- Email: [email protected]
Helium-3 blankets for tritium breeding in near-term fusion devices
Fusion Engineering and Design North-Holland, Amsterdam
HELIUM-3
BLANKETS
D. STEINER, Department
8 (1989)
121-125
FOR TRITIUM
M.J. EMBRECHTS,
of Nuclear
Engineering
121
BREEDING
G. VARSAMIS
and Engineering
Physics,
IN NEAR-TERM
FUSION
DEVICES
and R. VESEY
Rensselaer
Polyrechnic
Institute,
Troy, NY
12180-3590,
USA
P. GIERSZEWSKI Canadian
Fusion
Fuels Technology
Project,
Mississauga,
Ontario,
Canada
L.5J
]K3
C.P.C. WONG GA Technologies,
Inc., P.O. Box 85608,
San Diego,
CA 92138,
USA
The purpose of this paper is to propose a novel blanket concept for in-situ tritium breeding in a near-term device such as ITER. In this concept terrestial supplies of helium-3, rather than lithium, are used for tritium breeding. In order to assess the key characteristics of this concept, a reference configuration was adopted based on minor modifications to the helium-cooled blanket concepts considered in the Blanket Comparison and Selection Study. The chosen configuration assumes a ferritic steel for structure and cladding and beryllium for neutron multiplication. The helium-3 is contained in a loop separate from the helium-4 coolant loop and flows within the beryllium. The helium-3 blanket exhibits good tritium breeding potential and low tritium inventories and leakage rates. The helium-3 requirements are 25-50 kg of inventory and 3.4-8.5 kg makeup/yr, 95% of which is due to bumup. It is estimated that there is sufficient helium-3 from decay of present military tritium stockpiles to meet this requirement.
1. Introduction
2. Blanket
The purpose of this paper is to propose a novel blanket option for in-situ tritium breeding in a device such as ITER. In this concept terrestial suppliesof helium-3, rather than lithium, are used for tritium breeding, taking advantage of the very large (n, T) cross-sectionof helium-3 at thermal neutron energies. The useof 3He for tritium breedingyields a uniquely simpleand safe blanket option for near-term application. The critical issue associatedwith this concept appearsto be political, not technical,in nature; namely, will countrieswith the required3He reservesmakethose reservesavailableto the project. In order to assessthe potential of this concept, scopingstudieswere carried out in the areasof tritium breeding,tritium recovery and control, and 3He inventory and leakagecharacteristics.Implications for ‘He resourcesare also considered.The choice of blanket configuration is addressedfirst.
As a starting point for assessingthe 3He blanket concept, the modular helium-cooledbreeder blankets from the Blanket Comparison and Selection Study (BCSS) were considered[l]. Theseblanketsuseferritic steel (HT-9), helium cooling at 5 MPa, and a tritium carrier fluid at 0.1-0.6 MPa, and are designedfor three full power years at 5 MW/m’ neutron wall load and 1 MW/m’ surfaceheat load, operatingat helium temperatures from 250o C to 500oC. For an ITER driver blanket, the samegeneralconfigurations can be usedat the lower powersand fluence of ITER. A redesign for ITER applications was not performed, but could lead to improved tritium breeding, useof 316stainlesssteelstructure, and thinner blankets. Rather, the BCSSdesignswereusedto estimaterelative volume fractions of materials. If tritium self-sufficiency is required, a neutron multiplier such as beryllium must be used. The beryllium
0920-3796/89/$03.50 0 Elsevier Science Publishers B.V.
configuration
122
D. Steiner
et al. / ‘He blankets
can be in the form of plates or rods. Conceptually, the former is equivalent to replacing the Liz0 breeder [l] with beryllium inside steel claddings. The latter would be similar to the LiAlO, blanket design [l], but with clad beryllium rods only, or to the liquid lithium design [l] with the lithium replaced by beryllium. In order to minimize 3He inventories, and 3He and tritium leakage rates, the ‘He is placed separate from the main helium (4He-4) coolant, flowing through the beryllium in a manner similar to the helium purge through a solid breeder. This arrangement provides good tritium breeding and also follows for removal of the beryllium-bred tritium. The 3He circulates only fast enough to recover the tritium. Both the breeding and cooling loops operate at about 5 MPa, consistent with available helium experience in large system, and the design pressure of the BCSS modules. 3He should be at a slightly lower pressure in order to minimize leaks outwards. For 600 MWth, and an inlet temperature of 80” C, a helium coolant flow rate of 510 kg/s (130 kmol/s) gives a temperature rise of 225°C across the blanket. For BCSS power reactor conditions, the maximum temperature in 2 cm beryllium rods was 400°C, with only a 10°C temperature rise across the beryllium. For ITER, temperatures should be lower.
