Development of a Machine Learning Algorithm for Tomography reconstructions in RFX-mod

In magnetic fusion devices, tomography reconstruction is a technique that can be applied to Soft-X-Ray (SXR) data (brilliance) to obtain a spatial distribution of the plasma properties - in particular, the electron temperature. One of the challenges with the current tomographic reconstruction method is its time-intensive nature, requiring significant computational resources. In this context, the development and application of a machine learning model can help expedite the tomographic reconstruction process, enabling its reconstruction in real-time. The aim of this Thesis project is to analyze SXR data of RFX-mod plasmas (Consorzio RFX, Padova, Italy) and to create a database usable for machine learning. A machine learning procedure will be applied to obtain the coefficients of the Cormack-Bessel expansion, used to mathematically retrieve the 2D emissivity map.