- Email: [email protected]
Design, performance and manufacturing analysis for a compact permeator
Fusion Engineering and Design 87 (2012) 1495–1500
Contents lists available at SciVerse ScienceDirect
Fusion Engineering and Design journal homepage: www.elsevier.com/locate/fusengdes
Design, performance and manufacturing analysis for a compact permeator R. Sacristán a,∗ , G. Veredas b , I. Bonjoch a , I. Fernandez b , G. Martín a , M. Sanmartí c , L. Sedano b a
SENER Ingeniería y Sistemas, Provenza 392, 5a 08025 Barcelona, Spain EURATOM-CIEMAT Fusion Assoc., BBTU, Av. Complutense 22, 28040, Madrid, Spain c b FUS, Jardins de les Dones de Negre, 08930 Santa Adrià de Besòs, Barcelona, Spain b
a r t i c l e
i n f o
Article history: Available online 17 April 2012 Keywords: Permeator against vacuum LibPb loop Tritium recovery Efficiency analysis Process data sensitivity analysis Fabrication tests
a b s t r a c t A fast and efficient recovery of bred tritium is a major milestone of tritium breeding technologies R&D for the demonstration of a fusion reactor tritium self-sufficiency. Permeator against vacuum (PAV) runs as a single-step process for tritium on-line recovery, acts as passive systems allowing to be thermally governed can be easily in-pipe integrated in LiPb loop systems and can be conceived with high compactness. An optimal PAV design is proposed with detailed design parameterization of tritium recovery efficiency at different velocity ranges from numerical simulation based on properly developed Openfoam® CFD code BelFoam® customized solver. Diverse structural design options are being considered to manufacture the PAV component that presents diverse manufacturing concerns. Fabrication options are exploring rolling of thin plate with internal armor in the vacuum gap up to the coating of a porous controlled thick wall. The aim of this paper is to emphasize on the advanced performance of the PAV proposed concept, to present and discuss the different technical solutions that have been studied as well as the fabrication tests carried out. © 2012 Elsevier B.V. All rights reserved.
1. Introduction The design and construction of a permeator against vacuum (PAV) is one of the main challenges in tritium breeding technologies R&D due to the advantages that it has in front of the rest of techniques as ‘PB Sweepers + TRS’, ‘PB Strippers + TRS’ or immersed getters. Fuskite® [1], whose name is due to its spiral rolled shape, is being developed under HCLL blanket concept at loop operational ranges. The main advantage of this permeator in front of the US vacuum permeator one [2] is the compactness of fuskite® [1], as it is directly installed inside the pipes, so no additional space is necessary (see Fig. 1). The most significant strengths are the minimization of the tritium recovery time in blanket systems, the direct transfer to fueling systems in tritium plant, the high potential efficiency and compactness by design, the simple integration in a piping loop system, and the fact that it is a passive system. Furthermore it will bring the possibility to have the tritium breeding and the recovery system inside the Bioshield. This paper proposes the different options that have been considered for the PAV fabrication due to the difficulty in manufacturing
∗ Corresponding author. Tel.: +34 932 283 380; fax: +34 932 283 316. E-mail address: [email protected] (R. Sacristán). 0920-3796/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.fusengdes.2012.03.044
the geometry that is needed, in order to obtain enough efficiency in the system. It presents also the design options studied for the loop. An Openfoam® CFD sensitivity analysis in a simplified model depending on the temperature difference between the loop heater and the cold area has been carried out, in order to ensure natural circulation of the liquid PbLi and to obtain the process data needed to design and size the adequate equipments and measurement instruments. In addition a detailed model with fuskite® [1] inside the loop has been calculated to have the certainty that the pressure drop due to fuskite® [1] does not stop the natural circulation. Permeation efficiency analysis has also been performed to obtain the ratios between the quantity of tritium introduced in the system and the tritium recovered by fuskite® [1].
2. Manufacturing options and fabrication tests for the PAV The immersed PAV conceptual design is shown in Fig. 2 [3,4]. The geometry and dimensions needed are not easy to be obtained by means of standard fabrication methods. This, in addition with the fact that materials considered have to be specific ones with adequate permeation properties, has required a wide investigation in different fabrication methods available in the market, as well as some testing to evaluate the quality of the process. In this permeator, p = 4.5 mm, D = 125 mm and L = 200 mm. Fig. 3 shows the detailed layout of fuskite® [1] inside the loop.
