Search

As part the DTE2 campaign in the JET tokamak, we conducted a parameter scan in T and D-T complementing existing pulses in H and D. For the different main ion masses, type-I ELMy H-modes at fixed plasma current and magnetic field can have the pedestal pressure varying by a factor of 4 and the total pressure changing from β N = 1.0 to 3.0. We investigated the pedestal and core isotope mass dependencies using this extensive data set. The pedestal shows a strong mass dependence on the density, which influences the core due to the strong coupling between both plasma regions. To better understand the causes for the observed isotope mass dependence in the pedestal, we analysed the interplay between heat and particle transport and the edge localised mode (ELM) stability. For this purpose, we developed a dynamic ELM cycle model with basic transport assumptions and a realistic neutral penetration. The temporal evolution and resulting ELM frequency introduce an additional experimental constraint that conventional quasi-stationary transport analysis cannot provide. Our model shows that a mass dependence in the ELM stability or in the transport alone cannot explain the observations. One requires a mass dependence in the ELM stability as well as one in the particle sources. The core confinement time increases with pedestal pressure for all isotope masses due to profile stiffness and electromagnetic turbulence stabilisation. Interestingly, T and D-T plasmas show an improved core confinement time compared to H and D plasmas even for matched pedestal pressures. For T, this improvement is largely due to the unique pedestal composition of higher densities and lower temperatures than H and D. With a reduced gyroBohm factor at lower temperatures, more turbulent drive in the form of steeper gradients is required to transport the same amount of heat. This picture is supported by quasilinear flux-driven modelling using TGLF -SAT2 within Astra. With the experimental boundary condition TGLF -SAT2 predicts the core profiles well for gyroBohm heat fluxes > 15 , however, overestimates the heat and particle transport closer to the turbulent threshold. [ABSTRACT FROM AUTHOR]

Kinetic profile predictions of ITER PFPO-1 plasmas require high accuracy in the central electron temperatures to be applied to the calculation of third harmonic electron cyclotron absorption. Correctly predicting the transition from L-mode to H-mode further requires precise estimates of the ion heat flux in the periphery of the plasma. Recent versions of the quasi-linear transport models TGLF and QuaLiKiz were tested against an extensive set of experimental results from ASDEX Upgrade (AUG) and JET-ILW, where the focus was put on AUG plasmas heated by ECRH. Spectra obtained from TGLF are compared to a set of linear gyrokinetic simulations performed with GKW. Electron and ion temperature profiles obtained with TGLF-SAT1geo show good agreement with the experimental profiles, but there is a slight tendency to underpredict central Te and Ti at high ratios Te/Ti. QuaLiKiz yields reasonable results for Te and Ti profiles in plasmas where the ion temperature gradient mode is dominant, but predicts a significantly too weak transport in the presence of dominant trapped electron modes in conditions of strong central electron heating. [ABSTRACT FROM AUTHOR]

The reproducibility of diamagnetic flux measurements in vacuum field shots for the internal and external diamagnetic flux measurements in the ASDEX Upgrade tokamak is investigated. A comparison of diamagnetic flux measurements and predictions from equilibrium reconstruction, including pressure constraints from thermal and fast-ion contributions, is reported. The ideal magnetohydrodynamic model involved in interpreting diamagnetic flux measurements in a tokamak with anisotropic pressure is summarised. The plasma energy calculated from equilibrium reconstruction and inferred from diamagnetic flux measurements is compared with the respective values calculated by modelling with TRANSP and IDE/RABBIT. It is found that at low densities, in discharges with parallel neutral beam sources, the plasma energy inferred from the diamagnetic flux measurements is smaller than the plasma energy calculated from the equilibrium reconstruction. In addition, details of the modelling are discussed to stress the interdependence of the various modelling inputs and the interpretation of the results. A comparison of diamagnetic flux measurements with calculations from equilibrium reconstruction by IDE and fast-ion profiles from RABBIT and TRANSP simulations can identify anisotropic plasma discharges and provide a consistency check of the input data for modelling. [ABSTRACT FROM AUTHOR]

The I-mode confinement regime in non-deuterium plasmas has been investigated in the ASDEX Upgrade tokamak. We report the first experimental observations on the existence and the main characteristics of this regime in hydrogen and helium plasmas and compare them with deuterium I-modes. Hydrogen features a higher power threshold to enter I-mode and a higher electron edge pressure at the L- to I-mode transition with respect to deuterium. Furthermore, all the hydrogen I-modes obtained exhibit pedestal relaxation events (PRE). The I-mode power window in hydrogen is found to be 2–3 times larger than in deuterium. This property allows a better characterization of PRE and to differentiate them from type-III ELMs. Helium I-modes feature properties which are similar to those of deuterium I-modes. The analysis on the minimum of the edge radial electric field Er shows a correlation between the Er minimum, the net input power and the ion diamagnetic term in the ion radial force balance. Indications of the dominant role of the edge ion temperature in the evolution of the radial electric field with increasing input power are also reported. [ABSTRACT FROM AUTHOR]

