Joint European Torus Experiment Research Articles

Neural network surrogates of Bayesian diagnostic models for fast inference of plasma parameters.

We present a framework for training artificial neural networks (ANNs) as surrogate Bayesian models for the inference of plasma parameters from diagnostic data collected at nuclear fusion experiments, with the purpose of providing a fast approximation of conventional Bayesian inference. Because of the complexity of the models involved, conventional Bayesian inference can require tens of minutes for analyzing one single measurement, while hundreds of thousands can be collected during a single plasma discharge. The ANN surrogates can reduce the analysis time down to tens/hundreds of microseconds per single measurement. The core idea is to generate the training data by sampling them from the joint probability distribution of the parameters and observations of the original Bayesian model. The network can be trained to learn the reconstruction of plasma parameters from observations and the model joint probability distribution from plasma parameters and observations. Previous work has validated the application of such a framework to the former case at the Wendelstein 7-X and Joint European Torus experiments. Here, we first give a description of the general methodological principles allowing us to generate the training data, and then we show an example application of the reconstruction of the joint probability distribution of an effective ion charge Zeff-bremsstrahlung model from data collected at the latest W7-X experimental campaign. One key feature of such an approach is that the network is trained exclusively on data generated with the Bayesian model, requiring no experimental data. This allows us to replicate the training scheme and generate fast, surrogate ANNs for any validated Bayesian diagnostic model.

Gyrokinetic simulations of toroidal Alfvén eigenmodes excited by energetic ions and external antennas on the Joint European Torus

The gyrokinetic toroidal code (GTC) has been used to study toroidal Alfvén eigenmodes (TAEs) in high-performance plasmas. Experiments performed at the Joint European Torus (JET), where TAEs were driven by energetic particles arising from neutral beams, ion cyclotron resonant heating, and resonantly excited by dedicated external antennas, have been simulated. Modes driven by populations of energetic particles are observed, matching the TAE frequency seen with magnetic probes in JET experiments. A synthetic antenna, composed of one toroidal and two neighboring poloidal harmonics has been used to probe the modes’ damping rates and quantify mechanisms for this damping in GTC simulations. This method was also applied to frequency and damping rate measurements of stable TAEs made by the Alfvén eigenmode active diagnostic in these discharges.

The effects of impurities and core pressure on pedestal stability in Joint European Torus (JET)a)

The H-mode pedestal plays an important role in determining global confinement in tokamaks. In high triangularity H-mode experiments in Joint European Torus with the ITER-like wall (JET-ILW), significantly higher pedestal temperature and global confinement have been achieved with nitrogen seeding. The experimentally observed increase in pedestal height is inconsistent with the stability calculations using the experimental profiles. Numerically, we find that the consistency with stability improvement can be restored if we assume a shift of the pedestal inwards and increased ion dilution due to the impurity seeding. Significantly better confinement and pedestal height have been observed in JET-ILW plasmas when the core pressure is increased. The enhanced pedestal height can be linked to an improvement in edge stability arising from an increase in the Shafranov-shift, higher edge current, and pedestal widening in flux space.

Low-recycling conditions and improved core confinement in steady-state operation scenarios in JET (Joint European Torus)

In this paper, we discuss the phenomena that link particle recycling from the vessel walls in the L-mode during discharge start-up and the core confinement in the H-mode during the subsequent main heating phase. We consider available data of JET experiments that aimed at approaching fully non-inductive ITER-relevant steady-state conditions and show that the high electron temperature produced at the edge by a low recycling during start-up tends to favour the build-up of high normalized β(βN) regimes in the H-mode, the confinement being improved in a large plasma volume. To provide an insight into this complex phenomenon we have modelled the relation between particle recycling in the scrape-off layer and the evolution of plasma transport, plasma current density and shear as well as the stability properties for such experimental conditions. The results confirm the existence of a link between the confinement in the H-mode phase and the values at the edge of electron temperature, bootstrap current density and local magnetic shear during start-up. Such a link could favour these regimes to be self-sustained in time.

