The theoretical and experimental development of stellarators has removed some of the specific deficiencies of this configuration, viz., the limitations in β, the high neoclassical transport, and the low collisionless confinement of α particles. These optimized stellarators can best be realized with a modular coil system. The W7-AS experiment [Plasma Phys. Controlled Fusion 31, 1579 (1989)] has successfully demonstrated two aspects of advanced stellarators, the improved equilibrium and the modular coil concept. Stellarator optimization will much more viably be demonstrated by W7-X [Plasma Physics and Controlled Fusion Research, Proceedings of the 12th International Conference, Nice, 1988 (IAEA, Vienna, 1989), Vol. 2, p. 369], the successor experiment presently under design. Optimized stellarators seem to offer an independent reactor option. In addition, they supplement, in a unique form, the toroidal confinement fusion program, e.g., energy transport is anomalous in stellarators too, but possibly more easily understandable in the frame of existing theoretical concepts than in tokamaks.

No abstract

In many stellarators-envisagcd as fusion deviccs--any a-particle which cver gets reflected (all = 0) is collisionlcssly lost in a time which is orders ofmagnitude smaller than the typical slowing-down time of =IO-' s. Two classes of stellarators to which this general picture docs not apply arc dcscribcd: quasi-helically symmetric stellarators and a class ofstellarators with vanishing bootstrap current in which the collisionless r-particle confinement sufficiently improves at finile 8. The influence of the modular ripple in optimized coil systems realizing these configurations, the angular distribution of the kast e-particle losses, and the application of the rcsults to a-particle confinemcnt Simulation experiments in next-generation stellaialors are also discussed

No abstract

No abstract

The neoclassical bootstrap current properties of optimized stellarators are analyzed in the relevant mean-free-path regimes and compared with the neoclassical transport properties. Two methods—global Monte Carlo simulation [Phys. Fluids 31, 2984 (1988)], and local analysis with the drift kinetic equation solver code [Phys. Fluids B 1, 563 (1989)]—are employed and good agreement is obtained. Full consistency with the elimination of the bootstrap current and favorable neoclassical transport are found.

Neoclassical transport coefficients and confinement times in stellarators of general geometry and tokamaks with and without ripple are computed by Monte Carlo simulation over wide ranges of mean free paths, ratios of plasma to gyroradius, and radial electric fields. The results for monoenergetic particles can be represented by simple formulas using a transport coefficient normalized to the tokamak plateau value and a mean free path normalized to half the connection length. Transport coefficients obtained with monoenergetic particles subjected to pitch angle scattering and energy relaxation are convoluted with a Maxwellian energy distribution. The results are compared with theory and simulations using a particle distribution subjected to pitch angle as well as energy scattering. The overall agreement is good. Transport coefficients with Maxwellian energy distributions for l=2 stellarators and for various other stellarator configurations are shown. Particle transport coefficients, as well as energy transport coefficients, for these configurations are computed for ions (deuterons) as well as for electrons. Estimates of particle and energy confinement times are also obtained.

Experiments with the first complete high-6 stellarator torus with helical magnetic axis are reported. The Garching 2. 6-MJ capacitor bank was used to feed a toroidal theta coil with a major.diameter of 2. 7 m (ISAR Tl). Toroidal equilibrium was achieved by superposing helical fi = 1 and & = 2 (partly also £ = 0) fields on the slender (A = R T /r 0 » 150) toroidal pinch plasma such that the plasma surface facing the torus centre was more corrugated than the outer side (M-and-S effect). The toroidal equilibrium condition and the corresponding plasma distortions were consistent with sharp-boundary-model predictions. Effects in connection with initial dynamics, toroidal plasma currents and transverse magnetic fields could be explained by simple models. In agreement with sharp-boundary theory, short wavelength m = 1 and m = 2 modes were found to be stable and long wavelength m = 1 modes were unstable, limiting the plasma life-time by wall contact. Long-wavelength m £ 2 instabilities were not observed in contrast to sharp-boundary theory, i. e. this model is much too pessimistic for m > 2 modes, even if the finite gyroradius is included. No significant difference in the stability behaviour was -found, compared with previous linear and toroidal sector experiments.