ARIES Project Meeting Minutes
31 May - 1 June 2012
Documented by L. Waganer
Ref: Agenda and Presentation Links: Meeting Agenda
Welcome/Agenda - Les Waganer welcomed the ARIES team to the Gaithersburg Hilton Hotel and explained the facilities. He reviewed the day and a half agenda (see above link) and the topics to be presented. Al Opdenaker introduced Dr. James Van Dam, Director of Research Division, Office of Fusion Energy Sciences.
Summary of Fusion Energy Sciences Activities - Dr. James Van Dam joined the Department of Energy as the Division Director of the Research Division in the Office of Fusion Energy Sciences in September 2011. This is the first time that he has attended a Systems Studies Quarterly Meeting in his new position. He spoke about the coming reorganization of FES that is scheduled to go into effect on June 18th. As a part of that change, Al Opdenaker, the Program Manager for the Systems Studies program, will be moving into the Research Division and bringing the System Studies program with him. This will give Dr. Van Dam an opportunity to become more involved in the studies program. He expressed his appreciation for the excellent results that have been produced by the System Studies' team over the years. He is concerned about the impact of the FY 2013 budget cuts on the Systems Studies' program and he is hoping that some way can be found to mitigate the effects of that cut on the study's team.
Dr Van Dam also mentioned that materials research is another element of the Research Division's portfolio and is one of only two program areas in FES that will be receiving budget increases in FY 2013. He is expecting that the System Studies program will be able to help influence some of the specific research areas that DOE will need to pursue in the future.
Dr. Gene Nardella, the new FES Chief of Staff, explained some additional details about the reorganization the Office of Fusion Energy Sciences. These details will be available once the reorganization is in place on June 18th.
Next Meeting and Call - The next meeting will be held in about three months, however the date is TBD. Farrokh Najmabadi and Mark Tillack will decide the date and time for the next conference call to finalize the initial design configuration. (It has been decided that the next project call will be on Tuesday, July 3rd, 2012.)
Plans and General Scope
ARIES Project Update and Near-Term Objectives - Farrokh Najmabadi stressed that the present activity of defining the ACT-1 (aggressive physics and the SiC power core) design point is a critical milestone to be completed as soon as possible. The action items from this meeting (see last two pages of these minutes) are essential to complete this near term effort, as well as the longer range activities necessary to complete the design space of the four design points (aggressive and conservative approaches for physics and engineering) by the June 2013. The first half of next year will be devoted to finalizing the design details and documenting the results.
ARIES Task Results (Physics)
ARIES ACT-1 Core Physics Updates and Systems/1.5D Comparisons - Chuck Kessel informed the group that he has been obtaining plasma simulations with the TSC code, which closely match the JSOLVER equilibria solutions. These results are close to the ARIES-AT conditions, so he is recommending using the b/a ratio (plasma half height/half width) of approximately 0.33 with the vertical shell in the same location as ARIES-AT.
Chuck presented the graphical and tabular results of the 1.5D time-dependent simulation of the TSC operating point. These simulations are non-inductive, steady-state solutions within a range of desired physical plasma parameters. He also found that the plasma fueling and exhaust parameters are consistent with the UEDGE model analysis.
Chuck felt that the original plasma temperature profile is too broad. He found that discrepancies in ßN values were the result of T- and n-profiles at the plasma edge and slight volume differences due to the plasma shape. The 95% shape parameters yielded a viable 1.5D operating point. The 1.5D analysis suggests the normalized current drive is lower than assumed in ARIES-AT. The bootstrap current comparison indicates a correction is needed. Chuck presented and discussed the plasma parameters for the ACT-1 systems code results, the TSC 1.5D analysis and other ACT-1 systems code results. He discussed some new systems scans over a limited set of parameters with the new physics inputs and updates. Again, this is backed up with a table of the better results. He used the TSC results to determine the PF volt-seconds needed for startup, using conservative assumptions.
Chuck's remaining work is to finalize the assumptions for all four-corner concepts, rerun the physics database with new changes, incorporate edge plasma physics, rerun the engineering and cost models and focus on the desired operating point.
