ARIES Project Meeting Minutes
22-23 January 2013
UCSD, San Diego, CA
Documented by L. Waganer
Ref: Agenda and Presentation Links: Meeting Agenda
Administrative and General Information
Welcome/Agenda - Mark Tillack provided refreshments for the ARIES team for both days. Les Waganer summarized the day and a half agenda (see above link) with the topics to be presented.
Next Project Meeting and Call - The next ARIES conference call will be February 19 at the usual time of 9:30 PST. The next meeting will be held the latter part of May 2013. [Mark Tillack sent out a Doodle query on the available dates and many have responded.
Review of ARIES-ACT Project Status and Objectives - Farrokh Najmabadi asked the team to send him by 1 February the titles, authors, and general content of the technical papers that will comprise the ACT-1 final report. Farrokh intends to have all first drafts of the papers by 1 March and the final drafts by 1 April at which time the drafts can be provided to the Fusion Science Technology Journal editor.
The second detailed ACT study, ACT-2 should be finished and documented by the end of 2013. The other two ACT studies, ACT-3 and ACT-4 are to be parametric studies and they also will be completed by the end of 2013.
ARIES Task Results
Update of the ARIES ACT-1 Systems Analysis - Mark noted several significant changes that have been implemented in the ARIES Systems Code (ASC) since the September meeting. These changes included the power transition width from the divertor, revised radial builds from Laila El-Guebaly, revised edge radiation fractions and cost account corrections per Les Waganer's September presentation (corrections of this ASC subject remains an ongoing process). Mark Tillack has streamlined the production run process of the ASC thus allowing 412,000 runs overnight on the PPPL computers to update the current project database. From this large database of possible physics solutions, these results are filtered to determine viability and then subjected to engineering and economic filtering criteria. This results in 43 remaining points within 5% of the minimum COE.
Mark showed a table of seven representative 1000 MWe plant cases. The team recommended focusing on the two cases with a major radius of 6.25 m as this is the currently documented radius in the power core CAD definition and radial builds. These cases have heat fluxes on the divertor of about 13 MW/m2. Laila suggested comparing the ARIES-ACT COE to that of advanced fission systems and identify the differences.
ACT-1 Physics Results - Chuck Kessel presented several current ACT-1 physics results, including five 1.5D, time-dependent free-boundary plasma configurations depending on the pressure and density conditions. These cases relaxed at a ßNtotal around 5.72 when nearing steady-state conditions.
Chuck provided a table of 17 cases near the baseline operating point that may be suitable for lower stability profiles. In pursuing the ideal MHD stability condition, he determined the broad pressure case at ßNtotal ~ 5.72 is a high-n (density) ballooning stable case. Other cases will be analyzed later.
The PF coil solution is evolving with iterations between Wang and Kessel. An additional outboard coil was required as compared to the ARIES-AT solution.
Chuck outlined the general lower hybrid (LH) heating and current drive (CD) parameters proposed. He feels launching above the midplane provides improved CD and deeper deposition in the plasma (60 deg location is shown in his slide). He mentioned the launcher was only located in one sector, but a better solution might be to halve the launcher size (and power) and install two launchers in two sectors. Mark noted that it would not be feasible to have multiple LH launchers above each other in a single sector as cooling the blankets between multiple launchers could not be easily accomplished.
Chuck showed a fast particle MHD plasma parameter space diagram that agreed with the NOVA-K normalized 1.5D model, if radiative damping is not considered. The latter damping analysis may shift the loss diagram to allow a larger operating space without any significant α-particle loss. Since the 1.5D model is conservative, this is good news for ACT-1. Future analyses will be done.
Fast particle MHD stabilities are aggravated by high q and high Ti(0). Also the operating point has a higher major radius to limit the heat load in the divertor and the plasma density has been limited to ≤ 1.0 x nGr, where nGr =Ip/(πa2). Normally, reaching or exceeding nGr leads to plasma problems, however these may be controlled by fueling techniques or SOL modifications. It is advantageous in power plants to have n ≥ nGr, because the higher density is needed for higher fusion power.
