Advanced power core system for the ARIES-AT power plant

Fusion Engineering and Design 80 (2006) 79–98

Advanced power core system for theARIES-AT power plant

A.R. Raffraya,∗, L. El-Guebalyb, S. Malangc, I. Sviatoslavskyb,M.S. Tillacka, X. Wanga, and

The ARIES Teama University of California, San Diego, 460 EBU-II, La Jolla, CA 92093-0438, USA

b University of Wisconsin, Fusion Technology Institute, 1500 Engineering Drive, Madison, WI 53706-1687, USAc Forschungszentrum Karlsruhe, Postfach 3640, D-76021 Karlsruhe, Germany

Accepted 1 June 2005Available online 25 October 2005

Abstract

The ARIES-AT power core was evolved with the overall objective of achieving high performance while maintaining attractivesafety features, credible maintenance and fabrication processes, and reasonable design margins as a rough indication of reliability.The blanket and divertor designs are based on Pb-17Li as coolant and breeder, and low-activation SiCf/SiC as structural material.

imizeence of

ionalseconds. Thiscludingn

sd

Flowing Pb-17Li in series through the divertor and blanket is appealing since it simplifies the coolant routing and minthe number of cooling systems. However, Pb-17Li provides marginal heat transfer performance in particular in the presMHD effects and the divertor design had to be adapted to accommodate the peak design heat flux of 5 MW/m2. The blanketflow scheme enables operating Pb-17Li at a high outlet temperature (about 1100◦C) for high power cycle efficiency whilemaintaining SiCf/SiC at a substantially lower temperature consistent with allowable limits. Waste minimization and additcost savings are achieved by radially subdividing the blanket into two zones: a replaceable first zone and a life of plantzone. Maintenance methods have been investigated which allow for end-of-life replacement of individual componentpaper summarizes the results of the design study of the ARIES-AT power core focusing on the blanket and divertor and ina discussion of the key parameters influencing the design, such as the SiCf/SiC properties and the MHD effects, and a descriptioof the design configuration, analysis results and reference operating parameters.© 2005 Elsevier B.V. All rights reserved.

Keywords: Advanced tokamak; Power plant studies; Power core system; Blanket and divertor design; Power cycle

∗ Corresponding author. Tel.: +1 858 534 9720;fax: +1 858 822 2120.

E-mail address: [email protected] (A.R. Raffray).

1. Introduction

The ARIES-AT power plant was evolved to assesand highlight the benefit of advanced technologies an

0920-3796/$ – see front matter © 2005 Elsevier B.V. All rights reserved.doi:10.1016/j.fusengdes.2005.06.356

80 A.R. Raffray et al. / Fusion Engineering and Design 80 (2006) 79–98

of new physics understanding and modeling capabili-ties on the performance of advanced tokamak powerplants [1]. The design builds on over a decade ofexperience and effort by the ARIES team in advancedpower plant design (e.g.[2,3,4]) and reflects the overallbenefit from high-� operation, high temperature super-conducting magnet, high power cycle efficiency, andlower-cost advanced manufacturing techniques.Fig. 1shows the ARIES-AT power core andTable 1summa-rizes the typical geometry and power parameters of thedevice, emerging from the parametric system studies[1].

The ARIES-AT power core have been developedwith the overall objective of achieving high perfor-mance while maintaining attractive safety features,credible maintenance and fabrication processes, andreasonable design margins as an indication of reliabil-ity. The design utilizes Pb-17Li as breeder and coolantand low-activation SiCf/SiC composite as structuralmaterial. The Pb-17Li operating temperature is opti-mized to provide high power cycle efficiency whilemaintaining the SiCf/SiC temperature under reason-able limits.

Table 1Typical ARIES-AT machine and power parameters

Machine geometryMajor radius 5.2 mMinor radius 1.3 mOutboard FW location at midplane 6.5 mOutboard FW location at lower/upper end 4.9 mInboard FW location 3.9 mOn-axis magnetic field 5.9 TToroidal magnetic field at inboard FW 7.9 TToroidal magnetic field at outboard FW 4.7 TToroidal magnetic field at divertor (outer/inner) 7/7.9 T

Power parametersFusion powera 1719 MWNeutron power 1375 MWAlpha power 344 MWCurrent drive power 25 MWTransport power to the divertor 256 MWBlanket multiplication factor 1.1Maximum thermal power 1897 MWAverage neutron wall load 3.2 MW/m2

Outboard maximum wall load 4.8 MW/m2

Inboard maximum wall load 3.1 MW/m2

a In the course of design optimization, slightly different designparameters were used for various calculations, and in the case ofvery small differences, the calculations were not redone to save timeand effort. For example, the neutronics calculations were performedfor a fusion power of 1755 MW, whereas a fusion power of 1719 MWis assumed here. However, the neutron wall load values are consistentunder the assumptions of slightly different first wall surface areas.

2. Power cycle

The Brayton cycle offers the best near-term possi-bility of power conversion with high efficiency and ischosen to maximize the potential gain from high tem-perature operation of the Pb-17Li which after exitingthe blanket is routed through a heat exchanger withthe cycle He as secondary fluid[5]. The Brayton cycleconsidered is shown inFig. 2. It includes three-stagecompression with two intercoolers and a high effi-ciency recuperator. Its main parameters are set underthe assumption of state of the art components and/orwith modest and reasonable extrapolation and are asfollows:

• lowest He temperature in the cycle (heat sink)= 35◦C;

• turbine efficiency = 93%;• compressor efficiency = 90%;• recuperator effectiveness = 96%;
Fig. 1. ARIES-AT power core (radial dimension in m).

