Overview of the ARIES-RS reversed-shear tokamak power plant study

Transcription

1 Fusion Engineering and Design 38 (1997) 3 25 Overview of the ARIES-RS reversed-shear tokamak power plant study Farrokh Najmabadi a, *, The Aries Team: Charles G. Bathke b, Michael C. Billone c, James P. Blanchard d, Leslie Bromberg e, Edward Chin f, Fredrick R. Cole g, Jeffrey A. Crowell d, David A. Ehst c, Laila A. El-Guebaly d, J. Stephen Herring h, Thanh Q. Hua c, Stephen C. Jardin i, Charles E. Kessel i, Hesham Khater d, V. Dennis Lee g, Siegfried Malang a,j, Tak-Kuen Mau a, Ronald L. Miller a, Elsayed A. Mogahed d, Thomas W. Petrie f, Elmer E. Reis f, Joel Schultz e, M. Sidorov e, Don Steiner k, Igor N. Sviatoslavsky d, Dai-Kai Sze c, Robert Thayer k, Mark S. Tillack a, Peter Titus e,l, Lester M. Wagner g, Xueren Wang a, Clement P.C. Wong f a Fusion Energy Research Program, School of Engineering, Uni ersity of California, San Diego, La Jolla, CA 92130, USA b Los Alamos National Laboratory, Los Alamos, NM 87545, USA c Argonne National Laboratory, Argonne, IL 60439, USA d Uni ersity of Wisconsin, Madison, WI 53706, USA e Massachusetts Institute of Technology, Cambridge, MA 02139, USA f General Atomics, San Diego, CA 92186, USA g McDonnell Douglas Aerospace, St. Louis, MO 63166, USA h Idaho National Engineering Laboratory, Idaho Falls, ID 83415, USA i Princeton Plasma Physics Laboratory, Princeton, NJ 08543, USA j Forschungszentrum Karlsruhe, Gmbh, Karlsruhe, Germany k Rensselaer Polytechnic Institute, Troy, NY 12180, USA l Stone and Webster, USA Abstract The ARIES-RS tokamak is a conceptual, D T-burning 1000 MWe power plant. As with earlier ARIES design studies, the final design of ARIES-RS was obtained in a self-consistent manner using the best available physics and engineering models. Detailed analyses of individual systems together with system interfaces and interactions were incorporated into the ARIES systems code in order to assure self-consistency and to optimize towards the lowest cost system. The ARIES-RS design operates with a reversed-shear plasma and employs a moderate aspect ratio (A=4.0). The plasma current is relatively low (I p =11.32 MA) and bootstrap current fraction is high ( f BC =0.88). Consequently, the auxiliary power required for RF current drive is relatively low ( 80 MW). At the same time, the average * Corresponding author /97/$ Elsevier Science S.A. All rights reserved. PII S (97)

2 4 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) 3 25 toroidal beta is high ( =5%), providing power densities near practical engineering limits (the peak neutron wall loading is 5.7 MW m 2 ). The toroidal-field (TF) coil system is designed with relatively conventional materials (Nb 3 Sn and NbTi conductor with 316SS structures), and is operated at a design limit of 16 T at the coil in order to optimize the design point. The ARIES-RS design uses a self-cooled lithium blanket with vanadium alloy as the structural material. The V-alloy has low activation, low afterheat, high temperature capability and can handle high heat flux. A self-cooled liquid lithium blanket is simple, and with the development of an insulating coating, has low operating pressure. Also, this blanket gives excellent neutronics performance. Detailed analysis has been performed to minimize the cost and maximize the performance of the blanket and shield. One of the distinctive features of this design is the integration of the first wall, blanket, parts of the shield, divertor and stability shells into an integral unit within each sector. The maintenance scheme consists of horizontal removal of entire sectors. Prior to the initiation of the ARIES-RS study, a set of top-level requirements and goals for fusion demonstration and commercial power plants was evolved in collaboration with representatives from US electric utilities and from industry. The degree to which ARIES-RS reached these requirements and goals and the necessary trade-offs are described and the high-leverage areas and key R&D items are presented Elsevier Science S.A. Keywords: ARIES-RS; Reversed-shear; Tokamak power plant 1. Introduction The ARIES-RS tokamak power plant study continues in the tradition of the ARIES Program to establish the economic, safety and environmental potential of fusion power plants, and to identify physics and technology areas with the highest leverage for achieving attractive and competitive fusion power in order to guide fusion R&D. The ARIES Team is a US national effort with participation from national laboratories, universities and industry, and with strong international collaborations. The Team performs detailed physics and engineering analyses using the most current and detailed models available, and then uses the results to perform optimization and trade studies via a cost-based systems code. Prior to the initiation of the ARIES-RS study, a set of top-level requirements for fusion demonstration and commercial power plants was evolved in collaboration with representatives from US electric utilities and from industry [1 4]. These requirements were framed in a quantitative way to help establish the minimum necessary features of a fusion power plant that would lead to its likely introduction into the US and world Table 1 Top-level requirements and goals for commercial and demonstration fusion power plants Element Demonstration Commercial Must use technologies to be employed in commercial power plant Yes Yes Net electric output must be greater than 75% Commercial N/A COE must be competitive (in 1992 mill (kwe h) 1 ) 80 (Goal) 65 (Goal) 90 (Reqmt) 80 (Reqmt) No evacuation plan required for any credible accident: Total dose at site boundary 1 rem 1 rem Generate no rad-waste greater than Class C Class C Must demonstrate public day-to-day activity is not distrubed Yes Yes Must not expose workers to a higher risk than other power plants Yes Yes Must demonstrate robotic maintenance of power core Yes Yes Must demonstrate routine operation with less than (x) unscheduled shutdowns per year 1 1/10 including disruptions Demonstrate a closed tritium fuel cycle Yes Yes Must demonstrate oepration at partial load conditions at 50% 50%

