ARIES Project Meeting Minutes
27-28 July 2011
Documented by L. Waganer
Ref: Agenda and Presentation Links: Meeting Agenda
Welcome/Agenda - Les Waganer welcomed the ARIES team to the Gaithersburg Hilton hotel. Les reviewed the day and half agenda and discussed the topics to be presented.
Next meeting - The next meeting will be held in the Washington D.C. area on October 13-14, 2011, (Thursday and half a day on Friday). Al Opdenaker will make the meeting arrangements.
Plans and General Scope
Updates of the Fusion Budget and FES News and Rumors - Al Opdenaker mentioned that the current FES grant programs (UCSD, UW and GIT) should submit their 2-3 page summaries of their planned 2012 effort to get the contractual effort started. The current program will likely stay the same, but future efforts may be more aligned with the main FES development direction.
James Van Dam will replace John Willis as the Director of the Research Division. FESAC is meeting July 28 to address several new charges.
Updates of Current ARIES Project - Farrokh Najmabadi said that the current ARIES program has and will continue to address the ability of the proposed design points and supporting technologies to have effective tritium breeding, handle the anticipated high heat and particle fluxes and provide an integrated design solution. FESAC and other national and international parties view the ARIES work as very insightful and helpful in determining future trends in the fusion R&D and future machines. Assessing critical engineering issues are very worthwhile and ARIES is viewed as an unbiased agent. The designs should be as complete and well documented as possible. The project needs to begin ASAP to focus on our completed analysis and design points.
Chuck Kessel supported these thoughts and noted there is no current U.S. Demo study in place - this would be useful to help provide applicable and focused R&D. The current four-corner studies will be beneficial to help understand the benefits and detriments of using either conservative or advanced physics or technologies. Chuck suggested that it would be good to have one or two in-depth power plant studies based on the current results, perhaps with the DCLL blankets as it is the more favored approach. However, Al Opdenaker noted that the projected budget would be insufficient for this level of effort. Farrokh said that he has hoped that after the current analysis is complete, we can develop preliminary designs for each of the four corner parameter sets, however these design points have not converged yet. For example, we need to better understand the ELMs and disruption scenarios with suitable technical solutions.
ARIES Task Results
Generalization Changes to the ARIES Systems Code - Lane Carlson summarized the current progress to issue the April 2011 SiC strawmen for ACT-I and ACT-II. The code now contains a generalization of the radial build for the various blankets. The code also has better transparency for parametric analyses with more input parameter file capability. Several suggested changes have been incorporated. Most of his effort has been to incorporate the DCLL blanket scheme into the SiC blanket module. Pumping power definition has been updated, yet separation of helium and liquid metal pumping powers for the SiC blanket remains in work.
Lane presented the ARIES Systems Code (ASC) simulation capabilities and results at the June 2011 SOFE meeting. The significant results were the real-time filtering of key machine parameters.
Lane illustrated the difficulty to incorporate specific design approaches into the code via radial builds for the inboard, outboard and divertor regions. This effort remains in progress. He is also working on the capability to tailor the structural geometries of the blankets, shields, hot structure and vacuum vessels.
The plasma surface area has been corrected to better reflect the actual surface area related to common shape parameters. There had been a disagreement between the ASC and the CAD software solutions and this has been corrected.
The updated ACT-Ib has been issued which now includes volume and costs for the VV ports. The plasma surface and NWL is now more accurate. All hardwired parameters have been replaced with input values. Lane illustrated this change for the input variables.
He further compared the previously published ARIES-AT results with the new ACT-I (aggressive physics) and ACT-II (conservative physics). He noted that he will continue to enhance the ASC and after all the DCLL changes are implemented, he will issue the ACT-III and ACT-IV design points.
Breakdown of Elements Degrading TBR of ARIES-ACT DCLL - Laila El-Guebaly asserted that an accurate assessment of the tritium breeding ratio (TBR) can only be done with the DAG-MCNP code that couples the CAD with the 3-D MCNP code. Accuracy is necessary because a 1% error in TBR translates into 1.1 kg of tritium per full power year difference in a 2000-MW fusion power machine. Both overestimating and underestimating the TBR induces significant problems, thus the desirable operating window is quite small. There is also a concern that the TBR varies over time due to Li burnup, depending on the breeding materials. There is the possibility of online adjustment of the TBR for liquid breeders. Several factors influence the TBR margin depending on materials, fidelity of data, design uncertainties and modeling errors.
