ARIES Project Meeting Minutes
Princeton Plasma Physics Laboratory, Princeton, NJ
Documented by L. Waganer
Welcome - Rob Goldston welcomed the ARIES Team to PPPL. He feels the ARIES contributions help the fusion community visualize the future capabilities and developmental needs of promising fusion approaches. These quantified visions enable DOE to prepare a more directed and efficient program development plan. He pointed out several laser program efforts that are enabling activities. Rob emphasized the compact stellarator as a very promising concept being investigated for a proof-of-principle experiment. He further explained several of the key advantages and recent advances of the compact stellarator concept.
Status of DOE Systems Studies - Al Opdenaker discussed the FY03 DOE fusion budget for various programs and the System Studies program. The Financial Plan is assuming the smaller of the Congressional budgets. NCSX is planned for ~$11M to start design effort. The ITER decision will soon be made with inputs/recommendations from FESAC and NRC. To determine the cost impact to the US, PPPL will lead an ITER cost assessment of tasks the US may select.
Status of ARIES Program - Farrokh Najmabadi compared the ARIES budget for FY02 and FY03, noting a small decline in the total budget. He had planned on scaling down the IFE budget while initiating the new compact stellarator (CS) study at a significant level. However, it was recommended the IFE effort remain at a similar level, hence the initial investigation of the CS will be reduced. Distribution of effort within the ARIES team was provided
Farrokh detailed the IFE research plans for FY03 that would investigate design windows for the thin and thick wall protection schemes, materials and design of the hohlraum targets, and HI driver interface studies. The results from the past two years on the IFE studies will be documented in a special issue of Fusion Science and Technology with the deadline for paper submission is December 10, 2002. Five papers were identified for submission with primary authors.
Farrokh outlined the Compact Stellarator (CS) research plans in the ARIES program that will commence in FY03 and run for three years. This CS concept is a good candidate for investigation for a commercial power plant based on its improved feasibility considerations. It was noted that the NCSX and QPS experiments are being initiated in the US along with favorable results from LHD in Japan and W7X in Germany. Continuing advances in the simulation modeling and multi-dimensional optimization capabilities may yield an attractive CS power plant design. The three-year ARIES effort would investigate the development of the commercial plant optimization tools, explore the configuration design space, and develop a detailed and optimized system design.
Next Meeting/Conference Call - The next project meeting will be held at UCSD, tentatively on January 8, 9, 10 or January 13, 14, 15 for 2 ½ days. L. Waganer will poll the team for conflicts on those dates. [Subsequently, several members had conflicts on the Jan 13 -15 period, thus the next meeting will take place on January 8-10.] The meeting will follow the same format as the October 02 PPPL meeting by devoting one and a half days to IFE and one day to the new CS effort.
The next project conference calls dates were selected to be Tuesday 5 November and Tuesday 10 December. Les Waganer will notify the team of the conference call numbers.
ARIES-IFE Study Presentations
Systems Analysis and Integration
Status of Heavy Ion Driver Point Design Parameters - Wayne Meier informed the group of the continuing definition of the self-consistent HI power plant design, primarily the HI driver parameters. Since the July ARIES meeting, two changes have been implemented.
Liquid Wall Thickness, Life, Pumping Power, and COE Tradeoffs - Wayne Meier reiterated the HYLIFE-II design goals of protecting solid structures for full power plant life (30 FPY), low activation with ordinary steels (304 SS), and reduced material development needs. Wayne discussed his analysis of the action item question to reduce the thickness of the liquid metal protecting wall. He defined a detailed model of the flow rate and pumping power as a function of the protective wall thickness. He also defined a base case set of parameters and other assumptions on power and cost increments. At an effective thickness of 0.56 m (1.12 m at 50% density), the structural wall behind the liquid protective wall has a 30-year life. The costs of the pumps are proportional to mass flow rate. Parametric curves are given on wall life as a function of liquid thickness for different dpa assumptions. From these data, the pumping power was obtained as a function of desired lifetime. With the maintenance assumptions, the capacity factor and COE are determined as a function of FWS lifetime. Wayne concluded the liquid protection thickness sufficient for life of plant structure yielded the lowest cost approach. Wall lives greater than 10 years only decreased the COE by 4%. The 25-dpa damage limit increases the COE by ~ 7.5 to 19 percent.
Chamber and Final Optics Nuclear Analysis
Ferritic Steel Lifetime Assessment and Self-Consistent Nuclear Parameters for ARIES-IFE - Laila El-Guebaly stated the objectives to assess the nuclear performance of the HYLIFE thick liquid blanket design concept: evaluate the candidate breeders (FLiBe and FLiNaBe) and determine the lifetime of ODS (oxide dispersion strengthened) ferritic steels, waste disposal rating, and helium production. She also estimated reduction in the waste of the thick liquid wall concept.
