Non-linear MHD modelling of 3-D plasma edge with Resonant Magnetic Perturbations in DIII-D and ITER

International audience1.Introduction. Edge Localized Modes (ELMs) potentially represent an issue for the ITER divertor lifetime due to large transient heat and particle losses released in ELM relaxations [1]. The application of small Resonant Magnetic Perturbations (RMPs) demonstrated the possibility of total ELMs suppression or mitigation of their size [2], motivating the use of such method in ITER. The important drawback of RMPs and intrinsic error fields in general is the complex magnetic topology at the edge and formation of a 3D Scrape of Layer (SOL) leading to the splitting of the separatrix, seen in experiment as helical "lobes" near the X-point. When crossing the divertor plates they form finger-like structures ("footprints") and non-axisymmetric heat and particle fluxes observed in many RMP experiments [3]. The non-axisymmetric divertor fluxes with RMPs can represent an issue in ITER leading to local high heat fluxes ("hot spots") in the unprotected areas, additional material erosion and fatigue stress on the divertor components. In this work the characterisation of 3D plasma edge with RMPs was done using the non-linear MHD code JOREK [4,5]. In was shown in previous work [5,6] that a non-linear approach, two fluid diamagnetic effects, toroidal plasma rotation and neoclassical poloidal friction are the essential elements for the modelling of the non-linear plasma response to RMPs. In particular, the self-consistent interplay between RMP penetration, the evolution of radial electric field, the interaction with ELMs and heat and particle transport due to RMPs can be taken into account [6]. In this work firstly, the validation of RMP model implemented in JOREK was done comparing simulation results with measurements of divertor particle flux splitting during ELM suppression experiment on DIII-D. Secondly, the results of JOREK modelling of ITER divertor magnetic footprints and divertor fluxes with RMPs generated by ELM Coils are presented for standard H-mode scenario at 15MA/5.3T. 2. DIII-D. In DIII-D, the FASTCAM diagnostic measures D2 molecule emission near the divertor plates. It provides high temporal and spatial resolution of strike point splitting in particle flux due to RMPs in attached conditions [3]. In spite of the usually expected screening of RMPs [5], the measured particle flux splitting on the contrary, in many cases, exceeds even vacuum modelling predictions [3,7]. Note that more splitting is observed in the particle flux than in the heat flux as shown in Fig.1 for DIII-D pulse #166439 (Bt=-1.9T, Ip=1.52 MA) for the inner strike point at a toroidal angle ~75°. The parameters of this shot and the RMP spectrum with dominant toroidal modes n=1 and n=3 [7] were used for validation of the non-linear MHD plasma response model implemented in JOREK [5]. The initial equilibrium was calculated using data from the EFIT EQDSK file for #166439 at 3300ms in an ELM suppression phase. Then, after equilibrium with flows was established in JOREK modelling (at ~0.5ms), the vacuum RMPs were progressively (~0.5ms) increased at the computational boundary and JOREK ru