LDX: The Levitated Dipole Experiment
Updated Aug. 15, 2004 : A summary of the experimental plan and a description of the experimental apparatus is given in Garnier's 2003 APS poster.
The primary objective of this experimental program is to investigate the possibility of steady-state, high beta dipole confinement with near classical energy confinement. The experimental program will also provide important new research opportunities:
The rest of this document will give a description of the experimental plans after the following outline:
The design of the experiment was based on the careful consideration of the key physics issues, various techniques to create and heat high-beta plasmas, engineering, and costs. Our overriding design objective was to conceive of an experimental device having the lowest possible cost while at the same time having high confidence for the operational capabilities required for our proposed scientific studies. With this objective in mind, the following major design choices where made:
The figure above presents a cross-section view of the proposed LDX device (as of 5/1/97, some changes have been made in the design since then.) The basic machine consists of a large vacuum vessel, a superconducting ring, two superconducting charging coils, a water-cooled copper levitation coil, copper divertor shaping coils (consisting of a low current Helmholtz pair), and ring positioning control coils.
Cross section of the LDX superconducting ring showing both the primary solenoids (1.36 MA) and the reversed winding (-0.118 MA) used to shape the inner most flux surface to the ring cyrostat.
The levitated dipole ring is a superconducting magnet comprised of two solenoids connected in series, carrying 3200A in a persistent mode. The design of this conductor, coil and its cryostat will be based on many of the advances in superconducting magnet technology made over the past 25 years and now widely used in large numbers of commercial MRI, NMR and other high field magnet systems in reliable, long term operating service worldwide. These technologies include:
Optimum combination of these technologies allows for long levitation times. The design of the conductor is a 19 strand Rutherford cable, cable-in-channel conductor. The strand itself is an extremely high performance strand developed and used in the D20 dipole, a very high-performance particle physics magnet developed recently at the Lawrence Berkeley Laboratory. The conductor is layer-wound with two joints between the coils. The cable uses an internal tin Nb3Sn strand with a diameter of 0.6mm and a copper fraction of 0.45. The low copper fraction is desirable in this application, because it maximizes the current-sharing temperature at a given current, and therefore the experimental run time. It also minimizes levitated coil mass and the associated error field from the currents in the external levitation coils. In order to obtain the highest possible plasma pressure compression ratio a two solenoid design has been chosen. The coil includes an inboard coil with "positive current" and an outboard "negative current" coil to shape the flux surface to conform to the circular cross-section of the levitated cryostat.
The cryostat consists of a toroidal shaped titanium magnet/helium vessel surrounded by a 304L stainless steel vacuum vessel. Run time is also maximized by an optimized combination of helium ullage space in the inner helium pressure vessel and a lead shield at intermediate cryogenic temperature, similar in concept to the shield used in the Princeton FM-1 experiment . Other aspects of the novel cyrostat design cannot be devulged at this time do to patent concerns.
An innovative "floating sensor" telemetry system, based on the internal sensor/extraction systems developed for ITER, TPX, and KSTAR would monitor magnet performance. On-board voltage, temperature, and pressure sensor signals would be fed through a sealed feedthrough to a low-power surface telemetry system, based on ~ 10mW of battery or broadcast input power.
The superconducting ring will be supported and serviced by a mechanical support system located inside the vacuum vessel at the lower flange. It will consist of a mechanical system for holding the ring in the lowered position and orienting it for proper alignment of helium transfer lines and charging instrumentation leads. The ring will be cooled/recooled and energized/de-energized in the lowered position inside the coil charging well. Coil charging is accomplished by energizing the external superconducting charging coils, then cooling the floating coil to superconducting, and then de-energizing the charging coil. Once the ring is cooled and energized, the helium lines and leads will be retracted from the magnet cryostat. The ring will then be mechanically lifted in two stages to the vertical levitation position at the center of the vacuum vessel. The external levitation coil will then be energized to support the weight of the ring. The mechanical support will then be disengaged and retracted. All these actions will be accomplished while maintaining a high vacuum. When the ring needs to be lowered, the operations are reversed. The mechanical system extends to just under the floating ring. The external levitation coil is slowly de-energized until all the weight of the ring is carried by the support. When all external magnetic fields have been switched off, the ring within its support is retracted and returned to the lowered and locked position. Magnet servicing can again be performed. It is expected that during normal operation, the coil will remain levitated for several hours. In the morning, the ring will be energized and lifted. At the end of the day's experiments, the ring will be again lowered and de-energized. The ring will remain connected overnight to the external cryogenic system to reduce the cooling time needed for the next days operation.
