ARIES-CS Project Meeting Minutes
8-9 March 2004
University of California, San Diego
San Diego, California
Documented by L. Waganer
Ref: Agenda and Presentation Links: Project Meeting
Welcome - Farrokh Najmabadi welcomed the team to UCSD. He explained that several of the ARIES team members would be participating in the meeting via a phone hookup.
Status of ARIES Program - Farrokh Najmabadi told the team that he had a meeting with the VLT PAC regarding program status and funding. The ARIES CS study is a 4-year project that started in 2002. The budget has been reduced for FY04, thus the CS project will be extended an additional year to complete the effort.
Farrokh also briefed the team on the US ITER emphasis on magnets, cooling system, and baffles in the first wall/blanket system. Approximately half the money to support this effort will be diverted from the base program, which causes the shortfall in the Advanced Design area. The IFE program will also experience funding cuts, as they are not perceived to be an energy-related program.
Laila El-Guebaly summarized the status of the upcoming TOFE to be held at the University of Wisconsin – Madison in September 2004. ARIES will have a dedicated oral session for ARIES presentations. You should reserve your hotel ASAP, as the block of rooms will soon be released.
Next Meeting/Conference Call - The next project meeting will be held in June at the University of Wisconsin – Madison. It will be 1-½ days in length.
DOE Program Status- Al Opdenaker noted that the DOE budget discussions are continuing to address how to fund the ITER initiative while minimizing the impact on the base programs. A FESAC panel to address how ITER and the base programs can co-exist will be completed in June or July 2004. NCSX will stretch out and finish in 08. It is difficult to retain existing experimental facilities running periodically at a lower level. FIRE will wrap up in 04 and be held in contingency pending a commitment to move ahead with ITER. The Materials program will be reduced slightly. The 05 budget for Advanced Design will remain about the same as 04.
Compact Stellarator Reactor Engineering Assessment
Impact of Medium-Temperature Magnet and 2-Field Period Configuration on Radial Build - Laila El-Guebaly presented data on the new magnet design provided by Leslie Bromberg. The design approach combines medium temperature (15 K) and low temperature (4 K) magnet winding packs contained within an insulator surrounded by 316 SS, which is neutronically superior to Incoloy alloy steel. Laila said that the new magnet configuration has a standoff from the magnet to plasma (Dmin) 5 cm greater than the previous design. The new magnet design increases the fluence and local heating that requires the shield increase in thickness by roughly 1 cm. Laila updated the radial builds for all the blanket concepts to reflect the new magnet design basis. She presented the minimum radial distances (Dmin) for all concepts and noted the Flibe/Be option provides the thinnest blanket while retaining sufficient breeding margin. She provided data for the blanket and shield-only zones for all concepts. The present LiPb and Li blanket design concepts may not provide sufficient breeding margin for the two-field period configuration and more beryllium multiplier may be needed for the Flibe blanket. Radial builds may need to be updated as the MgB2 radiation limits are provided.
Preliminary Activiation Analysis for LiPb/SiC and LiPb/FS Systems - Laila El-Guebaly said that the SiC- and FS- based systems were selected for favorable activation results. They yielded first wall and blanket lifetimes of 6 and 5 full power years, respectively. The analyses were conducted at two radial crossections; one through the nominal breeding blanket and one through the tungsten carbide shield-only blanket. Data was provided on the radioactive activity and decay heat on the two blanket material systems. At intermediate times, the SiC systems offer 3-4 orders of magnitude less activity than the FS systems and the decay heat for the SiC systems drop sharply after shutdown. The WC filler generates the highest activity and 10-1000 times higher decay heat than the FW/blanket regions at one day after shutdown. Both blanket designs have Waste Disposal Ratings (WDR) that qualifies for a Low Level Waste (LLW) classification. The new magnet design offers a lower WDR than the older design and will qualify as LLW.