3 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) electric energy supply. These requirements constitute a common basis for quantitative evaluations of any candidate fusion concept. A similar set of criteria has been developed by the EPRI fusion working group [5]. These criteria and associated top-level requirements and goals (Table 1) can be divided into three general categories: (1) cost; (2) safety and environmental features; and (3) reliability, maintainability, and availability. Top-level requirements and goals for cost of electricity (COE) were adopted based on the estimated cost of competitive source of electricity at the time of introduction of fusion in the market place [1,6,7]. These requirements and goals for COE are also in line with projections of future power plant costs based on energy forecasting models [8]. Safety and environmental requirements are included to circumvent the difficulties experienced by fission and, to some degree, expected to be faced by fossil fuels in the future. Fusion should be easy to license by the national and local regulating agencies, and be able to gain public acceptance. Fusion power plants should only generate low-level waste (i.e. waste storage time less than a few hundred years, such as Class- A through C under US regulations). Realization of the full safety and environmental potential of fusion will also help fusion to achieve a cost advantage over other sources of electricity. Fusion power plants can be designed to achieve these criteria only through the use of low-activation materials and care in design. However, these requirements result in stringent constraints on the sub-system choices and design. Lastly, it should be demonstrated that Demo and commercial power plants can achieve the necessary degree of reliability. The conceptual design studies can partially address this issue (e.g. by including maintenance considerations in the design). This criterion, to a large degree, should be addressed in the development path of fusion power. Today s experiments are not intended to provide detailed engineering data to support the design, construction, and operation of a power plant. With the top-level requirements in hand, an assessment of various tokamak plasma operation modes and engineering design options was made [1,3,9]. In each area, this assessment was aimed at investigating: (1) the potential to satisfy the toplevel requirements and goals; and (2) the feasibility (e.g. critical issues) and credibility (e.g. degree of extrapolation required from present data base). Five different tokamak plasma regimes were considered: (1) steady-state operation in the firststability regime, e.g. ARIES-I [10]; (2) pulsedplasma tokamak operation, e.g. Pulsar [11]; (3) steady-state operation in the second-stability regime, e.g. ARIES-II and ARIES-IV [12]; (4) steady-state operation with reversed-shear profile; and (5) low-aspect ratio tokamak (spherical tokamak). Several physics figures of merits were identified in order to assess the extent of the plasma physics data base for each option [1,13]. An assessment of the five tokamak physics regimes was made based on economic performance and maturity of the data base [1,9,14]. The first-stability pulsed-plasma and steady-state regimes are closest to the present data base. Of course these regimes should be demonstrated at long-pulse discharges with burning plasmas. On the other hand, the economic performance of pulsed-plasma operation is poor. First-stability steady-state did not achieve the economic requirements for the Starlite project. High-field toroidalfield coils can improve the attractiveness of this regime of operation. The second-stability regime has better economic performance but the experimental data base for this regime is very small. The data base for spherical tokamaks is not mature yet and, in addition, many critical issues remain. Detailed design studies are needed before the true potential of spherical tokamak power plants can be assessed. The reversed-shear mode of operation offers the best economic performance. The data base for this regime, while small, is growing rapidly. Based on the superior economic performance and the growing experimental and theoretical data base, the reversed-shear plasma mode of operation was chosen for detailed design phase. Various classes of engineering design options were examined and their potential to meet the power plant requirements assessed [1,15,16]. These options included material choices for the structure, breeder and coolant. The design space for an attractive tokamak fusion power core is not unlimited; previous studies have shown that