3. Tritiom
breeding
The tritium breeding potential of the 3He blanket configuration was assessed using the ONEDANT [2] one-dimensional discrete ordinates transport code. Material cross-sections were produced using the TRANSX -CTR code [3] and the MATXSS cross-section library, with a s/S, approximation. The one-di-
for tritium
breeding
10”““““““” 0
2 HELIUIM-3
Fig. 1. Tritium
4
6 VOLUME
8 10 12 % lbd BER‘fLLlUM
I-l
breeding ratio versus percentage helium-3 in beryllium.
mensional cylindrical geometry and zone material compositions are given in table 1. Tritium breeding was examined varying the relative compositions of beryllium and 3He in zones 3 and 4. The total volume of beryllium and 3He was kept constant. The results are displayed in fig. 1, indicating that tritium breeding ratios above 1.5 are attainable when the 3He volume is about 10% of the beryllium plus 3He volume. This corresponds to an overall 3He volume fraction in the blanket of 6%. It is emphasized that the reference blanket configuration has not been optimized for use of 3He as a breeding material in ITER. For example, at the anticipated ITER heating loads, adequate thermal-hydraulic performance could be achieved at helium pressures of - 2.5 MPa (compared with the BCSS value of 5 MPa). This reduction in system pressure would allow a reduction in structural content which would enhance tritium breeding potential, yielding either higher tritium breeding at a fixed 3He content or reduced ‘He content for a fixed breeding ratio.
Table 1 ONEDANT geometry and zone compositions breeding calculations Zone
ThiCknCSS
1 (first wall) 2 (inlet plenum) 3 (breeder)
(cd 1.3 3.0 4.0
4 (breeder)
61.7
5 (outlet plenum)
15.0
for tritium
Composition 17.5% HT-9,9.3% 4He 14.1% HT-9,47.5% 4He 21.7% HT-9, 11.1% ‘He, 61% Be+3He 22.2% IjT-9,11.4% 4He, 66.3% Be+‘He 19.3% HT-9, 80.7% 4He
4. Tritium
recovery and control
For a 600 MW (fusion) machine with a tritium breeding ratio of unity, the tritium production rate of the blanket is 91 g T/d. This must be recovered from a helium carrier gas of 3He. The tritium pressure is set by tritium leakage &rough the clad breeder region (i.e. Be + 3 He) into the main coolant and then across the heat exchanger. The permeation of tritium from the main coolant circuit was determined by integrating the permeation rate along the exchanger tubes assuming a
D. Steiner
et 01. / ‘He blankets
linear temperature profile, and a permeation rate for 316SS of: P (mol HT/s)
for tritium
123
breeding
0.1 mol He/s. 5 MPa wth 2 Pa HTL25 Pa H2
= 0.0003438 exp( -8212/T(K))
Oxidizer
l
Drier
1 wb
with hydrogen addition (oxygen addition is an altemative approach for reducing permeation which should also be investigated). Owing to the low temperatures of this blanket, a 10 Ci/d loss rate across the heat exchanger is easily met without assuming oxide barriers by, for example, a HT partial pressure of 2 Pa and a H, partial pressure of 25 Pa in the main coolant. The tritium loss from the breeder across the cladding and into the main coolant is similarly expected to be low because of the low temperatures and because the breeder helium is at a slightly lower pressure than the coolant helium. Assuming a similar permeation rate as across the heat exchanger, partial pressure of 5 Pa HT/25 Pa H, in the 3He breeder gas will lead to overall tritium permeation rates of about 10 Ci/d, less if oxide barriers are present. Direct loss of tritium through leaks at valves and flanges will be very small at these tritium concentrations. Assuming 1% of the circuit volume leaks per year, and -400m34Heat2PaHTand -15m33Heat5Pa HT, the tritium leakage is 80 Ci/yr and 6 Ci/yr, respectively. The total tritium inventory in the breeder gas is then 0.06 g, and in the main coolant (assuming2 Pa HT) 0.8 g. In the event of an accidental releaseof this entire inventory, the maximumoff-site dosewould be only 50 mrem. The total tritium dissolvedin beryllium is about 0.5 g. The 1% of tritium directly producedin the beryllium leads to 45 g tritium inventory for a 1 cm thick beryllium layer, but up to 1 kg tritium by the end of ITER operation if a Be0 layer forms over much of the beryllium and inhibits diffusion. A related issueis the potential that someof the 3He bred tritium may be directly implanted in the beryllium. The extent of such implantation and its impact on the tritium inventory in beryllium need to be assessed. The tritium recovery systemmust remove91 g T/d from a 5 MPa helium carrier gas, where the tritium concentration is about 5 Pa HT/25 Pa Hz. The required helium gasflow rate is then 350 mol/s, or about 0.3 m3/s at 5 MPa. This flow rate is comparablewith the purge flow rate in solid breederblankets.The most proven approachfor recoveringtritium from suchhelium flows is by absorption onto molecularsieve beds. For
He HT H2 HTO H20
blankel 1 PPm moisture
350 mol/s. 5 MPa 0.00035 mo!Js. 5 Pa 0.0018 molds. 25 Pa 0.00007 molds 0.00028 mol/s
HTO M20 330 kg/d 2670 Ci/kg v
Fig. 2. Schematic of tritium recoverysystemfor the helium-3 blanket.
