Toroidal modelling of RMP response in ASDEX Upgrade: coil phase scan, q(95) dependence, and toroidal torques

Journal article, 2016

The plasma response to the vacuum resonant magnetic perturbation (RMP) fields, produced by the ELM control coils in ASDEX Upgrade experiments, is computationally modelled using the MARS-F/K codes (Liu et al 2000 Phys, Plasmas 7 3681, Liu et al 2008 Phys. Plasmas 15 112503), A systematic investigation is carried out, considering various plasma and coil configurations as in the ELM control experiments. The low q plasmas, with q95 3.8 (q95 is the safety factor q value at 95% of the equilibrium poloidal flux), responding to low a (n is the toroidal mode number) field perturbations from each single row of the ELM coils, generates a core kink amplification effect. Combining two rows, with different toroidal phasing, thus leads to either cancellation or reinforcement of the core kink response, which in turn determines the poloidal location of the peak plasma surface displacement, The core kink response is typically weak for the a = 4 coil configuration at low q, and for the n = 2 configuration but only at high q (q(95) similar to 5.5). A phase shift of around 60 degrees for low q plasmas, and around 90 degrees for high q plasmas, is found in the coil phasing, between the plasma response field and the vacuum RMP field, that maximizes the edge resonant field component, This leads to an optimal coil phasing of about 100 (-100) degrees for low (high) q plasmas, that maximizes both the edge resonant field component and the plasma surface displacement near the X-point of the separatrix. This optimal phasing closely corresponds to the best ELM mitigation observed in experiments. A strong parallel sound wave damping moderately reduces the core kink response but has minor effect on the edge peeling response. For low q plasmas, modelling shows that both the resonant electromagnetic torque and the neoclassical toroidal viscous (NTV) torque (due to the presence of 3D magnetic field perturbations) contribute to the toroidal flow damping, in particular near the plasma edge region. For high q plasmas, however, significant amount of torque is also produced in the bulk plasma region, and the contributions from the electromagnetic, the NTV, and the torque associated with the Reynolds stress, all