A.R. Raffray et al. / Fusion Engineering and Design 80 (2006) 79–98 81

Fig. 2. Brayton cycle assumed for power conversion.

• He fractional pressure drop in out-of-vesselcycle = 0.025 (this would probably require a He pres-sure of about 15–20 MPa; the benefit of a low frac-tional pressure drop would then have to be assessedagainst the structural design requirements and pos-sible safety impact of a higher pressure system).

The compression ratio is set to optimize the cycleefficiency while maintaining a reasonable He tempera-ture level at the heat exchanger inlet (which in turn setsthe power core Pb-17Li inlet temperature). As illus-trated inFig. 3, for a maximum cycle He temperatureof 1050◦C, a cycle efficiency of 58.5% can be achievedfor a compression ratio of 3[5]. The corresponding Hetemperature at the heat exchanger inlet is 604◦C result-ing in a power core Pb-17Li inlet temperature of 654◦Cunder the assumption of a heat exchanger effectivenessof 0.9 and a ratio of Pb-17Li to He thermal capacitiesof 1.

3. Material considerations

3.1. SiCf/SiC

Use of SiCf/SiC as a structural material is attractivesince it enables operation at high temperature and its

low decay heat facilitates the accommodation of loss-of-coolant (LOCA) and loss-of-flow (LOFA) events.However, there are some key issues influencing itsattractiveness, including: thermal conductivity; param-eters limiting the temperature of operation, such asswelling under fusion neutron irradiation and compati-bility with the liquid metal; maximum allowable stresslimits; lifetime parameters; fabrication and joining

Fig. 3. Cycle efficiency and He temperature at inlet of heat exchangeras a function of compression ratio for Brayton cycle.


procedures. These issues were addressed in detail inpresentations and discussions at the January 2000 Inter-national Town Meeting on SiCf/SiC Design and Mate-rial Issues for Fusion Systems[6] and in a relatedpublication[7]. The SiCf/SiC parameters and proper-ties used in the ARIES-AT analysis are consistent withthe suggestion from this meeting. They are summarizedin Table 2and are briefly commented on below.

3.1.1. Thermal conductivityMeasurement of transverse thermal conductivity of

unirradiated SiCf/SiC composite tends to fall in the10–20 W/(m K) range for the earlier low-quality SiCfibers. It is expected that the thermal conductivitywill be improved as better stoichiometry is achievedin SiCf/SiC. Measurement of the transverse thermalconductivity of MER CVR SiC/SiC samples yielded75 W/(m K) at room temperature and≈35 W/(m K) at1000◦C. Although there are no confirmed irradiationdata on SiCf/SiC conductivity, it is assumed that thedegradation would be comparable to or less than (ona percentage basis) that observed in pure SiC. Startingwith an unirradiated thermal conductivity close to theMER value, it seems reasonable to assume a transverseconductivity of about 15 and possibly up to 20 W/(m K)for irradiated material at 1000◦C in future generationsof SiCf/SiC.

3.1.2. Stress limitsan-

n Then -l sms

TS

DDY aPTTT )MMM

ME

must be taken into account by sophisticated analyses. Inthe absence of tools enabling such sophisticated anal-yses, it is suggested that designers using conventionalFEM analysis limit the total combined stress in plane to≈190 MPa as a conservative first order approximation[7].

3.1.3. Temperature limitsThe maximum SiCf/SiC temperature limit is set to

avoid the void swelling regime. It is not clear at exactlywhat temperature the transition from point defectswelling to void swelling occurs but from the few exist-ing results, this transition temperature is≈1000◦C.

The maximum interface temperature limit forSiCf/SiC in contact with Pb-17Li is also not known.Only one data point exists in the open literature on thechemical compatibility of SiC with Pb-17Li at hightemperature. From an experiment performed at ISPRA,no compatibility problem was reported for SiC exposedover 1500 h to static LiPb at 800◦C. However, somequestions exist regarding the quality and impurity levelof the SiC used in these tests. This is certainly an areawhere R&D is needed. In anticipation of such data, itseems reasonable not to prejudge the blanket limita-tions and to assume a maximum temperature limit inthe blanket zone set by irradiation-induced dimensionalchanges (≈1000◦C). In the future, as new compatibil-ity data become available, this limit can be revised.

3.1.4. Lifetime limiting factortic-

u dataa stionei ce-m o bec sis-tl e of1 fH rtieso

3

ereu hesep

It seems clear that conventional stress limits cot be directly applied to ceramic composites.on-isotropic behavior of SiCf/SiC as well as the non

inearity arising from the unique damage mechani

able 2iCf /SiC properties and parameters assumed in this study[6]

ensity ‰3200 kg/m3

ensity factor 0.95oung’s modulus ‰200–300 GPoisson’s ratio 0.16–0.18hermal expansion coefficient 4× 10−6 K−1

hermal conductivity ‰20 W/(m K)hermal conductivity through thickness ‰20 W/(m Kaximum allowable combined stress ‰190 MPaaximum allowable operating temperature ‰1000◦Caximum allowable SiC/LiPb interfacetemperature

‰1000◦C

aximum allowable SiC burnup ‰3%lectrical resistivity 2× 10−3 � m

SiCf/SiC lifetime is one parameter that is parlarly hard to characterize based on the limitednd understanding and for which no clear suggemerged from the town meeting[6]. However, it is an

mportant parameter affecting the lifetime and replaent cost of in-vessel component and needs t

haracterized for power plant studies. Here, conent with previous ARIES studies[2], a 3% burnupimit is assumed (corresponding to a neutron fluenc8 MW-a/m2), subject to R&D verification of impact oe and H production on thermomechanical propef SiCf/SiC.

