Space reactor power systems with no single point failures

Nuclear Engineering and Design 238 (2008) 2245–2255

Contents lists available at ScienceDirect

Nuclear Engineering and Design journal homepage: www.elsevier.com/locate/nucengdes

Space reactor power systems with no single point failures Mohamed S. El-Genk ∗ Institute for Space and Nuclear Power Studies and Chemical and Nuclear Engineering Department, The University of New Mexico, Albuquerque, NM 8713, USA

a r t i c l e

i n f o

Article history: Received 19 March 2007 Received in revised form 27 February 2008 Accepted 27 February 2008

a b s t r a c t Nuclear reactor power systems could revolutionize space exploration and support human outpost on the moon and Mars. This paper reviews various energy conversion technologies for use in space reactor power systems and provides estimates of the system’s net efficiency and specific power, and the specific area of the radiator. The suitable combinations of the energy conversion technologies and the nuclear reactors, classified based on the coolant type and cooling method, for best system performance and highest specific power, are also discussed. In addition, a number of power system concepts with both static and dynamic energy conversion, but with no single point failures in reactor cooling, energy conversion and heat rejection, and for nominal electrical powers up to 110 kWe , are presented. The first two power systems employ reactors cooled with lithium and sodium heat pipes, SiGe thermoelectric (TE) and alkali-metal thermal-to-electric conversion (AMTEC), and potassium heat pipes radiators. The reactors heat pipes operate at a fraction of the prevailing capillary or sonic limit, and in the case of a multiple heat pipes failure, those in the adjacent modules remove the additional heat load, thus maintaining the reactor adequately cooled and the power system operating at a reduced power. The third power system employs SiGe TE converters and a liquid metal cooled reactor with a divided core into six sectors that are neurotically and thermally coupled, but hydraulically decoupled. Each sector has a separate energy conversion loop, a heat rejection loop, and a rubidium heat pipes radiator panel. When a core sector experiences a loss-of-coolant, the fission power of the reactor is reduced, and that generated in the sector in question is removed by the circulating coolant in the adjacent sectors. The fourth power system employs a gas cooled reactor with a core divided into three identical sectors, and each sector is coupled to a separate Closed Brayton Cycle (CBC) loop with He–Xe binary mixture (40 g/mol) working fluid, a secondary loop with circulating liquid Nak-78, and two water heat pipes radiator panels. © 2008 Elsevier B.V. All rights reserved.

1. Introduction Future space exploration and lunar and Mars outposts would require a reliable power sources that operate 24/7 independent of the sun. Space nuclear reactor power systems could provide tens to hundreds of kilowatt electric continuously for 7–10 years, or even longer, and operate 24/7 independent of the sun. The availability of the solar option, at a rapidly increasing launch cost, practically diminishes with distance from earth. The solar isolation on Mars and Jupiter is ∼45% and only 4% of its value on earth surface. The absence of the solar option and the vast traveling distances to the planets and satellites in the solar system, strongly justify consideration of space reactor power systems to support future robotic and human exploration and outposts on the moon and Mars. These compact power systems could have a specific mass ≤50 kg/kWe (or specific power ≥20 We /kg), increasing the science

∗ Tel.: +1 505 277 5442. E-mail address: [email protected]. 0029-5493/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.nucengdes.2008.02.026

payload and the rate of data return, at a significantly reduced mission cost and duration to the farthest planets in the solar system. Examples are Jupiter and its icy moons Callisto, Ganymede, and Europa, Saturn, Neptune and the dwarf plant Pluto. Space reactor power systems would be started, for the first time, in an earth orbit, thus eliminating radiological concerns during launch from earth. In addition to the longevity and compactness, these systems could operate at multiple power levels and be designed for bimodal operation of electricity generation and propulsion (El-Genk, 2001). Desirable design and operation features of space reactor power systems are to avoid single point failures in reactor cooling, energy conversion, and heat rejection and the likelihood of a criticality of the bare nuclear reactors upon submersion in wet sand and being flooded with seawater, following a launch abort accident. The latter is typically accomplished using neutron poison materials, such as gadolinium, europium, etc., as additives to the nuclear fuel and a thin coating on the outer surface of the nuclear reactor vessel (King and El-Genk, 2006). The avoidance of a single point failure in reactor cooling could be accomplished by either: (a)

2246

M.S. El-Genk / Nuclear Engineering and Design 238 (2008) 2245–2255

dividing the nuclear reactor into a number of identical sectors that are thermally and neutronically coupled, but hydraulically decoupled, or (b) cooling the reactor with liquid metal heat pipes. Using separate energy conversion loops/modules, each with a separate radiator panels, not only enhances the power system reliability and redundancy, but also avoids single point failures in energy conversion and heat rejection (El-Genk, 2001; King and El-Genk, 2006; El-Genk and Tournier, 2004a,b, 2006). These design measures for avoiding single point failures in space reactor power systems come at the expense of increased system complexity and total mass, and may not be necessary for short duration missions and when operating at low temperatures. This paper reviews the various energy conversion technologies for use in space reactor power systems and quantifies the impact on the system’s net efficiency and specific power, and the specific area of the radiator. The suitable combinations of energy conversion technologies and nuclear reactor types, classified based on the coolant type and cooling method, for best system performance and highest specific power, are also discussed. In addition, a number of nuclear reactor power system concepts, with both static and dynamic energy conversion and no single point failures, for and nominal powers up to 110 kWe are presented (El-Genk and Tournier, 2004a,b, 2006). The first two systems have liquid–metal heat pipes reactors and thermoelectric (TE) and alkali-metal thermalto-electric conversion (AMTEC) units, for converting the reactor thermal power to electricity, and potassium heat pipes radiators. The third power system employs SiGe TE converters and a liquid metal cooled reactor with a divided core into six identical sectors. Each sector has a separate energy conversion loop, a heat rejection loop, and a rubidium heat pipes radiator panel. The fourth power system has a gas cooled reactor, with a sectored core. Each of the three sectors in the core is coupled to a separate Closed Brayton Cycle (CBC) loop with He–Xe (40 g/mol) working fluid and a Nak-78 secondary loop, and two separate water heat pipes radiator panels. 2. Choices of reactor type and energy conversion There are many possible combinations of reactor type and energy conversion technology, but there are a fewer matching choices, for best operation compatibility, namely: (a) Liquid metal heat pipes cooled reactors with energy conversion technologies of TE (El-Genk and Tournier, 2004a,b; El-Genk and Saber, 2003; Mondt et al., 2005), AMTEC (El-Genk and Tournier, 2004b; Cole, 1983; Tournier and El-Genk, 2003), or free-piston stirling engines (FPSEs) (Wang et al., 1992; Schreiber, 2001; Thieme et al., 2002; Docha, 1992). (b) Liquid-metals cooled reactors with TE, AMTEC, FPSEs, or thermionic (TI) energy conversion (Paramonov and El-Genk, 1994, 1996; El-Genk and Paramonov, 1999; Mills and Van Hagan, 1994). (c) Gas-cooled reactors and CBC, with single-shaft, centrifugal flow turbo-machines (Gallo and El-Genk, 2008; Harty and Mason, 1993; Marriot and Fujita, 1994; Lipinski et al., 2002; Gallo et al., 2007; El-Genk, 1994). In addition to the operation and integration compatibility, the selection of an energy conversion technology is based on considerations of safety, scalability, modularity, load following, system integration, radiation hardness, efficiency, reactor exit temperature and the average heat rejection temperature and size. The following subsections briefly review various static and dynamic conversion technologies for use in space nuclear reactor power systems.