Chuck mentioned the content of the TOFE paper that addressed steady-state heat loads, transient ELMs and off-normal disruptions. These results will be implemented in the ACT-1 design.
Ion Cyclotron (IC) and Electron Cyclotron (EC) Heating/Current Drive - Francesca Poli discussed the ICRH power level needed to drive a 350-kA current in the flat-top (steady-state). She showed that 10 MW of IC CD at f > 90 MHz can drive approximately 400 kA of current (TORIC full wave analysis was used). She also found that the 92-94 MHz IC frequency could provide sufficient ion heating during plasma start up and electron heating in the flat top period.
Francesca investigated electron cyclotron heating and current drive at 170 GHz (same as ITER) at several injection angles into the plasma to assess angular orientation of the antennas. The mid-plane orientation had good accessibility to the core plasma with the maximum toroidal steering of 32 degrees. The upper launcher positions offered good q-profile control and good CD efficiency.
In the future, she will use the EC profiles in the TSC simulation to assess the steering efficiencies for FW, EC and ECCD for both on-axis and off-axis positions.
Improved models for radiating edge plasmas for ACT-1 - Part 1. Kinetic Monte Carlo neutrals for pumping - Marv Rensink noted that he and Tom Rognlien had previously reported that the neutral pressure near the private-flux pumping region for fully-detached and partially-detached solutions would be ~100 times that of ITER's simulations. Now, using ITER parameters, they still found much larger pressures.
Marv noted that the UEDGE code models hydrogenic plasmas, impurities and neutrals. He has shown that partially-detached ACT-1 plasmas have very high neutral pressures in the PF region, near the wall, either with the ACT-1 or ITER UEDGE modeling or the ITER modeling. He said the UEDGE code analyzes fluid neutrals, whereas the more time- and computer-intensive ITER code analyzes kinetic neutrals, with many significant differences. He continued with neutral pressure diagrams of both model types. However, he noted the location and strength of the ionization source is not sensitive to the model used.
In summary, he said the fluid neutrals model overestimates the neutral pressure at the PF wall. Both models are consistent in the high heat flux region of the divertor. The PF region is decoupled from the strongly radiating region, which could be modeled with the kinetic model (smaller effort). The ACT-1 PF wall neutral pressure may be similar to ITER results or approximately 10 Pa at pump duct entrance.
Part 2. Multi-charge-stated impurities - Tom Rognlien noted that previous simulations of the UEDGE ACT-1 plasma used a basic coronal equilibrium impurity model with a specified impurity concentration. The more recent helium simulations with a highly-detailed multi-charge state (MCS) model shows strong variations in concentration. What effect does MCS have for neon or carbon radiation? Both models use the same electron line-radiation loss equation, d (3neTe/2) / dt + transport terms = - ε (Tene) ne nimp, where ε (Tene) is the impurity emissivity. In the first case, the fixed fraction model specifies nimp/ne and obtains ε from a coronal equilibrium model. This model executes very quickly. On the other hand, the MCS (multi-charge-state) model follows individual charge states and computes nimp/ne locally. These two models were compared graphically with contours of impurity radiation power and spatial variations for power and electron temperature.
In summary, for neutrals, the kinetic model is preferred to correctly model the neutral-dominated pumping region; however the plasma-dominated region can be adequately modeled with the UEDGE fluid model. For the impurity analysis, Tom mentioned the more detailed MCS models show strong variations in concentration, but radiation levels are adequate for Zeff ~ 2. Slowly evolving fully-detached plasmas are predicted to be capable of being feedback stabilized.
Nuclear Fusion Power Plant Reliability Processes and Application to ACT Divertor - Ali Yousefiani, a Boeing Research and Development Technical Fellow, discussed the general damage mechanisms that could cause environmental durability challenges in a MFE power reactor divertor system, primarily in the environmental and material degradation, stress/strain induced thermo-mechanical failure and synergistic damage phenomena, all with subcategories. Ali quickly summarized the He-cooled, W-armored ACT divertor current requirements, materials and fabrication processes.
