ARIES-CS Project Meeting Minutes
16-17 June 2004
University of Wisconsin - Madison
Documented by L. Waganer
Ref: Agenda and Presentation Links: Project Meeting
Welcome - Laila El-Guebaly welcomed the ARIES team to the University of Wisconsin -Madison. She had arranged for several of the ARIES team members to participate in the meeting via a phone hookup. She mentioned the no-host dinner plans for the first evening. Stewart Prager will present the status of the FESAC Priorities Panel as the first topic. Following the meeting on Thursday, Professor Dave Anderson will show the group the Helically Symmetric EXperiment (HCX) and explain its operation and the benefits of its results. Laila also updated the team on the status of the TOFE meeting. There will be a special topic just for ARIES-CS presentations.
Status Report on FESAC Panel on Priorities - Stewart Prager explained the Orbach charge to the FESAC Panel to help focus the DOE program in a more complete and fundamental way by identifying the major science and technology needs into campaigns and themes. This effort includes ITER commitments. He identified the panel members and the work plan. He then discussed the overarching themes and the topical questions. These themes and questions then are organized into research thrusts and prioritized campaigns.
Stewart also discussed the Center for Magnetic Self-Organization in Laboratory and Astrophysical Plasmas, which is trying to understand the major physics problems of these plasmas. This NSF center is established in partnership with DOE. Specific topics addressed are dynamo, magnetic reconnection, angular momentum transport, ion heating, magnetic chaos and transport, and magnetic helicity conservation and transport.
DOE Program Status - Al Opdenaker is occupied with the FY06 budget and could not discuss specifics at this point. Congress is presently dealing with the FY05 budget. Timing of the budget resolution is unknown. Of particular importance is on how to deal with inclusion of ITER, which has not been resolved and may have an adverse impact. The budget for the Advance Design effort is not expected to be impacted.
It was suggested that ARIES prepare a summary review presentation for DOE within a few months. Al thought that would be an excellent idea once the budget crisis is resolved.
Status of ARIES Program - Farrokh Najmabadi stated that the ARIES-CS effort is roughly half completed. It will have 11 papers at the upcoming TOFE meeting to convey our current technical progress and status. Shortly, we will be able to commence on a design point for the Compact Stellarator. We also should be considering the next potential confinement concept to assess.
Next Meeting/Conference Call - The Team tentatively decided to hold the next ARIES conference call on 20 July (Tuesday). We also plan to hold a short, half-day meeting at the end of the ANS/TOFE meeting being held in Madison, WI. This meeting is on September 16 (Thursday) from 1:00 to 5:00 CST after the end of the TOFE Plenary Session. Tentatively, the following ARIES meeting will be held at the University of California, San Diego on November 4 (Thursday) and half a day November 5 (Friday).
Compact Stellarator Reactor Physics Basis
Reactors with Stellarator Stability and Tokamak Transport - Paul Garabedian has been investigating more realistic simulations using the NSTAB code. The problem may be that force balance and stability are lost across islands if equilibrium equations are not in conservation form. He is investigating the two- and three-field period configurations, as they are most likely to meet the equilibrium conditions. He showed a two-field period configuration designed using the NSTAB equilibrium code. He showed energy confinement times, in milliseconds, for a Monte Carlo computation of a typical NCSX configuration. Poincaré maps of four cross-sections of the LHD configuration were presented at a beta of 4%. Similar maps for the DIII-D cross-sections suggest that an advanced tokamak reactor might be subject to disruptions and loss of alpha particles. Paul showed MHH2 and NCSX plots at beta values of 6%. Paul thinks at a beta of 6% may be at the limit due to the evidence of ballooning modes on the surface features.
Progress of Configuration Development for Compact Stellarator Reactors - Long-Poe Ku reminded the team of the new class of QA (quasiaxisymmetric) configurations introduced at the March meeting. These configurations had strong negative shear provided by the shaping coils to match the magnitude of the bootstrap current. This avoids the lower order resonance at the target beta (6%). These configurations have good QA with low residual non-axisymmetric fields and low effective ripple. These configurations also have low alpha particle confinement and an energy loss of less than 10%. They have deep magnetic wells with an average elongation of >1.8 and a triangularity of > 0.7. These data are better than those of advanced tokamaks.