2.1. Static energy conversion In addition to the absence of moving parts, static conversion technologies are inherently modular and load following (El-Genk and Tournier, 2004a,b, 2006; El-Genk and Saber, 2003; Fleurial et al., 1997; Saber and El-Genk, 2002; Calliat et al., 2000; El-Genk and Saber, 2005a,b; Mondt et al., 2005; Paramonov and El-Genk, 1994, 1996; El-Genk and Paramonov, 1999; Mills and Van Hagan, 1994; Virkar et al., 2000; Williams et al., 1999; Rayn et al., 2000; Tournier and El-Genk, 2003). The SiGe thermoelectric converters operating between 1273 and 790 K have an efficiency of ∼6%, representing ∼16% of Carnot. Thus, for a 100 kWe space reactor power system with SiGe TE, the surface area of the heat rejection radiator could exceed 100 m2 and the reactor exit temperature could be as high as 1373 K, requiring refractory alloys structure materials such as PWC11 (Nb–1%Zr, 0.1% C), Mo–Re, and TZM (El-Genk and Tournier, 2005). These alloys are heavy, and those of the niobium are incompatible with oxygen, thus could not be used on Mars without a protective coating. On the other hand, SiGe converters had been very reliable, with a space flight experience in >25 space missions powered by radioisotope thermoelectric generators (RTGs). The segmented thermoelectric (STE) converters are more efficient than SiGe because the materials of the segments in the n- and p-legs operate in the temperature range in which they possess the highest figure-of merit (FOM), or Z values. For example, a SiGe unicouple operating between a hot-side temperature, Th = 1273 K and a heat rejection temperature, TR = 700 K, Z ∼ 0.7 × 10−3 K−1 , while for STE operating between Th = 973 K and TR = 300 K, Z ∼ 1.15 × 10−3 K−1 and the efficiency ∼14.8%. A number of STE converters have been fabricated, using p-type CeFe4 CoSb12 and Bi2 Te3 -based alloys and n-type CoSb3 and Bi2 Te3 -based alloys, and tested at cold and hot shoe temperatures of 300 and 973 K, respectively (El-Genk and Saber, 2003; Fleurial et al., 1997; Saber and El-Genk, 2002; Calliat et al., 2000). At a cold shoe temperature of 300 K, the efficiency of a space reactor power system that uses STE could be ∼13% (Fig. 1), but the radiator’s specific area would ∼15.8 m2 /kWe (Fig. 2). Increasing the radiator temperature to 373 and 573 K decreases the projected system efficiency with STE converters to 12.4 and 7.8% (Fig. 1), but decreases the radiator specific area to 8.0 and 2.45 m2 /kWe (Fig. 2). The estimates in Figs. 1 and 2 assume a power system efficiency that is 90% of that of the converter. The cascaded SiGe converter, operating between 1273 and 973 K, with a STE bottom converter at average radiator temperature of 300 K, could increase the efficiency of the nuclear reactor power system to ∼18% (Fig. 1) and decrease the specific area of the radiator to 10.9 m2 /kWe (Fig. 2) (El-Genk and Saber, 2005a,b). Increasing the cold shoe temperature for these converters to 573 and 673 K, decreases the system’s efficiency to ∼12.9 and 10.2% (Fig. 1),

Fig. 1. Efficiency estimates for space reactor power systems with different energy conversion technologies.

M.S. El-Genk / Nuclear Engineering and Design 238 (2008) 2245–2255

Fig. 2. Estimates of radiator specific area for space reactor power systems with different energy conversion.