.2. Pb-17Li and W properties

Temperature-dependent properties of Pb-17Li wsed in the analysis. However, the expressions for troperties, as shown for example in Ref.[8], tend to be


Table 3Typical Pb-17Li properties (at 850◦C) and W properties used in thisstudy

Pb-17Li at 850◦CDensity 8560 kg/m3

Heat capacity 185 J/(kg K)Thermal conductivity 24.1 W/(m K)Dynamic viscosity 6.51× 10−4 kg/(m s)Prandtl number 5× 10−3

Electrical resistivity 1.5× 10−6 � m

TungstenDensity 18,950 kg/m3

Thermal conductivity 113 W/(m K)Poisson’s ratio 0.2Young’s modulus 370 GPaThermal expansion coefficient 4.5× 10−6 K−1

limited to rather low ranges of temperature, at mostabout 700◦C. In the absence of R&D data, it wasdecided to extrapolate from this property data base andto strongly recommend as future R&D an effort to mea-sure Pb-17Li properties at high temperature. The finaldesign would have to be re-optimized if Pb-17Li prop-erties (in particular the thermal conductivity) at hightemperatures were found to be substantially differentfrom the extrapolated values.

Table 3shows the typical Pb-17Li properties esti-mated for an average temperature of 850◦C. The Wproperties used in the divertor analysis are also shownin the table.

4. Blanket and coolant configuration

For waste minimization and cost saving reasons, theblanket is subdivided radially into two zones, as shownin Fig. 1: a replaceable first zone in the inboard andoutboard, and a life of plant second zone in the out-board. To simplify the cooling system and minimizethe number of coolants, the Pb-17Li is used to cool theblanket as well as the divertor and hot shield regions.The coolant is fed through an annular ring header sur-rounding the power core from which it is routed toeach of 16 reactor sectors through the following fivesub-circuits:

(1) series flow through the lower divertor and inboardblanket region (seeFig. 4; total thermal power and

( seg-

Fig. 4. Circuit 1—series flow through the lower divertor and inboardblanket region.

total thermal power and mass flow rate = 598 MWand 7270 kg/s);

(3) flow through the second segment of the first outerblanket region (seeFig. 6; total thermal power andmass flow rate = 450 MW and 5470 kg/s);

(4) series flow through the inboard hot shield regionand first segment of the second outer blanket region(seeFig. 7; total thermal power and mass flowrate = 182 MW and 4270 kg/s);

(5) series flow through the outboard hot shield regionand second segment of the second outer blanketregion (seeFig. 8; total thermal power and massflow rate = 140 MW and 1700 kg/s).

As illustrated inFigs. 9 and 10, the blanket design ismodular and consists of a simple annular box throughwhich the Pb-17Li flows in two poloidal passes. Posi-tioning ribs are attached to the inner annular wall

mass flow rate = 501 MW and 6100 kg/s);2) series flow through the upper divertor and one

ment of the first outer blanket region (seeFig. 5;


Fig. 5. Circuit 2—series flow through the upper divertor and one segment of the first outer blanket region.

forming a free floating assembly inside the outer wall.These ribs divide the annular region into a number ofchannels through which the coolant first flows at highvelocity to keep the outer walls cooled. The coolantthen makes a U-turn and flows very slowly as a secondpass through the large inner channel from which the

Pb-17Li exits at high temperature. This flow schemeenables operating Pb-17Li at a high outlet temperature(1100◦C) while maintaining the blanket SiCf/SiC com-posite and SiC/PbLi interface at a lower temperature(‰1000◦C). The first wall consists of a 4 mm SiCf/SiCstructural wall on which a 1 mm CVD SiC armor layer


Fig. 6. Circuit 3—flow through the second segment of the first outer blanket region.

is deposited. This is based on the assumption of nosignificant carbon erosion and resulting tritium co-deposition; if this were found to be problematic underAT operations, the possibility of coating the first wallwith tungsten (as in the divertor case) would have to beconsidered.

5. Blanket analysis

Detailed 3-D neutronics analyses of the power corewere performed yielding a tritium breeding ratio of1.1 and the energy multiplication and wall loading val-ues shown inTable 1 [9]. When incorporating the final


Fig. 7. Circuit 4—series flow through the inboard hot shield region and first segment of the second outer blanket region.

requirements from the stabilizing shell thickness, theTBR was reduced to <1.1. If required, a compensat-ing measure would be to increase the blanket regionthickness. Ref.[9] shows that an overall TBR of 1.1can be achieved in that case with a 35 cm inboardblanket region thickness, 30 cm first outboard blan-ket region thickness and 45 cm second outboard blan-ket region thickness. For consistency the neutronics-

related analyses are based on that radial build. How-ever, the power core design and analyses were alreadydone based on the original radial build (30 cm inboardblanket region thickness, 30 cm first outboard blanketregion thickness and 35 cm second outboard blanketregion thickness) and are believed to appropriatelyreflect the thermal-hydraulic and thermo-mechanicoperating parameters of the concept. To avoid unnec-


Fig. 8. Circuit 5—series flow through the outboard hot shield region and second segment of the second outer blanket region.

essary duplication of effort, they are described inthis paper with the understanding that the parame-ters and design details could be further optimized asrequired by TBR considerations. The volumetric heatgeneration profiles from the neutronics analyses wereused in the thermal investigations. Of the three blan-

ket regions, the first outboard region is subjected tothe highest heat loads. A typical module in an out-board segment cooled in series with the upper diver-tor was the focus of the thermal analyses which aredescribed below and whose results are summarized inTable 4.