example, mole sieve driers are in use for recovering tritiated water from air at flow ratesof 0.7 m3/s (TSTA) and higher(CANDU reactors).To keepthe heliumfeed rate down, the ITER tritium systemwould operateat 5 MPa, rather than 0.1 MPa as in theseexamples. The tritium recovery system would also need oxidizers to convert the inlet HT into the water form, and electrolyzers to separateback the oxygen once the HTO has been trapped on the mole sievesand separated from the helium. Oxidizers are available in the appropriate size range, but again at 0.1 MPa (e.g. 0.6 cm3/s at TSTA). The tritiated water generatedis about 330 kg/d at 2.67 kCi/kg with water swamping,and electrolysis cells for handling the small volumes of highly active tritiated water are under testing by JAERI/TSTA, Mol (Belgium) and CEA (France). A schematicof the tritium recovery systemis shown in fig. 2, along with the relevant flow rates of helium and hydrogen species. This design assumesa recombiner exhaust of 1 ppb hydrogen and a drier exhaust of 1 ppm moisture,typical of presentsystems. Other options for tritium recovery include lithium trapping, solid metal gettering, cold trapping, and permeation membranes.Except for the first, which hasnot beentested,theseare generallynot attractive at the flow ratesinvolves, although all have the advantageof avoiding highly tritiated water. The installed cost for the tritium recovery systemis about lo-15 M$ US [4].
5. Helium
inventory
and leakage
The 3He breeder gas circuit must maintain enough flow to extract the tritium and maintain the HT partial pressureat under 5 Pa. This requiresa flow rate of 0.35
D. Steiner
124 Table
et al. / ‘He blankets
2
Helium-3
breeder
gas circuit
volume
Section
Dimensions
Number
Total volume
Inboard blanket Outboard blanket
150mZX0.3mX5% 250 m2 x0.6 m X 5%
Module plena
10 m X 0.0013 m2
24
2.3 7.5 0.3
15.7 m x 0.016 m2
4
1.0
30 m x 0.032 m2
2
1.9
1
1.0
Cm’)
Inlet/exit
ring
headers Pipe to/from headers Tritium recovery systems Pumps/valves (2 x 100%) Miscellaneous
Total
0.4 -
m IDx0.2 mx702
2
0.1
-
0.5
14.6
km01 3He/s. Based on an average density of 3.3 kg/m3 (250°C, 4.9 MPa) and flow rate of 10 m/s, the total flow area is 0.032 m2. This gas flows through the blanket modules where it occupies about 5% by volume, through sector plena into a 10 m diameter ring header at the top of the reactor, and then through a pipe into the tritium recovery system. From here, it is pumped back into another ring header around the machine, and again into the blanket modules. The ring headers are center fed, so each half-ring carries only half the flow. The dimensions assumed for distances and flow areas, and the corresponding volumes, are summarized in table 2. The path lengths are based on draft facility layouts for NET and US ITER [5,6]. The total 3He circuit volume is about 15 m3, requiring 50 kg 3He. This volume can be reduced significantly. If breeding is confined to outboard only, then an inventory reduction of - 25% would result. If the purge flow rate is increased to 30 m/s, the external piping diameter could be decreased leading to a further reduction of - 10%. If, as suggested earlier, the blanket structural content is reduced by adopting a lower system pressure, an additional inventory reduction of 15% might be achieved. Thus, it appears that the 3He inventory could be as low as 25 kg. The dimension of the 4He coolant circuit were also estimated. For a helium flow velocity of 50 m/s carrying 600 MWth, the minimum helium volume is about 230 m3. The pressure drop is about 130 kPa, for a pumping power of 17 MWe (80% efficient). An average velocity of 30 m/s would increase the circuit volume to
for mritium breeding
- 400 m3, but decrease the pumping power to about 5 MWe. As part of the coolant circuit design and permeation calculations, the primary heat exchanger (HX) was sized [7]. The design uses a tube-in-shell approach. The helium coolant leaves the HX at 80 o C, and heats up by 225 ’ C across the reactors. The water would enter the HX at 30” C, and leave with a 10” C temperature rise, consistent with a cooling tower approach and summer conditions. These conditions represent a compromise between minimizing the blanket temperature, and minimixing the HX area for permeation. The tubes were assumed to be 1 cm diameter and 1.5 mm thick, and velocities of 30 m/s He and 1 m/s water were used. A fouling resistance on the water-side of 2800 W/m2 K was added, typical of cooling tower water [7]. The resulting HX has 56600 tubes, each 3.4 m long, with a total surface area of 6000 m2. The leakage rate of helium from the coolant circuit and from the breeder circuit is estimated as between 1 and 5%/yr. Higher rates have been reported in some industrial helium systems where leak-tightness was not required. The lower rates were assumed in BCSS []. For the ‘He blanket, we assume that the ‘He breeder circuit has a l%/yr loss rate due to its small size and assuming welded, leak-tight construction. This corresponds to a 3He loss rate of about 0.25-0.5 kg/yr, depending on the inventory. This blanket would require some capability for 3He/ 4 He separation, the magnitude depending on the leakage rate of 3He into the main coolant. Thermal diffusion has been historically used, but is not particularly efficient. At low 3He concentrations, the difference in permeability of 3He and superfluid 4He provides good separation, and at higher 3He concentrations (> I%), cryogenic distillation is very effective. These have been demonstrated experimentally on a 3 L/h scale [8].
