Simulation and analysis of radiation from runaway electrons
Electron runaway constitutes one of the primary threats to future tokamak fusion reactors such as ITER. Successful prevention and mitigation of runaways relies on the development of theoretical models which accurately describe the dynamics of runaway electrons, and these models must in turn be validated in experiments. Experimental validation of models is however often made difficult by the fact that the diagnostic signals obtained in experiments only depend indirectly on the particle dynamics. In this thesis, a synthetic diagnostic model is presented which has been implemented in the Synchrotron-detecting Orbit Following Toolkit (SOFT), and which bridges this divide between theory and experiment. The synthetic diagnostic calculates the bremsstrahlung and synchrotron radiation diagnostic signals corresponding to a given runaway electron population, which can be directly compared to camera images and radiation spectra obtained in experiments. Bremsstrahlung and synchrotron radiation from runaway electrons are particularly sensitive to the runaway dynamics and, as is shown in this thesis, they provide insight into the runaway electron distribution function.
This thesis focuses on geometric effects observed in the detected radiation when magnetic field inhomogeneities and detector properties are
taken into account, something which previous studies have neglected. The dependence of the observed radiation on magnetic field geometry, detector properties and runaway parameters is characterised, and it is explained how geometric effects limit the otherwise monotonic growth of the diagnostic response function with the runaway pitch angle. The synthetic diagnostic model is applied to experiments in the Alcator C-Mod and the DIII-D tokamaks and is used to validate kinetic theory predictions of the electron distribution function. It is found that the kinetic model agrees well in certain scenarios and fails in others. In the scenarios where it fails, the synthetic diagnostic model suggests that a mechanism causing a larger spread in pitch angle may be missing from the kinetic model.