Fig. 9. Cross-section of ARIES-AT outboard blanket segment (radial dimension in m).

Fig. 10. Cross-section of ARIES-AT outboard blanket module (alldimensions in cm).

For these analyses, the plasma heat flux profile wasestimated by considering bremstrahlung radiation, lineradiation and synchotron radiation with average andpeak values of 0.26 and 0.34 MW/m2, respectively[4].The analysis assumes no radiation to the first wall fromthe 256 MW transport power to the divertor. The designcan be re-optimized if such radiation becomes appre-ciable under future physics scenarios.

5.1. Thermal-hydraulic analysis

Even though the SiCf/SiC provides insulated wallsthereby minimizing MHD effects, the analysis conser-vatively assumes MHD-laminarized flow of the Pb-17Li in the blanket and heat transfer by conductiononly. This is also justified by the observation that theMHD Re for the transition to turbulent flow estimatedfrom the given Hartmann number,H (‰500H for atransverse magnetic field[10]) is higher than the actualflow Re, as shown inTable 3. The temperature profilethrough the blanket was estimated by a moving coordi-nate analysis which follows the Pb-17Li flow throughthe first-pass annular wall channel and then throughthe second-pass large inner channel. In general, MHDeffects would tend to flatten the velocity profile and,for simplicity, the analysis assumed slug flow in both


Table 4Summary of typical ARIES-AT blanket parameters (as representedby first outboard region)

Overall geometryInboard blanket first wall area 111 m2

Outboard blanket first wall area 244 m2

Total blanket first wall area 355 m2

Pb-17Li coolantPower core inlet temperature 654◦CInlet temperature to outboard blanket region I 764◦CBlanket outlet temperature 1100◦CBlanket inlet pressure 1 MPaBlanket pressure drop 0.25 MPaTotal mass flow rate 22,700 kg/sMass flow rate per module in O/B blanket region I 76 kg/s

Outboard blanket region INumber of sectors/segments 16/32Number of modules per outboard segment 6Module poloidal dimension 6.8 mAverage module toroidal dimension 0.19 mFirst wall SiCf /SiC thickness 4 mmFirst wall CVD SiC thickness 1 mmFirst wall annular channel thickness 4 mmAverage Pb-17Li velocity in first wall 4.2 m/sFirst wall channel Re 3.9× 105

First wall channel transverse Hartmann number 4340MHD turbulent transition Re 2.2× 106

First wall MHD pressure drop 0.19 MPaMaximum SiCf /SiC temperature 996◦CMaximum CVD SiC temperature 1009◦CMaximum Pb-17Li/SiC interface temperature 994◦CAverage Pb-17Li velocity in inner channel 0.11 m/s

channels. The annular wall rib spacing is used as MHDflow control to achieve a higher flow rate through thefirst wall (with larger toroidal spacing) than through theside and back walls. For example, having three chan-nels in the module first wall and 13 in the back wallallows for a high velocity of 4.2 m/s in the first wallchannels and a lower velocity of 0.66 m/s in the backwall channel for the same MHD pressure drop. Thesecond poloidal pass of the Pb-17Li through the largeinner channel is much slower with an average velocityof 0.11 m/s.

Fig. 11illustrates the results for a typical outboardmodule. Even though the average outlet Pb-17Li tem-perature is 1100◦C, this design results in a maximumSiC temperature at the first wall (radial distance = 0) of1009◦C, a maximum SiCf/SiC temperature of 996◦C.and a maximum blanket SiC/Pb-17Li interface tem-perature at the inner channel wall of 994◦C, which

Fig. 11. Temperature distribution through outboard blanket module.

satisfy the maximum temperature limits shown inTable 2.

The first wall and blanket pressure drops were esti-mated as a function of the Hartmann number,H, andwall to fluid conductance ratio,C, based on the follow-ing expression for the pressure gradient under MHDeffects,dP

dx[10]:

dP

dx= µv

a2

(H2 tanhH

H − tanhH+ H2C

1 + C− 3 + dP ′

dx′

)

(1)

C = σwtw

σLa(2)

H = aB

√σL

µ(3)

whereµ is the liquid viscosity;ν the liquid averagevelocity; a the channel half-width in the direction par-allel to the magnetic field,B; dP ′

dx′ the ordinary laminar-flow pressure drop (=7.11 for square channels);σwthe wall electrical conductivity;σL the liquid electricalconductivity;tw the wall thickness.

The resulting overall blanket pressure drop is about0.25 MPa.


5.2. Stress analysis

Stress analyses were performed both on the mod-ule outer and inner shells. A 1 MPa inlet pres-sure is assumed for the coolant which adequatelyaccounts for both the pressure drop through the blanket(‰0.25 MPa) and the hydrostatic pressure due to the‰6 m Pb-17Li (‰0.5 MPa) column. The outer wallis designed to withstand this pressure while the innerwall is designed to withstand the difference betweenblanket inlet and outlet pressures (0.25 MPa). There aresix modules per outboard segment as shown inFig. 9.These modules are brazed to one another and the sidewalls of all the inner modules are pressure balanced.However, the side walls of the outer modules must bereinforced to accommodate the 1 MPa coolant pressure.For example,Fig. 12shows that the maximum side wallpressure stress is 85 MPa for a 2 cm thick side wall.The side wall can be tapered radially and poloidallyby tailoring the thickness to maintain a uniform stress.This would reduce the SiC volume fraction and ben-efit tritium breeding. In addition, the thermal stress atthis location is small and the sum of the pressure andthermal stresses is well within the 190 MPa limit. Thismargin can be considered as a rough indication of reli-

ability and provides some flexibility if the final blanketdesign optimization shows that further reductions ofthe SiC volume fraction are needed for better tritiumbreeding.