He examined, in more detail, the current fabrication processes, developing advanced fabrication processes, and remaining challenges. He used the same technique to address machining, forming and joining processes. He then addressed NDE (non-destructive evaluation) considerations, requirements, NDE candidates, and suitable inspection component regions. He also discussed the currently utilized ultrasonic inspection and eddy current NDE methods. This NDE technology field is rapidly improving with advanced hardware and processes.
In summary, Ali offered a host of recommendations including fabrication automation with graded transitions from structure to armor materials, new forming and brazing methods and improved NDE methods and processes.
Correlations for the Plate Divertor and Update on the GT Helium Test Facility - Minami Yoda reiterated GT's research goals of evaluating the thermal performance of gas-cooled high heat flux divertor designs, providing design guidance and thermal data correlations and parametrically determining changes due to material choices and temperature limits. These goals were approached by conducting experiments spanning non-dimensional parameters for prototypical divertor conditions. To satisfy the design support, they developed power-law correlations for Nusselt numbers and thermal conductivities.
The primary divertor approach experimentally analyzed by GT was the plate-type, helium-cooled divertor that is intended for heat fluxes up to 10 MW/m2. Experiments for both HEMP-like and HEMJ-like divertor modules have been performed with air, helium and argon at the similar Reynolds numbers. Obviously, the data and simulations display variations based upon coolant and divertor materials, therefore the objective is to improve the thermal performance correlations for the plate divertor.
GT has developed a 3D half model (FEM) of the HCFP model with ANSYS Workbench and simulated the experiment with ANSYS Fluent 14. The simulations used air, He and Ar that are cooling brass and WL10 tungsten alloy. This has resulted in Nusselt number results for these numerical simulations. GT used a multi-linear regression to fit the numerical predictions that significantly reduces the maximum heat flux for the plate type divertor (from that previously predicted.)
The GT He loop has been employed to evaluate various divertor designs at prototypical performance conditions. Minami showed the He loop schematic and an actual picture of the loop. This loop will begin experiments on the HEMJ-like test section in February 2013 with a brass test section and then the tungsten section. The intent is to test up to a heat flux of 6 MW/m2.
Update of Heating of the ARIES Divertor - Jake Blanchard illustrated the plate and finger divertor designs to help frame his goal of parametrically analyzing the effects of ELM heating on the ACT divertors. He then clarified the properties of both the small and large ELMs. The "Old Small ELM" heating analysis of the divertor indicated a 20-μm melt depth. The new parametric analysis of single, 5 Hz and 10 Hz ELMs over a flux range of 2 to 10 MW/m2 would just achieve melting at an average ELM heat flux over a range from 400 MW/m2 to about 1050 MW/m2.
Jake showed the simplified finger ELM thermal model for a 10 MW/m2 heat flux reaching a peak surface temperature of 2124°C on the outer corners of the finger tiles. He then compared the plate and the finger designs with identical thermal loading and found similar maximum temperatures.
He did a similar parametric thermal analysis on the finger design and then compared the finger to the plate divertor elements. The finger performed slightly better (allowing higher heat fluxes before melting).
Vacuum Vessel Disruption Analysis - Jake Blanchard outlined the modeling of plasma currents and magnetic fields that will be transferred to the vacuum vessel during a plasma disruption.
Updates on ACT-1 Power Core Configuration and System Integration - Xueren Wang listed his tasks/assignments from the last project meeting. He tackled the first assignment, which was developing the layout of the lower vacuum ducts. He showed the ITER Torus Vacuum pumping system for inspiration. He then displayed his layout of the discrete ACT-1 vacuum pumping ducts, option 1, that relocated the PF coil sets, integrated the inboard and outboard blanket manifold, thickened the shield blocks and redefined the cryostat and coolant ring headers.
