《复合材料 Composites》课程教学资源（学习资料）第五章 陶瓷基复合材料_SiC-SiC-15

Fusion Engineering and design ELSEVIER Fusion Engineering and Design 61-62(2002)307-318 Progress in blanket designs using SiCSic composites L. Giancarli,*, H. Golfier, S. Nishio, R. Raffray, C. Wong, R. Yamada CEA-Saclay, DEN/CPT, 91191 Gif-sur-Yuette, france CEA-Saclay, DENIDMTSERMA, 91191 Gif-sur- Yrette, Fran JAERI, Depa f Fusion Plasma Research, Naka Fusion Research Establishment, 801- Mukouyama, Naka-imachi, Naka-gun Ibaraki-kenn 311-0193, Japan 458 EBU-l1, University of California, San Diego, 9500 Gilman Drice, La Jolla, CA 92093-0417, US.A General Atomics. San Diego, CA 92186-9784. USA This paper summarizes the most recent design activities concerning the use of SicSic composite as structural material for fusion power reactor breeding blanket. Several studies have been performed in the past. The most recent proposals are the tauRo blanket concept in the European Union, the ARIES-AT concept in the US, and DREAM concept in Japan. The first two concepts are self-cooled lithium-lead blankets, while DREAM is an helium-cooled beryllium/ceramic blanket. Both TAURO and aries-at blankets are essentially formed by a sicsic box acting as a container for the lithium-lead which has the simultaneous functions of coolant, tritium breeder, neutron multiplie and, finally, tritium carrier. The dream blanket is characterized by small modules using pebble beds of Be as neutron multiplier material, of Li2O (or other lithium ceramics) as breeder material and of Sic as shielding material. The He coolant path includes a flow through the pebble beds and a porous partition wall. For each blanket, this paper describes the main design features and performances, the most recent design improvements, and the proposed manufacturing routes in order to identify specific issues and requirements for the future R&d on SiCr/Sic c 2002 Published by elsevier Science B V. Keywords: Blanket designs: SiC/SiC composites; Self-cooled lithium-lead 1. Introduction The low activation and afterheat levels asso- ciated with SiC!Sic after long-term neutron The use of SiCdSic composites as structural irradiation allow a design of FPR nuclear compo material for in-vessel components permits to show nents showing high safety standards and simplifie the potential of D-T fusion power reactors(FPR) maintenance schemes. In addition, SicSiC has in terms of safety and environmental impact excellent chemical stability at high temperature, which minimizes mobilization of radioactive pro- Corresponding author. Tel. +33-1-69-08-21-37: fax: +33 Furthermore, the high temperature properties of 6908-58-61 SiCHSic improve energy handling capabilities, E-mail address: luciano. giancarli(@cea fr (L. Giancarli allowing the use of high temperature coolant 0920-3796/02/S- see front c 2002 Published by Elsevier Science B.V. PI:S0920-3796(02)00213-2

Progress in blanket designs using SiCf/SiC composites L. Giancarli a,
, H. Golfier b , S. Nishio c , R. Raffray d , C. Wong e , R. Yamada c a CEA-Saclay, DEN/CPT, 91191 Gif-sur-Yvette, France b CEA-Saclay, DEN/DMT/SERMA, 91191 Gif-sur-Yvette, France c JAERI, Department of Fusion Plasma Research, Naka Fusion Research Establishment, 801-1 Mukouyama, Naka-machi, Naka-gun, Ibaraki-kenn 311-0193, Japan d 458 EBU-II, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0417, USA e General Atomics, San Diego, CA 92186-9784, USA Abstract This paper summarizes the most recent design activities concerning the use of SiCf/SiC composite as structural material for fusion power reactor breeding blanket. Several studies have been performed in the past. The most recent proposals are the TAURO blanket concept in the European Union, the ARIES-AT concept in the US, and DREAM concept in Japan. The first two concepts are self-cooled lithium/lead blankets, while DREAM is an helium-cooled beryllium/ceramic blanket. Both TAURO and ARIES-AT blankets are essentially formed by a SiCf/SiC box acting as a container for the lithium/lead which has the simultaneous functions of coolant, tritium breeder, neutron multiplier and, finally, tritium carrier. The DREAM blanket is characterized by small modules using pebble beds of Be as neutron multiplier material, of Li2O (or other lithium ceramics) as breeder material and of SiC as shielding material. The He coolant path includes a flow through the pebble beds and a porous partition wall. For each blanket, this paper describes the main design features and performances, the most recent design improvements, and the proposed manufacturing routes in order to identify specific issues and requirements for the future R&D on SiCf/SiC. # 2002 Published by Elsevier Science B.V. Keywords: Blanket designs; SiCf/SiC composites; Self-cooled lithium/lead 1. Introduction The use of SiCf/SiC composites as structural material for in-vessel components permits to show the potential of D/T fusion power reactors (FPR) in terms of safety and environmental impact. The low activation and afterheat levels asso￾ciated with SiCf/SiC after long-term neutron irradiation allow a design of FPR nuclear compo￾nents showing high safety standards and simplified maintenance schemes. In addition, SiCf/SiC has an excellent chemical stability at high temperature, which minimizes mobilization of radioactive pro￾ducts. Furthermore, the high temperature properties of SiCf/SiC improve energy handling capabilities, allowing the use of high temperature coolant
Corresponding author. Tel.: /33-1-69-08-21-37; fax: /33- 1-69-08-58-61 E-mail address: luciano.giancarli@cea.fr (L. Giancarli). Fusion Engineering and Design 61/62 (2002) 307/318 www.elsevier.com/locate/fusengdes 0920-3796/02/$ - see front matter # 2002 Published by Elsevier Science B.V. PII: S 0 9 2 0 - 3 7 9 6 ( 0 2 ) 0 0 2 1 3 - 2