4 6 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) 3 25 advanced low-activation ferritic steel, vanadium alloy, or SiC/SiC composites are the only known viable candidates for the primary in-vessel structural material. In order to provide a framework for this assessment, these three material classes were used to distinguish engineering design choices, to determine if they meet the top-level requirements, and to characterize the remaining technical issues. Silicon-carbide composite has a unique potential for safety and high performance, but the database requires substantial improvement in order to identify and develop a satisfactory material composition. Significant improvements are needed in the basic properties, and several key material issues such as joining and hermeticity must be resolved. Ferritic steel has the largest database, and hence the smallest uncertainty in its performance. However, as with all materials, behavior after long-term operation in a complete fusion environment is highly uncertain. The features of ferritic steel which cause the greatest concern are its restricted temperature window, which limits the maximum achievable thermal conversion efficiency, and its potential loss of ductility under irradiation. It is not clear how one would operate a high-power device with pressure-bearing structures made of potentially brittle materials. In addition design options have to be developed to address the low maximum temperature capability of ferritic steels so that an acceptable power conversion efficiency can be achieved [15]. Vanadium offers significant advantages in its high temperature and high thermal performance capability, and its low activation. The combination of V-alloy structure and Li coolant appears particularly unique in material chemistry, offering good compatibility up to very high temperatures. The primary concerns with vanadium alloy are high cost and liquid metal MHD effects (including the need for and feasibility of insulating coatings). Minimization of vanadium in the shield and external systems is an important cost-reduction strategy. The absence of an established industrial base is also an important consideration, but not one of the top-level requirements as elaborated above. Due to its greater ultimate potential for attractive commercial power plants and a development path which appears practical within a 25 year time-frame, the combination of Li breeder/coolant and vanadium-alloy structure was chosen for full system design and analysis. While there is no unique design concept guaranteed to succeed, the ARIES Team choices (a reversed-shear plasma with in-vessel components made of high-performance vanadium-alloy structures cooled by lithium) represent trade-offs between a reasonable database and an acceptable performance. For the detailed engineering design, the blanket and shield concepts were based on ARIES-II [12]. The choice to develop an existing design, rather than explore entirely new concepts, was made in order to evolve a higher level of detail and sophistication in the design, which then allows a much firmer basis for evaluating the attributes and a clearer understanding of the issues and R&D needs. This paper provides an overview of the ARIES- RS design including physics (Section 2) and engineering (Section 3) analysis, design, and trade-offs. Safety and licensing issues are discussed in Section 4. Section 5 examines the extent to which the ARIES-RS has met the top-level requirements and describes the key R&D issues. The details of the ARIES-RS study can be found in [17 23] in this issue and in the ARIES-RS final report [24]. The ARIES-RS tokamak is a conceptual, D T- burning 1000 MWe power plant. As with earlier ARIES design studies, the final design of ARIES- RS was obtained in a self-consistent manner using the best available physics and engineering models. Detailed analyses of individual systems together with system interfaces and interactions were incorporated into the ARIES system code in order to ensure self-consistency and to optimize towards the lowest cost system. Fig. 1 shows the fusion power core cross section and Table 2 summarizes the key parameters for the final design point of ARIES-RS. The design employs a moderate aspect ratio (A=4.0). Preliminary observations indicated that the costof-electricity (COE) variation within the range 3 A 4 was not significant. Engineering design considerations which could not be quantified accurately for systems-code analysis led to the choice of A=4. The plasma current is relatively low (I p = MA) and bootstrap current fraction is high ( f BC =0.88). Consequently, the auxiliary power

5 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) Fig. 1. The ARIES-RS fusion power core. required for RF current drive is relatively low. At the same time, the average toroidal beta is high ( =5%), providing power densities near practical engineering limits (the peak neutron wall loading is 5.68 MW m 2 ). The toroidal-field (TF) coil system is designed with relatively conventional materials (Nb 3 Sn and NbTi conductor with 316SS structures), and is operated at a design limit of 16 T at the coil in order to optimize the design point. Following the ARIES-II/-IV [12] and PULSAR [11] designs, a level of safety assurance (LSA) rating [25,26] of LSA=2 has been assigned, and appropriate cost credits (J.G. Delene, LSA cost factors for fusion reactors, Oak Ridge National Laboratory, private communication, December 1990) have been taken. Table 3 summarizes the final set of economic parameters. 2. Plasma physics The ARIES-RS design operates in the reversedshear mode which has a considerable potential for providing the basis of an economically competitive fusion power plant. The benefits of this configuration are that it achieves both high N and high, it obtains large bootstrap current fractions with very good current profile alignment, and appears to provide the transport suppression necessary to sustain the pressure profile that is consistent with the higher and high bootstrap current [27 29]. Like all configurations with N values which exceed the first stability regime this plasma requires a conducting wall to stabilize low-n external kink modes. For ballooning modes the reversed shear plasma is in second stability in the region inside the minimum safety factor, and in first stability outside this region. An additional advantage of the reverse shear configuration is that the hollow current profile matches very closely with the natural shape of the plasma generated bootstrap current profile. This allows one to significantly reduce the amount of external current drive required. Recent experiments [30 32] have demonstrated

6 8 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) 3 25 transiently the improved MHD stability and suppression of plasma particle and energy transport. The experiments each observed strong transport suppression in differing degrees for energy and particles. Both experiments have determined that the plasma core is in the second stability regime for ballooning modes. In addition, both experiments obtain limits consistent with ideal MHD stability predictions, and these are expected to be responsible for the discharge terminations. The global energy confinement times obtained in these experiments have H-factors over ITER89-P scaling of 2 3. In addition, bootstrap current fractions between 50 and 75% have been inferred. Complete demonstration of all the favorable properties simultaneously will require steady state conditions, which are not accessible in present experimental devices. The reference plasma configuration is given in Fig. 2 and Table 4. The operating value is reduced to 90% of the maximum stable value to provide some margin to the stability boundary. Table 2 Operating parameters of the ARIES-RS tokamak power plant Aspect ratio 4.00 Major radius (m) 5.52 Minor plasma radius (m) 1.38 Plasma vertical elongation (x-point) 1.70 Plasma current (MA) Bootstrap current fraction 0.88 Current-drive power (MW) 81 Toroidal field on axis (T) 7.98 Peak field at the TF coils (T) 16 Toroidal 0.05 Average neutron wall load (MW m 2 ) 3.96 Primary coolant and breeder Natural lithium Structural materials Vanadium & steel Coolant inlet temperature ( C) 330 Coolant outlet temperature ( C) 610 Fusion power (MW) 2170 Total thermal power (MW) 2620 Net electric power (MW) 1000 Gross thermal conversion efficiency 0.46 Net plant efficiency 0.38 Recirculating power fraction 0.17 Mass power density (kwe ton 1 ) Cost of electricity (mill (kw h) 1 ) Table 3 Reverse-shear power-plant economic parameters (1992 $) Account Account title Cost number ($ M) 20. Land and land rights Structures and site facilities Reactor plant equipment First wall, blanket, and reflector Shield Magnets TF coils PF coils Supplemental heating systems Primary structure and support Reactor vacuum systems Power supplies Impurity control Direct energy conversion system ECRH breakdown system Reactor equipment Main heat transfer and transport Turbine plant equipment Electric plant equipment Miscellaneous plant equipment Special materials Total direct cost Construction services and equipment Home office engineering and services Field office engineering and services Owner s costs Project contingency Interest during construction (IDC) Escalation during construction (EDC) Total capital cost Constant $ [90] Unit direct cost, UDC ($ kwe 1 ) [94] Unit base cost, UBC ($ kwe 1 ) [99] Unit total cost, UTC ($ kwe 1 ) Capital return (mill (kwe h) 1 ) [40 47,51] O&M (1.4%) (mill (kwe h) 1 ) 9.16 [50] Blanket replacement (mill (kwe h) 1 ) 4.63 Decommissioning (mill (kwe h) 1 ) 0.50 [02] Fuel (mill (kwe h) 1 ) 0.03 COE (mill (kwe h) 1 ) MHD stability and bootstrap current Studies were done to determine the impact of plasma aspect ratio, triangularity, elongation, pressure and current profiles, and the kink stabi-

