Neoclassical islands, β-limits, error fields and ELMs in reactor scale tokamaks

An assessment is presented of the impact of recent magnetohydrodynamic research results on performance projections for reactor scale tokamaks as exemplified by the ITER Final Design Report (ITER/FDR) facility. For nominal ELMy H mode operation, the presence and amplitude of neoclassical tearing modes governs the achievable β value. Recent work finds that the scaling of β at which such modes onset agrees well with a polarization drift model, with the consequence that, with reasonable assumptions regarding seed island width, the mode onset β will be lower in reactor scale tokamaks than in contemporary devices. Confinement degradation by such modes, on the other hand, depends on relative saturated island size which is governed principally by β and secondarily by ν* effects on bootstrap current density. Relative saturated island size should be comparable in present and reactor devices. DT ITER demonstration discharges in JET exhibited no confinement degradation at the planned ITER operating value of β_N = 2.2. Theory indicates that electron cyclotron current drive can either stabilize these modes or appreciably reduce saturated island size. Turning to operation in candidate steady state, reverse shear, high bootstrap fraction configurations, wall stabilization of external kink modes is effective while the plasma is rotating but (so far) rotation has not been maintained. Recent error field observations in JET imply an error field size scaling that leads to a projection that the ITER/FDR facility will be somewhat more tolerant to error fields than thought previously. ICRF experiments on JET and Alcator C-Mod indicate that plasmas heated by central energetic particles have benign ELMs compared with the usual type 1 ELM of NBI heated discharges.