A volumetric neutron source for fusion nuclear technology ... Mechanical, Aerospace and Nuclear Engineering Department, University of California Los Angeles, ...
E.T. Cheng f, L.A. El-Guebaly d, M.Z. Hasan g, F. Najmabadi h, S. Sharafat g,. I.N. Sviatoslavski d ... accumulated in the shield during the 7200 s fusion power production phase. .... ing this differential leads to large storage units with .... Peak
Jul 9, 2008 - Vertical displacement events: A serious concern in future ITER ... tions such as space studies, discharge and laser-produced plasma .... forced flow of sub-cooled liquid, using of heat transfer principle ... important part in the heat t
Nov 5, 2008 - c Department of Nuclear Engineering, Kyoto University, Yoshida, ... requires good heat transfer from heated surface to the bulk flow. ... In the preceding research, free surface flows were investigated .... reference book .
Fusion nuclear technology and materials: status and R&D needs. Mark Tillack a, Shahram ......  G.R. Longhurst, S.L. Harms, E.S. Marwill and B.G.. Miller ...
Dec 5, 2016 - the world to feature calendar days made from tea leaves. Calendars are ... example, setting some obstacles for reading, setting the audience in ...
lyzing the vast amounts of multi-project software engineer- ing data. ... An essential component of Multi-Project Software Engi- neering is ... ences of the Open Source and Free Software communities ... is, one can issue queries such as âwhich othe
May 27, 2017 - paper, or from several separated sheets (a technique known as modular origami), drastically reduces the .... but with two diagonal possibilities each time, a crude enumerating approach consists in selecting .... semiperimeter is probab
Validation, and Reporting of Indoor Environment CFD analyses" developed by ASHRAE. The details of the manual are ... assessments of the value and quality of the CFD work. The manual suggests that the .... mesh system, the numerical procedures with fi
design and ensured its competitiveness with other power generation systems. ... parameters (thickness, composition, and Li enrichment). ... that the TBR of the LiSn/SiC and FLiBe/SiC blankets remained below 0.9 ..... 51-52 (2000) 325-330.
Oct 9, 2008 - proposes a strategic conclusion derived from the review of blanket designs for advanced fusion reactors. .... fusion machine. An alternative high temperature compatible solid neutron multiplier option would be highly preferable on the v
Feb 20, 2009 - The estimation is done for KU-1 silica transparency degradation during ... first tests allow selecting of Russian KU-1 fused silica glass as a can-.
Jul 11, 2008 - Coupled transient thermo-fluid/thermal-stress analysis approach ... Mechanical and Aerospace Engineering Department, UCLA, Los Angeles, CA 90095, .... analysis system (ANSYS structural code was used for this paper).
Fusion Engineering and Design 29 ( 1995) 3- 17. Contributions of the ... heating, activation and tritium breeding in relevant materials. IFE target ..... of 100-400 kJ. A NIF operational issue may ..... tion selection, target transport from the targe
Nov 2, 2015 - of two synthetic test problems. However synthetic, these tests are aiming to solve one of the long stand- ing challenges in petroleum geosciences, i.e. to fuse or merge structure data from macro, micro and nanoscale into a single, multi
Sep 9, 2007 - Abstract Curriculum design is a concern in European Universities as they face the ..... and components, PCBs design, devices protection, electromechanic drives technolo- gies, PCBs ..... collaborative curriculum development.
Separate read/write bitlines. â Single-port reads. â Dual-port write. â¢ Enable different design constraints. â Cell sizing. â Different pre-charge of the read-port.
CEE 3150 â Reinforced Concrete Design â Fall 2003. Design the flexural (including cutoffs) and shear reinforcement for a typical interior span of a six span continuous beam with center-to-center spacing of 20 ft. Assume the supports are 12 inches
and performance, and of validating one's library of target models. A more powerful realization of the role of data fusion â and, indeed, of resource management, ...
ADS by Agilent. â¢ Ansoft Serenade (RF package is called. Harmonica). Student version is available at www.ansoft.com/about/academics/sersv/ind ex.cfm Note: ...
The technical information engineering system (TIES) is a computer-as- sisted methodology to help achieve significant improvements in prod- uct quality ...
Mar 16, 2017 - ... Fritze, Uwe MÃ¶nks *, Christoph-Alexander Holst and Volker Lohweg ...... by a servomotor, but consists of an additional linear drive for vertical ...
... to disc disorders. Spinal discs are located between each vertebra in the spine and are designed to act ... For many disc disorders it is difficult to treat them through non-surgical meth- ods. .... results are compared to the best previous implan
Fusion Engineering and Design 84 (2009) 2145–2157
Contents lists available at ScienceDirect
Fusion Engineering and Design journal homepage: www.elsevier.com/locate/fusengdes
An example pathway to a fusion power plant system based on lead–lithium breeder: Comparison of the dual-coolant lead–lithium (DCLL) blanket with the helium-cooled lead–lithium (HCLL) concept as initial step S. Malang a , A.R. Raffray b,∗ , N.B. Morley c a
Fliederweg 3, D76351 Linkenheim-Hochstetten, Germany Mechanical and Aerospace Engineering Department and Center for Energy Research, University of California, San Diego, 9500 Gilman Drive, 458 EBU-II, La Jolla, CA 92093-0417, USA c Mechanical and Aerospace Engineering Department, University of California, Los Angeles, CA 90095, USA b
a r t i c l e
i n f o
Article history: Received 14 January 2009 Received in revised form 23 February 2009 Accepted 24 February 2009 Available online 5 April 2009 Keywords: Dual-coolant lead–lithium blanket He-cooled lead–lithium blanket Pathway to power plant concept
1. Introduction In considering the pathway for the development of an attractive fusion energy source, it is helpful to consider the R&D needed to develop the various fusion systems from ITER onwards. ITER will play a critical role in the development of fusion energy, contributing valuable information on most of the elements of a magneticallyconﬁned burning plasma fusion facility. However, ITER is not designed as a power-producing plant, and lacks essential features of an attractive power plant, such as a tritium producing, high temperature blanket system. Even blanket test module experiments in ITER, which themselves are being designed as the ﬁrst tests of many integrated aspects of blanket operation needed for a commercial power plant , will be operated only for relatively short periods of time under conditions and constraints that prevent the complete testing of all nuclear components and system responses. In particular, the effects of signiﬁcant radiation damage on middle and end of life behavior of nuclear components cannot be addressed in ITER. In the US, a demonstration power plant (DEMO) is traditionally considered as the ﬁrst commercial demonstration to investors and regulators that fusion is an economic and viable power source. It follows that there can be little difference between the essen-
tial features of a commercial plant and a DEMO, which thus must have the same physics characteristics, the same coolant, blanket, magnet and structural materials, and the same power conversion system as the commercial device. Thus, an information gap exists between ITER and DEMO, in different areas such as the aforementioned use and operation of prototypical blanket systems and materials, integration, reliability and maintenance. Filling this gap will require another facility or facilities that can be used to provide an integrated, prototypical environment to create the database and experience needed to ﬁnalize design and materials choices for DEMO . The pathway to the successful development of an attractive power plant implies a successful strategy in developing a number of key fusion systems forming part of the power plant (see e.g. Ref. ). Here, an example strategy and pathway in developing a blanket system based on the lead–lithium alloy breeder is described as an illustration. 2. Pb–Li class of blanket concepts A number of blanket concepts utilizing the eutectic liquid metal alloy, Pb–17Li,1 have been developed with a range of performances
1 Denoted Pb–17Li alloy in this work, sometimes referred to as Pb–15.3Li and Pb–16Li in other works depending on the speciﬁc alloy used.
