Table of contents

Measurements of ion temperature profiles, obtained on a single-shot basis, have been coupled with transport analysis to obtain spatial profiles of the ion thermal diffusivity χ_i for a series of 18 neutral-beam-heated discharges in the Doublet III tokamak. The expanded boundary divertor discharges examined in this study had the following range of parameters: B_t = 2.53 T, I_p = 0.35–0.89 MA, vertical elongation = 1.4–1.9, _e = (3.7–7.9) × 10¹³ cm⁻³, P_T = 3.5–6.6 MW, and ν_i* = 0.03–0.22 , for r/a = 0.2–0.7. The experimentally inferred values of χ_i have been compared with the values of the ion thermal diffusivity predicted by neoclassical theory. On the basis of a detailed error analysis, the results for all discharges examined are the same: χ_i is approximately equal to near the plasma centre, but for r/a between about 0.2 and 0.8, χ_i is both larger than and almost certainly has a spatial variation different from that of . The differences between χ_i and are larger than can be explained by errors in the experimental measurements and indicate that there are processes operative for the loss of ion heat which are not included in the usual transport models for ion confinement.

In the overall design of an inertial fusion reactor driven by ion beams or lasers the target design plays a central role. The concept of central ignition is used to reduce the input energy of the driver as much as possible. In this respect, the range of the 3.5 MeV alpha particles at 1–20 keV released by the fusion reactions i n D-T is crucial for estimating the driver parameters. Further, for the calculation of the pellet gain and the burn processes, the ranges of alpha particles at temperatures up to 200 keV and at densities up to 1000 g·cm⁻³ must be accurately known. The 14.1 MeV neutrons produced during the D-T reaction can collide with the deuterium and tritium ions and produce suprathermal knock-on ions which then slow down in the background plasma. This effect must also be calculated, especially for reactor-size pellets for which the pellet ρR is comparable to the neutron mean free path. The stopping power of low atomic fusion products and deuterium and tritium ions in D-T is calculated, using the dielectric function theory for the stopping power of electrons. The theory of the energy deposition of ions in fully ionized, quantum and classical, ideal and non-ideal dense plasmas is reviewed. The GORGON computer code is used for the numerical calculations. The results obtained are compared with the results of other authors and with their theoretical methods.

A simple model of energy confinement in tokamaks is proposed. In this model, the temperature and current profiles are determined by tearing mode stability considerations. As a result, the global energy confinement becomes strongly dependent on the boundary energy transport. A higher edge temperature ensures a higher core temperature and hence a higher global energy confinement.

Plasma position control in a tokamak reactor in the phase approaching ignition is closely related to burn control. If ignited burn corresponds to a thermally unstable situation the plasma becomes sensitive to the thermal instability already in the phase when ignition is approached so that the trajectory in the position-pressure (R, p) space becomes effectively unpredictable. – For example, schemes involving closed cycles around ignition can be unstable in the heating-cooling phases, and the deviations may be cumulative in time. – Reliable plasma control in pressure-position (p, R) space is achieved by beforehand constraining the p, R trajectory rigidly with suitable feedback vertical field stabilization, which is to be established already below ignition. – A scheme in which ignition is approached in a stable and automatic way by feedback stabilization on the vertical field is proposed and studied in detail. – The values of the gain coefficient ensuring stabilization and the associated p and R excursions are discussed both analytically, with a 0-D approximation including non-linear effects, and numerically, with a 1-D code in cylindrical geometry. Profile effects increase the excursions, in particular above ignition.

The loss flux to an azimuthally segmented limiter in the central cell of the cusp end cell stabilized, axisymmetric mirror machine, RFC-XX, shows that non-axisymmetric illumination by either ICH or ECH can induce azimuthal non-uniformities in the plasma. Measurements of the floating potential in the vicinity of the limiter and by the limiter itself suggest that a large stationary azimuthal electric field is induced which can convectively transport plasma across magnetic field lines. The local radial component of the electric drift velocity is as large as 7 × 10⁴ cm·s⁻¹. Examinatio n of the plasma end loss versus azimuth and probe measurements deep in the plasma show that the non-uniformities extend into the plasma and are not localized near the outer boundary. The data suggest that the RF fields are driving an anomalous radial loss.

The paper extends and augments earlier research concerning the effect that a finitely conducting wall has on the linear stability of various plasma models. If a plasma is stable to a particular mode when surrounded by a perfect wall but unstable to it when the wall is removed, then theory indicates that surrounding the plasma with an imperfect wall can lead to instability of that mode on the long time constant of the wall. Analysis of this effect requires that the various modes possible be classified as to whether the plasma is regarded as ideal or resistive, and whether the pitch of the perturbation resonates with that of the equilibrium. The effect of both bulk plasma rotation and mode rotation (plasma diamagnetism) on these various modes is discussed.

