Table of contents

Strong m = 1 MHD activities are observed in the HL-1M tokamak during off-axis electron cyclotron resonance heating (ECRH) when the cyclotron resonance location is placed just outside the q = 1 surface at the high-magnetic-field side of the magnetic surface. Addition of lower-hybrid waves to ECRH significantly enhances the MHD excitation, but lower-hybrid waves alone cannot excite or sustain the mode. This result is a clear demonstration of the suprathermal trapped electron effect on the instability because of the absence of energetic ions in the plasma.

The influence of certain equilibrium flows on the stability of tearing modes in the reversed-field pinch is investigated. By solving the linearized magnetohydrodynamic equations in cylindrical geometry, the tearing mode stability factor Δ' is calculated for a variety of axial flow profiles which have nonzero shear away from the rational surface, including flows localized entirely in the external, ideal region of the tearing mode. It is found that both m = 1 and m = 0 modes are destabilized by an axial flow localized near the edge of the plasma. This is the kind of flow that might be generated by any physical process creating an edge-localized radial electric field. A global flow profile with shear over the middle region of the plasma, simulating the differential rotation of core and edge modes observed in some reversed-field pinch discharges, is found to have a destabilizing effect on the m = 1 mode, while leaving the stability parameter of the m = 0 mode practically unchanged. The possible connection of these results with features of the spontaneous enhanced confinement regime in the Madison Symmetric Torus is discussed.

Double transport barrier plasmas comprised of an edge enhanced D_α (EDA) H-mode pedestal and an internal transport barrier (ITB) have been observed in Alcator C-Mod. The ITB can be routinely produced in ICRF heated plasmas by locating the wave resonance off-axis near |r/a|~0.5, provided the target plasma average density is above ~1.4×10²⁰ m^-3, and can develop spontaneously in some Ohmic H-mode discharges. The formation of the barrier appears in conjunction with a decrease or reversal in the central (impurity) toroidal rotation velocity. The ITB foot is located near r/a~0.5, regardless of how the barrier was produced. The ITBs can persist for ~15 energy confinement times (τ_E), but exhibit a continuous increase of the central electron density, up to values near 1×10²¹ m^-3 (in the absence of an internal particle source), followed by collapse of the barrier. Application of additional on-axis ICRF heating arrests the density and impurity peaking, which occurs along with an increase (co-current) in the core rotation velocity. Quasi-steady state double barrier plasmas have been maintained for 10τ_E or longer, with a bootstrap fraction of 0.13 near the ITB foot.

Analysis of experiments performed at different heating powers and magnetic field intensities shows that the existence of internal transport barriers in JET can be inferred in regions of space-time where the ratio of the ion gyroradius to the local gradient scale length exceeds some critical value. A possible interpretation leading to the theoretical relevance of this dimensionless parameter as a local indicator of a bifurcated plasma state is the stabilization of turbulence by the E×B shear flows associated with large pressure gradients and plasma rotation. Large database analysis and real time plasma control are envisaged as attractive applications.

Ion cyclotron resonance heating in a toroidal plasma not only increases the perpendicular energy of the resonating ions but also results in their spatial transport. Depending on the direction of propagation of the waves, the ions will either drift inwards or outwards giving rise to an RF induced rotation with the toroidal torque component in the co-current or counter-current directions, respectively. It is found that the spatial transport induced by the RF field, the topology of the ion drift orbits and a wave field consistent with ion absorption are important for determining the distribution function of the heated species. Studies of ICRF heating with the self-consistent code SELFO reveal new features such as the formation of non-standard passing orbits residing on the low field side of the magnetic axis. For a symmetric spectrum the drift terms will in general not cancel. Some classes of orbit will be subjected only to an inward drift and others only to an outward drift. The lack of cancellation of the drift terms is further enhanced by the self-consistent coupling, increasing the absorption for waves propagating parallel to the plasma current, but not for waves propagating in the antiparallel direction. This results in a strong inward pinch also for symmetric wave spectra as well as for typical experimental spectra, with the dominant peak in the counter-plasma-direction.