1496
R. Sacristán et al. / Fusion Engineering and Design 87 (2012) 1495–1500 Table 1 Main manufacturing options list with problems and advantages for the PAV. Manufacturing/fabrication method
Problems and advantages description
Electroforming–electroplating
The electroforming/electroplating has to be done in the piece not rolled yet. Rolling has to be done later: very difficult process without damaging the material. It can be manufactured by laser-cusing method with stainless steel 316L. Welding is not avoided, one cover will have to be welded, the PAV cannot be manufactured in one piece. Tests are necessary to know the conformability by HIP. A special machine is needed to curve the spiral. An analysis of the behavior of thin plates under pressure and temperature brought by the process is necessary. It can be very expensive due to the tests and the need to manufacture the special machine to form the spiral. With this method welding of thin plates with 0.5 mm thickness can be done. Testing has been carried out in alpha-iron plates. It seems the best welding option. In the forming operation, the intermediate reinforcement will make internal stresses appear. Geometry too complex for this method.
Laser-cusing
HIP Fig. 1. Compact design with fuskite® permeator on-line developed in TECNO FUS project.
Laser welding
Thin plates forming
Extrusion (delta-ferrite or ceramic material) Solid block machining Wax mould Microfusion Spark erosion
Fig. 2. Conceptual design of compact immersed PAV.
2.1. Main manufacturing options considered for the PAV The main materials that have been considered, due to its adequate permeation properties [5,6], for the PAV are alpha-iron, niobium, palladium–silver and 304L or 316L stainless steel. Also porous ceramic manufacturing options have been studied. Table 1
Fig. 3. Detailed layout of fuskite® .
Geometry too complex for this method. Geometry too complex for this method. Geometry too complex for this method. Geometry too complex for this method.
sums up the more remarkable ones [7–9]. From this study it can be clearly deduced that there is no option that fulfills fully all the requirements that the PAV needs. The more restrictive and difficult ones are to ensure that there is no leakage (PbLi must not enter inside the vacuum area, see Fig. 4), that all the parts and joins are resistant enough and that the materials used in joins (or base materials) have a correct and controlled behavior in working conditions (not corrosion in contact with liquid PbLi, properties maintenance at 350 ◦ C, etc.). The most complicated part to manufacture is the area shown in the bottom part of Fig. 4 due to the difficulty to obtain this geometry in one piece and the
Fig. 4. Detail of a PAV possible configuration and detail of longitudinal section.
R. Sacristán et al. / Fusion Engineering and Design 87 (2012) 1495–1500
Fig. 5. Connection between the PAV and the vacuum pump.
complexity in joining two pieces with the quality necessary to avoid further problems when working. The vacuum gap has a thickness of 4.5 mm and in order to maintain it in spite of liquid metal pressure, some stiffeners have been included and calculations have been performed to ensure it withstand the applied loads. Fig. 5 shows how the permeator is connected to the vacuum pump. Vacuum provided is enough to ensure that it arrives to the whole permeator volume. 2.2. PAV fabrication tests Some fabrication tests have been carried out in order to check the possible difficulties for some of the options listed in Table 1. The first test has consisted in forming and welding two stainless steel plates with 1 mm thickness by standard methods as TIG (Tungsten Inert Gas) or MIG (Metal Inert Gas) welding. Fig. 6 shows a picture of this test. This first prototype had leakage problems, so it was concluded that, for plates with a thickness of 0.5 mm or less, that are the ones needed in the real PAV, a higher quality welding method has to be considered. That is why a second test has been performed: laser welding 300 mm × 300 mm in alpha-iron steel plates of 0.5 mm thickness has been carried out. Figs. 7 and 8 shows the weldings performed with filler metal and without it. In order to evaluate the two laser weldings, dye-penetrant inspection in the two samples has been done (see Fig. 9). The inspection showed that the laser welding process with and without filler metal in thin alpha-iron material can be done with quality results, without porous or cracks. The last check needed was to know if the laser cusing process maintained the crystallographic structures of the stainless steel 316L material: an austenitic structure is necessary to allow permeation. This structure is shown in Fig. 10. The structure appears to be formed by an austenitic matrix, so this process could be used to fabricate the PAV final prototype. As the PAV geometry cannot be obtained with a single method, several of them will be considered at the same time.