S. Malang et al. / Fusion Engineering and Design 84 (2009) 2145–2157
and associated development risk. This write-up focuses on this class of concept. In general, Pb–17Li is preferred over pure lithium because it is much less chemically reactive with water, air and nitrogen. Blanket concepts based on this breeder can be divided into those utilizing Pb–17Li for tritium breeding only with a separate coolant and those utilizing Pb–17Li for both breeding and self-cooling (sometime in addition to another coolant). A short historical review of the development of these different classes of Pb–17Li blankets is given in the next section, along with a description of the key features and issues driving the design. 2.1. Self-cooled Pb–17Li blanket concepts In this class of concepts, the liquid breeder serves also as a coolant, extracting the heat generated by the fusion reaction through an external heat exchanger. In principle, such “self-cooled” blankets represent the simplest design because no large cooling surfaces are required inside the blanket to transfer the volumetric heat generated in the liquid breeder to another coolant, such as helium or water. The liquid breeder can ﬂow with relatively low average velocity in large ducts through the blanket, extracting in this way all the heat and all the tritium generated in the blanket to external systems outside the irradiation environment. One of the ﬁrst concepts of such a self-cooled lead–lithium blanket was proposed in the Mirror Advanced Reactor Study (MARS)  where the blanket was composed of circular tube bundles in which the breeder ﬂowed in a direction perpendicular to the main magnetic ﬁeld. This resulted in a rather large pressure drop caused by the Lorenz forces since the conducting tube walls were in direct contact with the ﬂowing liquid metal. Fortunately, the anticipated heat ﬂux at the ﬁrst wall (FW) surface was much lower than the one anticipated in Tokamaks, and allowed in principle the use of such a simple geometry. The ﬁrst design of a self-cooled lead–lithium blanket for Tokamaks was proposed in the frame of the US Blanket Comparison and Selection Study (BCSS) . The idea behind this implementation was to utilize a fast ﬂow of the liquid metal only in relatively small channels oriented in the toroidal direction (parallel to the main magnetic ﬁeld) to provide sufﬁcient cooling of the FW while avoiding the large magneto-hydrodynamic (MHD) pressure drop. The liquid metal would then ﬂow slowly through large poloidal channels in the main breeding region where it is heated up to the blanket exit temperature by volumetric heat generation. The MHD pressure drop was kept to a tolerable range by maintaining low average velocity and minimizing the wall thickness of the structure in these poloidal channels In principle, this ﬂow scheme looks attractive, but more sophisticated MHD analyses showed later that the velocity proﬁles in the FW channels were unfavorable for sufﬁcient cooling of the large surface heat ﬂux, and that there was a large multi-channel effect caused by the ﬂow transition between radial and toroidal direction, resulting in a much larger overall pressure drop than originally predicted (“Madarame Effect” ). A European effort proposed a modiﬁcation to the BCSS concept to reduce the pressure drop by including ﬂow channel inserts (FCIs)  to electrically decouple the ﬂowing liquid metal from the load carrying walls. These FCIs were made of a steel–alumina–steel sandwich, with the steel cover sheets welded together at all sides and corners to prevent any contact between the insulator and the liquid metal. With these FCIs arranged in all channels with a ﬂow direction perpendicular to the main magnetic ﬁeld, the MHD pressure drop became proportional to the thickness of the inner FCI wall (typically ∼0.5 mm), which is much lower than that of the load carrying walls. It was proposed to minimize primary stresses in these FCIs by including pressure equalization slots or holes in the inserts, theoretically making the liquid metal pressure in the
small gap between the FCI and structural wall nearly identical to the pressure in the bulk ﬂow. However, this pressure equalization under developing ﬂow conditions was never tested, installation of FCIs in all radial and poloidal channels was judged as being rather complicated, and the issue of insufﬁcient cooling of the ﬁrst wall still remained. In order to avoid large pressure drop and the relatively low temperature possible with ferritic steel structures, another idea was advanced. The use of SiCf /SiC composites as structural material was proposed, which allowed for much higher operating temperature opening the possibility of not only utilizing the fusion plant for power production but possibly for other applications such as hydrogen production. Two examples are the French TAURO  and the ARIES-AT  blanket concepts, both utilizing a selfcooled lead–lithium breeding zone and SiCf /SiC composite. In these concepts, the SiCf /SiC composite served as a high temperature structural material with low electrical conductivity to reduce MHD effects, allowing lead–lithium exit temperatures up to about 1000 ◦ C. When coupled to a Brayton power cycle, this would allow for electricity production with a cycle efﬁciency of ∼55%. It also would provide the possibility of hydrogen generation, typically requiring coolant temperatures above ∼800 ◦ C. The European PPCS advanced Model D [10,11] concept was based on a combination of the TAURO and ARIES-AT concepts. Although these concepts provide the possibility of high performance, they rely on the development and qualiﬁcation of SiCf /SiC as structural material with a relatively high thermal conductivity in a fusion environment, and carry a higher degree of development risk than concepts based on ferritic–martensitic steels. They tend to be considered as second-generation advanced concepts for fusion power plants. 2.2. The dual-coolant lead–lithium concept To avoid the need for insulating coatings or the use of SiCf /SiC composite as structural material in self-cooled Pb–17Li blankets, the ﬁrst dual-coolant lead–lithium blanket concept was proposed, based at that time on the use of sandwich FCIs . In this concept the entire steel structure including the ﬁrst wall was helium-cooled, and the FCIs were needed only in the large inner poloidal channels. At that time, different methods for electrically insulating coatings on steel walls were under development, most of them based on alumina applied either by hot dipping of steel in liquid aluminum, powder pack cementation, or plasma spraying. These methods were judged as promising, and replacing the FCIs by alumina coatings was suggested to simplify the DCLL blanket concept. This was one of the four competing blanket concepts in the European Blanket Comparison and Selection Exercise (BCSE). For a number of reasons, the water-cooled lead–lithium (WCLL) concept was evaluated with higher marks . A major consideration in this evaluation was the high development risk associated with the DCLL alumina coatings; the possible solution of utilizing sandwich ﬂow channel inserts was not then considered. As a result, the European program concentrated on the WCLL blanket. The DCLL concept was considered and further evolved in the US as part of the ARIES Spherical Tokamak (ST) study [14,15]. An ST imposes two key constraints on the blanket choice: (1) the blanket concept must be compatible with the water coolant in the center post, ruling out the use of liquid lithium; and (2) a high efﬁciency in the power conversion system is needed to compensate for the large re-circulating power in such a power plant. The DCLL concept was chosen but with the sandwich FCIs being replaced by an insert made of SiC-composite  which could serve as thermal insulator and standoff, in addition to electrical insulator, thereby allowing for much higher Pb–17Li exit temperatures (and correspondingly higher power cycle efﬁciency) than feasible with
S. Malang et al. / Fusion Engineering and Design 84 (2009) 2145–2157
steel–alumina–steel FCIs. This concept became the reference ARIESST blanket concept, as illustrated in Fig. 1. It was also evaluated later in the frame of the European Power Plant Conceptual Study (PPCS) as Model C [10,11]. It was judged as an interesting blanket concept with beneﬁcial performance and feasibility issues somewhere between the near term separately cooled blanket (WCLL Model A ) and the very advanced, entirely self-cooled lead–lithium blanket with SiCf /SiC composite as structural material (Model D [10,11]). The DCLL concept was further evolved as part of the ARIESCS study , including substantial progress in the areas of ﬂow routing, module attachment and coolant connection, and system integration. Concurrently with ARIES-CS, the DCLL concept was also chosen as US reference ITER TBM concept, and the US ITER TBM team has made important contributions to the concept including detailed MHD modeling and the introduction of a permeator for tritium extraction [17,18]. The ARIES-CS DCLL concept is shown schematically in Figs. 2 and 3 and the FCI in Fig. 4 .
2.3. Separately cooled lead–lithium blanket concepts For a long time the water-cooled lead–lithium blanket was the favorite liquid breeder concept in the European blanket development. This concept was considered to offer the following advantages when compared to self-cooled liquid breeder concepts and to ceramic breeder blanket concepts: - no neutron damage in the breeder;
Fig. 2. ARIES-CS DCLL blanket module .
S. Malang et al. / Fusion Engineering and Design 84 (2009) 2145–2157
Fig. 4. Flow channel insert (FCI) in Pb–17Li channel .
- easy tritium extraction outside the blanket; - no serious MHD issues caused by liquid metal ﬂow in the strong magnetic ﬁeld; - very modest chemical reaction potential between the coolant and breeder/multiplier (avoiding the safety issue of Li/water or Be/water reactions); - limited extrapolation of technology from the proven technology of ﬁssion pressurized water reactors (PWRs). However, there are disadvantages associated with the WCLL concept, namely: - complicated blanket structure with a large number of cooling tubes inside the blanket (large steel fraction, reliability issue); - slow circulation rate of the liquid metal, which in combination with the extremely low tritium solubility of Pb–17Li results in a high tritium partial pressure, leading potentially to large tritium permeation rates into the cooling water; - relatively low water temperature, which limits the efﬁciency in the power conversion system to ∼32%; - the need to design the ﬁrst wall box to withstand the full water pressure (15 MPa) under accidental conditions in order to limit water ingress into the liquid metal pool as well as the spill of water or LM to the vacuum vessel in case of internal water leaks. Such a WCLL blanket concept was under development in the EU for tests in ITER until 2002. A decision was then made to replace it by a helium-cooled lithium–lead (HCLL) concept in order to beneﬁt from the large similarities with the helium-cooled ceramic breeder (HCCB) blanket concept being developed in parallel, especially in the overall blanket structure and ancillary coolant systems (anticipated cost savings) [19,20]. Additional reasons for that change were some safety concerns with water cooling, and the lack of potential for achieving higher efﬁciency. This concept has been evaluated as model AB in the EU PPCS. It is illustrated in Figs. 5–7, which show the box structure, the breeder unit and the Pb–17Li ﬂow scheme, respectively.
