Demonstrating improved confinement of energetic ions is one of the key goals of the Wendelstein 7-X (W7-X) stellarator. In the past campaigns, measuring confined fast ions has proven to be challenging. Future deuterium campaigns would open up the option of using fusion-produced neutrons to indirectly observe confined fast ions. There are two neutron populations: 2.45 MeV neutrons from thermonuclear and beam-target fusion, and 14.1 MeV neutrons from DT reactions between tritium fusion products and bulk deuterium. The 14.1 MeV neutron signal can be measured using a scintillating fiber neutron detector, whereas the overall neutron rate is monitored by common radiation safety detectors, for instance fission chambers. The fusion rates are dependent on the slowing-down distribution of the deuterium and tritium ions, which in turn depend on the magnetic configuration via fast ion orbits. In this work, we investigate the effect of magnetic configuration on neutron production rates in W7-X. The neutral beam injection, beam and triton slowing-down distributions, and the fusion reactivity are simulated with the ASCOT suite of codes. The results indicate that the magnetic configuration has only a small effect on the production of 2.45 MeV neutrons from DD fusion and, particularly, on the 14.1 MeV neutron production rates. Despite triton losses of up to 50 %, the amount of 14.1 MeV neutrons produced might be sufficient for a time-resolved detection using a scintillating fiber detector, although only in high-performance discharges.

After completing the main construction phase of Wendelstein 7-X (W7-X) and successfully commissioning the device, first plasma operation started at the end of 2015. Integral commissioning of plasma start-up and operation using electron cyclotron resonance heating (ECRH) and an extensive set of plasma diagnostics have been completed, allowing initial physics studies during the first operational campaign. Both in helium and hydrogen, plasma breakdown was easily achieved. Gaining experience with plasma vessel conditioning, discharge lengths could be extended gradually. Eventually, discharges lasted up to 6 s, reaching an injected energy of 4 MJ, which is twice the limit originally agreed for the limiter configuration employed during the first operational campaign. At power levels of 4 MW central electron densities reached 3  ×  1019 m−3, central electron temperatures reached values of 7 keV and ion temperatures reached just above 2 keV. Important physics studies during this first operational phase include a first assessment of power balance and energy confinement, ECRH power deposition experiments, 2nd harmonic O-mode ECRH using multi-pass absorption, and current drive experiments using electron cyclotron current drive. As in many plasma discharges the electron temperature exceeds the ion temperature significantly, these plasmas are governed by core electron root confinement showing a strong positive electric field in the plasma centre.

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.

Predictive capabilities better than 95%, and very limited false alarms, are demanding requirements for reliable disruption prediction systems in tokamaks such as JET or, in the near future, ITER. The prediction of an upcoming disruption has to be provided sufficiently in advance in order to apply effective disruption avoidance or mitigation actions preventing the machine to be damaged. In this paper, following the typical machine learning workflow, a Generative Topographic Mapping (GTM) of the operational space of JET has been built using a set of disrupted and regularly terminated discharges. In order to build the predictive model, a suitable set of dimensionless, machine-independent, physics-based features have been synthesized, which make use of 1D plasma profiles information, rather than simple zero-D time series. The use of such predicting features, together with the power of the GTM in fitting the model to the data, allows obtaining, in an unsupervised way, a 2-dimensional map of the multi-dimensional parameter space of JET, where it is possible to identify a boundary separating the region free from disruption from the disruption region. In addition to helping in operational boundaries studies, the GTM map can also be used for disruption prediction exploiting the potentiality of the developed GTM toolbox to monitor the discharge dynamics. Following the trajectory of a discharge on the map throughout the different regions, an alarm is triggered depending on the disruption risk of these regions. The proposed approach to predict disruptions has been evaluated on a training and an independent test set, allowing to achieve very good performance with only one tardive detection and a limited number of false detections. The warning times are suitable for avoidance purposes and, more important, the detections are consistent with physics causes and mechanisms that destabilize the plasma leading to disruptions.

Neural networks are trained to evaluate the risk of plasma disruptions in a tokamak experiment using several diagnostic signals as inputs. A saliency analysis confirms the goodness of the chosen inputs, all of which contribute to the network performance. Tests that were carried out refer to data collected from succesfully terminated and disruption terminated pulses performed during two years of JET tokamak experiments. Results show the possibility of developing a neural network predictor that intervenes well in advance in order to avoid plasma disruption or mitigate its effects.