Table of contents

Neoclassical transport is studied for large helical device (LHD) configurations in which the magnetic axis has been shifted radially by determining the mono-energetic transport coefficient and the effective helical ripple. With respect to the transport in the long mean-free-path collisionality region—the so-called 1/ν transport—the optimum configuration is found when the magnetic axis has a major radius of 3.53 m, which is 0.22 m inward shifted from the `standard' configuration of LHD. In the optimized case, the effective helical ripple is very small, remaining below 2% inside (4/5) of the plasma radius. This indicates that a strong inward shift of the magnetic axis in the LHD can diminish the neoclassical transport to a level typical of so-called `advanced stellarators'.

The possibility of fulfilling the condition of poloidal closure of the contours of the second adiabatic invariant for all reflected particles is studied for systems with poloidally closed contours of the magnetic field B on the magnetic surfaces through computational stellarator optimization. It is shown that by adjusting the geometry this is possible in a major fraction of the plasma volume. The most salient characteristic (as compared to previous quasi-isodynamic configurations) is a magnetic axis whose curvature vanishes in all cross sections with an extremum of B on the magnetic axis and renders possible a three-dimensional structure of B with unprecedentedly high collisionless α-particle confinement.

In response to social demand, this paper investigates the breakeven price (BP) and potential electricity supply of nuclear fusion energy in the 21st century by means of a world energy and environment model. We set the following objectives in this paper: (i) to reveal the economics of the introduction conditions of nuclear fusion; (ii) to know when tokamak-type nuclear fusion reactors are expected to be introduced cost-effectively into future energy systems; (iii) to estimate the share in 2100 of electricity produced by the presently designed reactors that could be economically selected in the year. The model can give in detail the energy and environment technologies and price-induced energy saving, and can illustrate optimal energy supply structures by minimizing the costs of total discounted energy systems at a discount rate of 5%. The following parameters of nuclear fusion were considered: cost of electricity (COE) in the nuclear fusion introduction year, annual COE reduction rates, regional introduction year, and regional nuclear fusion capacity projection. The investigations are carried out for three nuclear fusion projections one of which includes tritium breeding constraints, four future CO₂ concentration constraints, and technological assumptions on fossil fuels, nuclear fission, CO₂ sequestration, and anonymous innovative technologies. It is concluded that: (1) the BPs are from 65 to 125 mill kW⁻¹ h⁻¹ depending on the introduction year of nuclear fusion under the 550 ppmv CO₂ concentration constraints; those of a business-as-usual (BAU) case are from 51 to 68 mill kW⁻¹h⁻¹. Uncertainties resulting from the CO₂ concentration constraints and the technological options influenced the BPs by plus/minus some 10–30 mill kW⁻¹h⁻¹, (2) tokamak-type nuclear fusion reactors (as presently designed, with a COE range around 70–130 mill kW⁻¹h⁻¹) would be favourably introduced into energy systems after 2060 based on the economic criteria under the 450 and 550 ppmv CO₂ concentration constraint, but not selected under the BAU case and 650 ppmv CO₂ concentration constraint, and (3) the share of electricity in 2100 produced by the presently designed tokamak-type nuclear fusion reactors (introduced after 2060) is well below 30%. It should be noted that these conclusions are based upon varieties of uncertainties in scenarios and data assumptions on nuclear fusion as well as technological options.

The current ramp down phase of the Frascati Tokamak Upgrade plasma has been found to exhibit very different particle transport than that which is observed in the Ohmically heated steady state phase. The current ramp down phase can be characterized by a central (r/a<0.5) region that has anomalous transport similar to that seen in the steady state phase and an outer region (r/a>0.5) that has transport two to three times less. The temperature in the inner region had a Gaussian-like shape as a function of minor radius. In the outer region, the temperature had roughly a single value (∼100 eV) with no sharp gradients which implied no gradient driven transport. To study the particle transport, iron was introduced into the ramp down phase by using the laser blow off technique. The time behaviour of several charge states of iron have been measured spectroscopically to determine the transport in the different spatial locations in the plasma. The time histories of the charge state line emission in the ramp down phase have been observed to decay much more slowly than in the steady state phase which implied the existence of a transport barrier at the half radius. The behaviour of the impurity charge states has been modelled with the MIST transport code using up to date atomic physics rates to infer the particle transport.