L Giancarli et al Fusion Engineering and Design 61-62(2002)307-318 with the potential for high energy conversion neutron multiplier and breeder, respectively, and efficiency(>50%) SiCSiC for joints(e.g. bolts). The difficulty in this Starting from the different strategies which can case is to keep the same low-activation require- be adopted for FPR safety and from the r&d ment for all the other in-vessel components next section, this paper presents an assessment of reactivity, low afterheat materials should be used the most recent proposals of breeding blanket This strategy has been adopted by the TauRo designs with particular focus on fabrication issues. and ARIES-AT blanket designs, which use low In particular, two self-cooled lithium-lead (SCLL) pressure, low reactivity, low afterheat Pb-17Li as blanket designs, ARIES-AT [ and TAURO [2]. coolant, neutron multiplier and breeder. The and one helium-cooled be/ceramic(HCBC)design, difficulty in this case is to fulfill the same require- DREAM 33], will be considered ments for the other in-vessel components which imply for instance the use Pb-17Li(or equivalent) as coolant for divertor and shield 2. Attractiveness and development risks for SiCr Sic structures 2. 2. High plant efficiency The attractiveness of Fpr breeding blankets using SiCSic structures is based on the achiev- Maximum acceptable working temperature of able high safety standards and high plant effi- SiC/Sic under irradiation is about 1000C. de- ciency. These significant advantages of SiC!Sic have been developed with the aim of exploiting this favorable feature for having high compared to other structural materials can be fully coolant outlet temperature and, as a consequence, exploited by Dy making coherent design choices concerning the other materials required in the high overall plant efficiency. Moreover, high blanket temperature coolant gives the potential of an efficient hydrogen production in combination 2.1. High safety standards with the standard electricity production. The three designs considered in this paper have High safety standards can be potentially aid particular attention to this aspect. In part achieved because of the low short term activation cular, He-coolant outlet temperature in DREAM blanket is about 900C leading to a net thermal and decay heat which minimize accidental releases, efficiency greater than 45%. For the Tauro facilitates the accommodation of loss-of-coolant (LOCA)and loss-of-flow (LOFA)events, and blanket the Pb-17Li parameters have been opti- simplifies maintenance procedure mized in order to reach an outlet temperature of In particular, in order to limit to an acceptable about 950C and a corresponding net thermal level the accidental release of activation products efficiency of about 55%. In case of ARIES-At the two different strategies can be envisaged (4), that choice of having an annular Pb-17Li flow allows is. either to minimize the in-vessel overall activa to reach an outlet temperature of about 1100C tion inventory and control the release, or to leading to a net thermal efficiency as high as 58.5% minimize the available energy within the safety vessel and keep the activation products confined In the first case, all materials present within the 2.3. R&d requirements and development risks vessel should have low activation characteristics and for the SiCsic a minimization of the Present-day SiCASic composites are not ade- impurity contents should be pursued. This strategy quate to be used directly as structure of nuclear has been adopted by the dream blanket design, components. A comparison between measured which uses only low-activation materials, such as properties on present-day Sic!SiC and require- high-pressure He as coolant, and Be and Li2O ments are given in Table 1. In fact, there are some

with the potential for high energy conversion efficiency (
/50%). Starting from the different strategies which can be adopted for FPR safety and from the R&D needs for SiCf/SiC structures, summarized in the next section, this paper presents an assessment of the most recent proposals of breeding blanket designs with particular focus on fabrication issues. In particular, two self-cooled lithium/lead (SCLL) blanket designs, ARIES-AT [1] and TAURO [2], and one helium-cooled be/ceramic (HCBC) design, DREAM [3], will be considered. 2. Attractiveness and development risks for SiCf/ SiC structures The attractiveness of FPR breeding blankets using SiCf/SiC structures is based on the achiev￾able high safety standards and high plant effi￾ciency. These significant advantages of SiCf/SiC compared to other structural materials can be fully exploited by making coherent design choices concerning the other materials required in the blanket. 2.1. High safety standards High safety standards can be potentially achieved because of the low short term activation and decay heat which minimize accidental releases, facilitates the accommodation of loss-of-coolant (LOCA) and loss-of-flow (LOFA) events, and simplifies maintenance procedure. In particular, in order to limit to an acceptable level the accidental release of activation products, two different strategies can be envisaged [4], that is, either to minimize the in-vessel overall activa￾tion inventory and control the release, or to minimize the available energy within the safety vessel and keep the activation products confined. In the first case, all materials present within the vessel should have low activation characteristics and for the SiCf/SiC a minimization of the impurity contents should be pursued. This strategy has been adopted by the DREAM blanket design, which uses only low-activation materials, such as high-pressure He as coolant, and Be and Li2O as neutron multiplier and breeder, respectively, and SiCf/SiC for joints (e.g. bolts). The difficulty in this case is to keep the same low-activation require￾ment for all the other in-vessel components. In the second case, only low pressure, low reactivity, low afterheat materials should be used. This strategy has been adopted by the TAURO and ARIES-AT blanket designs, which use low pressure, low reactivity, low afterheat Pb/17Li as coolant, neutron multiplier and breeder. The difficulty in this case is to fulfill the same require￾ments for the other in-vessel components which imply for instance the use Pb/17Li (or equivalent) as coolant for divertor and shield. 2.2. High plant efficiency Maximum acceptable working temperature of SiCf/SiC under irradiation is about 1000 8C. De￾signs have been developed with the aim of exploiting this favorable feature for having high coolant outlet temperature and, as a consequence, high overall plant efficiency. Moreover, high temperature coolant gives the potential of an efficient hydrogen production in combination with the standard electricity production. The three designs considered in this paper have paid particular attention to this aspect. In parti￾cular, He-coolant outlet temperature in DREAM blanket is about 900 8C leading to a net thermal efficiency greater than 45%. For the TAURO blanket the Pb/17Li parameters have been opti￾mized in order to reach an outlet temperature of about 950 8C and a corresponding net thermal efficiency of about 55%. In case of ARIES-AT the choice of having an annular Pb/17Li flow allows to reach an outlet temperature of about 1100 8C leading to a net thermal efficiency as high as 58.5%. 2.3. R&D requirements and development risks Present-day SiCf/SiC composites are not ade￾quate to be used directly as structure of nuclear components. A comparison between measured properties on present-day SiCf/SiC and require￾ments are given in Table 1. In fact, there are some 308 L. Giancarli et al. / Fusion Engineering and Design 61/62 (2002) 307/318