Laila showed the key parameters for the liquid wall concept, the radial build for the liquid wall and liquid-cooled shield, and coolant parameters along with the related ARIES requirements and design limits. She noted FLiBe is a better breeder than FLiNaBe, thus wall thicknesses of 83 cm is required for FLiBe and 150 cm for FLiNaBe for a breeding ratio of 1.08. FLiBe is a better shield, so FLiNaBe blankets will be slightly thicker for the same shielding requirement at the innermost FS or SS shield structure. Laila said that the FS reweldability limit (1 He appm) is greatly exceeded for both FLiBe and FLiNaBe blankets, thus the innermost shield or nozzles cannot be rewelded at any time during the plant operation. The nuclear source terms as a function of radial depth for both materials was provided for aerosol production.
All the current FS and SS alloys generate high-level waste at the inner shield surface and the flow nozzles. For the outer shield structure, only the ODS-MF82H-FS will meet Class C limits with FLiNaBe shield thicknesses greater than 45 cm. Potential solutions to meet waste disposal requirement (WDR < or = 1) for the outer shield wall when using other material combinations are: 1) thicken blanket and/or 2) control Mo and Nb content in steels. Theoretically, tailoring the FS materials to reduce the Nb and Mo concentrations will help reduce the WDR. Practically and economically, complete removal of these materials can never be accomplished. The only WDR solution for the inner wall and nozzles is to thicken the blanket. The thick liquid wall concept offers a small reduction in waste as solid blankets represents only 2-4% of the life-cycle waste.
Recycling of Target Materials versus One-Shot Use Scenario - Laila El-Guebaly reviewed the HIB ID target parameters and the selection criteria for the hohlraum wall materials. Even though the HIB ID hohlraum materials are thought to be a problem, they represent less than 1% of the life cycle waste stream. Recycling is not a requirement unless the materials have a significant cost or resource limitation. Recycling adds cost and complexity by remote handling, radioactive storage, and special handling/cooling of materials.
In keeping with the ARIES requirements of a WDR of < or = 1, recycling dose of < or = 3000 Sv/h, and an accident dose at the site boundary of < or = 1 rem, Laila examined the cooling period and how it controls the WDR and recycling dose. All materials qualify as Class A (or Class C) LLW after one shot (no recycling). If recycled with no cooling period, all hohlraum materials generate HLW except for a few materials (W, Ta, and Xe). The combination of gold/gadolinium generated the highest level of HLW. When cooled for < 18 days, all materials except for Au/Gd and solid Kr would meet waste and dose requirements.
It seems feasible to try to reach a point to "clear" the materials (release to commercial market) for the one shot use materials, as the cooling period required would range from 25 to 225 y for Au, Hg, and Ta. Other materials may never reach a clearance index of 1 before 300 y. At present, there is no US market for cleared materials.
Laila presented data on the economic impact of recycling the close-coupled target materials. The results indicated the costs of recycling and material costs may exceed the benefits of using Au/Gd as target materials. Comparing the one-shot approach versus recycling target materials indicated a preference for the one-shot approach and low-cost materials, offering attractive safety features, radiation-free hohlraum fabrication facility, less complex design, and lowest COE.
Final Focus Magnet Shielding Update -Wayne Meier presented Jeff Latkowski's results on magnet shielding design parametrics. Wayne summarized the general requirements of shielding the final focus magnets with a half angle of 24° beam array: magnets should not quench on a single-shot basis, have a reasonable recirculating power, achieve a suitable radiation damage lifetime, and generate no Class C waste. The final focusing magnet radiation lifetime is limited by the total dose to the S/C insulators of 100 MGy, which is dominated by the gamma-rays and a conservative limit for the fast neutron flux of 1019 n/cm2. Use of high temperature superconductors (HTS) may help ease these limits.
Point design results for the final S/C magnets are shown below.
In response to an action item to define the target chamber response to an off center target ignition, an analysis was conducted with the target off-center by 0.5 cm. The average insulator is insensitive to final target position in this range. There was slight change in the magnet superconductor lifetimes, but it is probably statistical noise. There was no change in the WDR.
An analysis was conducted in response to the effect of eliminating the crossed jets. It was assumed the wetted wall protects the structural materials from x-ray and ion damage. Removing the liquid crossed jets increases the neutron damage to the port plug by a factor of 5. This increase is also likely to increase the magnet dose rate, which can be compensated by thickening the port plug.
Liquid Chamber Analyses Results
Vapor Conditions Including Ionization State in Thick Liquid Walls - Don Haynes stated that the approximate 5 MJ of prompt x-ray energy (out of the 120 MJ released from the target) would vaporize 8.4 kg of FLiBe in a microsecond, which amounts to the 1 cm of surface thickness. This cloud of FLiBe vapor shields the later arriving debris ions and prevents additional evaporation of the surface. The BUCKY code analyzed the wall in both of its heat capacity and plasma equation of state modes. As the vapor cloud shock moves toward the chamber center, it begins to interact with the expanding target shock, beginning around 10 microseconds. After 20 µs, the 1-D BUCKY code begins to lose fidelity. As the wall blowoff moves inward, it cools by expansion, but remains ionized as aerosolization begins.
The narrow gaseous shell of the blowoff plasma shields the remainder of the liquid wall from the later target emissions. As these emissions are absorbed in the gaseous shell, the shell re-radiates to the rest of the wall on a longer time scale.
