This paper was produced for the 2019 NAFEMS World Congress in Quebec Canada

Resource Abstract

Nuclear fusion is the process that heats the stars by the collision of atomic nuclei which fuse together to form heavier elements and release energy. The generation of energy using this process has several advantages: no carbon emissions, abundant fuel supplies, efficiency, reliability and operationally safe. One way to achieve the necessary conditions for producing fusion energy on earth is by controlling a hot gas of fully ionized hydrogen isotopes (plasma) with strong magnets in a ring-shaped magnetic chamber known as tokamak. The real-time control of this hot plasma requires magnetic diagnostic and actuators which must be designed to be reliable and immune to undesirable interferences. The heating and stabilization of the plasma partly rest on Radio-Frequency (RF) antennas which must be designed and controlled carefully to avoid undesired plasma-wall interactions that can produce excessive heat-loads and endanger the integrity of the machine. Also, the safe installation of the different diagnostics, devices and structures in and around the fusion machine requires the knowledge of the Lorentz forces induced by the time-varying electromagnetic fields present during the operation of the machine.



As can been expected from the above examples, nuclear fusion engineering and design can be greatly benefited from the use of computational electromagnetic software tools. In this work we are going to present how the open-source finite element tool ERMES has been upgraded to tackle nuclear fusion-related problems. The updated version of ERMES can solve problems from the static regime (electrostatic and magnetostatic), to the high-frequency regime (interaction of radio frequency waves with plasma and walls), passing through the quasi-static low frequency regime (induced eddy currents). A novel finite element formulation has been implemented to compute the interaction of the electromagnetic waves with the inhomogeneous anisotropic cold plasma present in the scrape-off layer close to the RF antennas. New numerical models have been developed to estimate the probability of arcing under different failure scenarios (e.g. unmitigated superconducting coil quench). The generality of these new developments allows a straightforward application in engineering problems outside the nuclear fusion environment as in the design of helicon plasma thrusters, high voltage engineering, inductive heating, and bio-electromagnetism.