L Giancarli et al. Fusion Engineering and Design 61-62(2002)307-318 309 Table I Comparison between SiCSic properties assumed in the analysis and typical measured values on present-day industrial composites Key SiC/Sic properties and parameters SCLL blankets(agreed drEAM blanket Typical measured value 3000kg/m3 2500kg/m3 ≈2500kg/m Porosity 00-300GI 200 GPa 0.16-0.18 20 0.18 pansion coefficient 4×10-°FC 4×10-6FC Thermal conductivity in plane(1000 C) ≈20WmK(EOL) 15 and 60 W/m k R 15 WIm K(BOL) (EOL) Thermal conductivity through thickness (1000 C) <20 W/m K ( EOL) 15 and 60 W/m K ≈7.5W/mK(BOL (EOL) Electrical conductivity 500/Q2 m(under irradia- Not applicable 500/@2m(out of irradia Tensile strength 300 MPa 300 MPa 300 MPa Trans-laminar shear strength 200 MPa Inter-laminar shear strength 44 MPa Maximum allowable tensile stress Not used 200 MPa Maximum allowable temperature(swelling basis) 1000 C ≈1100°C Maximum allowable interface temperature with 1000°c( nowing) 800°C( statIc) breeder Minimum allowable temperature( thermal conduc- 600C 600° tivity basis) Cost ≤S400/kg ≈10 times larger Assumed design criteria are slight different for each design. They are given in the appropriate chapters. No validated experimental data are yet available key nfluencing its attractiveness, which can development, testing and validation of accep be identified as 'development risks' and which table joining techniques. Different joining tech define the required r&d program. Most R&D niques can be envisaged: (i) assembling by requirements on SiCSic are common to both He- sewing at textile stage to join the stiffeners to cooled and Pb-l7Li cooled systems [5]. The most the side walls; (ii) sticking and co-infiltration to Important common requirements are join the second wall to the first stiffener;(iii) brazing of finished components to join the improvement of thermal conductivity, espe bottom and the top closure plates and the cially through the thickness, at high tempera different sub modules. A promising brazing under technique using a braze material compatible determination and possible improvement of with SiC, the Brasic@, is currently under devel- maximum working temperature under irradia- 46 tion(swelling, compatibility ); development and validation of appropriate de- Specific r&d items concerning SCll blankets sign criteria (e.g. maximum allowed stresses) are which could ensure reasonable component re liability; determination of the electrical conductivity determination and improvement of the lifetime under irradiation capability of fabrication of components with establishment of the maximum interface tem homogeneous properties and reasonable dimen- perature with Pb-17Li under representative sions, with particular attention to the minimum flowing conditions and irradiation level; in particular verification that no Pb-17Li infiltra

key issues influencing its attractiveness, which can be identified as ‘development risks’ and which define the required R&D program. Most R&D requirements on SiCf/SiC are common to both He￾cooled and Pb/17Li cooled systems [5]. The most important common requirements are: . improvement of thermal conductivity, espe￾cially through the thickness, at high tempera￾ture and under neutron irradiation; . determination and possible improvement of maximum working temperature under irradia￾tion (swelling, compatibility); . development and validation of appropriate de￾sign criteria (e.g. maximum allowed stresses) which could ensure reasonable component re￾liability; . determination and improvement of the lifetime; . capability of fabrication of components with homogeneous properties and reasonable dimen￾sions, with particular attention to the minimum and maximum thickness; . development, testing and validation of accep￾table joining techniques. Different joining tech￾niques can be envisaged: (i) assembling by sewing at textile stage to join the stiffeners to the side walls; (ii) sticking and co-infiltration to join the second wall to the first stiffener; (iii) brazing of finished components to join the bottom and the top closure plates and the different sub modules. A promising brazing technique using a braze material compatible with SiC, the Brasic†, is currently under devel￾opment [4,6]. Specific R&D items concerning SCLL blankets are: . determination of the electrical conductivity under irradiation; . establishment of the maximum interface tem￾perature with Pb/17Li under representative flowing conditions and irradiation level; in particular verification that no Pb/17Li infiltra￾Table 1 Comparison between SiCf/SiC properties assumed in the analysis and typical measured values on present-day industrial composites Key SiCf/SiC properties and parametersa SCLL blankets (agreed values) DREAM blanket Typical measured value Density :/3000 kg/m3 2500 kg/m3 :/2500 kg/m3 Porosity :/5% :/10% :/10% Young’s modulus 200/300 GPa :/200 GPa :/200 GPa Poisson’s ratio 0.16/0.18 0.20 0.18 Thermal expansion coefficient :/4/106 /8C 3.3/106 /8C 4/106 /8C Thermal conductivity in plane (1000 8C) :/20 W/m K (EOL) 15 and 60 W/m K (EOL) :/15 W/m K (BOL) Thermal conductivity through thickness (1000 8C) :/20 W/m K (EOL) 15 and 60 W/m K (EOL) :/7.5 W/m K (BOL) Electrical conductivity :/500/V m (under irradia￾tion) Not applicable :/500/Vm (out of irradia￾tion) Tensile strength 300 MPa 300 MPa 300 MPa Trans-laminar shear strength / / 200 MPa Inter-laminar shear strength / / 44 MPa Maximum allowable tensile Stress Not useda 200 MPaa Unknowna Maximum allowable temperature (swelling basis) :/1000 8C :/1100 8C :/1000 8C Maximum allowable interface temperature with breeder :/1000 8C (flowing) / :/800 8C (static) Minimum allowable temperature (thermal conduc￾tivity basis) :/600 8C :/600 8C :/600 8C Cost 0/$400/kg / :/10 times larger a Assumed design criteria are slight different for each design. They are given in the appropriate chapters. No validated experimental data are yet available. L. Giancarli et al. / Fusion Engineering and Design 61/62 (2002) 307/318 309