Edge Plasma/Neutral Modeling Results for ACT-1 Divertor (out of presentation order) - Tom Rognlien discussed his and Marv Rensink's results for the detached plasma modeling with the defined divertor plates and pumping. They are modeling the non-orthogonal shaped divertor plates, which is generally representative of the ACT plasma conditions. Per the action items from the Jan 2012 meeting, a few simulations have been done with argon as the impurity gas species instead of neon with no major difference observed other than roughly 25% more radiated power from argon for the attached plasma case at the same Zeff. They have been able to obtain stable plasma conditions with acceptable electron temperatures, neutral density and ion density. One of the new issues to arise is that the ratio of helium in the private-flux pump region to that in the midplane region is a factor of 3-4 smaller than modeled in ITER. Consequently, more DT may need to be pumped (and reprocessed) to remove the helium ash.
In summary, Tom and Marv believe the detached, radiative divertor remains the best option for ARIES-ACT, with certain caveats noted. There are also issues to be resolved, namely high neutral density near separatrix, identification and development of suitable diagnostics, startup scenario, maintaining the neutral pressure via a sufficiently closed divertor and how to do helium pumping.
Tom also noted that weakly pumped, high-density scrape off layers yield stable, highly-radiating edge plasmas. He supported this view with Slide 15 containing four substantiating points.
Progress on Optimization Results for Steady-State Plasma Solutions (out of presentation order) - Alan Turnbull reviewed previous results that showed the plasma profiles collapsed using the GLF transport model, however the TGLF model with real geometry might yield better results. Alan also proposed to implement the Multi-Mode model before this present meeting.
Alan and Holger St. John have implemented the faster Multi-Mode Gyro-fluid transport model, which did simulate the ARIES base case to a steady-state solution with no collapse. Two cases with different densities and plasma broadnesses were found, however the axis temperature overheats causing increased on-axis current. The Multi-Mode anomalous transport model can predict the temperature, density and rotation profiles. Te and Ti temperatures evolved to steady-state for both density cases without collapse. The Multi-Mode has a reduced flux in the outer regions.
In the future, Alan plans to benchmark this solution, try to eliminate the axis overheating, obtain steady-state solutions and iterate to obtain a new starting equilibrium state.
ARIES Task Results (Engineering)
Radial/Vertical Builds for ARIES-AC-SiC Power Core - Laila El-Guebaly compared the new design changes to those on the ARIES-AT. The former SiC-based LiPb-cooled HT shield is now a 20-cm thick He-cooled steel-based ring. The vacuum vessel (VV) is thinner and is cooled with high temperature helium (550°C) instead of water. The ACT design has a lower temperature, water-cooled LT shield with WC or B-FS filler, which is now located outside the VV. There is also a change to the magnet technology. The ARIES-ACT low-temperature magnet operates at 4K whereas the ARIES-AT high-temperature magnet operates at 70K.
The TBR for the SiC design is currently adequate. She also provided the damage limits to the SiC and FS structure, but W limits are unknown. The VV and the steel structural ring are not reweldable after operation. She provided the radiation limits for the superconducting magnet materials. Lifetime is estimated to be 40 full power years, which may allow an availability of 85% or better for the 10th of a kind plant. Operational dose to workers and the public should be < 2.5 mrem/h.
For the radial build, Laila presented the 2006 ARIES-AT inboard radial build for reference. She then showed the new ACT-SiC inboard radial build. The new build is 6 cm thicker and the steel structural ring needs to be replaced every 10 FPY. (Note that calendar years will be longer due to time in the Hot Cell and Availability effects). The increase in inboard radial build is mainly due to cooling the structural ring with He instead of LiPb. None of the inboard components are reweldable during the 40 FPY of operation.
For the outboard radial build, again the AT and the ACT-1 designs are compared. The ACT concept is thicker by 14 cm than the AT version. Most of the inboard comments also apply here.
The structural ring and VV are not reweldable after operation. Without gaps, the structural ring, VV, LT shield and magnets are life of plant components. The gap neutronic analysis will be assessed in the future.
The ACT vertical build is also thicker by 10 cm as compared to AT. The LT shield has the same composition as the IB region. The lifetime of the divertor is TBD.
Laila provided the SiC blanket composition by region.
