Search

A local island divertor (LID) experiment has begun on LHD, with the aims of controlling edge recycling and improving the plasma confinement. The fundamental divertor functions of the LID have been demonstrated in the recent experiments. From the particle flux profile measurements on the LID head it was found that the particles diffusing out from the core region are well guided along the island separatrix to the LID head. Owing to the closed configuration around the LID head, evidence of the high efficient pumping was observed, together with a strong capacity to screen impurities. The first results of edge modeling using the EMC3-EIRENE code are also presented.

Divertor footprints have been identified within a few mm in accuracy after 10 000 shots. This is the large merit of the large helical device divertor for erosion/deposition studies due to high reproducibility with an external superconducting coils system. Helical distribution of divertor erosion is compared with the prediction from magnetic field characteristics. The measured net erosion depth is found to be about a factor 3 less than the estimated one. Numerical simulations have revealed the net erosion to be very sensitive to deposition of C impurity in the plasma. Eroded carbon atoms are mainly redeposited near the divertor tiles, and partly deposited near the divertor strike point, forming a mixed layer with promptly deposited metals. Deposited metals accumulate locally at the edge of microscale open pores and around grains of graphite. This kind of metals sink possibly plays an important role as an impurity source after the tiles installation. This aspect of ‘microscopic-PSI study’ is very informative for understanding macroscopic-PSI.

The first design of Carbon Sheet Pump (CSP) proposed in the Large Helical Device (LHD) project to achieve high plasma performances with low overall recycling ratio is presented. The CSP sheets of C/C composite are designed to be mounted on the vessel wall near the divertor plate, and operated in the cycle of (1) pumping of charge-exchanged (CX) fast H neutrals by trapping them in the sub-surface layers of the carbon at temperatures below 200°C in the total duration time of about 100 s, (2) degassing of trapped hydrogen by direct Joule heating of the carbon sheet at a temperature above 800°C in a period between main discharges, and then (3) cooling down to the operation temperature below 200°C. The first demonstration of CSP carried out successfully in a hydrogen glow discharge chamber is presented.

The Large Helical Device (LHD) is a heliotron/torsatron-type confinement device (B = 4 T, R = 3.9 m) equipped with a helical divertor. In the LHD configuration, the plasma region is shifted inwards approximately a third of the plasma minor radius relative to the center of the two pairing helical coils, thus making the distance between the coil center and the edge plasma small, 328 mm on the small major radius side of the torus. Within this space, many components must be installed, such as the superconducting helical coil (163), its coil can (45), the thermal shield (30), the vacuum gap (35), the plasma vacuum vessel (15), the first wall (25) (the numbers in the parentheses are allocated radial space in mm for the respective components). Effective edge plasma control by the divertor requires a space of more than 15 mm between the plasma and the first wall. To meet the above space requirement, the thickness of the helical coil is designed to be as small as possible, yet the coil current density and maximum field strength are within the limits of reliable coil operation.

Irregular magnetic fields resonate with the rational surface of the magnetic confinement systems, form magnetic islands and ergodic layers, and destruct the plasma confinement. To avoid this confinement destruction the requirement of an accuracy of 10 −4 in the magnetic field is adopted as the magnetic-accuracy design criterion for the LHD machine. Following this criterion the width of the undesirable magnetic island is kept less than one tenth of the plasma radius. The irregular magnetic field from the superconducting (SC) helical and poloidal coils is produced by winding irregularity, installing irregularity, cooling-down deformations and electromagnetic deformations. The local irregularities such as feeders, layer connections, adjacent-conductor connections of the coils also produce an error field. The eddy currents on the supporting shell structure of SC coils, the cryostat, etc. are also evaluated. All irregular effects are analyzed using Fourier decomposition and field mapping methods for the LHD design, and it is confirmed that the present design of the superconducting coil system satisfies the design criterion for these field irregularities.

A steady-state H-mode plasma with frequent ELMs (edge localized modes) was obtained by applying magnetic fields with high poloidal mode numbers resonant in the plasma edge. The induced ELMs limit the density and impurity accumulation normally observed during ELM-free H-mode. An EML coil with numerous resonant modes ( m 10) in the plasma edge appears the most effective. This method of control is extended to higher auxiliary heating power by increasing the magnitude of the magnetic perturbations.