Fig. 7. Laser welding with filler metal.
Fig. 8. Laser welding without filler metal. Fig. 6. First built prototype of a rolled double membrane.
1497
1498
R. Sacristán et al. / Fusion Engineering and Design 87 (2012) 1495–1500
Fig. 9. Dye-penetrant inspection in the two samples.
3. Design options for the loop The conceptual design of the PAV demonstrator prototype is shown in Fig. 11. At early states of the development, one main issue is to choose the appropriate material for the loop. The material and the way to fabricate the loop must fulfill permeation requirements: no tritium, or the minimum quantity, should leave the loop: fuskite® [1] must be the way the tritium takes to leave the circuit [10,11]. Table 2 Sums up the main materials that have been considered for the loop construction. One of these three options will have to be chosen, after a more complete study of them. Assembly easiness, permeation prevention and cost will be the main factors that will be considered in the final election.
Fig. 11. PAV demonstrator prototype. Table 2 Sums up the main materials that have been considered for the loop construction. Material
Problems and advantages description
Stainless steel with an alumina coating
Different methods to of lumina coating: aluminizing, plasma projection, high velocity projection, CVD, PVD, etc. Difficulties ensuring access to the whole loop interior. Not adequate thermal conduction properties: difficulties in heating and cooling. Expensive and not easy to form and weld ensuring no permeation.
Quartz
4. Process data obtained from CFD analysis Tungsten
In order to know how the prototype will behave in case that the temperatures difference between the cold are and the hot area
Fig. 10. Microstructures (at different magnifications) in stainless steel AISI 316L after laser-cusing process.
R. Sacristán et al. / Fusion Engineering and Design 87 (2012) 1495–1500
Fig. 12. Results for 50 ◦ C T.
1499
Fig. 14. Results for 30 ◦ C T.
Fig. 13. Results for 10 ◦ C T. Fig. 15. Detailed model results.
is modified, a sensitivity analysis has been performed in a simplified model with the loop without the fuskite® [1] inside. PbLi properties introduced in the model have been extracted from [12]. Figs. 12, 13 and 14 shows flow velocities and temperatures distribution for 50 ◦ C, 10 ◦ C and 30 ◦ C temperature range respectively. The average velocity varies from 0.15 m/s with a T of 10 ◦ C to 0.25 m/s with a T of 50 ◦ C.
All the process data extracted from the CFD analysis is necessary to size the cold and heat equipment, as well as all the instrumentation. The velocity inside the loop increases as the temperature difference rises. In order to ensure that the fuskite® [1] does not block the natural circulation, a detailed simulation has been performed. Results are
Fig. 16. Hydrogen inventories and total amounts in each component.
Fig. 17. Efficiency evolution (solubility ratio: Fe/LiPb = 10 [3], v: ∼1 mm/s, TW = 500 ◦ C).
1500
R. Sacristán et al. / Fusion Engineering and Design 87 (2012) 1495–1500
shown in Fig. 15. Actually, PbLi velocity increases when passing through fuskite® [1] and the behavior is the expected one. Anyway, in order to be able to adapt the process conditions to the ones in HCLL or DCLL concepts, in the real experiment T can be modified and this allows a control of the rest of variables. 5. Efficiency analysis The tests are planned to check the efficiency of the component. The system loop is closed, e.g. hermetic for hydrogen. So the efficiency will be measured and demonstrated from very low partial pressures. The membrane will be thin enough and will transfers tritium across into vacuum quickly. Therefore, the tritium inventory in the membrane is not considered a primary issue. In the preliminary study performed [4] with Openfoam simulations [13] showed that a 2 m long permeator with 45 annular channels and 1 mm/s PbLi inlet velocity reached an efficiency of 99.5%. In the present work a 0.2 m permeator with 6 annular channels and an inlet velocity of 0.2 m/s is analyzed. Present analysis pretends to be a rough approach to be used as a reference for the experiment design. MathCAD® modeling results give similar scalable results for this kind of systems. Fig. 16 shows the evolution of the H mass inside the permeator component. The efficiency evolution is shown in Fig. 17. A stationary state is reached before the injection stops, reaching a stable efficiency. The efficiency strongly depends on the