310 L Giancarli et al. Fusion Engineering and Design 61-62(2002)307-318 tion through the SiC SiC surface will occur, the technologies and of physics understanding major risk being an increase of the wall and modeling capa on the performance of electrical conductivity advanced tokamak plants [7]. The blanket compatibility of brazing material with Pb-17Li. design was developed to achieve high performance while maintaining attractive safety features, simple Specific r&d items concerning HCBC blankets design geometry, credible maintenance and fabri cation processes, and reasonable design margins as hermeticity to high-pressure Helium; an indication of reliability [1] compatibility with Be and Li2O The Pb-I7Li operating temperature is opti- mized to provide high power cycle efficiency while Most of these issues were addressed in detail in maintaining the SiCrSic temperature under rea- presentations and discussions at the January 2000 sonable limits. The Brayton cycle offers the best International Town Meeting on SiC/SiC Design near-term possibility of power conversion with and Material Issues for Fusion Systems and in a high efficiency and is chosen to maximize the related publication [ 5] potential gain from high temperature operation of the Pb-17Li which after exiting the blanket is routed through a heat exchanger with the cycle He 3. Self-cooled Pb-17Li blankets as secondary fluid [8]. The maximum He cycle temperature is 1050C, resulting in a high cycle The safety strategy for SCll blankets is based efficiency of about 58.5% on the minimization of the energy inventory in the The Sic/sic parameters and properties used in vessel. This strategy, in principle, allows the use of he ARIEs-AT analysis are summarized in Table materials for in-vessel components which are not I. For thermo-mechanical analyses it has been low activation(except for long-term waste man- assumed that the maximum allowed combined agement considerations) and which could be of stress(primary and thermal stresses)is 190 MPa particular interest for developing high erior mance joining techniques or for designing other 3.1.2. Blanket description components(such as divertor or shielding) For waste minimization and cost saving reasons. scli blankets use the eutectic Pb-17Li whose the blanket is subdivided radially into two zones: a eplaceable first zone in the inboard and outboard melting point is 235C. Because of the high and a life of plant second zone in the outboard To coolant temperature, it is probably necessary to have the whole coolant circuit made of ceramics simplify the cooling system and minimize the composites(SiC,SiC or equivalent) and it is then number of coolants the pb-17Li is used to cool required to develop specific Pb-17Li/helium heat the blanket as well as the divertor and hot shield exchanger regions. As illustrated in Fig. I and Fig. 2 for the Among advantages one can also note the outboard region, the blanket design is modular and consists of an assembly of simple annular relatively easy tritium extraction to be performed boxes through which the Pb-17Li flows in two outside the reactor and the use of only two basic materials,SiC/SiC and the liquid Pb-17Li which. poloidal passes. Positioning ribs are attached to at least in theory, should allow to reach good he inner annular wall forming a free-floating liability assembly inside the outer wall. These ribs divide he annular region into a number of channels through which the coolant first flows at high 3.1. ARIES-AT blanket velocity to keep cooled both inner and outer walls The coolant then makes a U-turn and flows very 3. .1. General background slowly as a second pass through the large inner The ARIES-at power plant was evolved channel from which the Pb-17Li exits at high assess and highlight the benefit of advanced temperature. This flow scheme enables operatin

tion through the SiCf/SiC surface will occur, the major risk being an increase of the wall electrical conductivity; . compatibility of brazing material with Pb/17Li. Specific R&D items concerning HCBC blankets are: . hermeticity to high-pressure Helium; . compatibility with Be and Li2O. Most of these issues were addressed in detail in presentations and discussions at the January 2000 International Town Meeting on SiCf/SiC Design and Material Issues for Fusion Systems and in a related publication [5]. 3. Self-cooled Pb/17Li blankets The safety strategy for SCLL blankets is based on the minimization of the energy inventory in the vessel. This strategy, in principle, allows the use of materials for in-vessel components which are not low activation (except for long-term waste man￾agement considerations) and which could be of particular interest for developing high-perfor￾mance joining techniques or for designing other components (such as divertor or shielding). SCLL blankets use the eutectic Pb/17Li whose melting point is 235 8C. Because of the high coolant temperature, it is probably necessary to have the whole coolant circuit made of ceramics composites (SiCf/SiC or equivalent) and it is then required to develop specific Pb/17Li/helium heat exchanger. Among the advantages one can also note the relatively easy tritium extraction to be performed outside the reactor and the use of only two basic materials, SiCf/SiC and the liquid Pb/17Li which, at least in theory, should allow to reach good reliability. 3.1. ARIES-AT blanket 3.1.1. General background The ARIES-AT power plant was evolved to assess and highlight the benefit of advanced technologies and of new physics understanding and modeling capabilities on the performance of advanced tokamak power plants [7]. The blanket design was developed to achieve high performance while maintaining attractive safety features, simple design geometry, credible maintenance and fabri￾cation processes, and reasonable design margins as an indication of reliability [1]. The Pb/17Li operating temperature is opti￾mized to provide high power cycle efficiency while maintaining the SiCf/SiC temperature under rea￾sonable limits. The Brayton cycle offers the best near-term possibility of power conversion with high efficiency and is chosen to maximize the potential gain from high temperature operation of the Pb/17Li which after exiting the blanket is routed through a heat exchanger with the cycle He as secondary fluid [8]. The maximum He cycle temperature is 1050 8C, resulting in a high cycle efficiency of about 58.5%. The SiCf/SiC parameters and properties used in the ARIES-AT analysis are summarized in Table 1. For thermo-mechanical analyses it has been assumed that the maximum allowed combined stress (primary and thermal stresses) is 190 MPa. 3.1.2. Blanket description For waste minimization and cost saving reasons, the blanket is subdivided radially into two zones: a replaceable first zone in the inboard and outboard, and a life of plant second zone in the outboard. To simplify the cooling system and minimize the number of coolants, the Pb/17Li is used to cool the blanket as well as the divertor and hot shield regions. As illustrated in Fig. 1 and Fig. 2 for the outboard region, the blanket design is modular and consists of an assembly of simple annular boxes through which the Pb/17Li flows in two poloidal passes. Positioning ribs are attached to the inner annular wall forming a free-floating assembly inside the outer wall. These ribs divide the annular region into a number of channels through which the coolant first flows at high￾velocity to keep cooled both inner and outer walls. The coolant then makes a U-turn and flows very slowly as a second pass through the large inner channel from which the Pb/17Li exits at high temperature. This flow scheme enables operating 310 L. Giancarli et al. / Fusion Engineering and Design 61/62 (2002) 307/318

L Giancarli et al. Fusion Engineering and Design 61-62(2002)307-318 3.1.3. Analysis Detailed analyses of the ARIES-at blanket were performed and the results are summarized are summarized below [1I A tritium-breeding ratio of 1. I was calculated from 3D neutronics analyses of the power core. Thermal-hydraulic analyses conservatively as- MHD-laminarized Pb-17Li flo showed that for an average outlet Pb-17Li emperature of 1100C, both the maximum SiC!Sic temperature at the FW and the max imum blanket SiC/Pb-17Li interface tempera ture at the inner channel wall are maintained at 1000C, which satisfy the maximum tem perature limits shown in Table 1. The corre sponding blanket pressure drop is about 0.25 Fig. 1. ARIES-AT outboard first wall and blanket segment. Stress analyses were performed both on the module outer and inner shells indicating that Pb-17Li at a high outlet temperature (1100C) the maximum combined stress in all cases is less while maintaining the blanket SiC/SiC composite than the assumed conservative limit of 190 and Sic/Pbli interface at a lower temperature MPa, often with significant margin(as a 1000C). The first wall consists of a 4-mm SiCd positive measure of reliability) SiC structural wall on which a 1-mm chemical The activation, decay heat, and waste disposal vapor deposition (CVD) Sic armor layer is analyses performed in support of the ARIF deposited at design are described in Ref [9]. The decay First Wall R685 Fig. 2. Cross-section of ARIES-AT outboard blanket segm