Laila is concerned about the assembly gaps, the maintenance ports without inner doors, and penetrations for pumping, heating, and fueling. Laila showed the 3-D model of the maintenance port without the shielding door at the inner region of the port. For this concept, there was only a small triangular shield wedge to protect the OB legs of the TF magnets and this was rather ineffective as the 3-D analysis showed excessive nuclear heating in the OB magnet legs. It was suggested that for an option, Laila should analyze a larger wedge of shielding material and possibly changing the aspect ratio of the OB magnet to add extra shielding along the sides of the magnet. An alternative approach is to add an inner shielding door at the plasma-side of the maintenance ports (as in ARIES-AT) to help protect the sides of the OB legs of the TF magnets and other externals.
ACT-1d Design Point Definition - Mark Tillack has assumed leadership on the ARIES Systems Code and is becoming proficient in its coding and execution. He said the ACT-1d systems code run will be issued with the new physics and the existing code infrastructure. New builds and compositions from Laila have been incorporated into the code. New code utilities have been implemented to allow pre-filtering of the design space. Mark illustrated the code results with the existing physics.
Mark outlined the process for the design point selection for the SCLL power core. The aggressive physics will be regenerated and will be filtered to reduce the design space (ACT-1d). Mark then presented his approach and schedule for obtaining the remainder of the four corner concept design parameters.
Final 3-D Tritium Breeding Results for ARIES-ACT-SiC Power Core - - Laila El-Guebaly provided an overview of the ARIES-ACT-SiC design concept she has been analyzing - namely LiPb-cooled SiC blanket with LiPb manifolds for inboard blanket and no blanket behind the divertor. The UW team has been running the 3-D neutronics code in a step-wise fashion to fully understand the importance of each design element: FW, side/top/bottom/back walls, cooling channels, assembly gaps, stabilizing shells, and penetrations. The goal is to achieve a TBR of ? 1.05 by varying either or both the blanket thickness or the lithium-6 enrichment. The result was that the present radial builds achieve a TBR of 1.05 using a 55% to 60% lithium-6 enrichment. The main findings are:
Her next steps will be to update the cost of lithium enrichment and to analyze the DCLL blanket to determine its TBR.
Updated ARIES-ACT Power Core Definition and SiC Blanket - Xueren Wang began by showing the key components of the power core sector, including the inboard and outboard blankets, divertor and structure and the structural ring. Xueren showed that, if needed, a two-zone (plate and finger) divertor concept can be used in this design. Xueren also presented an alternate design option.
Xueren showed the power core design integration, but the team was concerned that there is little or no detail about how the sector will be removed from the core to the maintenance ports.
A backup option for the local water-cooled shield was shown with local shielding on the walls of the maintenance port. He also showed the arrangement of the cooling access pipes as connections to the ring headers.
Xueren provided the schematic of the ACT blanket along with the respective pressure drops. Then he presented the IB, OB-I and OB-II blanket definition and composition. He had also computed the primary pressure stresses on the three blanket regions. Only locally in the corners of the blanket modules did the stresses approach the design allowables. He also conducted the thermal stress calculations for the three types of blanket modules. The combined pressure and thermal stress remained below the allowable design stress.
Plate-Type Divertor 3-D Printing - Xueren Wang has been translating the plate-type divertor CAD models into computer files that can be used in a 3-D printer to better visualize the fabrication processes and determine how to improve the divertor design. Xueren summarized the current helium-cooled tungsten plate divertor design concept that is a leading candidate for the ARIES-ACT divertor. Xueren then described the fabrication processes that might be used to construct the divertor including the tungsten face sheets. The 3-D printed model was a half-size model constructed with a plastic resin fabricated in multiple successive layers. The inner cartridges were fabricated separately and could be inserted inside the plate assembly. Transition joints were also modeled. The only model fabrication cost for the project was the material. It was printed (constructed) in 51 hours.
Update of Vacuum Vessel Analysis - Farrokh Najmabadi began by summarizing the design and operational aspects of the ARIES-AT Vacuum Vessel. In addition to being the primary vacuum vessel, it also provides shielding with water and WC inside a double-walled vessel that was 40-cm thick on the inboard and top/bottom. (Additional information - The ARIES-AT VV spool walls were 2-cm thick. The vessel thickness was 25 cm on the outboard doors and frames with 3-cm thick walls. The ports were 10 cm thick with 2 cm thick walls. LMW). The ARIES-AT VV operated with 50°C water that may absorb tritium. Also during an accident scenario, the power core afterheat would raise the water temperature above boiling and induce steam inside the vessel.