solubility ratio between PbLi and Fe, the solved tritium residence time in the permeator area related with flowing velocity and mass transfer regimes and local temperatures. Aforementioned variables must be taken into consideration for design purposes. For solubility ratio up to 10 [5] and low velocities 0.1 mm/s, the experimental efficiency can reach values above 20%, that justifies high efficiency high values of scaled extrapolations to DEMO. The transfer mechanisms as diffusion in the flow boundary layer between bulk and membrane surface, diffusion in the membrane itself, recombination of tritium atoms to T2-molecules at the inner side of the membrane and pressure drop in the T2 flow between membrane and the connection to the vacuum pump are considered at first release-rate order. The qualification of the efficiency will be established both, in transient (concentration increase in the liquid-metal and changing permeation flux) and in steady-state. In first case, all the release-rate mechanisms accounted in the model are dynamically balanced and in the second case balanced and in-equilibrium. Acting in-series, the slowest one establishes the kinetics. In relation to the overcoming of permeation reducing layers on both surfaces of the membrane, it is expected to be controlled through experimental procedures by the control of PbLi-eutectic
oxidation as it depends on the choice of the membrane material, quality of the eutectic. 6. Conclusions PAV represents the simplest, cost effective and reliable technology to be qualified for tritium recovery from liquid-metal breeders. An advanced design of PAV (fuskite® [1]) is being developed. Different manufacturing options are being taken into consideration and a mixture of them will be necessary. The same applies to the loop manufacturing. CFD analysis proves that fuskite® [1] will operate properly with natural convection and that tritium extraction is possible. The results show the technical attractiveness of PAVs as tritium processing solution for liquid metal breeding blankets. Experiments for prototype development and qualification are ongoing. Parameters for such a permeator in a power plant with either HCLL or DCLL breeding blankets will be calculated in further steps. References [1] Fuskite® , Registered design. [2] S. Malang, M. Tillack, C.P.C. Wong, N. Morley, S. Smolentsev, Development of the lead lithium (DCLL) Blanket Concept, TOFE-19 Conference Las Vegas, November 2010, 2010. ˜ L. [3] I. Martínez, B. Herrazti, G. Veredas, J. Fradera, I. Fernández, L. Batet, I. Penalva, Mesquida, J. Abellà, J. Sempere, L.A. Sedano, A demonstrator of a PAV for tritium recovery from LLE at HCLL TBM loop operational ranges, in: IEA International Workshop on Liquid Metal Breeder Blankets, 23–24 September 2010 (Madrid), Spain, 2010. ˜ [4] G. Veredas, J. Fradera, I. Fernández, L. Batet, I. Penalva, L. Mesquida, J. Abellà, J. Sempere, I. Martínez, B. Herrazti, L. Sedano, Design and Qualification of an On-line Permeator for the Recovery of Lead–Tritium from Lead–Lithium Eutectic Breeding Alloy, 26th Symposium on Fusion Technology (SOFT-2010), 2010. [5] F. Reiter, K.S. Forcey, G. Gervasini, A Compilation of Tritium-Material Interaction Parameters in Fusion Reactor Materials, Report EUR 15217 EN, 1993. [6] J.E. Bringas, Handbook of Comparative World Steel Standards, ASTM DS67A 2nd ed., 2002. [7] L. Castillo, Study about the rapid manufacturing of complex parts of stainless steel and titanium, TNO and AIMME report, 2005. [8] G. Masters, A benchmark of RM technologies for metal components. Report for Perkins Group of Companies, TCT-Magazine–feature: Rapid Manufacturing, 2008. [9] D.M. Bromley, Hydrogen Embrittlement Testing of Austenitic Stainless Steels SUS 316 and 316L, The University of British Columbia, April 2008. [10] S. Malang, R.F. Mattas, Comparison of Lithium and The Eutectic Lead Lithium Alloy, Two Candidate Liquid Metal Breeder Materials for Self-Cooled Blankets, ISFNT-3, international symposium on fusion nuclear technology, Los Angeles, 1994. [11] G.W. Hollenberg, A. Terlain, E.P. Simonen, G. Kalinin, Tritium/hydrogen barrier development, in: Third International Symposium on Fusion Nuclear Technology, June 27–July 1, 1994, Los Angeles, California, 1994. [12] E. Mas de les Valls, L.A. Sedano, L. Batet, I. Ricapito, A. Aiello, Lead–lithium eutectic material database for nuclear fusion technology, Journal of Nuclear Materials 376 (2008) 353–357. [13] OpenFOAM® version 2.0.0.