at which the specific areas of the radiator are only ∼1.35 and 0.85 m2 /kWe (Fig. 2). The relatively lower hot-side temperature for AMTECs decreases the reactor exit temperature to ∼1180 K (El-Genk and Tournier, 2004b; Mills and Van Hagan, 1994; Virkar et al., 2000; Williams et al., 1999; Rayn et al., 2000; Tournier and El-Genk, 2003), making it possible to use relatively lighter super-steels, titanium alloys, or mechanically alloyed oxides dispersed steels (MA-ODS) as structure materials (El-Genk and Tournier, 2005). While the AMTECs technology is currently at a technology readiness level3 (TRL-3) and could have a conversion efficiency in excess of 20% (El-Genk and Tournier, 2004b; Tournier and El-Genk, 2003), they have never been used in space. Thus, in order to be considered for deployment sometime within a decade, AMTEC technology needs to be advanced to TRL-5. AMTEC units typically operate at moderate hot-side temperatures, Th , of 1000–1123 K and relatively high radiator temperatures, TR , of 550–650 K, with potassium and sodium working fluid at a net conversion efficiency of 22–27% (El-Genk and Tournier, 2004b; Tournier and El-Genk, 2003). Depending on the type of alkali-metal working fluid (K or Na), the AMTEC units employ thin (≤0.5 mm) membrane of K-beta “alumina or Na-beta” alumina solid electrolyte (BASE) mills. The BASE material has a “spinel” crystal structure of extended alumina (Al2 O3 ) blocks separated by ion conductions planes of loosely packed alkali-metal ions and equal number of O2− ions (Cole, 1983; Virkar et al., 2000). The spinel blocks are low activation energy barriers for the alkali-metal ions, which jump from one site to the next in the conduction planes and in the direction of the applied pressure difference across the BASE (Cole, 1983). The spinel crystal structure causes the electron conductivity of the BASE to be very low and the ionic conductivity to be very high. The vapor pressure on the anode side of the BASE (40–80 kPa) is much higher than on the cathode side (20–60 Pa). This pressure difference is balanced by the electrochemical potential resulting from the diffusion of the alkali-metal ions to the cathode side of the BASE. The electrons from the anode side circulate through the external load to the cathode side where they recombine with the alkali-metal ions emerging from the BASE, producing lowpressure sodium vapor (<80 Pa). This vapor diffuses through the porous cathode electrode (∼1 ␮m thick WRh1.5 ) and traverses the low-pressure cavity to a remote condenser. The resulting liquid condensate is then circulated back through a porous artery by the capillary action generated in the surface pores (<4 ␮m in radius) of the evaporator wick in the high-pressure cavity (Tournier and El-Genk, 2003). The electrochemical potential generated across the BASE is typically ≤0.5 V and the electrical current is proportional to the cathode

2247

Fig. 3. Specific power estimates for space reactor power systems with different conversion technologies.

electrode area and the circulation rate of the liquid metal working fluid in the AMTEC units. Typical operating current density for the cathode electrode is ∼0.2–0.5 A/cm2 . Structure materials, which are most compatible with liquid metal working fluids, the BASE, and the electrode materials are molybdenum-rhenium (Mo–Re) alloys (El-Genk and Tournier, 2005). A Na-AMTEC unit (4–6 kWe each), operating between 1123 and 650 K, could have an efficiency of ∼26% (El-Genk and Tournier, 2004b; El-Genk and Tournier, 2003a,b; Tournier and El-Genk, 2003), representing ∼60% of Carnot, and the radiator area for a 100 kWe nuclear reactor power system with these units would be <25 m2 (Figs. 1 and 2). 2.2. Dynamic energy conversion Dynamic energy conversion technologies of CBC and FPSE, with rotating and linear alternators, typically operate at low radiator temperature (<400 K), but high conversion efficiencies (23–35%), and are inherently radiation hard (El-Genk and Tournier, 2006; Wang et al., 1992; Schreiber, 2001; Thieme et al., 2002; Docha, 1992; Harty and Mason, 1993; Marriot and Fujita, 1994; Lipinski et al., 2002; Mason et al. 2002). The low radiator temperature significantly increases its size and mass and those of the power system (Figs. 1 and 2). Fig. 3 provides estimates of the specific power of space reactor power systems with static and dynamic energy conversion technologies. With TE and TI, the specific power of the power system ranges from 5 to 15 We /kg (Fig. 3) and may exceed 30 We /kg with AMTEC (El-Genk and Tournier, 2004b; Tournier and El-Genk, 2003). With dynamic conversion technologies that include K-Rankine cycle, CBC, FPSE, the system’s specific power could be 10–30 We /kg (Fig. 3). 3. Space reactor power systems This section presents four space nuclear reactors and power systems for the avoidance of single point failures in reactor cooling, energy conversion, and heat rejection. The reactors presented are cooled with either liquid metal heat pipes or circulating liquid metal and noble gas binary mixture of He–Xe (40 g/mol). For the latter, the reactor cores are divided into six and three sectors. The sectors are thermally and neutronically coupled, but hydraulically decoupled, and each has its own energy conversion primary loop, secondary loop, and heat rejection radiator panels. Fig. 4 presents several integration options of space reactor power systems with sectored and non-sectored gas and liquidmetal cooled reactors. Sectored the cores avoid a single point failure in reactor cooling (Fig. 4a and b) but, unless each sector has its own coolant loop (Fig. 4a), the power system could have a single point failure (Fig. 4b), if all coolant loops have common plenums. Thus, a break in any of the loops would cause a complete loss of

2248

M.S. El-Genk / Nuclear Engineering and Design 238 (2008) 2245–2255

Fig. 4. Several layouts of space reactor power systems.

the coolant from the system. With non-sectored reactors (Fig. 4c), using multiple primary loops increases system redundancy, but the common hot and cold plenums represent single point failures in the system. The avoidance of single point failures increases the operation redundancy (Fig. 4a), but at the expense of increased mass and complexity of the power system. The system integration in Fig. 4a may be considered, however, for long duration missions of more than a few years. For shorter missions, in which the reactor operates below 900 K and employs fuel and structural materials with well known properties, a simple system integration may be adequate. The following section presents reactor designs and space power systems developed at the University of New Mexico’s Institute for Space and Nuclear Power Studies for the avoidance of single point failures (El-Genk, 2001; El-Genk and Tournier, 2004a,b, 2006).

3.1. SAIRS—scalable AMTEC integrated reactor space power system Fig. 5a and b shows cross-sectional views of the SAIRS power system (El-Genk and Tournier, 2004b), developed with no single point failures. It employs a fast neutron spectrum nuclear reactor, cooled with a multitude of sodium (Na) heat pipes, and 18, 5.6 kWe Na-AMTEC units, or 24, 4.2 kWe Na-AMTEC units (Tournier and ElGenk, 2003). The AMTEC units are divided into six blocks of 3 or 4 units and each block is cooled by a multitude of potassium (K) heat pipes in a separate radiator panel (Fig. 5b). The AMTEC units, placed behind the radiation shadow shield, are heated by a multitude of sodium heat pipes, which are thermally coupled, using heat pipe locks, to the reactor’s Na heat pipes. The six AMTEC blocks in the power system are connected electrically in parallel, and the units

Fig. 5. SAIRS, a 100-kWe space power system with heat pipes cooled reactor and AMTEC units.

