Table of contents

A potential new standard in stationary tokamak performance is emerging from experiments on DIII-D. These experiments have demonstrated the ability to operate near the free boundary, n = 1 stability limit with good confinement quality under stationary conditions. The normalized fusion performance is at or above that projected for Q_fus = 10 operation in the International Thermonuclear Experimental Reactor (ITER) design over a wide operating range in both edge safety factor (2.8–4.7) and plasma density (35–70% of the Greenwald density). Projections to ITER based on this data are uniformly positive and indicate that a wide range of operating options may be available on ITER, including the possibility of sustained ignition. Recent experiments have demonstrated the importance of a small m = 3, n = 2 neoclassical tearing mode in avoiding sawteeth and the effect of edge localized modes on tearing mode stability at an edge safety factor near 3. Transport studies using the GLF23 turbulence transport code suggest that E × B shear stabilization is important in reproducing the measured profiles in the simulation. Yet, even in cases in which the toroidal rotation is moderate, confinement quality is robustly better than the standard H-mode confinement scalings.

We have initiated an experimental programme to address some of the questions associated with the operation of a tokamak with high bootstrap current fraction under high performance conditions, without assistance from a transformer. In these discharges, stationary (or slowly improving) conditions are maintained for up to 3.7 s at β_N ≈ β_p approaching 3.3. The achievable current and pressure are limited by a relaxation oscillation, involving growth and collapse of an internal transport barrier at ρ ≥ 0.6. The pressure gradually increases and the current profile broadens throughout the discharge. Eventually the plasma reaches a more stable, high confinement (H_89P ∼ 3) state. Characteristically these plasmas have 65–85% bootstrap current, 15–30% neutral-beam-driven current and 0–10% driven by electron cyclotron frequency electromagnetic waves.

Using a global Landau-fluid code in toroidal geometry, an electromagnetic ion temperature gradient (ITG) driven turbulence–zonal mode system is investigated. Two different types of zonal flows, i.e. stationary zonal flows in a low q (safety factor) region and oscillatory ones in a high q region, which are called geodesic acoustic modes, are found to be simultaneously excited in a torus. The stationary flows efficiently suppress turbulent transport, while the oscillatory ones weakly affect the turbulence due to their time varying nature. Therefore, in the low q region where the zonal flows are almost stationary, they are dominant over the turbulence. On the other hand, the turbulence is still active in the high q region where the zonal flows are oscillatory.

A Nb₃Al superconductor inherently has the outstanding features of a high critical field and an excellent strain tolerance in critical current performance. The Japan Atomic Energy Research Institute developed the world's first large Nb₃Al coil using jelly-roll processed Nb₃Al strands. This coil was successfully operated up to the nominal current and field of 46 kA and 13 T, respectively. The test results demonstrated that a Nb₃Al conductor is suitable for application in high field, large magnets, such as the toroidal field (TF) coils in a fusion reactor. In parallel with this work, a Nb₃Al strand having a high critical current at high field has been developed by the National Institute of Materials Science. This strand cannot incorporate enough copper stabilizer, resulting in poor stability, since it is heat-treated at a much higher temperature than the melting temperature of copper. In order to find a solution to this issue, we performed an analytical study, and it showed that the externally incorporated copper after the high temperature heat treatment is effective for stabilization when the electric conductance and heat transfer coefficient between the Nb₃Al strand and the external copper are more than 10 MS m⁻¹ and 10 kW m⁻² K⁻¹, respectively. Since they seem to be attained by using present conductor technologies, the development of a TF coil operated at a high field around 16 T seems promising.

Most tokamak experimental results (Bessenrodt-Weberpals M. et al1993 Nucl. Fusion33 1205) indicate dependence of the ion thermal conductivity on the isotopic mass close to , i.e. inverse gyro-Bohm in mass scaling. This is in stark contradiction to most present theoretical models predicting Bohm or gyro-Bohm scaling. A basic physics experiment (Sokolov V. and Sen A.K. 2002 Phys. Rev. Lett.89 095001) on the anomalous ion thermal conduction due to ion temperature gradient instabilities in two different gases (hydrogen and deuterium) closely confirms the tokamak results.

We now report another series of experiments designed to explore the physical basis of this scaling, which appears to lead to a new model for this scaling based on three-wave coupling of two ion temperature gradient radial harmonics and an ion acoustic (IA) wave. The resulting isotopic scaling of transport is , dictated primarily by the IA damping. This basic physics is deemed to be extrapolatable to tokamaks resolving the paradox and is tantamount to a new paradigm for plasma turbulent transport.

