Turbulence and associated transport degrade the confinement of tokamak plasmas, reducing the expected performance in terms of energy gain. Experimentally, several regimes of improved confinement have been observed, notably those where turbulent transport is strongly reduced at the plasma edge. These external transport barriers lead to strong density and/or temperature gradients that maximize the energy content of the confined plasma. These spontaneous bifurcations result from the self-organization of turbulence under the forcing of different sources, of particles and heat. Their mechanism is poorly understood, not least because of the topological complexity of this outer region and the wealth of probable processes at play. These regimes represent a major opportunity for achieving the best performance in ITER plasmas. It is therefore crucial to gain a better understanding of them, so as to be able to predict their transition thresholds and, if possible, control them.
The proposed PhD thesis falls within this framework. It is based on state-of-the-art numerical modeling of fusion plasmas, the five dimensional (in phase space) gyrokinetic approach. Recent developments have made it possible to treat self-consistently both transport of matter and heat in this peripheral region. What remains to be done is to implement a source of neutral particles which, through ionization, will constitute the plasma's density forcing. We already know, thanks in particular to reduced models, that this dynamic source plays a crucial role in self-organization processes. The aim of this thesis work is to couple a reduced fluid model of neutrals to kinetically described electrons and ions, and to study their impact on turbulent transport and self-organization using high-performance computing (HPC) simulations with the GYSELA code.