6. Helium-3
resource implications
Wittenberg et al. [9] recently estimated terrestial resources of 3He. Table 3 summarizes these estimates, showing helium-3 reserves that could be available in the year 2000, a time-frame consistent with the ITER schedule. The naturally occurring helium-3 resources are very dilute and, thus,’ are not economically attractive for ITER applications. The CANDU reactor resource is insufficient given an inventory requirement of - 25-50 kg and a yearly makeup rate of - 3.4-8.5 kg/yr (based on lo-25% availability/yr). Thus, tritium decay in
D. Steiner Table
et al. / ‘He blankets
3
Reserves [91
of helium-3
that could
Source
be available
Cumulative amount to year 2000 (kg)
Natural gas wells Underground storage Known reserves Decay of T, CANDU reactors US weapons (approximate) Total
weapons is the only could satisfy ITER
in the year
2000
Production rate post year 2000 (k&r)
29 187 10
source
15 17
125
to terrestial supplies of 3He), it will provide reactor-relevant information on helium-cooled blankets. While the technical features of the helium-3 blanket option are extremely attractive, the ability to guarantee the required helium-3 resources emerges as the major issue associated with the ultimate viability of this concept. Perhaps the initiation of the ITER project combined with the eventual ratification of the INF (Intermediate Nuclear Forces) treaty could provide the catalyst for an international agreement to supply the helium-3 requirements of the proposed blanket in ITER.
of helium-3
Acknowledgements
which
needs. One would expect that the US and the USSR have comparable sources of 3He associated with stores of thermonuclear weapons. Thus, the ultimate viability of the 3He blanket is tied to a political issue: would the relevant countries agree to make their 3 He resources available to the project. Clearly any opinion on the likelihood of such an agreement is speculative. Nevertheless, it is worth noting that the current international climate seems well suited to a proposal of this type.
7. Concluding
breeding
2
300 500 to600
potential
for tririum
remarks
A helium-3 blanket offers an attractive option for tritium breeding in near-term devices. The concept exhibits good tritium breeding potential, low tritium inventories, and low tritium leakage rates. Moreover, the concept retains the desirable safety and operational features inherent in gas-cooled blankets, including full compatibility with all reactor test blankets and the ability to operate the first walI at high temperatures for outgassing. There is no need to mechanically access the blanket in order to control breeding - the same blanket hardware can provide shielding or breeding controlled only by out-of-core addition of helium-3. The blanket performance is not adversely affected by variations in power level. Finally, although the concept is not extrapolatable to power reactors (assuming we are limited
This work was supported by the DOE, Office of Fusion Energy and by the Canadian Fusion Fuels Technology Project.
References
111C.P.C.
Wong, R.F. Bourque, E.T. Cheng, R.L. Creedon, 1. Maya, R.H. Ryder and K.R. Schultz, Helium-cooled blanket design, Fusion Technology, Vol. 8 (July 1985). PI R.D. O’Dell, F.W. Brinkley and D.R. Marr, User’s manual for ONEDANT: A code package for one-dimensional diffusion accelerated neutral particle transport, Los Alamos National Lab. Rep. LA-9184-M (Feb. 1982). TRANSX-CTR: A code for interfacing 131 R.E. MacFarlane, MATXS cross section libraries to nuclear transport codes for fusion system analysis, Los Alamos National Lab. Rep. LA-9863-MS (Feb. 1984). et al., Tritium systems concepts 141 S.K. Sood, O.K. Kveton for the Next European Torus (NET), CFFTP = G-86020 (Sept. 1986). draft report 151 G.S. Shaw et al., NET facility layout, CFFTP (1987). to US ITER Design Review Meet[61 J. Blevins, Contributions ing, CFFTP (1988). Heat Transfer, 2nd edn. (McGraw-Hill, [71 B. Gebhart, Toronto, 1971). distillation apparatus, MLM-2005, PI W.R. Wilkes, 3He-4He US Atomic Energy Commision (Mar. 1973). J.F. Santarius and G.L. Kulcinski, Lunar 191 L.J. Wittenberg, source of ‘He for commercial fusion power, Fusion Technology, Vol. 10 (Sept. 1986).