Update on IFE Aerosol Analysis - Rene Raffray presented Phil Sharpe's presentation on aerosol conditions. Phil indicated he was expecting some results, but he did not quite finish the analysis. The 1-D radiative gas dynamics model has not been improved since the last presentation, but Phil intends to add ionized FLiBe species properties and ion heating from target debris. He is working with UCSD on the wall condensation model to add enhanced vaporization and phase explosion to liberate mass and evaporation from surface leading to heterogeneous particle formation. He is considering extending the aerosol model to multiple component materials to simulate FLiBe. He may need to add ion-induced nucleation. The wall thermal model now solves the energy and mass transportation between the wall and chamber in a self-consistent matter. He has added a volumetric heating source from x-ray attenuation to help drive wall response, but may need to add ion heating.
A detailed thermal wall response model has been added and soon he will run both liquid lead and FLiBe cases with the coolant data supplied by UCSD.
HEIGHTS Integrated Models for Thick Liquid Walls in IFE Systems - Ahmed Hassanein described his modeling of liquid jets response to neutron deposition, fragmentation, and aerosol formation. In modeling of the liquid oscillations, equations of mass, momentum, and energy conservation were solved. The equations of state for the three components were also solved assuming the internal energy consists of the thermal motion of the nuclei, thermal excitation of the electrons, and their elastic energy.
He investigated the formation of cavities within the liquid from a tensile relief wave. There is a critical radius, r*, below which the cavities disappeared and above which the cavities grew freely. He also defined the rate at which these cavities formed and the pressures where liquids can be fractured due to the growth of cavities. The fracture pressure for FLiBe is 8.7 atm. Impurities reduce the actual fracture pressure when compared to theoretical values.
Neutron beam volumetric deposition of > 100 j/gm in the liquid jets results in a thermal pressure rise of ~ 2-3 atm that creates a negative pressure of the same magnitude in the center of the jet. Liquid metals subjected to negative pressure are metastable, with possible spontaneous formation of cavities. After neutron deposition and during an initial period of a few µs, the liquid negative tension reaches values near the fracture pressure.
A relaxation shock wave is then created due to the formation of the cavities in the stretched media. The vapor pressure inside the cavity is ~ 0, but the liquid outside the cavity is at a negative pressure. Such a jump in pressure creates a shock wave (relaxation wave) as the media recovers to normal density. In usual cavitation theory, vapor bubbles grow with time until they come into a region of higher pressure where they collapse and are accompanied by acoustic noise. Formation by negative pressure is quite different. Vapor bubbles grow due to heat conduction from surrounding overheated liquid. The negative pressure bubble grows due to the unloading of the stretched liquid. The shock wave brings the media back into a normal state with normal density. Therefore, any arising negative cavity will not disappear.
Liquid jets will oscillate after instant energy deposition. Because the frequency of radial oscillations is much higher than axial oscillations, the cavity dynamics are governed by the radial oscillations. The number of cavities arises mostly near the jet axis where the magnitude of the negative pressure is maximum. Cavities expand freely initiating a shock wave of relaxation ahead of the cavity surface, expanding in axial and radial directions. On a longer time scale, these cavities will join to form larger cavities with a longer axial scale length.
Oxygen Control in a FLiBe Blanket - Dai-Kai Sze said that oxygen would be present in the target debris and it is also a transmutation product from FLiBe. The oxygen will combine with the FLiBe coolant to form BeO. The BeO will saturate the coolant in approximately 25 days. If not combined, oxygen is very corrosive to the structural materials. BeO will form particulate that can easily be filtered out. O2, as a gas, can be removed in a side stream. Other materials such as carbon must also be considered in the contamination and corrosion control process.
Discussion of Structural and Nozzle Material Selection - Dai-Kai Sze presented Mike Billone's discussion on materials selection. The HYLIFE coolant conditions are 630°C and 150 Pa. If 304 SS is used as a structural material, there will be radiation damage above 550°C with high amounts of swelling and creep. Tritium inventory would be a safety problem. The primary stress can be lowered to reasonable levels by imposing design limits. Ferritic steel can be used up to 600°C, but the tritium inventory is higher by a factor of 5. Vanadium would have an even higher tritium inventory. The advanced ferritic steel can go to 630°C, but tritium inventory is still a problem. Dai-Kai Sze thought FLiNaBe is a better choice than FLiBe as a coolant.
The selection of the HYLIFE structural material was not decided at this time. Instead a conference call would be convened next week to discuss the material issues and recommend a structural material for the thick liquid wall protected chamber concept.
Coolant Cleanup and Process Requirements - Laila El-Guebaly presented the Igor Sviatoslavsky and El-Guebaly results on coolant cleanup and process requirements as based on the prior Ralph Moir work on HYLIFE-II on FLiBe cleanup. Coolant cleanup encompasses removal of target debris, corrosion products, atmospheric contaminates, and tritium. Cleanup can be accomplished by filtration, distillation/evaporation, centrifugal separation, and reductive extraction (in the order of decreasing ease and attractiveness). Each of these processes were described and evaluated as to their capabilities for cleanup of the coolant. A combination of the cleanup processes will probably be recommended. The likely cost for the cleanup system would be around $100M.
