ARIES Project Meeting Minutes
21-22 May 2013
Hampton Inn, Germantown, MD
Documented by L. Waganer
Ref: Agenda and Presentation Links: Meeting Agenda
Administrative and General Information
Welcome/Agenda - - Les Waganer thanked Al Opdenaker for the meeting room and the hotel arrangements. Al and his staff also arranged for a light breakfast, morning snack, lunch, and an afternoon snack. Les Waganer summarized the day and a half agenda (see above link) with the topics to be presented. [The meeting was concluded by the end of the first day and the next morning was provided for individual interaction.]
Next Project Meeting and Call - The next ARIES conference call will be June 18th at the usual time of 9:30 PST.
The next meeting will be held sometime in the last two weeks of September 2013. [Les Waganer sent out a Doodle query on the available dates, Doodle Site ]
Review of ARIES ACT Project Status and Objectives - Farrokh Najmabadi noted that most of the ACT-1 technical papers are in hand. Those remaining uncompleted papers have been identified and have an acceptable completion date. After all papers are received and approved, Farrokh will forward them to Nermin Uckan for publishing in a special issue of Fusion Science and Technology Journal.
The intent is to complete and document the second, less detailed ACT-2 study by the end of 2013. The other two ACT studies, ACT-3 and ACT-4, are parametric studies and they will be completed and documented by the end of 2013.
ARIES ACT Task Results
Comparison of ACT and AT Cost Modeling - Les Waganer summarized the ARIES Systems Code economic subroutines – they have been updated to reflect the new costing algorithms recommended by Les Waganer and approved by the ARIES Team. Some coding errors were found by Les and Mark and those have been corrected by Mark Tillack..
On or about 22 February 2013, Les Waganer was asked to compare the current economic results for the ARIES-ACT-1d (on the ARIES web site) with the previously published ARIES-AT cost results. The ACT-1d run of 26 July 2012 had a cost escalation error, so a corrected version was used. The detailed comparison results have been provided to Mark Tillack and Farrokh Najmabadi for inspection. The following presentation illustrates the highlights.
ACT-2 Systems and Physics Progress - Chuck Kessel began with a comparison of the ARIES-I operating point as well as the type of points we are seeing now for ACT-2 from the systems code. Observations are that the full amount of oxygen impurity could not be reached, only up to 0.7%, likely due to differences in radiation models. Overall, the point is reproduced reasonably well, with the standard approach to shaping approximations using values lower than the X-point values. The resulting energy confinement multiplier of H98 = 1.49 is outside of the range where we are looking in ACT-2, since we are presently restricting our search to H98 ≤ 1.3.
Examining the power terms and plant power balance, the fusion gain is 19.9 for ARIES-I and we are getting 17 with the present code, due to the lower stored energy and therefore lower fusion power (ARIES-I is 1925 MW, present systems code for ARIES-I is 1826 MW). The current drive efficiency assumed in ARIES-I was 0.32, and it was necessary to use 0.37 in the present systems code to get a similar CD power, likely due to using different reference points for the density in the efficiency expression. This current drive efficiency is not necessarily wrong for ICRF FW, but the configuration assumes one can distribute FWCD across the minor radius, based on ray-tracing calculations with multiple reflections off the separatrix. This approach is not accepted as viable today based on experiments, since this provides a strong power loss mechanism and also distorts wave properties at each reflection. Single pass approaches are ALWAYS preferred; however, this means we can only get ICRF FWCD at the plasma center. For our scanning to identify an operating point for ACT-2, we are using 0.15 for the CD efficiency based on LHCD, as was used in ACT-1 (since the CD was dominated by LH). We do not currently know what the CD distribution will be for an ideal MHD stable plasma for ACT-2.
The ARIES-I design used a SiC structural material, and assumed a thermal conversion efficiency of 0.49 (ACT-2 assumes 0.44 +/- factors on neutron wall load and surface heat flux with DCLL). The ARIES-I wall-plug efficiency for the H/CD systems was assumed to be 0.72 (ACT-2 is assuming 0.40). Pumping power is 54 MW which is about 3% of Pfusion (ACT-2 is assuming 4%) and all other subsystems are 52 MW (ACT-2 is assuming 4% of Pel,gross). The neutron multiplication for ARIES-I is 1.3, and ACT-2 is using 1.1. The values Chuck gives are what his engineering module is assuming. If the ARIES-I power terms are assumed with the ARIES-I reproduced physics operating point, the code gives 916 MW of net electric power, pretty close to the 1000 MW from ARIES-I report. So the present systems code can reproduce the operating point overall including the engineering features. The peak heat flux on the divertor from the ARIES-I report is 3.88 MW/m2, while our present definition gives 15.3 MW/m2, which is expected.