Progress in Assessing CS Maintenance Approaches - Xueren Wang reviewed the field-period maintenance scheme that was presented by Siegfried Malang last September. Clearances for the blanket and shield module inside the coil support tube were determined for a reduced-size 100-cm thick blanket and shield module. Laila said the 40-cm shield behind the 60 cm blanket is not sufficient to protect the magnet for the plant life. Even with the 20 cm thinner shield, there were a few zones that would require minor shaving of the interfering regions. This time, Xueren refined and verified the clearances for removing a whole field period in a radial direction, then pulling out the two blanket segments from both ends. As suggested, he illustrated the actual shape of the coil supporting tube as opposed to a schematic cylindrical torus. Superimposed on the tube were the grooves for winding the superconductor – six per field period. It was noted that a planar parting section is not possible, as it would intersect some coils. Instead, a non-planar parting line could be used that would retain the ability to move the sector radially from the remaining two stationary field period segments.
Xueren discussed the configuration and maintenance of the two field period configuration (12-coil case). This configuration has more room between the individual coils and no interferences were found between adjacent coils. This would enable removal of segments of a partial field period. Removing a blanket sector between fixed coils, per the ARIES-AT maintenance scheme, is not possible due to the reduced space between coils. However, large maintenance ports are possible. The coil support and winding scheme is also applicable to this two field period configuration.
Xueren then discussed the radial distances between power core elements, as provided by Laila El-Guebaly. He then showed the plasma and coil cross-section at 15-degree increments. Two minimum distances exist per field period and are 1.28 m.
He summarized that it is possible to remove an entire field period. The new MHH2 coil configuration (12 coil case) provides adequate room for maintenance ports and spaces for the blanket. Sector-based maintenance seems impossible for MHH2, however port maintenance seems possible with the larger ports.
Some of the Major Considerations in Designing a Ceramic Breeder Blanket for ARIES-CS - René Raffray reviewed the ceramic blanket design pressure that impacts the module design, weight, size, complexity, and tritium breeding ratio. The high-pressure helium can be routed through coolant tubes in the module and a low-pressure purge gas will be used. An EU study of accident scenarios were examined for this module. The present design is compared to the EU blanket design.
René compared the design features and performance of the Rankine and Brayton power cycles. The Brayton cycle would solve the safety issues that were previously raised. The cycle components and related thermal conditions are examined to determine the expected cycle efficiency. The advantages and disadvantages of using an intermediate heat exchanger is examined along with the expected pumping power.
Two Tube Failure Event for the Helium-Cooled Blanket - Lee Cadwallader presented the background data on a two-tube failure that would result in an overpressurization and failure of the blanket module. This chain of events would lead to a failure of a steam generator (SG) tube. This failure scenario allows steam to interact with the high temperature beryllium, which produces hydrogen and possible hydrogen explosions. There are several mitigation measures that could be applied to reduce the likelihood and consequences. Three failure cases were analyzed: independent tube failures, failure of a blanket tube combined with pre-existing SG tube leak, and failure of a blanket tube and a dependent SG tube break. The independent failure is considered to be an occurrence frequency of 5 x 1010/y. The pre-existing tube leaks are not considered as a significant concern. The dependent SG tube failures are not considered to be a credible result of a small tube breach event.
Alpha Particle Loss Assessments and Status of Coil Optimization for 2-Field Period Compact Stellarator - T. K. Mau chose to investigate the alpha particle losses for the 2-field period configuration as it offers maintenance and fabrication advantages. Specifically, he examined the 16-coil case. He presented the physical, geometry, and operational parameters. He used the ORBIT-3D particle trajectory to generate the lost alpha footprint at the last closed magnetic surface (LCMS). The early lost alphas are concentrated in a toroidal strip from -0.5 < qb< 0. Lower energy alphas are lost along the toroidal direction with a weak periodic structure. The heat load distribution was also presented.
TK noted there is a USCD objective to design and optimize stellarator coils. UCSD has obtained two codes and are learning their operation.
