Progress On Cost Modeling- Cost Accounts 20 and 21

1. Progress On Cost Modeling- Cost Accounts 20 and 21 L. Waganer 14-15 June 2007 ARIES Project Meeting at GA

2. It is recommended that cost modeling be updated and fully documented Should ARIES Systems Code Cost Modeling Be Revised? • C. Bathke, and subsequently R. Miller, constructed ARIES cost models based on prior work and DOE recommended modeling practices in early 90s • Costing model is not documented to any great extent (some in SPPS) • Published cost data in final reports are limited in depth • Some cost data are not reported, e.g. 22.03 through 22.08 • 22.03 Cryogenic Cooling • 22.04 Radioactive Waste Treatment and Disposal • 22.05 Fuel Handling and Storage • 22.06 Reactor Plant Maintenance Equipment • 22.07 Reactor Plant Instrumentation and Control • 22.08 Other Reactor Plant Equipment • Only FWBSD and magnet costs have been benchmarked and updated on a semi-regular basis

3. General Costing Groundrules • PNL recommended initial spare parts and contingency be applied at major level-2 accounts, e.g., 21, Structures and Site Facilities • However, it is more reasonable to estimate initial spares be included within each cost account • Replacement costs for FWBD components are documented in annual replacement costs, Account 50, Scheduled Replacements • Contingency is applied at the plant level in Account 96, Project Contingency for unanticipated cost increases • Process (or Design) contingency can be reported in Account 95, but all design studies have assumed mature (tenth of a kind) plant, hence, the process contingency should be zero. For the TNS or Demo, some amount of process (design) contingency should be considered, perhaps 2%?

4. General Costing Background • Starfire Conceptual design (1980) was the most complete and best documented conceptual fusion power plant design. Ralph M. Parsons was the A&E firm that developed the designs and costs for all plant buildings and facilities. (1980$) • Generomak (Delene, 1986) developed a basis for a parametric reactor design with modeling algorithms based largely on Starfire and a few other designs (UW-Mad). • In the early 90’s, four IFE reactor designs were completed, Prometheus-L, Prometheus-HI, Osiris, and Sombrero (1991). An advisory commission developed a common set of costing guidelines for all four reactors to make sure all estimates were comparable and consistent. These studies are useful because the costing basis is fairly well documented. Although these were IFE designs, much of the BOP and other systems are relevant. (1991$ basis) • In the mid 90’s, the ARIES project began conceptual designs and costing analyses. ARIES adopted much of the prior guidelines and bases for their usage. The ARIES designs are fairly consistent, design to design, but the level of detail and documentation is somewhat lacking. So comparison with ARIES is rather limited. ARIES-SPPS final report did provide some algorithms that do apply to the ARIES series of designs. All costs are indexed to 1992$

5. Level of Safety Assurance • The Generomak introduced the concept of Safety Assurance Credits for buildings, reactor components, and indirect cost factors. • ARIES adopted the Level of Safety Assurance (LSA) with four levels of cost credits for both direct and indirect costs. It is implied that the cost algorithms do not contain the LSA credits and these are applied when the costs are summed. • When the recommended LSA credits are applied on Buildings, the LSA 1 and LSA 2 costs are very low (0.6, 0.9, 0.96, 1.00 for Reactor and Hot Cell; 0.60, 0.67, 0.67, and 1.0 for all other buildings) • I do not believe the FWBSD and coil benchmarked costs were developed with LSA guidelines considered • I would recommend ARIES discontinue the use of LSA credits. Just state it is an LSA of 1 or 2 and show the cost estimate.