Pb/17Li at a high outlet temperature (1100 8C) while maintaining the blanket SiCf/SiC composite and SiC/PbLi interface at a lower temperature (/ 1000 8C). The first wall consists of a 4-mm SiCf/ SiC structural wall on which a 1-mm chemical vapor deposition (CVD) SiC armor layer is deposited. 3.1.3. Analysis Detailed analyses of the ARIES-AT blanket were performed and the results are summarized are summarized below [1]: . A tritium-breeding ratio of 1.1 was calculated from 3D neutronics analyses of the power core. . Thermal/hydraulic analyses conservatively as￾suming MHD-laminarized Pb/17Li flow showed that for an average outlet Pb/17Li temperature of 1100 8C, both the maximum SiCf/SiC temperature at the FW and the max￾imum blanket SiC/Pb/17Li interface tempera￾ture at the inner channel wall are maintained at /1000 8C, which satisfy the maximum tem￾perature limits shown in Table 1. The corre￾sponding blanket pressure drop is about 0.25 MPa. . Stress analyses were performed both on the module outer and inner shells indicating that the maximum combined stress in all cases is less than the assumed conservative limit of 190 MPa, often with significant margin (as a positive measure of reliability). . The activation, decay heat, and waste disposal analyses performed in support of the ARIES￾AT design are described in Ref. [9]. The decay Fig. 1. ARIES-AT outboard first wall and blanket segment. Fig. 2. Cross-section of ARIES-AT outboard blanket segment. L. Giancarli et al. / Fusion Engineering and Design 61/62 (2002) 307/318 311

312 L Giancarli et al. Fusion Engineering and Design 61-62(2002)307-318 heat results were used to perform 2D safety Pb-17Li acts as coolant, breeder, neutron multi nalyses of the power core, which showed that plier, and tritium carrier. Each inboard and out the low decay heat of SiC enables accommoda- board segment is poloidally divided into several tion of any LOCA or LOFA scenarios without straight modules attached on one common thick serious consequences to the blanket structure back plate but cooled independently. The coolant enters from a backside collector and flows in parallel into five channels for cooling five corre 3. 1. 4. Fabrication and maintenance sponding sub-modules as shown in Fig 3. In each As a reliability measure, minimization of the sub-modules the Pb-17Li flows first through a number an nd length of brazes was a major factor in thin channel at high velocity for cooling the Fw evolving the fabrication procedure for the blanket. and then slows down for cooling in series the rest The proposed fabrication scheme requires three- of the box radial/toroidal coolant-containment brazes per In the latest reference design [2], whose design module, as illustrated by the following fabrication point is summarized in Table 2, the Pb-17Li outlet steps for an outboard segment consisting of 6 temperature is 950"C which leads to an estimated efficiency of approximately 55% for a compression ratio of about 1.77 1) Manufacturing separate halves of the SiCHSiC poloidal module by SiCr weaving and Sic chemical vapor infiltration(Cvi) or polymer 3.2.2. Assumed boundary conditions The blanket design optimization and structural 2)Inserting the free-floating inner separation assessment of the tauro blanket were per all in each half module 3) Brazing the two half modules together at the Pb 4) Brazing the module end cap 5)Forming a segment by brazing six modules together (this is a joint which is not in contact with the coolant); and 6) Brazing the annular manifold connections to one end of the segment Note that if handling size is an issue the fabrication steps could also proceed with smaller poloidal units but with additional joining steps Maintenance methods have been investigated which allow for end-of-life replacement of indivi- dual components. These are discussed in Ref [ll] 3.2 TAURO Blanket 3.2.1. Blanket description The tauRo blanket offers the capability of heat extraction at high coolant temperatures and promises favorable conversion efficiencies. The TAURO blanket is essentially formed by a SiCp Sic stiffened box with an indirectly cooled F which acts as a container for the pb-17Li. The Fig 3. TAURO outboard blanket module

heat results were used to perform 2D safety analyses of the power core, which showed that the low decay heat of SiC enables accommoda￾tion of any LOCA or LOFA scenarios without serious consequences to the blanket structure [10]. 3.1.4. Fabrication and maintenance As a reliability measure, minimization of the number and length of brazes was a major factor in evolving the fabrication procedure for the blanket. The proposed fabrication scheme requires three￾radial/toroidal coolant-containment brazes per module, as illustrated by the following fabrication steps for an outboard segment consisting of 6 modules: 1) Manufacturing separate halves of the SiCf/SiC poloidal module by SiCf weaving and SiC chemical vapor infiltration (CVI) or polymer process; 2) Inserting the free-floating inner separation wall in each half module; 3) Brazing the two half modules together at the midplane; 4) Brazing the module end cap; 5) Forming a segment by brazing six modules together (this is a joint which is not in contact with the coolant); and 6) Brazing the annular manifold connections to one end of the segment. Note that if handling size is an issue, the fabrication steps could also proceed with smaller poloidal units but with additional joining steps. Maintenance methods have been investigated which allow for end-of-life replacement of indivi￾dual components. These are discussed in Ref. [11]. 3.2. TAURO Blanket 3.2.1. Blanket description The TAURO blanket offers the capability of heat extraction at high coolant temperatures and promises favorable conversion efficiencies. The TAURO blanket is essentially formed by a SiCf/ SiC stiffened box with an indirectly cooled FW which acts as a container for the Pb/17Li. The Pb/17Li acts as coolant, breeder, neutron multi￾plier, and tritium carrier. Each inboard and out￾board segment is poloidally divided into several straight modules attached on one common thick back plate but cooled independently. The coolant enters from a backside collector and flows in parallel into five channels for cooling five corre￾sponding sub-modules as shown in Fig. 3. In each sub-modules the Pb/17Li flows first through a thin channel at high velocity for cooling the FW and then slows down for cooling in series the rest of the box. In the latest reference design [2], whose design point is summarized in Table 2, the Pb/17Li outlet temperature is 950 8C which leads to an estimated efficiency of approximately 55% for a compression ratio of about 1.77. 3.2.2. Assumed boundary conditions The blanket design optimization and structural assessment of the TAURO blanket were per￾Fig. 3. TAURO outboard blanket module. 312 L. Giancarli et al. / Fusion Engineering and Design 61/62 (2002) 307/318