ICRH has proved to be an efficient method of heating the plasma in toroidal devices. The high voltages needed at the coupling structure are, however, a severe handicap of this method. The possibility is investigated of having the highest voltages between the bars of the Faraday shield (FS), where they are both necessary and easier to maintain. For this purpose a resonant Faraday shield (RFS) antenna where the first and last bars of the FS are connected by an inductive strip is proposed. In front of this strip there is a second strip, fed, as in a conventional antenna, by an RF generator. It is shown that if the toroidal length of the FS is larger than λ/2 the strip connecting the bars of the FS acts as the secondary coil of a tuned transformer, the strip fed by the generator being the primary. It is therefore possible, by varying the frequency and the distance between the two strips, i.e. the coupling coefficient, to match the impedance of the primary to that of the generator.

A `common basis' systems study of superconducting (SC) and normal conducting (NC) DT burning fusion power and materials testing reactor designs is presented. The figures of merit for power and materials testing reactors are, respectively, projected cost of electricity (COE) and direct cost. A common 0-D plasma modelling basis is used and the plasma geometry and engineering aspects of the SC and NC designs are treated in an equivalent manner that is consistent with the limitations of their respective magnet technologies. Aspect ratios A in the range 1.2⩽A⩽6 and plasma elongations κ in the range 1.5⩽κ⩽3 are explored and an MHD stability (beta limit) physics basis that accurately describes the increase of normalized beta β_N and toroidal beta β_T with decreasing A and/or increasing κ is incorporated. With this MHD basis taken into account and with the usual reactor geometry, physics and engineering constraints and costing bases applied, the results of the study show that for SC power reactor designs with κ = 2 the COE has a minimum for 2⩽A⩽3 and increases with a further increase in A(A>3). For NC power reactors the COE has minima at A≈2. For both SC and NC power reactors, the minima are more apparent with lower κ. While SC options appear to offer lower COE for power plants, the direct cost for NC test reactors with similar fusion power output is significantly lower. Within the NC category, test designs that combine modest A and maximum elongation show promise for achieving ITER-like testing capabilities at low direct cost. For example, an NC coil design with A = 2, κ = 3 could produce fusion power of 200 MW at 1.23 MW/m² average neutron wall loading at a total direct cost of about 650 million US dollars. This NC design with a fissile blanket could also convert about 1270 kg of fission reactor waste per full power year. A possible cost effective development scenario for fusion power is identified for NC A = 2 toroidal devices for physics and material testing studies for use in the near future. The selection of the SC or NC coil option could then be made for the construction of the demonstration power reactor.

The effects of variations in net toroidal current on the Mercier criterion are studied in a heliotron configuration under the free boundary condition. The plasma column and the magnetic axis are shifted horizontally by the net toroidal current because the free boundary condition changes the effective vertical field. The direction of the shift depends on the direction of the current. In the case of a subtractive current density peaked at the axis, which decreases the rotational transform, the equilibrium is more stable than the no net current equilibrium, although the plasma axis is shifted inwards by the current itself. This is because the Shafranov shift is greater than the inward axis shift caused by the current itself, and the magnetic well is enhanced. In the case of additive currents which increase the rotational transform, there is a tendency for the stability to be improved as the peak position of the current density varies from the axis to the peripheral region. This improvement is due to the enhancement of the magnetic shear at low beta and the magnetic well at high beta. As in the no net current case, the equilibrium with an additive hollow current under the free boundary condition is more stable than that under the fixed boundary condition because magnetic well stabilization is enhanced. When the subtractive current peaked at the axis and the additive hollow current flow simultaneously so that the total current is zero or slightly additive, the interchange mode can be stabilized even if the corresponding equilibrium with no net current is strongly Mercier unstable.

The variation of equilibrium in the tokamak plasma is derived by solving the integral form of the linearized Grad-Shafranov (GS) equation. The state space description of the evolution of the plasma equilibrium is obtained by combining the solutions to the linearized GS equation with the circuit equation of flux conservation. Vertical displacements of plasmas are analysed using the equilibrium of the compact ITER, and the results are compared with a rigid model of plasma evolution. In the case of the compact ITER, the linear growth rates derived by the linearized GS equation are larger than those derived by the rigid model. The quantitative characterization of the non-rigid motion of a plasma is discussed. To assess the validity of the simulation method, a minor disruption observed in a JT-60U experiment is simulated and good agreement between the simulation and the experiment is found.