7 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) Fig. 2. The reference plasma equilibrium parallel current profile as a function of poloidal flux, the plasma flux surface contours, and toroidal current density as a function of the major radius. lizing wall position. The plasma current and pressure profiles were determined to provide a high stable and large well-aligned bootstrap current simultaneously. This includes maintaining the minimum value of the safety factor, q min, above 2.0 everywhere, maintaining q 95 above 3.5, moving the location of q min as close to the plasma boundary as possible (to increase the high pressure plasma volume), maintaining a degree of negative shear given by q o q min 0.3, and maximizing the match between the bootstrap current profile and the desired MHD stable current profile. The reference case plasma profiles are given in Fig. 3. The plasma geometry plays a significant role in MHD stability as well as the overall power plant design. The aspect ratio variations showed the well known trend that increased as the aspect ratio was lowered, however, the current drive power increased as well. System analysis indicated that the COE variation within the range 3 A 4 was not significant. Engineering design considerations which could not be quantified accurately for systems-code analysis led to choice of A=4. Scans of the plasma triangularity from 0.2 to 0.6 showed that a triangularity above 0.4 was necessary for high and that increased further with increasing triangularity. For ARIES-RS, the triangularity was limited to 0.5 because of the inboard divertor requirements (mainly space requirements for sufficient neutron shielding). Plasma elongation improves, but makes the plasma unstable to vertical motion. The maximum elongation is limited by the passive-stabilization conducting structure and the feedback control system. The conducting structure was designed to provide a stability safety margin ( f s =1+ g / L/R ) of 1.2, and was made of tungsten 4.0 cm thick. The tungsten structure was made electrically continuous in the toroidal direction and located in the gap between the reflector and shield. The plasma elon-

8 10 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) 3 25 gation at the separatrix (and 95% surface) was limited to 1.9 (and 1.7). Feedback control simulations for the plasma vertical position showed that 15 MVA was required if the control coils were located in the gap between the reflector and shield, and 40 MVA was required if they were located behind the shield (note that most of this power is reactive power). It is important to emphasize that a conducting wall is required to stabilize low-n kink modes for these configurations to obtain high values. Since no second stability regime (similar to that for the n ballooning modes) has been found for the low-n kink modes, it is necessary to utilize a conducting wall to access the higher pressures where a fusion power plant can become significantly more attractive. Recent experiments and theory indicate that a resistive wall can stabilize the low-n kink modes if plasma rotation and a dissipation mechanism in the plasma are present. Predictions from ideal MHD theory indicate that the plasma must rotate at about 5 10% of Table 4 Reference reversed shear plasma configuration Plasma current, I p (MA) 11.3 On-axis toroidal field (T) 7.98 Major radius (m) 5.52 Minor radius (m) 1.38 Elongation Triangularity Poloidal 2.28 a Toroidal 4.96% a Toroidal * 6.18% a Normalized N 4.83% a Maximum N 5.35% a Bootstrap current (MA) 10.0 Driven current, I CD (MA) 1.2 On-axis safety factor, q o 2.80 Minimum safety factor, q min 2.49 Location of q min, (q min ) 0.69 Edge safety factor, q e 3.52 Edge safety factor, q* 2.37 Internal inductance, l i (3) 0.42 Density profile peaking factor, n 0 / n 1.36 Temperature profile peaking factor, T 0 / T 1.98 Pressure profile peaking factor, p 0 / p 2.20 Location of kink-stabilization wall b a The ARIES-RS operates at 90% of maximum theoretical. b Distance fom the plasma edge normalized to minor radius. the Alfven speed, which for ARIES-RS is m s 1. Neutral beams are capable of providing such plasma rotation speeds. On the other hand, ARIES-RS uses only RF current drive. The highest plasma rotation speeds from RF were observed for ICH at m s 1. Whether these RF driven plasma rotation speeds will be sufficient for kink mode stabilization is not clear, and this is presently an active area of research. In principle, low-power beams can be added to the system for plasma rotation if RF systems cannot provide sufficient rotational speed. The theory also gives some criteria for the wall conductivity and thickness. For the ARIES-RS design, the criteria is / , where is the thickness and is the wall resistivity. Scans for the distance of the wall from the plasma showed that N was limited to 5.35 for wall locations closer than 0.25a from plasma separatrix (a is the plasma minor radius). This was due to the bootstrap current becoming too large, even though the could actually be increased as the wall was made closer. For ARIES-RS, the kink wall is integrated in the blanket design by slightly thickening the structure behind the first wall cooling channels Current dri e In the ARIES-RS reference equilibrium, the bootstrap current fraction is 0.88, using density and temperature profiles which are characteristic of reversed shear plasmas. External non-inductive techniques are generally required to drive currents at the plasma center and off-axis near the shear reversal region. A number of non-inductive current-drive techniques have been considered for the reversed shear power plant. Depending on the reference equilibrium, a combination of these techniques is required to drive currents in different locations within the plasma. Because of their natural tendency to propagate radially towards the plasma center, ICRF fast waves are most suitable for current drive on the magnetic axis. On the other hand, with a proper launched spectrum, the lower-hybrid wave can be made to drive currents off-axis at the shear reversal region. However, lower-hybrid waves typically cannot