Measurements using a recently installed edge fast-ion D-alpha (FIDA) diagnostic at the ASDEX Upgrade tokamak show a clear effect of edge localised modes (ELMs) on the passive FIDA signals. While a reduction in the passive FIDA emission is observed in the scrape-off layer (SOL) region, measurements close to the last closed flux surface show an increase in signals shortly after ELMs, followed by a decrease. The decrease provides a clear sign of fast-ion losses in the SOL, while the increase can be explained by an enhanced neutral density during ELMs inside the plasma. In addition, small ELMs are observed, which barely change the neutral density and plasma position but still cause significant changes in the passive FIDA signals. A comparison of the measurements with forward modelling shows that 60% to 80% of the fast ions are lost by ELMs outside the last closed flux surface. In addition, a 20% decrease of the fast-ion density in a range up to 4 cm within the last closed flux surface can be inferred. This range agrees well with the latest modelling results of ELMs using the non-linear MHD code JOREK and shows that less than 0.3% of all fast ions are lost by ELMs. [ABSTRACT FROM AUTHOR]

H-mode plasmas in ASDEX Upgrade (AUG) using different hydrogen isotopes are analysed with respect to their core transport properties. The experimental results are discussed and we present gyrokinetic simulations which are able to reproduce the experimental observations. A novel strategy allows us to disentangle core and pedestal physics by mitigating the isotopic dependence of pedestal properties while keeping the heat and particle sources the same. Matched pedestal profiles are obtained between hydrogen (H) and deuterium (D) plasmas when increasing the triangularity in H plasmas with respect to D plasmas. In the core of these plasmas little isotopic dependence is observed when the fast-ion content is low Wfast/Wth < 1/3. Quasi-linear modelling with TGLF reproduces the experimental trends under these conditions. For larger fast-ion fractions an isotope dependence is observed in the core heat transport. This is related to a difference in fast-ion stabilization of turbulent transport. The fast-ion pressure in H and D plasmas is different due to the mass dependence in the fast-ion slowing down time as well as to operational restrictions when heating with H neutral beam injection (H-NBI) or D-NBI. Typically, Wfast,H < 1/2Wfast,D for comparable NBI heating powers in AUG. The gyrokinetic analysis shows that linear growth rates of ITG modes do not show a pure gyro-Bohm mass dependence, but follow the experimentally observed mass dependence when taking collisions, EM-effects and fast ions into account. Non-linear gyrokinetic simulations reproduce the experimental heat fluxes for different isotopes when fast ions are included. This highlights the role of the fast-ion pressure as a key element to explain the observed differences in the core of H and D plasmas. [ABSTRACT FROM AUTHOR]

Confinement scaling laws such as IPB98(y, 2) are widely used to extrapolate the performance of present tokamaks to next-step devices such as ITER or DEMO. The thermal energy of the plasma (Wth), which is used to determine the energy confinement time for most scaling laws, is difficult to measure, due to the sizeable uncertainties in the experimental kinetic profiles. The common approach in the tokamak community is to derive Wth as the difference between the measured magnetohydrodynamic (MHD) energy and some simulation-based estimate of the fast ion energy Wfi. In H-mode plasmas Wfi can be as high as Wth, in presence of neutral beam injection (NBI) or ion cyclotron radio frequency heating (ICRF), therefore an accurate assessment of Wfi is crucial to have a somewhat reliable H-factor, regardless of the power-scaling of a given scaling law. In this paper we aim at evaluating the current approach to estimate Wfi, by comparing its predictions with a wide database of calculations using validated NBI codes. Systematic deviations and trends, as well as statistical scatter are discussed. We use a comprehensive database of AUG H-mode deuterium plasmas, with significant variations of plasma current, NBI power and plasma density. We neglect thereby the fast-ion losses caused by MHD modes and the synergy effect between NBI and ICRF. A new approach is proposed based on the newly developed fast NBI code RABBIT. [ABSTRACT FROM AUTHOR]

Here we present a new method that allows the determination of background neutral density profiles based on measurements from neutral particle analyzers (NPA). Bayesian optimization is used to obtain a reliable 5-parameter representation of the inferred profiles. The method has been benchmarked using forward modelling from FIDASIM against measured data. The systematic errors coming from assumptions made in the analysis were evaluated and they are lower than ≲5%. The new method has been tested using data from the ASDEX Upgrade tokamak. When using reconstructed neutral density profiles, good agreement is found between the measured NPA fluxes of neutralized fast ions with predictions based on TRANSP and FIDASIM. Moreover, a clear drop of neutral density is observed at the plasma boundary after edge localized mode (ELM) activity. As suggested by KN1D simulations, this drop is mainly due to an increase of scrape-off layer (SOL) ionization rate, resulting from higher SOL plasma densities and temperatures after the ELM crash. Moreover, the new method allows calculating the local plasma ionization source which will be important for future transport studies. [ABSTRACT FROM AUTHOR]

A confinement database with mainly electron-heated hydrogen plasmas from ASDEX Upgrade and Wendelstein 7-X was assembled. Stellarator confinement scaling expressions describe both standard discharges in the stellarator and L-mode plasmas in the tokamak similarly well and indicate a similar quality of energy confinement in both devices. While the energy confinement time in ASDEX Upgrade benefits from the smaller aspect ratio of the device, the transport coefficients in Wendelstein 7-X appear to be smaller possibly due to reduced average magnetic field curvature. A physics based confinement scaling is derived from a model that successfully describes transport in tokamaks. The dimensionally correct scaling has very similar parameter dependencies as the stellarator scalings and reproduces also the trends in the data from ITER L- and H-mode databases reasonably well. On the basis of this scaling, which represents the confinement times of the present data base, average tokamak L-mode and H-mode confinement is 7% lower and 76% higher, respectively. [ABSTRACT FROM AUTHOR]