Finally the ARIES-I power balance assumptions were used with the ACT-2 database (the original one with large scan over R, Bt, etc). These assumptions are 49% thermal conversion efficiency, 72% H/CD wall plug efficiency, 3% of fusion power for pumping and 3% of gross electric for all other subsystems. Also the neutron multiplication is set to 1.3. When we apply the engineering assessment to the original ACT-2 physics database, we find that low R configurations are obtained, with R as low as 6.0 m. The filters used were electric power between 950 and 1050 MW, H98 < 1.6, and n/nGr < 1.6. Using the power balance assumptions we are using now for ACT-2, the lowest major radius solution is at 7.0 m. If we further restrict the allowed operating points with peak divertor heat flux 15 MW/m2, H98 < 1.3, and n/nGr < 1.3 (which are typical of those Mark and Chuck are using right now for ACT-2), then the ARIES-I power balance reaches R's as low as 7.0 m, while the ACT-2 power balance reaches R's as low as 8.5 m.
So it is clear that the power balance terms influence the major radius solution space. ARIES-I values and the ACT-2 values have been applied to the same physics database we constructed for ACT-2. ARIES-I assumptions are more optimistic, especially the H/CD wall plug efficiency. So in general, we should never see an ARIES-I like operating point within our scans, as it requires a series of parameter assumptions we are not making in ACT-2, e.g.,. H98 ~ 1.49, ηCD = 0.72, ηth= 0.49, Mn = 1.3, Ppump = 0.03 * Pfus, and Psub = 0.03 * Pel,gross. In addition, ARIES-I had a high peak divertor heat flux, which we would prefer to avoid. The high CD efficiency in the plasma, assumed in ARIES-I, allows higher Ip solutions, and so lower q95 values. It is not clear what the benefit of the very high B-field is, although we certainly see the higher BTmax values at the lower plasma major radii in our ACT-2 scans, so it probably contributes to smaller devices to produce the needed fusion power. Of course this magnet assumption (21 T at the magnet) in ARIES-I is totally unsupported from present experience.
Chuck and Mark scanned the extensive parameter space within the broad constraints of a net electric power around 1000 MWe , (betaNth + betaNfast) < 3.25, H98 ≤ 1.3, η/η Gr ≤ 1.3 and qdiv ≤ 15 MW/m2. These results generally had major radii, R, in the range of 8.5 to 9.0 meters.
Chuck also showed his preliminary ACT-2 work, assuming a R = 9.5 m plasma, an equilibria, ideal MHD stability, current drive and using the ARIES-I solution, and the pedestal pressure. Ideal MHD stable solutions are difficult to find with no stabilizing wall, due to the restriction of current profiles from bootstrap and external CD sources. The ARIES-I CD solution is considered unfeasible based on common experience on tokamak experiments.
Chuck's concluding remarks identified the subsections of the ACT-1 Physics document he is preparing.
System Analysis of ACT-2 Design Space - Mark Tillack (with support from Chuck Kessel) has analyzed many power plant design points, literally millions of point designs. Some of the code features were modified to accommodate the large number of design points and the low power density of ACT-2 conditions. Mark has incorporated the costing updates suggested by Les Waganer as well as selecting more desirable material choices. Some coding errors remain and those are being addressed.
Mark noted it was difficult to find a modest major radius machine with the other desired parameters. To examine more points, he increased the parameter search resolution that yielded more points in the major radius range of 8 to 9 meters. From the resultant 2.8 M points, he filtered those down to 2,006 points having the most attractive physics and engineering parameters. Within that range, the smaller machines require higher H98 factors and Greenwald factors ranging from 1.1 to 1.2. The COE increases with BT,max over a range of 70 to 95 mills/kWhr, but is invariant with respect to qdiv ob.
Mark picked a few representative design points over the range of major radii to illustrate how the significant physics and engineering parameters were influenced. He selected a few points with attractive COE values for more detailed examination.
Mark was interested in the intensity of the PF fields in the liquid metal flow regions (approximately 1.0 to 1.5 Tesla with moderate gradients (see his ACT-1 side view for a graphical representation)). This result was encouraging for the LM MHD effects.
Action item [Mark T]: Compare a 9-m ACT-2 case to a representative ACT-1 case. Presently it appears to have a COE difference of 10%, which seems too low.