From Fig. 12, the pressure stress at the first wallis quite low, ‰60 MPa. The corresponding thermalstress is also very reasonable and the design providessubstantial margin in this regard being able to accom-modate more than twice the reference heat flux values.For example,Fig. 13 shows the 3-D thermal stressdistribution in the module for a case assuming a sim-ple linear heat flux poloidal distribution (with averageand maximum values of 0.6 and 0.7 MW/m2, respec-tively) and an effective heat transfer coefficient of15,000 W/(m2 K) in the Pb-17Li channel. The maxi-mum thermal stress in this case is 114 MPa resulting ina modest combined stress of 174 MPa still well withinthe 190 MPa limit.

Fig. 14shows the pressure stress in an inner mod-ule wall of 8 mm side thickness. The thickness of theupper and lower wall is 5 mm. The maximum stressis 116 MPa again well within the maximum allowablestress. In addition, the maximum pressure differentialof ‰0.25 MPa occurs at the lower poloidal location.The inner wall thickness could be tapered down to

stributio
Fig. 12. Pressure stress di n in blanket module outer wall.


Fig. 13. Thermal stress distribution in toroidal half of blanket mod-ule.

‰5 mm at the upper poloidal location if needed to min-imize the SiC volume fraction.

The design has been optimized to the given dimen-sions to provide reasonable margins on stress limits.As mentioned above, it also provides the possibility ofreducing the SiC volume fraction for improved tritiumbreeding by tapering the side wall thickness for bothinner and outer shells. If further improvement in the tri-tium breeding is required, the blanket thickness can alsobe adjusted. For example, the second outboard blanketregion could be increased by ‰10 cm. However, theside wall thicknesses of the outer module in each seg-ment and of the inner shell in each module would haveto be adjusted accordingly to meet the stress limits.

5.3. Safety analysis

The activation, decay heat, and waste disposal anal-yses performed in support of the ARIES-AT designare described in Refs.[11,12]. The decay heat resultswere used to perform 2-D safety analyses of thepower core which showed that the low decay heatof SiC enables accommodation of any loss-of-coolant(LOCA) or loss-of-flow (LOFA) scenarios without

Fig. 14. Pressure stress distributio
n in blanket module inner wall.


serious consequences to the blanket structure[12,13].Interestingly, because of the low decay heat of SiCas compared to Pb-17Li, the analyses showed that theresulting maximum power core temperature was about50◦C higher in the case of a LOFA than in the case ofa LOCA, but still well within the limits.

6. Divertor configuration

Pb-17Li has a relatively low thermal conductivityand tends to offer limited heat removal performance inparticular in the presence of a magnetic field. In orderto accommodate MHD effects, the proposed design:

• minimizes the interaction parameter (<1) which rep-resents the ratio of MHD to inertial forces;

• directs the flow in the high heat flux region parallelto the toroidal magnetic field;

• minimizes the Pb-17Li flow path and residence timein the high heat flux region.

The design is illustrated inFig. 15, which shows across-section of a divertor plate. It consists of a numberof 2 cm× 2.5 cm SiCf/SiC poloidal channels. The frontSiCf/SiC wall is very thin (0.5 mm) in order to main-tain the maximum temperature and combined stresslimits to ≤1000◦C and≤190 MPa, respectively. Asplasma facing component (PFC), a 3.5 mm plasma fac-ing layer of sputter-resistant W alloy (with 0.2 wt%TiC) is bonded to the thin SiCf/SiC to provide addi-t 17Lip

this design, as in previous ARIES studies, the philos-ophy is to operate with a∼10% margin within theplasma� limit to avoid disruptions, which is also con-sistent with the understanding that a utility companyis not likely to buy or operate a power plant with fre-quent disruptions. However, the design should be ableto handle a few disruptions without loss of investmentand the≈1 mm sacrificial W armor is provided for thispurpose.

In each channel, a T-shaped flow separator isinserted. The Pb-17Li flows poloidally through one halfof the channel which acts as an inlet header. The flowis then forced to the PFC region through small holes atone side of the channel. The flow through these smallholes is inertial with an interaction parameter <1. ThePb-17Li then flows toroidally to cool the high heat fluxregion through a very short flow path (2 cm). It is thenrouted back to the other side of the poloidal channelserving as outlet header. The Pb-17Li velocity throughthe toroidal PFC channel can be adjusted by chang-ing the dimension of this channel or by increasing thenumber of toroidal passes through a plate. The refer-ence design uses a 2 mm thick channel and 2-pass flowresulting in toroidal channel velocities and residencetimes of about 0.4 m/s and 0.05 s, respectively

7. Divertor analysis

/m2

a

diverto

ional structure to accommodate the 1.8 MPa Pb-ressure and to provide sacrificial armor (≈1 mm). In

Fig. 15. Cross-section of

The average heat flux on the divertor is 1.75 MWnd, in analogy with the analysis shown in Ref.[3], a

r plate (all dimensions in cm).