M.S. El-Genk / Nuclear Engineering and Design 238 (2008) 2245–2255

2249

Fig. 6. Cross-sectional views of SAIRS reactor and radiation shadow shield.

in each block are connected in series, providing excellent redundancy, while supplying electrical power to the load at high (>176 V) dc voltage. Thus, with a failure of more than one AMTEC unit in one or more blocks, the power system will continue to operate, but at reduced reactor thermal power (El-Genk and Tournier, 2004b; Tournier and El-Genk, 2003). When the AMTEC units operate nominally at ≤95% of their peak electrical power, the net efficiency of the power system (Fig. 5) is 22.7–27.3%, the terminal voltage ≥50 V dc and the current ≤100 A (Tournier and El-Genk, 2003). The estimate of the system’s specific power is 29.7–34.8 We /kg and of the radiator area is 21.1–26.9 m2 . The heat rejection radiator panel for each AMTEC block is made of two sections, a stationary forward section attached to the AMTEC units, and a rear conical deployable section, which is folded onto the stationary section in the stowed launch configuration (Fig. 5a). The K-heat pipes in the two sections of the radiator panels are connected using flexible joints and the surface of the radiator is armored to protect against impact by meteoroids (El-Genk and Tournier, 2004b). The specific mass of the potassium heat pipes radiator of 7.67 kg/m2 is the same as that for the SP-100 system (Marriot and Fujita, 1994). The SAIRS hexagonal, fast-spectrum nuclear reactor is cooled by 60, 1.5 cm o.d., Mo–14%Re sodium heat pipes. Each heat pipe serves a module comprised of three UN, Re clad, fuel pins arranged in a triangular lattice and brazed to a centrally positioned heat pipe, with a Mo–14%Re wall of the same o.d. (1.5 cm). Six, Re tri-cusps braze the heat pipes wall (0.4 mm thick) to the Re cladding of the fuel pins along their active length (Tournier and El-Genk, 2003). The UN fuel modules placed in the reactor core are delineated in Figs. 6a and b. Fig. 5a shows the routing of the reactor sodium heat pipes after exiting the reactor. The shadow radiation shield (Fig. 6a) consists of a 2.5-cm thick tungsten layer for attenuating the high energy gammas rays, followed by a 50 cm thick LiH, for attenuating the fast neutrons from the nuclear reactor. To minimize the physical penetrations through the shield, the heat pipes exiting the reactor vessel and the axial BeO reflector are bent around the shadow shield before entering the radiator cavity (Figs. 5a and 6a). These heat pipes are structurally supported using a 2-mm thick graphite disk placed in front, but thermally insulated from, the radiation shadow shield (Fig. 6a) (El-Genk et al., 2005). The heat deposited in the shield (<6 kW) by the attenuation of the fast neutrons and the primary gammas from the reactor is dissi-

pated by thermal radiation into space from the conical side surface of the shield, keeping the LiH temperature within a desirable range of 600–680 K (Barattino and El-Genk, 1985). The reactor heat pipes are kept 3.0 cm away from the side surface of the shield, to allow the emitted thermal radiation to stream into space and, do not obstruct the drive shafts (or actuators) of the 12 BeO/B4 C control drums in the radial reflector (Fig. 6a and b). The section of the Na-heat pipes that extends from the reactor to the AMTEC blocks is thermally insulated using multi-foils insulation (MFI). The length of the evaporator section of these heat pipes is the same as the active height of the UN fuel in the reactor core (42 cm), and the length of the condenser section is the same as that of the AMTEC units block (1.23–1.64 m). Depending on the radial location in the reactor core, the length of the adiabatic section of the reactor Na-heat pipes is 1.84–1.9 m. For a 100 kWe space power system, the reactor thermal power is quite low, 452–542 kW (Tournier and El-Genk, 2003). In order to minimize the diameter and the mass of the radiation shadow shield (628 kg) as well as the minor diameter of the radiator, the shadow shield is placed as close to the reactor as possible (25 cm), and still maneuver the reactor heat pipes around the shield before entering the radiator cavity (Figs. 6a and 7). 3.2. Heat pipe cooled reactor with segmented thermoelectric module converters (HP-STMCs) A heat pipe cooled nuclear reactor, similar to that of SAIRS, has been developed and integrated to STE energy conversion modules and a heat pipes radiator in the HP-STMCs space reactor power system, nominally generates 110 kWe (Fig. 8) (El-Genk and Tournier, 2004b; Mondt et al., 2005). The hexagonal heat pipe reactor core in HP-STMCs power system is comprised of 126 heat pipe-fuel modules (El-Genk and Tournier, 2004b). Each module consists of three UN, Re clad fuel pins arranged in a triangular lattice and brazed to a central lithium heat pipe with a Mo–14%Re wall of the same o.d. (1.5 cm) as the fuel pins. The evaporator length of the lithium heat pipes is the same as that of the UN fuel stack in the pins (45 cm). The thickness of the radial reflector surrounding the Mo–14%Re reactor vessel varies from 7.1 to 9.9 cm and the half-cone angle for the system is 15◦ (Fig. 8). The lithium heat pipes reactor, including the radial and the axial BeO reflectors, has a 56 cm o.d. and is 66.1 cm tall (El-Genk and Tournier,

2250

M.S. El-Genk / Nuclear Engineering and Design 238 (2008) 2245–2255

Fig. 7. Isometric view of 100-kWe SAIRS power system with approximate dimensions.

