ERROR FIELD CONTROL IN MAGNETIC FUSION DEVICES

In magnetic fusion devices, undesired non-axisymmetric magnetic field perturbations, typically called error fields, have been observed to have a detrimental effect on plasma stability and confinement. The main strategies that can be adopted to reduce error fields are: a careful alignment of the field coils, i.e. applying coil shift and tilt, when assembling a fusion device, and the installation of error field correction coils capable of suppressing resonant components of the error fields by superposition. In this thesis, useful methods to study the error field topology in MAST-U and ITER are presented. In MAST-U the study is performed through the compass scan technique, complemented with a new method based of transfer functions compensation to precisely resolve the EF measure. The results suggests that the intrinsic error field source is relatively small with respect to MAST. For ITER, a similar compensation routine has been developed using an electromagnetic model to estimate the coupling between each actuator coil and the array of magnetic probes. This information plays a crucial role in error field characterization and correction, which are essential steps in preparing for ITER plasma startup. However, despite the efforts to avoid the onset of locked magnetic instabilities, disruption mitigation systems are considered essential during ITER operations and in the view of the next fusion reactors such as the DEMO. Thanks to the large availability of data, machine learning methods show to be a promising tool towards the achievement of this task. Predicting the proximity to locked mode formation with sufficient time margin allows to modify the plasma state promoting a detour from such dramatic outcome or at least, reach a machine safety regime. For this purpose, two machine learning models have been successfully train and tested on a large database of MAST-U discharges proving their prediction reliability. Finally, the last pages of this work offer an overview on the scientific feasibility of alternative fusion concepts. In reaching the goal of a burning plasma reactor, every fusion approach, being either old or new, introduces advantages and drawbacks. Some of the most limiting and significant ones will be addressed in this thesis.