Laila showed the current ARIES-ACT design that will be modeled with the most current 3-D neutronics code (DAG-MCNP with FENDL-2.1 point-wise library) to determine the necessary blanket thickness, lithium enrichment and capacity for excess breeding, if needed. The present intent is to design the blanket for a TBR of 1.05.
Laila then illustrated the step-wise approach to assess the several parameters that determine the blanket TBR capability (blanket dimension and coverage, first and side walls, cooling channel and flow channel inserts, stabilizing shells, assembly gaps, and penetrations). The recommended parameters for the ARIES ACT-DCLL blanket are a 0.65-m thick IB blanket, a 1.0-m thick OB blanket, 70% enrichment, all of which yields a TBR of 1.056 without outboard penetrations. She then provided additional information about the geometry and composition of the IB and OB blankets. Laila discussed adding breeding blankets behind the divertors, which can provide an additional 5% breeding, if necessary. The current margins should be retained until Demo can provide more definitive breeding data. The intent is to breed only a net amount of 1% excess tritium (TBR=1.01). Online TBR adjustment is a necessary control feature to assure the correct breeding amount over time.
Laila noted future efforts will concentrate on improving the plasma neutron source, the effect of the tungsten stabilizing shell, assessment of the envisioned penetrations and future updates of the blanket and divertor design. These effects will be assessed to update the TBR analysis.
Farrokh Najmabadi mentioned that the DCLL blanket is not ready to be incorporated into the ACT design basis yet and we should proceed in the future with the SiC blanket as the baseline for our first two detailed design points.
ARIES ACT-DCLL NWL Distribution and Revised Radial Build for ARIES-ACT-DCLL Case - Laila El-Guebaly reviewed the ACT-DCLL machine parameters used to evaluate the NWL analysis. She also showed the neutron source distribution for the 3-D NWL model and the results she presented (for inboard, outboard and dome). This analysis provided the peak NWLs for the revised radial builds on the inboard, outboard and divertor regions. The essence of the new radial build involves a thicker IB back wall (more protection for reweldable support ring and reweldable VV everywhere). All materials with high decay heat potential were eliminated. The builds also include a 2-cm thick helium-cooled thermal shield between the VV and magnets. The impact of the proposed changes is ~25 cm increase to the radial standoff.
More details were provided on the specified radiation limits, IB and OB radial builds and compositions, two OB cross-sectional views, magnet composition, and the VV radiation damage at the end of life (40 FPY). Again, she stressed the issues with the neutron streaming down assembly cracks. This can be mitigated with developing more detailed designs in the crack regions and mapping the local radiation fields.
Farrokh again stressed we need to put most of our efforts to bring together the SiC blanket design as the primary baseline for the near future. This means putting the DCLL blanket on the back burner for the time being.
Correlations For Divertor Thermal-Hydraulic Performance At Prototypical Conditions - Minami Yoda discussed their objectives and motivation to investigate the high-performance divertor heat flux and pumping power experimental and analytical modeling of helium-cooled flat plate divertors. She reviewed their experimental and analytical approach to generate parametric design curves. This approach was followed to develop the Plate Test Module used to determine the effective and actual heat transfer coefficients (HTC) which can be extrapolated to helium coolants. The maximum heat flux and loss coefficients were calculated using the experimental data. In turn, these data were used to prepare parametric design curves.
Turbine Plant Equipment Cost Graph Action Item - A question was raised to Les Waganer during the April ARIES meeting regarding the appropriateness of graphing the Turbine Plant Equipment (TPE) cost versus the plant thermal power. It was suggested he relate the TPE cost to the gross electric power generated. Les presented the cost accounts for the TPE system and related these costs to the most influential parameters. It turned out that only generator cost items were directly related to the gross electric power. Les presented the prior cost chart relating to thermal power as well as the new chart relating to the gross electric power. This latter approach clustered the parametric cost curves generally to single wide curve area and obscured the underlying cost determining factors. Therefore, Les recommended retaining the existing TPE cost algorithm related to the thermal power. He cautioned that these estimated costs are very parametric in nature and will need to be refined for more detailed studies and designs.
