Table of contents

A time dependent lower hybrid current drive tokamak simulation code has been developed. This code utilizes the BALDUR tokamak simulation code and permits the study of the interaction of lower hybrid current drive with neutral beam heating in shaped cross-section plasmas. Examples of simulations are shown for a high aspect ratio tokamak which illuminate the effect of cross-section shaping on current drive

The impurity content of the plasma during the steady state phase of ICRF minority heating in TEXTOR is found to be, for boronized wall conditions, a function of the thermal load on the limiter only. This new experimental result, which is to be related to the RF power absorption characteristics of the TEXTOR plasmas, has been established by comparing, at different RF power levels, the plasma effective charge with the thermal deuteron flux to the limiter. During the transient phase of the RF pulse, however, relevant RF-specific effects in the boundary plasma have been observed. These effects, which are accompanied by a release of impurities, are suppressed almost completely if a high energy RF induced tail of minority ions is previously built up in the target plasma. This is due to a large increase of the damping of the fast wave. The RF pre-heating at the moderate power level of 0.8 MW has been found experimentally to be very effective, in agreement with the predictions of a full wave Fokker-Planck simulation code. Comparison of Z_eff with spectra in the UV, VUV and SXR ranges leads to the conclusion that the carbon influx dominates the plasma contamination

The bootstrap current and electron thermal conductivity have been investigated in ECRH plasmas on the CHS heliotron/torsatron. The neoclassical transport coefficients have been externally controlled over a wide range by changing the ellipticity of the plasma cross-section. While local thermal conductivity, as well as global confinement, have shown inconsistency with the neoclassical predictions, the observed toroidal net current has been successfully explained by considering the three dimensional geometry with the available neoclassical theory. The present experimental results indicate that the transport parallel to the magnetic field lines (bootstrap current) is likely to be neoclassical but that the perpendicular transport (thermal conductivity) is anomalous, which is common in current less plasmas as well as tokamaks

The runaway diffusion coefficient, D_r, for the TJ-I tokamak has been deduced using two different methods: in the first, D_r is obtained using a steady state approach for values of the runaway confinement time tau _r deduced from hard X ray (HXR) bremsstrahlung spectra; in the second, D_r is deduced from sawtooth oscillations of HXR flux, and from the soft X ray intensity, recorded simultaneously. Data have been taken in a scan with the toroidal magnetic field B_T. Averaged values obtained for D_r=5-10 m² s^-1, decreasing with B_T, are in both cases consistent. Assuming that magnetic turbulence is responsible for their transport, from D_r an averaged value of b/B_T, the magnetic fluctuations level, can be deduced. The thermal conductivity coefficient chi _e obtained using other diagnostics in TJ-I, is compared with D_r. The results are in good agreement with those obtained in TJ-I using probes, spectroscopic methods, power balance analysis and a coupled analysis for temperature and density pulses

Asymmetries of various plasma parameters, both equilibrium and fluctuating, are studied in the edge of the TEXT-U plasma using multiple Langmuir probe arrays. A substantial up-down asymmetry in potential is found, suggesting the related local radial steady state convection could be significant. However, the dependence of this asymmetry on density suggests that its flux surface average is not important except at low densities. Results during electrode biasing experiments, in which improved confinement is observed at low density, support this view

Plasma propagation along the magnetic field lines after pellet evaporation is simulated. The character of the plasma expansion depends strongly on the ambient plasma parameters. For medium plasma parameters, T_e < 1 keV, n_e = 3 × 10¹⁹ m^-3, the ablatant expansion, calculated with respect to the ambient temperature, is supersonic. Shock formation at the plasma front is observed. The results of the numerical simulation are in reasonable agreement with the self-similar solution obtained for the isothermal plasma. For reactor type parameters, the ablatant heating is a much slower process, and plasma expansion is subsonic

A time dependent quasi-two-dimensional ablation model has been developed. The ablation of a pellet injected into a magnetically confined plasma and the temporal variation of the properties of the cloud surrounding it are calculated in a self-consistent manner. The energy deposition in the shielding cloud is determined by stopping length calculations applied along the magnetic field lines with allowance for electrostatic shielding effects. The stopping length calculations are supplemented by thermal diffusion calculations, thus redistributing the energy deposited in the discrete energy group approximation. The neutral cloud is allowed to expand also in the direction perpendicular to the magnetic field. Finite rate equations are used for determining the time history of the ionization state of the ablated substance. The deceleration and full stopping of the cross-field motion is calculated by means of an MHD model (iteratively with the axial expansion dynamics), thus determining the transient variation of the lateral cloud dimension and the corresponding modifications of the field aligned density and temperature distributions used in the stopping length calculations. The expulsion of the magnetic field lines by the expanding and ionized ablatant and their rediffusion into the partial magnetic cavity thus formed are also calculated. The variation of the ablation rate along the pellet path (i.e. that of its value averaged over the residence time of a pellet of given injection velocity in a flux tube defined by the local ionization radius of the ablatant) are calculated at a sequence of flux surfaces, thus determining the particle deposition profile. Results of penetration depth calculations are compared with values stemming from experiment and from other ablation models. Good correspondence is found at moderate plasma temperatures. Predictive calculations performed for a reactor-grade plasma yield, owing to magnetic shielding effects, substantially larger penetration depths than those obtained with the standard NGS a