Fig. 5. Box structure of the HCLL blanket module .
needed information to qualify them for deployment in a DEMO and/or move to a higher performance DCLL concept using FCIs based on SiC for DEMO if R&D progresses rapidly enough. The information obtained from operation of a higher performance DCLL using SiC as non-structural FCIs could be an attractive blanket solution itself, or might then open the way for the more advanced self-cooled concepts using SiCf /SiC as structural material for the second generation power plants (post-DEMO). Such a pathway is illustrated in Fig. 8. Additionally, a formulation of development pathway depends also on overall strategies for fusion development and deployment such as (1) what degree of efﬁciency, cost, reliability and safety will be acceptable to utilities that will deploy fusion and government agencies that will license it, and (2) whether there exists a target deployment date that is needed to play a role in greenhouse gas emission reduction plans for climate change considerations. To a large degree, in the US, answers to these two questions are in ﬂux and remain to be fully crystallized. In any event, before an informed decision regarding the most desirable development pathway and initial starting point can be made, it is critical that one understands the relative performance, safety and reliability advantages of the various concepts, weighed in perspective with the additional development difﬁculties and risks and compatibility with the testing environment available in ITER,
2.4. Example development path Based on the discussion and blanket concept descriptions highlighted in the previous sections, one might suggest that a reasonable pathway for the development of a Pb–17Li-based blanket concept would be to start with the lower-risk, lower-performance concepts, such as the HCLL and/or a DCLL concept with sandwich-FCIs based on mature technology. Ideally these concepts would be tested in ITER and complementary facilities (as required) providing the
Fig. 6. Breeder unit in HCLL blanket module .
S. Malang et al. / Fusion Engineering and Design 84 (2009) 2145–2157
doubts are substantiated, considering the fact that the requirements on the SiCf /SiC composite are much smaller for FCIs than for the case of a helium-cooled SiCf /SiC composite ﬁrst wall, for example. However, even if this is the case, the original DCLL concept  based on low temperature inserts made of a steel–alumina–steel sandwich can be considered as a ﬁrst step and will be used here as a basis for the comparison. Since this sandwich has a much higher thermal conductivity than SiC, and the allowable Pb–17Li/steel interface temperature is much lower than that of Pb–17Li/SiC, the achievable Pb–17Li exit temperature is considerably lower than in the reference DCLL blanket, and the modiﬁed concept is called therefore LT-DCLL (or lower-technology/temperature DCLL). The comparison is performed on the two classes of concepts (the HCLL shown schematically in Figs. 5–7; and the DCLL in Figs. 2–4) but effectively covers three concepts since both DCLL and LT-DCLL are included. The following key areas are used as basis for the comparison:
Fig. 7. Flow scheme for Pb–17Li in HCLL blanket .
1. 2. 3. 4. 5. 6. 7. 8. 9.
engineering complexity of the design, magneto-hydrodynamic (MHD) issues, tritium extraction and control, compatibility issues, pumping power, achievable efﬁciency in the power conversion system, required Li-6 enrichment to achieve tritium self-sufﬁciency, potential for liquid metal (LM)/water reaction, required extrapolation of the present technologies (development risks), 10. potential for extrapolation to more advanced concepts. The serious, but more general issue of the fabrication technology for the complicated ferritic steel blanket structure is not discussed in detail because the development risks and requirements are nearly identical for both types of blankets. 3.1. Engineering complexity of the design
Fig. 8. Illustration of development pathway for Pb–17Li blanket concepts.
IFMIF, and additional medium term devices. This is the question that is addressed in the next section through a detailed comparison of the HCLL and DCLL concepts. Finally, environmental conditions likely to be available in ITER must be considered, and additional testing needs beyond ITER identiﬁed. The strategy to reach DEMO and eventually a 2nd generation power plant blanket concept should make the most of ITER and of its environmental conditions (large size, prompt nuclear heating, in-pile spectrum dependence, magnetic ﬁeld shape dependence). 3. Comparison of HCLL and DCLL blanket concepts The focus in this section is on the overall comparison of the two blanket concepts based on a number of key parameters and issues, and not so much on proposing speciﬁc design solutions. As explained above, the present DCLL concept (denoted simply as DCLL in the remainder of this article) [16–18] is based on the use of FCIs fabricated from a SiC-composite (or potentially a SiC foam) that serve as both an electrical and thermal insulator. The feasibility of such a FCI material in the time frame of ITER is questioned sometimes, and some critics have doubts if such FCIs can really be developed for power plant applications. It is not clear that these
3.1.1. Blanket modules The overall geometry of the blanket modules can be very similar for the three concepts. Typical dimensions of these modules are about 2 m in the toroidal direction and 2–3 m in the poloidal direction. These dimensions are ﬂexible and can be adjusted as needed for a variety of optimization criteria. The radial thickness of the breeding zone is ∼0.6 m for the outboard zone and ∼0.4 m for the inboard zone in order to achieve tritium self-sufﬁciency and required shielding. All concepts employ a U-shaped FW panel with cooling channels in radial–toroidal–radial direction connected to distribution manifolds in the rear of the module. The module box is strengthened by helium-cooled vertical stiffening plates connecting the FW panel with a strong back wall. There are additional stiffening plates installed in the DCLL concepts to separate the 2 or 3 rows of poloidal liquid metal channels (separation walls), and to support the side walls (see Fig. 3). In the HCLL concepts there are additional horizontal stiffening plates used to support the side walls (see Fig. 5). Altogether, the ﬁrst wall and all the stiffening plates add about the same amount of steel and helium for the three concepts, and the fabrication, assembling, and cooling of these plates are quite similar for both classes of concepts. However, an important difference between the HCLL and the DCLL concepts is in the need for additional cooling plates in the HCLL concept. These cooling plates must extract about 50% of the blanket heat generated in the Pb–17Li, resulting in a considerable increase in the steel and helium content in the module and, correspondingly, in the fabrication cost. They also complicate the distribution of the helium coolant, increase tritium permeation from the Pb–17Li into helium, and have an impact
S. Malang et al. / Fusion Engineering and Design 84 (2009) 2145–2157
on the tritium breeding rate as well as the reliability requirements of the module itself. For a comparison among the three blanket module concepts, the disadvantage of the HCLL concept caused by the need of these additional cooling plates has to be compared with the requirement for FCIs in both DCLL concepts. These FCIs can be fabricated in parts, assembled with overlapping zones like stove pipes. They are ﬁtted loosely into the large poloidal channels with a gap of 1–3 mm, minimizing in this way the required geometrical precision. The degree of complexity added by the FCIs will be described in Section 3.9 (as required extrapolation of present technologies and associated development risks). The Pb–17Li ﬂow scheme in the DCLL is rather simple (see Fig. 3). The liquid metal enters at the lower back of the module, fed through the annular region of a concentric access pipe; it is then distributed to a toroidal manifold and fed to the front row of poloidal channels. From there, for example it ﬂows in an upward direction, turns around at the top through 180◦ , and ﬂows downward in parallel in the back 1 or 2 rows of poloidal channels (only 1 back row is shown in Fig. 3). At the bottom of the module, there is a second toroidal manifold collecting all the liquid metal ﬂow and feeding it to the inner tube of the access pipe (as module outlet). Other ﬂow patterns are also possible which have certain advantages and disadvantages in terms of heat transfer and FCI thermal stress considerations. Such ﬂows are still being considered in the context of the US TBM program R&D effort. In any event, typical Pb–17Li average velocities in the large poloidal channels are in the range of 0.05–0.2 m/s. The LM ﬂow rate in the HCLL concept is about 2 orders of magnitude lower than in the DCLL concepts since the Pb–17Li is not used for heat extraction but only for tritium extraction. This means that typical Pb–17Li velocities in the HCLL module are of the order of mm/s. However, the ﬂow path has to be designed so that there are no “pockets” with stagnant Pb–17Li in order to avoid excessive tritium concentrations there. Usually, a meander-shaped ﬂow pattern is employed with the liquid metal ﬂowing around all cooling plates. Since such a ﬂow path includes ﬂow perpendicular to the main magnetic ﬁeld in small channels and numerous changes between parallel and perpendicular directions relative to the magnetic ﬁeld, MHD pressure drop is still an important issue in spite of the low velocities. 3.1.2. External systems for heat and tritium extraction An obvious difference between the two classes of liquid breeder blankets is the need for large primary liquid metal loops for the DCLL concepts. In these concepts, about 50% of the heat is extracted by the liquid metal and about 50% by the helium cooling the steel structure. Usually, such a dual ﬂuid cooling system is considered as the main disadvantage of the DCLL concepts, and it will certainly engender higher base equipment costs. However, a closer look is necessary to judge the true added complexity and reliability issues. Any liquid metal breeder blanket has to be designed for the following start-up and off-normal conditions: (a) pre-heating the module and ancillary liquid metal ﬂow systems before ﬁlling with liquid metal breeder; (b) keeping the breeder in liquid form during any downtime with no fusion power; and (c) providing reliable after heat removal if the helium cooling system is not operational. All these three functions can be provided in the DCLL concept by one of the two cooling systems. This is a large practical and safety advantage compared to entirely self-cooled blankets. Compared to the HCLL, the inherent possibility for redundant and reliable after heat removal in the DCLL concept (either with the helium or the