Preliminary TBR/Shielding Analysis and Build Definition for ACT-2 - Laila El-Guebaly has been working on the TBR analysis to help determine the size and extent of the inboard and outboard DCLL blankets for ACT-2 and -3 along with the radial and vertical builds. Within the preliminary design envelope (final design details remain TBD), she has determined the TBR could be maintained with lower Li-6 enrichment, on the order of 70%. A new, wider OB divertor slot may reduce the TBR slightly, pending new design information from UCSD.
Laila presented the current ACT-2 and -3 engineering parameters she has been using to derive the peak NWLs over the IB and OB first walls and divertor surfaces. She then illustrated the OB, IB and vertical radial builds with salient features of each. She provided the material compositions of all of the power core elements. These data will be updated when the ACT-2 and -3 design points are selected.
Preliminary ACT-2 Power Core Design Definition - Xueren Wang explained most of the ACT-1 design features will be incorporated in the ACT-2 design approach. The major changes involve the FWB material choices, coolants and plumbing routing. Some of the lessons learned in the ARIES CS DCLL design will be adopted in this design. Xueren thought the ACT-1 manifold design could be used. The initial layout of ACT-2 was shown along with key design features. Xueren explained there are two design approaches for the DCLL blanket: 1) large blanket box concept and 2) small modular concept. The large blanket box concept is applied to each of 16 sectors with manifolding in each sector, whereas the small module concept subdivides the sector into four individual blanket modules that results in more plumbing connections. A lot of detail work has been developed on the design of the manifolding and plumbing connections.
Design Curves for the Plate Divertor and the Status of the GT Helium Test Facility - Minami Yoda explained the recent efforts at Georgia Tech to estimate the maximum heat flux allowable on the divertor and the resultant pumping power requirements for the helium-cooled flat-plate (HCFP) divertor. This work builds on the 2010 experiments by Hagemann at GT on an air-cooled electrically heated brass test section. Using numerical simulations with the commercial ANSYS code, GT has looked at the effect of thermal conductivity ratio on the thermal performance of an older version of the HCFP design which matches the experimental studies. Minami reported new correlations of the HCFP test module over a range of temperatures and Reynolds numbers. GT plans to perform more experiments on a new air-cooled electrically heated stainless steel test section to validate this correlation at different values of the thermal conductivity ratio later this calendar year.
The ongoing experiments in the GT helium test facility with the helium-cooled divertor with multiple-jet cooling (HEMJ) test section have achieved heat fluxes up to 2.2 MW/m2 from an oxy-acetylene torch with no oxidation of the tungsten surface observed. The experimental data continue to show good agreement with the previously developed Nu correlation. GT will continue to expand the range of experimental parameters, especially in terms of the incident heat flux. There was some discussion of whether the Densimet D185 tungsten-alloy used in the current HEMJ test section was a good choice. Per discussion with Arthur Rowcliffe, Minami will contact Michael Rieth at KIT regarding sources for WL10.
Progress in the Development of Nano-Structured Steels - Arthur Rowcliffe explained the composition of the nano-structured steels and their potential advantages to increase the fusion power plant operating temperature window while accommodating much higher levels of atom displacement damage and helium concentrations. There is an ongoing initiative to develop a best-practice processing technology for the 14YWT ferritic alloy and, to date, this project has successfully demonstrated reproducible process control with 22 kg batches while scale-up to 60 kg batches is in progress. Recent tensile and fracture properties demonstrating improved isotropy were presented.
Arthur summarized the recent demonstration by UK researchers of the production of tonnage quantities of a nanostructured low steel alloy in which alternating nano-scale platelets of austenite and ferrite are formed by a bainitic transformation during continuous cooling from the austenite phase. The very high density of interfaces gives rise to an interface surface area to volume ratio of the same order as the ratio that is characteristic of Nano-structured Ferritic Alloys (NFA) such as 14YWT. The high strength combined with favorable ductility and fracture properties indicates that this type of alloy could form a basis for the development of alloys with mechanical, corrosion and radiation damage resistance properties for fusion applications.
Power Core Safety Analysis Results - Paul Humrickhouse discussed his revisions and updates to the MELCOR model since the February ARIES project meeting (radial build updated, shield plugs added, blanket structures added, kink shell added, divertor dome and structure added, modified decay heats, and helium cooling system resized). Paul noted both the nuclear and decay heating have increased from the prior results.
As a reference point, Paul showed both the model schematic for the prior model and the current model with the aforementioned additions. One of the accident scenarios being analyzed is an ex-vessel, double guillotine break in the helium outlet piping. His primary concern is an over-pressure in the cryostat that, if compromised, would result in release of radioactive materials. A previous analysis (described at the January 2013 ARIES meeting) had indicated a pressure rise to 0.65 MPa, which is in excess of the ARIES-CS cryostat design limit of 0.3 MPa. The coolant volumes were resized after the January meeting and presently this accident results in acceptable cryostat pressures of no greater than ~0.23 MPa.