In the case of a coil quench, levitation coil power supply failure, or other loss of coil position control, a coil crash system is being developed. In the current design, involving a net of small SS cables, will be rapidly deployed to limit the coil acceleration to < 5g.
In general a coil that is magnetically levitated has six degrees of freedom: vertical motion, horizontal translation, tilt and precession about its axis. In the dipole confinement approach there are no toroidal or vertical fields and the superconducting ring must only be levitated against its own weight (in our case, equal to 450 kg). In LDX, we have located the levitation coil 1.6 m above the superconducting ring, and, thereby, we have selected for our base-case configuration a simple levitation scheme which is stable to translation and tilt. Only a single low power feedback system is required to stabilize the vertical mode. This approach works for limited plasmas and for ring null diverted configurations. We have solved the linearized equations of motion of a rigid ring current to obtain the characteristic frequencies and growth rates of the levitated ring . Characteristic frequencies and growth rates are typically < 10 Hz.
At this time, we expect to monitor the position of the ring optically, by reflecting a scanning laser beam from a toroidal machined scratch in the cryostat wall. Six such measurement will allow for 1 redundant measurement, in case of detector failure. A seventh scanner, observing a series of stripes etched on the surface will diagnose the final degree of freedom (spin).
In order to investigate the coupling between the edge plasma and the hot plasma core within a dipole-confined plasma, low-current copper shaping coils are attached to the outside of the vacuum vessel to shape the outer flux surfaces. By applying magnetic fields parallel to the ring's dipole moment, a ring null divertor configuration can be produced. When the applied field is anti-parallel to the ring's dipole, two point null divertors can be created above and below the ring along the axis of symmetry. LDX has been designed to investigate the stability and confinement properties of both divertor types. By energizing the Helmholtz coils from -12kA to +60kA, the position of the ring null divertor can be positioned from the outer vacuum chamber wall to a nearby radius of 1m without significantly changing ring stability. This allows systematic investigation of the role of flux compression in plasma stability and confinement. The point null divertor requires installation of two additional shaping coils and providing gravitational support from below the superconducting ring. In this case, the ring is unstable to tilt and horizontal displacements. Tilt stabilization will require four low-power windings of approximate circular cross-section, and horizontal stability will require two larger quadrupole windings. Both windings will be made from appropriately sized insulated cable and wound on the outside of the vacuum vessel. Since this divertor option requires feedback control of four dynamical variables (instead of one), we plan to investigate point null divertor configurations near the end of our five-year research plan.
The vacuum vessel is a 5 m diameter by 3 m high non-magnetic, 304L stainless steel vessel. Two 1.5m diameter center ports, one on top and one on the bottom, are provided for stage mounting and access. The vessel also incorporates an array of six 10-inch diameter radial diagnostic ports at its center axis and a set of four 10-inch diameter ports on the top head. A separate 10-inch port is located on the bottom head for mounting of vacuum pump components. The vessel is supported under the bottom head by three 304L stainless steel legs. The vacuum pumping system consists of a gate valve for isolation, an 8-inch turbo pump and a 60 CFM roughing pump for initial pump down and backing of the turbo pump. Vacuum readout and controller are also included. A 12-inch cryo pump will be used to maintain a base-pressure without plasma to 10-8Torr. All vacuum flanges are high-temperature copper gaskets, and the machine will be well insulated to allow bake-out to 200°C. Glow discharge cleaning will be performed in the usual manner.
We have selected resonant microwave heating of electrons as the plasma heating and formation technique best suited for the production of high beta plasmas in LDX. Four microwave power sources are available to the LDX experiment.These include a 28 GHz Varian gyrotron (200 kW, 100 ms and 5 kW CW) originally used for the Tara tandem mirror experiment, an 18 GHZ Varian klystron (6 kW CW) also used for Tara, a 9.3 GHz klystron (4 kW CW) originally used in the Constance mirror experiment, and a 6 GHz Varian klystron (3.4 kW CW) originally used at Columbia University for plasma processing. We have used these microwave sources successfully in previous plasma experiments, and we expect little difficulty adapting them to LDX. As in the Constance mirror and CTX dipole experiments, the 6, 9.3, and 18 GHz microwave power will be coupled into the vacuum chamber through quartz windows and using single mode waveguide. For these frequencies, we intend to rely on "cavity heating" to maintain uniformity during microwave heating. This will limit plasma density approximately one-half of the cut-off density for each frequency. Since the stationary density profile scales as r-4 while the cutoff density scales as r-6 along the dipole midplane, microwave accessibility to the fundamental and first harmonic electron cyclotron resonance will probably determine the maximum density compatible with hot electron formation. For the 28 GHz gryrotron, we intend to use the same Vlasov antenna used in Tara which couples the TE02 mode to a gaussian beam and direct this beam to the high-field regions within the ring.