2004b). The active core is 45 cm long and 35.4 cm wide flat-to-flat (Fig. 9a and b). The control of the reactor is accomplished using a total of 12 BeO/B4 C rotating drums in the radial reflector. The B4 C segments (5 mm-thick 120◦ sectors) in the drums face the reactor core during launch and at startup and face away from the reactor core at the end of life (Figs. 6b and 9a). The nominal thermal power of the reactor is 1.6 MW and the average design power throughput per lithium heat pipe is 12.7 kW. This power throughput represents an operation design margin of at least 28% relative to the prevailing wicking limit (Tournier and El-Genk, 2004). An auxiliary radiator with a surface area of 1.0-m2 cools the instrumentation and control equipment placed behind the shield (Fig. 8). The radiator of the HP-STMCs power system (Fig. 8) is comprised of two sections: (a) a front section, in which the evaporators of a multitude of potassium heat pipes are conductively coupled to the cold side of STMCs and radiates into space from the opposite side and (b) a rear section, in which the evaporators of the potassium heat pipes are conductively coupled to the bottom

surface of a second set of STMCs that are thermally insulated on the opposite side (Fig. 8). The number of the K-heat pipes (162) in the front and the rear sections of radiator are the same, but their diameter in the rear section is slightly larger to ensure their nominal operation at 43.6% of the sonic limit (El-Genk and Tournier, 2004a; Tournier and El-Genk, 2004). During nominal operation, the hotside temperature of the STMCs is constant at equal 1300 K, for a SiGe hot junction temperature ≥1270 K, and the evaporator temperature of the K-heat pipes in the radiator is 776 K. At these temperatures, the efficiency of the power system is 6.7% and the net electrical power is 110 kWe (Fig. 8). For a total system mass of 4261 kg, this electrical power corresponds to a specific power of ∼26.8 We /kg for the intergraded power system. 3.3. Sectored, compact reactor (SCoRe) space power system The SCoRe is convectively cooled with a circulating liquid metal of NaK (78% Na), sodium, or lithium, in order of increasing reac-

Fig. 8. A layout of the reference 110 kWe , HP-STMCs space reactor power system.

M.S. El-Genk / Nuclear Engineering and Design 238 (2008) 2245–2255

2251

Fig. 9. Radial and axial cross-sectional views of HP-STMCs lithium heat pipes reactor.

tor temperatures of 850, 1100, and >1100 K. The hexagonal reactor core is divided into six, thermally and neutronically coupled, but hydraulically decoupled sectors, and surrounded by a 16-cm thick radial BeO reflector and has 4.0 cm thick axial BeO reflectors. The reactor control is accomplished using 6, BeO rotating control drums, in the radial reflector, with 0.5 mm thick, 120◦ , B4 C segments (Fig. 10a). The liquid metal coolant enters the reactor and flows upward in an annulus on the inside of the reactor vessel wall then reverses direction at the opposite end to flow through the core sectors, then to the lower plenums from which it exits the reactor core (Fig. 10b) (El-Genk et al., 2005). The coolant inlet annuli, the exit plenums, and the dome of the six sectors of the reactor core are physically separated. In the event of a break in the inlet or the exit pipe of a sector, resulting in a loss of coolant (LOC), the reactor continues to operate, but at a lower power. The common dividers between the sectors in the reactor core and the inner wall of the coolant inlet annulus are made of flatplate, liquid metal heat pipes. These heat pipes facilitate the passive

cooling of the reactor sector experiencing a LOC, by transporting the fission generated heat in it to the circulating coolant in the two adjacent sectors and flowing in the inlet annulus (Fig. 10a and b). The SCoRe core is loaded with UN fuel pins, 0.74 cm o.d., arranged in a triangular lattice with a pitch of 0.8 cm. The UN fuel pellets have as-fabricated porosity of 10%, and the Mo–14%Re cladding has a wire wrap (0.6 mm o.d.) brazed in a helical pitch onto its surface. The Mo–14%Re wires, which run the length of the fuel pins, maintain uniform spacing between the pins, provide structural support, and create a spiraling coolant flow path for enhanced convective cooling. The volume of the gas plenum is ∼23% of the UN fuel volume in the pins, and together with the radial gap (0.05 mm) between the pellets and the cladding, accommodates the fission gases released during reactor operation, reducing the induced stresses in the cladding (El-Genk et al., 2005). Fig. 11a shows an isometric view of the space power systems with SCoRe-11 and the radiation shadow shield. This shield consists of a front layer of depleted-LiH, fol-

Fig. 10. Cross-sectional views of SCoRe for space power systems.

2252

M.S. El-Genk / Nuclear Engineering and Design 238 (2008) 2245–2255

Fig. 11. An isomeric view and a radial cross-section of SCoRe-S.

lowed by a thin layer of tungsten, then a thicker layer of natural-LiH. Fig. 11b shows a radial cross-section of the SCoRe-S11 core loaded with a total of 1026 UN fuel pines, 171 pins in each of the six core sectors. When these pins operate at an average power of 2.5 kW, SCoRe-S11 generates 2.86 MW of thermal power. Each sector in the SCoRe-S11 core is thermally coupled to a separate pair of primary and secondary liquid metal loops and a rubidium heat pipes radiator panel for waste heat rejection. Fig. 12 shows an isometric view of the fully integrated space reactor power system powered by a SCoRe-S11 and employing SiGe power conversion assemblies (PCAs), placed between the primary and the secondary liquid metal loops. The liquid metal coolants in these loops are circulated using separate electro-magnetic pumps powered by SiGe thermoelectric conversion assemblies (TCAs), also placed between the primary and secondary loops (Fig. 12). The nominal working fluids in the secondary and primary loops could be any combinations of Li/Li, Li/NaK, Na/Na, or Na/NaK, depending on the selected operation temperatures of the power system (El-Genk and Tournier, 2004b). Each primary and secondary loop has its own liquid-metal accumulator (Tournier and El-Genk, 2006), that is designed to maintain appropriate coolant pressures and accommodates the volume expansion of the liquid metal coolants with increased temperature during the system startup and operation transients. At nominal operation, the accumulators in the Li primary and secondary loops are 63 and 68% of full capacity, providing enough operation margins (Fig. 13). The

performance results presented in Fig. 13 are for a power system with Li/Li, SCoRe-S11 and SiGe converters. This space power system has no single point failures in either reactor cooling or energy conversion (El-Genk and Tournier, 2006a). 3.4. Submersion-subcritical safe space (S4 ) reactor power system The S4 reactor are cooled with a He–Xe binary mixture having a molecular weight of 40 g/mol and nominally operate at 471 kWth . Fig. 14a presents a radial cross-section of the S4 reactor core and Fig. 14b presents a coolant channels-fuel elements unit cell in the core (King and El-Genk, 2007). The S4 reactor has a hexagonal, Mo–14Re (molybdenum with 14 wt.% rhenium) solid core with cavities loaded with uranium nitride fuel pellet and surrounded by coolant channels (Fig. 14b). The UN fuel stacks in the reactor core are 1.25 in diameter and arranged in a triangular lattice with a pitch of 1.779 cm. Each fuel stack is cooled by a total of nine coolant channels that are 3 mm in diameter (Fig. 14b) (King and El-Genk, 2007). The S4 reactor core is divided into three sectors, that are thermally and neutronically coupled, but hydraulically decoupled. Each reactor sector has its own CBC loop with a separate Nak secondary loop and heat rejection radiator panels (Fig. 15) (El-Genk and Tournier, 2006). The high thermal conductivity of the solid core block (>68 W/m-K) facilitates the transfer of the heat generated by fission from a sector experiencing a LOC to the adjacent sectors by conduction. In this case, the reactor and the power sys-

Fig. 12. SCoRe-S11 , space reactor power system with SiGe TE converters for 111.5 kWe .