Preliminary Assessment of Porous Gas-Cooled and Thin Liquid Protected Divertors - Said Abdel-Khalik noted the objective to predict the performance of porous cooled divertors. He defined a test problem with heat flux on the exterior of a hollow annular porous divertor surface with the coolant media flowing through the media, as shown in the attached graphic. He defined the effective heat transfer coefficient, temperature distribution, and pressure drop. He concluded the heat transfer coefficient predicted by MERLOT and FLUENT appear consistent. The pressure drop predicted by heat transfer coefficient predicted by MERLOT is significantly lower than FLUENT. The acceleration pressure drop cannot be ignored.
A maximum allowable surface temperature for a thin liquid protected divertor has been established for evaporation and plasma contamination (~ 380° for Li). The objective is to establish the operational limits to prevent film rupture due to thermo capillary effects related to temperature gradients of the substrate. Said presented the analytical problem definition and the governing equations. The starting condition is a cosine temperature distribution of the substrate and a uniform thin liquid layer. In a few hundredths of a second, the film thins at the highest temperature location and then thickens to a slightly thinner steady-state solution. With sufficient temperature gradient and thickness, the film will rupture and recover, or completely rupture and never recover. He concluded a methodology has been defined to determine the limiting temperature gradients, which might be more restrictive than surface temperature limits for plasma contamination. Future effort will be to explore the applicable parametric spaces.
Scoping Study of the Helium-Cooled Porous Media for the ARIES-CS Divertor - John Pulsifer listed possible helium-cooled divertor approaches: simple pipe or tubes with or without micro channels, porous media, or other heat transfer enhancement techniques (pins, fins, and fibers). He chose the porous media because the tools (MERLOT) were available to calculate the performance. He outlined the capabilities of MERLOT when applied to this problem along with previous study results. The performance of packed spherical particle beds is presented. Tailored foam structures and fibers are better. For constant pressure drop and helium mass flow, narrower channel widths provide better heat transfer coefficient (heff) at higher velocity and higher porosity. There are also local porosity variation effects (perhaps radially) that affect the heat transfer coefficient (heff). John then listed the approaches to size the tube for lower pressure drops for the ARIES divertor solution. Using the pressure drops in the open tube and the porous media as constraints, an optimum inner radius of 11 mm is determined. Parametric results were developed for a 10 MW/m2 tungsten divertor with a possible increase to 12 MW/m2.
Preliminary Coil Definition and Fabrication Assessment of Stellarator Magnets - Leslie Bromberg described the design approach to incorporate a dual temperature superconductor, such as MgB2 and NbTi. He showed a preliminary design of a coil cross-section and key design parameters. He developed coil performance over a range of magnetic field strengths. He also addressed the manufacturing approach for the superconducting coils. He then explained the approach and benefits of using a dual temperature superconducting magnetic system mainly lower cost solution with a more complex design.
Leslie presented the design and costing algorithms for the superconducting coil systems. See his charts for detailed data and algorithms. Leslie discussed future work areas and expected results.
Compact Stellarator Reactor Integrated Systems Assessment
NCSX, MHH2, and HSR Reactor Assessment Results - Jim Lyon has been focusing on the highest leverage areas, namely improved density and impurity profiles, high Tc superconducting coil systems, reoptimized HSR modeling, and testing models and assumptions for inclusion into the systems code.
Impurity effects were examined for other stellarators, as well as ARIES-CS. Power balance modeling was also examined. Three stellarator configurations are explored and characterized: NCSX-R (QAS with 18 coils), MHH2-R (QAS with 16 coils), and HSR (QP with 50 coils). He examined three superconducting coil configurations: high temperature superconductor (HSC), MgB2 at 15K, and NbTi. See Jim’s slides for specific result details.
In summary, he has developed an improved treatment of ne(r) and impurities, provided revised performance parameters for three superconducting coil types on three reactor approaches, and continues to improve the systems code model.