Conceptual Design, Development and Preliminary Tests of a Compact Neutron Spectrometer for the JET Experiment

A Compact Neutron Spectrometer (CNS) has been developed to measure the neutron emission spectra in Joint European Torus (JET) fusion plasma experiments. The spectrometer, based on a liquid scintillation detector (BC501A), is equipped with a Digital Pulse Shape Discrimination (DPSD) acquisition system for neutron (n) and gamma-ray (γ) separation. The CNS enables recording the n and γ pulse height spectra (PHS) up to total count rates of ~10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sup> s <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-1</sup> . Energy resolution, after PHS unfolding, will be <; 2% for 14 MeV neutrons and <; 4% for 2.5 MeV neutrons. The work done by ENEA-Frascati and Physikalisch-Technische Bundesanstalt (PTB) respectively in the assembly and test of DPSD and scintillation detector, along with the first results obtained by the spectrometer in JET plasma experiments are presented. The experience obtained with CNS in JET will contribute to the development of neutron spectrometers suitable for applications in the International Thermonuclear Experimental Reactor (ITER).

Initial results from the lost alpha diagnostics on Joint European Torus

Two devices have been installed in the Joint European Torus (JET) vacuum vessel near the plasma boundary to investigate the loss of energetic ions and fusion products in general and alpha particles in particular during the upcoming JET experiments. These devices are (i) a set of multichannel thin foil Faraday collectors and (ii) a well collimated scintillator, which is optically connected to a charge-coupled device. Initial results including the radial energy and poloidal dependence of lost ions from hydrogen and deuterium plasmas during the 2005-06 JET restart campaign will be presented.

Simulation studies of the role of reconnection in the “current hole” experiments in the Joint European Torus

Injection of lower hybrid current drive into the current ramp-up phase of Joint European Torus (JET) discharges has been observed to produce an annular current distribution with a core region of essentially zero current density [Hawkes et al., Phys. Rev. Lett. 87, 115001 (2001)]. Similar “current holes” have been observed in Japan Atomic Energy Research Institute Tokamak 60 Upgrade discharges with off-axis current drive supplied by bootstrap current [T. Fujita et al., Phys. Rev. Lett. 87, 245001 (2001)]. In both cases, the central current does not go negative, although current diffusion calculations indicate that there is sufficient noninductive current drive for this to occur. This is explained by the multilevel 3D code (M3D) nonlinear 2D and 3D resistive magnetohydrodynamic (MHD) simulations in toroidal geometry, which predict that these discharges undergo n=0 reconnection events—“axisymmetric sawteeth”—that redistribute the current to hold its core density near zero. Unlike conventional sawteeth, these events retain the symmetry of the equilibrium, and thus are best viewed as a transient loss of equilibrium caused when an ι=0 rational surface enters the plasma. If the current density profile has a central minimum, this surface will enter on axis; otherwise it will enter off-axis. In the first case, the reconnection is limited to a small region around the axis and clamps the core current at zero. In the second case, more typical of the JET experiments, the core current takes on a finite negative value before the ι=0 surface appears, resulting in discrete periodic axisymmetric sawtooth events with a finite minor radius. Interpretation of the simulation results is given in terms of analytic equilibrium theory, and the relation to conventional sawteeth and to a recent reduced-MHD analysis of this phenomenon in cylindrical geometry [Huysmans et al., Phys. Rev. Lett. 87, 245002 (2001)] is discussed.

Neutron spectrometry at JET (1983–1999)

An account is given of the neutron spectrometric measurements on tokamak plasmas that have been performed at the Joint European Torus (JET) Joint Undertaking. The original restrictions for physical access to the tokamak and the performance projections are described. The actual characteristics of JET plasmas as intense but highly transient sources of neutrons are then presented. Next, the various neutron spectrometers that have been deployed at JET are listed and their success in meeting the demands of the JET experiment is appraised. Finally, there is a discussion of the plasma physics considerations that determine the detailed shapes of the d–d and d–t spectral lines under the various plasma conditions and spectrometer viewing directions and of the results obtained.

