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.

No abstract

A disruption prediction system, based on neural networks, is presented in this paper. The system is ideally suitable for on-line application in the disruption avoidance and/or mitigation scheme at the JET tokamak.A multi-layer perceptron (MLP) predictor module has been trained on nine plasma diagnostic signals extracted from 86 disruptive pulses, selected from four years of JET experiments in the pulse range 47830–57346 (from 1999 to 2002).The disruption class of the disruptive pulses is available. In particular, the selected pulses belong to four classes (density limit/high radiated power, internal transport barrier, mode lock and h-mode/l-mode).A self-organizing map has been used to select the samples of the pulses to train the MLP predictor module and to determine its target, increasing the prediction capability of the system.The prediction performance has been tested over 86 disruptive and 102 non-disruptive pulses. The test has been performed presenting to the network all the samples of each pulse sampled every 20 ms. The missed alarm rate and the false alarm rate of the predictor, up to 100 ms prior to the disruption time, are 23% and 1%, respectively.Recent plasma configurations might present features different from those observed in the experiments used in the training set. This ‘novelty’ can lead to incorrect behaviour of the predictor. To improve the robustness and reliability of the system, a novelty detection module has been integrated in the prediction system, increasing the system performance and resulting in a missed alarm rate reduced to 7% and a false alarm rate reduced to 0%.

A power-balance model, with radiation losses from impurities and neutrals, gives a unified description of the density limit (DL) of the stellarator, the L-mode tokamak, and the reversed field pinch (RFP). The model predicts a Sudo-like scaling for the stellarator, a Greenwald-like scaling, , for the RFP and the ohmic tokamak, a mixed scaling, , for the additionally heated L-mode tokamak. In a previous paper (Zanca et al 2017 Nucl. Fusion 57 056010) the model was compared with ohmic tokamak, RFP and stellarator experiments. Here, we address the issue of the DL dependence on heating power in the L-mode tokamak. Experimental data from high-density disrupted L-mode discharges performed at JET, as well as in other machines, are taken as a term of comparison. The model fits the observed maximum densities better than the pure Greenwald limit.

The JET 2019-2020 scientific and technological programme exploited the results of years of concerted scientific and engineering work, including the ITER-like wall (ILW: Be wall and W divertor) installed in 2010, improved diagnostic capabilities now fully available, a major Neutral Beam Injection (NBI) upgrade providing record power in 2019-2020, and tested the technical & procedural preparation for safe operation with tritium. Research along three complementary axes yielded a wealth of new results. Firstly, the JET plasma programme delivered scenarios suitable for high fusion power and alpha particle physics in the coming D-T campaign (DTE2), with record sustained neutron rates, as well as plasmas for clarifying the impact of isotope mass on plasma core, edge and plasma-wall interactions, and for ITER pre-fusion power operation. The efficacy of the newly installed Shattered Pellet Injector for mitigating disruption forces and runaway electrons was demonstrated. Secondly, research on the consequences of long-term exposure to JET-ILW plasma was completed, with emphasis on wall damage and fuel retention, and with analyses of wall materials and dust particles that will help validate assumptions and codes for design & operation of ITER and DEMO. Thirdly, the nuclear technology programme aiming to deliver maximum technological return from operations in D, T and D-T benefited from the highest D-D neutron yield in years, securing results for validating radiation transport and activation codes, and nuclear data for ITER.

No abstract