9 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) Fig. 3. The reference plasma equilibrium profiles for the safety factor, pressure, temperature, and density as a function of poloidal flux. penetrate deeper into the plasma and the choice of another RF technique for off-axis (mid-plasma location) current drive is necessary. Several candidates have been considered, including high-frequency fast waves (HFFW), mode conversion current drive (MCCD), minority heating current drive (MHCD), and electron-cyclotron waves. All these techniques have the potential capability of localized off-axis current drive in a fusion plasma core, each using a distinct physical mechanism. For the ARIES-RS design, HFFW was chosen to aid the lower-hybrid wave for off-axis current generation. Other candidates may be equally attractive, but have not been fully explored in this study mainly because of the lack of resources within the project. Indeed, seeking the most attractive technique to drive currents off-axis in the plasma core is presently a topic of intense research interest. A large number of current drive calculations were performed to scale the current drive efficiency as a function of plasma aspect ratio, temperature, and Z eff for the reversed shear plasma configurations. Shown in Fig. 4 is the reference plasma equilibrium with the various current drive components. All three systems were used in this particular case, even though in some instances, only two or even one system may be sufficient. All required wave spectra have values of N that can be launched by reasonable launcher designs. In this particular case study, the aggregate current-drive efficiency is found to be B = AW 1 m 2, and the current-drive system details are given in Table 5. A radiative divertor was utilized to disperse the transport power and keep the heat-flux at the divertor region to an acceptable level. The impact of the radiative divertors on the choice of MHD equilibrium and current-drive power constitutes a new feature of the ARIES-RS physics analysis and is described in Section 2.3 below.

10 12 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) 3 25 Fig. 4. Matching of the driven (dotted line labeled T) current density profile to the target profile (solid line labeled E), using ICRF fast wave (FW), high-frequency fast wave (HFFW), and lower-hybrid wave (LH) power, for a reversed shear equilibrium similar to the ARIES-RS design point Di ertor physics Because of the high power density of the ARIES- RS design, the heat fluxes on divertor structures can be very high. Material surfaces under such high heat loads are subject to severe thermal-related structural damage. Thus, to avoid compromising the structural integrity of the divertor it is necessary to reduce the heat fluxes to more manageable levels. To avoid overheating the divertor surfaces, the power flow into the scrape-off can be dispersed over a wider area. One way this can be done is by enhancing the radiated power from the divertor Table 5 Current drive requirements for ARIES-RS-like equilibrium and main plasmas, often referred to as radiating divertor and radiating mantle methods, respectively. In the former, excessive power flowing into the divertors is dissipated largely by radiation in the divertor channel. In the latter, this power flow is radiatively dissipated at the edge of the plasma before it passes into the scrape-off layer (SOL). Both methods require injection of high-to-medium Z impurities into the plasma (which increases Z eff ) and operation at a high edge density (n ea /n e0 0.2). Analysis indicates that increased Z eff severely degrades the overall current-drive efficiency. In addition, the substantial edge density implies that a larger fraction of the current-drive power will be in the lower hybrid system, which has the lowest system efficiency and the highest unit power cost among the three RF systems. The lower-hybrid power fraction increases at higher Z eff and lower temperatures. Analysis showed that the radiating mantle option (which requires a higher Z eff and/or a higher edge density) leads to systems with unacceptable current-drive power requirements. Therefore, a radiative divertor option was chosen which required a lower edge density, n ea /n e0 0.2 and a lower Z eff 1.7. For the radiative divertor case, the impurity chosen was neon, since this atom radiates efficiently in the cooler scrape-off and divertor regions. While the higher Z impurities can also radiate efficiently, their impact on the core plasma is more severe than the lower Z neon. These studies indicated that adding neon to the plasma system, under conditions consistent with other analyses (i.e. MHD stability, current drive, bootstrap current, and transport), would radiate a System Frequency N Power (MW) Launcher position ( ) a ICRF fast wave 91.0 MHz High-frequency fast wave 1.0 GHz GHz Lower hybrid wave 4.6 GHz GHz GHz GHz GHz a In poloidal angle above (+) or below ( ) outboard mid-plane.