liquid metal loop, and even by natural convection) avoids the need for any additional measures. Overall, the helium loops for the heat extraction system are nearly identical in both classes of concepts but with about half the capacity for the DCLL concepts. For the liquid metal part, the heat exchange between liquid metal and secondary helium would require about the same heat transfer surface area as the HCLL blanket module itself. However, for the HCLL concept, this heat exchange occurs inside the neutron environment where repair, replacement, and/or redundancy is either not possible or at best extremely time consuming. The critical issue for the external DCLL liquid metal heat exchanger is to ﬁnd a suitable material for the heat exchanger (HX) tubes. This is a serious issue for the high temperature DCLL because it requires a material compatible with Pb–17Li and He at temperatures up to 700 ◦ C. Possible solutions are either refractory metals (Nb or Ta alloys), corrosion barrier coatings on ODS-steel or Nibased tubes, or possibly even SiC tubes. However, in the case of the LT-DCLL, the design of the HX is even simpler than the design of cooling plates in the HCLL since the liquid metal temperatures are roughly equal in these two concepts, but the material selection and fabrication is easier in the absence of neutron irradiation. The size and number of coolant access pipes to the blanket modules is about the same for the two classes of blankets. Concentric pipes for the Pb–17Li ﬂow are proposed for the DCLL so that an outer cold supply ﬂow (∼450 ◦ C) can cool the wall of the inner pipe containing hot (∼700 ◦ C) return ﬂow. A FCI will be required inside the inner tube to thermally isolate the hot ﬂow from the steel pipe wall. Without the use of concentric pipes, or some similar arrangement, some other high temperature compatible pipe material will be required for the high temperature DCLL. If, for example, a total thermal power of 25 MW has to be extracted from a module, concentric tubes with an outer diameter of ∼0.3 m can be used for both coolants in the DCLL case; in the HCLL case, two parallel helium access tubes of similar size (∼0.3 m) would be needed. Combining the two He tubes of the HCLL design into one concentric tube would require a larger diameter tube (∼0.5 m), making it more challenging to avoid excessive neutron streaming. 3.2. Magneto-hydrodynamic issues On this subject, there are large differences between the HCLL and DCLL concepts. The impact of the strong magnetic ﬁeld on the ﬂowing liquid metal (ﬂow distribution, velocity ﬁelds, pressure drop) is the most crucial issue for the design of self-cooled blankets. Adequate FW cooling in particular is a feasibility issue in this class of blankets. This problem is avoided completely with the DCLL and HCLL concepts since the FW is cooled with helium. The pressure drop in the large poloidal ducts of the DCLL concept can be minimized by the use of ﬂow channel inserts, either fabricated from a SiCf /SiC composite, some other SiC form such as SiC-based foams, or a steel–alumina–steel sandwich . A comparison of the results of the MHD analyses performed for the US ITER TBM indicates that ﬂow distribution and heat transfer is less of a problem for the HCLL concept, but the pressure drops are approximately of the same order. The increase in pressure drop in the DCLL blanket is marginal if the SiC FCIs are replaced by the sandwich FCIs, but in both cases, the pressure drop is not a major concern. While such inserts can reduce the MHD pressure drop in ﬂows perpendicular to the magnetic ﬁeld by orders of magnitude, they are less efﬁcient in regions with 3D current ﬂow. This is critical particularly at the transition between toroidal and poloidal ﬂow, such as in supply manifolds, and it remains to be seen how feasible measures can be taken to ensure equal ﬂow rates in all parallel channels [17,18]. Unequal ﬂow distribution is somewhat less of an issue for the concept LT-DCLL with sandwich FCIs than for the DCLL concept with SiC FCIs since the steel wall in the sandwich option
S. Malang et al. / Fusion Engineering and Design 84 (2009) 2145–2157
Table 1 Example tritium concentrations and partial pressures in Pb–17Li for HCLL case (estimates are based on Reiter’s solubility formula ). Pb–17Li Recirculation rate (1/day) Rise in T-concentration/pass (gT /gLM ) T-concentration at blanket inlet (wppb) T-partial pressure at blanket inlet (Pa) T-concentration at blanket exit (wppb) T-partial pressure at blanket exit (Pa)
10 1.85 × 10−8 3.4 100 22.1 3861
causes higher pressure drops in the long poloidal ducts, leading to a more uniform ﬂow distribution. Unequal ﬂow distribution has a much smaller impact on heat transfer in the HCLL concept since conduction to the coolant plates is sufﬁcient to keep temperatures within an acceptable limit. Considerable MHD research is required to develop models and computer codes describing 3D current regions. This is also true for the issue of heat transfer in the large DCLL ducts where turbulence and natural convection have to be taken into account for the calculation of velocity and temperature ﬁelds determining the heat transfer from the liquid metal to the helium coolant . It is not at all certain at this time how vulnerable the SiC based FCIs will be to thermal stresses, and what the acceptable limit is on the FCI temperature difference. The higher the acceptable temperature difference, the less critical are the MHD issues affecting the Pb–17Li ﬂow balance and MHD mixed-convection heat transfer [17,18]. To deal with thermal stress, design solutions such as nested or two layer FCIs have been proposed that appear to mitigate this concern. The heat transfer is much easier to analyse for the HCLL concept, because the liquid metal is laminar and a heat transfer governed by conduction is sufﬁcient. 3.3. Tritium extraction and control For a long time, the two candidate liquid metal breeders (lithium and lead–lithium) have been compared on the basis of tritium extraction and control. The observations were that the extremely low tritium solubility in the Pb–17Li eutectic alloy facilitates tritium extraction, but that it is extremely challenging to control tritium losses by permeation to an acceptable level. In contrast, the high solubility of tritium in lithium avoids any permeation problem, but makes tritium extraction very challenging. For both the HCLL and DCLL concepts, it can still be concluded that the total tritium inventory in the lead–lithium in a power plant is low, <50 g in (e.g. see ), and, therefore, not an important issue. However, a satisfactory tritium extraction method from the liquid metal breeder has still not been demonstrated, and tritium permeation losses either into the helium coolant or to the environment from the external loop remain a critical issue. In this regard, there are large differences between the two classes of blanket concepts. Since there is an important interaction between extraction method and permeation losses, these issues are described separately for the two different blanket concepts. 3.3.1. Proposed methods for the HCCL concept This blanket concept is characterized by a low circulation rate of the liquid metal breeder, of the order of 10–100 times the total liquid metal inventory in the blanket per day typically limited by MHD
31.62 5.84 × 10−9 3.4 100 9.4 692
100 1.85 × 10−9 3.4 100 5.3 224
pressure drop (discussed above). If one assumes a module with a Pb–17Li inventory of 2 m3 and a corresponding fusion power of 20 MW, the corresponding increase in the tritium concentration in the Pb–17Li is between 20 and 2 wppb per pass through the blanket. For a perfectly efﬁcient tritium extraction device installed in this liquid metal loop there would be a zero T-concentration at the blanket inlet and the concentration at the exit would then be between 20 and 2 wppb, corresponding to a partial pressure between 2800 and 28 Pa (the estimates are based on Reiter’s formula for Sievert’s constant, Ks (mol m−3 Pa−0.5 ) = 1.31 exp(−1350 (J/mol)/RT) ). The tritium extraction methods proposed for the HCLL blanket are either bubble towers or packed columns with counter ﬂow of lead–lithium and helium. Experiments performed with such devices indicate that a tritium partial pressure of about 100 Pa at the exit from the extractor is about the lower limit of the achievable range . The corresponding T-concentration is about 3.4 wppb. If one assumes this value at the blanket inlet, the concentration at the exit would be ∼22 wppb for the case of a Pb–17Li circulation rate of 10 times a day, and ∼5.3 wppb for the high circulation rate of 100 times a day. The corresponding partial pressures at the blanket exit would be ∼3861 Pa for the low circulation rate and ∼224 Pa for the high circulation rate, as illustrated in Table 1. This means that the driving force for tritium permeation from the Pb–17Li into the helium coolant inside the blanket module and in the piping of the liquid metal circulation loop would be of the order of at least 100–4000 Pa, depending on the circulation rate and the location. To control these permeation losses, efﬁcient tritium permeation barriers as well as efﬁcient tritium extraction systems from the helium coolant are necessary for the HCLL concept . Compared to clean ferritic steel surfaces, the permeation has to be reduced by at least two orders of magnitude to keep tritium permeation into the helium coolant to a tolerable range. The development of coatings to help achieve this reduction in permeation has been ongoing for many years but the results are still not convincing . A sufﬁciently high permeation reduction factor has to be maintained over the entire blanket and piping life time, which is very difﬁcult to achieve considering that the coating is in contact with Pb–17Li at temperatures of about 350–530 ◦ C, and has to survive a neutron ﬂuence of up to 200 dpa in steel as well as cyclic stresses and strains in the steel wall. The main candidate for such a coating is alumina or aluminides, applied either by hot dipping of the steel into an aluminium pool, by powder pack cementation, or by plasma spraying . If sufﬁciently effective coatings cannot be developed, a very efﬁcient tritium extraction system in the helium cooling loop has to be deployed, which would be expensive, power consuming and cause safety concerns because the tritium inventory in the helium and in the tritium extraction system could become rather large.