corresponding peak heat flux of 5 MW/m2 is assumedfor the analysis. The heat transfer performance isslightly more challenging for the lower divertor thanfor the upper divertor since the former is part of a Pb-17Li circuit with a somewhat lower mass flow rate (seeSection4) than the latter. Consequently, the divertoranalyses described below are focused mostly on thelower divertor. However, some key parameters for theupper divertor are also included in the overall sum-mary of divertor design parameters shown inTable 5.The parameters for the inner and outer divertor platesare also shown separately when there are significantdifferences. This was found not to be the case for themaximum material temperature calculations, for exam-

Table 5Summary of typical ARIES-AT divertor parameters

Divertor geometryDivertor plate poloidal dimension

(outer/inner)1.5/1.0 m

Divertor channel toroidal pitch 2.1 cmDivertor channel radial dimension 3.2 cmNumber of divertor channels (outer/inner) 1316/1167Total number of divertor channels

(upper + lower)4967

SiCf /SiC plasma-side thickness 0.5 mmW thickness 3.5 mmPlasma facing channel thickness 2 mmNumber of toroidal passes 2Toroidal dimension of inlet and outlet

feeder slots1 mm

Inboard divertor plate area 40.9 m2

P

ag/s

ple, as the effect of the higher Pb-17Li temperature inthe inner divertor (the Pb-17Li flows in series throughthe outer and then inner divertor plates) is compensatedby the effect of higher Pb-17Li velocity resulting fromthe geometry and lower number of tubes in the innerdivertor region.

7.1. Thermal-hydraulic analysis

The 2-D moving coordinate method was used forthe flow analysis under the assumption of MHD-laminarized slug flow in the toroidal channel.Fig. 16shows the typical results for an upper divertor case withan inlet Pb-17Li temperature of 653◦C, a W thicknessof 3 mm, a SiCf/SiC first wall thickness of 0.5 mm, aPb-17Li toroidal channel thickness of 2 mm, a SiCf/SiCinner wall thickness of 0.5 mm, a Pb-17Li velocity of0.35 m/s and a surface heat flux of 5 MW/m2. The max-imum W temperature is≈1150◦C, and the maximumSiCf/SiC temperature is≈950◦C. The lowest W tem-perature is≈800◦C for this case and will be about700–800◦C for lower local heat fluxes. This is some-what lower than the expected value to avoid severe irra-diation embrittlement based on the scarce data available[14]. However, a margin of about 50◦C exists for theSiCf/SiC maximum temperature and, as new R&D dataon W mechanical properties under irradiation becomeavailable, the design and coolant conditions can be

Outboard divertor plate area 54.9 m2

Divertor dome area 50.8 m2

Total divertor area 146.6 m2

b-17Li coolant and thermal-hydraulicsDivertor inlet temperature 654◦CDivertor outlet temperature (upper/lower) 785/764◦CDivertor inlet pressure 1.8 MPaDivertor pressure drop (lower/upper) 0.55/0.7 MPCircuit mass flow rate (lower/upper) 6100/7270 kVelocity in lower divertor PFC channel

(outer/inner)0.35/0.6 m/s

Velocity in upper divertor PFC channel(outer/inner)

0.42/0.71 m/s

Velocity in inlet and outlet slot to PFCchannel

0.9–1.8 m/s

Interaction parameter in inlet/outlet slot 0.46–0.73Maximum SiCf /SiC temperature 970/950◦C

(lower/upper)Maximum W temperature (lower/upper) 1145/1125◦CW pressure + thermal stress ≈25 + 45 MPaSiCf /SiC pressure + thermal stress ≈25 + 165 MPa

F n ofo

ig. 16. Temperature distribution through plasma-facing regiouter divertor plate.


reoptimized to increase the minimum W operating tem-perature.

The resulting Pb-17Li MHD pressure drop throughthe divertor for the proposed flow configuration is esti-mated from Eqs.(1)–(3) to be about 0.7 MPa. This issignificantly larger than the blanket pressure drop. Tominimize pressure stresses in the piping and blanketsystem, the Pb-17Li in the inlet manifold is kept at1.1 MPa. The Pb-17Li is then pressurized to 1.8 MPajust before flowing though the divertor by means of anE-M pump making use of the existing toroidal magneticfield. The outlet flow from the divertor then rejoins theblanket inlet manifold at about 1.1 MPa.

7.2. Stress analysis

Stress analyses were performed to size the diver-tor. For example, the thermal stress distribution isshown inFig. 17for a surface heat flux of 5 MW/m2.From the figure, the maximum thermal stress atthe plasma-facing wall is low in the W (≈45 MPa)but is of more concern in the SiCf/SiC (165 MPa).However, the corresponding pressure stress in theSiCf/SiC also calculated from ANSYS is modest

(25 MPa for an inner Pb-17Li pressure of 1.8 MPa).The resulting combined stress is 190 MPa which sat-isfies the maximum allowable stress criterion forSiCf/SiC.

8. Fabrication and maintenance

As a reliability measure, minimization of the num-ber and length of brazes was a major factor inevolving the fabrication procedures for the blanketand divertor. For the blanket, the proposed fabri-cation scheme requires three radial/toroidal coolant-containment brazes per module, as illustrated by thefollowing fabrication steps for an outboard segmentconsisting of six modules (seeFig. 18(a–d)):

1. manufacturing separate halves of the SiCf/SiCpoloidal module by SiCf weaving and SiC chem-ical vapor infiltration (CVI) or polymer process;

2. inserting the free-floating inner separation wall ineach half module;

3. brazing the two half modules together at the mid-plane;

4. brazing the module end cap;

ss dist
Fig. 17. Thermal stre ribution in divertor plate.