Large and rapid plasma density increases will result from both fast gas and Li pellet injections. The Li pellet injector will be be a reproduction of a simplified version of the Alcator C-Mod Li pellet injector . Lithium pellets are fabricated in a "cookie" cutter fashion from Li metal contained in ribbon form. The major elements of the pellet injector are a high pressure gas gun used to accelerate the pellets, a guide tube used to aim the pellet and differentially pumped vacuum systems used to isolate the plasma from the propellant gas. The high speed pressure valves follow a solenoid design developed at ORNL. They are capable of operation from 300-1000 psi and are driven by a low-energy transistor controlled capacitive discharge. A typical pulse length for reliable injections ~ 1 ms. The velocity of the pellet is related to the sound speed of the propellent gas and so the highest speeds are attained with the use of hydrogen gas at a typical pressure of 650 psi. The fast gas values will also be used for direct gas injection in order to reduce the total plasma density rise.
The diagnostics for LDX have been selected to provide complete information on global plasma equilibrium, plasma density, hot electron energy, plasma fluctuations and instabilities, neutral particle sources, and edge plasma characteristics. These diagnostics are primarily low-cost and take advantage of the outstanding diagnostic access provided by the dipole geometry and of the ability to reconstruct plasma equilibrium from magnetic data in a manner analogous to equilibrium reconstruction in tokamaks. Basic diagnostics installed during the third program year will include magnetic flux loops, multi-element Hall effect probes, XUV diode arrays, hard x-ray detector, photodiodes with deuterium line filters, and a 90 GHz microwave interferometer. Key measurements are the hot electron energy distribution (NaI pulse-height analysis), hot electron profile (XUV diode arrays), plasma density, and magnetic measurements of the hot-electron diamagnetic currents. After the first year of plasma operation, additional diagnostics will be installed in order to measure the effectiveness of hot electron energy thermalization. These will include additional hard x-ray detectors, a charge-exchange energy analyzer, and perhaps additional interferometer cords appropriate to the high-density plasmas created with fast gas puffing or Li pellet injection. (see also our collaboration page.)
The LDX experiment will be located in the south end of the west cell in MIT PSFC building NW21. The experimental area will comprise approximately 1/3 of the experimental hall previously used for the TARA experiment. Besides the usual electrical, water and HVAC services already provided within the building, the LDX experiment will take advantage of a significant additional experimental facilities already in place, including: (1) TARA power supplies comprised of 10 converters (each rated 800 V, 4 kA), (2) 10 ton overhead crane, and (3) 10,000 gallon LN2 storage tank and vacuum insulated transfer line. The Alcator project, located in the same building, is supported by dedicated staff, technicians and shops providing machining, electronics, welding, and vacuum support services. These resources are also made available to other projects within the PSFC on a charge per use basis, with adequate advance scheduling, when it does not interfere with operation or maintenance of the Alcator C-Mod. Adjacent to the LDX experimental area within the same experimental hall is the Pulse Test Facility (PTF) which is used on a limited basis for ITER superconductor cable and joints testing. The LDX project will take advantage of the PTF data acquisition and control system computers and software for operation of the LDX facility and superconducting dipole magnet. The LDX project will provide its own CAMAC, signal conditioning and PLC hardware while utilizing the PTF hardware and software in the control room.
The proposed LDX experiments will yield new data on high-beta magnetic plasma confinement in a dipole magnetic field. It will provide the first systematic investigation of the use of MHD compressibility for plasma stability. In addition, LDX will provide the basic understanding of energetic particle confinement and stability in a dipole magnetic field, of the relation between edge plasma and a hot plasma core, and of the possible elimination of drift-wave turbulence to produce plasmas with classical confinement. By the end of fifth program year, we will have had two years of study of high beta dipole plasmas and be able to report on the feasibility of the dipole confinement concept as a potential route to an innovative fusion power source. The LDX schedule reflects the need for experimentation with both long-pulse, energetic electron plasmas produced by ECRH and for relatively transient plasmas produced by fast gas and Li pellet injection. The program allows 33 months for the design, procurement, assembly, and evaluation required to complete hardware systems. Our plan targets initiation of plasma studies during the last quarter of the third fiscal year. Based on our previous experience with microwave heating of plasmas confined by magnetic mirrors and mechanically-supported dipoles, we expect little difficulty with the production of relativistic electrons shortly after the start of plasma operation. During the fourth project year, our primary research activity will be the study and optimization of hot electron beta. At this time, we will also begin fast deuterium gas puffs and Li pellet injection order to observe density build-up and transient thermal plasma confinement. We have summarized below the major activities involved with the LDX experiment. Specific dates are based upon a chronology which begins with project funding at the start of FY '98. The five activities are presented in chronological order beginning with machine design and ending with dipole concept optimization.