Option 2 has no discrete pumping ducts, but rather relies on the open space behind the structural ring and the vacuum vessel maintenance ports for an integrated pumping region. The vacuum is pumped from the maintenance ports into ring headers above and below the maintenance ports. A question was raised about the retained tritium inventory in the vacuum vessel and maintenance ports when this method is used (see related action item for Paul Humrickhouse). Laila raised the issue of tritium losses and impact on the TBR requirement.
Xueren showed the ITER LH launcher design to deliver 20 MW of power per port. Xueren followed with a preliminary design (and dimensions) for the ACT LH launcher design including the routing of the blanket coolant. [See an action item for Xueren related to resizing launchers.] Two power core sectors will house the two LH launchers, diametrically opposite of each other. He further illustrated how the blanket plumbing might be arranged. It was noted that there does not seem to be a method to accommodate a second launcher directly above the first launcher. Xueren provided additional schematics of the blanket coolant flows in the LH launcher sector.
Xueren presented a pictorial of a method to electrically connect/disconnect the conductive vertical stability coils that are located immediately outside of and structurally connected to the Outboard Blanket II. This could be accomplished with a tapered W or WC shield block with conformal conducting material on the tapered surface. This conducting, shielding block can be held in place with a bolted joint that also provides structural support.
ARIES-ACT 1 SCLL Blanket and Manifolding - Christina Koehly explained her objectives to help develop a more detailed design of the ACT-1 blanket and manifolding/flow distribution system for the SCLL design. She is to concentrate on the fabrication, manufacturing, disassembly and reassembly of the blankets, manifolds and piping. Siegfried Malang, Xueren Wang, and Mark Tillack provided preliminary designs for the blankets and manifolding.
Christina has determined that manufacturing of single piece, concentric SiCf/SiC pipes is possible with today's technology. At the end of the presentation, Christina showed a SiCf/SiC composite woven structure for a cooled rocket nozzle, however the size and development status were not provided. Manufacturing separate inner and outer parts is easier and more affordable, but assembly would be difficult and small leakage from pipe to pipe may occur. Cutting and joining technologies inside the power core are under development and not validated [probably not with SiC].
Christina illustrated several candidate methods to manifold designs for consideration. Most of the design options had serious flaws and were not recommended - designs V3 and V4 seemed to be most promising. The team made several suggestions for improvements.
3D Finite Element Analysis for a Ribbed Structure Vacuum Vessel - Hamed Hosseini outlined the present vacuum vessel (VV) design, with maintenance ports. He is analyzing the stresses in the VV structure due to pressure and gravity loads, with the goal of reducing the thickness of the structure. The VV is constructed of 316 SS with a yield stress of 140 MPa and a working temperature of 550 K (277°C).
Hamed provided the analysis boundary and loading conditions. His first analysis was for a thin (5 cm) single walled VV that resulted in many areas in excess of the yield stress. Increasing the single wall thickness to 10 cm did reduce the maximum stresses below the yield stress level; however a double-walled design is preferable to allow cooling of the walls.
Hamed then showed typical double-walled ribbed structures and sandwich panels. He discussed several design options, such as direction of ribs, total thickness of panel, wall thicknesses and distance between ribs. Hamed provided analysis results for each of these options with recommendations. He also addressed the transition of rib direction at the sides and top of the ports along with the termination between the sides and top. He then presented his recommended final port design parameters followed by the general configuration of, and stresses for, the whole vacuum vessel.
Thermal Analysis of Various LOCA/LOFA Scenarios for ARIES-ACT-1 - Carl Martin explained the scenario for the worst case accident, namely the loss of coolant (LOCA) for both water and helium combined with the loss of flow (LOFA) for the LiPb coolant. The superinsulation (multi-layer insulation (MLI)) on the exterior of the inboard low-temperature shield will limit the heat removal following the accident. The maintenance ports are assumed to be the primary means for decay heat removal, whereas the inboard region will present a problem. [PS. Siegfried Malang noted after the meeting that he remembered a technical paper he coauthored entitled "Passive system for cooling the inboard region in the case of a severe accident", Fusion Engineering and Design 63_/64 (2002) 251-256 that may be applicable in this case.]