Contents lists available at SciVerse ScienceDirect
Fusion Engineering and Design journal homepage: www.elsevier.com/locate/fusengdes
Design, performance and manufacturing analysis for a compact permeator R. Sacristán a,∗ , G. Veredas b , I. Bonjoch a , I. Fernandez b , G. Martín a , M. Sanmartí c , L. Sedano b a
SENER Ingeniería y Sistemas, Provenza 392, 5a 08025 Barcelona, Spain EURATOM-CIEMAT Fusion Assoc., BBTU, Av. Complutense 22, 28040, Madrid, Spain c b FUS, Jardins de les Dones de Negre, 08930 Santa Adrià de Besòs, Barcelona, Spain b
a r t i c l e
i n f o
Article history: Available online 17 April 2012 Keywords: Permeator against vacuum LibPb loop Tritium recovery Efficiency analysis Process data sensitivity analysis Fabrication tests
a b s t r a c t A fast and efficient recovery of bred tritium is a major milestone of tritium breeding technologies R&D for the demonstration of a fusion reactor tritium self-sufficiency. Permeator against vacuum (PAV) runs as a single-step process for tritium on-line recovery, acts as passive systems allowing to be thermally governed can be easily in-pipe integrated in LiPb loop systems and can be conceived with high compactness. An optimal PAV design is proposed with detailed design parameterization of tritium recovery efficiency at different velocity ranges from numerical simulation based on properly developed Openfoam® CFD code BelFoam® customized solver. Diverse structural design options are being considered to manufacture the PAV component that presents diverse manufacturing concerns. Fabrication options are exploring rolling of thin plate with internal armor in the vacuum gap up to the coating of a porous controlled thick wall. The aim of this paper is to emphasize on the advanced performance of the PAV proposed concept, to present and discuss the different technical solutions that have been studied as well as the fabrication tests carried out. © 2012 Elsevier B.V. All rights reserved.
1. Introduction The design and construction of a permeator against vacuum (PAV) is one of the main challenges in tritium breeding technologies R&D due to the advantages that it has in front of the rest of techniques as ‘PB Sweepers + TRS’, ‘PB Strippers + TRS’ or immersed getters. Fuskite® [1], whose name is due to its spiral rolled shape, is being developed under HCLL blanket concept at loop operational ranges. The main advantage of this permeator in front of the US vacuum permeator one [2] is the compactness of fuskite® [1], as it is directly installed inside the pipes, so no additional space is necessary (see Fig. 1). The most significant strengths are the minimization of the tritium recovery time in blanket systems, the direct transfer to fueling systems in tritium plant, the high potential efficiency and compactness by design, the simple integration in a piping loop system, and the fact that it is a passive system. Furthermore it will bring the possibility to have the tritium breeding and the recovery system inside the Bioshield. This paper proposes the different options that have been considered for the PAV fabrication due to the difficulty in manufacturing
∗ Corresponding author. Tel.: +34 932 283 380; fax: +34 932 283 316. E-mail address: [email protected] (R. Sacristán). 0920-3796/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.fusengdes.2012.03.044
the geometry that is needed, in order to obtain enough efficiency in the system. It presents also the design options studied for the loop. An Openfoam® CFD sensitivity analysis in a simplified model depending on the temperature difference between the loop heater and the cold area has been carried out, in order to ensure natural circulation of the liquid PbLi and to obtain the process data needed to design and size the adequate equipments and measurement instruments. In addition a detailed model with fuskite® [1] inside the loop has been calculated to have the certainty that the pressure drop due to fuskite® [1] does not stop the natural circulation. Permeation efficiency analysis has also been performed to obtain the ratios between the quantity of tritium introduced in the system and the tritium recovered by fuskite® [1].
2. Manufacturing options and fabrication tests for the PAV The immersed PAV conceptual design is shown in Fig. 2 [3,4]. The geometry and dimensions needed are not easy to be obtained by means of standard fabrication methods. This, in addition with the fact that materials considered have to be specific ones with adequate permeation properties, has required a wide investigation in different fabrication methods available in the market, as well as some testing to evaluate the quality of the process. In this permeator, p = 4.5 mm, D = 125 mm and L = 200 mm. Fig. 3 shows the detailed layout of fuskite® [1] inside the loop.
