A Bayesian formulation for perturbed current tomography in tokamaks

Abstract

Abstract An initial investigation into the application of Bayesian inference to the reconstruction of the spatial distribution of current perturbations in tokamaks from diagnostic signals is presented. Previous work in Bayesian equilibrium current tomography is extended to the case of a complex phasor representation of harmonically time varying current distributions. A forward function to predict the response of magnetic diagnostics is constructed using only electrodynamics and not reduced models of ideal MHD. The extension of this forward function to incorporate a fully kinetic model of the plasma state is suggested. The response of soft x-ray diagnostics, and the motional Stark effect diagnostic to the current perturbations are also predicted and the integration of all diagnostics into a single estimate of the current perturbation is proposed. Simulations with synthetic diagnostics in simple geometry demonstrate that the perturbed current distribution can be reconstructed subject to prior assumptions regarding solution smoothness.