The wave induced particle transport during the ion cyclotron resonance heating is studied in collisionless toroidal plasmas. It is shown that the previously neglected non-conservation of the toroidal angular momentum _ϕ caused by the toroidal wave component E_ϕ, is necessary to allow particle diffusion and yields the leading diffusive contribution. While the induced ion transport for the RF power in contemporary experiments is of the order of the neoclassical value, that of fast alpha particles is quite large if resonance is present.

The paper discusses the results of 1½-D computer simulations used in a study of the influence of the power spectrum bandwidth of the incident laser wave on stimulated Raman scattering in laser-irradiated plasmas. Purely sinusoidal pump waves and non-monochromatic pump waves are used. The non-monochromatic irradiation corresponds to a multiple-line pump. Parameters of Raman scattering, such as the growth rate of the plasma wave, laser light-absorption, backscattering and forwardscattering rates, are compared. The generation of suprathermal electrons is examined. Reduction of the efficiency of Raman scattering and weakening of its ability to produce fast electrons are observed when a large laser bandwidth is used.

It is attempted to predict the consequences of adding a cylindrical inductive current transformer on the small major radius side of the present S-1 Spheromak experiment. Axisymmetric modelling with only classical dissipation shows an increase of toroidal current and a shrinking and hollowing of the current channel, conserving toroidal flux. These unstable profiles will undergo helical reconnection, conserving helicity while increasing the toroidal flux and decreasing the poloidal flux so that the plasma relaxes toward the Taylor state. This flux rearrangement is modelled by a new current viscosity term in the mean field Ohm's law which conserves helicity and dissipates energy.

A new method for measuring divertor configuration parameters (plasma position, location of separatrix surface, β_p + ℓ_i/2) for feedback control in JT-60 is presented. By using the MHD equilibrium database, it is possible to derive statistical relations expressed by simple functions between these parameters and the measured values of quantities such as poloidal magnetic fields, plasma current and external coil currents. These functions can be used as the measurement formulas of these parameters. In JT-60, the second order polynomial functions of these magnetically measured values, which are obtained as the measurement formulas, are sufficiently accurate to be applicable to feedback control.

A method for investigating the non-equilibrium kinetic processes which determine the formation of a non-Maxwellian electron distribution in tokamak plasmas is discussed. The case of a non-Maxwellian fast electron tail is considered in detail. It is shown that the study of the transient phase after modification of the tail by a short pulse of electron cyclotron wave energy yields valuable information on the mechanisms of tail formation.

Internal disruptions occurring in an inductively driven discharge are suppressed by lower hybrid waves that drive the plasma current in the same direction as the inductively driven current. Upon suppression, in the central region a finite-amplitude, non-growing m = 1 oscillation is observed and strong electron heating is measured.

The magnetic field configuration of a helical axis stellarator, having the advantage of both magnetic well and field-line shear, can be realized by means of a simple coil system of the torsatron type, even without the help of special toroidal field coils. Such a system of the torsatron type, the 'Vintotron', may provide experimental information relevant to nuclear fusion which is complementary to the information obtained by another type of helical axis stellarator, the Heliac.

An upper bound for the local electron heat coefficient χ(r) in electron cyclotron heated (ECH) discharges is determined experimentally. The method applied uses power modulation of the ECH and measures the time delay of the electron temperature response as a function of radius by means of a multi-channel radiometer for second-harmonic electron cyclotron emission (ECE). The evaluation is restricted to plasma radii between 3 and 6.5 cm, where the pre-conditions of the method are best verified experimentally. The theoretical transport model considers only processes related to the electron temperature gradient and neglects the dependence of χ on T_e.

With the progress in controlled thermonuclear fusion research, experiments having dimensions approaching those of future fusion reactors and discharge times of the order of 100 to 1000 s have become a necessity. Parameters close to those of ignited plasma are anticipated or achieved in today's four large tokamaks TFTR (Princeton), JET (Culham), JT-60 (Naka-machi) and T-l 5 (Moscow); the burning of a fusion plasma is to be investigated in the next generation of experiments. In these large experiments the load from the plasma on the surrounding vessel walls is considerably increased in duration and intensity compared to that of most previous plasma experiments. The wall load is approaching the values foreseen in a fusion reactor and the lifetime of the wall components represents a major physical and engineering problem.