1496
R. Sacristán et al. / Fusion Engineering and Design 87 (2012) 1495–1500 Table 1 Main manufacturing options list with problems and advantages for the PAV. Manufacturing/fabrication method
Problems and advantages description
Electroforming–electroplating
The electroforming/electroplating has to be done in the piece not rolled yet. Rolling has to be done later: very difficult process without damaging the material. It can be manufactured by laser-cusing method with stainless steel 316L. Welding is not avoided, one cover will have to be welded, the PAV cannot be manufactured in one piece. Tests are necessary to know the conformability by HIP. A special machine is needed to curve the spiral. An analysis of the behavior of thin plates under pressure and temperature brought by the process is necessary. It can be very expensive due to the tests and the need to manufacture the special machine to form the spiral. With this method welding of thin plates with 0.5 mm thickness can be done. Testing has been carried out in alpha-iron plates. It seems the best welding option. In the forming operation, the intermediate reinforcement will make internal stresses appear. Geometry too complex for this method.
Laser-cusing
HIP Fig. 1. Compact design with fuskite® permeator on-line developed in TECNO FUS project.
Laser welding
Thin plates forming
Extrusion (delta-ferrite or ceramic material) Solid block machining Wax mould Microfusion Spark erosion
Fig. 2. Conceptual design of compact immersed PAV.
2.1. Main manufacturing options considered for the PAV The main materials that have been considered, due to its adequate permeation properties [5,6], for the PAV are alpha-iron, niobium, palladium–silver and 304L or 316L stainless steel. Also porous ceramic manufacturing options have been studied. Table 1
Fig. 3. Detailed layout of fuskite® .
Geometry too complex for this method. Geometry too complex for this method. Geometry too complex for this method. Geometry too complex for this method.
sums up the more remarkable ones [7–9]. From this study it can be clearly deduced that there is no option that fulfills fully all the requirements that the PAV needs. The more restrictive and difficult ones are to ensure that there is no leakage (PbLi must not enter inside the vacuum area, see Fig. 4), that all the parts and joins are resistant enough and that the materials used in joins (or base materials) have a correct and controlled behavior in working conditions (not corrosion in contact with liquid PbLi, properties maintenance at 350 ◦ C, etc.). The most complicated part to manufacture is the area shown in the bottom part of Fig. 4 due to the difficulty to obtain this geometry in one piece and the
Fig. 4. Detail of a PAV possible configuration and detail of longitudinal section.
R. Sacristán et al. / Fusion Engineering and Design 87 (2012) 1495–1500
Fig. 5. Connection between the PAV and the vacuum pump.
complexity in joining two pieces with the quality necessary to avoid further problems when working. The vacuum gap has a thickness of 4.5 mm and in order to maintain it in spite of liquid metal pressure, some stiffeners have been included and calculations have been performed to ensure it withstand the applied loads. Fig. 5 shows how the permeator is connected to the vacuum pump. Vacuum provided is enough to ensure that it arrives to the whole permeator volume. 2.2. PAV fabrication tests Some fabrication tests have been carried out in order to check the possible difficulties for some of the options listed in Table 1. The first test has consisted in forming and welding two stainless steel plates with 1 mm thickness by standard methods as TIG (Tungsten Inert Gas) or MIG (Metal Inert Gas) welding. Fig. 6 shows a picture of this test. This first prototype had leakage problems, so it was concluded that, for plates with a thickness of 0.5 mm or less, that are the ones needed in the real PAV, a higher quality welding method has to be considered. That is why a second test has been performed: laser welding 300 mm × 300 mm in alpha-iron steel plates of 0.5 mm thickness has been carried out. Figs. 7 and 8 shows the weldings performed with filler metal and without it. In order to evaluate the two laser weldings, dye-penetrant inspection in the two samples has been done (see Fig. 9). The inspection showed that the laser welding process with and without filler metal in thin alpha-iron material can be done with quality results, without porous or cracks. The last check needed was to know if the laser cusing process maintained the crystallographic structures of the stainless steel 316L material: an austenitic structure is necessary to allow permeation. This structure is shown in Fig. 10. The structure appears to be formed by an austenitic matrix, so this process could be used to fabricate the PAV final prototype. As the PAV geometry cannot be obtained with a single method, several of them will be considered at the same time.
Fig. 7. Laser welding with filler metal.