Long-Poe reports additional study results of the new configurations. The alpha loss fraction is now < 5% (KJC167m (modified) is an example). Robust flux surfaces are found for three-field period with intermediate and high iota cases. The QA is correspondingly improved as well as the effective ripple. The angular alpha loss distribution shows structures that correlated poloidal and toroidal angles. These results are obtained with more complex shaping of the coils and plasma. The PIES code predicts excellent flux surface quality for KJC167m. The configuration KKD863 also has good flux surface qualities over the entire plasma.
A second area was to determine if “reasonable” coil solutions may exist. “Reasonable” is defined to be conditions devoid of obvious adverse features with coil spacing consistent with needed plasma-coil spacing, coil-to-coil spacing, and number of coils. Long-Poe developed a coil set for the improved KJC167m plasma. They are much larger with larger spacing as compared to the NCSX configuration. However, when viewed on the winding surface, KJC167m coils shows two coils that have a long toroidal segment that folds back on itself and nearly touches itself, as shown below. This feature may be mitigated with a more twisted coil in that region or placing a coil at the half period.
Plasma KKD863, with an average rotational transform of approximately 0.7, requires more complex coils, but the design solution may be more reasonable.
The two-field period solution KDB124, shown at the right, has an average iota of 0.42 and an alpha loss fraction of ~ 7.3%. This configuration only has four coils per period and a wide spacing between coils. The coil shapes are more planar, but have higher order shaping that might be difficult to wind.
In summary, Long-Poe has made progress in reducing the alpha losses while maintaining or improving the plasma performance (beta). Rotational transforms can be modified to avoid higher order rational surfaces. Modular coils with reasonable shapes can be determined.
Status of Alpha Losses and Heat Load Assessments for Two-Field Period Compact Stellarator Reactors - T. K. Mau provided the background why it is important to develop a capability to do alpha loss and heat load assessments on fusion reactors. He adopted the MHH2 plasma with 16-coil set for the baseline assessment and discussed the physical and plasma parameters used. He used the ORBIT3D code with 4000 particles tracked to determine the alpha loss fraction. His lost alpha footprint plots correlated very well with those of Long-Poe Ku. He also varied the beta for several cases but the alpha loss maps showed very little difference.
He also showed heat loads on the last closed magnetic surface (LCMS) due to lost alphas. He attempted to calculate the heat load distribution by mapping from Boozer coordinates to real coordinates. Ku pointed out that the Boozer coordinates are not the same as those used in the VMEC equilibrium that T. K. assumed. Several sources were identified that have defined correct Boozer coordinates for these types of cases, and T. K. will make the corrections.
Compact Stellarator Integrated Systems Assessment
Studies of NCSX and MHH2 Reactor Configurations - Jim Lyon noted he has been focusing on a more accurate treatment of the 16-coil MHH2 configuration and has added the 8-coil configuration, recalculated the reactor parameters, corrected impurity radiation profiles, working on optimizing impurity radiation, and testing models and assumptions. He has studied the NCSX-R, MHH2 (8), and the MHH3 (16) configurations. The model has been modified to reflect the design features of the three options.
Jim reviewed the data provided at the last meeting. His new approach involves fixing the peak wall loading at 5 MW/m2 with the average at 2.5 MW/m2. For a given electrical power output and fixed thermal efficiency, this establishes the thermal and fusion power, thus establishing the reactor power core size. He then illustrated B/Bmax plots for varying magnetic coil parameters as related to the three reactor types. He discussed the radial gap considerations for the coil geometry assumptions. He then showed plots that determined Bmax for <ß> of 4% and 6% and wall loadings for different reactor types. Farrokh thought we should not be optimizing on Bmax because in all prior studies, optimal economic systems always chose the highest allowable field for Bmax. Farrokh recommended Jim immediately begin to implement costing algorithms in his systems code to commence optimizations based on COE, rather than assume some other optimization basis (Bmax or NWL).
Jim presented his treatment approach for the plasma impurities. He will be using a neoclassical impurity transport model, which is a conservative approach. The density of high Z particles will be peaked at the edge and the electron density will be minimal (hollow) in the center for stellarators. He showed some data from the W7-AS experiment to support his approach. From his assumed electron density model shape, he calculates density, temperature, and pressure profiles to impute an impurity density profile.
Compact Stellarator Reactor Engineering Assessment
Radial Build Definition for Solid Breeder System - Laila El-Guebaly summarized the candidate ARIES-CS breeding blanket options, highlighting the current one being defined – the Li4SiO4 solid breeder (SB) with ODS FS structure. This approach uses a beryllium multiplier, helium-cooled first wall and blanket, and a water-cooled vacuum vessel. The EU is proposing this design approach and Siegfried Malang has improved the designs and provided technical advice. The key findings from the UW investigations are that the peak neutron wall loading should not exceed 4.5 MW/m2, a beryllium to solid breeder ratio should be around 2:1, and a 65-cm thick blanket will breed sufficient tritium and overprotect the shield.