M.S. El-Genk / Nuclear Engineering and Design 238 (2008) 2245–2255

2253

Fig. 13. A pair of primary and secondary coolant loops for one sector of SCoRe-S11 space power system.

tem continues to operate, but a lower power level (El-Genk and Tournier, 2006a). Fig. 15 presents the layout of one of the CBC loops of the space power system that employs an S4 reactor (Fig. 14) and a water heat pipes radiator (El-Genk and Tournier, 2006). At nominal steady state operation, the S4 space power system generates a net electrical power to the load of 93.6 kWe , at a net efficiency of 21.4%. The nominal pressure at the exit of the compressor units in the CBC loops is 2.0 MPa and the CBC loops employ single-shaft, centrifugal flow turbo-machines, rotating at 36,000 rpm (Mason et al., 2002; Gallo et al., 2007; Gallo and El-Genk, 2008) (Fig. 15) and He–Xe (40 g/mol) working fluid (El-Genk and Tournier, 2006). The power system continues to operate with only two CBC engines, but at a lower reactor power, following a failure of one of the turbomachines, a loss-of-cooling, or a break in one of the CBC loops. The fission power generated in the reactor sector connected to the failed CBC loop is transported by conduction and/or radiation to the

dividers with the two adjacent sectors, where it is then removed by forced convection of the circulating gas in these sectors. The power system layout is identical to that in Fig. 12, with a 30◦ cone angle and a minor radiator diameter of 1.08 m. The six panels of the radiator have an external surface area of 168.9 m2 and an effective heat rejection area of 203 m2 . The two radiator panels dedicated to each of the three CBC loop in the power system are hydraulically connected in parallel to reduce pressure losses and the NaK-78 inventory in the secondary loop and hence, the volume and mass of the liquid metal accumulator (Fig. 15) (Tournier and El-Genk, 2006). The performance values shown in Fig. 15 are for gas cooler and recuperator effectiveness of 0.97 and 0.95 (El-Genk and Tournier, 2006). The NaK-78 (78 wt.% Na and 22 wt.% K), circulating in the secondary loop using an alternative linear induction pump (ALIP), transports the thermal power extracted from the He–Xe in the gas

Fig. 14. A radial cross-section of and the arrangement of UN fuel and coolant channels in S4 reactor.

2254

M.S. El-Genk / Nuclear Engineering and Design 238 (2008) 2245–2255

Fig. 15. Operation parameters of one of the three CBC in S4 reactor power system for nominal 93.6 kWe .

cooler to be rejected into space by the water heat pipe radiator panels. The NaK-78 coolant enters the radiator panels at 530 K and exits at 395 K. Each radiator panel nominally rejects 54 kWth and the total thermal power generated in the three sectors of the reactor is 437 kWth , which slightly lower that its nominal design value of 471 kWth (El-Genk and Tournier, 2006). The turbo-machines in each of the three CBC loops deliver a net electric power to the load of 31.2 kWe , after accounting for 10% transmission and inversion losses and the electrical power supplied to the ALIP. This electrical power represents at a power system net efficiency of 21.4% (El-Genk and Tournier, 2006). 4. Summary and conclusions Future space exploration and human outposts would require high power sources that operate 24/7 for more than 10 years, independent of the sun. Such power sources are likely to use fast neutron spectrum, nuclear fission, reactors and either static or dynamic energy conversion. Owing to the need for high reliability, these power systems would be based on reactor and energy conversion technologies and nuclear fuel and structure materials with demonstrated performance and well known properties. Other desirable attributes for such power systems are the absence of a single point failure and, preferably, load-following characteristics. This paper reviewed the various choices of energy conversion technologies for use in space nuclear power systems, and quantified the impact on the system’s net efficiency and specific power, and the specific area of the radiator. Also discussed are the suitable combinations of energy conversion technologies and reactor types, classified based on the coolant type and cooling method, for best system performance and highest specific power. Static conversion technologies of TE, TI and AMTEC are inherently modular and load following and have no moving parts, thus adaptable for operation at varying electrical power levels, depending on the space mission profile. The estimates of the system net efficiency with SiGe converters is 4.5%, 6–8% with TI, 8–10% with STE, 10–12% with cascaded SiGe-STE and 20–26% with high power NaAMTEC units. Static converters, connected electrically in series in several parallel strings, increase redundancy in electrical power

