Dependence of a current driven ELM self-amplification process on the plasma shape

Abstract

The numerical model of the non-linear evolution of edge-localized modes (ELMs) in tokamaks being used in this paper assumes that thermoelectric currents flow in short connection length flux tubes, initially established by error fields or other non-axisymmetric magnetic perturbations. The additional magnetic perturbation of the current filaments changes the magnetic topology. In a self-amplification process, more flux tubes are created which eventually allow more thermoelectric current to flow through the plasma edge. The process of flux tube formation is highly sensitive to the position of the secondary X-point in typical single null discharges in DIII-D. A new scenario for cases with large distances Δs between the secondary X-point and the primary separatrix is presented. In the numerical simulations, as Δs is increased the current evolution through short connection length flux tubes changes significantly. Ultimately, a final state with large stripe structures is found that results in footprints on the vessel wall which are similar to those found when Δs is small (Wingen et al 2010 Phys. Rev. Lett. 104 175001).

Resonant magnetic perturbations of edge-plasmas in toroidal confinement devices

“…Simulations of divertor footprints, due to homoclinic and heteroclinic separatrix tangles, show that they rotate toroidally and poloidally due to their helical structure as the safety factor is varied and the size of the interaction of the lobes, making up the tangle, with plasma-facing surface is significantly altered at each toroidal and poloidal location. These simulations suggest that elevated divertor heat fluxes result from force-free currents that flow through the tangle from one target plate to another and result in an amplification of the lobes [75,143,153,289]. Based on these simulation results it is reasonable to assume that the lobes of the = m n / 3/1 island tangle could have been amplified by the LHCD resulting in an enhanced level of boundary layer stochasticity.…”

“…• Uncompensated equilibrium coil leads; • Transient image current in conducting wall components due to core MHD and RWMs; • Thermoelectric scrape-off layer [152]; • Homoclinic [103,143] and heteroclinic tangle currents [75,153]; • Halo currents during disruptions [154] referred to as Hiro [152,155] and Evans [156] currents by Zakharov [157]; • Resonant field amplification (RFA) due to stable kink modes [158]; • Ferromagnetic materials; • Insulating gaps in the vacuum vessel [107]; and • Ports in the vacuum vessel [159]. Time-dependent 3D perturbations fields generated by currents in the plasma, often in response to 3D vacuum perturbation fields, can combine with external perturbation fields in the plasma leading to the possibility of a complex nonlinear plasma response triggered by the external 3D vacuum fields.…”

“…Correlations between the divertor heat flux footprints and currents measured with the TTCA suggested that these currents are driven by thermoelectric potentials due to plasma sheath and pre-sheath potential asymmetries caused by T e differences between the HFS and LFS divertor target plates [102,287,288]. Based on the TTCA measurements of the currents during locked modes, ELMs and disruptions it was proposed that homoclinic tangles and stochastic layers associated with these tangles will be amplified by the forcefree parallel currents flowing inside coherent helical flux tubes embedded inside the lobes of the tangles [75,103,143,153]. Numerical modeling studies of homoclinic and heteroclinic tangles using currents measured with the TTCA have verified that the simulated divertor footprints are consistent with heat flux measurements during ELMs in DIII-D [75,143,289].…”

“…The results shown in figure 31 demonstrate that all of the analytic models underestimate the experimental results by at least an order of magnitude except for the fluid model χ e f , which is about a factor of three too small and has the wrong radial dependence. This is to be expected since not encompass a full, highfidelity, description of all the perturbation sources, such as those used in numerical field line integration codes, and do not include the physics effects associated with changes in the background turbulence or those due to interactions caused by field lines that connect to plasma-facing surfaces such as the generation of thermoelectric currents [103,152,153]. Recent RMP transport modeling with a full-f transport code xgc0 has shown the importance of including kinetic effects such as trapped particles and a self-consistent radial electric field E r [313].…”