Table of contents

Relaxation oscillations found in the divertor of the ASDEX Upgrade tokamak by means of spectroscopy and probe diagnostics are reported. They are observed in both electron density and temperature with frequencies in the range from 45 to 80 Hz. In the line emission they are detected in C III, but not in D_alpha light. Interpretation of the experimental data indicates that the relaxation oscillations are associated with periodic movements of the thermal front and the power loss zone against each other

The magnitude and radial profiles of non-inductive currents driven by fast magnetosonic waves in tokamaks have been calculated directly from the wave induced quasi-linear flux in a toroidal geometry and a Green's function for the current. An expression for the quasi-linear flux has been derived that accounts for coupling between modes in the spectrum of waves launched from the antenna. A Fokker-Planck code for the Green's function and a full wave code for the electric field in the quasi-linear flux are used to evaluate the current in a specified toroidal geometry

An analysis is presented of runaway electron formation and its evolution during disruptions in large tokamaks, where avalanche phenomena play a crucial role. A simplified, but sufficiently accurate, analytical model suitable for one dimensional (1-D) transport codes is proposed. Validation of the model was done by comparison with Monte Carlo calculations

The magnetohydrodynamic (MHD) ballooning stability of the TJ-II heliac standard configuration is examined using a criterion derived for three dimensional (3-D) general configurations. From the ideal limit of this criterion, a critical ⟨β⟩ approximately=1.3% is obtained. This critical value is very close to that obtained from the Mercier criterion and is almost unaffected when those terms associated with compressibility are retained. The inclusion of resistivity is shown to have almost no effect on the attainable critical ⟨β⟩ even though unstable modes persist for lower ⟨β⟩ values due to their small growth rate. For parameters relevant to this configuration, the resistive pressure convection limit predicts unstable modes, scaling as γ~m²η for all ⟨β⟩ below the ideal limit, which are shown to disappear when compressibility terms are retained in the calculations. Unstable resistive ballooning modes with γτ _A~10^-3 are then found for ⟨β⟩ values down to 90% of the ideal critical ⟨β⟩. The growth rate rapidly falls if the pressure gradient is further decreased owing to the stabilizing effect of compressibility for γτ_A << 1

The hydrogen and deuterium inventories of the ASDEX Upgrade divertor tiles were measured after the experimental period from December 1994 to July 1995 by thermal desorption spectroscopy (TDS) of samples cut out of the divertor tiles. The samples were heated by electron bombardment up to 2100 K; the released gases were measured by means of a calibrated quadrupole mass spectrometer. The measured hydrogen or deuterium inventories are of the order of 10²³ m^-2. They are larger for samples of the inner divertor than of the outer divertor by a factor of about 2. The largest inventory was found at the separatrix position of the inner divertor. Most of the released hydrogen (H) can be attributed to water adsorbed in the near surface region during the air exposure prior to the TDS measurements. The total inventories measured by TDS exceed the inventories in the near surface region (<25 mu m) measured by ion beam analysis methods by a factor of up to 10. Hence, the total hydrogen retention is governed by the diffusion out of the near surface region deep into the material. The hydrogen and deuterium inventories decrease with increasing surface temperature

Experiments have been performed on the TCV tokamak to evaluate the response of ohmic, L mode, limited, vertically unstable plasmas to changes in all the poloidal field coil voltages. The resulting closed loop plasma responses have been compared with the CREATE-L linearized MHD equilibrium model of the TCV tokamak. All the responses in both the time domain and the frequency domain show excellent agreement both for directly measured quantities and for derived parameters. No modifications to the CREATE-L model were made to achieve this quality of agreement, indicating that the underlying physical assumptions are appropriate

Fluctuations produced by beam-driven toroidicity-induced Alfven eigenmode (TAE) activity in the DIII-D tokamak are measured by a poloidal array of magnetic probes and compared with the wavefields computed by two theoretical models. Fluid resistive models compute continuum damped TAEs. A kinetic plasma model that retains Landau damping and finite Larmor radius effects computes global drift-kinetic Alfven eigenmodes. The phases of the probes disagree with both theoretical predictions, while the amplitudes agree best with the kinetic model

Magnetohydrodynamic (MHD) operational limits are computed for tokamaks with negative central shear (NCS). Beta optimized profiles are generated, imposing stability to ideal n = 1, 2, 3 and infinity modes without a conducting wall. In addition, the profiles are constrained so that no negative current drive is needed to counterbalance the bootstrap current in steady state operation. Under this last condition, the highest stable values of both β and β_N are found for high current and broad current profiles. Beta limits significantly above the semi-empirical scaling β_N ⩽ 4l_i are found at low inductances, in particular for strong shaping. The broadness of useful current profiles is limited by the appearance of `ravines', where the beta limit falls drastically for q_a below integer values. Low-n modes, in particular n=1, limit the peaking of the pressure, and the optimal pressure peaking factors are in the range of 2.5 to 3. The beta limit increases significantly when both elongation kappa and triangularity delta are increased, but high elongation is not favourable at low triangularity. At low-q operation with about 40% bootstrap fraction, a JET shaped cross-section, κ = 1.6, δ = 0.3, gives a β* limit of 6.2% while stronger shaping, κ = 2.0 and δ = 0.7, gives a limit of 9.8%. At a bootstrap fraction of 65%, the corresponding β* limits are rather low, about 2.3% for a JET shaped cross-section and 3.5% for κ = 2.0, δ = 0.7