generation, with no or little structural complexity and increase in mass. Static energy conversion options, typically operate at high heat flux (8–10 W/cm2 for AMTEC and 10–30 W/cm2 for TE), thus are most suitable to use in conjunction with a liquid metal cooled reactors, operating at an exit temperature ≤1200 K with AMTEC and up to 1400 K with TE, an average radiator temperature of 500–600, and 500–800 K, respectively. TE conversion devices operate at a small fraction (0.1–0.18) of the ideal Carnot cycle efficiency and are sensitive to exposure to fast neutrons and gammas photons from the reactor, thus are placed at a distance behind the radiation shadow shield. The AMTEC units, however, may not be as sensitive to radiation exposure, bending experimental confirmation, and operate at the highest fraction (up to ∼0.6) of the Carnot efficiency of all static and dynamic technology technologies discussed in this paper. The dynamic conversion technologies of FPSE, K-Rankine cycle and CBC are inherently radiation hard. Although operate at high conversion efficiencies (20–30%), they reject waste heat at low temperatures (<400 K), resulting in voluminous and massive heat rejection radiator and power system. Because they are not load following, dynamic conversion technologies could not easily operate at variable power levels, without risking potential instabilities and structural stresses and vibrations. Instead, space reactor power systems with dynamic conversion dump excess electrical power on board using electrical heating elements with an auxiliary radiator, adding to the total mass of the power system and increasing the fuel burnup in the reactor. The increase in the fuel burnup, decreases the operation lifetime of the reactor and/or increase the fissile fuel loading for longer duration missions. A key objective of this paper was to discuss methods for avoiding single point failures in space reactor power systems and present conceptual designs of nuclear reactor power systems, developed at the University of New Mexico, with both static and dynamic energy conversion, but no single point failures in reactor cooling, energy conversion, and heat rejection. The first two power systems employ reactors cooled with lithium and sodium heat pipes and SiGe thermoelectric (TE) and AMTEC static energy conversion units and potassium heat pipes radiators. The reactors heat pipes operate

M.S. El-Genk / Nuclear Engineering and Design 238 (2008) 2245–2255

at a fraction of the prevailing capillary or sonic limit, and in case of a multiple heat pipes failure, those in the adjacent modules remove the additional heat load, thus maintaining the reactor adequately cooled and the power system operating at a reduced power. The third power system employs a liquid metal cooled reactor with a divided core into six identical sectors that are neurotically and thermally coupled, but hydraulically decoupled. Each sector has a separate energy conversion loop, a heat rejection loop, and a rubidium heat pipes radiator panel. When a core sector experiences a loss-of-coolant, the fission power of the reactor is reduced, and that generated in the sector in question is removed by the circulating coolant in the adjacent sectors. The fourth power system employs a gas cooled reactor with a core divided into three identical sectors. Each sector is coupled to a separate CBC loop with He–Xe binary mixture (40 g/mol) working fluid, a secondary heat rejection loop with a circulating Nak-78, and two water heat pipes radiator panels. References Barattino, W., El-Genk, M.S., Voss, S., 1985. Review of previous shield analysis for space reactors. In: El-Genk, M.S., Hoover, M.D. (Eds.), Space Nuclear Power Systems 1984, 2. Orbit Book Company, Inc., Malabar, FL, pp. 329–339 (CONF840113). Calliat, T., et al., 2000. High efficiency segmented thermoelectric unicouples. In: Proceedings of the Space Technology and Applications International Forum (STAIF-2000), AIP Conference Proceedings No. 504, American Institute of Physics, Melville, NY, pp. 1508–1512. Cole, T., 1983. Thermoelectric energy conversion with solid electrolytes. Science 221, 915–920. Docha, G., 1992. Free piston stirling component power converter test results and potential applications. In: Proceedings of the 27th Intersociety Energy Conversion Engineering Conference, San Diego, CA. El-Genk, M.S., 2001. A high-energy utilization, dual-mode system concept for mars missions. J. Propul. Power 17 (2), 340–346. El-Genk, M.S., Tournier, J.-M., 2003a. High Power AMTEC Converters for Deep-Space Nuclear Reactor Power Systems. In: Proceedings of the Space Technology and Applications International Forum (STAIF-03), AIP Conference Proceedings No. 654, American Institute of Physics, Melville, NY, pp. 730–739. El-Genk, M.S., Tournier, J.-M., 2003b. Conceptual Design of a 100-kWe Space Nuclear Reactor Power System with High-Power AMTEC. In: Proceedings of the Space Technology and Applications International Forum (STAIF-03), AIP Conference Proceedings No. 654, American Institute of Physics, Melville, NY, pp. 397– 407. El-Genk, M.S., Tournier, J.-M., 2004a. Conceptual design of HP-STMCs space reactor power system for 110 kWe. In: Proceedings of the Space Technology and Applications International Forum (STAIF-2004), AIP Conference Proceedings No. 669, American Institute of Physics, Melville, NY, pp. 658–672. El-Genk, M.S., Tournier, J.-M., 2006a. DynMo-TE: dynamic simulation model for space reactor power systems with thermoelectric converters. J. Nucl. Eng. Des. 236 (23), 2501–2529. El-Genk, M.S., Tournier, J.-M., 2004b. SAIRS—scalable AMTEC integrated reactor space power system. Prog. Nucl. Energy 45 (1), 25–69. El-Genk, M.S., Tournier, J.-M., 2006. Selection of noble gas binary mixtures for Brayton space reactor power systems. In: Proceedings of the International Energy Conversion Engineering Conference, IECEC-2006, Paper No. AAIA-2006-4168, San Diego, CA. El-Genk, M.S., Saber, H.H., 2003. High efficiency, segmented thermoelectric for operation between 973 K and 300 K. J. Energy Conv. Manage. 44 (7), 1069–1088. El-Genk, M.S., Saber, H.H., 2005a. Performance analysis of cascaded thermoelectric converters for advanced radioisotope power systems. J. Energy Conv. Manage. 46, 1083–1105. El-Genk, M.S., Saber, H.H., 2005b. Performance and mass estimates of CTM-ARPSs with 4-GPHS bricks. In: Rowe, D.M. (Ed.), CRC Thermoelectrics Handbook: Macro to Nano. CRC Press, NY, pp. 541–5414. El-Genk, M.S., Paramonov, D., 1999. Thermionic conversion. In: Webster, J.G. (Ed.), Encyclopedia of Electrical and Electronics Engineering, 22. John Wiley & Sons, Inc., pp. 49–67. El-Genk, M.S., 1994. A Critical Review of Space Nuclear Power and Propulsion 1984–1993. AIP Press, New York, NY, pp. 21–86. El-Genk, M.S., Tournier, J.-M., 2005. Review of refractory metal alloys and mechanically alloyed-oxide dispersion strengthened steels for space nuclear power systems. J. Nucl. Mater. 340, 93–112. El-Genk, M.S., Hatton, S., Fox, C., Tournier, J.-M., 2005. SCoRe—concepts of liquid metal cooled space reactors for avoidance of single-point failure. In: Proceedings of the Space Technology and Applications International Forum (STAIF-2005), AIP Conference Proceedings No. 746, American Institute of Physics. Melville, NY, pp. 473–484.