A Potpourri of Engineer Topics - 1) Mark Tillack first compared the AT, ACT-I and ACT-II machine parameters, highlighting those that are similar and dissimilar. He presented the proposed advanced SiC blanket to be used in the advanced technology design concepts. He highlighted the various elements of the SiC/LiPb blanket system. He discussed the thermal hydraulic and MHD considerations for the design approach. Flow conditioning will be an important element in keeping the MHD drag within reasonable bounds.
2) Mark reviewed the ARIES-AT coolant circuits along with major parameter considerations. These considerations are translated into design details for the coolant manifolds.
3) The selection of the material for the vacuum vessel has been passionately discussed for the past months. Mark summarized the recent history on this subject and noted that there is an assessment underway regarding the requirements, material choices, activation and R&D material needs. Mark noted that ITER chose 316SS as their VV structural material as it is relatively easy to fabricate and weld, is compatible with their water coolant and did not have significant embrittlement issues. However this material will not be suitable for a power plant application because of the high neutron activation (need to be low level waste). Swelling could also be an issue for regions subject to neutron streaming if doses approach ~20dpa. He then reviewed and compared the suitable material choices and presented some conclusions.
4) Beginning with ARIES-ST, one-sided roughening of the helium-coolant passages has been assumed, which would provide an HTC improvement of about two times with only added friction on one wall. Several design options have been considered for helium HTC improvements. However, Mark showed that HTC enhancements come with additional friction factor penalties, sometimes sizable. He showed more details about the impact of roughness on various wall surfaces. Dimpling also helps HTC enhancement. Mark suggested future analysis of the dimples with CFD analysis tools.
5) The impetus is to move to higher structural temperatures to gain thermal/electric power conversion efficiencies. Tantalum (or a tantalum alloy) is a possible structural material to consider for fusion applications. It has been assessed in previous MFE and IFE applications. In addition to high temperature operation, it has high ductility, after irradiation, at temperatures above ~700°C. Mark compared the properties of both tantalum and tungsten.
Activation Concerns for Candidate Vacuum Vessel Materials - Laila El-Guebaly mentioned that the steels being considered for the vacuum vessel can meet the current shielding and activation requirements relating to clearance, recycling and geological disposal. Siegfried Malang had addressed the vacuum vessel fabrication requirements: 1) can be a low-strength steel, 2) may operate at lower temperatures (150-200ºC), 3) exhibit little or no embrittlement at low temperatures, 4) be compatible with water cooling, 5) be easily rewelded, 6) require no PHWT and 7) have an acceptable level of neutron-induced swelling. Per Arthur Rowcliffe, the best candidate material seems to be the newly developed 3Cr-3WV reduced-activation Bainitic ferritic steel, which has the potential advantage of not requiring a post-weld heat treatment (PHWT). ORNL has developed this steel for which 50-ton heats have been produced. An ASME Code Case is being developed for this steel. Laila also mentioned this steel along with several other steels that have been proposed for this application. The level and composition of the impurities are very acceptable for all candiates. An example analysis of the ARIES-CS radial build was used to evaluate the material activation. Although all candidate VV materials are recyclable, none can be qualified for clearance even after 100 years of cooling. All VV materials (except for 316 SS) can be classified as Class C low-level waste at 100 years after shutdown. All VV materials (except for 316 SS and RAAS) can be classified as Class A low-level waste at 100 years after shutdown (by more than an order of magnitude).
Arthur Rowcliffe suggested the use of steel with a standard set of impurities, which would be achievable with relatively modest effort (cost). Laila showed the list of impurities that Arthur proposed. The use of these sets of impurities did reduce the activation levels for Class C and Class A low-level wastes.
Based on these analyses, Laila recommended the use of 3Cr-3WV ferritic steel for the vacuum vessel of ARIES.
Update on the Fracture Analysis of the Flat Plate Divertors - Jake Blanchard mentioned at the last meeting, he discussed the results of his analysis of the T-tube divertor design. At this meeting, he reported his analysis of the plate-type divertor with a tungsten helium-cooled flat plate cooling design. He used a 2-D planar model to analyze the design loads with simple convective cooling. He found that the fracture results are insensitive to small changes in temperature. Structural boundary conditions were adjusted to not over-constrain the model. The resultant model identified two areas of high stress, principally at the inner surface of the plasma-facing structure.