Laila El-Guebaly provided the outboard and inboard radial builds and decay heats for a period of time after the accident. The maintenance ports and port shield plug also will have decay heat. Carl constructed a 2D ANSYS model and a set of assumptions for the LOCA/LOFA. The analysis assumed the plasma heating would continue for an additional three seconds beyond the LOCA/LOFA. Les Waganer thought that the 10th of a kind reactor would have sufficient instrumentation to commence an immediate plasma shutdown, subject to the necessary sequence of shutdown events. With his assumptions, Carl determined that the temperatures of the steel structures in both the inboard and outboard regions after the LOCA/LOFA exceeded their reuse limits of 750°. Carl approximated the cross-channel radiation (inboard first wall to outboard first wall) that may help cool the inboard region. The surface emissivities of the radiating and receiving surfaces may be significant in the cooling effect, but determining their likely values (ranging form 02 to 0.9) may be hard to predict.
Future work will include adding the divertors, designing the WC-free inboard LT shield, investigating other coolant avenues, and predicting when water flow will be necessary to preclude structural damage.
Progress on Low Activation Structural Steels - Arthur Rowcliffe discussed the "new" nano-structured ferritic steel alloys (NFAs) that have nanograins and nanoclusters. These materials have very large interfacial areas that enhance recombination of point defects as well as accommodate helium atoms or gases when irradiated. In the ferritic phase, the nanograins and nanoclusters are exceptionally stable up to very high temperatures (good to increase design margins and component reuse after LOCA/LOFA accidents). These NFAs have outstanding strength and creep properties. This class of materials was patented in 1978, rediscovered in 1999 and has spawned newer alloy developments from 2001 to the present. These alloys are not yet commercially available and only exist in small quantities. Fracture toughness of NFA 14YWT challenges the high strength, low toughness paradigm for structural alloys. Arthur showed HIFR irradiation data for two representative alloys. With limited testing, no DBTT shift has been observed. Materials of this type were patented by International Nickel in 1978 and the production of ferritic alloys such as MA956 and MA957 were successfully scaled-up to ~3000 lb quantities for aerospace applications.
Following the discovery in 1998 of nano-scale solute clusters in experimental alloy 12YWT, the application of advanced characterization techniques revealed that the high strength of the commercial alloy MA957 was also related to a very high number density of similar nano-features. The mechanical behavior of alloys such as NFA 14YWT challenges the high strength/low fracture toughness paradigm for structural alloys. An NFA variant using 9Cr promises improved high temperature fracture resistance. Creep performance of these NFAs are also very much improved over the existing 9Cr-WMoVNb steel.
[It should be noted that these data reflect current developments and much improved material advances are likely before the advent of the future power plants.]
ACT-1 LOFA model updates and LOCA analyses - Paul Humrickhouse explained the general updates to his MELCOR safety analysis model since May 2012. Specifically, the coolant loops have been reconfigured, water pressures in the low temperature shield were reduced and corrections to the steel structural ring have been incorporated. Paul presented the current version of the heat transfer schematic for the ACT power core.
Paul summarized his updated LOFA analysis. At TOFE, he explained that there were no problems in removing the decay heat during a LOFA and the structural temperatures were reasonable, except in the divertor where the temperature reaches 2500°C and may melt the ODS FS structure. The water in the loop did boil, but the heat was transferred to the roof HX via water convection. The HX on the roof has been modified to remove additional heat. A graph of structural temperatures over a period of days was presented. The fluctuating mass flow rate on the right of the slide is due to flow variations caused by the behavior of the boiling liquid.