Laila showed a plot of all candidate solid breeder materials for TBR vs. neutron energy multiplication and concluded all solid breeder materials in actual designs must use a beryllium multiplier to meet tritium breeding requirements of overall TBR ³ 1.1.
Laila provided the design and performance groundrules in defining the solid breeder radial build. She concluded the upper temperature limit on the beryllium and solid breeder limits the radial thickness of the front SB and Be layers to 1 and 2 cm, respectively. Lithium enrichment also increases the heat deposited in the SB layers. She affirmed a 65-cm thick blanket of 32 layers would satisfy the breeding requirement of TBR of 1.1. She also listed the material composition of the blanket, shields, vacuum vessel, and winding packs along with a preliminary blanket cost estimate.
Laila compared the candidate blanket thicknesses and performance parameters. The solid breeder blanket has the highest energy multiplication, but it is the thickest, has one of the lower thermal efficiencies, has one of the shortest service lifetimes, and is one of the more complex, expensive blankets. In her opinion, liquid breeder blankets are superior to the other approaches, particularly for stellarators.
Ceramic Breeder Blanket Conceptual Design for ARIES-CS - Xueren Wang described the solid breeder blanket design considerations: design pressure impacts structure, cooling requirements determine helium coolant pressure (~8 MPa), and the solid breeder packing fraction affects volume requirements. A Brayton power cycle is adopted to avoid water interaction problems inherent with a Rankine cycle. Xueren showed a design of a box structure to house the SB blanket. Consistent with a radial build provided, he designed a blanket arrangement and coolant plumbing geometry. He explained the helium coolant flow patterns in the inlet and outlet flow channels with higher flow velocities in the first walls and slower velocities in the blanket regions. Concentric tubes would enable remote maintenance of the modules with sliding seals and remote welding machines. Xueren then examined the remote handling aspects of the modules.
Xueren discussed the results of the thermal hydraulic analysis and design constraints, pressure stress analysis, and primary membrane stress intensity during a LOCA. He felt the low-pressure requirement on the module with meandering coolant channels results in a simpler design. The modular blanket can accommodate the varying stellarator first wall surface. The pumping power is approximately 3.35% of thermal power.
Ceramic Breeder Blanket Parameters and Cycle Efficiency - René Raffray reviewed the considerations on the choice of the module design and power cycle: coolant design pressure and flow path through breeder. Coupling the blanket with a steam cycle raises the issue of possible Be/steam reaction in the case of tube failure in the steam generator followed by tube failure in the module. In this case, the module must be designed to withstand the high pressure in order to avoid module failure. Consideration of a Brayton cyle avoids this safety concern and allows for a simpler, lower-pressure blanket module design (the lower steel volume fraction also provides for better tritium breeding). René outlined the coolant flow path through the modules for the proposed concept.
Two Brayton cycle configurations were shown: I) a cycle with 3-stage compression and two inter-coolers and a single expansion stage and II) a more advanced cycle with 4-stage compression with inter-coolers and 4 stage expansions and reheaters. T-S diagrams were provided for both versions. The method of calculating the cycle thermal conversion efficiencies was examined. Both types of thermal conversion cycles were assessed assuming blanket FS temperature limits of 550°C (regular FS) and 700°C (ODS FS). The maximum allowable ceramic breeder and Be temperatures were set at 950°C and 750°C, respectively. The more advanced cycle (Brayton II) yielded the highest thermal efficiency, but at the price of unacceptable fractional pumping power, thus effectively ruling it out for this application. The Brayton I cycle optimization indicated that the maximum cycle efficiency is obtained for a neutron wall loading of ~3-3.5 MW/m2 and reaches 36.5% and 44% for FS temperature limits of 550°C and 700°C, respectively (for a fractional pumping power <5%).
A key parameter affecting the cycle efficiency is the plasma surface heat flux, as the He coolant temperature must then be lowered to accommodate the maximum FS temperature at the FW. On this basis, it becomes very challenging to couple a Brayton cycle with the ceramic breeder design for plasma surface heat flux higher than about 0.5 MW/m2.