Jet tritium inventory-control and measurement during DT, shutdown and DD operational phases

DT plasma operations took place at Joint European Torus (JET) from May to November 1997. The site tritium inventory was 20 g, and this was recycled several times using the active gas handling system (AGHS) to recover tritium from mixed exhaust gases. The amount of tritium available for use at any time was determined by the hold-up in the in-vessel components, in-process inventory, and losses via permeation and transfer to waste. It was therefore necessary, to continually assess the tritium inventory through on-line and batch measurements. During the shutdown periods, tritium continued to be evolved from the machine and its trapped volumes. This meant that a large number of specific measurements were necessary to ensure that the safety requirements were met and that the waste could be properly characterised. During DD operations, the exhaust gas from the machine contained quantities of tritium, too great to be discharged to atmosphere and must continue to be processed by the AGHS. The techniques for tritium inventory assessment are reviewed, taking into account continuing uncertainties and plans for resolutions of these. The JET experiment is used to highlight issues of tritium inventory control, which will apply to future fusion reactors.

Comparison of hydrogen and tritium uptake and retention in JET

During previous Joint European Torus (JET) deuterium-to-tritium change-over experiments, subdivertor tritium concentrations were compared with those measured at the strike point region and found to differ significantly during the first few discharges, which was correlated with wall saturation. New deuterium-to-hydrogen fueling experiments in JET have been made and are compared to these previous experiments. Rates of hydrogenic species exchange are similar to those found in previous tritium experiments, granting differences in divertor configuration and mass ratio. In the new experiments, measurements of the CD and CH molecular band intensities near the divertor strike point monitor an intermediate stage of particle exchange between the plasma and wall. The CD/CH ratio correlates well with both the plasma and subdivertor concentration. The neutral transport code EIRENE and the wall hydrogen trapping and diffusion code WDIFFUSE have been used to evaluate the wall saturation. It appears that chemically-related processes play a role in mediating the plasma–wall exchange.

Effect of isotope mass on transport simulations of Joint European Torus high-mode plasmas with Edge Localized Modes

The effect of isotopic mass on heat and particle transport in Joint European Torus (JET) [P.-H. Rebut et al., Nucl. Fusion 25, 1011 (1985)] plasma discharges is studied using the Multi-Mode model in the BALDUR predictive transport code [Bateman et al., Phys. Plasmas 5, 1793 (1998)]. Temperature and density profiles from these simulations generally agree with the experimentally measured profiles for high-mode JET discharges with Edge Localized Modes in hydrogen, deuterium, and tritium discharges. It is surprising that a purely gyro-Bohm transport model, used in these simulations, correctly predicts the experimentally observed improvement in confinement as the isotope mass is increased—given the fact that gyro-Bohm diffusion coefficients increase with isotope mass when the shapes of all the plasma profiles are held fixed. However, in the JET experiment, it was found that the electron and ion temperature at the top of the edge pedestal increases systematically as the isotope mass in increased (J. G. Cordey et al., Report No. JET-P (98)53, 1998). The numerical simulations reported here show that this increase in the edge temperatures and subsequent broadening of the temperature profiles account for the improvement in confinement as the isotope mass is increased.