Experimental Results for Thin and Thick Liquid Walls - Minami Yoda described the experimental objectives for both the thin and thick wall protection approaches for the IFE chambers.
The experiments on thin, high-speed, forced film or tangentially injected, liquid wall protection schemes are aimed at defining design windows, mainly for the downward facing surfaces, which are the most severe application. Minami showed the test apparatus to investigate circular ports of varying heights and diameters in the flow stream. Other fluid parameters considered were initial film thickness, flow velocity, and Reynolds number. In the cases where the port cylinder height exceeded the flow height (H > d ), there were two distinct detachment locations: a) at port leading edge and b) at the trailing inner port surface. In both cases, significant fluid flow went over the port and the film detached from the first wall even though the height of the port was above the flow surface. For cases where the port height was less than the flow thickness (H < d), more fluid flow went over and blocked the port, but wit and the film detached from the first wall h much less disturbance. From these experiments, cylindrical openings are not compatible with high-speed thin film protection. Streamline port fairings will be investigated in the future to see if other geometries are compatible with forced films.
Surface wettability was investigated to determine its effect on detachment of an inverted parallel flow on a glass surface. At two flow rates and Reynolds numbers, a non-wetting surface (treated with Rain-X®) detached from the surface at a shorter distance from the nozzle. This indicates a wetted surface is desirable for maximum film attachment.
The thick liquid protective approach can use either thick oscillating liquid slabs to form a target pocket or stationary jets arranged in an orthogonal array to allow beam and target penetration while providing protection to the chamber structure. One of the key parameters is the expected ripple on the liquid sheet or tube to determine the closest positioning of the driver beams without adversely effecting the beam propagation. The other question is how robust is the liquid sheet geometry to flow disturbances or blockages. The GT researchers are trying to study and quantify the impact of nozzle design and blockage on surface ripple on the liquid sheets. Variations of nozzle geometries were tested to determine the probability of liquid (Liquid Probability Distribution, LPD) at any spatial location and the mean surface ripple as sz (average standard deviation of z position of the free surface). The LPD data was generated photographically, by averaging results over 100 images, to obtain the probability density functions. The ripple was roughly similar in magnitude on all three nozzle types, with the rounded corners being the worst and the 5th order polynomial as being the best. The rounded corner nozzle does not reduce surface ripple. With the best nozzle configuration, side and edge ripple increased with Reynolds number. In the scaled experiments, the average sz is 0.10 mm at x = 10 cm and 0.19 mm (or 1.9% of the jet thickness at the nozzle) at x = 25 cm. For the larger prototypical slab jets planned for HYLIFE-II, the maximum surface ripple of 1.4 mm is predicted, corresponding to 3 sz, assuming dynamic similarity and the same sz of 1.9% of the jet thickness at the nozzle.
A flow straightener was constructed of a perforated plate of holes with an open area of 50%, a 25-mm section of honeycomb cells, and a fine mesh screen. The flow area through the straightener is 10 cm by 3 cm. The overall length of the straightener component was 62 mm and it is located 195 mm from the last screen to the nozzle exit. To determine the consequences of flow blockage (due, for example, to debris in the coolant), blockages consisting of a rectangle 1.5 cm by 0.5 cm (2.5% of total flow area) were placed on the final fine screen. The blockage rectangles were placed either in the center of the screen or adjacent to the short edge of the screen. The blockage drastically increased ripple on the flow, with the edge blockage altering the shape of the flow more than the center blockage.
Liquid Wall Ablation - René Raffray presented his results of investigating the ablation of Pb and FLiBe liquid wall when exposed to repeated IFE target emissions and debris. Properties of these two coolants were shown. The ABALATOR code analyzed these two fluids with the x-ray spectra from the target. X-ray spectra and cold opacities were used in the aerosol source term estimate. The very rapid heating rate of the liquid results in phase explosions or explosive boiling from superheating the metastable liquid state. Photon energy deposition density profiles were shown for both Pb and FLiBe in the film and explosive boiling region.
René discussed the physical processes leading to an aerosol formation following the high-energy deposition on a short time scale. The ABLATOR computer code was used to model the energy deposition from x-ray source. Comparison of the results from ABLATOR code and the simple volumetric model showed very similar results for both lead and FLiBe. The similar results suggest the simpler model would suffice in most instances. Some uncertainty remains about the ablation source term.
FLiBe Properties - René Raffray described the FLiBe thermodynamic properties currently being used over the expected range of temperatures. He also showed the ionization equilibrium and validity of the local thermal equilibrium (LTE) conditions. Also included were the thermodynamic functions for pressure, internal energy, enthalpy, specific heat, adiabatic exponent, and sound speed. These data were compared to the results from Cheng, et. al. The pressure and internal energy results were significantly different between the two sources.
