Development and Validation of a software for the Analysis of Antennas for Controlled Magnetically Confined Nuclear Fusion

Ion-Cyclotron Resonance Heating (ICRH) and Lower-Hybrid (LH)Resonance Heating are key parts of all present-day experiment toward the realization of controlled nuclear fusion with magnetic confinement. Both auxiliary heating systems are essentially antennas made up of complex plasma facing components, charged with the difficult mission of delivering extremely high RF powers to the plasma with typically poor loadings (as compared to antennas in radar or broadcasting), and high near-field reactive energies that generate serious mismatches to the feeding power lines. Because of the impossibility of testing these antennas in plasmas outside the actual experiments for which they have to be designed, availability of accurate simulation tools is a key factor in assisting their design: this work is concerned with the numerical analysis of these plasma facing antennas, through the upgrade of an existing code called TOPICA and its recent extension named TOPLHA. The antenna simulation code must be able to handle the actual geometry of both antennas (including their housing in the experiment, the protective screens, etc.), which have witnessed a constant increase in complexity. On the other hand, plasma loading on the antenna is extremely sensitive to the plasma profile, especially near the antenna itself: the antenna code must therefore be able to correctly account for the plasma conditions, which makes it necessary to include non-cold plasma terms (that affect resonances). More specifically, since the frequency range of the two heating systems is quite different (below 100 MHz for ICRH and around few GHz for LH), the two antennas differ not only in the geometrical features (ICRH antennas are essentially metal loops while LH antennas are open-ended waveguide arrays), but also in the way the wave propagates in the plasma and the heating process happens. The problem can be solved with a considerable numerical efficiency by formally separating it into two parts: the vacuum region around the antenna conductors and the plasma region. This approach leads to the problem formulation via a set of two coupled integral equations, further discretized via the Method of Moments (MoM). The MoM is used in a hybrid form: spatial-domain approach is employed for the antenna and other conductors (with high geometrical complexity), while a spectral-domain approach is used for the plasma region (as plasma description is naturally available in this domain). Numerical tests have already been performed for simple ICRH launchers, and results compared with available experimental data both in vacuum and with real plasmas. Eventually, this work has extended the existing capabilities of TOPICA in two directions: the efficient handling of IC antennas housed in near, but distinct recesses and the LH range. Both problems are multi-cavity, in the sense that the antennas are recessed in a modular way. Starting from the validated version of TOPICA, a new approach has been developed to allow the code handle a much greater number of unknowns (from 10,000 to more than 150,000). In the IC range, this is dictated primarily by geometrical complexity, while the overall electrical length of one recess does not exceed one half free-space wavelength; in the LH range, conversely, geometry is smoother, but electrical size is larger. The multi-cavity approach addressed in this work exploits the fact that the inner parts of the individual cavities are coupled one to each other only through the equivalent currents on their apertures, and accordingly solves the global MoM linear system block-wise, with significant memory and time saving. Furthermore, this modular approach led to a heavy parallelization of the code, with an astonishing increase in the overall performances