L Giancarli et al Fusion Engineering and Design 61-62(2002)307-318 Table 2 The obtained maximum stresses largely satisfy Design point for a typical blanket module of the TAURO the TaURo criteria in the whole sub-module reference design structure including the first wall: its maximum Module height 2 m value of 0.75 is reached in the composite plane Module width 0.3m which indicates reasonable design margins FW thickness For the tauro blanket both core sector Thickness first Pb-17Li channel Surface heat flux on Fw 0.5 MIm maintenance or core segment maintenance are Neutron wall loading on Fw 2.5 MW/m possible. In fact, the choice of maintenance Maximum Pb-17Li velocity procedure depends on many other considera- Pb-17Li inlet/outlet temperatures 800°C tions concerning the whole reactor plant; there- AT through Fw fore, this item has not yet been addressed aximum Temperature in SiCf/SiC However, a reactor study involving SCLL Maximum pb-I7Li-SiCiSiC interface tem. 915 blanket is being performed in 2002 within perature in the first channel the eu. and more information on blanket integration within the reactor will be available The MhD pressure drops are maintained at formed in 3D, assuming square meshes and an olerable levels by the relatively low electrical orthotropic model for SiCH/SiC. Due account was conductivity of SiC compared to steel in spite of taken of the poloidal variation in heat transfer coolant conditions. mechanical loads and thermal the relatively high velocity of the flowing Pb 17Li. loads. Conduction through Pb-I7Li has been considered as the dominant and therefore unique heat transfer mechanism. Assumed Sic/SiC prop- erties and parameters are those given in Table 1. 3.2. 4. Manufacturing and mounting scheme The TaURO design criteria [12] have been applied The different basic mponents have with limits described hereafter manufactured and then assembled separately The manufacturing of the module depends on Tensile and compressive stresses in plane are the possibility of joining SiC/SiC components. In limited to 145 and 580 MPa, respectively; order to be efficient, joining techniques require a Tensile and compressive stresses through the relatively large contact region between the differ hickness are limited to 110 and 420 MPa ent sub-components. Different joining techniques respectively; and can Shear stresses through the thickness are limited be envisaged: (i)assembling by sewing at 0 45 MPa textile stage to join the stiffeners to the side walls (ii) sticking and co-infiltration to join the second 3.2.3. Performances and modeling wall to the first stiffener; (iii) brazing of finished components n the bottom and the top closure Some of the main design features and perfor mance parameters of the reference design are plates and the different sub modules Therefore the stiffeners could be manufactured highlighted below with a T shape or an L shape surface at the end. A 3D neutronics analysis yields a tritium- The proposed fabrication scheme needs 10 coolant breeding ratio of 1.1 containment brazes per sub-module. For a module Assuming laminar Pb-17Li flow and neglecting consisting of 5 sub-modules this corresponds to onvection, the maximum SiC temperature at braze length of about 20 m(not including the top the first wall is 995C, and the maximum SiC/ and bottom caps) Pb-17Li interface temperature at the inner The fabrication scheme of a sub-module could channel wall is9l5° be the one illustrated in Fig. 4 where the different

formed in 3D, assuming square meshes and an orthotropic model for SiCf/SiC. Due account was taken of the poloidal variation in heat transfer, coolant conditions, mechanical loads and thermal loads. Conduction through Pb/17Li has been considered as the dominant and therefore unique heat transfer mechanism. Assumed SiCf/SiC prop￾erties and parameters are those given in Table 1. The TAURO design criteria [12] have been applied with limits described hereafter: . Tensile and compressive stresses in plane are limited to 145 and 580 MPa, respectively; . Tensile and compressive stresses through the thickness are limited to 110 and 420 MPa, respectively; and . Shear stresses through the thickness are limited to 45 MPa. 3.2.3. Performances and modeling Some of the main design features and perfor￾mance parameters of the reference design are highlighted below: . A 3D neutronics analysis yields a tritium￾breeding ratio of 1.1. . Assuming laminar Pb/17Li flow and neglecting convection, the maximum SiC temperature at the first wall is 995 8C, and the maximum SiC/ Pb/17Li interface temperature at the inner channel wall is 915 8C. . The obtained maximum stresses largely satisfy the TAURO criteria in the whole sub-module structure including the first wall; its maximum value of 0.75 is reached in the composite plane which indicates reasonable design margins. . For the TAURO blanket, both core sector maintenance or core segment maintenance are possible. In fact, the choice of maintenance procedure depends on many other considera￾tions concerning the whole reactor plant; there￾fore, this item has not yet been addressed. However, a reactor study involving SCLL blanket is being performed in 2002 within the EU, and more information on blanket integration within the reactor will be available soon. . The MHD pressure drops are maintained at tolerable levels by the relatively low electrical conductivity of SiC compared to steel in spite of the relatively high velocity of the flowing Pb/ 17Li. 3.2.4. Manufacturing and mounting scheme The different basic components have to be manufactured and then assembled separately. The manufacturing of the module depends on the possibility of joining SiCf/SiC components. In order to be efficient, joining techniques require a relatively large contact region between the differ￾ent sub-components. Different joining techniques can be envisaged: (i) assembling by sewing at textile stage to join the stiffeners to the side walls; (ii) sticking and co-infiltration to join the second wall to the first stiffener; (iii) brazing of finished components to join the bottom and the top closure plates and the different sub modules. Therefore, the stiffeners could be manufactured with a T shape or an L shape surface at the end. The proposed fabrication scheme needs 10 coolant containment brazes per sub-module. For a module consisting of 5 sub-modules this corresponds to a braze length of about 20 m (not including the top and bottom caps). The fabrication scheme of a sub-module could be the one illustrated in Fig. 4 where the different Table 2 Design point for a typical blanket module of the TAURO reference design Module height 2 m Module width 0.3 m FW thickness 3 mm Thickness first Pb/17Li channel 8.5 mm Surface heat flux on FW 0.5 MW/m2 Neutron wall loading on FW 2.5 MW/m2 Maximum Pb/17Li velocity 2.25 m/s Pb/17Li inlet/outlet temperatures 800 8C/ 957 8C DT through FW 115 8C Maximum Temperature in SiCf/SiC 995 8C Maximum Pb/17Li/SiCf/SiC interface tem￾perature in the first channel 915 8C L. Giancarli et al. / Fusion Engineering and Design 61/62 (2002) 307/318 313