HELIUM-3
BLANKETS
D. STEINER, Department
8 (1989)
121-125
FOR TRITIUM
M.J. EMBRECHTS,
of Nuclear
Engineering
121
BREEDING
G. VARSAMIS
and Engineering
Physics,
IN NEAR-TERM
FUSION
DEVICES
and R. VESEY
Rensselaer
Polyrechnic
Institute,
Troy, NY
12180-3590,
USA
P. GIERSZEWSKI Canadian
Fusion
Fuels Technology
Project,
Mississauga,
Ontario,
Canada
L.5J
]K3
C.P.C. WONG GA Technologies,
Inc., P.O. Box 85608,
San Diego,
CA 92138,
USA
The purpose of this paper is to propose a novel blanket concept for in-situ tritium breeding in a near-term device such as ITER. In this concept terrestial supplies of helium-3, rather than lithium, are used for tritium breeding. In order to assess the key characteristics of this concept, a reference configuration was adopted based on minor modifications to the helium-cooled blanket concepts considered in the Blanket Comparison and Selection Study. The chosen configuration assumes a ferritic steel for structure and cladding and beryllium for neutron multiplication. The helium-3 is contained in a loop separate from the helium-4 coolant loop and flows within the beryllium. The helium-3 blanket exhibits good tritium breeding potential and low tritium inventories and leakage rates. The helium-3 requirements are 25-50 kg of inventory and 3.4-8.5 kg makeup/yr, 95% of which is due to bumup. It is estimated that there is sufficient helium-3 from decay of present military tritium stockpiles to meet this requirement.
1. Introduction
2. Blanket
The purpose of this paper is to propose a novel blanket option for in-situ tritium breeding in a device such as ITER. In this concept terrestial suppliesof helium-3, rather than lithium, are used for tritium breeding, taking advantage of the very large (n, T) cross-sectionof helium-3 at thermal neutron energies. The useof 3He for tritium breedingyields a uniquely simpleand safe blanket option for near-term application. The critical issue associatedwith this concept appearsto be political, not technical,in nature; namely, will countrieswith the required3He reservesmakethose reservesavailableto the project. In order to assessthe potential of this concept, scopingstudieswere carried out in the areasof tritium breeding,tritium recovery and control, and 3He inventory and leakagecharacteristics.Implications for ‘He resourcesare also considered.The choice of blanket configuration is addressedfirst.
As a starting point for assessingthe 3He blanket concept, the modular helium-cooledbreeder blankets from the Blanket Comparison and Selection Study (BCSS) were considered[l]. Theseblanketsuseferritic steel (HT-9), helium cooling at 5 MPa, and a tritium carrier fluid at 0.1-0.6 MPa, and are designedfor three full power years at 5 MW/m’ neutron wall load and 1 MW/m’ surfaceheat load, operatingat helium temperatures from 250o C to 500oC. For an ITER driver blanket, the samegeneralconfigurations can be usedat the lower powersand fluence of ITER. A redesign for ITER applications was not performed, but could lead to improved tritium breeding, useof 316stainlesssteelstructure, and thinner blankets. Rather, the BCSSdesignswereusedto estimaterelative volume fractions of materials. If tritium self-sufficiency is required, a neutron multiplier such as beryllium must be used. The beryllium
0920-3796/89/$03.50 0 Elsevier Science Publishers B.V.
configuration
122
D. Steiner
et al. / ‘He blankets
can be in the form of plates or rods. Conceptually, the former is equivalent to replacing the Liz0 breeder [l] with beryllium inside steel claddings. The latter would be similar to the LiAlO, blanket design [l], but with clad beryllium rods only, or to the liquid lithium design [l] with the lithium replaced by beryllium. In order to minimize 3He inventories, and 3He and tritium leakage rates, the ‘He is placed separate from the main helium (4He-4) coolant, flowing through the beryllium in a manner similar to the helium purge through a solid breeder. This arrangement provides good tritium breeding and also follows for removal of the beryllium-bred tritium. The 3He circulates only fast enough to recover the tritium. Both the breeding and cooling loops operate at about 5 MPa, consistent with available helium experience in large system, and the design pressure of the BCSS modules. 3He should be at a slightly lower pressure in order to minimize leaks outwards. For 600 MWth, and an inlet temperature of 80” C, a helium coolant flow rate of 510 kg/s (130 kmol/s) gives a temperature rise of 225°C across the blanket. For BCSS power reactor conditions, the maximum temperature in 2 cm beryllium rods was 400°C, with only a 10°C temperature rise across the beryllium. For ITER, temperatures should be lower.