6. It is recommended that the cost accounts be defined and reported to a reasonable depth • PNL-2648, “Fusion Reactor Design Studies – Standard Accounts for Cost Estimates”, 1979, defined a set of standard accounts that all fusion conceptual plant studies have followed • Reasonable depth is probably a second level definition on all major accounts (e.g., Cryogenic Cooling System) except for third level Reactor Equipment (e.g., RF Heating and Current Drive) • This definition was followed by Starfire and the 1990’s IFE plant designs. Whereas, ARIES used a simplified one-page summary. • Lower level detail is shown in account listing and used to develop account costs, but lower level costs are not reported • Recommended cost account definition is shown in attached spreadsheet, Representative Cost Accounts.xls

7. Land and Land Rights, 20 • The reference plant site was chosen to be 1000 acres in a Midwestern location. The land requirements are less severe than for an LWR in regard to exclusion boundaries, therefore 1000 acres are deemed adequate. In the case of constructing multiple plants at the common site, sufficient space is provided. • The cost associated with the land and privilege acquisition is estimated at 1000 acres times the cost per acre. Most of the prior fusion studies estimated the cost of land escalating with general land values for a Mid Western site, per PNL-2648 • This data shows a steady progression of escalating land values, which is historically correct. An examination of 2006 good cropland values along the Missouri or Mississippi rivers is currently in the range of $2500/A to $3000/A. Assuming there would be a sizable premium to obtain a contiguous site of 1000 A, it is not unreasonable to assume that the current land price would be $10,000/A. Thus the cost of Land and Privilege Acquisition, C20.01 is $10.00 M. • The Relocation of Buildings, Utilities, Highways, and Other Services is typically 10% of the land costs, or C20.02 is $1.00 M

8. Structures and Site Facilities, 21 This account covers all direct costs associated with the physical plant buildings such as reactor, turbine, electrical equipment, cooling system structures, site improvements and facilities, miscellaneous structures and building work, and ventilation stack. All provisions for cooling, site access and security will be provided. The facility is located in a secured area within the site. The site is adjacent to the "North River" which supplies adequate water for cooling purposes. The river is assumed to be navigable by barge traffic throughout the year to provide a means to ship in the large modules and equipment. Highway access is also provided by eight kilometers of secondary road leading to a state highway. The secondary road requires no improvement to permit overland shipments. Railroad access will be provided by constructing a five mile railroad spur from the main line to the plant site. Other site‑related assumptions have been established as follows: • Incoming power will be provided by two independent EHV power sources, probably 345‑kV or high voltage lines. • Power and water for construction are available at the site boundary. • Communication lines will be provided at the site boundary. • Sanitary sewage system will be available for tie‑in at the site boundary. • An auxiliary boiler furnishing plant auxiliary steam is included in the facility design. • Plant utility systems including compressed air, inert gas storage and distribution, and portable and de-mineralized water are included in the facility design. • Personnel parking will be located outside the facility perimeter close to the guard station that will control incoming and outgoing personnel, vehicles and rail cars. • The facility will be located on level ground at an elevation unaffected by potential flooding. • Seismic criteria UBC Zone 2 will be assumed for all structures.

9. Structures and Site Facilities, 21(Cont’d) • The direct cost for the Structures and Site Facilities, Acct 21, represents a very significant portion of the total facility cost and maybe higher than that for a comparable PWR. However the cost increases can be identified and are reasonable. • The Reactor Building much larger than a PWR containment building but contains more equipment. The remote handling features in the Reactor and Hot Cell Building contributes to increased size and cost. Tritium containment and detritiation contributes to the cost. • The Fuel Handling and Storage (Tritium Reprocessing) Building is additional as is the handling and containment of tritium. The premise that this facility should have very low release rates of tritium, low material activation, and a very high factor of safety, which greatly impacts the overall facility cost. • The Hot Cell is another facility which is additive to a PWR system. Because of the large sizes of the components and large number of components handled within the Hot Cell, a large building is required designed with remote handling and hazardous waste handling considered. • Several prior studies enlisted A&E firms for the buildings and facilities, e.g., Starfire (RM Parsons Company), Ebasco (Prometheus), and Bechtel (Osiris/Sombrero)