2255

Fleurial, J.-P., et al., 1997. New materials and devices for thermoelectric applications. In: Proceedings of the 32nd Intersociety Energy Conversion Engineering Conference, AIChE, New York, pp. 1080–1085. Gallo, B.M., El-Genk, M.S., Tournier, J.-M., 2007. Compressor and turbine models of Brayton units for space nuclear power systems. In: Proceedings of the Space Technology and Applications International Forum (STAIF-07), AIP Conference Proceedings No. 880, American Institute of Physics, Melville, NY, pp. 472– 482. Gallo, B.M., El-Genk, M.S., 2008. Performance analysis of 38 kWe turbo-machine unit for space reactor power systems. In: Proceedings of Space Technology and Applications International Forum (STAIF-08), AIP Conference Proceedings No. 969, American Institute of Physics, Melville, NY, pp. 625– 636. Harty, R., Mason, L., 1993. 100-kWe Lunar/Mars surface power utilizing the SP-100 reactor with dynamic conversion. In: Proceedings of the Symposium on Space Nucl. Power and Propulsion, vol. 2, American Institute of Physics, New York, NY, pp. 1065–1071 (AIP CP-271). King, J.C., El-Genk, M.S., 2006. Submersion criticality safety of fast spectrum space reactors: potential spectral shift absorbers. J. Nucl. Eng. Des. 236 (3), 238–254. King, J.C., El-Genk, M.S., 2007. Thermal-hydraulic analyses of the submersionsubcritical safe space (S4) reactor. In: Proceedings of the Space Technology and Applications International Forum (STAIF-2007), AIP Conference Proceedings No. 880, American Institute of Physics, Melville, NY, pp. 261–270. Lipinski, R., et al., 2002. Small fission power for NEP. In: Proceedings of the Space Technology and Applications International Forum (STAIF-2002), American Institute of Physics Proceedings No. 608, American Institute of Physics, Melville, New York, pp. 1054–1062. Marriot, A., Fujita, T., 1994. Evolution of SP-100 system designs. In: Proceedings of the Symposium on Space Nuclear power and Propulsion, American Institute of Physics, New York, NY, pp. 157–169 (AIP CP-301). Mason, L., et al., 2002. Status of Brayton Cycle Power Conversion development at NASA GRC. In: Proceedings of the Space Technology and Applications International Forum (STAIF-2002), American Institute of Physics Proceeding No. 608, American Institute of Physics, Melville, New York, p. 865–871. Mondt, J., Johnson, K., Fleurial, J.-P., El-Genk, M.S., Frye, P., Determan, W., 2005. Segmented thermoelectric multicouple converter technology development. In: Proceedings of the Space Technology and Applications International Forum (STAIF-2005), AIP Conference Proceedings No. 746, American Institute of Physics, Melville, NY, pp. 495–502. Mills, J., Van Hagan, T., 1994. S-PRIME-SNP conceptual design summary. In: Proceedings of the Symposium on Space Nuclear power and Propulsion, American Institute of Physics, New York, NY, pp. 695– 700 (AIP CP-301). Paramonov, D.V., El-Genk, M.S., 1994. Development and comparison of a TOPAZ-II system model with experimental data. J. Nucl. Technol. 108 (2), 157–170. Paramonov, D.V., El-Genk, M.S., 1996. A review of cesium thermionic converters with developed emitter surfaces. J. Energy Conv. Manage., 38. Rayn, M., et al., 2000. Lifetime of AMTEC electrodes: molybdenum, rhodium–tungsten, and titanium nitride. In: Proceedings of the Space Technology and Applications International Forum (STAIF-2000), AIP Proceedings No. 504, American Institute of Physics, Melville, NY, pp. 1377–1382. Saber, H.H., El-Genk, M.S., 2002. Optimization of segmented thermoelectric for maximizing conversion efficiency and electric power density. In: Proceedings of the International Thermoelectrics Conference, Long Beach, CA. Schreiber, J., 2001. Power characteristics of a stirling radioisotope power system over the life of the mission. In: Proceedings of the Space Technology and Applications International Forum (STAIF-2001), American Institute of Physics Proceeding No. 552, American Institute of Physics, Melville, New York, pp. 1011–1016. Thieme, L., Schreiber, J., Mason, L., 2002. Stirling Technology Development at NASA GRC. In: Proceedings of the Space Technology and Applications International Forum (STAIF-2002), American Institute of Physics Proceedings No. 608, American Institute of Physics, Melville, New York, pp. 872–879. Tournier, J.-M., El-Genk, M.S., 2003. Design optimization of high-power, liquid anode AMTEC. In: Proceedings of the Space Technology and Applications International Forum (STAIF-2003), AIP Conference Proceedings No. 654, American Institute of Physics, Melville, NY, pp. 740–750. Tournier, J.-M., El-Genk, M.S., 2004. Reactor lithium heat pipes for HP-STMCs space reactor power system. In: Proceedings of the Space Technology and Applications International Forum (STAIF-2004), AIP Conference Proceedings No. 669, American Institute of Physics, Melville, NY, pp. 781– 792. Tournier, J.-M., El-Genk, M.S., 2006. Bellows-type accumulator for liquid metal loops of space reactor power systems. In: Proceedings of the Space Technology and Applications International Forum (STAIF-2006), AIP Conference Proceedings No. 813, American Institute of Physics, Melville, NY, pp. 730–742. Wang, W., et al., 1992. NASA Lewis Stirling SPRE Testing and Analysis with Reduced Number of Cooler Tubes. Report No. NASA TM-105767. Williams, R.M., et al., 1999. The thermal Stability of Sodium Beta”-Alumina Solid Electrolyte Ceramic in AMTEC Converters. In: Proceedings of the Space Technology and Applications International Forum (STAIF-1999), AIP conference Proceedings No. 458, American Institute of Physics, Melville, NY, pp. 1306–1311. Virkar, A., Jue, J.-F., Fung, K.-Z., 2000. Alkali-Metal-␤- and ␤ -Alumina and Gallate Polycrystalline Ceramics and Fabrication by a Vapor Phase Method, US Patent No. 6,117,807.