11 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) Fig. 5. Plasma operating contours of the auxiliary power (in increments of 10 MW) required for steady-state plasma power balance displayed in ion density and temperature space for the following assumptions: ITER-89P confinement scaling with a multiplier H=2.339, I p =11.3 MA, I p /a p B T =1.028 MA mt 1, a p =1.38 m, f Ne =0.0055, and p */ E =10. sufficient amount of power in the SOL and divertor regions to make a feasible divertor solution. In this case, the sum of particle and radiation heat flux does not exceed 6 MW m Plasma operating regime A Plasma OPerating CONtour (POPCON) plot provides a rudimentary indication of a start-up trajectory and thermal stability of the steady-state operating point. The contours of the auxiliary power required for steady-state plasma power balance as a function of fuel density and temperature are displayed in Fig. 5. Ignited plasmas are possible for fuel temperatures in the range of 9 T i 15 kev and fuel densities n i m 3. In regions outside of the ignition region transport and radiation losses exceed the fusion power, and auxiliary power is required for steady-state operation. An ARIES-RS-like operating point is maintained with an auxiliary power of 81 MW (current-drive power). The start-up trajectory to fusion burn that requires the minimum auxiliary power follows the valley floor (in auxiliary power) between a radia- tion (i.e. bremsstrahlung and line radiation) hill at low temperature and a transport hill at high temperature that results from increasing power levels, decreasing confinement times, and increasing ash concentration with increasing temperature. The minimum-power trajectory requires a maximum of 15 MW of auxiliary power at T i 7 kev and n i m 3. The start-up trajectory to the ARIES-RS operating point requires four times more power, but the maximum auxiliary power occurs at the final operating point, because the operating point resides on the side of the hightemperature transport hill. As was mentioned before, plasma transport is not explicitly modeled. Plasma temperature and density profiles are only made to roughly agree with theoretical and experimental observations for reversed shear configurations. The global plasma energy confinement time is determined by the systems code to provide plasma power balance. The energy confinement time is 1.34 s, which corresponds to an ITER-89P scaling multiplier of This value is within the present experimental bounds for reverse shear plasmas, as well as other plasma operating modes. The particle confinement time in the main plasma is expected to track the energy confinement time. Experiments on helium removal which used significant active pumping indicate that the ratio p */ E is in the range of 7 15 ( p *is the effective particle confinement time including recycling). For the ARIES-RS design a value of p */ E =10 is used. Controlling the edge plasma density, n(a) at the separatrix boundary is important for avoiding disruptions. For ARIES-RS density limits are predicted to be in the range of m 3. The edge density ratio n(a)/n 0 of 0.2 satisfies the lower limit and is consistent with the divertor, ideal MHD stability, and current drive solutions. 3. Fusion power core engineering The ARIES-II power core [12] has been used as a starting point from which additional design detail and improvements can proceed. The ARIES-RS design process was driven by the need

12 14 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) 3 25 to meet the top-level system requirements. Meeting economic goals while maintaining the safety and environmental attractiveness possible with a fusion energy source was a major thrust. For example, minimization of the use of vanadium to only those areas where its use is necessary for performance reasons made a significant impact on the COE. Plant availability is another significant factor in the COE. Modern fission plants are aiming at 80% plant capacity factors [33], implying availability of 85% or more. Availability is especially important with high capital-cost facilities such as a fusion plant. The layout of the power core, as well as the detailed design of the internal components, was continuously influenced by the goal to reduce the down-time for sector repairs and replacement to about 1 month. Finally, attempts were made to develop design solutions which integrate both physics and engineering constraints. For example, the requirements for both passive and active stability systems were incorporated into the sectors from the initial strawman and revised as necessary during the design process. The divertor is another area where efforts to develop physics solutions consistent with engineering limitations, and engineering solutions consistent with physics constraints, were made throughout the design study. In the following subsections, the overall design features of the power core are summarized and the principle features and conclusions from the individual components are described. These include the first wall and blanket, radiation shielding, divertor, magnet systems, and current drive, heating and fueling systems Configuration and maintenance Achieving a high plant availability goal was a primary influence on the overall configuration and on the major power core design decisions. This requirement is met by emphasizing: (a) rapid replacement time (1 month is the goal) via configuration design; (b) component reliability via design simplicity and operating margins; (c) long lifetime via material selection; and (d) horizontal maintenance of full sectors is performed through large ports (Fig. 6) using a rail system together with transporter casks. Port doors on the back of the shield and at the cryostat prevent the spread of radioactive contamination to the confinement building. All coolant connections are made in the evacuated port area where the radiation field is low. This design allows rapid disconnections of the piping, which could be either mechanically sealed or cut and rewelded. One of the distinctive features of this design is the integration of the sectors. The first wall, blanket, parts of the shield, divertor and stability shells form an integral unit within each sector (Fig. 7). The integrated sector construction eliminates in-vessel maintenance operations and provides a very sturdy continuous structure able to withstand large loads. Cavity loads are supported at the bottom through the vacuum vessel. Sectors are disassembled and reusable parts maintained in hot cells after the plant returns to operation. No rewelding is needed for elements located within the radiation environment. Fig. 6 highlights the significant impact of the horizontal maintenance of a complete sector as a unit on the reference design. The outboard TFcoil leg must be displaced 2.56 m from the back of the vacuum vessel to provide the necessary space for single-piece removal of the breeding blanket and structural ring out through the access tunnel between adjacent TF coils. As the TF coil location is moved further from the vacuum vessel to accommodate the horizontal maintenance scheme, the PF coils are also displaced further from the plasma, thereby increasing the PF-coils current and mass. To minimize cost penalties associated with a larger TF coil (i.e. increased TF and PF coil costs), the TF coil was deformed from a constant-tension shape, as is shown in Fig. 6. Although the TF coils are supported by a cap, the deformation is limited by the magnitude of the bending stress that can be tolerated across the TF-coil structure. The most severe penalty of single-piece sectors is the increased size of both the TF and PF coil systems, needed to allow adequate space for sector removal. With an optimized design, the cost of the increased size of the TF and PF coils is 2 3 mill (kw h) 1, which is substantially smaller than the cost savings due to the increased availability.

