BENEFITS OF RADIAL BUILD MINIMIZATION AND REQUIREMENTS IMPOSED ONARIES COMPACT STELLARATOR DESIGN

L. El-Guebaly1, R. Raffray2, S. Malang3, J.F. Lyon4, L.P. Ku5, and the ARIES Team

1University of Wisconsin, Fusion Technology Institute, Madison, WI, elguebaly@engr.wisc.edu2University of California-San Diego, 9500 Gilman Drive, La Jolla, CA, raffray@fusion.ucsd.edu

3Consultant, Fliederweg 3, D 76351 Linkenheim-Hochstetten, Germany, smalang@web.de4Oak Ridge National Laboratory, PO Box 2008, MS-6169, Oak Ridge, TN, lyonjf@ornl.gov

5Princeton Plasma Physics Laboratory, PO Box 451, Princeton, NJ, lpku@pppl.gov

It is widely recognized among stellarator researchersthat the minimum distance between the plasma boundaryand the middle of the coil (Dmin) is of great importance forstellarators as it impacts the machine parametersconsiderably. Techniques for minimizing the radial buildhave made impressive progress during the first year of theARIES-CS study. A novel approach has been developedfor ARIES-CS where the blanket at the critical areasurrounding Dmin has been replaced by a highly efficientWC-based shield. As a result, an appreciable 20-90 cmsavings in the radial build has been achieved, reducingthe major radius by more than 20%, which is significant.The economic benefit of this approach is yet to bedetermined and the added engineering problems andcomplexity will be addressed during the remaining periodof the study. This paper covers the details of the radialbuild optimization process that contributed to thecompactness of ARIES-CS. Compared with previousdesigns, the major radius of ARIES-CS has more thanhalved, dropping from 24 m to less than 10 m, making astep forward toward the feasibility of a compactstellarator power plant.

I. INTRODUCTION

After two decades of stellarator power plant studies,it was evident that a new design that reflects theadvancements in physics and improvements in technologywas needed. To realize this vision, the multi-institutionalARIES team has recently launched a study1 to provideperspective on the benefits of optimizing the physics andengineering characteristics of the so-called compactstellarator (CS) power plants. The primary goal of thestudy is to develop a more compact machine that retainsthe cost savings associated with the low recirculatingpower of stellarators, and benefits from the higher beta,smaller size, and higher power density, and hence lowercost of electricity, than was possible in earlier studies.Relative to tokamaks, stellarators promise disruption-free,steady-state operation with reduced recirculating power

due to the absence of current-drive requirements. Suchadvantages could be offset, however, by the morecomplex coil system, nonuniform blanket/shield/divertorconfiguration, and challenging maintenance scheme.

The ARIES team has explored the potential of twoconfigurations for the ARIES-CS power plant: two fieldperiods (FP) with 16 coils and three field periods with 18coils. Figures 1 and 2 illustrate both options showing thelast plasma closed flux surface and the individual coilsthat create the compact configuration. The limited spaceassigned to the internals (blanket, shield, and vacuumvessel) calls for a well optimized, highly compact radialbuild, particularly at Dmin as it controls the minimummajor radius and the maximum field at the coil. Duringthe first phase of the study, several blanket/shield systemshave been considered employing advanced ferritic steel(FS) structure (such as IEA MF82H) and SiC/SiCcomposites. The list of candidates includes one solidbreeder-based system (Li4SiO4/FS/Be/He) and four liquidbreeder-based systems (self-cooled Flibe/FS/Be andLiPb/SiC, and dual-cooled LiPb/FS/He and Li/FS/He).The newly proposed dual-cooled Flibe/FS/Be/He systemwill be examined later. As predicted, each concept offersadvantages and drawbacks and an integrated study withguidance from the economic analysis and maintenancescheme will later identify the preferred concept during thesecond phase of the study.

The nuclear assessment proceeded interactively withguidance from the thermo-mechanical analysis. Westarted the analysis by examining the breeding capacity ofthe candidate breeders, and then defined the blanketparameters (thickness, composition, and Li enrichment).Next, the shield was designed to protect the VV and bothcomponents were simultaneously sized to essentiallyprotect the superconducting magnets. Finally, wespecified the radial build and identified the key nuclearparameters: tritium breeding ratio (TBR), neutron energymultiplication (Mn), radiation damage to structuralcomponents and thei r service l i fe t imes .

Fig. 1. Two field period design option for ARIES-CS.