Fig. 18. (a–d) Step-wise blanket fabrication procedure.


5. forming a segment by brazing six modules together(this is a joint which is not in contact with thecoolant);

6. brazing the annular manifold connections to one endof the segment.

The divertor fabrication scheme also aims at mini-mizing brazing and can be summarized as follows (seeFig. 19):

1. manufacture separate SiCf/SiC toroidal halves ofthe divertor plate by SiCf weaving and SiC CVIor polymer process; maintain constant channeltoroidal dimensions but tapered side wall thick-nesses to account for torus geometry;

2. insert the inner SiCf/SiC separation wall in eachdivertor channel;

3. braze the two toroidal halves of the divertor platetogether;

4. braze the end cap and manifold on each end;5. bond the W layer to the SiCf/SiC front wall by

plasma spray (the W does provide some structuralfunction and the bonding to the SiCf/SiC must pro-vide the required material strength; this needs to beaddressed through R&D).

Maintenance methods have been investigated whichallow for end-of-life replacement of individual compo-nents. These are discussed in detail in Ref.[15].

9. Out-of-reactor cooling system and tritiumissues

An annular Pb-17Li coolant manifold is used to feedthe divertor and blanket, with the lower temperatureinlet flow in the outer channel and the higher temper-ature outlet flow in the inner channel. In this way, anyeffect of the high SiC/Pb-17Li interface temperatureon the manifold inner wall would only result in a leakto the manifold outer channel, which would not be ofmajor consequence. However, the structural integrityof the configuration would be ensured by the low tem-perature outer channel.

One issue relates to the difficulty of maintainingthe maximum SiC/Pb-17Li interface temperature at theoutlet side of the manifold well below the Pb-17Li out-let temperaure (1100◦C). This can only be done bymaintaining a heat flux between the outlet and inlet flowto provide a temperature gradient at the outlet wall suchthat the wall interface temperature would be lower thanthe outlet Pb-17Li temperature. This could be achievedto some extent by adjusting the annular manifold geom-etry (affecting the velocity and heat transfer coefficientin the inner and outer channels) and inner wall thermalconductivity.Fig. 20illustrates this effect for a mani-fold servicing the outboard blanket segment from thering header. For example, the maximum Pb-17Li/SiCinterface temperature can be reduced to about 1000◦Cfor manifold inner and outer radii of 0.05 and 0.12 mand an inner wall thermal conductivity of 10 W/(m K).H out-l lkt han1

or int f thec tureo ccepta them turea min-i hilemA /SiCc uldb rre-s .06
Fig. 19. Assembled divertor plate segment.
owever, the corresponding heat flux from the hotet Pb-17Li to the cold inlet Pb-17Li results in a buemperature increase of the inlet flow of more t00◦C which negates the initial effect.

All manifold flows are in regions with very low or nadiation and with the higher temperature outlet flowhe inner channel such that the structural integrity oonfiguration depends mostly on the low temperauter channel wall. Thus, it seems reasonable to amaximum Pb-17Li/SiC interface temperature inanifold comparable to the outlet Pb-17Li temperand to set the manifold annular dimensions so as to

mize the heat flux between the hot and cold legs waintaining a reasonable pressure drop (seeFig. 21).s data become available on unirradiated Pb-17Liompatibility at high temperature, this approach coe revised. For the above case, this would copond to manifold inner and outer radii of ‰0


Fig. 20. Maximum Pb-17Li/SiC interface temperature in manifoldand corresponding Pb-17Li temperature as a function of annular man-ifold inner radius for one of 16 manifolds feeding an outboard blanketsegment from the ring header (Pb-17Li mass flow rate = 45.4 kg/s;Pb-17Li inlet/outlet temperatures = 653/1100◦C; pipe length = 8 m;outer radius,ro, = 0.12 m; SiCf /SiC outer and inner wall thick-ness = 5 mm). Cases with assumed SiCf /SiC inner wall conductivitiesof 1 and 10 W/(m K) are shown.

and 0.12 m and an inner wall thermal conductivity of1 W/(m K).

Tritium has a high vapor pressure over Pb-17Li andoperation of SiCf/SiC at high temperature raises thequestion of whether tritium permeation and inventorywould be an issue. For the annular manifold, tritiumpermeation is less of a concern since some level of per-meation through the high temperature SiCf/SiC innerchannel wall would be allowable as long as the cor-responding tritium inventory is acceptably low andpermeation through the colder outer channel wall isacceptably low. Of more concern would be the poten-tial permeation in the heat exchanger between the Pb-17Li primary side and the He secondary side. SiCf/SiCwould be a logical candidate as heat exchanger mate-rial although no detailed design of the heat exchangerhas been performed and no final decision on the heatexchanger material made. Ref.[16] includes a reviewof tritium solubility and diffusivity in SiC, highlightingthe experimental difficulties encountered in attemptingto absorb hydrogen isotopes into the bulk of the�-SiCeven at temperatures as high as 1000◦C. This suggeststhat tritium retention in the SiCf/SiC walls of the power

Fig. 21. Total manifold pressure drop as a function of annular man-ifold inner radius for fixed outer radius, length and flow rate (seeFig. 20; entry/exit loss coefficient = 2).

core channels should be small. It also suggests that tri-tium permeation in the heat exchanger should not bea major issue in particular if a coating of fully dense�-SiC (such as by CVD) is present, as is proposed forthe current design.