The fracture toughness of tungsten used a representative data set as the baseline. With this data, he predicted there might be a crack that could develop at the outer surface of the structure, near the root of the castellated divertor surface.
Jake plans on doing more 3-D work on the divertor modeling, examining the cracks in the tile and armor, and obtain the appropriate fracture toughness for the divertor material.
Possibility of Helium-cooled SiC Composite Divertor Design - Xueren Wang reviewed the two ARIES-SiC composite divertor designs. The helium-cooled SiC divertor of ARIES-I was designed to removed 4.5 MW2(peak) of surface heat. Pumping power was calculated to be 10% of the thermal power handled to achieve the temperature and stress limits. Based on recommendations from the ARIES SiC/SiC Town meeting of January 18-19, 2000, the combined ARIES-I stress limit for the SiC/SiC might be too conservative. The later ARIES-AT LiPb cooled SiC composite divertor was designed to handle 5 MW/m2. Since that design basis, the operating temperature revised to between 700°C and 1200°C, yet the stress limit remains at 190 MPa per ARIES-I and ARIES-AT. A SiC composite finger design can handle heat fluxes up to 5 MW/m2, but the pumping power and stress limits are exceeded. However, it the original ARIES-finger configuration or the SiC/SiC tube target plate with jet impingement are used, the pumping power and stress limits are satisfied.
A Low Technical Risk Assessment (TRL) Reliability, Availability, Maintainability and Inspectability (RAMI) Process for Fusion Energy Research and Development Backup - Tom Weaver proposed tailoring the successful Boeing RAMI process to the fusion arena to help identify and quantify the highest priority development needs. Tom identified the scope of the process and highlighted where his current activity has concentrated. Data mining and engineering-guided experimental planning may be the most promising areas. The aerospace industry has been using reliability-centered maintenance as an analytical process for a decade or more to determine appropriate failure management strategies. Six patterns of failure were identified to define and improve maintenance plans.
A Fusion Power Plant Reliability-Centered Maintenance Program for ARIES - Tom Weaver noted that aerospace and aviation businesses have been using new analytical processes to more efficiently develop tailored maintenance plans. This has been driven by major technology innovations in aviation, such as all composite major structures, thinner-low drag wings, all electric flight equipment and fiber-optic control and instrumentation systems. Faster and more efficient testing of these systems with improved maintenance processes was implemented. Increased statistical analysis was used to speed up and enhance the process. These methods and processes can be readily applied to fusion as many of the same elements are applicable.
ELM Loading Updates, TSC Simulations, Tritium Burn-up Fraction, and Plasma Equilibria - Chuck Kessel said that he generated a new equilibrium to remove the "wave" on the IB side of the plasma, using only two solenoid coils rather than five. He is progressing on the TSC simulation but it is not complete yet. He plans on using the new equilibria in a time-dependent scenario to keep the plasma diverted. Then he will increase the plasma parameters to the power plant level. He will then investigate disruption scenarios based on ITER prescriptions to assess the structural and coil configurations for feedback loops.
Chuck is concerned about the burn-up fraction of tritium in the plasma, especially the tritium confinement time (τT), which is largely undefined for large DT devices. The helium confinement time is also not well characterized. The plasma will also contain unplanned and planned impurities, which will have to be characterized.
Chuck examined the pedestal parameters as a way to characterize the energy released in the ELMs. Preliminary results indicate the ELM dissipated energy is in the range of 0.15 to 0.20 of the pedestal energy. Chuck noted ITER has a pedestal pressure database and scaling law and he estimated the pressure to be 210 kPa, which is similar to the prior EPED1 results. Using the pressure and the volume and accounting for the species, this yields the energy for the pedestal of 136 MJ and suggests an ELM energy 20-27 MJ. This result is lower than the prior estimate assuming the ELM energy is based on a fraction of the input power. Chuck then subdivided the energy loss into density and temperature parts. He can determine the amount of energy reaching the divertor with the remainder being deposited on the first wall. He examined the conductive and convective components of the available energy along with the impact on the divertor. He discussed the anticipated particle loss and radiation deposition (from the plasma) during an ELM event.