3. Tritiom
breeding
The tritium breeding potential of the 3He blanket configuration was assessed using the ONEDANT [2] one-dimensional discrete ordinates transport code. Material cross-sections were produced using the TRANSX -CTR code [3] and the MATXSS cross-section library, with a s/S, approximation. The one-di-
for tritium
breeding
10”““““““” 0
2 HELIUIM-3
Fig. 1. Tritium
4
6 VOLUME
8 10 12 % lbd BER‘fLLlUM
I-l
breeding ratio versus percentage helium-3 in beryllium.
mensional cylindrical geometry and zone material compositions are given in table 1. Tritium breeding was examined varying the relative compositions of beryllium and 3He in zones 3 and 4. The total volume of beryllium and 3He was kept constant. The results are displayed in fig. 1, indicating that tritium breeding ratios above 1.5 are attainable when the 3He volume is about 10% of the beryllium plus 3He volume. This corresponds to an overall 3He volume fraction in the blanket of 6%. It is emphasized that the reference blanket configuration has not been optimized for use of 3He as a breeding material in ITER. For example, at the anticipated ITER heating loads, adequate thermal-hydraulic performance could be achieved at helium pressures of - 2.5 MPa (compared with the BCSS value of 5 MPa). This reduction in system pressure would allow a reduction in structural content which would enhance tritium breeding potential, yielding either higher tritium breeding at a fixed 3He content or reduced ‘He content for a fixed breeding ratio.
Table 1 ONEDANT geometry and zone compositions breeding calculations Zone
ThiCknCSS
1 (first wall) 2 (inlet plenum) 3 (breeder)
(cd 1.3 3.0 4.0
4 (breeder)
61.7
5 (outlet plenum)
15.0
for tritium
Composition 17.5% HT-9,9.3% 4He 14.1% HT-9,47.5% 4He 21.7% HT-9, 11.1% ‘He, 61% Be+3He 22.2% IjT-9,11.4% 4He, 66.3% Be+‘He 19.3% HT-9, 80.7% 4He
4. Tritium
recovery and control
For a 600 MW (fusion) machine with a tritium breeding ratio of unity, the tritium production rate of the blanket is 91 g T/d. This must be recovered from a helium carrier gas of 3He. The tritium pressure is set by tritium leakage &rough the clad breeder region (i.e. Be + 3 He) into the main coolant and then across the heat exchanger. The permeation of tritium from the main coolant circuit was determined by integrating the permeation rate along the exchanger tubes assuming a
D. Steiner
et 01. / ‘He blankets
linear temperature profile, and a permeation rate for 316SS of: P (mol HT/s)
for tritium
123
breeding
0.1 mol He/s. 5 MPa wth 2 Pa HTL25 Pa H2
= 0.0003438 exp( -8212/T(K))
Oxidizer
l
Drier
1 wb
with hydrogen addition (oxygen addition is an altemative approach for reducing permeation which should also be investigated). Owing to the low temperatures of this blanket, a 10 Ci/d loss rate across the heat exchanger is easily met without assuming oxide barriers by, for example, a HT partial pressure of 2 Pa and a H, partial pressure of 25 Pa in the main coolant. The tritium loss from the breeder across the cladding and into the main coolant is similarly expected to be low because of the low temperatures and because the breeder helium is at a slightly lower pressure than the coolant helium. Assuming a similar permeation rate as across the heat exchanger, partial pressure of 5 Pa HT/25 Pa H, in the 3He breeder gas will lead to overall tritium permeation rates of about 10 Ci/d, less if oxide barriers are present. Direct loss of tritium through leaks at valves and flanges will be very small at these tritium concentrations. Assuming 1% of the circuit volume leaks per year, and -400m34Heat2PaHTand -15m33Heat5Pa HT, the tritium leakage is 80 Ci/yr and 6 Ci/yr, respectively. The total tritium inventory in the breeder gas is then 0.06 g, and in the main coolant (assuming2 Pa HT) 0.8 g. In the event of an accidental releaseof this entire inventory, the maximumoff-site dosewould be only 50 mrem. The total tritium dissolvedin beryllium is about 0.5 g. The 1% of tritium directly producedin the beryllium leads to 45 g tritium inventory for a 1 cm thick beryllium layer, but up to 1 kg tritium by the end of ITER operation if a Be0 layer forms over much of the beryllium and inhibits diffusion. A related issueis the potential that someof the 3He bred tritium may be directly implanted in the beryllium. The extent of such implantation and its impact on the tritium inventory in beryllium need to be assessed. The tritium recovery systemmust remove91 g T/d from a 5 MPa helium carrier gas, where the tritium concentration is about 5 Pa HT/25 Pa Hz. The required helium gasflow rate is then 350 mol/s, or about 0.3 m3/s at 5 MPa. This flow rate is comparablewith the purge flow rate in solid breederblankets.The most proven approachfor recoveringtritium from suchhelium flows is by absorption onto molecularsieve beds. For
He HT H2 HTO H20
blankel 1 PPm moisture
350 mol/s. 5 MPa 0.00035 mo!Js. 5 Pa 0.0018 molds. 25 Pa 0.00007 molds 0.00028 mol/s
HTO M20 330 kg/d 2670 Ci/kg v
Fig. 2. Schematic of tritium recoverysystemfor the helium-3 blanket.