Table 2 Pb–17Li ﬂow and tritium parameters for DCLL case (estimates are based on Reiter’s solubility formula ). Pb–17Li temperature rise (K) Pb–17Li ﬂow rate (kg/s) Number of cycles (1/day) Rise in T-concentration/pass (gT /gLM ) T-partial pressure at exit (for zero T-concentration at inlet) (mPa)
100 638 3015 6.13 × 10−11 30
150 426 2021 9.18 × 10−11 66
250 256 1227 1.53 × 10−10 172
S. Malang et al. / Fusion Engineering and Design 84 (2009) 2145–2157
Table 3 Example tritium concentrations and partial pressures in Pb–17Li for DCLL case (estimates are based on Reiter’s solubility formula ). LM temperature rise (K) T-concentration at blanket inlet (wppb) T-partial pressure at blanket inlet (mPa) T-concentration at blanket exit (wppb) T-partial pressure at blanket exit (mPa)
100 0.059 30 0.122 117
150 0.059 30 0.154 179
250 0.059 30 0.218 334
3.3.2. Proposed methods for the DCLL concept The Pb–17Li circulation rate in the DCLL concept is much larger and determined by the desired temperature rise in the blanket and the fraction of the total power to be removed by the Pb–17Li. For a module size similar to that used for the HCCL (2 m3 LM inventory, 20 MW power of which 50–60% is removed by the Pb–17Li), Pb–17Li ﬂow and tritium parameters can be estimated for different assumed Pb–17Li temperature rises, as shown in Table 2. A very promising tritium extraction method for the high liquid metal ﬂow rates of the DCLL concept is the use of a permeator. In this device, the liquid metal ﬂows through tubes which are located in a vacuum chamber. The tritium diffuses there through the LM boundary layer and the tube wall, and is then extracted by a vacuum pump. Such a permeator is installed in the Pb–17Li main stream (for example bundles of tubes with a diameter of 10 mm, a wall thickness of 0.5 mm, and a length of 5 m). Candidate tube materials include Nb or Ta alloys. Scoping calculations showed that it should be possible to reduce the tritium concentration in such tubes with a liquid metal velocity of 5 m/s down to a value of about 0.06 wppb, corresponding to a partial pressure of 30 mPa. If one assumes this value at the blanket inlet (identical to the concentration at the exit from the extractor), the resulting concentrations and partial pressures at the liquid metal exit from the blanket are very low, as shown in Table 3. Such low tritium partial pressures in the Pb–17Li have very positive consequences. Tritium permeation in the blanket from the liquid metal into the helium coolant or in the HX from the liquid metal into the helium loop of the Brayton cycle would not be of concern even when allowing for a build up in the tritium concentration in the helium until the partial pressure is equal in both ﬂuids. Tritium permeation losses in the piping of the liquid metal loop would also be negligible, especially for concentric pipes with the cold Pb–17Li in the annulus. One could ask if such a permeator extractor could also be used in a HCLL system. However, the potential for an improvement there is marginal. As shown in the previous section, even for a maximum liquid metal circulation rate and with zero tritium concentration at the blanket inlet, the tritium partial pressure at the blanket exit would be ∼35 Pa, which is at least two orders of magnitude larger than anticipated for the DCLL concept. 3.4. Compatibility issues 3.4.1. Compatibility of ferritic–martensitic steels with Pb–17Li This is a common issue of HCLL and DCLL blanket concept. Usually a maximum allowable interface temperature of 470 ◦ C is used for the system steel/Pb–17Li [25,26]. However, it is not clear how justiﬁed this value is based on the following reasons: 22.214.171.124. Extrapolation of the results from the corrosion tests to the blanket conditions is made on too simple a basis. Usually for fusion applications, the corrosion rate is described as a function of temperature and velocity only. However, - in blanket cases the temperature gradient in the ﬂow direction has a large impact;
- even in tests with turbulent ﬂow without magnetic ﬁeld, the channel dimensions have an important inﬂuence on the corrosion rate; thus, for a start, the velocity in the correlation should be replaced by the Reynolds number; - extrapolation of results from tests with turbulent ﬂow to cases with laminar ﬂow or with magnetic ﬁeld are of very limited value (different boundary layers, molecular diffusion completely different from turbulent material transport). 126.96.36.199. The criteria used for determining the maximum allowable interface temperature in a blanket are not adequate. - A value of 20 m/year has been speciﬁed as maximum allowable corrosion rate in order to avoid plugging in “cold” regions with small cross-sections (i.e. valves). This was based on experiences with sodium loops under completely different conditions . - If plugging would really be the determining step, not only would corrosion rates have to be evaluated but also deposition rates in colder parts of the loop to determine the parameters with the largest impact on the corrosion and deposition processes (certainly not the maximum corrosion rate somewhere in the loop, but the super saturation of the Pb–17Li with the wall material, and the distribution of “hot” and “cold” surfaces in the entire loop). 188.8.131.52. What could be a more suitable criterion?. As part of the US BCSS study , three candidate criteria had been envisaged for the determination of the maximum allowable interface temperature: - hands-on maintenance, - wall thinning, - risk of plugging. Of these, the maximum allowable wall thinning seems to be the ultimate criterion to determine the allowable interface temperature. In some elements of fusion blankets, the wall thickness is as low as 2 mm. If one assumes a blanket lifetime of 5 years, and an allowable wall thinning of 10%, the allowable corrosion rate would then be about 40 m/year. This is not too different from the previously assumed 20 m/year criterion, but has been derived on a much better basis. The local limit of the interface temperature depends on the ﬂow conditions and the actual wall thickness at that location, and is in general not identical with the maximum interface temperature in the entire blanket. 184.108.40.206. What kind of models are required for analysing corrosion tests and extrapolating the results to the conditions in blankets?. Simple rate equations or one-dimensional models are probably not sufﬁcient. Models are needed to calculate the concentration proﬁle perpendicular to the interface considering the local ﬂow conditions (i.e. boundary layer thickness, diffusivity of wall material in this boundary layer, turbulent material transport in the core of the ﬂow region, temperature gradients in ﬂow direction, magnetic ﬁeld strength and material transport in the ﬂow direction). Such models can also be used to calculate deposition rates. The present basis for the design of the DCLL concept is to limit the maximum interface temperature to ∼470 ◦ C. For the HCLL blanket the following extrapolation is made to determine the maximum allowable interface temperature: - Assume basis of 20 m/year allowable corrosion rate criterion. - This value is achieved in corrosion tests with turbulent ﬂow of ∼0.3 m/s at a temperature of 470 ◦ C. - From the corrosion test results, an Arrhenius-like correlation between temperature and corrosion rate has been derived with velocity as a parameter; it shows that the corrosion rate is proportional to velocity to the power of 0.8.
S. Malang et al. / Fusion Engineering and Design 84 (2009) 2145–2157
- This correlation is used to determine the maximum allowable interface temperature for the velocities down to mm/s, a typical value for the ﬂow inside the blanket. - This leads to a temperature limit of 510–530 ◦ C in the layout of the HCLL test blanket modules for ITER . Such an extrapolation is questionable because it is far outside the experimental range, and it is based on a debatable criterion. In summary, better models are required for a more appropriate assessment and comparison of the compatibility issue for both classes of blankets. 3.4.2. Compatibility of SiCf /SiC composites with Pb–17Li In power plant studies with self-cooled lead–lithium blankets (e.g. EU PPCS Model D and US ARIES-AT) interface temperatures of up to 1100 ◦ C were allowed. The only experimental basis for this is based on capsule tests with interface temperatures up to 1000 ◦ C at CEA  and up to 1100 ◦ C at ORNL . In both experiments, no reaction was found. However, such capsule tests with a small pool of stagnant Pb–17Li are of limited value. The results have to be veriﬁed by loop experiments with a temperature gradient in the ﬂow direction. Nevertheless, it seems reasonable to expect that the maximum interface temperature of <800 ◦ C estimated for the DCLL blanket would pose no serious compatibility problem. 3.4.3. Compatibilities in the multi-material system of the DCLL primary loop In the primary loop, Pb–17Li is in contact with ferritic–martensitic steel and SiCf /SiC composite in the blanket module, with Nb or Ta alloys in the permeator for tritium extraction, and with the tubes of the liquid metal/He heat exchanger. There is a large temperature variation along this loop, ranging from about 700 ◦ C at the blanket exit down to ∼450 ◦ C at the HX exit. The material losses at some locations, their transport by the liquid metal into other regions, and the depositions and possible reactions on other surfaces are issues which have to be investigated with dedicated loop experiments. 3.5. Pumping power The ratio of the pumping power to the heat extracted is a crucial issue in any helium-cooled system. In the design of such systems, usually an upper limit for the pumping power is set at ∼3% of the thermal power corresponding to ∼8% of the electrical power generated in the power conversion system in order to limit the impact of the pumping power on the efﬁciency of the power conversion system. Water and liquid metal cooled systems have much lower pumping power requirements rendering this point moot for these coolants. A related point to be considered in the layout of cooling systems is the maximum pressure drop to be overcome with the circulation pump. To keep such pumps as simple as possible, the pressure drop in lead–lithium systems should be less than ∼2 MPa, and the ratio between pressure drop and system pressure in the helium loops less than ∼5%. If such guidelines are followed, the DCLL concepts have in this area an important advantage over the HCLL concept since in the latter case 100% of the heat is extracted with He, whereas in the DCLL case only ∼50% of the heat is extracted with He and the other ∼50% with the liquid metal, which requires much less pumping power. This means that for identical conditions in the power conversion system (e.g. associated with HCLL blankets and LT-DCLL blankets), and the same He cooling conditions in the two blanket modules, the HCLL blanket requires roughly twice more of the generated electrical power to drive its blowers than the LT-DCLL.