Beryllium-Steam Reactor in Ceramic Blanket - Brad Merrill thought the main safety concern with the solid breeder blanket is the beryllium steam reaction that produces chemical heat and produces more blanket failures. He modeled the solid breeder blanket to analyze the safety problem. He explained the double tube accident progression that allows the coolant to pressurize the tritium extraction system, leading to a steam generator tube failure with radioactive decay and beryllium-steam reactions with heat generation. The MELCOR code evaluates the accident scenario. Brad shows many temperature time histories for different blanket locations. No thermal runaway occurred because the quantity of beryllium is small and the solid breeder surface area is small, which produces a small amount of hydrogen. Varying degrees of decay heat and size of tube break are examined to see if there is any significant influence. More accurate analyses are planned.
Temperature Gradient Limits for Liquid-Protected Divertors - Said Abdel-Khalik defined the design approach for a liquid surface that protects plasma-facing components. ALPS and APEX programs have investigated this approach and have established operating temperature windows for different materials based on allowable plasma impurity limits due to vapor pressures. This ARIES thrust has established limits for maximum allowable temperature gradients to prevent film rupture due to thermo capillary effects. This work examines spatial variations in the wall and liquid surface temperatures that might create thermo capillary forces that would lead to film rupture and dry spot formation. The initial condition is a uniform thickness liquid film of height, ho, that is superimposed on a 2-D Cartesian surface having a cosine thermal distribution. The governing equations are conservation of mass, momentum, and energy along with the long-wave theory with surface tension. Generalized non-dimensional charts have been created as a function of Weber and Froude numbers. The fluids of lithium, lithium-lead, flibe, tin, and gallium have been analyzed at their expected operating temperatures. For the maximum temperature gradient (that would cause a dry spot), the temperature gradient varies from 13 to 211 K/cm depending on the liquid material. A numerical solution yielded similar results and trends. Said concluded there are limiting values for temperature gradients, depending on the liquid coolant material. Generalized design charts have been formulated. Spatial thermal gradients of thin films may be more restrictive than surface temperature limits. Experimental validation needs to be done. Axisymmetric geometries will be more restrictive.
Power Core Materials and Fabrication Unit Cost Estimation - Jerry Wille from The Boeing Company described the need to develop the power core material unit costs and the basis for the estimate. These will apply to commercial fusion power and would exclude the developmental costs (10th of a kind units). He is trying to validate the costs with similar fabrication techniques. The costs are in 2004$ and validates with prior similar estimates. Jerry presented a table with preliminary unit costs for typical fusion materials.
In the future, he will assess more materials that will be used in the ARIES-CS designs.
Compact Stellarator Coil Design Definition/Fabrication Approach - Leslie Bromberg provided global deflection and stress analysis results for the NCSX modular coil structure. The models show good margin for the stress values even with stress concentrations around discontinuities and constraints. Based on this analysis, Leslie predicts the stresses for the ARIES-CS should be around 180 MPa for 4 Tesla at the coil. However, we want to go to 16 T, which will require more structure, especially at the outside perimeter. Strains were around 1% for the NCSX experimental hardware. With a 600 MPa stress limit, 16 T at the coil, and strains proportional to forces and inversely to stresses, this would suggest a deflection of ~ 5 cm for ARIES-CS.
The ARIES-CS coil set have no thermal loads, have a better space allocation, does not need TF coils, and are smaller coils on a relative basis. Leslie is trying to evaluate the Wind and React versus the React and Wind approaches.
Leslie addressed the thickness of the superconductor wire sizes and how the field scales. He is developing a simple code to calculate the magnetic field intensity for sizing, maximum field, and body forces. He is also evaluating a process to estimate the poloidal and toroidal magnetic fields.
Neutron Wall Loading Profile Using CAD/MCNP Interface (progress report) - Mengkuo Wang explained the current method of using the user-unfriendly MCNP code to estimate the NWL of a first wall surface. For more complex first wall geometries (stellarators), it is desired to couple the capabilities of a neutron generation code with a CAD code. A converter is possible, but compatibility issues limit its use and applicability.
To demonstrate, he assumed a simple torus shape with a simple outboard plasma shift. The results were fast, but not accurate. A more complex stellarator shape was used with an outboard magnetic shift modeling both plasma and first wall. On the very complex stellarator shape of ARIES-CS, the first wall could not be physically defined due to intersecting normals. The future plan is to fix the geometry errors, speed up the calculations, and construct a viable model of the first wall for a stellarator.
Tour of The Helically Symmetric Experiment
Professor Dave Anderson explained the theory and the history of his experiment before leading the ARIES team on a guided tour of the HSX experimental facilities.