example, mole sieve driers are in use for recovering tritiated water from air at flow ratesof 0.7 m3/s (TSTA) and higher(CANDU reactors).To keepthe heliumfeed rate down, the ITER tritium systemwould operateat 5 MPa, rather than 0.1 MPa as in theseexamples. The tritium recovery system would also need oxidizers to convert the inlet HT into the water form, and electrolyzers to separateback the oxygen once the HTO has been trapped on the mole sievesand separated from the helium. Oxidizers are available in the appropriate size range, but again at 0.1 MPa (e.g. 0.6 cm3/s at TSTA). The tritiated water generatedis about 330 kg/d at 2.67 kCi/kg with water swamping,and electrolysis cells for handling the small volumes of highly active tritiated water are under testing by JAERI/TSTA, Mol (Belgium) and CEA (France). A schematicof the tritium recovery systemis shown in fig. 2, along with the relevant flow rates of helium and hydrogen species. This design assumesa recombiner exhaust of 1 ppb hydrogen and a drier exhaust of 1 ppm moisture,typical of presentsystems. Other options for tritium recovery include lithium trapping, solid metal gettering, cold trapping, and permeation membranes.Except for the first, which hasnot beentested,theseare generallynot attractive at the flow ratesinvolves, although all have the advantageof avoiding highly tritiated water. The installed cost for the tritium recovery systemis about lo-15 M$ US [4].
5. Helium
inventory
and leakage
The 3He breeder gas circuit must maintain enough flow to extract the tritium and maintain the HT partial pressureat under 5 Pa. This requiresa flow rate of 0.35
D. Steiner
124 Table
et al. / ‘He blankets
2
Helium-3
breeder
gas circuit
volume
Section
Dimensions
Number
Total volume
Inboard blanket Outboard blanket
150mZX0.3mX5% 250 m2 x0.6 m X 5%
Module plena
10 m X 0.0013 m2
24
2.3 7.5 0.3
15.7 m x 0.016 m2
4
1.0
30 m x 0.032 m2
2
1.9
1
1.0
Cm’)
Inlet/exit
ring
headers Pipe to/from headers Tritium recovery systems Pumps/valves (2 x 100%) Miscellaneous
Total
0.4 -
m IDx0.2 mx702
2
0.1
-
0.5
14.6
km01 3He/s. Based on an average density of 3.3 kg/m3 (250°C, 4.9 MPa) and flow rate of 10 m/s, the total flow area is 0.032 m2. This gas flows through the blanket modules where it occupies about 5% by volume, through sector plena into a 10 m diameter ring header at the top of the reactor, and then through a pipe into the tritium recovery system. From here, it is pumped back into another ring header around the machine, and again into the blanket modules. The ring headers are center fed, so each half-ring carries only half the flow. The dimensions assumed for distances and flow areas, and the corresponding volumes, are summarized in table 2. The path lengths are based on draft facility layouts for NET and US ITER [5,6]. The total 3He circuit volume is about 15 m3, requiring 50 kg 3He. This volume can be reduced significantly. If breeding is confined to outboard only, then an inventory reduction of - 25% would result. If the purge flow rate is increased to 30 m/s, the external piping diameter could be decreased leading to a further reduction of - 10%. If, as suggested earlier, the blanket structural content is reduced by adopting a lower system pressure, an additional inventory reduction of 15% might be achieved. Thus, it appears that the 3He inventory could be as low as 25 kg. The dimension of the 4He coolant circuit were also estimated. For a helium flow velocity of 50 m/s carrying 600 MWth, the minimum helium volume is about 230 m3. The pressure drop is about 130 kPa, for a pumping power of 17 MWe (80% efficient). An average velocity of 30 m/s would increase the circuit volume to
for mritium breeding
- 400 m3, but decrease the pumping power to about 5 MWe. As part of the coolant circuit design and permeation calculations, the primary heat exchanger (HX) was sized [7]. The design uses a tube-in-shell approach. The helium coolant leaves the HX at 80 o C, and heats up by 225 ’ C across the reactors. The water would enter the HX at 30” C, and leave with a 10” C temperature rise, consistent with a cooling tower approach and summer conditions. These conditions represent a compromise between minimizing the blanket temperature, and minimixing the HX area for permeation. The tubes were assumed to be 1 cm diameter and 1.5 mm thick, and velocities of 30 m/s He and 1 m/s water were used. A fouling resistance on the water-side of 2800 W/m2 K was added, typical of cooling tower water [7]. The resulting HX has 56600 tubes, each 3.4 m long, with a total surface area of 6000 m2. The leakage rate of helium from the coolant circuit and from the breeder circuit is estimated as between 1 and 5%/yr. Higher rates have been reported in some industrial helium systems where leak-tightness was not required. The lower rates were assumed in BCSS []. For the ‘He blanket, we assume that the ‘He breeder circuit has a l%/yr loss rate due to its small size and assuming welded, leak-tight construction. This corresponds to a 3He loss rate of about 0.25-0.5 kg/yr, depending on the inventory. This blanket would require some capability for 3He/ 4 He separation, the magnitude depending on the leakage rate of 3He into the main coolant. Thermal diffusion has been historically used, but is not particularly efficient. At low 3He concentrations, the difference in permeability of 3He and superfluid 4He provides good separation, and at higher 3He concentrations (> I%), cryogenic distillation is very effective. These have been demonstrated experimentally on a 3 L/h scale [8].