3.6. Achievable efﬁciency in the power conversion system Three different kinds of power conversion systems can be considered for fusion power plants with HCLL and DCLL blankets: (A) Rankine steam cycle with modest reheat (max. steam pressure ∼7 MPa, max. steam temperature ∼470 ◦ C, efﬁciency ∼40%. (B) Advanced Rankine steam cycle (supercritical pressure, steam temperature ∼550 ◦ C, 2–3 reheats, efﬁciency ∼45%). (C) Closed Brayton helium cycle (helium pressure ∼12 MPa, upper temp. in recuperator ∼650 ◦ C, lower temp. in recuperator ∼330 ◦ C, efﬁciency ∼44%. The coolant conditions would determine the more appropriate cycle to be combined with the particular blanket concept: 3.6.1. System with HCLL blankets In this system the helium temperature at the exit from the blanket is limited to ∼500 ◦ C at a helium pressure of ∼8 MPa. This means that the power conversion system (A) with a theoretical efﬁciency of ∼40% is the only adequate one for this blanket concept. If the pumping power is included in this estimate, the net efﬁciency of the blanket system is reduced to about 36%. 3.6.2. System with LT-DCLL blanket, characterized by steel–alumina–steel sandwich FCIs In this system the helium inlet/outlet temperatures for the blanket are ∼350/450 ◦ C. The Pb–17Li enters the blanket at 350 ◦ C and exits at ∼500 ◦ C. This translates into that about the same conditions for the power conversion system (A) as with the HCLL blanket with an efﬁciency of ∼40%. Including the lower pumping power results in a net efﬁciency of ∼37%. 3.6.3. System with the high temperature DCLL blanket, characterized by SiCf /SiC composite FCIs The high liquid metal temperature at the blanket exit provides the option of using either the advanced steam turbine cycle (B) or the closed cycle helium turbine cycle (C). The achievable efﬁciencies in these two systems are very close (∼45% and ∼44%, respectively) but for safety reasons the Brayton cycle (C) is preferred (to avoid the high pressure steam cycle associated with cycle (B)). The gross thermal efﬁciency is then ∼44% and the net efﬁciency ∼41% (e.g. see ). It should be noted that the values presented in this assessment are based on simpliﬁed evaluations and would need to be conﬁrmed through more detailed calculations. However, even though these values are estimates, it is believed that the relative performances of the different blanket concepts are fairly represented by the results. 3.7. Required Li-6 enrichment to achieve tritium self-sufﬁciency The additional cooling plates in the HCLL blanket increase the steel and helium content in the module quite considerably as already mentioned in Section 3.1.1. This has a major impact on tritium breeding. To achieve tritium self-sufﬁciency, a higher Li6 enrichment and/or a thicker breeding zone would be required in the HCLL concept as compared to the DCLL concept (this may amount to a 5–10% difference in Li-6 concentration, for example 90% in HCLL and 80% in DCLL, or in about a 5–10 cm difference in blanket thickness to achieve the same tritium breeding ratio). Both points have an important impact on the cost of electricity. 3.8. Potential for LM/water reaction The much lower chemical activity of Pb–17Li with water was one of the main reasons for selecting this breeder material. However, if
S. Malang et al. / Fusion Engineering and Design 84 (2009) 2145–2157
such a potential cannot be completely avoided, one has to make a distinction between different contact modes of these two materials. There is a large difference in the reaction rate for the three possible contact modes:
from a failure in some other component. This is a common issue, but the DCLL variants may introduce signiﬁcantly more Pb–17Li during a ﬁxed detection/valve closure time than the HCLL due to their higher ﬂow rates.
(a) A lump of liquid Pb–17Li thrown into a pool of water. (b) Jets of water or steam injected into a pool of Pb–17Li. (c) Pb–17Li droplets sprayed into a pool of water.
3.9. Required extrapolation of the present technologies (development risks)
For the contact mode (a), a number of experiments have shown that the amount of lithium reacting with water was in all cases much smaller than 50% of the Li inventory in the alloy, and the liquid metal solidiﬁed rather fast [29,30]. The contact mode (b) was especially investigated in Hanford, and the experiments there showed that nearly all the water in such jets was reduced to hydrogen, leading to a considerable temperature increase in the pool. Actually, this temperature increase was nearly identical for Pb–17Li, Li, or Na [30,31]. The contact mode (c) has not yet been investigated in sufﬁcient details. Scoping estimates indicated that up to 70% of the Li inventory in the Pb–17Li droplet spray may react, resulting in a considerable hydrogen generation which may become a safety issue. The following subsections address the relevancy of these contact modes in relation to the different blanket concepts. 3.8.1. Power plant with steam cycle power conversion system and LT-DCLL blanket If there is a tube rupture in the liquid metal/steam HX (steam generators), steam jets could enter the liquid metal (contact mode b), resulting in a pressurization of the blanket modules and considerable hydrogen generation. To limit the amount of water ingress, the module has to be designed for the full operating pressure in the steam cycle (∼7 MPa). 3.8.2. Power plant with steam cycle power conversion system and LT-DCLL or HCLL blanket In such a system a tube rupture in a He/steam HX could cause a pressure pulse transmitted through the helium pipe into the blanket module. If such a pressure pulse would trigger a failure of a cooling plate inside the blanket module, there could be a connection between the steam volumes in the steam plant and the liquid metal volume in the blanket module, and the module could be pressurized up to the operating steam pressure. To avoid larger amounts of steam ﬂowing into the module, the module has to be designed for the full steam pressure. 3.8.3. Power plant with closed cycle helium turbine power conversion system and DCLL blanket In this case any potential for liquid metal/water reaction coming from the power conversion system is eliminated. 3.8.4. Common issue for all three blanket concepts In case of large helium leaks inside the blanket module the module can be pressurized up to the helium operating pressure (usually ∼8 MPa). If this would lead to a leak in the blanket box, a spray of liquid metal droplets could enter the plasma chamber. If there is water from other components present in the vacuum vessel in the course of such an accident, a liquid metal/water reaction with contact mode (a) or (b) could occur. The potential for such an accident and the results of it are identical for the HCLL and the DCLL blanket concepts. Disruptions loads may also provide the mechanism for a guillotine break in Pb–17Li inlet pipes or connections between manifolds and individual modules. In this case, a large stream or jet of Pb–17Li can enter the vacuum vessel and interact with any water present
There are two critical factors involved in the development of these blanket concepts: (a) The ﬂow channel inserts made of SiCf /SiC composite for the DCLL blanket. (b) The permeator tritium extraction system for both types of DCLL blankets. Other development issues, such as the LM manifold design to ensure acceptably uniform ﬂow distribution for the Pb–17Li under a variety of expected operating conditions, the fabrication of the rather complicated blanket structure from martensitic–ferritic steels, the fabrication of sandwich FCIs made of steel–alumina– steel, the design of high temperature Pb–17Li/He HX, or the closed cycle helium turbine power conversion system still need to be addressed, but the risk that these developments will not be successful seems smaller. 3.9.1. Development of ﬂow channel inserts made of SiCf /SiC composite The mission of these inserts, the requirements on the material, and the status of the development is as follows: 220.127.116.11. Characteristics of the ﬂow channel inserts. These FCIs have the following main functions: (a) Decouple electrically the ﬂowing liquid metal from the load carrying steel walls in order to reduce the impact of the strong magnetic ﬁeld on the ﬂow (MHD pressure drop, velocity proﬁles). (b) To minimize heat losses from the volumetrically heated liquid breeder to the helium-cooled steel walls. The FCIs with a typical thickness of 5–10 mm are loosely ﬁtted into all liquid metal channels of the blanket and can have the shape of cylindrical ducts (circular or square cross-section), bends, ﬂow distribution manifolds and so on. The entire insert structure is usually composed of a number of elements with overlapping connections between them (like stovepipes in a home). There is a liquid metal ﬁlled gap between the inserts and the steel structure (1–3 mm wide), allowing for differential thermal expansion and irradiation induced swelling (see Fig. 4). These exterior gaps are connected to the interior ﬂow region at the FCI overlap regions and potentially by small axial slots or rows of small holes on one wall if necessary, to equalize interior/exterior pressure differences and minimize primary load on the FCI structure. If there are forces between the FCI and the liquid metal core caused by electromagnetic transients, the FCI should transfer these forces to the steel walls without experiencing itself high stresses (“inserts are mechanically supported by the steel walls”). As a result of the thermal insulation by the FCIs, the liquid metal exit temperature can be made to be 100–200 ◦ C higher than the maximum interface temperature between the liquid metal and the steel, resulting in a substantial increase in the thermal efﬁciency of the power conversion system. Typical temperatures in lead–lithium blankets with helium-cooled RAFS-walls are
S. Malang et al. / Fusion Engineering and Design 84 (2009) 2145–2157