Preliminary considerations of possible tritiumrecovery processes indicate that for the high tritiumpartial pressure (‰20 Pa) and low tritium concentra-tion (inventory) associated with the low solubility oftritium in Pb-17Li and its low permeation through SiC,a liquid-gas contactor system would be possible. Forexample, a fraction of the Pb-17Li (‰10%) flow wouldbe diverted to the tritium recovery process where tri-tium would be removed by a stream of He gas bubblingthrough the Pb-17Li. The He purge could then be routedto a molecular sieve bed for recovery of tritium andregeneration as pure hydrogen isotopes. Although sucha system appears feasible in principle, detailed studiesare required to confirm its applicability and to thor-oughly address the design issues.

10. Conclusions

The ARIES-AT power core utilizes high tempera-ture Pb-17Li as breeder and coolant and low-activationSiCf/SiC composite as structural material. High power


cycle efficiency (‰58.5%) is achieved while thein-reactor material limits are accommodated by thedesign. Both the blanket and divertor designs are basedon a simple geometry. Credible fabrication procedureshave been evolved which minimize the coolant contain-ing joints and enhances reliability. The design processstrives as much as to maintain comfortable stress limitmargins as an additional reliability measure.

Key issues requiring R&D attention are mostlylinked with the SiCf/SiC material. They include devel-opment of low-cost high-quality material and joiningmethods and characterization of key SiCf/SiC prop-erties and parameters at high temperature and underirradiation, in particular thermal conductivity, temper-ature limits (based on strength degradation, compati-bility with Pb-17Li, and He swelling), and lifetime. Inaddition, Pb-17Li properties at high temperature needto be measured.

Acknowledgement

This work was supported by U.S. Department ofEnergy under Contract DE-FC03-95ER54299.

References

[1] F. Najmabadi and The ARIES Team, The ARIES-AT advancedEng.

san,inal

RSDes.

[4] F. Najmabadi and the ARIES team, Spherical torus concept aspower plant – The ARIES-ST study, Fusion Eng. Des. 65 (2003)143–164.

[5] R. Schleicher, A.R. Raffray, C.P. Wong, An assessment of theBrayton Cycle for high performance power plant, Fusion Tech-nol. 39 (2) (March 2001) 823–827.

[6] International Town Meeting on SiC/SiC Design and MaterialIssues for Fusion Systems, Oak Ridge National Laboratory,January 18–19, 2000.http://aries.ucsd.edu/LIB/MEETINGS/0001-SiCSiC/.

[7] A.R. Raffray, R. Jones, G. Aiello, M. Billone, L. Giancarli, H.Golfier, et al., Design and material issues for high performanceSiCf /SiC-based fusion power cores, Fusion Eng. Des. 55 (1)(2001) 55–95.

[8] Property data base from ARIES web site,http://aries.ucsd.edu/LIB/PROPS/.

[9] L.A. El-Guebaly and the ARIES Team, Nuclear performanceassessment for ARIES-AT, Fusion Eng. Des. 80 (2006) 99–110.

[10] M.A. Hoffman, G.A. Carlson, Magnetic field effects in fusionreactor blankets, in: Proceedings of the 13th Intersociety EnergyConversion Energy Conference, SAE P-75, IEEE 78-CH1372,San Diego, August 20–25, 1978, pp. 1096–1108.

[11] D. Henderson, L. El-Guebaly, A. Abdou, P. Wilson, the ARIESTeam, Activation, decay heat, and waste disposal analyses forARIES-AT power plant, Fusion Technol. 39 (2) (March 2001)444–448.

[12] D.A. Petti, B.J. Merrill, R.L. Moore, G.R. Longhurst, L. El-Guebaly, E. Mogahed, et al., ARIES-AT safety design andanalysis, Fusion Eng. Des. 80 (2006) 111–137.

[13] E. Mogahed, L. El-Guebaly, A. Abdou, P. Wilson, D. Hender-son, the ARIES Team, Loss of coolant and loss of flow accidentanalyses for ARIES-AT power plant, Fusion Technol. 39 (2)(March 2001) 462–466.

[14] S.J. Zinkle, N.M. Ghoniem, Operating temperature windows1–52

[ andDes.,

[ M.-ST

tokamak, advanced technology Fusion power plant, FusionDes. 80 (2006) 3–23.

[2] F. Najmabadi, R.W. Conn, P.I.H. Cooke, S.P. Grotz, M.Z. HaE. Ibrahim, et al., The ARIES-I tokamak reactor study, FReport, UCLA-PPG-1323, UCLA, CA, 1991.

[3] F. Najmabadi, the ARIES Team, Overview of the ARIES-reversed-shear tokamak power plant study, Fusion Eng.38 (1997) 3–25.

for fusion reactor structural materials, Fusion Eng. Des. 5(2000) 55–71.

15] L. Waganer and the ARIES Team, ARIES-AT configurationmaintenance system definitions and analysis, Fusion Eng.2005, in this Proceedings.

16] M.S. Tillack, X.R. Wang, J. Pulsifer, S. Malang, D.K. Sze,Billone, et al., Fusion power core engineering for the ARIESpower plant, Fusion Eng. Des. 65 (2003) 215–261.

http://aries.ucsd.edu/lib/meetings/0001-sicsic/

http://aries.ucsd.edu/lib/props/

Advanced power core system for the ARIES-AT power plant

Documents

Transcript of Advanced power core system for the ARIES-AT power plant