6. Helium-3
resource implications
Wittenberg et al. [9] recently estimated terrestial resources of 3He. Table 3 summarizes these estimates, showing helium-3 reserves that could be available in the year 2000, a time-frame consistent with the ITER schedule. The naturally occurring helium-3 resources are very dilute and, thus,’ are not economically attractive for ITER applications. The CANDU reactor resource is insufficient given an inventory requirement of - 25-50 kg and a yearly makeup rate of - 3.4-8.5 kg/yr (based on lo-25% availability/yr). Thus, tritium decay in
D. Steiner Table
et al. / ‘He blankets
3
Reserves [91
of helium-3
that could
Source
be available
Cumulative amount to year 2000 (kg)
Natural gas wells Underground storage Known reserves Decay of T, CANDU reactors US weapons (approximate) Total
weapons is the only could satisfy ITER
in the year
2000
Production rate post year 2000 (k&r)
29 187 10
source
15 17
125
to terrestial supplies of 3He), it will provide reactor-relevant information on helium-cooled blankets. While the technical features of the helium-3 blanket option are extremely attractive, the ability to guarantee the required helium-3 resources emerges as the major issue associated with the ultimate viability of this concept. Perhaps the initiation of the ITER project combined with the eventual ratification of the INF (Intermediate Nuclear Forces) treaty could provide the catalyst for an international agreement to supply the helium-3 requirements of the proposed blanket in ITER.
of helium-3
Acknowledgements
which
needs. One would expect that the US and the USSR have comparable sources of 3He associated with stores of thermonuclear weapons. Thus, the ultimate viability of the 3He blanket is tied to a political issue: would the relevant countries agree to make their 3 He resources available to the project. Clearly any opinion on the likelihood of such an agreement is speculative. Nevertheless, it is worth noting that the current international climate seems well suited to a proposal of this type.
7. Concluding
breeding
2
300 500 to600
potential
for tririum
remarks
A helium-3 blanket offers an attractive option for tritium breeding in near-term devices. The concept exhibits good tritium breeding potential, low tritium inventories, and low tritium leakage rates. Moreover, the concept retains the desirable safety and operational features inherent in gas-cooled blankets, including full compatibility with all reactor test blankets and the ability to operate the first walI at high temperatures for outgassing. There is no need to mechanically access the blanket in order to control breeding - the same blanket hardware can provide shielding or breeding controlled only by out-of-core addition of helium-3. The blanket performance is not adversely affected by variations in power level. Finally, although the concept is not extrapolatable to power reactors (assuming we are limited
This work was supported by the DOE, Office of Fusion Energy and by the Canadian Fusion Fuels Technology Project.
References
111C.P.C.
Wong, R.F. Bourque, E.T. Cheng, R.L. Creedon, 1. Maya, R.H. Ryder and K.R. Schultz, Helium-cooled blanket design, Fusion Technology, Vol. 8 (July 1985). PI R.D. O’Dell, F.W. Brinkley and D.R. Marr, User’s manual for ONEDANT: A code package for one-dimensional diffusion accelerated neutral particle transport, Los Alamos National Lab. Rep. LA-9184-M (Feb. 1982). TRANSX-CTR: A code for interfacing 131 R.E. MacFarlane, MATXS cross section libraries to nuclear transport codes for fusion system analysis, Los Alamos National Lab. Rep. LA-9863-MS (Feb. 1984). et al., Tritium systems concepts 141 S.K. Sood, O.K. Kveton for the Next European Torus (NET), CFFTP = G-86020 (Sept. 1986). draft report 151 G.S. Shaw et al., NET facility layout, CFFTP (1987). to US ITER Design Review Meet[61 J. Blevins, Contributions ing, CFFTP (1988). Heat Transfer, 2nd edn. (McGraw-Hill, [71 B. Gebhart, Toronto, 1971). distillation apparatus, MLM-2005, PI W.R. Wilkes, 3He-4He US Atomic Energy Commision (Mar. 1973). J.F. Santarius and G.L. Kulcinski, Lunar 191 L.J. Wittenberg, source of ‘He for commercial fusion power, Fusion Technology, Vol. 10 (Sept. 1986).