An attempt is made to find an optimal quasi-helically symmetric (QHS) four period magnetic configuration with a Heliac-like structure of the magnetic surface cross-sections. This means that the cross-sections rotate in phase with the magnetic axis principal normal in contrast to the Helias-like systems where the magnetic surface cross-sections rotate half as fast as the principal normal. The equilibrium and local stability are studied with the VMEC and TERPSICHORE codes. It is shown that, in the configuration found, the equilibrium beta limit is about 9%, the Mercier beta limit is about 4% and the ballooning modes restrict the beta value to 1%. The accuracy of the fulfilment of the QHS condition is given by X~3 (i.e. the largest amplitude of the magnetic field strength B harmonic that violates the quasi-symmetry is a third that of the main helical harmonic of B in Boozer co-ordinates)

The problem of helium ash removal from a fusion reactor is treated with spatially varying plasma parameters, and the allowance of variations is found to strongly influence the necessary conditions for sustained deuterium-tritium (DT) ignition. In particular, a one dimensional (1-D) particle transport model is used to calculate the core helium concentration on the basis of a suitably parametrized source profile. It is estimated that the necessary exhaust efficiency for sustained DT ignition is of the order of 2%, rather than the familiar 10% figure deduced from uniform plasma calculations. This is due to: (1) the fact that tau _P and tau _E are not equal; and (2) the fact that fusion alphas are produced only in a fraction of the plasma volume. It is also shown that, in order for the ash concentration to be effectively controllable by pumping, it is necessary that the particle diffusion coefficient be sufficiently high and that the reactor's power density not be excessive. In order to determine the exact pumping requirements for a machine such as ITER, a more detailed and self-consistent model is needed

The scrape-off layer (SOL) plasma is studied over a wide range of Coulomb collisionality via particle Monte Carlo simulation with focus on the cooling effect of secondary electrons in the high temperature divertor operation. The electron distribution function in the collisionless regime shows a large departure from a Maxwellian distribution. In the presence of strong secondary electron emission the electron sheath energy transmission factor in the collisionless regime is found to be significantly smaller than that in the collisional regime, which indicates that a higher temperature is obtainable in collisionless SOL plasmas. In both collisional and collisionless SOL plasmas, the change in the conductive energy flux from the core region into the SOL has little influence on the normalized potential drop and electron energy transmission factor, while the electron temperature is raised as the conductive energy flux increases

Linear magnetohydrodynamic (MHD) and equilibrium evolution approaches describe the linear and non-linear axisymmetric displacement dynamics of free boundary plasma equilibrium configurations surrounded by conductors in an external magnetic field. A comparison of the two different approaches was made using DIII-D-like free boundary equilibria. Good agreement was found for up-down symmetric configurations. However, a considerable difference in growth rates is found for up-down asymmetric equilibria. The difference can be explained by taking into account surface current perturbations in the MHD model. Common and specific features of the two approaches are discussed

The effects of finite beta and radial electric fields on neoclassical transport in LHD are investigated with the Drift Kinetic Equation Solver (DKES) code. In the finite- beta configuration, even orbits of deeply trapped particles deviate significantly from magnetic flux surfaces. Thus, neoclassical ripple transport coefficients in the finite- beta configuration are several times larger than those in the vacuum configuration under the same conditions of temperatures and radial electric fields. When the plasma temperature is several kiloelectronvolts, a bifurcation of the electric fields appears under the ambipolarity condition. The E*B drift due to the resultant sufficiently large radial electric fields causes the orbits to curve so that they more closely follow the pressure surfaces and improves significantly the transport coefficients in the finite- beta configuration

The power handling and coupling of a folded waveguide (FWG) antenna array in the ion cyclotron range of frequencies (ICRF) are analysed numerically for reactor plasma conditions. The fields inside the FWG unit are approximated in the model of the usual rectangular waveguide that will result from unfolding the FWG, keeping the corresponding dimensions. Including the surface impedance for a hot reactor sized plasma and the external current simulating the short probe source inside the waveguide allows one to investigate the coupling between units, the input impedance and the matching of a system consisting of several FWG units. Reconstructing the field structure both inside the waveguides and in the vacuum layer (between the coupler front and the plasma), the poloidal and toroidal spectra of the waves launched into the plasma are computed in the case where the effect of the poloidal magnetic field is neglected. Assuming 60 MHz frequency and specific ITER relevant waveguide and plasma parameters with 5 cm scrape-off decay length and 10 cm vacuum gap between the plasma and the waveguide, the present model with idealized non-perturbed feeder excitation predicts a voltage of less than 40 kV along the current probe source, a maximum electric field below 30 kV.cm^-1 and a maximum voltage of the order of 250 kV inside the FWG with ten folds for 10 MW radiated power per unit. For a waveguide complex composed of eight units, the mutual coupling between the units is found to be non-negligible, indicating a need for tuning in the system circuit