Description
Type-I edge localised modes (ELMs) repetitively expel heat and particles from the confined plasma volume to the Scrape-Off Layer (SOL) and divertor targets of tokamak devices operating in high-confinement conditions. Future devices like ITER, SPARC, and STEP are foreseen to require ELM control to avoid excessive damage to the divertors and other plasma facing components. An option to achieve it is to operate in the so-called Quasi-Continuous Exhaust (QCE) regime, which relies on cross-field transport localised near the last closed flux surface (separatrix) to regulate the H-mode pedestal and, thus, avoid large ELMs. Access to the QCE regime has been observed in several tokamaks operating with sufficiently high shaping (elongation/triangularity) and high separatrix density, while high ne,sep at weak shaping can first lead to poor confinement and eventually to H-L back-transition.
In this work, JOREK simulations of AUG #36165, which features a transition from type-I ELMs to the QCE regime by means of stepped increases of neutral gas fuelling, are described. A kinetic particle-in cell model for atomic hydrogenic neutrals two-way coupled to the non-linear extended MHD JOREK code is used to assess the non-linear pedestal stability with respect to increasing gas puff rate. The initial magnetic equilibrium is obtained from an experimental reconstruction, which is used to build a flux-surface aligned grid including the private flux region and SOL and extended all the way to the AUG first wall. The kinetic PIC model is time dependent and includes wall recycling, reflection, simplified neutrals pumping as well as the following atomic processes: ionisation, recombination, charge-exchange, and continuum + effective line radiation. With this approach, the transition from type-I ELMs at low gas fuelling rate to QCE regime at high fuelling rate is reproduced and characterised. The main mechanisms responsible for the transition will be analysed and discussed.