Preliminary Cost Information for Nuclear Components - Lee Cadwallader provided limited data on the incremental cost of providing nuclear-grade materials for fusion. Some data is available from INEL cost estimators. The basic nuclear-grade steel products typically cost up to 10 times that of commercial grades of steel, mainly in the alloying process, certified material test reports and quality assurance procedures. Complete components can cost at least 2X and 3X of commercial mechanical components. For electrical equipment, there are three grades: residential, commercial and industrial. A nuclear grade essentially does not exist due to lack of demand. The nuclear plants typically use industrial grade components and the additional cost is more than 2X.
Edge Model Results and Issues For Using A Radiating Divertor In ACT-I - At the last meeting, Marv Rensink showed results from a single null ARIES-RS configuration and a few double-null ARIES-AT results. Today, Tom Rognlien discussed the ACT-I double-null edge plasma modeling results including geometry and impurity injection. Tom began by trying to accommodate the ACT- heat loads without a strong impurity radiation factor. Tilting the divertor plates helps reduce the heat load on the divertor, but impurity radiation is a needed factor. UEDGE is being used to analyze the divertor. The transport barrier is just inside the separatrix, which regulates the pedestal plasma. Several attractive cases were found using high density neon impurity assumptions. A 280-MW case is acceptable and a 400-MW case is modestly over the 10-MW/m2 limit. The neon impurity case results in a significant core concentration. Tom foresees a wide operating range for the ACT cases.
Update on Self-Consistent Plasma Modeling of ARIES Baseline Design Points - Alan Turnbull reported that he has made progress in implementing the self-consistent simulation for ARIES-AT baseline. The results were a self-consistent transport simulation using the new TGLF model, which were more pessimistic in that a steady-state solution was not found and the profiles tended to collapse. Proposed efforts involve coupled equilibrium, transport, current drive and stability calculations to obtain a steady-state solution using the latest core transport and edge-pedestal models. Through April, Alan set up the initial cases and simulated the electron and ion densities and temperatures with LH and IC current drive system parameters using GENRAY
Next he will work on a complete self-consistent optimization of the ARIES-AT design point including a pedestal, current drive and heating. The iteration parameters will include ß, q-profile and (plasma) cross sections. Resistive wall mode stability may be considered. The LH wave frequencies may be varied to assist the optimization via GENRAY. The pedestal temperature and density parameters should be consistent with the EPED1 model, H-mode profile. Alan has obtained initial steady state results (~1 sec) using GLF23 and TGLF transport models with LHCD from GENRAY. The TGLF model incorporates cross-section shaping, but is significantly more pessimistic, perhaps not even reaching steady-state conditions. This may be due to full geometry modeling of the code.
Future plans are to complete the optimization for self-consistent steady-state solutions including ß, q-profile, plasma shape and size. He intends to use IMFIT to automate the core loop. The GLF12 transport model may also be used as it is faster.
Recent Progress on Scaling of the Power Scrape-Off Width in Low-Gas-Puff H-modes - Rob Goldston had previously reported that the plasma power scrape-off (SO) layer may be quite thin on larger fusion devices. This meeting, he reported that gas puffing might help thicken the SO region. Tom Rognlien asked how this gas puffing will impact the gas recycling. Rob noted that the SO region of JET has recently been reduced by a factor of 2.
Discussion of Future ARIES Plans - Farrokh Najmabadi wanted to make sure the systems code was fully validated and self-consistent before we embark on more detailed definition of the four corner regions of parameter space. Alan Turnbull should validate the 2006 physics data base before proceeding. Tom Rognlien should complete the divertor modeling. Chuck Kessel needs to complete stability, current drive, and system powers for the aggressive physics case. INEL should determine if water is acceptable in the vacuum vessel. Lane should complete the DCLL code results in 3 months. Lee provided some additional cost factors to consider, including an increment from first of a kind to an n'th of a kind plant. The conclusion is there no readily available cost databases or cost breakdown for fabrication and installation - any preliminary cost estimate will have large error bounds.
Review of Past and Current Action Items - - Les Waganer and Mark Tillack reviewed the past and the current action items (taken during the meeting) for the group. As each item was discussed, Mark edited each item, the responsible people and the due date. Les Waganer will update the prior action items and then post the action items. [Les later sent out a note to all responsible people about the action items and posted replies. Lane Carlson took the updated action item list and installed it on the web as an interactive spreadsheet. Les sent the spreadsheet link to the team for updating as required.]