Predictive transport modelling and gyrokinetic stability analyses of demonstration hybrid (HYBRID) and advanced tokamak (AT) discharges from the International Tokamak Physics Activity (ITPA) profile database are presented. Both regimes have exhibited enhanced core confinement (above the conventional ITER reference H-mode scenario) but differ in their current density profiles. Recent contributions to the ITPA database have facilitated an effort to study the underlying physics governing confinement in these advanced scenarios. In this paper, we assess the level of commonality of the turbulent transport physics and the relative roles of the transport suppression mechanisms (i.e. E × B shear and Shafranov shift (α) stabilization) using data for select HYBRID and AT discharges from the DIII-D, JET and AUG tokamaks. GLF23 transport modelling and gyrokinetic stability analysis indicate that E × B shear and Shafranov shift stabilization play essential roles in producing the improved core confinement in both HYBRID and AT discharges. Shafranov shift stabilization is found to be more important in AT discharges than in HYBRID discharges. We have also examined the competition between the stabilizing effects of E × B shear and Shafranov shift stabilization and the destabilizing effects of higher safety factors and parallel velocity shear. Linear and nonlinear gyrokinetic simulations of idealized low and high safety factor cases reveal some interesting consequences. A low safety factor (i.e. HYBRID relevant) is directly beneficial in reducing the transport, and E × B shear stabilization can dominate parallel velocity shear destabilization allowing the turbulence to be quenched. However, at low-q/high current, Shafranov shift stabilization plays less of a role. Higher safety factors (as found in AT discharges), on the other hand, have larger amounts of Shafranov shift stabilization, but parallel velocity shear destabilization can prevent E × B shear quenching of the turbulent transport, and only E × B suppression is achieved.

Experimental investigation of the Scrape-Off Layer (SOL) turbulence and the perpendicular anomalous particle transport in the high density regime in the T-10 tokamak is presented. Comparison of the plasma parameters measured at the low-field side (LFS) and the high-field side (HFS) of the tokamak shows that the inboard turbulence essentially differs from the outboard one. Turbulence has an intermittent character but the fluctuation level of the inboard ion saturation current is about two times lower than at the LFS. The radial turbulent particle flux and the effective perpendicular diffusion coefficient measured at the LFS are about 3–5 times higher than those observed at the HFS. Enhanced perpendicular particle flux at the LFS is defined by the high density plasma structures formed in the vicinity of the last closed flux surface. The poloidal electric field inside the high density structures has different directions at the inboard and outboard sides of the plasma column. The radial electric drift induced by the poloidal electric field is directed towards the vacuum vessel at both the LFS and the HFS. These experiments support the hypothesis that SOL parallel flow might be caused by an additional mechanism such as 'ballooning' transport, which generates a larger flux of particles from the core into the SOL at the outer mid-plane.

The results of predictive self-consistent modelling of plasma parameters in low (L) and high (H) confinement modes by the one-dimensional transport code RITM, with particular emphasis on the properties of the edge transport barrier, are presented and discussed. The same transport model is used under both L- and H-mode conditions and includes contributions from ion temperature gradient (ITG), trapped electron, drift Alfvén (DA) and drift resistive ballooning instabilities described in the fluid approximation. The computations predict the formation of the edge transport barrier at a high enough heating power due to stabilization of ITG and DA modes, dominating the edge transport in the L-mode, through the effects of the density gradient and the pressure gradient at low collisionality, respectively. The calculated radial profiles and scalings for pedestal and confinement characteristics are compared with measurements on JET, DIII-D and JT-60U tokamaks.

The change in nonlinear energetic particle mode (EPM) dynamics that accompanies the transition from weak to strong energetic ion transport is discussed in this work. It is demonstrated that the nonlinear threshold in fast ion energy density for the onset of strong convective transport occurring in avalanches is close to the linear EPM excitation threshold. This phenomenology is strictly related to the resonant character of the modes, which tend to be radially localized where the drive is strongest. After the convective loss phase, during which nonlinear EPM mode structure is displaced outwards, fast ion transport continues owing to diffusive processes. Theoretical analyses, presented here, are the basis for consistency analyses of operation scenarios in proposed burning plasma experiments. Comparisons between theoretical predictions and both simulation and experimental results are also briefly discussed.

High-power ion–cyclotron resonance heating (ICRH) can produce centrally peaked fast ion distributions with wide non-standard drift orbits exciting Alfvén eigenmodes (AEs). The dynamics of the AE excitation depends not only on the anisotropy and the peaking of the fast ion distribution but also on the decorrelation of the AE interactions and the renewal of the fast ions resonant with the AE by ion–cyclotron interactions. A method of self-consistently including the evolution of the distribution function of fast ions during excitation of AEs and ICRH has been developed and implemented in the SELFO code. Numerical simulations of the AE dynamics and ICRH give a variation of the AE amplitude consistent with the experimentally observed splitting of the mode frequency. The experimentally observed fast damping of the mode as the ICRH is switched off is also evident in the simulations.

Two experiments on DIII-D have been performed with the purpose of searching for evidence of a critical electron temperature gradient or gradient scale length. Both experiments employed off-axis electron cyclotron heating to vary the local value of inverse temperature scale length while holding the total heating power and thus edge temperatures constant. The incremental or heat pulse thermal diffusivity was monitored as a means of identifying the existence of a threshold or critical value of , k_crit, where a large increase in heat pulse diffusivity may occur while going from below to above k_crit. No evidence of a critical value was found, but the existence of one cannot be ruled out by the experimental results. The results can be used to place an upper bound on k_crit. If k_crit exists, the experimental results indicate k_crit < 3.8 m⁻¹ at ρ = 0.45 and k_crit < 2.5 m⁻¹ at ρ = 0.29 corresponding to temperature scale lengths relative to the plasma minor radius of and 0.65, respectively. Over most of the plasma radius (ρ ⩾ 0.3) the temperature profile in the L-mode discharges studied exceeds these upper bounds on k_crit. A transport model based on a critical gradient is consistent with experimental observations but is not unique. Models other than one based on the existence of a critical gradient are also consistent with experimental observations.