Chamber/Beam Physics, Beam Transport, and Chamber Clearing
Gas Transport and Control in Thick-Liquid Target Chambers and Heavy Ion Beam Lines: Past, Present, and Future - Christophe Debonnel describe the difficulty in shielding the beamlines from charged particles and debris from the target chamber. In the 80's, the Hyball-II concept identified many of beam tube issues, estimated the clearing times, and postulated some vacuum and debris control concepts. The TSUNAMI code was developed to estimate the gas dynamics behavior during the venting process by solving the Euler equations for compressible gas flows. This code evolved throughout the 90's to solve more challenging problems. In the early 2000's, the HYLIFE-II concept refined the liquid geometries to better protect the solid structures and proposed approaches to control the gas/debris contamination of the beam lines.
The current approach to controlling the debris entry in the beamlines include efficient design of chamber structures, minimize mass and energy flows at beam ports, venting of debris toward condensing areas, and analyzing the target chamber pressures to determine inlet boundary conditions for the beam tubes using the latest TSUNAMI code. The newest approach to minimize or eliminate target debris entering the beam tubes is to employ both a magnet divertor to direct the charge particles into a trap region. Also a liquid vortex using FLiNaBe on the inside of the beam tubes between the final focus magnet and the chamber wall will capture particles and gases. Simulations were shown of the magnetic divertor and the vortex tube.
3-D Simulations of Magnetic Shutters - David Rose expanded on the analysis of a magnetic divertor or shutter to capture and control the debris ionized particles. The baseline conditions are modeled as a plasma being blown from the target into the beam driver ports. David had previously conducted a 2-D simulation of this phenomenon, but now has upgraded it to a 3-D analysis. He had test cases, both with and without a magnetic field of 1 kG. The results indicated a 10-cm long magnetic chamber with a modest field strength could capture virtually all the charged particles. The few remaining particles may be an artifact of the code methodology. He proposed a modest tabletop experiment to verify the analytical results.
Beam Chamber Transport Requirements for FLiBe Vapor Pressure and Aerosol Conditions - Craig Olson refreshed the group on the variety of possible ion beam transport techniques across the power core chamber. Neutralized ballistic remains the baseline approach with the pinch modes as attractive backup approaches. Craig showed a matrix of transport schemes for the dry-wall, wetted-wall, and thick liquid wall with key parameters and concerns for each combination. He explained several physical quantities that scale with the line charge density. He also informed the group of a series of scaled experiments at LBNL that would address the science issues of HIF.
Aerosols may adversely affect neutralized ballistic transport (NBT) with thick liquid walls. The aerosol effects for BNT include loss of energy per distance ( JE/Jx), scattering, stripping, charged droplets, micro-breakdown, and plasma-like effects of the "charged aerosol". To scope the problem, Craig established a parametric fill fraction for droplets of different sizes, defined by the parameter naerosol · r3 . Then each of the NBT effects was examined with respect to the fill fraction, qualitative relationships defined, and limiting values for each case. The stripping case is the dominant effect at a fill fraction of naerosol · r3 = 1 x 10-9 (equivalent to an integrated line density of 1 mTorr).
In the preformed channel transport, stripping is also the dominant limit around 1-Torr chamber pressure. For the self-pinched transport, the self-pinch process is dominant around 100 mTorr.
Beam Interaction with Chamber Aerosols - David Rose explained that the purpose his analysis was to assess the effects of aerosol droplets on HIF beam propagation. The major areas of investigation were the time-dependence effect and the electrostatic charging of the droplets. Impact ionization, beam stripping and slowing down of co-moving electrons can remove or deposit additional electrons onto the droplet. He used the 2-D LSP simulation code to model the beam and droplet initial conditions and to simulate the beam/aerosol physics. He showed a time sequence from 67 ps to 375 ps of a 0.2-µm-droplet expansion from collisional interaction with the beam plasma ions. There is plasma expansion around the droplet driven by ambipolar diffusion form the 6 eV ionization electrons. Initially the stripping is large for a few beam ions, but later more beam ions are only weakly stripped. This same behavior is exhibited by larger 1- µm droplet. In the regimes of interest, simulations suggest that the effect of droplets can be calculated purely from their line-integrated density.
Status and Objectives of Pinched Mode Studies - Simon Yu highlighted the robust physics and technology achieved on the mainline ballistic neutralized transport (BNT) approach. He also summarized the advantages, accomplishments, and plans for both the assisted pinch and the self pinch transport approaches. Both have identifiable and beneficial advantages with demonstrated progress toward their respective goals. Both pinch approaches will continue to be developed in the coming year.
Assisted Pinch Mode for Heavy Ion Beam Transport - Stephen Neff explained the basic physics for assisted pinch transport in the power core chamber. There must be sufficient rarefaction of the plasma on axes to prevent breakdown to the walls, a plasma density less than a upper limit value to prevent energy loss in the channel, and a plasma density greater than a lower limit for j x B forces. Therefore the goal is to map out an acceptable plasma pressure window. Ranges for each of the conditions are identified. An experimental setup with appropriate diagnostics to validate the calculated ranges was shown.
To create an assisted channel, a laser beam of 5 J was used to heat the plasma to the required rarefaction on the axis. Then a prepulse beam of 40 J will increase the rarefaction and ionization of the chamber gas. Finally, the main discharge of 3 kJ will create the plasma channel. A graph confirmed good agreement between the simulation and the experimental measurements. The prepulse was observed to stabilize the main discharge. When channels were orthogonally crossed, the discharge beam followed the crossed laser paths.
