[en] The prediction of Q=10 for ITER is based on standard ELMy H-mode performance. In recent years robust operating scenarios have been developed which extrapolate for ITER to significantly higher Q-values at full current or longer inductive pulse lengths at reduced plasma current with a non-inductive current drive fraction above 50%, called hybrid operation. In particular the stationary improved H-Mode, realized at ASDEX Upgrade in 1998, combines improved core confinement and stability with an H-mode edge. Further developments of this regime by AUG and DIII-D, its demonstration at JET and the similarity to the high-beta pol plasmas at JT-60U have proven that it is a strong candidate for ITER hybrid operation. This scenario is stationary for several current re-distribution times and is characterized by a central q above 1, low central magnetic shear, fish bones or low (m,n) neoclassical tearing modes (NTM's) replacing sawteeth, and a confinement factor with respect to H-mode of up to 1.4. (3,2) NTM's remain small, enabling routine operation up to .N 3 at ITER relevant collisionalities for a broad range of q95 =3.2 4.5. Operation at low q maximises performance, while higher q95 maximise bootstrap current fractions and pulse length. Fusion performances in terms of HP-98.N / q95 20.4 have been stationary maintained over a broad density range compared with 0.2 assumed for the standard ITER scenario. Even at densities close to the Greenwald density needed for optimum exhaust . values up to .N3.5 and a fusion performance of 0.35 were achieved. Temperature profiles are still ITG/TEM turbulence dominated while density profiles are peaked even at high densities enhancing confinement. Extrapolated to ITER one still expects from quasi-linear models peaked density profiles (R/Ln 3). A further contribution to global confinement can arise from an increase of edge pedestal pressure combined with stiff temperature profiles. A main focus is to obtain improved H-modes over the widest range of non-dimensional parameters extending to ITER values. In AUG at low collisionality, .N close to 3 and HP-98>1 were achieved for .* ranging from 7.5 to 11-10-3 at fixed q954. At the lowest .* a heating power of 20 MW was needed. This result indicates a more favourable scaling of the maximum .N to ITER compared to standard H-modes. No insurmountable problem arose from impurity accumulation despite the peaked density profiles. Using tailored heat deposition with central wave heating and broad NI deposition a compromise in density peaking was found allowing enhanced confinement and keeping the tungsten concentration below 10-5 even with tungsten coated first wall and divertor structures. Experiments with dominant ICRH heating demonstrated operation slightly below .N3 with reactor relevant heating schemes without particle input. Finally, benign type II ELMs have been combined with improved H-modes at high densities and close to double null configurations. We present the existing database and the status of theoretical understanding of the underlying transport and the self-regulating nature of the current density profile, proceeding evidently through (possibly differing) benign instabilities. (Author)

[en] Characteristics of edge-localized mode (ELM) energy and particle losses in type-I ELMy H-mode plasmas on ASDEX Upgrade are investigated by a large data set with as much as possible independent variations in engineering parameters. A statistically unbiased estimate for ELM energy and particle losses has been employed to remove the diagnostic noise as much as possible. The data set shows that the ELM energy loss decreases with the pedestal density or collisionality. However, the role of plasma shape, especially, triangularity, as an explicit parameter is revealed. Elevated triangularity leads to lower ELM frequency and larger ELM energy loss significantly exceeding that simply expected from the increased pedestal pressure in high triangularity plasmas at a fixed power. The observed larger ELM energy drop at higher triangularity involves the ELM perturbations of the electron temperature profile across an ELM that extend radially more inward, suggesting that there is a direct effect of plasma shape on ELM energy losses. It is found that the fraction of ELM loss power does not remain constant but the increased pedestal collisionality enhances the transport level between ELMs and reduces the ELM loss power. The ELM particle flux at fixed power rises with increasing gas fuelling rate. On the other hand, at fixed gas puff, an inverse proportionality between ELM frequency and ELM particle loss roughly holds despite of a large variation of power. When the particle flux near the separatrix is enhanced, the increase of ELM frequency and the reduction of ELM loss power caused by the increased collisionality lead to the reduction of ELM energy loss so that the energy balance could be sustained

[en] During the combined plasma heating with neutral beam injection (NBI) and waves in the ion cyclotron (IC) range of frequencies, the NBI fast ions are preferentially accelerated by IC waves close to the IC harmonics, as a consequence of finite Larmor radius (FLR) effects. Since the NBI fast ions are expected to have a strong influence on the wave absorption and propagation, we have implemented a NBI source in the quasilinear Fokker-Planck SSFPQL code, interfaced with the toroidal full-wave TORIC solver. In this implementation the NBI ionization sources are obtained from the output of a Monte Carlo code, such as FAFNER. The numerical scheme adopted in the TORIC-SSFPQL package allows to describe very anisotropic sources, such as NBI, and to iterate the solution of Maxwell's equation taking into account selfconsistently the fast ion tails. As a first application, we present modeling of an ASDEX-Upgrade discharge with combined NBI and ICRF heating.

