Experiments were carried out in the JET tokamak to determine the critical ion temperature inverse gradient length (R/LTi=R|nablaTi|/Ti) for the onset of ion temperature gradient modes and the stiffness of Ti profiles with respect to deviations from the critical value. Threshold and stiffness have been compared with linear and nonlinear predictions of the gyrokinetic code GS2. Plasmas with higher values of toroidal rotation show a significant increase in R/LTi, which is found to be mainly due to a decrease of the stiffness level. This finding has implications on the extrapolation to future machines of present day results on the role of rotation on confinement.

New transport experiments on JET indicate that ion stiffness mitigation in the core of a rotating plasma, as described by Mantica et al. [Phys. Rev. Lett. 102, 175002 (2009)] results from the combined effect of high rotational shear and low magnetic shear. The observations have important implications for the understanding of improved ion core confinement in advanced tokamak scenarios. Simulations using quasilinear fluid and gyrofluid models show features of stiffness mitigation, while nonlinear gyrokinetic simulations do not. The JET experiments indicate that advanced tokamak scenarios in future devices will require sufficient rotational shear and the capability of q profile manipulation.

No abstract

We present an ultrafast neural network (NN) model, QLKNN, which predicts core tokamak transport heat and particle fluxes. QLKNN is a surrogate model based on a database of 300 million flux calculations of the quasilinear gyrokinetic transport model QuaLiKiz. The database covers a wide range of realistic tokamak core parameters. Physical features such as the existence of a critical gradient for the onset of turbulent transport were integrated into the neural network training methodology. We have coupled QLKNN to the tokamak modelling framework JINTRAC and rapid control-oriented tokamak transport solver RAPTOR. The coupled frameworks are demonstrated and validated through application to three JET shots covering a representative spread of H-mode operating space, predicting turbulent transport of energy and particles in the plasma core. JINTRAC-QLKNN and RAPTOR-QLKNN are able to accurately reproduce JINTRAC-QuaLiKiz T i,e and n e profiles, but 3 to 5 orders of magnitude faster. Simulations which take hours are reduced down to only a few tens of seconds. The discrepancy in the final source-driven predicted profiles between QLKNN and QuaLiKiz is on the order 1%-15%. Also the dynamic behaviour was well captured by QLKNN, with differences of only 4%-10% compared to JINTRAC-QuaLiKiz observed at mid-radius, for a study of density buildup following the L-H transition. Deployment of neural network surrogate models in multi-physics integrated tokamak modelling is a promising route towards enabling accurate and fast tokamak scenario optimization, Uncertainty Quantification, and control applications.

The wall loads due to fusion alphas as well as neutral beam injection- and ICRF-generated fast ions were simulated for ITER reference scenario-2 and scenario-4 including the effects of ferritic inserts (FIs), test blanket modules (TBMs), and 3D wall with two limiter structures. The simulations were carried out using the Monte Carlo guiding-centre orbit-following code ASCOT. The FIs were found very effective in ameliorating the detrimental effects of the toroidal ripple: the fast ion wall loads are reduced practically to their negligible axisymmetric level. The thermonuclear alpha particles overwhelmingly dominate the wall power flux. In scenario-4 practically all the power goes to the limiters, while in scenario-2 the load is fairly evenly divided between the divertor and the limiter, with hardly any power flux to other components in the first wall. This is opposite to earlier results, where hot spots were observed with 2D wall (Tobita et al 2003 Fusion Eng. Des. 65 561–8). In contrast, uncompensated ripple leads to unacceptable peak power fluxes of 0.5 MW m−2 in scenario-2 and 1 MW m−2 in scenario-4, with practically all power hitting the limiters and substantial flux arriving even at the unprotected first wall components. The local TBM structures were found to perturb the magnetic field structure globally and lead to increased wall loads. However, the TBM simulation results overestimate the TBM contribution due to an over-simplification in the vacuum field. Therefore the TBM results should be considered as an upper limit.