Modeling of Assisted and Self-Pinch Transport - David Rose elaborated on his analysis of the assisted and self-pinch beam transport modes. He first summarized the advantages of each approach and the hardware required to implement the transport mode. He described how the assisted pinch transport (APT) would be configured to handle the hybrid target foot and main beam timing and energy demands. The foot and main beams were simulated and both proved to be well focused. The APT also has some self-fields that may enhance beam confinement. The simulations showed about an 85% of the beam energy (efficiency) were within a 5-mm diameter spot. An ideal case with no self-fields would yield a beam efficiency of 94%. Beam efficiency is best for a current around 75 kA and begins to degrade below 50 kA. The APT degrades for beam divergence less than 1.5 milli-radians. The APT is insensitive to the beam ion species.
David explained the principle of the self-pinch transport (SPT) scheme. Results from the SPT experiment in the NRL GAMBLE II device showed a better charge per beam energy within a 5?cm spot than predicted with the IPROP code. The code results indicated pinched equilibrium after one meter of beam propagation in the 50-mTorr-chamber pressure. At 65 mTorr, the beam develops an equilibrium current radius of 0.5 cm and contains ~5 kA of net current. Radial ion velocities were found to be consistent with channel hydro limits. Simulations with finite-mass beam ion (Pb+65) and radial temperature profiles suggest less evaporative losses.
Beam Steering Capability for Target Position Uncertainty - Simon Yu explained there are three types of beam steering capabilities with heavy ion drivers:
The steering capability requirements for target injection of 200 m/s include corrections of ± 2 mm and response times of < 25 ms. Simon showed a sample dipole layout for the final focus magnet with a beam envelope generally of 9-10 cm with peaks up to 20 cm that had a centroid position variation of 2-3 mm. He showed the setup and results of the first beam through the NTX magnetic transport conducted on August 2002. He presented the results of the first beam through the neutralized drift section without and with a plasma plug - the latter produced a smaller beam spot size. He concluded that the dipole corrector requirements are easy to meet as the requirement for pulsed dipoles was ~ 1 kG within 25 ms and it has been demonstrated that NTX quadrupoles can achieve ~7 kG in 1 ms.
Pulsed Normal Quadrupoles for a Heavy Ion Fusion Driver Final Focus Section - Derek Shuman presented the detailed design requirements for the last quadrupole magnet set. The quadrupole must also minimize the power loss, be highly radiation resistant, and not create high-level waste products. He showed the HIF final focus driver quadrupole field plots and the Von Mises stress plots. His suggested quadrupole design parameters were shown. The proposed design advantages and disadvantages were discussed.
The power losses are particularly important and were analyzed to determine the power loss scaling to determine the optimum cable diameter. This diameter is proportional to the square root of the pulse width. The total coil power loss is then a linear function of strand diameter only. The total power loss is proportional to the cube of the winding radius.
After determining the optimal cable diameter and strand diameters, the operating parameters were derived. Derek chose a single power supply per coil to reduce the voltage by 75%. The relative size and arrangement of the leads were shown. A single turn helical Litz-type coil is feasible. He showed how the cross-section of the coaxial cable would transition into the coil cross-section.
Target Fabrication, Injection, and Testing
Indirect Drive Target Aerosol Limits, Foam Mechanical Properties, and Target Injection Accuracy - Ron Petzoldt reviewed the ID target aerosol limits, maximum size droplet size, and vapor density, to avoid the need for in-chamber target tracking. The maximum droplet size is approximately 0.29 mm for a 1-g/m3 chamber density. This set of parameters would correspond to ~ 1% loss of ion beam energy for a 3.5 GeV Pb ion beam.
Optical scattering and extinction efficiency is important to track the target in the chamber. Ron showed the results of scattering and extinction for the considered chamber vapors. These data showed that very small particles of salt (FLiBe) have less effect on laser light than Pb particles. Beam extinction places limits on the particle number and the mass density.
GA is starting to study the mechanical properties of ID target foam materials as they relate to and limit target acceleration. This work is just beginning.
Ron advised that perhaps a better target accuracy capability might be possible than the original requirement of ± 5 mm at the center of a chamber for indirect drive targets. The study was requested to allow the thick liquid closer to the target path, thus better shielding the magnets and smaller indirect drive chamber radius. It is desirable to reduce the accuracy to ±1 mm to allow improved magnet shielding. He compared this to the demonstrated capabilities of a match air rifle (± 0.7 mm at 10 m) and the LBNL gas gun (± 6 mm at 10 m). Ron listed the possible causes for poor accuracy. The team pointed out that air rifles use rifling to spin stabilize the projectile that is quite a bit heavier than the DD or ID target. Spin stabilization is necessary for indirect drive targets to keep the axis of symmetry aligned with the driver beams and probably has less affect on target accuracy in a vacuum environment.
Ron identified the recommended materials for the near term ID targets, but the ARIES team is looking at target design simplification and material substitutions to ease material costs and reduce fabrication and assembly costs. These have to be compared to the possible reductions in target gains. This is recommended as a high priority task for the ARIES team next year.