An integrated radiative high performance scenario has been established at ASDEX Upgrade based on simultaneous feedback control of the average divertor neutral particle and power flux in combination with a high, pellet induced frequency of edge localized modes (ELMs). This approach is fully compatible with the present tungsten wall coating covering about 65% of the plasma facing components and is intended for application in the envisaged full-tungsten experiment. In these experiments, divertor recycling and effective divertor temperature (derived from thermoelectric currents) were tuned by acting on fuel gas puff and argon injection rates. The ELM frequency (f_ELM) was kept high by repetitive injection of small cryogenic deuterium pellets to avoid the radiative instabilities seen at low f_ELM and high radiated power, and to control the ELM energy. No confinement loss is observed in this radiative type-I ELMy scenario with relatively flat density profiles. In contrast, similar type-III ELM scenarios achieved in hydrogen show a confinement loss of 25% as compared to the type-I phase. In parallel to pellets, alternative ELM trigger techniques have been investigated as well. Fast vertical plasma oscillations are able to synchronize the ELM frequency to values higher and lower than the intrinsic f_ELM, but remain to be tested in the integrated scenario. Supersonic gas injection showed better fuelling efficiencies than usual gas puffing but instantaneous ELM release has not been achieved. A particular experimental challenge for AUG conditions is to obtain a high pace making frequency, to establish scalings of confinement and energy loss as a function of controlled ELM frequency.

The ITER divertor vertical target has to sustain heat fluxes up to 20 MW m⁻². The concept developed for this plasma facing component working at steady state is based on carbon fibre composite armour for the lower straight part and tungsten for the curved upper part. The main challenges involved in the use of such components include the removal of the high heat fluxes deposited and mechanically and thermally joining the armour to the metallic heat sink, despite the mismatch in the thermal expansions. Two solutions based on the use of a CuCrZr hardened copper alloy and an active metal casting (AMC®) process were investigated during the ITER EDA phase: the first one called 'flat tile geometry' was mainly developed for the Tore Supra pumped limiter, the second one called 'monoblock geometry' was developed by the EU Participating Team for the ITER project. This paper presents a review of these two solutions and analyses their assets and drawbacks: pressure drop, critical heat flux, surface temperature and expected behaviour during operation, risks during the manufacture, control of the armour defects during the manufacture and at the reception, and the possibility of repairing defective tiles.

Recent experiments in the Current Drive Experiment-Upgrade (CDX-U) provide a first-ever test of large area liquid lithium surfaces as a tokamak first wall to gain engineering experience with a liquid metal first wall and to investigate whether very low recycling plasma regimes can be accessed with lithium walls. The CDX-U is a compact (R = 34 cm, a = 22 cm, B_toroidal = 2 kG, I_P = 100 kA, T_e(0)∼ 100 eV, n_e(0) ∼ 5 × 10¹⁹ m⁻³) spherical torus at the Princeton Plasma Physics Laboratory. A toroidal liquid lithium pool limiter with an area of 2000 cm² (half the total plasma limiting surface) has been installed in CDX-U. Tokamak discharges which used the liquid lithium pool limiter required a fourfold lower loop voltage to sustain the plasma current, and a factor of 5–8 increase in gas fuelling to achieve a comparable density, indicating that recycling is strongly reduced. Modelling of the discharges demonstrated that the lithium limited discharges are consistent with Z_effective < 1.2 (compared with 2.4 for the pre-lithium discharges), a broadened current channel and a 25% increase in the core electron temperature. Spectroscopic measurements indicate that edge oxygen and carbon radiation are strongly reduced.

The influence of toroidal sheared equilibrium flows on the nonlinear evolution of classical and neoclassical tearing modes is studied through numerical solutions of a set of reduced generalized MHD equations that include viscous force effects based on neoclassical closures. In general, differential flow is found to have a strong stabilizing influence leading to lower saturated island widths for the classical tearing mode and reduced growth rates for the neoclassical mode. Velocity shear, on the other hand, is seen to make a destabilizing contribution.

The concept of a power plant for a fast ignition heavy ion fusion is briefly reviewed. The reactor chamber response to x-ray irradiation, ion debris and neutron flux is considered. Ablation of a thin liquid film at the first wall is simulated by means of a one-dimensional hydrodynamic code. Ion debris reheating of the vaporized coolant is evaluated. Vaporization and subsequent condensation of the coolant is computed with the use of a kinetic model of fast condensation. Heating of the blanket due to neutron deposition and corresponding generation of pressure and stress waves are simulated by means of the Monte Carlo (MCNP) code and by the strength media mechanics code, respectively.