B. McCormack

Exploration of Spherical Torus Physics in the NSTX Device

The National Spherical Torus Experiment (NSTX) is being built at the Princeton Plasma Physics Laboratory to test the fusion physics principles for the Spherical Torus (ST) concept at the MA level. The NSTX nominal plasma parameters are R {sub 0} = 85 cm, a = 67 cm, R/a greater than or equal to 1.26, B {sub T} = 3 kG, I {sub p} = 1 MA, q {sub 95} = 14, elongation {kappa} less than or equal to 2.2, triangularity {delta} less than or equal to 0.5, and plasma pulse length of up to 5 sec. The plasma heating/current drive (CD) tools are High Harmonic Fast Wave (HHFW) (6 MW, 5 sec), Neutral Beam Injection (NBI) (5 MW, 80 keV, 5 sec), and Coaxial Helicity Injection (CHI). Theoretical calculations predict that NSTX should provide exciting possibilities for exploring a number of important new physics regimes including very high plasma beta, naturally high plasma elongation, high bootstrap current fraction, absolute magnetic well, and high pressure driven sheared flow. In addition, the NSTX program plans to explore fully noninductive plasma start-up, as well as a dispersive scrape-off layer for heat and particle flux handling.

Engineering design of the National Spherical Torus Experiment

NSTX is a proof-of-principle experiment aimed at exploring the physics of the ‘spherical torus’ (ST) configuration, which is predicted to exhibit more efficient magnetic confinement than conventional large aspect ratio tokamaks, among other advantages. The low aspect ratio (R:a, typically 1.2‐2 in ST designs compared to 4‐5 in conventional tokamaks) decreases the available cross sectional area through the center of the torus for toroidal and poloidal field coil conductors, vacuum vessel wall, plasma facing components, etc., thus increasing the need to deploy all components within the so-called ‘center stack’ in the most efficient manner possible. Several unique design features have been developed for the NSTX center stack, and careful engineering of this region of the machine, utilizing materials up to their engineering allowables, has been key to meeting the desired objectives. The design and construction of the machine has been accomplished in a rapid and cost effective manner thanks to the availability of extensive facilities, a strong experience base from the TFTR era, and good cooperation between institutions.

A Neutral Beam Injector Upgrade for NSTX

The National Spherical Torus Experiment (NSTX) capability with a Neutral Beam Injector (NBI) capable of 80 kiloelectronvolt (keV), 5 Megawatt (MW), 5 second operation. This 5.95 million dollar upgrade reused a previous generation injector and equipment for technical, cost, and schedule reasons to obtain these specifications while retaining a legacy capability of 120 keV neutral particle beam delivery for shorter pulse lengths for possible future NSTX experiments. Concerns with NBI injection included power deposition in the plasma, aiming angles from the fixed NBI fan array, density profiles and beam shine through, orbit losses of beam particles, and protection of the vacuum vessel wall against beam impingement. The upgrade made use of the beamline and cryo panels from the Neutral Beam Test Stand facility, existing power supplies and controls, beamline components and equipment not contaminated by tritium during DT [deuterium-tritium] experiments, and a liquid Helium refrigerator plant to power and cryogenically pump a beamline and three ion sources. All of the Tokamak Fusion Test Reactor (TFTR) ion sources had been contaminated with tritium, so a refurbishment effort was undertaken on selected TFTR sources to rid the three sources destined for the NSTX NBI of as much tritium as possible. An interconnecting duct was fabricated using some spare and some new components to attach the beamline to the NSTX vacuum vessel. Internal vacuum vessel armor using carbon tiles was added to protect the stainless steel vacuum vessel from beam impingement in the absence of plasma and interlock failure. To date, the NBI has operated to 80 keV and 5 MW and has injected requested power levels into NSTX plasmas with good initial results, including high beta and strong heating characteristics at full rated plasma current.

NSTX high temperature sensor systems

The design of the more than 300 in-vessel sensor systems for the National Spherical Torus Experiment (NSTX) has encountered several challenging fusion reactor diagnostic issues involving high temperatures and space constraints. This has resulted in unique miniature, high temperature invessel sensor systems mounted in small spaces behind plasma facing armor tiles, and they are prototypical of possible high power reactor first-wall applications. In the Center Stack, Divertor, Passive Plate, and vessel wall regions, the small magnetic sensors, large magnetic sensors, fluxloops, Rogowski coils, thermocouples, and Langmuir probes are qualified for 600 /spl deg/C operation. This rating will accommodate both peak rear-face graphite tile temperatures during operations and 350/spl deg/C bake-out conditions. Similar sensor systems including flux loops, on other vacuum vessel regions are qualified for 350 /spl deg/C operation. Cabling from the sensors embedded in the graphite tiles follows narrow routes to exit the vessel. The detailed sensor design and installation methods of these diagnostic systems developed for high-powered ST operation are discussed.

