Tokamak disruptions can convert a large fraction of the plasma current into a beam of relativistic runaway electrons. In a reactor-scale device such as ITER,a runaway electron beam could carry several megaamperes and, if left uncontrolled, could cause severe damage to plasma-facing components. Predicting whether a given disruption scenario leads to a dangerous runaway beam, and designing injection schemes that prevent it, requires models that capture the interplay between material injection, rapid plasma cooling, electric field evolution, and the various mechanisms by which runaway electrons are born,multiply, and are lost. This thesis addresses runaway electron physics from seed formation to disruption mitigation through numerical modelling.

A synthetic electron cyclotron emission (ECE) framework is developed and applied to vertical ECE measurements on the TCV tokamak, combining Fokker-Planck calculations of the electron distribution function with ray tracing and radiative transfer. The analysis demonstrates that vertical ECE can resolve the energy-dependent dynamics of suprathermal electrons in the 20–100 keV range, providing constraints on the nascent runaway seed that are difficult to obtain with conventional diagnostics.

The disruption simulation framework Dream is then extended with several physics models relevant to ITER: runaway electron losses from vertical plasma displacement, cross-field drift of pellet ablation material, stochasticity driven current-profile relaxation, and an updated Compton scattering source for the ITER first wall. These are applied to a systematic study of shattered pellet injection scenarios in ITER showing that avoiding a multi-megaampere runaway beam depends sensitively on the thermal quench timescale, the injected material composition, and the competition between runaway multiplication and scrape-off losses. Finally, a viable theoretical pathway that limits the runaway current to tolerable levels even in the presence of nuclear runaway sources is identified.