10. Site Improvements and Facilities, 21.01 This account includes all site improvements and facilities necessary for the complete power plant. This includes the general site improvements including site work, fencing, storm sewer, earth moving equipment, tank and pump foundations, fire protection, and sanitary sewers. Transportation access is provided by highway and railway access, but no waterfront improvements were considered. In fusion conceptual studies, this is typically a fixed cost item. The estimates escalated to 2007$ are STARFIRE at $24.66 M, Generomak at $24.8 M, Prometheus-L at $29.7 M and ARIES-SPPS at $25.5 M. It should be noted that ARIES-AT did not identify sub-accounts under Account 21, so the ARIES-SPPS is being used as the primary ARIES reference. It is thought this site improvement and facilities account only scales weakly with land size, as evidenced by the higher price for the Prometheus estimate. It is assumed that the cost of the site improvements and facilities should be $25M. (Additional detail on cost elements under this account is available in the Starfire or Prometheus reports.)

11. Reactor Building, 21.02 • The ARIES-AT Reactor Building is very compact design, a right circular cylinder that is 25.4 m in radius and a height of 54.3 m, inside dimensions for a volume of 110,000 m3. There is a lower level that is circular, 12.5 m in radius and 7.5 m high. The total volume is around 113.7 m3. (We may need to further refine this value later). The walls are 2.6 m thick for radiation shielding. This is thicker than most fusion power plant designs. The access corridors are included but the hot cell is estimated as a separate building.

12. Reactor Building, 21.02 (cont’d) • Several other fusion reactor conceptual power plant designs were considered in determining the selected cost basis, namely Starfire, Generomak, Prometheus-L/-HI, ARIES-SPPS, and ARIES-AT. These have a range of enclosed volumes and related costs. Cost is usually related to interior volume. Concrete thickness is normally similar. Generomak is scaled to the fusion island volume^0.6 • Starfire was a very large reactor building with a lot of lay-down space for reactor components. It also housed a very large reactor with larger than current thermal output. Since that time, the tokamak power cores have become smaller and the reactor buildings more compact. • For comparison, the Prometheus-L and –HI and Osiris designs are also analyzed. The reactor building for Prometheus-L is almost double the volume of the large Starfire design, but the cost per unit volume is the lowest of the buildings. Prometheus designs used the scaling relationship of $274.5*Vol+$36.08M in 2007$, which is a much lower cost that is probably an aggressive estimate. The Prometheus constant cost of $36M is due to a base cost associated with any size reactor building. Osiris is a very low cost building, but is expensive per unit volume. In the figure shown on the next page, a scaling relationship connect these three cost estimates.

13. 2007$ Calculated 3 ARIES - AT Vol ~ 114,000m Starfire SPPS 0.62 Prom - L C =$106.7 x (V /8e5) Ron’s rb rb Basis C =$ 274.5 .7 x Vrb + $36.1M rb Prom - HI Osiris Reactor Building, 21.02 (cont’d) With the ARIES algorithm, the cost for the nominal ARIES-AT reactor building with an interior volume of 113,700 m3 is $132.69M. For comparison, the Prometheus and Starfire scaling would be upper and lower bounding cases. I believe we should adopt the current ARIES costing algorithm for the Reactor Building, which estimates the Reactor Building cost to be $132.69M ($112/m3). No LSA factor included. Note: ARIES-AT Acct 21 is $350.9M (2007$)

14. Turbine Building, 21.03 • The Turbine Building may be the largest building, except for the Reactor Building. It houses the turbines and all the auxiliary equipment for the turbines. The surface condenser is located on the sub-grade level, the feed-water heaters are on the ground level and the turbine generator is on the upper level. The steel framed, truss roofed building is of a conventional construction consistent with current power plant installations. (PS, I need to determine where the heat exchangers (or steam generators) are located. In this building?) • The size of the Turbine building is dependent on the size of the turbines and ancillary equipment. All estimates are scaled based on the gross electric power generated raised to the 0.5 power for Prometheus and .75 for ARIES. Generomak uses the gross thermal power raised to the .5 power (conversion efficiency not considered). There is no difference noted for traditional or advanced Rankine or Brayton cycles. ARIES adds in a fixed cost of $7.92M in addition to the variable costs.