The JET neutron emission profile monitor is used to measure the 2.5 MeV and 14 MeV neutron emission line integrals before and after sawtooth crashes in high d-d neutron yield, hot ion H mode plasmas in the Joint European Torus (JET). Deuterium-deuterium (d-d) fusion produces 2.5 MeV neutrons and 1 MeV tritons (t) at nearly equal rates from its two reaction channels. A plasma current of 3 MA is sufficiently high to contain most of the fusion product tritons, which have birth orbit gyroradii and velocity space distributions similar to those of the 3.5 MeV or particles from d-t fusion. By examining neutron emission line integrals and tomographically deduced local emissivity profiles, an upper limit of 10% can be placed on the net fraction of fusion product tritons which are displaced from the plasma axis by those sawtooth crashes studied. This is a much smaller net fraction than that typically observed, 35-55%, for displaced injected neutral beam deuterium ions. A study of the response of beam injected deuterium ions to a sawtooth crash shows that the change in their axial density depends on the precrash spatial width of the neutron emissivity profile. The fusion product tritons, which have a large precrash spatial width, respond weakly to a crash. This weak response is consistent with the behaviour of the analogous d-d beam thermal neutrons when extrapolated to the corresponding emissivity spatial width. The implication of these observations is that beam ions and 3.5 MeV alpha particles in JET may be relatively resilient to sawtooth clashes, when the spatial width of their density is sufficiently large

Ion temperature profiles of neutral beam heated plasmas in JT-60 and TFTR have been simulated using models of ion thermal diffusivity, χ_i, based on ion temperature gradient mode (η_i mode) turbulence and drift wave turbulence (trapped electron mode and circulating electron mode). The ion temperature profiles measured by charge exchange recombination spectroscopy are compared to predicted T_i profiles calculated from three theoretical models for ion heat transport by Dominguez and Waltz (1987), Lee and Diamond (1986), and Romanelli (1989). All three χ_i models can reproduce the measured T_i profile over a wide range of parameters in JT-60 L mode plasmas, except for two 1 MA limiter cases. With the use of transport models, which were adjusted to fit JT-60 plasmas, it was found that the Dominguez and Waltz and Romanelli models agree with measured T_i, profiles in TFTR L mode discharges in the region a/4 < r < a/2. The observed central peaking of T_i near the magnetic axis of TFTR L mode plasmas cannot be reproduced. It was found that the Lee and Diamond model does not fit me data as well. The χ_i models studied cannot consistently reproduce the measured T_i in the peripheral region of either JT-60 or TFTR plasmas. In the high ion temperature (high T_i) plasmas of JT-60 and supershot plasmas of TFTR, the predicted T_i profiles are much broader than the measured profile. In supershot plasmas, the measured central ion temperature greatly exceeds the predicted temperature, although there is reasonable agreement near the edge

The local 3-D ideal MHD stability properties for two different H-1 Heliac configurations with values of magnetic well of 1.3 and 3.2%, respectively, have been investigated. The pressure profile has been determined that approaches marginal stability to 3-D ballooning modes almost uniformly across the plasma radius for each case. In spite of the difference in well depths, the beta limits for both configurations are almost the same (at β ≃ 0.8%). The optimal pressure profile for the deep well configuration is slightly more peaked than that for the standard configuration. The ballooning imposed beta limits are more stringent than those of the Mercier criterion (β ≃ 1%) and the ballooning optimized pressure profiles are more peaked than the corresponding profiles which are marginal according to the Mercier criterion

Helium-3 ions with energies from thermal values up to 300 keV produced by ICRF heating have been spectroscopically observed by the α-CHERS diagnostic on the Tokamak Fusion Test Reactor (TFTR). The shape of the spectrum and the temporal and spatial behaviour of the energetic ³He ion emission are consistent with expectations for this ICRF heating case. Because the ³He ion density was similar to the alpha particle density predicted for D-T supershots on TFTR, these observations give confidence that confined alphas produced by D-T fusion reactions can be observed by α-CHERS during D-T operation of TFTR

The interaction between waves having frequencies much lower than the cyclotron frequency and trapped electrons is considered in the limit α ≡ ω_w²/λω_bo² >> 1 (here ω_w is the particle oscillation frequency in the wave field, ω_bo is the bounce frequency of trapped particles and λ = k_||qR). In this case, stochasticity arises due to the separatrix crossings experienced by electrons when they pass through a wave-particle resonance on the bounce trajectory. According to general theory, for successive resonance passages, changes in the adiabatic invariant are uncorrelated. As a result, diffusion in velocity space should arise. The calculated diffusion coefficient scales as D ∝ α^-1ln²α while in the opposite limit, α << 1, D ∝ α². This degradation of diffusion coefficient is favourable for current drive efficiency, since a smaller fraction of the total RF power is absorbed by trapped electrons

Electron cyclotron assisted breakdown is discussed. The investigation is motivated by the interest in low voltage plasma formation in next generation tokamaks requiring an efficient and reliable non-inductive method of plasma ionization to assist ohmic breakdown. Two basic problems for solving the relevant transport equations, the electron cyclotron power deposition and the ionization rate in a large major radius tokamak, are investigated. It is shown that wave damping occurs at the electron gyrofrequency without mode conversion at the upper hybrid frequency, and that the ionization rate is nearly independent of the shape of the electron distribution but only depends on the average thermal energy

Report on the Fifth IAEA Technical Committee Meeting and Workshop, held at the University of California, Los Angeles, California, United States of America 13-17 September 1993.