ARIES -IFE Summary, Action Items, and Plans - Farrokh Najmabadi provided budget detail on the money provided to the Advanced Design Research in FY 2002 and the currently envisioned FY03 budget. He noted the increased emphasis on IFE work as compared to the earlier reduced IFE scope. The MFE work will be ramped up to address the Compact Stellarator viability as a power plant energy source.
The ARIES-IFE FY03 work will focus on the design window definition for both the thin and thick liquid chamber protection schemes. The thin liquid effort will focus on film generation, coverage, and stability. The chamber clearing issue will continue to be analyzed, especially the aerosol source terms. The design window for the injection and tracking will be examined for an aerosol size and density. The propagation and focusing of beams will be analyzed for acceptable solutions. The thick liquid protection scheme will continue to draw from the HYLIFE team for design and technology expertise. Aerosol production and interaction with the beams will be further investigated. Protective flow arrangements will continue to be optimized. Material selection will be refined and justified. Additional work on the HIF target hohlraum material selection and cleanup will be refined along with the waste or recycling issues. Heavy ion driver interface studies will focus on the beam propagation into the chamber.
Farrokh said the ARIES-IFE papers already written would be submitted to the Fusion Science and Technology journal.
ARIES-Compact Stellarator Study Presentations
Compact Stellarators as Power Plants
Review of Modern Stellarator Reactor Studies - Jim Lyon of ORNL discussed the advantages (ignited and inherently steady-state) and disadvantages (complex non-axisymmetric coils) of stellarators as fusion core for a power plant. The Compact Stellarator (CS) device has distinct advantages because it combines the better stellarator and tokamak features. Two new CS experiments (NCSX and QPS) are being designed to verify the CS promised performance and features. Jim described several recent and current reactor studies and experimental facilities.
Jim discussed the stellarator coil systems, the divertor regions, and blanket systems as adapted to the unique axisymmetric geometries. Maintenance is a particular difficult issue to solve. He illustrated that the CS device might combine the best features of both a tokamak and traditional stellarator, e.g., compact power density without disruptions, feedback controls, or external current drives. Major parameters for several point design power reactors were discussed. Jim concluded with a set of lessons-learned to apply to the new ARIES-Compact Stellarator study.
Steps Toward a Compact Stellarator Reactor - Hutch Neilson proposed that compact stellarators might improve our vision of attractive magnetic fusion power plants. He views ARIES as critical to optimize the CS as a power plant. Stellarators are 3-D toroidal configurations that have up to 100% of the rotational transform generated by external coils, hence no current drive, rotation drive, or instability feedback systems. The CS devices offer a lower aspect ratio, higher power density, lower physics risk, and shorter development time.
Recent stellarator physics results from Wendelstein 7-AS (Germany), Large Helical Device (Japan), and Helical Symmetric Experiment (Univ. of Wisc.) have shown increased beta performance, enhanced confinement, improved density control, longer pulses, and validation of quasi-symmetry operation. Two new US experiments are being proposed: National Compact Stellarator Experiment (NCSX), an axi-symmetric compact stellarator and Quasi-Poloidal Symmetry (QPS) stellarator experiments. The NCSX, scheduled to begin fabrication in FY03, has modular stellarator, TF, PF, and trim coils to investigate field configuration variations.
Hutch's vision of the CS reactor is a steady state device with no disruptions, no conducting structures or active feedback control, no current drive, and a high power density. The optimum configuration involves tradeoffs of many configuration, plasma, and engineering features. He intends to build upon the rapidly growing stellarator physics database. Then drawing upon the ARIES expertise, the combined team would explore the parameter spaces and trade the features to arrive at an attractive power plant design configuration. The configuration space need not be limited to those previously proposed as some innovative designs may offer promise.
Physics Issues and Optimization Codes- Mike Zarnstorff discussed the inherent advantages of a stellarator as a power reactor. There are two methods of attractive 3-D orbit confinement: quasi-symmetry (e.g., NCSX) and quasi-poloidal (e.g., W-7X and QPS).
Mike outlined the generalized stellarator design process as an iteration of the fixed boundary equilibrium conditions and defining the coil design. When optimized, a free boundary analysis will be conducted to determine the robustness/flexibility and discharge evolution. Mike also explained the recent advances in the ability to design for orbit confinement and good flux surfaces while obtaining the desired plasma physics properties. This results in stellarators that are steady state devices with no current drive, no disruptions, and plasma shaping for good plasma physics parameters.
Stellarators have a very large number of free parameters with a lot of configuration space to explore. Contrary to many other magnetic fusion concepts, stellarators cannot simply be scaled in size or power level. Rather, they have to be reoptimized for each significant increase in power output and physical size. High-speed numerical optimization processes now allow an efficient examination of configurations. Mike then explained the detailed process and tools to examine each step of the optimization process. Based on the NCSX experience, Mike outlined the steps necessary to begin the examination of a compact stellarator reactor.
Initial Evaluation of Computational Tools for Stability of Compact Stellarator Reactor Designs - Alan Turnbull was unable to attend the meeting. He plans to meet with Farrokh Najmabadi as soon as possible and discuss this subject.