The new single-walled vacuum vessel will use helium cooling at higher temperature of 300-500°C, depending on the tritium permeation and inventory requirements. The single wall is comprised of two thin sheets separated with internal ribs. The helium coolant passes between the two sheets.
The initial design assumes there is no inner door on the maintenance ports. Therefore the vacuum vessel is an integral vessel of the inner spool and all the ports, with the port door on the outer perimeter of the ports. Shielding of the sides of the outer TF coil legs by only the port walls remains a concern.
The scoping analysis initially used only atmospheric and gravity loads. Presently, the inner power core is only supported on the bottom of the vacuum vessel floor and no disruption loads are transmitted to the vacuum vessel. The initial stress analysis assumed a solid 10-cm thick vacuum vessel, which had acceptable stresses, but the design was probably too massive. There were a few stress concentrations with the flat port walls being a difficult area. Cooling of this design had not been addressed.
The next step in the analysis is to consider a double-walled structure with internal ribs. This would allow helium cooling of the structure. The thickness of the walls and spacing are TBD, but an initial trial run used two 3-cm thick walls and a 2-cm thick inner space containing ribs. This resulted in a substantially higher stress on the inner plate. Therefore the next trial was a 4-cm thick inner plate (wall) and a 2-cm thick outer plate (wall), still with a 2-cm spacing for a total wall thickness of 8 cm. These changes for the 8-cm thick ribbed wall reduced the stresses close to the solid 10-cm wall case. The optimization process will continue to refine the VV wall design details.
There was some question if the existing pressure vessel codes are proper and if there are any applicable vacuum vessel codes that are available - Tom Weaver accepted the action item.
Manifolding and MHD Issues - Mark Tillack reviewed his action items from the January project meeting, namely providing further guidance on MHD issues, temperature boundary conditions for stress analysis and initiating DCLL blanket analysis. He added the definition of ACT-1 flow loop parameters, lower manifold design definition and assessment of "flow balancing".
Mark felt they have minimized the 3-D MHD effects throughout the SCLL blanket except for the inlet FW manifolds in the lower power core. He and Xueren have routed the inlet and outlet pipes to regions that have reasonably low PF strengths, therefore low MHD effects. Central manifolds use successive (equal and symmetrical) splits with toroidal displacements to keep flow channels equivalent. On the other hand, FW manifolds remain a problem and electrodes may be needed.
Mark mentioned the use of "flow balancing", that is, a self-correcting process in which higher velocity streams in parallel paths induce currents that helps to increase the velocity in the slower flow channels, thus inherently balancing the channels.
Mark showed a flow diagram that illustrated the pressures and pressure drops around the SCLL inboard blanket and heat exchanger loop. There are some semi-empirical formulations of the 3D MHD effects on simple geometric cross-sections that were considered in the design of the ARIES blanket and coolant loop design. Mark presented the final design of the SCLL blanket designs for the inboard, outboard I and outboard II modules.
Mark then began to show the analytical approach he used for the thermal performance and MHD power losses in the blankets and the coolant loop. He defined the approach for the variable flows in the blankets. He also analyzed the MHD losses in curved pipes in a magnetic field. This allowed him to plot the radial and axial temperature profiles in the blanket (figures on pages 14 and 15 are unlabeled but the abscissas are intended to be temperatures in °C). Mark determined the HX temperatures on the primary and secondary sides. He presented the helium and LiPb power flows and temperatures in the power conversion cycle, which yielded an efficiency of 58%. Mark showed the thermal conversion gain of greater than 1% by increasing the turbine inlet temperature from 1000°C (nominal for the example on the previous page) to 1050°C.
Mark's next steps will be to update the ACT-1 MHD, thermal-hydraulic and geometric parameters for the updated strawman for ACT-1, as well as the other design cases provided for assessment.
The Planning of Reliability-Centered Maintenance Programs for Nuclear Fusion Power Plant - Tom Weaver summarized his objectives to define the RAMI plan such that the DEMO plant can demonstrate a sufficiently high plant availability to proceed to build a commercial plant with acceptable availability risk. This would translate into a 50% or greater plant availability for the (U.S.) DEMO. Achievement of this availability level in DEMO will require quantified levels of reliability in component design, RAMI design integration into the power core, maintainability and system monitoring/inspectability of all components, subsystems and systems.