13 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) Fig. 6. The elevation view of the ARIES-RS fusion power core First wall, blanket and shield The ARIES-RS design uses a self-cooled lithium design with vanadium alloy as the structural material. It was judged by the ARIES team that this blanket has the best potential to fulfil the plant requirements with a moderate extrapolation of today s technology. The V-alloy has low activation, low afterheat, high temperature capability and can handle high heat flux. Also, this blanket gives excellent neutronics performance. One main concern with liquid-metal cooled systems is the MHD pressure drops. In typical tokamak power plant designs, the MHD pressure drop for the inboard regime of a tokamak results in pressure stresses beyond the material limits [12]. To reduce this MHD pressure drop, some form of insulating wall is required, either with a direct coating or with a sandwich-type structure. With the assumption of reliable insulating coatings, the MHD pressure drop is no longer a major concern. The design of the first wall, blanket, and shield can be optimized to improve heat transfer and to simplify the configuration. The first wall and breeding blanket use a simple box-like structure, with lithium coolant flowing in poloidal paths. The outboard cross section is shown in Fig. 8. The design uses an insulator coating such as calcium oxide on all coolant channel interior surfaces. This can be achieved by adding 0.5% Ca in the flowing Li. In principle, this insulator can be replaced by any other self-healing insulator without any impact on the overall performance of the design. The development of insulating coatings is at a very early stage and much more R&D is required. The improvements and design flexibility that coatings provide for self-cooled liquid metal blankets make this a very high leverage item.

14 16 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) 3 25 Fig. 7. The elevation view of an integrated ARIES-RS sector. Table 6 summarizes the heat loads and peak temperatures in the blanket and divertor. Multiple flow passes in the blanket provide the capability for removing up to 1 MW m 2 of surface heat flux, although the radiative divertor mode of operation of ARIES-RS results in a peak of only 0.4 MW m 2. The full coolant flow is passed first through the front zone, where the surface heat flux creates large temperature gradients, and then through the back zones were the bulk temperature can be raised by volumetric heating without exceeding any structure temperature limits. Segmentation of the shield into a hot and a cold zone allows partial utilization of the heat deposited, and also provides further capability for superheating the coolant away from the high heat flux region. An important feature of the blanket and shield is the radial segmentation into four zones: two blanket zones and two shield (high- and low-temperature) regions to maximize the lifetime of the structures, reduce the replacement cost, and minimize the waste stream. Scheduled replacement occurs after 2.5 full power years (FPY) when the V-alloy reaches 200 dpa (for ARIES-RS, this translates to 15 MW m 2 ). At that time, the front portion of the blanket is disposed, but the rear portion of the blanket can be used until it reaches its own radiation lifetime, at about 7.5 FPY. This rear portion also serves as the structural ring, which provides poloidal continuity to the sectors and attachment points for the inner blanket segments. Location of the coolant connections outside the vacuum vessel allows for easy disassembly of the segments. Radial segmentation creates safety concerns, since radial heat transport pathways are critical in loss-of-coolant scenarios. Design solutions have been proposed; however, blanket response to coolant loss remains an important concern. Detailed shielding analysis has raised no serious difficulties with ARIES-RS shields. A dedicated effort was devoted to the bulk shield in particular, as it represents a major cost item for advanced tokamak designs. Significant savings in shield cost were obtained by these strategies: (1) Expensive V

15 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) Fig. 8. The cross section of the ARIES-RS outboard blanket and shield. alloy is limited to those regions where high temperature operation is absolutely necessary. (2) Shield is segmented into high temperature and low temperature zones. The nuclear heat deposited in the high-temperature shield ( 20% of total heating) is recovered as high grade heat to enhance the overall power balance of the machine. The low-temperature shield, vacuum vessel, and other external components (having low levels Table 6 Power flows and peak temperatures Multiplied neutron heating (MW) 2092 Total transport power (MW) 431 Bremsstrahlung power (MW) 56 Core line radiation (MW) 25 Power reradiated in divertor (MW) 341 First wall surface heating (MW) 165 Divertor total surface heating (MW) 348 Divertor particle power (MW) 88 Blanket bulk outlet temperature ( C) 610 Divertor bulk outlet temperature ( C) 610 Peak V temperature in first wall ( C) Peak V temperature in divertor ( C) of nuclear heating) can employ cheap stainless steel as the main structural material, instead of V. (3) Cheap steel filler (rather than V filler) in the shield reduces the cost tremendously. Fillers have no structural role and thus have lower unit costs compared to structures. (4) Highly efficient, expensive materials, such as WC and B 4 C, are used only in the space-constrained inboard side to reduce the overall size and cost of the machine while less efficient, cheaper materials is used in the divertor and outboard sides. The ARIES-RS design uses both active and passive stabilization systems for vertical displacement and kink-mode stabilization. These systems were integrated into the sectors (Figs. 1 and 7). Radiatively-cooled tungsten shells are located between the blanket and shield on the inboard and outboard sides to provide passive stabilization of vertical displacement modes. They are augmented by actively-cooled coils which are placed outboard between the low-temperature shield and the vacuum vessel. A thickened vanadium second wall behind the first wall cooling channel was shown to provide sufficient conductivity to stabilize kink modes.