[en] Large scale fluctuations in between edge localized modes (ELMs) are the main source for the scatter in plasma edge H mode profiles of electron density and temperature, as measured by high precision, high resolution Thomson scattering. These large scale fluctuations are also observed with electron cyclotron emission. They are quantitatively analysed by 2D poloidal snapshots of electron density and temperature, based on a 5 x 10 matrix of scattering volumes provided by the Thomson scattering system. Fluctuations with a quasi-periodic structure are found in a 2D snapshot with a frequency of about 61%. When interpreted as field-aligned helical structures toroidal quasi-mode numbers of 6-48 are found. The amplitudes of the fluctuations decrease with increasing quasi-mode number and edge profile gradient lengths. The amplitudes of the large scale structures in the steep gradient region are anti-correlated with the divertor D_α-intensity. The particle loss during an ELM is at least to a significant fraction due to the electron density 'blobs' observed in the scrape-off layer. The large scale fluctuations also perturb the measurement of 1D radial profiles. In the middle of the steep gradient region the perturbations are symmetric, but asymmetric both further inside (more minima) and further outside (more maxima)

[en] For identical edge pedestal conditions in plasmas with gas puff and with pellet injection, edge localized mode (ELM) mitigation during fully pellet-controlled ELMy H-modes has been demonstrated. When the ELM frequency is set to the pellet frequency, the ELM energy losses are significantly reduced at fixed pedestal collisionality, while pellets only weakly enhance the ELM loss power. The ELM affected region of the density profile is reduced by pellet injection, while the temperature perturbation profile does not change significantly. For constant source particle flux into the plasma edge at the same pedestal parameters, the reduction in convective losses per ELM during pellet-induced ELMs is balanced by increased ELM frequency and pellet fuelling of the pedestal region

[en] Linear and nonlinear gyrokinetic simulations with the GENE code [F. Jenko et al., Phys. Plasmas 7, 1904 (2000)] for tokamak edge plasmas as well as for stellarator core plasmas are presented, shedding light on the behavior of plasma microturbulence under near-separatrix or nonaxisymmetric conditions. To this aim, the required geometric coefficients are inferred directly from the magnetohydrodynamic equilibria of three different devices via the newly developed GIST code. It is found that the residual electron heat transport level in the H-mode edge can be explained in terms of high-wave-number fluctuations driven by electron temperature gradient modes. Moreover, the study of adiabatic ion temperature gradient turbulence in optimized stellarators points to the possibility of a systematic geometric optimization with respect to anomalous transport in nonaxisymmetric devices.

[en] Results of linear gyrokinetic simulations of ASDEX Upgrade [O. Gruber et al., Nucl. Fusion 39, 1321 (1999)] edge plasmas, with experimentally determined geometry and input parameters, are presented. It is found that in the near-edge region, microtearing modes can exist under conditions found in conventional tokamaks. As one enters the steep-gradient region, the growth rate spectrum is dominated--down to very low wavenumbers--by electron temperature gradient modes. The latter tend to peak near the X-point(s) and possess properties which may explain the ratios of the density and temperature gradient scale lengths that have been observed in various experiments over the last decade.

[en] In the ASDEX Upgrade tokamak, the power deposition structures on the divertor target plates during type-I edge localized modes (ELMs) have been investigated by infrared thermography. In addition to the axisymmetric strike line, several poloidally displaced stripes are resolved, identifying an ELM as a composite of several subevents. This pattern is interpreted as being a signature of the helical perturbations in the low field side edge during the non-linear ELM evolution. Based on this observation, the ELM related magnetic perturbation in the midplane can be derived from the target load pattern. In the start phase of an ELM collapse, average toroidal mode numbers around n ∼ 3-5 are found evolving to values of n ∼ 12-14 during the ELM power deposition maximum. Further information about the non-linear evolution of the ELM mode structure is obtained from statistical analyses of the spatial distribution, heat flux amplitudes and number of single stripes