The helium LOCA analysis now assumes a double guillotine break in the outlet helium piping. Paul is concerned about an overpressure in the cryostat. The team felt this would not be a problem as a previous analysis had been done for the smaller vacuum vessel for a prior ARIES study. Paul had assumed the ex-vessel (power core) helium volume was three times that of the in-vessel volume. Les Waganer corrected Paul as the ARIES project typically assumes the ex-vessel volume is 1.5 times the in-vessel volume in lieu of a detailed main heat transfer and transport system. This change in volume assumption may account for the predicted overpressure.
Paul is planning to accomplish a water LOCA analysis in the new future. There was a question about the heat transfer in the overheated MLI and he offered to do a literature search to determine if the MLI heat transfer properties are degraded when exposed to excess temperatures. Slight changes in the radial builds will be revised. Laila suggested considering the 3-second plasma heating after the initiation of any accident. Further, Paul will document the three accident cases in the final report.
Additional ACT-2 System Scans - Chuck Kessel provided additional data for the ACT-2 power plant with conservative physics and the DCLL blanket system. Chuck presented the physics and engineering parameters he used to prepare and filter the generated data point sets. He then reexamined the solutions with higher n/nGr and q95 parameters and generated a table of candidate sets of data. Major radii ranged from 9.5 m to 11.0 m with BT fields of 7.5 T to 8.5T. This set of cases is improved with respect to those generated in September 2012.
Chuck noted he wants to further examine the parameter space, including COE "optimization" and compare the ACT results with EU power plant (not Demo) parameters. He also wants to implement a TSC run on one case to examine the viability of the plasma configuration including the heating and current drive parameters. He intends also to identify stable profile configurations with no stabilizing wall.
After Chuck's presentation, Les Waganer asked if the team thought it appropriate to baseline the ODS FS material for the DCLL blanket design and the more advanced nano-structured ferritic steel alloys (NFAs) for structural elements in the SCLL blanket and divertor design. This prompted an action item on this issue.
Preliminary TBR Analysis and Radial/Vertical Build Definition for ARIES ACT-2/3 (DCLL) Design - Laila El-Guebaly identified changes relative to her July 2011 presentation, namely 3D TBR modeling of the OB blanket, a thinner 45-cm thick IB blanket, and updated OB penetrations.
Laila reviewed the results of their new 3D TBR analysis that starts with the UCSD FW and divertor shape and blanket details, and then builds the CAD model for the neutronics analysis. The TBR analysis then couples the UW CAD model with the MCNP code using the DAGMC code. She then provided the predicted breeding results beginning with an infinite cylinder and then added, stepwise, design elements to the final 3D design concept, that resulted in a TBR of 1.05 using an 85% 6Li enrichment. The similar ACT-3 power core achieves the same TBR with higher 6Li enrichment.
In summary, she found that limiting the blanket coverage radially and vertically will substantially impact the TBR by 20%. Adding FS structure and flow channel inserts significantly reduces the TBR by 23%. The use of stabilizing shells will decrease the TBR by 2-3%. Adding penetrations and assembly gaps will have a 2% negative change in the TBR. She recommended thickening the IB blanket by 20 cm to allow a reduction in lithium enrichment. For the moment, she is using a placeholder penetration area of 7 m2 in the outboard region. Laila provided a new set of ACT-2/3 NWLs, radial builds and component compositions needed for the CAD drawings and economic analysis.
Plasma Steady-State Solutions - Holger St. John summarized the recent progress in his self-consistent modeling of the ACT-1 plasma scenario. A first iteration for the new 6.8-m major radius equilibrium to steady state has been completed. While previous simulations required a lower hybrid current to yield near-zero ohmic current, the new simulation, with a corrected initial MHD equilibrium, requires relatively little RF current drive. Neoclassical confinement predicted from the multimode model leads to electron and ion temperature profiles that yield high bootstrap current. The assumed density profiles further enhance this effect. The evolved system is substantially changed from the initial equilibrium so that further iterations, including re-establishment of force balance, are required to ensure that a self-consistent steady state scenario is established.
Jan 22-23 ARIES Project Meeting Action Items