JET Takes a Step Closer to Break-Even

NUCLEAR FUSION R esearchers at the Joint European Torus (JET), the European fusion test reactor at Abingdon near Oxford, announced last week that they have come closer than ever before to break-even, the point at which a fusion reactor produces as much energy as it consumes. Burning the same mixture of hydrogen isotopes that would fuel an actual fusion power station, JET produced 50% of the energy supplied to the reactor—close to twice the previous record. ![Figure][1] Fusion first. JET sets records with a 50-50 mix of deuterium and tritium fuel. JET In a fusion reactor, deuterium or tritium nuclei (hydrogen atoms with one and two extra neutrons in the nucleus, respectively) fuse and form helium nuclei and neutrons. A small amount of mass is lost in the reaction and is converted into energy. However, fusion takes place only when the deuterium and tritium nuclei are confined inside a torus by powerful magnetic fields and heated to temperatures of over 100 million degrees. Heating the gas to such temperatures with beams of neutral particles, radio waves, and electric currents running through the plasma requires a huge amount of energy. Until recently, experimental fusion reactors all used a fuel consisting mainly of deuterium and a smaller proportion of tritium (about 11%), which yields less energy. In 1995, however, the Tokamak Fusion Test Reactor (TFTR) in Princeton, New Jersey, burned a 50-50 deuterium/tritium mixture to produce 28% of the energy needed to heat its plasma. Now JET, also burning an equal mix of deuterium and tritium, has moved halfway to break-even, in the process generating more than 12 megawatts of fusion power. “We hope to improve this [output/input ratio] somewhat over the next weeks,” says JET director Martin Keilhacker. JET deputy director Alan Gibson explains that they plan to confine the plasma within several different magnetic field configurations: “So far, we have only used the first of these, and some of the others we think probably have promise to go to higher values of this ratio.” Gibson says that the output of TFTR, which was shut down earlier this year, was limited by geometry. “The shape of the cross section of the plasma in JET is D-shaped, while in the TFTR the cross section is circular,” he explains. “The D shape has advantages from the point of view of stability,” Gibson adds, and allows the use of a diverter, a circular chamber in the bottom of the torus vessel that siphons off impurities that would otherwise slow the reaction. Gibson believes that it may be possible to reach break-even with JET, which will operate until at least 1999. “It depends on how successful we are in fine-tuning the plasma behavior,” he says. Experiments with an output/input ratio closer to one are interesting because “the self-heating in the plasma by the fusion power becomes progressively more important, and one of the things we plan to do at JET is to measure this self-heating,” Gibson says. The results so far, continues Gibson, “are very encouraging for ITER”—the proposed $10 billion International Thermonuclear Experimental Reactor, which, if built, would use the same fuel mixture and would also use a D-shaped plasma and a diverter. However, Michael Mauel of Columbia University, who works on General Atomics' DIII-D tokamak in San Diego, stresses that today's tokamak experiments still aim at improving the physics of fusion rather than demonstrating fusion power. “The importance of the JET experiment is not to demonstrate ignition, but to demonstrate that we understand the physics of plasma at 100 million degrees,” he says. “And what is so exciting is that our expectations and our understanding appear to be right.” Other fusion experts are also excited by the news. “It is wonderful that they have achieved high performance, and I hope they continue their experiments,” says William Dorland, a researcher at the Institute of Fusion Studies of the University of Texas, Austin. “It is too soon to know what it means for ITER, but it could be good.” [1]: pending:yes

Magnetohydrodynamic spectroscopy: Large scale computation of the spectrum of waves in plasmas

A systematic study of the spectrum of magnetohydrodynamic (MHD) waves in thermonuclear and astrophysical plasmas is presently undertaken at the FOM Institute for Plasma Physics. This study is based on the fact that the macroscopic waves and instabilities in both thermonuclear confinement machines and in astrophysical magnetic flux loops are described by the MHD equations. The linearised interaction between plasma and magnetic field leads to large-scale non-symmetric eigenvalue problems for free oscillations and to temporal evolution problems for forced (externally driven) oscillations. These problems have been cast in the form of a set of two modular computer programs, called CASTOR for tokamaks and POLLUX for coronal flux tubes, exploiting identical solvers for the resulting very large systems. CASTOR is extensively used at JET (the Joint European Torus experiment at Culham) for the interpretation of measured MHD spectra, whereas POLLUX is presently used for the study of Alfvén wave heating of the solar corona and will be used in the near future for the interpretation of observed x-ray emissions from the SOHO satellite which will be launched by ESA and NASA in 1995. For the interpretation of actual plasma dynamics in tokamaks and solar coronal loops it is crucial to understand the non-linear phase of the evolution of MHD phenomena as well. This adds two complications: non-linear mode coupling of hundreds of small-scale modes and enormous discrepancies between the different dynamical time scales. We have developed numerical codes that simulate the non-linear temporal evolution of plasma columns under external excitation. These codes have been modelled after the linear ones, with a finite difference discretisation in the radial direction and a spectral discretisation in the two other directions. A relatively simple and flexible semi-implicit predictor-corrector scheme has been used for the time stepping. This allows time steps up to a factor 1000 larger than the largest time steps allowed for explicit methods. Because of the many modes involved, memory requirements and CPU time requirements soon become prohibitive. Hence, the new physics in the non-linear plasma dynamics regime requires parallel computing.