[en] The EM induced power load on the divertor occurring in the type-I ELMy H-mode reference scenario is one of the main areas of concern for ITER. Extrapolating the results of present day tokamaks, an acceptable lifetime for the divertor cannot be realised. ELM mitigation by externally induced ELMS aim on a modification of the ELM frequency f_ELM, causing a reduction of the ELM energy ΔW_ELM by enhancing f_ELM. Various methods, usually targeting the ELM drive terms edge pressure gradient p_edge and current j_edge, have been tested. Some have shown to trigger prompt ELMs, establish f_ELM as a free parameter and mitigate ELMs according to the relation f_ELM x ΔW_ELM= const. The deleterious impact of ELM mitigation on the confinement is quite small, a relation for the plasma energy W∼f''-0.2_ELM was found. This is significantly less than W∼ f''-0.2_ELM observed in the case of intrinsic ELMs. Meanwhile, ELM frequency control already became a part of the toolkit for plasma control and is incorporated e. g. into integrated plasma scenarios as candidate regimes for ITER. Detailed investigations are performed at ASDEX Upgrade aiming to resolve the following puzzles connected with ELM trigger attempts relying on particle injection. This is the impact of the different trigger attempts relying on particle injection. This is the impact of the different trigger techniques, how the triggered ELM evolves temporally and spatially and why the correlation between local edge parameters like the collisionality ν and ΔW_ELM observed for intrinsic ELM can be broken. Local 3D perturbations were imposed by Deuterium (D) pellet injection, supersonic D gas jets, and Carbon and Aluminium micro pellet laser blow off. It is found that the gas jet, unlike the pellet, is insufficient to trigger prompt ELMs although the applied particle flux is of the same amount. On the other hand, investigations showed pellet masses reduced by more than a factor of ten-technically not feasible to date-would be still sufficient to release prompt ELMs. This shows that the perturbation required for ELM triggering has to fulfil local susceptibility criteria. Experiments conducted in order to uncover possible relations are reported as well as approaches to resolve the temporal and spatial evolution of a triggered ELMs with maximised resolution and compare it to its intrinsic counterparts. Furthermore, a fast framing camera system was set into operation for dedicated studies on the ELM trigger dynamics. The pellet based ELM pacing approach will be discussed with respect to its potential extension towards higher rations f_ELM /f_o(f_o: intrinsic ELM frequency). Experimental efforts at ASDEX Upgrade and its application to JET will be described. Finally, a possible scheme for pellet ELM pacing at ITER will be presented. (Author)

[en] A low density H-mode plasma has been selected for detailed inter-ELM modelling by the SOLPS code package, with the coupled treatment of its plasma (fluid code B2) and neutral (Monte-Carlo code Eirene) parts. Good quality measured midplane density and temperature profiles, covering the pedestal region and stretching far into the SOL, as well as several other parameters and profiles measured in the divertor, have enabled testing the consistency of code solutions with experiment. Once the upstream, midplane profiles have been fitted, and the global parameters (e.g. input power into the computational grid, radiated power) matched, the code reproduces experimental profiles and control parameters in the divertor with an accuracy within a factor of 2. Deviations of modelled parameters from the experiment were found around the strike point position where most of the power was deposited on the target. The deviations are consistent among themselves and all point to one common problem with the modelling: the predicted divertor electron temperature is too low and the density too high, compared with the experiment. The largest inconsistency between the code and experiment was in the magnitude of the peak H_α radiation in the outer divertor, which was larger by a factor of 2 in the code simulations. In addition, the code predicts a somewhat higher sub-divertor neutral flux but lower carbon impurity content in the edge plasma than in the experiment, as well as lower CIII emission. The discrepancy between H_α profiles can to a large degree be attributed to profile effects: the simulated H_α emission profiles are narrower than in the experiment, reflecting the tendency of the neutral-plasma mix to congregate excessively around the strike point in the modelling. At the same time, the integrated H_α emission matches very well with the experiment. Extensive sensitivity studies of the influence of variations in input parameters and assumptions of the code on the modelled divertor conditions have been conducted. They have not resulted in an identification of any SOLPS input/control parameters capable of removing the main disagreement between the code output and experiment. A possibility of parallel transport effects related to low collisionality to increase the effective plasma temperature near the strike point position or of increased perpendicular transport by neutrals (due to some missing reactions in Eirene) to widen the target profiles, will be explored in the future