314 L Giancarli et al Fusion Engineering and Design 61-62(2002)307-318 二 DETAIL C SEE DETAIL A DETAIL A 二SAIE21二 DETAIL B二 Fig 4. TAURO blanket design, breakdown in elementary components required sub-components are identified; the corre- Brazing the annular manifold connections to sponding mounting sequences could be as follows he back plate Manufacturing separate SiCfSic poloidal parts (box, second wall, stiffeners, top and bottom 4. Helium-cooled ceramic beryllium blankets cover)by SiCr weaving and Sic CVi or polymer The blanket concept described here was origin- Manufacturing back plate by C/SiCr weaving ally proposed for the DREAM reactor [13] and Assembling by sewing the second wall with the thereafter it was applied for A-SSTR2 [14]with first stiffener, and then sticking and co-infiltra- only little modifications tion the overall and the box together: Sticking and co-infiltration the others stiffeners 4.1. blanket description ith the Brazing the sub-module end cap at upper The power core torus structure is radially divided into equal sectors and each sector forms Brazing the sub-module end cap at lower an assembling unit. Each sector has itsown poloidal end; horizontal maintenance port, allowing replace Forming a blanket module by brazing 5 sub- cryostat or disassembling other components such modules together(this is a bond which is not in as the coil system. One blanket sector is divided contact with the coolant); and into 16 sub-sectors(rings)in the toroidal direction

required sub-components are identified; the corre￾sponding mounting sequences could be as follows: . Manufacturing separate SiCf/SiC poloidal parts (box, second wall, stiffeners, top and bottom cover) by SiCf weaving and SiC CVI or polymer process; . Manufacturing back plate by C/SiCf weaving; . Assembling by sewing the second wall with the first stiffener, and then sticking and co-infiltra￾tion the overall and the box together; . Sticking and co-infiltration the others stiffeners with the box; . Brazing the sub-module end cap at upper poloidal end; . Brazing the sub-module end cap at lower poloidal end; . Brazing the back plate with the box; . Forming a blanket module by brazing 5 sub￾modules together (this is a bond which is not in contact with the coolant); and . Brazing the annular manifold connections to the back plate. 4. Helium-cooled ceramic beryllium blankets The blanket concept described here was origin￾ally proposed for the DREAM reactor [13] and thereafter it was applied for A-SSTR2 [14] with only little modifications. 4.1. Blanket description The power core torus structure is radially divided into equal sectors and each sector forms an assembling unit. Each sector has its own horizontal maintenance port, allowing replace￾ment of the entire sector without opening the cryostat or disassembling other components such as the coil system. One blanket sector is divided into 16 sub-sectors (rings) in the toroidal direction. Fig. 4. TAURO blanket design, breakdown in elementary components. 314 L. Giancarli et al. / Fusion Engineering and Design 61/62 (2002) 307/318

L Giancarli et al. Fusion Engineering and Design 61-62(2002)307-318 The cooling ring(pipe) also works as support inner wall and the partition wall made of porous structure ceramic material The width (toroidal length) of the blanket modules is <500 mm based on consideration of 4.2. Assumed boundary conditions and thermo- fabricability and maintainability. The height(po- mechanical analyses lodal length) and thickness are 500 and 650 mm respectively. a typical blanket module is shown in The thermo-mechanical Fig. 5. It consists of the first wall, the tritium DREAM blanket was performed in 2D, assuming breeding zone, and the high-temperature shielding square meshes and an orthotropic model for SiC+ zone and is connected to the cooling ring with SiC composite. The assumed material properties bolts, as shown in the figure. The wall of the are listed in Table 1 module includes cooling paths. Neutron multiplier (Be), tritium breeding material (lithium oxide 4.2.1. Thermo-mechanical analysis of blanket LiO or other lithium ceramics) and shielding module first wall material(silicon carbide, Sic) are packed in the The integrity of the SiCSiC first wall during module. These are small-size pebbles, of diameter normal operation was examined the thermo- I mm for Be and Li,o. and 10 mm for sic mechanical analysis. Operating tions of the Cooling gas, helium( He)supplied from the inlet first wall are pipe of the cooling ring first flows through the cooling paths in the side wall to the first wall. Then nuclear heating rate: 16.5 MW/m3 helium flows into the module through the porous surface heat load: 0.5 MW/m- partition wall, cools the internal materials and coolant temperature: 700C returns to the outlet pipe of the ring. The first wall heat transfer coefficient: 6000 W/m/K configuration consists of a double-wall structure coolant pressure: 10 MPa The plasma side wall is 4-mm thick and the inner The maximum temperatures are 823C for a wall is 8-mm thick. Rectangular cooling channels, SiC!/SiC thermal conductivity, i=60 W/m/K and 3 10 mm, are provided between the two walls The total thickness of the first wall is 15 mm 954C for i=15 W/m/K. These values are below Rectangular cooling channel, 5 x 10 mm, which the operating temperature limit of SiC/ Sic lead helium into the breeding zone, lie between the posite, 1100C. The maximum Tresca stress values are 75.4 MPa for i=60 W/m/K and 137 High Temp. Cooling channel MPa for =15 W/m/K. These calculated values are below the assumed allowable stress of SiCsic composite of 200 MPa. Shield Pebble(SIC o1 4.2.2. Thermo-mechanical analysis of the module Tritium Breeder(LI2O 1) internal regions Neutron Multiplier(B The selected helium coolant pressure is 10 MPa nd the inlet/outlet temperatures are 600/900C. The pressure drop in the pebble bed was estimated Partition Wall using the Erguns equation and the film tempera ture difference between the helium coolant and the pebbles (Li,O) was estimated by the Shirai's First Wall equation. Since each blanket module is cooled in /Branch P parallel, the pressure drop is negligible small even for the smallest pebble diameter of I Fig. 5. Schematic 3D-view of a typical blanket module in restricting conditions to select the design para- DREAM reactor meters are: the mum SiCdSic temperature at