The first experimental test of the ergodic magnetic limiter concept is being conducted on the TEXT device. Modular coils have been installed on TEXT allowing the external generation of resonant helical field perturbations. Preliminary results are presented.

An instrumented scoop limiter probe is being operated on TEXT to test the concept of limiter cooling and improved particle removal efficiencies using an externally-applied resonant magnetic field perturbation (the resonant helical divertor concept). Cooling of the limiter face has been demonstrated for limiter positions ranging from r L = 29.0 cm inward to r L = 25.5 cm (the Text primary poloidal hoop limiter radius r a = 27.0 cm). Pressure rises in the limiter throat of approximatel 40% are observed under optimized conditions. Interchangable limiter heads with thicknesses of 1.0 cm and 0.3 cm have been used to examine particle ducting into the scoop aperture. Experimental results are discussed along with observations of the limiter floating potential, H α recycling emissions, pressure measurements, and edge density and temperature measurements.

Discharges with improved energy confinement are produced in Doublet III when the major recycling location is shifted away from the main plasma by forming an expanded boundary (XB) divertor configuration with a separatrix to limiter separation Δ > 1.5 cm. Monte Carlo calculations indicate that over a wide range in density the divertor plasma effectively limits the penetration of recycled neutrals to the main plasma. This results in large particle recycling rates at the divertor plates and a reduction of particle recycling near the main plasma. This divertor function apparently plays an important role in obtaining improved energy confinement. Particle containment, as inferred from variations in Dα emission near the main plasma, is found to decrease with n e and BT, and increases rapidly (faster than quadratic) with IP whereas energy confinement is relatively insensitive to variations in n e and BT and increases linearly with IP. This improved energy confinement with the XB divertor configuration is suggestive of particle confinement playing an important role. However, in view of their rather different parametric dependences, the correlation between τP and τE, if any, remains unclear.

The ergodic magnetic limiter coils on TEXT have been reconfigured to produce the primary helical perturbation resonance at m = 7 n = 3 . The experiments continue to demonstrate that the weak resonant perturbations modify the edge conditions in keeping with model predictions. We observe a reduction in the intrinsic impurity levels accompanying the helical current pulse, presumably the result of a reduction in the electron temperature in the edge. Heat follows the perturbed field lines to the limiter, generating heat load patterns which reflect the geometry of a magnetic island-limiter intersection. A strong spatial modulation of the electron density in the scrape-off-layer also reflects the helical mode structure.

A series of initial resonant helical divertor (RHD) experiments have been carried out in ohmically and auxiliary heated JIPP T-IIU plasmas. Disruptive and MHD instabilities make the interpretation of the RHD results difficult but an apparent increase in the energy confinement time is observed when the helical magnetic perturbation is applied. This may be due to the suppression of MHD activity or to a reduction in the edge convective heat losses. Magnetic island effects have been observed on the floating potential of a Langmuir probe array and energy scrape-off layer widths have been measured with and without helical perturbations during ICRF operation. Basic pump limiter data is presented including ion temperatures and C4+ impurity profiles. Energy confinement times are reported in ohmically and NBI heated discharges.

The improved energy confinement of the H-mode is invariably coupled to improved particle confinement, In other experiments, increased particle confinement resulted in impurity accumulation and severe power losses. In DIII-D, similar to the results of other experiments, H-mode confinement is associated with improved particle confinement. But in contrast to other devices, in most DIII-D H-mode plasmas there is no indication of impurity accumulation. In high current shots the impurity content of plasma during the H-mode is lower than the L-mode phase. Two factors most responsible for the low impurity content of DIII-D H-mode plasmas are the nearly flat or hollow electron density profiles and the presence of giant Edge Localized Modes (ELMs). Giant ELMs also cause large amplitude modulations in particle and power flow to the divertor target plates, resulting in large scale changes in the parameters of the the scrape-off layer plasma. Our analysis of temperature and density profiles during ELMs show that particle loss during ELMs is primarily due to perpendicular transport across the field lines.