- Maximum interface temperatures between steel and Pb–17Li <500 ◦ C. - Maximum interface temperatures between SiC and Pb–17Li 700–800 ◦ C. - Pb–17Li exit temperature from the blanket ∼700 ◦ C. 18.104.22.168. Material requirements. To perform the functions described above, the FCIs have to be built of a material which meets the following requirements: (a) Electrical conductivity: especially important is the conductivity perpendicular to the wall surfaces (“through the thickness”). This conductivity need only be <500 S/m to provide adequate pressure drop control, but the optimal value may be <20 S/m over the entire lifetime of the blanket, depending on desired velocity proﬁles for heat transfer reasons and/or pressure drop sensitivity which can affect ﬂow distribution . The conductivity in the direction of the wall (“in-plane”) is less important. An indication of the maximum allowable conductivity is the experience from analyses of a previous blanket design where sandwich-type FCIs were considered, fabricated from two steel sheets with an alumina layer in between. If in that case the inner steel layer was <0.5 mm thick, the MHD pressure drop was considered as manageable . The electrical conductivities given above are maximum values, which must not be exceeded during operation under neutron irradiation. However, the electrical conductivity of unirradiated SiC is lower than these values by many orders of magnitude. (b) Thermal conductivity: the conductivity through the thickness of the insert should be as low as possible. This is necessary to minimize the heat losses to the steel walls since the heat extraction by helium cooling is less effective than the heat extraction by the circulation of the liquid breeder (He has a lower exit temperature and requires more pumping power and larger manifolds). A value of less than 5 W/(m-K) is desirable. In-plane thermal conductivity has nearly no impact. (c) Reliable sealing layer on all surfaces of the FCI: low electrical conductivity through the thickness of the FCI has to be maintained during the entire lifetime of the blankets. This means that the porous structure of the composite has to be covered at all surfaces with a dense layer, which is compatible with the ﬂowing lead–lithium at temperatures up to 800 ◦ C over 3–5 years of operation; can withstand the stresses in the layer caused by temperature gradients across the wall without developing cracks which could give pathways for LM to penetrate into the porous structure of the insert; and does not spall-off during neutron irradiation. A major candidate material for such a sealing layer is pure SiC, applied by chemical vapor deposition. (d) Density of the composite: high density is not required, but it is desirable to have a structure with closed porosity in order to avoid the liquid metal soaking into a larger region of the insert in case a sealing layer fails locally. (e) Young’s modulus of the composite: this modulus in the in-plane direction should be as low as possible. The lower this modulus, the lower the stresses in the FCI and the sealing layer, caused by a thermal gradient through the wall. (f) Crack growth: cracks growing through the entire wall thickness of the FCI must be avoided since they would give a direct current path from the ﬂowing liquid metal to the steel walls. This could lead to large increases in the MHD pressure drop, offsetting the ﬂow balance between neighboring channels. Fortunately, thermal gradients through the wall cause in general tensile stresses on the cold side and compression stresses on the hot side. Since cracks can grow only in regions of tensile stress, cracks through the entire wall are unlikely.
22.214.171.124. Fabrication of the FCI-materials and testing issues. A distinction should be made between the measurement of basic properties and testing the behavior of the FCIs. Most important, however, for both classes of tests is to use materials suitable for the functions of a FCI. As outlined in the previous section, these functions require a material which is largely different from composites under development for structural functions in fusion FW/blankets. For FW applications as an example, three-dimensional woven composites with a large fraction of ﬁbers running perpendicularly through the wall are required to get a high thermal conductivity in this direction. In that application, a high structural density is required in addition to high conductivity ﬁbers in order to combine high thermal conductivity with high strength. For FCIs, however, a better choice is probably to use multi-layers of simple two-dimensional woven ﬁber-fabrics, and to inﬁltrate SiC to a low density only in order to obtain low through-the wallconductivities (electrical and thermal). Since this inﬁltration stops usually if the pores in a region are closed, it is hoped that a structure can be produced with mainly closed pores and with increasing density from the middle of the wall to the surfaces. Such a structure should be easier to fabricate than high-quality, high-density composites with maximized through-the-wall conductivity. Development of such a material is needed so that measurements of its thermal and electrical conductivity through the wall, Young’s modulus, and other mechanical properties such as interlaminar shear strength can be made. In addition, SiC-foam-based FCIs can also be considered. They have inherently lower thermal and electrical conductivity and the lower density absorbs few neutrons. Serious concerns exist about Pb–17Li inﬁltration into any open porosity should the dense sealing layer be breached. Ideas for closed-cell porous SiC foams and secondary ﬁllers of closed pore (alumina or similar materials) are being examined in this regard. For the investigation of the chemical compatibility of the FCI in Pb–17Li, it is crucial to have a material with the proposed sealing layer on all surfaces. For example, if pure SiC deposited by CVD on the composite serves this function, it would be misleading if the corrosion specimens are machined after the sealing layer has been deposited on the composite. If capsules made of solid SiC are used as corrosion specimen, the same “sealing layer” as envisaged for the composite should be applied to all capsule surfaces in contact with the Pb–17Li. In other words, for compatibility tests only the typical sealing layer is of importance, not the underlying structure. However, the right composite with the right sealing layer is mandatory for testing the behavior of the FCI under prototypical operating conditions. Only with such a material can the risk of sealing layer failures leading to Pb–17Li soaking into the composite structure be evaluated. 3.9.2. Development of a permeator for tritium extraction from Pb–17Li The high Pb–17Li circulation rate in the DCLL blanket results in an extremely low increase of the tritium concentration in the Pb–17Li during one pass through the blanket. A typical value is in the order of 10−4 wppm (see Section 3.3.2). For this reason, it would be extremely valuable to have a tritium extraction system capable of extracting tritium down to a tritium partial pressure <100 mPa (corresponding to a concentration <0.112 wppb calculated from Reiter’s solubility formula shown in Section 3.3.1) because this would avoid any problem of tritium permeation to the helium coolant or the environment. A rather simple extraction system based on tritium permeation from the Pb–17Li into a vacuum chamber is described in Section 3.3.2. However, tubes made of refractory metals (i.e. Nb- or Ta-alloys) are needed as the “permeation windows” owing to their compatibility with Pb–17Li
S. Malang et al. / Fusion Engineering and Design 84 (2009) 2145–2157
at temperatures up to 700 ◦ C, and their high tritium permeability. Such materials are widely used for high temperature application, but their high afﬁnity to oxygen, nitrogen, carbon, and hydrogen is a critical issue. They can be used only in inert gas systems or vacuum with extremely low impurity level. Whether such low levels can be maintained in real operation is a major challenge for a development program. For the LT-DCLL blanket with FCIs made of a steel–alumina–steel sandwich, steel tubes would be possible for such a permeator, substantially reducing the development risk associated with this component. 3.10. Potential for extrapolation to more advanced concepts Nearly all past blanket and fusion power plant studies have concentrated on plants for electricity production. A major reason for this was that other applications for fusion energy would require considerably higher coolant exit temperatures than achievable with ceramic breeders and with the ferritic–martensitic steel used in most of the concepts. The only exceptions proposed in some studies are blankets based on the use of SiCf /SiC composites as structural material and a self-cooled lead–lithium breeding zone, for example the French TAURO  and the ARIES-AT  blanket concepts. In these two concepts, the SiCf /SiC composite serves as high temperature structural material with low electrical conductivity, allowing lead–lithium exit temperatures of up to about 1000 ◦ C. This would allow electricity production with an efﬁciency in the power conversion system of ∼55%; it would also allow the use of fusion energy for hydrogen generation, a goal with increasing attractiveness but requiring coolant temperatures higher than about 800 ◦ C. However, many in the material community question whether such a material can be qualiﬁed for ﬁrst-generation fusion power plant blankets, and claim that ferritic–martensitic steels are the only realistic choice for this application. The DCLL concept appears to be a good compromise between more near term concepts such as helium- or water-cooled blanket concepts with steel structure, and the very advanced solutions with SiCf /SiC composite structures. It is emphasized that the requirement on the SiCf /SiC composite as an FCI are much lower than as structural material. One can even envisage that, with further development, a higher liquid metal exit temperatures (up to ∼800 ◦ C) could become feasible, enabling the use of the DCLL concept for hydrogen generation. In the mean time the water-cooled lead–lithium blanket has been replaced in Europe by the HCLL concept (see Section 2.2) allowing for an increase in the efﬁciency of the power conversion system from ∼32% to ∼36%. However, there is no known possibility to further increase the performance of this concept because the achievable helium exit temperature is limited to ∼500 ◦ C by the maximal allowable temperature at the interface of the Pb–17Li/steel as well as by the maximum allowable ferritic steel temperature of ∼550 ◦ C (for strength reasons). Even if one could increase the limit on the interface temperature by coatings and the maximum steel temperature by going to ODS-ferritic steel, the beneﬁt would be marginal. The situation is different for the DCLL concept. With SiC-based FCIs, this concept can enable an electricity production about 10% greater than the HCLL blanket with the same fusion power. If the development of SiCf /SiC composites for ﬁrst wall application (including neutron irradiation testing) would be successful, many of the results obtained in the development of the DCLL blanket could be extrapolated to blankets of the ARIES-AT class, which represent a really attractive solution. This would increase the efﬁciency of the power conversion system to ∼55% and would allow to use the heat for the generation of hydrogen.
