Heating optimization studies at JET in support of ITER

Due to the large gap (∼ 20 cm) between the plasma separatrix and the antenna straps, it will be challenging to couple the foreseen 20 MW ICRF power to ITER plasmas. A method to enhance coupling is needed for ITER, even in the case of an ELM resilient ICRF antenna system, as also in that case the coupling in between ELMs will be low. Applying D₂ gas puffing in the plasma edge in H-Mode plasmas we found: (i) increase in coupling by a factor of ∼ 6 compared to similar plasmas without gas puff up to ∼ 19 cm distance between plasma and straps, with up to 8 MW of ICRF power coupled so far under such conditions; (ii) the influence of ELMs on the antenna coupling is reduced with increasing ROG. Applying D NBI on top of D majority ICRF heating at the fundamental cyclotron resonance frequency increased the neutron rate ∼ 1000 times, turning a very poor heating scheme into a more viable one. Not only the lower collisionality of the pre-heated plasma but mostly the Doppler-shifted ion-cyclotron resonance absorption of the fast beam ions is responsible for this. Simulations using a combined full wave solver / quasi-linear Fokker-Planck code (taking into account the influence of non-Maxwellian distributions of the ICRF heated particles and the injected beam ions in the wave equation) show that maximum ICRF power absorption was found at ∼ 0.5 m off-axis, to the high field side. modelling confirms that the fast D particles, which represent only 5 - 10% of the total plasma particles, are responsible for the absorption of as much as 30 - 40 % of the ICRF power. These results show that absorption of the ICRF waves can take place even in scenarios (e.g second harmonic T heating in ITER) where the D cyclotron layer is near the plasma edge or outside the plasma. Anomalous behaviour of the beam ions when using off-axis NBI was reported earlier on ASDEX Upgrade, suggesting the possibility of a reduction of the current drive capabilities of the NBI system on ITER. A detailed analysis of JET experiments which used injection of short on- and off-axis tritium beam blips (∼ 300 ms), have led to the following conclusions: (i) at low to intermediate plasma densities, the behaviour of fast particles is consistent with neoclassical theory; (ii) at high densities, fast particle losses are anomalous. These results show that for high density plasmas on ITER anomalous losses for fast ions could occur. (author)