15. Turbine Building, 21.03 (cont’d) The table above summarizes the cost bases of the different designs and respective costs for the turbine building. In the last row, all algorithms are used with the ARIES-AT parameters. Starfire, Prometheus, Osiris, and Sombrero used engineer/contractor firms to provide a bottoms-up estimates, which should provide a high fidelity solution. Prometheus, Sombrero, Osiris, and ARIES used gross electrical power which is a good scaling parameter. Since it is the largest building, second only to the reactor building, I would recommend we adopt the Prometheus scaling relationship. This results in a cost of $73.893M (Acct 21.03). That is a 57% increase. Question: Since there is tritium in the primary heat transfer media and there may be only one HX between primary and secondary heat transfer media, should the turbine building have an containment criteria? If it should be considered as a containment building, the higher costs are more reasonable

16. Remainder of Str and Site Facl, 21.04-17 • In the interest of brevity, I will address all of the remaining buildings and facilities at the same time • Heat Rejection scales to rejected thermal power^.5 • Electrical Equipment & PS scales to net electrical power^.3 • Plant Auxiliary Systems scales to gross electrical power^.3 • Hot Cell Building are usually not scaled. It is recommended it scale to Reactor Building at 41% • Reactor Service Building is similar to Starfire and Prometheus at $4.16 M • Service Water Building is similar to Starfire, et. al at $1.46 M • Fuel Handling and Storage Building is around $15 M and scaled to tritium usage or fusion power ^.3 • Control Room Building is similar to Starfire, Generomak and Prometheus at $6.8 M • On-Site AC Power Building is similar to Starfire, Generomak and Prometheus at $4.5 M • Administrative Building is similar to Starfire, Generomak and Prometheus at $1.9 M • Site Service Building is similar to Starfire, Generomak and Prometheus at $1.9 M • Cryogenic and Inert Gas Storage Building is similar to Starfire, Generomak, and Prometheus at $2.0 M • Security Building is similar to Starfire and Prometheus at $0.68 M • Ventilation Stack is similar to Starfire, Generomak, and Prometheus at $4.00 M

17. Remainder of Str and Site Facl, 21.04-17 Adopted Algorithm Low Waste Power

18. Summation of Accounts 21 The recommended values for the cost estimating relationships can be summed to provide a total Account 21 cost of $381.26M. Table below shows several detailed cost estimates from prior conceptual design studies. The SPPS report defines several fixed cost buildings and facilities and algorithms that can be used to estimate the reactor and turbine buildings. The items in green are very similar in costs and, in most cases, will be applied to the Demo cost estimate. In a few cases, although there were some costs that were similar, a recommended cost is different. Smaller Building Incld Target Factory

19. Comparison of Accounts 21 Looking at the total Account 21 costs, Starfire is the most detailed, was the physically largest plant and produced more net electrical power (1200 MWe vs a nominal 1000 MWe). Since Starfire was designed, the size of the reactor island has been getting more compact. Also the trend is to higher efficiency thermal conversion, which reduces the thermal power, but at a higher temperature. The recirculating power fraction has been going down, hence lower gross electrical power. The other extreme is Osiris, which is a very simple power core and very optimistic estimates (the reactor building is only estimated at $46M in 2007$). The remainder is in the range of $350M to $490M. The summation of the recommended algorithms for ARIES-DEMO is $381 (based on ARIES-AT parameters), is toward the lower end of the prior range. This is reasonable as the ARIES-AT reactor building is very compact (i.e., low cost).

20. Next Actions If this is a reasonable approach, I would continue on to benchmark and compare the remaining cost accounts. The other accounts will involve soliciting the subject matter experts to determine the current and future cost estimates including learning curve factors.