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.

In the last campaign, the TJ-II heliac has been operated under lithium-coated walls, representing the first stellarator ever working under these boundary conditions. Enhanced density control and discharge reproducibility, leading to the drastic enlargement of the operational window, have been obtained. A strong decrease in recycling together with changes in the shot by shot fuelling characteristics and in the wall particle inventory have been recorded. These changes, associated with the new wall scenario, had led to a long-lasting good density control. The new conditions were also mirrored in the plasma profiles under NBI heating scenarios with increased peaking of the electron density profiles. Fuelling rates corresponding just to the nominal beam current were obtained for the first time, and transitions from bell to dome-type plasma profiles, with different collapsing limits, were observed and tentatively ascribed to changes in the local edge power balance. ELM-type activity was observed in concomitance to reduced fluctuation levels and confinement improvement. Record values of plasma energy content were measured at central densities up to 8 × 10 19 m −3 under Li-coated walls.

No abstract

Some techniques with a long tradition in the plasma technology field have already been successfully applied to research in plasma-wall interactions of fusion devices. They have produced important advances in the control of particle and energy exhaust. In this paper, the possible application of these techniques to the problem of tritium inventory control in fusion reactors with carbon-based plasma facing materials, as in ITER, is proposed. It is based on a critical analysis of relevant information obtained in the field of hard CN film deposition and consists of the use of chemical scavengers for the inhibition of tritium-rich carbon-film formation in hidden areas of the divertor. The practical implementation of the technique, however, requires a detailed knowledge of the physio-chemical processes involved, and, to date, experiments in cold and divertor plasmas have been performed. Very recent experiments in the ASDEX Upgrade device have shown that the injection of nitrogen in the sub-divertor region can lead to a drastic decrease in the level of deposited material with no significant effects in the performance of the main plasma. This and other findings are interpreted in the light of recent results from laboratory and divertor plasma experiments and the extrapolation to new divertor scenarios is discussed.

The edge parameters of electron cyclotron resonance heated plasmas in the TJ-II stellarator are reported. Data from atomic beam diagnostics and electrical probes have been used for edge and scrape-off layer characterization. Scans in heating power and plasma density for H and He plasmas have been performed, for a given magnetic configuration. A linear increase of the diffusion coefficient at the last-closed magnetic surface with the ratio of injected power to plasma density and a similar value of that parameter for the two atomic species investigated were obtained. Global particle confinement times between 3 and 15 ms have been deduced, and transition to an enhanced confinement mode in H plasmas has been observed under some conditions. The role of high-energy particle losses, due to trapping into the relatively high magnetic ripple, in the global energy balance of TJ-II plasmas is addressed.