Making of the NSTX facility

The NSTX (National Spherical Torus Experiment) facility located at Princeton Plasma Physics Laboratory is the newest national fusion science experimental facility for the restructured US Fusion Energy Science Program. The NSTX project was approved in FY 97 as the first proof-of-principle national fusion facility dedicated to the spherical torus research. On Feb. 15, 1999, the first plasma was achieved 10 weeks ahead of schedule. The project was completed on budget and with an outstanding safety record. This paper gives an overview of the NSTX facility construction and the initial plasma operations.

Analysis of ion source arc chamber failure and grid damage sustained during high power operation

Several ion sources failed during the last months of TFTR operation. Four suffered are chamber vacuum leaks which admitted SF/sub 6/ into the source. One of these also had warped accelerator rails which resulted in damage to a water cooled scraper. The arc chamber leaks occurred at the boundary between the probe plate and either the bucket or interface plate. Vacuum seals at those locations consist of a mylar insulator sandwiched between two o-rings. At the failure point, the mylar had been punctured back to the o-ring. Are tracks were observed along the plasma facing surfaces of the probe plate and were funneled into every other gap between magnet. This evidence is suggestive of J/spl times/B forces driving the arcs toward the interface between the various plates and toward the o-ring. Damage is correlated with high arc power and the failure to expeditiously extinguish the arc. Operation with additional anode area is suggested as a means to avoid are creation. On the source with the warped grid rails, a section of a nearby scraper was melted. TFTR ion sources are masked down versions of the US Common Long Pulse Ion Source. Four rails on either end of the source are masked creating an ion beam that is /spl sim/43 cm tall. Rails under the mask were the most deformed (0.055 out of tolerance). Damage is believed to have occurred during 3 s pulses when 15 MJ were extracted over 3 s under slightly overdense conditions. The rails were deflected back towards the arc chamber causing the edge beamlets to be directed away from the axis.

The TFTR neutral beam pulse length limit

Early operation of the TFTR long pulse ion sources identified the full-energy ion dump as the critical constraint for long pulse operation. Full power, two second operation resulted in severe stress cracking of this component. Subsequent to this discovery, the dumps on all beamlines were replaced and new operating procedures implemented that restricted the surface temperature to that consistent with 120 kV for 0.65 s. In November, 1996 the ion dump in a tritium contaminated beamline was remotely inspected. No new damage was observed and the pulse length limit was relaxed to allow 95 kV, 5 A operation in support of enhanced reversed shear experiments. Two techniques have been proposed and partly tested that reduce ion dump power densities and permit longer pulse lengths. These are: 1) rastering of the ion beam across the ion dump, and 2) a vertical adjustment of the impact point on the full-energy dump. The latter of these techniques has applicability to the proposed Korean tokamak, KSTAR, which is currently using the TPX beamline (an upgraded TFTR beamline) as the basis for its design. Rastering consists of moving the ion beam laterally across the dump, thereby increasing the heated area. Based on static rastering experiments, the power density reduction due to several rastering patterns have been modeled. A 40% decrease in power density is predicted. Technique 2 stems from the curved nature of the full-energy ion dump-ions from the outer two ion sources strike lower on the dump than do ions from the center source. A small increase in the magnetic field moves the impact point higher up on the ion dump and farther from the magnets focal point. The result is a 20% reduction in power density.

Tritium consumption and retention in TFTR neutral beams

750,000 Ci of tritium have passed through the TFTR neutral beamlines since December, 1993. During the course of 725 tritium heated discharges, 47,000 Ci have been extracted as ions from the ion sources and 27,000 Ci injected into TFTR as neutral atoms. Prior to the commencement of deuterium-tritium (DT) experiments, Los Alamos and Sandia National Laboratories made estimates of tritium retention in each beamline due to implantation in the 2 m/sup 2/ of copper beam absorbers and absorption on the 250 m/sup 2/ of internal surface area. Their estimates were: 500 Ci embedded per beamline in beam absorbers after 1000 DT Shots; and 100 Ci adsorbed per beamline after exposure to 1 torr of tritium for 1 hour. Both estimates were revised downward since the estimates assumed pure tritium operation, whereas the neutral beams used orders of magnitude more deuterium than tritium. Deuterium competes with tritium for implantation and adsorption sites, reducing the uptake of tritium relative to the use of pure tritium. Data from neutron detectors indicated that the quantity of tritium implanted is equivalent to the revised estimate. The amount of adsorbed tritium exceeds the prediction. While tritium operation was highly successful, there were problems with the failure of several ion sources and gas injectors. Ion source failures were not due to the use of tritium as the working gas. However, their removal yielded information regarding tritium retention. Ten tritium injectors failed during the three and a half years of tritium experiments; six were replaced. At the conclusion of TFTR operation, all working tritium injectors had throughput leaks. By comparison, the deuterium injectors, which used the same valves, had only two minor fill valve leaks, and no repairs were necessary. Tritium contaminated component removal required purging until the concentration of tritium in any released air was <20 /spl mu/ Ci/m/sup 3/. For ion source removal this necessitated 50 to 100 purges and the release of several Ci. Upon removal, source surfaces had to be further decontaminated, from several million dpm/100 cm/sup 2/ to levels at which repairs could be effected.