Tom outlined an integrated decision evaluation and analysis system flow diagram for reliability assessment. This generic tool is being tailored for the fusion application. Part of this process involves methods to handle large technology leaps from the present state of knowledge to that required for DEMO. All this methodology will be implemented into a Reliability Estimating Tool to not only estimate reliability at all levels of the plant, but also to help prepare availability estimates, sensitivity analyses and guidance for the supporting experiments and the future facilities.
Tom and I-Li Lu are modeling the flat plate divertor design being developed by Mark Tillack and Xueren Wang, who are providing updated design details and critical performance data and likely failure modes.
Tom explained the life history of operational parts, which may experience premature failures, normal operational lifetimes and extended life. The key is to predict failure modes for premature failures so that the normal operational lifetimes can be achieved or even extended. The optimum scenario would be to correctly predict incipient failure and plan maintenance actions accordingly. There are statistical methods that may be used to support this assessment.
Update on Structural Analyses: ELMS, Thermal Creep and ITER Seismic Analysis - Jake Blanchard summarized that he is analyzing thermal conditions on the tungsten-covered, helium-cooled flat plate divertor. One of the major thermal conditions of interest will be the ELM events with cycle times from 1 to 100 ms and heating levels varying +/- 20%. Jake started with a nominal heat flux with a +/- 20% variation and a pulse time of 1 ms. He noted that the surface temperature varies ~5°C, but no thermal variation is observed at 1/4 of the tile depth. At a pulse time of 10 ms, the surface temperature varies ~11°C, but no thermal variation is observed at the tile base. At a pulse time of 100 ms, the surface temperature varies ~35°C and a small thermal variation is observed at the tile base.
Low cycle fatigue data for tungsten is documented in the ITER handbook, but the high cycle fatigue data are lacking. Some data are available for other tungsten alloys (without irradiation).
There is some data on thermal creep of tungsten at elevated temperatures. Jake is adding the power law creep model to ANSYS to evaluate the creep behavior. He has observed significant relaxation of the thermal stresses due to thermal creep even after a few days of operation. After prolonged creep, significant deformation may occur. Cracking may occur at cool-down. Jake cautioned that the assumed creep strain rate may not be applicable in this application.
In summary, the primary stresses are from the coolant pressure. Thermal stresses are probably relaxed due to thermal creep, but this might cause stresses and cracking at shutdown. These creep rates are strongly influenced by stress and temperature, so small changes may have dramatic effects.
MELCOR LOFA Analyses - Paul Humrickhouse stated that the MELCOR input deck has been modified to reflect the latest SiC blanket and power core configuration. The vacuum vessel is now a thin, high temperature (500°C) He-cooled structure that externally radiates to the room-temperature water-cooled shield. The exterior surface of the LT shield has a layer of super insulation, thus it is considered to be an adiabatic boundary. Air injection is not presently considered for emergency heat removal.
Paul commented that the simulation of dual coolants of LiPb and water is a difficulty in the MELCOR modeling. Paul models the LiPb primarily with the code. The water mass flow rate is taken from ITER calculations for natural convection heat removal (for accident conditions).
Paul illustrated both modules for both the previous and the current model. He presented the results for the Loss Of Flow Accident (LOFA) for all power core coolants. The code predicts that significant natural circulation in the LiPb redistributes heat between the inboard and outboard. The vacuum vessel initially is at 510°C, increased to a peak temperature of around 625°C and over a period of 2 days, falls back to the 500°C region and continues to decline at a slow rate. (Note that the graphs are labeled in Kelvin.)Likewise, the water-cooled shield starts out around 70°C, climbs to 120°C and then steadily declines. The First Wall temperatures quickly spike and then quickly decline. The upper divertor rises slightly and then continues to fall (the lower divertor is similar). The IB HTS rises for an hour or so and then falls, while the OB HTS closely tracks the temperature of the upper divertor (OB HTS line is hidden beneath the upper divertor line, LMW).
Paul emphasized that he needs better data on the decay heat in the power core components.