The cooling ring (pipe) also works as support structure. The width (toroidal length) of the blanket modules is B/500 mm based on consideration of fabricability and maintainability. The height (po￾loidal length) and thickness are 500 and 650 mm, respectively. A typical blanket module is shown in Fig. 5. It consists of the first wall, the tritium breeding zone, and the high-temperature shielding zone and is connected to the cooling ring with bolts, as shown in the figure. The wall of the module includes cooling paths. Neutron multiplier (Be), tritium breeding material (lithium oxide, Li2O or other lithium ceramics) and shielding material (silicon carbide, SiC) are packed in the module. These are small-size pebbles, of diameter 1 mm for Be and Li2O, and 10 mm for SiC. Cooling gas, helium (He) supplied from the inlet pipe of the cooling ring first flows through the cooling paths in the side wall to the first wall. Then helium flows into the module through the porous partition wall, cools the internal materials and returns to the outlet pipe of the ring. The first wall configuration consists of a double-wall structure. The plasma side wall is 4-mm thick and the inner wall is 8-mm thick. Rectangular cooling channels, 3/10 mm2 , are provided between the two walls. The total thickness of the first wall is 15 mm. Rectangular cooling channel, 5/10 mm2 , which lead helium into the breeding zone, lie between the inner wall and the partition wall made of porous ceramic material. 4.2. Assumed boundary conditions and thermo￾mechanical analyses The thermo-mechanical assessment of the DREAM blanket was performed in 2D, assuming square meshes and an orthotropic model for SiCf/ SiC composite. The assumed material properties are listed in Table 1. 4.2.1. Thermo-mechanical analysis of blanket module first wall The integrity of the SiCf/SiC first wall during normal operation was examined by the thermo￾mechanical analysis. Operating conditions of the first wall are: nuclear heating rate: 16.5 MW/m3 surface heat load: 0.5 MW/m2 coolant temperature: 700 8C heat transfer coefficient: 6000 W/m2 /K coolant pressure: 10 MPa. The maximum temperatures are 823 8C for a SiCf/SiC thermal conductivity, l/60 W/m/K and 954 8C for l/15 W/m/K. These values are below the operating temperature limit of SiCf/SiC com￾posite, 1100 8C. The maximum Tresca stress values are 75.4 MPa for l/60 W/m/K and 137 MPa for l/15 W/m/K. These calculated values are below the assumed allowable stress of SiCf/SiC composite of 200 MPa. 4.2.2. Thermo-mechanical analysis of the module internal regions The selected helium coolant pressure is 10 MPa and the inlet/outlet temperatures are 600/900 8C. The pressure drop in the pebble bed was estimated using the Ergun’s equation and the film tempera￾ture difference between the helium coolant and the pebbles (Li2O) was estimated by the Shirai’s equation. Since each blanket module is cooled in parallel, the pressure drop is negligible small even for the smallest pebble diameter of 1 mm. The restricting conditions to select the design para￾meters are: the maximum SiCf/SiC temperature at Fig. 5. Schematic 3D-view of a typical blanket module in DREAM reactor. L. Giancarli et al. / Fusion Engineering and Design 61/62 (2002) 307/318 315

L Giancarli et al. Fusion Engineering and Design 61-62(2002)307-318 the plasma side 45%. the other hand, treating time is long and there is less experience. The fabrication process using the CVI method is as follows 4.3. Manufacturing and mounting scheme 1) fabricating textile preform of structural parts The blanket module fabricated by the unit of the module forming method is shown in Fig. 6. The process 2) densification by CVD in furnace to get dense SiCSic composite 1) assembling the cooling ring(low-temperature 3)machining and finishing pipe) which is divided into several parts The use of Sic coatings may be also needed to 2)temporary setting of branch high-temperature rovide He-hermeticity 3)inserting the high-temperature pipe and con- necting with the branch pi 4)completing the ring by joining the divided parts 5. Main requirements for other in-vessel 5)fixing the blanket module to the ring with components 4 bolts using the maintenance hole 6) closing the maintenance hole with the lid Other important in-vessel components are the divertor and shield. As a consequence of the The structural material, SiC!SiC composite, selected reactor strategy, as discussed in Section must be treated to get high density. There are 2, Pb-17Li coolant should be used for compo two methods to compose ceramics matrix directly nents associated with SCLl blankets, and He- from gaseous raw materials: CVD and CVI. The coolant for components associated VI method has advantages such as (1)no cooled ceramic beryllium blankets with helium First Wall 原a Exhaust Textile Preform →( Densification Machining &Finishi Fig. 6. Manufacturing process of a DREAM blanket module

the plasma side B/1100 8C, the maximum tem￾perature of the breeder pebble B/1000 8C, and the net thermal efficiency
/45%. 4.3. Manufacturing and mounting scheme The blanket module fabricated by the unit￾forming method is shown in Fig. 6. The process is as follows: 1) assembling the cooling ring (low-temperature pipe) which is divided into several parts 2) temporary setting of branch high-temperature pipes 3) inserting the high-temperature pipe and con￾necting with the branch pipes 4) completing the ring by joining the divided parts 5) fixing the blanket module to the ring with bolts using the maintenance hole 6) closing the maintenance hole with the lid. The structural material, SiCf/SiC composite, must be treated to get high density. There are two methods to compose ceramics matrix directly from gaseous raw materials: CVD and CVI. The CVI method has advantages such as (1) no pressurization, (2) close adherence between fiber and matrix, (3) possibility of complex shaping. On the other hand, treating time is long and there is less experience. The fabrication process using the CVI method is as follows: 1) fabricating textile preform of structural parts of the module 2) densification by CVD in furnace to get dense SiCf/SiC composite 3) machining and finishing The use of SiC coatings may be also needed to provide He-hermeticity. 5. Main requirements for other in-vessel components [4] Other important in-vessel components are the divertor and shield. As a consequence of the selected reactor strategy, as discussed in Section 2, Pb/17Li coolant should be used for compo￾nents associated with SCLL blankets, and He￾coolant for components associated with helium￾cooled ceramic beryllium blankets. Fig. 6. Manufacturing process of a DREAM blanket module. 316 L. Giancarli et al. / Fusion Engineering and Design 61/62 (2002) 307/318