation up to the q=3 surface while leaving the core largely unaffected, thereby creating conditions consistent with ELM mitigation. Overall, the Thesis contributes to the understanding of how 3D coils, producing magnetic perturbations, can be used as flexible actuators for plasma control. By integrating experimental evidence with predictive modeling, this work refines strategies for error-field correction, ELM mitigation, and momentum transport control. The results clarify how advanced use of RMPs may enhance discharge reliability, extend operational boundaries, inform actuator design and scenario preparation in next-generation devices. In this way, the work supports the consolidation of magnetic perturbations as a central element of plasma control systems, contributing to the development of ITER, DEMO, and future fusion reactors.