Pellet penetration experiments on JET

The measured penetration depth lambda of deuterium pellets injected into the Joint European Torus (JET) confirms some features of the neutral gas shielding (NGS) model, but not others. The scaling of lambda with plasma and pellet parameters agrees with the NGS model, as in earlier ASDEX studies. Pellet velocity was varied over the range 0.46-1.35 km/s in the JET experiments to test specifically the scaling of lambda with velocity. This scaling also agrees with the NGS model. However, the penetration is deeper in JET than in ASDEX when it is corrected for the expected machine size dependence. Furthermore, the penetration depths measured in JET are greater (by nearly a factor of two) than those predicted by local ablation calculations using the NGS model with an incident Maxwellian distribution of electrons. Plasma shielding used in previous modelling of the JET penetration data can account for the additional shielding, but it also removes the observed velocity dependence. The implications of both the scaling observations and the penetration depths for improvements in ablation theory and in the models are discussed

JET experiments with 120 keV 3He and 4He neutral beam injection and neutron diagnostic applications

Preliminary experiments were carried out in the Joint European Torus (JET) using 120 keV helium (He) neutral beam injection (NBI). Injected power levels up to 5 MW with 3He and 7 MW with 4He, lasting up to 3 s were reached. The 3 s helium NBI produced efficient ion heating and similar global and local energy confinements to those obtained with deuterium (D) NBI, in both limiter and X-point plasma geometries, in L-mode and H-mode plasma regimes. The elimination of beam-plasma and beam-beam fusion reactions by replacing D NBI with He NBI extended the range for measuring ion temperatures with the JET neutron profile monitor and neutron spectrometers.

A Comparison of Hydrogenic Pellet Ablation Models with Experiment

This paper reports on several theories developed over the past 15 years to describe the ablation of a solid hydrogenic pellet injected into a hot plasma. The most widely accepted theory is the neutral gas shielding model. This model has been expanded to include ablation by fast ions (as well as electrons), realistic particle distribution functions, self-limiting ablation, and a cold ionized plasma shield beyond the ablating gas. Ablation measurements, including absolute pellet penetration and ablation profiles, from the Impurity Study Experiment, Poloidal Divertor Experiment, Doublet-III, Alcator-C, Tokamak Fontenay-aux-Roses, T-10, Texas Experimental Tokamak, Tokamak Fusion Test Reactor, and Joint European Torus experiments are compared with variations of the neutral gas shielding model under a range of input assumptions.

Ion-cyclotron-frequency stabilization of internal kink mode and sawtooth oscillations in tokamaks.

It is proposed that the ponderomotive force due to applied ion-cyclotron resonance-frequency waves can stabilize the internal kink mode in tokamaks. The sufficient stability criterion is derived and the necessary power estimated. It is concluded that at the rf power level, present in the Joint European Torus experiment, the ponderomotive force effects are significant and may be responsible for the modification of the sawtooth activity observed in recent experiments.

The JET Experiment: Evolution, Present Status, and Prospects

The background to the decision to build the Joint European Torus (JET) is described, and a brief introduction to the main aims, overall design philosophy, and the planned parameter range of the lar...