Speaker
Description
Achieving burning fusion plasmas relies on the confinement of energetic fast ions, yet this remains challenging due to the gaps in our understanding of the fast-ion distribution function. However, the reconstruction of high-dimensional fast-ion distribution function is a severely ill-posed inverse problem. As a result, to find physically meaningful solutions, the experimental data must be augmented by prior information.
In this work, we present a method for reconstructing four-dimensional fast-ion distributions in JET using synthetic data. The approach employs a large set of basis functions, which are energetic particle distributions simulated by ASCOT. In these simulations, energetic ions are injected on a grid in constants-of-motion space and tracked in time obeying neoclassical collision physics. Because this domain describes all topologically allowed orbits in the plasma, it serves as a natural framework to capture all physically possible reconstructions of the fast-ion distribution function. The basis functions are evolved only over a partial slowing-down time to capture correlations between neighboring basis functions. We characterize the properties of these basis functions, analyze the resulting correlation structures, and demonstrate that the approach improves the fidelity of reconstructed four-dimensional fast-ion distributions.