Fig. 8. Laser welding without filler metal. Fig. 6. First built prototype of a rolled double membrane.
1497
1498
R. Sacristán et al. / Fusion Engineering and Design 87 (2012) 1495–1500
Fig. 9. Dye-penetrant inspection in the two samples.
3. Design options for the loop The conceptual design of the PAV demonstrator prototype is shown in Fig. 11. At early states of the development, one main issue is to choose the appropriate material for the loop. The material and the way to fabricate the loop must fulfill permeation requirements: no tritium, or the minimum quantity, should leave the loop: fuskite® [1] must be the way the tritium takes to leave the circuit [10,11]. Table 2 Sums up the main materials that have been considered for the loop construction. One of these three options will have to be chosen, after a more complete study of them. Assembly easiness, permeation prevention and cost will be the main factors that will be considered in the final election.
Fig. 11. PAV demonstrator prototype. Table 2 Sums up the main materials that have been considered for the loop construction. Material
Problems and advantages description
Stainless steel with an alumina coating
Different methods to of lumina coating: aluminizing, plasma projection, high velocity projection, CVD, PVD, etc. Difficulties ensuring access to the whole loop interior. Not adequate thermal conduction properties: difficulties in heating and cooling. Expensive and not easy to form and weld ensuring no permeation.
Quartz
4. Process data obtained from CFD analysis Tungsten
In order to know how the prototype will behave in case that the temperatures difference between the cold are and the hot area
Fig. 10. Microstructures (at different magnifications) in stainless steel AISI 316L after laser-cusing process.
R. Sacristán et al. / Fusion Engineering and Design 87 (2012) 1495–1500
Fig. 12. Results for 50 ◦ C T.
1499
Fig. 14. Results for 30 ◦ C T.
Fig. 13. Results for 10 ◦ C T. Fig. 15. Detailed model results.
is modified, a sensitivity analysis has been performed in a simplified model with the loop without the fuskite® [1] inside. PbLi properties introduced in the model have been extracted from [12]. Figs. 12, 13 and 14 shows flow velocities and temperatures distribution for 50 ◦ C, 10 ◦ C and 30 ◦ C temperature range respectively. The average velocity varies from 0.15 m/s with a T of 10 ◦ C to 0.25 m/s with a T of 50 ◦ C.
All the process data extracted from the CFD analysis is necessary to size the cold and heat equipment, as well as all the instrumentation. The velocity inside the loop increases as the temperature difference rises. In order to ensure that the fuskite® [1] does not block the natural circulation, a detailed simulation has been performed. Results are
Fig. 16. Hydrogen inventories and total amounts in each component.
Fig. 17. Efficiency evolution (solubility ratio: Fe/LiPb = 10 [3], v: ∼1 mm/s, TW = 500 ◦ C).
1500
R. Sacristán et al. / Fusion Engineering and Design 87 (2012) 1495–1500
shown in Fig. 15. Actually, PbLi velocity increases when passing through fuskite® [1] and the behavior is the expected one. Anyway, in order to be able to adapt the process conditions to the ones in HCLL or DCLL concepts, in the real experiment T can be modified and this allows a control of the rest of variables. 5. Efficiency analysis The tests are planned to check the efficiency of the component. The system loop is closed, e.g. hermetic for hydrogen. So the efficiency will be measured and demonstrated from very low partial pressures. The membrane will be thin enough and will transfers tritium across into vacuum quickly. Therefore, the tritium inventory in the membrane is not considered a primary issue. In the preliminary study performed [4] with Openfoam simulations [13] showed that a 2 m long permeator with 45 annular channels and 1 mm/s PbLi inlet velocity reached an efficiency of 99.5%. In the present work a 0.2 m permeator with 6 annular channels and an inlet velocity of 0.2 m/s is analyzed. Present analysis pretends to be a rough approach to be used as a reference for the experiment design. MathCAD® modeling results give similar scalable results for this kind of systems. Fig. 16 shows the evolution of the H mass inside the permeator component. The efficiency evolution is shown in Fig. 17. A stationary state is reached before the injection stops, reaching a stable efficiency. The efficiency strongly depends on the solubility ratio between