Compact Stellarator Reactor Physics Basis
Reactors with Stellarator Stability and Tokamak Transport - Paul Garabedian noted that the W7-AS and LHD codes have demonstrated beta values in excess of those theoretically predicted. Thus it is recommended the ARIES-CS explore high beta value cases. Compact stellarators with either 2 or 3 field periods are predicted to have good plasma properties. Paul is considering cases with an MHD force balance in the conservation form over the closed control surfaces. He showed a case where the force balance is not in the center with differing slopes on opposite sides of the current sheet. He also showed Pontcaré maps of the LHD flux surfaces. The MHH2 flux plots show a low order ballooning mode in the solution. Results for a quasineutrality in a Monte Carlo computation of the MHH2 energy confinement time for reasonable plasma conditions show excellent QAS and the radial field rise to a potential twice as big as the temperature.
MHH2 coils are shown that exhibit with ample room for maintenance between coils. There are a few points that are close and can be reshaped to eliminate any interference or close tolerances. The MHH2 stellarator at A=3.5 was constructed with the NSTAB equilibrium code.
Energy confinement times were provided for the KG3 configuration, which has good QAS, similar to the MHH2 coil set. Coils for this configuration are also shown.
Paul also showed flux cross-sections for a bi-furcated ITER plasma condition. It shows a large m=3, n=2 magnetic island suggesting the advanced tokamak may be close to being non-linearly stable.
Progress Towards Beta Limits for Compact Stellarators - Alan Turnbull summarized his recent work on equilibrium and stability analyses to cover more configurations with higher fidelity and verification to other codes and experiments. Convergence tests were shown on three-field period scaled NCSX equilibria that confirmed previous stability results. Fixed boundary MHH2 two-field period equilibrium cases were constructed from VMEC. He described the sequence of fixed boundary MHH2 equilibria as constructed from VMEC by uniformly scaling pressure. The VMEC calculations suggests MHH2 may be close to the equilibrium limit at beta of 6.92%. Contours at high beta are not strongly deformed from the low beta case. Both two-field period MHH2 and three-field period NCSX equilibria were shown and the main shape difference is the lower order shaping near Nfpf = 180°. Alan showed charts that indicated rotation transform profiles are different between scaled NCSX and MHH2 plasmas. He tested MHH2 equilibria with increasing beta values.
Alan noted his initial MHH2 stability results are suspicious – growth rates are small and insensitive to wall position, stability calculations used a limited number of harmonics, Terpsichore inputs need to be verified, computed mode structure need to be validated, and convergence studies conducted at higher radial mesh and poloidal points added. The VMEC Mercier criterion indicates stability up to a beta of 5%. Alan then summarized all progress since the last meeting. Tony Cooper is reluctant to allow any updating to the Terpsichore code. It seems Terpsicore problems running on Linus are related to memory allocation specifications.
Modeling of Particle and Power Control - Arthur Grossman discussed the general issues of particle and power control on reactor sized divertors, noting that divertors rather than limiters are the preferred solution for both tokamak and stellarator reactors. Many of the issues of power and particle confinement are similar in tokamaks and stellarators. However, the 3D nature of the stellarator magnetic fields often implies that the magnetic field structure in the scrape off layer is much less ordered and the separatrix may not be as well defined as in a tokamak. He noted that as a result, stellarators can have a natural divertor action, whereas tokamak reactors need special divertor coils to achieve a divertor configuration and control the flow of particles to specific divertor locations.