Key Engineering Issues and Constraints - Phil Heitzenroeder explained how the stellarator engineering design process differs from a more traditional magnetic concept. The designers work closely with the physicists and optimizers to mutually agree upon compromise solutions. The plasma surface is quite unsymmetrical although it is usually repeatable in modes. The coils are typically non-planar and close to the plasma. Access space for blankets, diverters, and remote maintenance is quite constrained between the coils. The coil systems are very complex and are the most challenging to design and fabricate. One key constraint is the minimum bend radii of the coil (out of plane) that can be fabricated. In addition to the stellarator coils, it is possible to have TF, PF, and trim coils that share some of the field generation function of the stellarator coils so the stellarator coils can be simpler. Determination of the coil forces is quite involved.
The design of vacuum vessel is challenging to be modular, allow sufficient penetrations, and support the loads. Remote maintenance is very difficult due to tight access constraints and 3-D geometries for access to life-limited components. New geometries should be considered to help ease the engineering of the compact stellarators.
Stellarator Reactor Optimization and Assessment - Jim Lyon stressed there are no simple scaling laws for beta limits and confinements as both depend on the details of the magnetic configuration. As Phil Heitzenroeder explained, the divertor and maintenance requirements are much more complex than for axisymmetric systems. Any systems code must incorporate the ability to model complex coil geometries and stellarator physics. Jim distinguished the differences between optimizing the reactor core that leads to fixed plasma and coil geometries and optimizing the plasma parameters that lead to plasma parameters, profiles, fields, and component sizes.
The minimum reactor size is determined by the closest distance from the plasma surface to the center of the coil winding (D). The average plasma radius is < a >. A configuration is characterized by the parameters of AD = R0/ D, Ap = R0/< a >, and Bmax/B0. Jim cautioned that the lowest < R >/< a > design may not lead to the most compact reactor design.
Jim has a 0-D systems code spreadsheet that handles fixed plasma and coil geometries with a large set of input parameters. It has been shown to be useful for size scaling for fixed plasma and coil geometries and comparing reactor configurations. This code may be used to examine the initial range of parameters for the reactor study. After the initial scan, a more detailed 1-D code can be used to complete the optimization process. Jim outlined the process to lead to an optimized stellarator power plant. The COE is the most important parameter for a power plant and the coils are critical to the success of the power plant.
NCSX Configuration Optimization Process - Long-Poe Ku reiterated the stellarator configuration space is vast and filled with local minima. There are over 30 variables for plasma shape optimization and greater than 200 variables for coil geometry optimization. In 3-D magnetic field topology, particle drift trajectories depend only the strength of the magnetic field, not the space of the magnetic flux surface. Quasi-axisymmetry cannot be attained simultaneously on all surfaces, but can be maximized along with other properties.
Long-Poe described the geometry and function of the NCSX coil sets. He also described the self-consistent pressure and bootstrap current profiles used in the optimization. Non-axisymmetric components in the magnetic spectrum are minimized that lead to good thermal ion confinement, low helical ripple, and acceptable ion losses. He then detailed the process for the plasma optimization and the process for coil optimization and definition. After these are determined as being reasonable, the island healing is accomplished for good plasma surface quality. Long-Poe again emphasized the importance of the distance from the plasma surface to the coil midpoint, D. Feasible values have been determined to be 17% of the major radius, R. Another important variable is beta. NCSX has a beta value of 4%, which is limited by the available heating power. Raising beta increases the bootstrap current.
Compact Stellarator Configuration Development Planning - Hutch Neilson proposed a strawman approach to develop the CS power plant configuration. The FY03 effort would 1) emphasize optimization tool development aimed specifically for larger power plant configurations and 2) identify/explore configuration space to identify attractive design space regions. The tools to be developed would determine the figure of merit for alpha confinement, engineering metric goals, cost of electricity metric, and stellarator systems code improvements. The configuration space metrics would include iota, R/< a >, R, B, ß, and QA/QP/QH symmetry plus typical coil representations, such as modular and tilted coils. He proposed that PPPL people would concentrate on configuration optimization and exploration of alternative configurations. ORNL would investigate alpha loss calculations and systems code improvements. The University of Wisconsin would look at physics and engineering inputs. NYU and University of Montana would look at concept optimization. Meanwhile, the core ARIES Team would provide feedback to the designers on the engineering criteria and engineering implications for the CS reactor core.
The second year would involve systematic trade studies leading to a reference configuration. The third year (and fourth) would be dedicated to design development and cost estimation and optimization.
Compact Stellarator Summary, Action Items, and Plans - Farrokh Najmabadi said that the ARIES program is commencing on a three-year study of compact stellarators as power plants. Such a study will advance the physics and technology of the CS concept and address concept attractiveness issues. The first year would be devoted to developing the plasma and coil configuration optimization tools for a CS power plant reactor core, determining engineering requirements and constraints, and exploring attractive coil topologies. The second year would be a more in-depth exploration of the configuration design space with the selection of a configuration for future study. The third year would be devoted to detailed system design and optimization. Responsibilities for key activities over the next three months were identified.