Laila El-Guebaly suggested that the worst case accident scenario is not a LOFA for all power core coolants; rather it is a LOFA for the LiPb and a LOCA for the other coolants. Paul agreed to model this new scenario.
Paul reported that the MELCOR model is now in place and functional. The next step can increase the fidelity of the safety modeling by defining the decay heats for all power core components, the material composition of all components, specific build details and dimensions and nominal operating temperatures of coolants and structures. Laila agreed to provide the decay heats from the 1-D activation analysis for the reference radial builds.
Updated Thermal Performance of Finger-Type Divertors - Minami Yoda presented an update on the Georgia Tech group's ongoing studies on the thermal performance of helium-cooled finger-type divertors with the objective of developing generalized design charts giving maximum incident heat flux and coolant pumping power. In these studies, the results from dynamically similar experiments using air, helium and argon with a finger-type divertor cooled by a single central impinging round jet and the same design with an array of cylindrical pin-fins are extrapolated to prototypical conditions.
The results for the finger-type divertor show that the fraction of the incident heat removed by the coolant at the surface where the jet impinges varies with coolant and that this fraction, characterized by the ratio of the coolant and divertor wall thermal conductivities, must be considered. The helium and argon results, along with earlier results for air, have been used to determine how the non-dimensional heat transfer coefficient, or Nusselt number, depends upon the non-dimensional coolant mass flow rate, or Reynolds number, and this thermal conductivity ratio. The results for the finger-type divertor with fins, however, suggest that the Nusselt number depends only upon the Reynolds number.
The results for both types of finger-type divertors were used to develop correlations for the Nusselt number over a wide range of Reynolds numbers that span the prototypical value. Correlations were also developed for the pumping power loss coefficients based on experimental measurements of the pressure drop across the divertor as a function of Reynolds number.
The maximum heat flux that can be accommodated by the finger-type divertor at prototypical conditions was found to be about 17 MW/m2 for a maximum pressure boundary temperature of 1200°, assuming circular tiles 12 mm in diameter, or a tile area of 1.1 cm2. This value is lower than that reported previously by the GT group because of the effects of the thermal conductivity ratio. The maximum heat flux that can be accommodated by the finger-type divertor with fins at prototypical conditions for the same maximum temperature is about 22 MW/m2, but the pumping power will then increase by about 18% because of the additional pressure drop due to the coolant flowing through the fin array.
ARIES Project Planning
Gantt Chart for ARIES ACT Project - Farrokh Najmabadi asked Mark Tillack to prepare a Gantt chart (annotated schedule) for the remainder of the year to define the key actions to be accomplished by the end of CY2012. This schedule can be found on the ARIES web page, May 31-June 1 Meeting Agenda, last item, Near Term Planning (xls).
Mark first defined the ARIES ACT designs envisioned to be at the four corners of plasma and engineering design space. Namely these are identified as:
Mark thought the ACT-1 design could completed with baseline physics parameters and design details by the end of July 2012, so the technical details could be finalized for the TOFE meeting in late August. Then the attention could be turned to the less aggressive physics and engineering of the DCLL design (ACT-4), with the technical details finalized by the end of November 2012. Since the other two designs, ACT-2 and ACT-3 are parametric designs incorporating the significant features of the other two design concepts; these data could be worked in parallel with the ACT-4, with parametric results by the end of the CY12. Final physics and engineering design definition would be forthcoming by second quarter of CY13.
General Project Goals - Farrokh Najmabadi emphasized the need to continue making significant progress on defining the ACT-1 and ACT-4 physics technical bases and the power core component and integrated designs so the plant designs and features can be finalized for the key publication dates. Chuck Kessel recommended the project adopt the detached divertor concept (a key enabling technology). We need afterheat data and the neutronic analysis of the DCLL design so all supporting technical efforts can proceed. The VV design and analysis needs to be completed prior to the TOFE preparations. This includes the disruption loads and analyses on the VV.
Farrokh Najmabadi said the intent is for the project (members) to publish a set of technical journal papers on the ARIES-ACT designs by November-December 2012 timeframe. There will also be an annual project report of this work provided to FES at the end of the CY12 period.
The meeting Action Item list can be found at this link, 31 May - 1 June 2012 Action Item List. Note that there are two sheets, one in number order and the second in Due Date order.