Abstract:JET underwent a transformation from a full carbon-dominated tokamak to a full metallic device with the ITER-like wall combination for the activated phase with Beryllium main chamber and Tungsten divertor. The ITER-Like Wall (ILW) experiment at JET provides an ideal test bed for ITER and shall demonstrate as primary goals the plasma compatibility with metallic walls and the reduction in fuel retention. We report on a set of experiments ( = 2.0 , = 2.0 − 2.4 , = 0.2 − 0.4) in different confinement and plasma conditions with global gas balance analysis demonstrating a strong reduction of the long term retention rate by a factor ten with respect to carbon references. All experiments have been executed in a series of identical plasma discharges in order to achieve maximum plasma duration until the analysis limit of the active gas handling system has been reached. The composition analysis shows high purity of the recovered gas, typically 99% D. For typical L-mode discharges ( = 0.5 ), type III ( = 5.0 ), and type I ELMy H-mode plasmas ( = 12.0 ) a drop of the retention rate normalised to the operational time in divertor configuration has been measured from 1.27 × 10 has been obtained with the ILW. The observed reduction by one order of magnitude confirms the expected predictions concerning the plasma-facing material change in ITER and widens the operation without active cleaning in the DT phase in comparison to a full carbon device.

Disruptions are a major operational concern for next generation tokamaks, including ITER. They may generate excessive heat loads on plasma facing components, large electromagnetic forces in the machine structures and several MA of multi-MeV runaway electrons. A more complete understanding of the runaway generation processes and methods to suppress them is necessary to ensure safe and reliable operation of future tokamaks. Runaway electrons were studied at JET-ILW showing that their generation dependencies (accelerating electric field, avalanche critical field, toroidal field, MHD fluctuations) are in agreement with current theories.

Real-time simultaneous control of several radially distributed magnetic and kinetic plasma parameters is being investigated on JET, in view of developing integrated control of advanced tokamak scenarios. This paper describes the new model-based profile controller which has been implemented during the 2006–2007 experimental campaigns. The controller aims to use the combination of heating and current drive (H&CD) systems—and optionally the poloidal field (PF) system—in an optimal way to regulate the evolution of plasma parameter profiles such as the safety factor, q(x), and gyro-normalized temperature gradient, . In the first part of the paper, a technique for the experimental identification of a minimal dynamic plasma model is described, taking into account the physical structure and couplings of the transport equations, but making no quantitative assumptions on the transport coefficients or on their dependences. To cope with the high dimensionality of the state space and the large ratio between the time scales involved, the model identification procedure and the controller design both make use of the theory of singularly perturbed systems by means of a two-time-scale approximation. The second part of the paper provides the theoretical basis for the controller design. The profile controller is articulated around two composite feedback loops operating on the magnetic and kinetic time scales, respectively, and supplemented by a feedforward compensation of density variations. For any chosen set of target profiles, the closest self-consistent state achievable with the available actuators is uniquely defined. It is reached, with no steady state offset, through a near-optimal proportional-integral control algorithm. Conventional optimal control is recovered in the limiting case where the ratio of the plasma confinement time to the resistive diffusion time tends to zero. Closed-loop simulations of the controller response have been performed in preparation for experiments, and typical results are shown. Finally, in the last section of the paper, the first experimental results using this dynamic-model approach to control the plasma current and the safety factor profile on JET, either with the three H&CD systems or also with the PF system as an additional actuator, are presented and discussed.

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.