The first major activity of the LDX project is design finalization of the superconducting ring and associated cryogenic and handling systems. Within the first three months of the project, the conceptual design of the superconducting ring will be completed and the most critical project pacing item, the superconducting wire, will be ordered. After six months, we will prepare a detailed design of the LDX superconducting ring, ring control, ring charging, and cryogenic systems, and we will convene a detailed design review from experienced engineers and plasma scientists. After the design review, the superconducting coil, levitation and control coils, and the vacuum vessel will be fabricated. The ring cyrostat, charging, and liquid helium transfer systems will be fabricated during the second program year.
The machine assembly will begin after completion of site preparation and during the summer of 1999. Site preparation occurs in parallel with the final design and procurement of the vacuum assembly and levitation magnetics. During machine assembly, we initially emphasize only those components required for the initial tests of the superconducting ring. These include the vacuum chamber, ring handling systems, retractable ring support and recovery system, levitation coils, coil position diagnostics and axial feedback system. The first major machine milestone is set for January, 2000 (27 months after the start of LDX funding) when the initial testing of superconducting ring levitation begins. At this time, ring motion will be limited with safety straps but the ring will otherwise be maneuvered and levitated as required for base-case operation with the single upper levitation coil and the low-frequency (10 Hz) axial position control system.
The initial phase of plasma experiments will establish a high beta hot electron plasma using multi frequency ECRF heating. This research extends the results already obtained using electron cyclotron resonance heating to create energetic electrons at the Columbia CTX device . In contrast with CTX, LDX will have no losses from a mechanical ring support structure nor plasma loss from pitch angle scattering. We expect particles to be very well confined in LDX, and an important issue to be investigated is will be density (and neutral gas) control. Since the loss of plasma in a levitated dipole is expected to be small, we anticipate significant reductions in the deuterium gas feed and operating pressures as compared with the CTX experience. This should also lead to improved hot electron production and increased plasma pressure. After formation of energetic electrons, we will focus on b enhancement. As described previously, hot electron interchange instabilities are expected if the hot electron pressure gradient becomes excessive. We will utilize multi-frequency heating to broaden the radial profile of the energetic electrons and to increase beta. Experiments in STM indicated that a substantial increase of stored energy was obtained in a hot electron plasmas when multiple frequencies were applied probably due to the elimination of superadiabatic effects which can create phase space barriers during single-frequency heating. Further experiments will:
After the formation of a high beta, energetic electron plasma, fast deuterium gas puffs and Li pellets will be injected into LDX in order (1) to raise plasma density by more than a factor of ten to n > 1019m-3, and (2) to transfer and thermalize hot electrons energy. As the density increases, the efficiency of thermalization will be detected by changes in b measured magnetically and the ion energy spectrum measured with the charge-exchange analyzer. These experiments will focus on the measurement of global energy and particle confinement and the observation of plasma instability or fluctuations.
The final research phase of the LDX program is devoted to the investigation of techniques to increase plasma pressure and energy confinement. This will include systematic investigation of the role of compressibility and the influence of the topology of the scrape-off layer on global performance. At this time, we will report our evaluation of the feasibility of dipole confinement as a possible fusion power source.
The LDX project is a joint venture among two well established groups within the fusion community and it will be a part of the Physics Research Division at the PSFC and Department of Applied Physics at Columbia University. Columbia University will provide scientific leadership in experimental operations and ECRH. The MIT Plasma Science and Fusion Center will provide engineering and technical expertise and lead the development of dipole confinement physics and the integration of theory with experimental results. We are actively seeking the participation of additional outside collaborative groups and individual researchers for both experimental and theory support programs.
The LDX project will be jointly directed by Drs. M. Mauel and J. Kesner. Project activities will be further organized into three key areas.
Mike Mauel <firstname.lastname@example.org>
Jay Kesner <email@example.com>
Darren Garnier <firstname.lastname@example.org>
Joe Minervini <email@example.com>
|Webmaster: D. Garnier||
Last updated: Mon, Aug 16, 2004