16 18 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) 3 25 Fig. 9. The ARIES-RS divertor. The design of the primary loop and power conversion system are critical to the attractiveness of the power plant. This system is responsible for efficient power conversion, isolation of radioactive products within the nuclear island, and reliable operation of the power plant. The choice of an advanced Rankine cycle conservatively offers 46% gross thermal conversion efficiency. High thermal efficiency is desirable to partially offset the high capital cost of fusion. A double-walled intermediate heat exchanger (IHX) with an Na secondary loop is used to isolate the activated Li primary coolant from the stream side. The IHX is also the location where the transition from V to stainless steel is made. The piping which connects the blanket to the IHX uses a double-walled structure with a thin V liner to minimize the added cost of vanadium. The tritium recovery process proposed here is based on cold trapping. For the liquid lithium system, the hydrogen solubility at the cold trap temperature of 200 C is 440 appm, which is far above the design goal of 1 appm. For this reason, cold traps have not been considered previously as a candidate process for recovering tritium from lithium. However, the concept developed here is modified to add protium in the lithium so that the total hydrogen concentration in the lithium is higher than the 200 appm saturation value. At 200 C, Li(H+T) will be supersaturated and precipitate out together. The Li(T+H) can be separated from lithium using a meshless cold trap process which was developed by the breeder program to separate NaH from Na by gravitational force. The Li(T+H) can then be heated up to 600 C for decomposition. The hydrogen stream will then be fed to the main isotope separation system (ISS) to separate tritium from protium Di ertor The divertor region of the sector, highlighted in Fig. 9, consists of two principal parts: the target plates and the structures. The structures fulfil several essential functions: (1) it provides for mechanical attachment of the plates through adjustable screw-bolts which provide for module alignment and offer lateral flexibility for thermal expansion together with strong support against EM events (e.g. disruptions); (2) it shields the magnets; (3) it provides coolant routing paths for the plates as well as the inboard blanket and replaceable shield; (4) divertor-plate coolant is routed through this region and is superheated.

17 F. Najmabadi et al. / Fusion Engineering and Design 38 (1997) Low outlet temperature of divertor plate coolant allows optimization of surface heat removal while a high outlet temperature for the divertor system is maintained; and (5) it contributes to the breeding ratio, since the coolant is Li. The target plates include three pieces: inboard, outboard, and dome plates. The plasma flows through the scrape-off layer and enters the divertor, where enhanced line radiation from injected neon impurity allows much of the power to be distributed along the plates and also partially redirected out to the first wall. Most of the unradiated particle energy strikes the outboard plates, but the peak heat flux (Combined Radiation and Transport) has been maintained below 6 MW m 2. The strike points are located close to the coolant inlet in order to maintain the vanadium structures below 700 C. A 2-mm thick castellated W coating is applied to the coolant channel front surface, which is only 1-mm thick V to satisfy temperature limits. Thermal stresses are reduced by using a relatively thick solid back on the target plates. The plates are connected to the rear zone via strong adjustable screw-type attachments. These attachments can be designed to react to the full force of disruptions and also accommodate thermal expansion. They also permit precise alignment to adjacent surfaces and removal of individual plates in the hot cells. Vacuum pumping ducts are placed behind the dome near the strike points for efficient exhaust. Radial channels then direct the gas to a single set of cryopumps at the bottom of the machine. Sufficient conductance between top and bottom divertor pumping ports are provided by using the inter-sector vessel volume underneath the TF coils Magnet systems The ARIES-RS toroidal field (TF) coil set consists of 16 coils using multi-filamentary Nb 3 Sn and NbTi superconductors with a peak field of 15.8 T at the coil. The TF coils are flattened vertically away from a constant-tension D-shape (Fig. 6) in order to reduce the size of both TF and poloidalfield (PF) coils and the peak fields. The design optimized the superconductor, copper, helium, insulation, and structural ratios by using four grades of conductor. Advanced magnet design techniques maximize the utilization of structural materials by using structural forms with grooves into which the conductor is wound. Rather than winding all of the materials in the magnet, only the conductor requires winding. Support of out-of-plane loads has been provided without intercoil structure in the outer legs of the TF coils, using caps and outer straps. This makes full-sector maintenance possible. The cap and strap structures also have been shown to accommodate off-normal events (e.g. a single coil short during a dump) and the bending stresses. The PF coil set consists of 22 coils: eight form the center stack, and the remaining 14 elongate the plasma, provide equilibrium and form the divertor magnetic configuration. An attempt has been made to keep the field at the PF coils 8 T, such that less expensive NbTi conductors can be used. This is accomplished by shaping the TF coils and by allowing a larger number of closely-spaced PF coils Heating and current dri e systems The three RF current-drive systems make use of ICRF fast waves (98 MHz), high-frequency fast waves (1.0 GHz), and lower hybrid waves ( GHz) to drive currents in the on-axis, off-axis (inside the reversed-shear region), and edge regions, respectively. The power from these systems can be used to heat the plasma from start-up to its final operating conditions. The total power delivered to the plasma from these systems is 102 MW, of which 81 MW is used in driving useful currents. The wave launcher and transmission systems are designed and configured so as to minimize intrusions into the periodic blanket and shield structures. All RF launchers can fit with a single blanket sector occupying 0.58% of the first wall area. As a result, the engineering impact are modest, and the effects on shielding and waste disposal are manageable. The launchers are located around the outboard mid-plane. The transmission lines and waveguides are routed behind the shielding structures to minimize neutron streaming and irradiation of ex-vessel components.