On the other hand, if the development of SiCf /SiC composites for FCI application would not be successful, there is still the back-up solution to go back to a LT-DCLL with steel–alumina–steel sandwich FCIs. Such a low-risk blanket concept would seem at least as attractive as the HCLL concept. 4. Conclusions In this paper an assessment of the HCLL and DCLL blanket concepts has been presented in the context of an example development pathway for Pb–17Li based blankets. A historical perspective of the development of the various Pb–17Li based blankets has been summarized and a thorough description of the key issues driving the blanket concept assessment has been provided, including: • • • • • • • • •
engineering complexity, MHD effects, tritium extraction and control, compatibility, pumping power, achievable efﬁciency in the power conversion system, required Li-6 enrichment to achieve tritium self-sufﬁciency, potential for liquid metal/water reaction, required extrapolation of the present technologies (development risks), • potential for extrapolation to more advanced concepts. If one assumes a progressive overall strategy, consisting of 1. starting with the lower-risk lower-performance concept to be tested in ITER; 2. moving up to higher performance but higher development risk concepts, such as the DCLL concept with SiC-based FCIs for DEMO; 3. and ultimately developing a more advanced self-cooled concepts using SiCf /SiC as structural material for second generation power plant blankets. then two possible concepts can be considered as starting point: the HCLL concept and the low technology (LT) DCLL concept using sandwich FCIs. Based on this assessment overall, it seems more advantageous to start directly with the LT-DCLL concept; it can provide information which can be better extrapolated to the next step concepts (DCLL) at roughly the same level of risk as the HCLL concept. With infrastructure in place at ITER for a LT-DCLL system, testing of SiC-based FCIs and refractory-based permeators can still be performed in ITER test modules as the technology matures. Such tests might be limited to an outlet temperature of 500 ◦ C, with possible short duration excursions at higher temperatures, unless an internal by-pass is provided in the TBM or hot leg piping and heat exchanger are also upgraded to prototypic DCLL designs. Still, even at conservative outlet temperatures, valuable data can be obtained toward accelerating the development of the high temperature concept. References  L. Giancarli, V. Chuyanov, M. Abdou, M. Akiba, B.G. Hong, R. Lässer, C. Pan, Y. Strebkov, TBWG, Breeding blanket modules testing in ITER: An international program on the way to DEMO, Fusion Eng. Des. 81 (1–7) (2006) 393–405.  M.A. Abdou, S.E. Berk, A. Ying, Y.K.M. Peng, S. Sharafat, et al., Results of an international study on a high-volume plasma-based neutron source for fusion blanket development, Fusion Technol. 29 (1) (1996) 1–57.  M.A. Abdou, P.J. Gierszewski, M.S. Tillack, M. Nakagawa, J. Reimann, et al., Technical issues and requirements of experiments and facilities for fusion nuclear technology, Nucl. Fusion 27 (4) (1987) 619–688.  Mirror Advanced Reactor Study (MARS), Interim Design Report, Lawrence Livermore National Laboratory, Livermore, CA, UCRL-53333 (April 1983).
S. Malang et al. / Fusion Engineering and Design 84 (2009) 2145–2157  D.L. Smith, et al., Blanket Comparison and Selection Study (BCSS), Final Report, Fusion Power Program, Argonne National Laboratory, Argonne, IL, ANL/FPP-841 (September 1984).  H. Madarame, M.S. Tillack, MHD ﬂow in liquid metal blankets with helical vanes, J. Faculty of Eng., Univ. Tokyo (B) 38 (4) (1986) 1–18.  S. Malang, V. Casal, K. Arheidt, U. Fischer, W. Link, K. Rust, Liquid metal cooled blanket concept for NET, in: Proceedings of the 14th Symposium on Fusion Technology, Avinion, France, vol. 2, September 1986, pp. 1273–1280.  L. Giancarli, J.P. Bonal, A. Caso, G. Le Marois, N.B. Morley, J.F. Salavy, Design requirements for SiC/SiC composites structural material in fusion power reactor blanket, Fusion Eng. Des. 41 (1998) 165–171.  A.R. Raffray, L. El-Guebaly, S. Malang, I. Sviatoslavsky, M.S. Tillack, X. Wang, et al., ARIES-AT blanket and divertor, Fusion Technol. 39 (March (2)) (2001) 429– 433.  D. Maisonnier, I. Cook, P. Sardain, L. Boccaccini, E. Bogusch, K. Broden, et al., The European Power Plant Conceptual Study, Fusion Eng. Des. 75–79 (2005) 1221–1227.  D. Maisonnier, I. Cook, S. Pierre, B. Lorenzo, D.P. Luigi, G. Luciano, et al., DEMO and fusion power plant conceptual studies in Europe, Fusion Eng. Des. 81 (February (8–14)) (2006) 1123–1130.  S. Malang, K. Schleisiek (Ed.), Dual coolant blanket concept, KfK-Report 5424 (November 1994).  L. Giancarli, M. Dalle Donne, W. Dietz, Status of the European breeding blanket technology, Fusion Eng. Des. 36 (1) (1997) 57–74.  M. Tillack, S. Malang, High performance PbLi blanket, in: 17th IEEE/NPSS Symposium on Fusion Energy, San Diego, CA, October, 1997.  M.S. Tillack, X.R. Wang, J. Pulsifer, S. Malang, D.K. Sze, et al., Fusion power core engineering for the ARIES-ST power plant, Fusion Eng. Des. 65 (2) (2003) 215–261.  A.R. Raffray, L. El-Guebaly, S. Malang, X.R. Wang, L. Bromberg, T. Ihli, et al., Engineering design and analysis of the ARIES-CS power plant, Fusion Sci. Technol. 54 (3) (2008) 725–746.  N.B. Morley, Y. Katoh, S. Malang, B. Pint, A.R. Raffray, S. Sharafat, et al., Recent US research and development for the dual coolant blanket concept, Fusion Eng. Des. Fusion Eng. Des. 83 (2008) 920–927.
 C.P.C. Wong, S. Malang, M. Sawan, M. Dagher, S. Smolentsev, B. Merrill, et al., An overview of dual coolant Pb–17Li breeder ﬁrst wall and blanket concept development for the US ITER-TBM design, Fusion Eng. Des. 81 (2006) 461–467.  L.V. Boccaccini, L. Giancarli, G. Janeschitz, S. Hermsmeyer, Y. Poitevin, A. Cardella, et al., Materials and design of the European DEMO blankets, J. Nucl. Mater. 329–333 (2004) 148–155.  Y. Poitevin, L.V. Boccaccini, A. Cardella, L. Giancarli, R. Meyder, E. Diegele, et al., The European breeding blankets development and the test strategy in ITER, Fusion Eng. Des. 75–79 (2005) 741–749.  S. Smolentsev, R. Moreau, M. Abdou, Characterization of key magnetohydrodynamic phenomena in PbLi ﬂows for the US DCLL blanket, Fusion Eng. Des. 83 (5–6) (2008) 771–783.  F. Reiter, Solubility and diffusivity of hydrogen isotopes in liquid Pb-17, Fusion Eng. Des. 14 (1991) 207–211.  N. Alpy, A. Terlain, V. Lorentz, Hydrogen extraction from Pb–17Li: results with a 800 mm high packed column, Fusion Eng. Des. 49–50 (2000) 775–780.  I. Ricapito, A. Ciampichetti, G. Benamati, Tritium management in the HCLLDEMO blanket, ENEA Report, February 2004.  S. Zinkle, N.M. Ghoniem, Operating temperature windows for fusion reactor structural materials, Fusion Eng. Des. 51–52 (2000) 55–71.  J. Sannier, T. Flament, A. Terlain, Corrosion of martensitic steels in ﬂowing Pb–17Li, in: Proceedings of 16th Symposium on Fusion Technology, London, UK, September 3–7, 1990, pp. 901–905.  Ph. Deloffre, A Terlain, T. Dufrenoy, F. Herbert, V. Lorentz, Compatibility of SiCf /SiC composite with liquid Pb–17Li alloy at high temperature, Commissariat à l’Energie Atomique Report RT-SCCME 586, 2001.  B. Pint, Compatibility issues for a high temperature dual-coolant blanket, Fusion Sci. Technol. 52 (2007) 829–833.  L.S. Nelson, A. Farahani, J.D. Krueger, M.L. Corradini, The behaviour of molten Pb and Pb–17Li alloy when impacted by a vertical column of water, UWFDM-1031, July 1996.  M.L. Corradini, D.W. Jeppson, Lithium alloy chemical reactivity with reactor materials: current status of knowledge, Fusion Eng. Des. 14 (1991) 273–288.  D.W. Jeppson, Summary of lithium–lead alloy safety compatibility tests, Westinghouse Hanford Company Report WHC-EP-0202, DE89 07310, January 1989.