The peculiarities of the electron temperature profile under dominant on- and off-axis heating in the Rijnhuizen Tokamak Project (RTP) were successfully described by a `shell model'. This model features electron transport barriers that are associated with simple rational values of the safety factor. In the present work, this model is tested against JET data. To this end, the model is incorporated into the JETTO code, which is used to simulate JET optimized shear discharges with pronounced electron transport barriers. It is found that the model is successful in reproducing the time and location of barrier formation, as well as the subsequent evolution of barrier location and strength.

Adequate helium enrichment and exhaust have been achieved in reactor relevant ELMy H mode plasmas in JET performed in the Mark II divertor geometry. These quantities, describing the retention of impurities in divertors, have been experimentally inferred from CXRS measurements in the core plasma, and from a spectroscopic analysis of a Penning gauge discharge in the exhaust gas. The retention of helium was found to be sufficient with respect to the requirements in a next step device, with helium enrichment factors exceeding 0.1 in high density ELMy H mode discharges. With increasing core plasma density the helium partial pressure in the exhaust channel increases. While in L mode plasmas the helium enrichment decreases with increasing core plasma density, it remains almost constant in ELMy H mode plasma. The noble gas neon is better enriched in the divertor at high core plasma densities in both confinement regimes. These experimental results can be explained by the significant differences between the penetration depths of the impurity neutrals and by their subsequent different impurity transport mechanisms. Analytical and numerical analyses of these plasmas using the impurity code package DIVIMP/NIMBUS support the proposition that, owing to their much longer ionization mean free path, helium particles can escape from the divertor chamber as neutrals, while neon escapes by means of ion leakage. Consequently, the divertor plasma conditions strongly influence the noble gas compression and enrichment. Variations of the divertor plasma configuration and modifications to the divertor geometry have enhanced the pumping capabilities of the tokamak, but have been found not to affect the helium enrichment.

In LHD discharges, the NBI heated plasmas are terminated in two ways: (a) thermal decay (TD) after the termination of NBI and (b) radiative collapse (RC) during the NBI heating. The basic characteristics of the TD and RC discharges are compared. It is found that the decay and collapse of the plasma are mainly governed by the heating power and the plasma density. The critical density _c for the collapse of RC plasmas is similar to the scaling laws obtained in other helical devices, i.e. _c∝(PB/V)^0.5, where P, B and V denote heating power, magnetic field and plasma volume, respectively. Moreover, measurements using multichannel bolometric diagnostics indicate that the total radiation profiles in TD and RC plasmas are usually inboard-outboard symmetric and asymmetric, respectively, at the end of the discharge. In RC discharges, the total radiation profile develops in several phases. Before the onset of the thermal instability (TI), the radiation profile is rather symmetric, while after that, the radiation profile evolves from being symmetric in the initial period towards being asymmetric eventually with high radiation on the inboard side. Corresponding variations are shown in the time evolutions of the density and temperature profiles, and a substantial contraction of the plasma column is observed immediately after TI onset. The spatial and temporal coincidence of the asymmetries in the radiation, density and temperature is similar to that observed with multifaceted asymmetric radiation from the edge (MARFE) in tokamaks. But, unlike MARFEs, the asymmetric radiation (AR) in LHD is rather transient since it appears just before the end of RC discharges. The underlying cause for the development of radiation asymmetry was investigated and compared with existing instability models. The result suggests that the high inboard radiation is a manifestation of an enhanced local thermal instability, and the AR results from asymmetric developments of TI on the inboard-outboard sides during the final stage of RC discharges.

The DIII-D tokamak evolved from the earlier Doublet III device in 1986. Since then, the facility has undergone a number of changes including the installation of divertor baffles and pumping chambers in the vacuum vessel, the addition of a radiation shield, the development of extensive neutral beam and RF heating systems, and the addition of a comprehensive plasma control system. The facility has become the focus of a broad fusion plasma science research programme. An integrated picture of the facility and its capabilities is presented.

Owing to a coding error in the electronic source file of this paper, certain expresssions were printed incorrectly. These are corrected at the following places.

Page 6, section 2, line 3: β_N 1

Page 6, section 2, paragraph 2, line 20: 0 m 4

Page 8, nine lines above section 3: β_N (> 1.5)

Page 8, section 3.2, line 8: (6.107 t 6.1083 s)

Page 10, section 4.1, lines 4, 5: (κ 1.8) and/or high triangularity (δ 4.5)

Page 10, caption of figure 8, last line: (d/a 1.3)

Page 11, section 4.2, line 6: d/a 1.3

These corrections have been made in the online version.