PbLi and Fe, the solved tritium residence time in the permeator area related with flowing velocity and mass transfer regimes and local temperatures. Aforementioned variables must be taken into consideration for design purposes. For solubility ratio up to 10 [5] and low velocities 0.1 mm/s, the experimental efficiency can reach values above 20%, that justifies high efficiency high values of scaled extrapolations to DEMO. The transfer mechanisms as diffusion in the flow boundary layer between bulk and membrane surface, diffusion in the membrane itself, recombination of tritium atoms to T2-molecules at the inner side of the membrane and pressure drop in the T2 flow between membrane and the connection to the vacuum pump are considered at first release-rate order. The qualification of the efficiency will be established both, in transient (concentration increase in the liquid-metal and changing permeation flux) and in steady-state. In first case, all the release-rate mechanisms accounted in the model are dynamically balanced and in the second case balanced and in-equilibrium. Acting in-series, the slowest one establishes the kinetics. In relation to the overcoming of permeation reducing layers on both surfaces of the membrane, it is expected to be controlled through experimental procedures by the control of PbLi-eutectic
oxidation as it depends on the choice of the membrane material, quality of the eutectic. 6. Conclusions PAV represents the simplest, cost effective and reliable technology to be qualified for tritium recovery from liquid-metal breeders. An advanced design of PAV (fuskite® [1]) is being developed. Different manufacturing options are being taken into consideration and a mixture of them will be necessary. The same applies to the loop manufacturing. CFD analysis proves that fuskite® [1] will operate properly with natural convection and that tritium extraction is possible. The results show the technical attractiveness of PAVs as tritium processing solution for liquid metal breeding blankets. Experiments for prototype development and qualification are ongoing. Parameters for such a permeator in a power plant with either HCLL or DCLL breeding blankets will be calculated in further steps. References [1] Fuskite® , Registered design. [2] S. Malang, M. Tillack, C.P.C. Wong, N. Morley, S. Smolentsev, Development of the lead lithium (DCLL) Blanket Concept, TOFE-19 Conference Las Vegas, November 2010, 2010. ˜ L. [3] I. Martínez, B. Herrazti, G. Veredas, J. Fradera, I. Fernández, L. Batet, I. Penalva, Mesquida, J. Abellà, J. Sempere, L.A. Sedano, A demonstrator of a PAV for tritium recovery from LLE at HCLL TBM loop operational ranges, in: IEA International Workshop on Liquid Metal Breeder Blankets, 23–24 September 2010 (Madrid), Spain, 2010. ˜ [4] G. Veredas, J. Fradera, I. Fernández, L. Batet, I. Penalva, L. Mesquida, J. Abellà, J. Sempere, I. Martínez, B. Herrazti, L. Sedano, Design and Qualification of an On-line Permeator for the Recovery of Lead–Tritium from Lead–Lithium Eutectic Breeding Alloy, 26th Symposium on Fusion Technology (SOFT-2010), 2010. [5] F. Reiter, K.S. Forcey, G. Gervasini, A Compilation of Tritium-Material Interaction Parameters in Fusion Reactor Materials, Report EUR 15217 EN, 1993. [6] J.E. Bringas, Handbook of Comparative World Steel Standards, ASTM DS67A 2nd ed., 2002. [7] L. Castillo, Study about the rapid manufacturing of complex parts of stainless steel and titanium, TNO and AIMME report, 2005. [8] G. Masters, A benchmark of RM technologies for metal components. Report for Perkins Group of Companies, TCT-Magazine–feature: Rapid Manufacturing, 2008. [9] D.M. Bromley, Hydrogen Embrittlement Testing of Austenitic Stainless Steels SUS 316 and 316L, The University of British Columbia, April 2008. [10] S. Malang, R.F. Mattas, Comparison of Lithium and The Eutectic Lead Lithium Alloy, Two Candidate Liquid Metal Breeder Materials for Self-Cooled Blankets, ISFNT-3, international symposium on fusion nuclear technology, Los Angeles, 1994. [11] G.W. Hollenberg, A. Terlain, E.P. Simonen, G. Kalinin, Tritium/hydrogen barrier development, in: Third International Symposium on Fusion Nuclear Technology, June 27–July 1, 1994, Los Angeles, California, 1994. [12] E. Mas de les Valls, L.A. Sedano, L. Batet, I. Ricapito, A. Aiello, Lead–lithium eutectic material database for nuclear fusion technology, Journal of Nuclear Materials 376 (2008) 353–357. [13] OpenFOAM® version 2.0.0.