The Flinders ST is a spherical tokamak project which has concentrated on the production by induction, of a low resistivity target plasma suitable for rotating magnetic field (RMF) current drive tests. Antennas for the production of the RMF fields were installed on the spherical tokamak device, so that it could operate as a device with pure RMF current drive (Rotamak), or purely as an inductively driven ST. Experiments with combinations of inductive and RMF current drive have been made. Experimental results obtained in this small Flinders ST device indicate areas of concern for the RMF current drive technique. First, the experiments show that the strong steady magnetic fields required for the spherical tokamak configuration degrade, and probably extinguish the RMF created plasma, when these transverse steady fields exceed some critical value, which is small compared to that required for ST operation. Second, a measure of the resistance of the pure RMF current drive device, using the ST solenoid and OH coils to create a small loop voltage purely as a diagnostic, shows that the RMF steady current is driven in a very resistive path.

It is shown that in the Helias reactor [1], depending on the ratio γ_L/γ_d with γ_L(d) the linear growth (damping) rate of the Alfvén eigenmode (AE), there are three regimes of the transport of α-particles resonant with AEs. (i) For 1<γ_L/γ_d<<10 the transport is determined solely by the pitch-angle scattering of αs trapped by an AE. In this regime τ_α>>τ_s, with τ_α(s) the α-particle confinement (slowing down) time. (ii) For γ_L/γ_d∼10 the dominant transport mechanism is the adiabatic convection of the slowing-down αs trapped by an AE, with particle `life time' in resonance with AE limited by pitch-angle scattering. In this regime τ_α∼τ_s. (iii) For γ_L/γ_d>> 10 the transport become stochastic with τ_α<<τ_s.

Measurements of deuterium retention in samples of lithium exposed in the liquid state to deuterium plasma are reported. Retention was measured as a function of plasma ion dose in the range 6×10¹⁹–4×10²² D atoms and exposure temperature between 523 and 673 K using thermal desorption spectrometry. The results are consistent with the full uptake of all deuterium ions incident on the liquid metal surface and are found to be independent of the temperature of the liquid lithium over the range explored. Full uptake, consistent with very low recycling, continues until the sample is volumetrically converted to lithium deuteride. This occurs for exposure temperatures where the gas pressure during exposure was both below and slightly above the corresponding decomposition pressure for LiD in Li.

A new scenario to delay or prevent neoclassical tearing mode (NTM) onset is presented. By active sawtooth destabilization, short period and low amplitude sawteeth are generated, such that the sawtooth produced NTM seed island is reduced and the threshold normalized plasma pressure for triggering of NTMs, β_Nonset, is increased. The scenario has been explored experimentally in the Joint European Torus (JET). Ion cyclotron resonance frequency (ICRF) waves tuned to the 2nd harmonic H-minority resonance have been used for sawtooth control. Whereas ICRF waves generally induce sawtooth stabilization, favouring the triggering of NTMs and reducing β_Nonset, the present experiments show that by toroidally directed waves, ion cyclotron current drive is produced, and that sawteeth can be destabilized by careful positioning of the 2nd harmonic H resonance layer with respect to the sawtooth inversion radius. As a result, NTM onset is delayed and β_Nonset is increased above its value obtained in discharges with additional heating from neutral beam injection alone.

Rapid plasma toroidal rotation, sufficient for stabilization of the n = 1 resistive wall mode, can be sustained by improving the axisymmetry of the toroidal magnetic field geometry of DIII-D. The required symmetrization is determined experimentally both by optimizing currents in external n = 1 correction coils with respect to the plasma rotation, and by use of the n = 1 magnetic feedback to detect and minimize the plasma response to non-axisymmetric fields as β increases. Both methods point to an intrinsic ∼7 G (0.03% of the toroidal field), m/n = 2/1 resonant helical field at the q = 2 surface as the cause of the plasma rotation slowdown above the no-wall β limit. The drag exerted by this field on the plasma rotation is consistent with the behaviour of `slipping' in a simple induction motor model.

The effect of nuclear elastic scattering (NES) of deuterium and tritium in close collisions with hot alpha-particles is considered. Calculations of the perturbed deuterium and tritium distribution functions are based on numerical solution of the three-dimensional trajectory averaged kinetic equation. The equation includes the full drift trajectory averaged Coulomb collision operator and the NES particle source. Modelling is carried out for parameters relevant to JET DT experiments and results of calculations are compared with typical measurements of NES deuterons. In addition, we evaluate the maximum achievable NES perturbation to the DT plasma. Finally, a potential opportunity for estimation of alpha-particle pitch-angle anisotropy, using measurements of NES perturbed deuterium or tritium distributions is discussed. We show that it is important to account for NES effects in order to accurately diagnose the alpha-particle distribution.