He adapted the W7-X codes, modified for bootstrap current contribution to the magnetics as well as interface with the U.S. VMEC2000 (rather than NEMEC) for the Compact Stellarator assessment. For the Compact Stellarator, which combines the best features of the tokamak and stellarator, he is analyzing the free boundary finite beta magnetic field structure using MFBE2001-VMEC2000-GOURDON codes to determine if compact modular coils can provide a natural divertor structure as in more conventional large aspect ratio stellarators. This is a first step in scoping the divertor parameters. He used an NCSX configuration adapted for the ARIES-CS studies to look for magnetic islands and ergodic structures, which could be the basis for a compact stellarator divertor with neutral pumping enhanced by baffles and divertor particle recycling. He examined NCSX configurations at 0%, 2% and 4% beta solutions and showed Poincaré plots that indicated an ergodic structure with island remnants comprising the scrape off layer, with SOL magnetics that do indeed change with finite beta.
The larger major radius ARIES-CS versions of NCSX-like configurations require many more toroidal planes (with each plane handled by a separate processor) for comparable accuracy. The initial Poincaré plots also indicate a mostly ergodic structure, but comparable numerical accuracy has not yet been attained and may require many more parallel processors than is commonly available. The alpha particle plot results are not random patterns, but structured solutions. Calculating the heat flux is possible, as has been done for NCSX via statistical GOURDON code field line peaking factor calculations on a conformal first wall. This technique provides a prediction of minimum and maximum heat fluxes regions and the angle field lines and particles make when they cross walls and divertor targets.
Development of a New Class of QA Stellarator Reactor Configurations - Long-Poe Ku highlighted that flux surface integrity is an important issue in the design of a QA plasma configuration because the bootstrap current may cause the rotational transform to cross lower order rational surfaces. The difficulty imposed by lower order rational surfaces can be minimized by resonance healing or choosing appropriate iota profiles. He proposed to investigate plasma configurations with large, externally generated negative shear such that at the target beta and current, the rotational transform will have small (but positive) shear through the plasma, thus making it possible to choose the iota to stay in a region devoid of the low order rational surfaces.
Long-Poe Ku shows that such configurations do exist, with the general characteristics:
Specific cases include runs KJC167, KJB277, KKD863, KBF122, and KDB124. The case KJC167 was used as a primary example for which the plasma boundary shape and VMEC flux surfaces are explored. The rotational transform for KJC167 has no N = 1 resonance at a beta of 6%. Good QA is achieved despite the large shear in vacuum iota and is not significantly affected by plasma pressure. The effective ripple is less than 0.6% everywhere and a good magnetic well is obtained. Alpha losses are localized into clustered regions.
Long-Poe showed several other configurations with similar properties in different regions of iota for both 2 and 3 field period cases. Configurations KKD863, KJC167, and KBF122 were compared for different iota values. KKD863 shows nearly nested flux surfaces for 6% beta with no low rational surfaces. He showed the increasing complexity of the plasma shaping as the demand of vacuum magnetic shear increases. He compares KKD863, KJC167, and KBF122 configurations for optimized QA. He also compares these for differing degrees of bootstrap current. The configurations KJC167 and KBF122 are very similar with only small differences in the outer plasma configuration. Additionally, he compared these new configurations with configurations having small, but positive, externally generated magnetic shear as well as configurations of NCSX type. In particular, the NCSX class configuration B5D was compared to KJC167. B5D has more complex plasma shape somewhat better QA, but comparable alpha confinement. B5D is stable to N=1 external kinks and infinite-n ballooning modes at 4% beta using ideal, linear MHD theory. It has a very shallow vacuum magnetic well, ~ 0.8%. It is unstable at 6% beta, with a Terpsichore calculated eigenvalue 1.5·10-3 for the N=1 kinks. The configuration KJC167 is unstable to external kinks and infinite-n ballooning using the same types of calculations at 4% beta. It is ballooning stable at beta ~ 2% and has ~ 4% vacuum magnetic well. At 6% beta, it has a Terpsichore calculated eigenvalue 6.0·10-3 for the N=1 kinks.
Long-Poe Ku thinks he has a good initial set of new QA configurations with good flux surface quality at high beta. These have good quasi-axisymmetry, a good confinement characteristic, and deep magnetic well. The configuration space also appears to be broad enough to encompass various iota regions, aspect ratios, and field periods.