Paul displayed the pressure and temperature time profiles inside the cryostat during a loss of flow accident (LOFA). The temperatures resulting in this case were higher than in previous analyses, in part due to the revised model and nuclear/decay heats and in part due to some (since resolved) problems with the MELCOR deck. He thought the combined Pb/He LOFA and water LOCA accidents would result in more severe consequences. Paul also noted some interesting natural circulation phenomena in these cases. It was apparent that certain structures (e.g., the IB water-cooled shield and divertor steel structure) would require some enhanced heat transfer capabilities to operate within the desired temperature ranges.
Updated Thermal Response during (a) LOCA/LOFA Event in ACT-1 - - Laila reported the progress Carl Martin has been making on his safety analysis. He assumed a complete loss of both water and helium coolants in the power core as well as loss of flow of the LiPb power core coolant. Additionally, the plasma remains on for a full three seconds after the LOCA/LOFA event. To help radiate the heat from the inboard components, he has assumed an emissivity of 0.9 on all facing surfaces. The heat is assumed to be removed from the shield plug to the port door.
The 2-D analysis assumed the time history of the heat decay for the power core components as shown in the provided graph. Carl changed the inboard shield filler from W-C to borated-FS, which reduced the maximum temperature by 46°C. However the FS still exceeds its reusability temperature limit of 700°C.
Carl enhanced the model by modeling the divertors and incorporating cross-channel radiation using computed view factors. The inboard maximum temperatures for the ODS-FS Structural Ring approached 900°C, but the outboard Structural Ring remained below 700°C. The ODS-FS structure of the divertor closely approaches 900°C.
In conclusion, changing the inboard shield filler did help. The divertor heating only had a slight impact on temperatures. The inboard FS will likely exceed its reuse temperature of (750°C). There are several solutions for these temperature excursions. There are several solutions for these temperature excursions, including the use of NFA that can survive such accidents if the temperature remains below 1050°C.
Preliminary Disruption Analyses - Jake Blanchard, with support from Carl Martin, has been analyzing the impact of a plasma disruption considering only the quenching of the plasma current, but with no plasma displacement. He is assuming the plasma current decreases linearly to zero in 30 ms. He is modeling a continuous, electrically-connected, core vacuum vessel structure (not including the first wall, blanket, or shield). It is assumed the TF coils are not active, but the PF coils are active. The vacuum vessel is modeled as a solid shell, including the large vacuum ports. Jake showed the graphs of the total magnetic flux (BSUM), induced currents, and resulting forces before and after the current ramp-down. The presence of the steel ring alters the currents in the vacuum vessel. Overall, the vacuum vessel forces seem manageable. Future work will include a more accurate representation of the plasma current distribution and stabilizing shells.
3-D Estimate of SR (Structural Ring) and VV (Vacuum Vessel) Lifetimes for ACT-1 - - Laila El Guebaly has been concerned about neutron streaming through the assembly gaps. They are nominally 2-cm wide at room temperature on a new set of sectors. As the power core heats up, these gaps will partially close. Also during operation, neutron-induced swelling will contribute to the gaps decreasing in width. As the designers and operators become more experienced with the gaps during operation, they will design the gaps so there will be no adverse interference or stress at the contact points.
Laila has examined the radiation damage for a range of gap widths from 0 to 2 cm. Her analysis determined the radiation damage peaking in the gap, the dpa and He production in the Structural Ring and Vacuum Vessel materials as well as their estimated lifetime (related to a 200 dpa limit).
Laila noted that for any gap thickness, the top, middle and bottom of the power core sector gaps behave differently due to the non-uniform temperature and NWL. She explained the 3-D model and the reference radial build that she used. She then discussed the damage parameters expected on the inboard and outboard SR and VV. From these analyses, she concluded the IB Structural Ring should be replaced one time during the plant's lifetime and it would not be reweldable (given the present material properties). Les Waganer noted that this would be appropriate as the current maintenance approach employs a spare set of populated Structural Rings in the Hot Cell to expedite the maintenance procedure. On the other hand, the entire VV should survive as a lifetime component (dpa << 200) although it is not reweldable at the inboard midplane even in the absence of a gap. An improved re-weldability limit for FS (> 1 He appm) could allow rewelding the OB VV at locations away from gaps and penetrations. Laila also evaluated the He/dpa ratio for the FS structure and showed a range of 0.3-2.0 for the SR and 0.1-8.0 for the VV.