II. HISTORIC BACKGROUND

Although the stellarator concept has been around forsome time, very little in the way of conceptual designstudies has been performed compared to tokamaks, ofwhich many studies have taken place in the US andabroad. During the decade of the 1980s and continuing tothe present, six stellarator power plants have beendeveloped. These include UWTOR-M2, ASRA-6C3,SPPS4, and ARIES-CS in the US, HSR5 in Germany, andFFHR6 in Japan. The studies varied in scope and depthand encompassed a broad range of configuration options.The 1982 UWTOR-M design2 has 18 modular twistedcoils with only two different coil geometries arranged in atoroidal configuration. The blanket employs FS as themain structure and LiPb for tritium breeding. Initiated inthe mid-80s, the ASRA-6C study designed all 30 coilsand internal components (FW, LiPb/FS blanket, andshield) to have identical elliptical bore. Next came theStellarator Power Plant Study (SPPS) initiated in 1995 bythe ARIES team to address key issues for stellaratorsbased on the modular helias-like heliac approach. Thebaseline configuration has four field periods produced by32 modular coils of four distinct types. Vanadiumstructure and lithium breeder are the reference materialsfor SPPS. On the international level, a helias stellaratorreactor (HSR) study was initiated in Germany in the late1990s based on the Wendelstein 7-X experiment that iscurrently under construction in Greifswald, Germany. Themost recent HSR4/18 design5 has four field periods with40 coils and LiPb/FS blanket. Alternatively, the stellaratorconfiguration can be produced using continuous helicalcoils. An example of this approach is the Forced FreeHelical Reactor6 (FFHR) presently under study in Japan.Vanadium structure, Flibe breeder, and berylliummultiplier are the materials of choice for FFHR. Note thatall designs developed to date employed liquid breeders(Flibe, LiPb, or Li) for breeding and cooling to cope withthe complex geometry of stellarators.

Fig. 2. Three field period design option for ARIES-CS.

III. DESIGN PARAMETERS AND LIMITS

The limited space for the internals (blanket, shield,and vacuum vessel) calls for a well optimized, highlycompact radial build and stresses on the crucial role of theneutronics activity. As such, certain features of theseactivities focused on a unique area to compact stellarators,mainly the minimum radial standoff between the plasmaand magnet. A novel approach has been developed forARIES-CS where the blanket at the critical areasurrounding Dmin has been replaced by a highly efficientWC-based shield. This approach places a premium on theblanket to supply all the tritium needed for plasmaoperation. For each blanket concept, the nuclear analysisdetermined Dmin that meets the top-level requirements forthe ARIES power plants. These requirements along withthe adopted radiation limits and key design parameters aresummarized in Table I. The reference design has a powerlevel of 1000 MWe. The blanket must breed sufficienttritium for plasma operation, recover ~90% of the neutronenergy, and protect the shield for the entire plant life. The10% energy leaking into the high-temperature (HT) shieldmust be recovered as high-grade heat to enhance thepower balance. The blanket and shield help protect thevacuum vessel (VV) and all three components protect thesuperconducting magnets for life. An alternate designoption has been proposed to facilitate the maintenancescheme where the VV is relocated outside the magnet. Ifso, a low-temperature (LT) shielding component shouldfollow the HT shield to adequately protect the magnet.The nuclear heating deposited in the internal VV (or LTshield) is so low (< 1% of the total heating) to the extentthat it could be dumped as low-grade heat. Theimplication for the design could be significant as coolantswith high shielding performance (such as water) could beused for the internal VV and LT shield. The magnetdesign7 calls for two consecutive winding packs runningat 15 K and 4 K with MgB2 and NbTi superconductors,respectively. At 15 K, the magnet cooling requirement isgreatly relaxed and the cryogenic load is not a concern.

TABLE I. Design Parameters, Requirements, andRadiation Limits

Fusion powerMajor radius

Minor radius

Peak neutron wall loadingAverage neutron wall loadingOverall TBRPeak damage to structure

He production at VVMagnet: Fast neutron fluence (En > 0.1 MeV) Peak nuclear heating dpa to Cu stabilizer Dose to GFF polyimideMachine lifetimeAvailability

~ 2000 MW8.25 m – 3 FP case7.5 m – 2 FP case1.85 m – 3 FP case2 m – 2 FP case~ 3 MW/m2

~ 2 MW/m2

1.1200 dpa - FS3% burn-up - SiC1 appm

1019 n/cm2

5 mW/cm3

6x10-3 dpa1011 rads40 full power years85%

IV. BLANKET DESCRIPTION

The team has examined five blanket systems. A briefdescription is given here and the accompanying paper8

should be consulted for the details. The five blankets are:- Self-cooled Flibe with beryllium multiplier and

FS structure,- Self-cooled LiPb with SiC/SiC composites,- Dual-cooled LiPb or Li with He and FS

structure, and- He cooled solid breeder (Li4SiO4) with beryllium

multiplier and FS structure.

For each system, we developed two radial builds:shield-only that fits in the local areas where the radialspace between the plasma and magnet is highlyconstrained and blanket and shield for the nominal areaselsewhere. At this stage of the design, a few designparameters that are essential for the accuracy of thebreeding and shielding parameters were missing. In orderto move forward with the nuclear analysis, we had tomake several educated assumptions. For instance, weassumed the peak to average neutron wall loading is ~1.5,the penetrations occupy 2% of the FW area, and thedivertor plates/baffles cover 15% of the FW area.Moreover, we had to rely heavily on the simple 1-Dpoloidal cylindrical model using an average plasma minorradius of 1.85 m and make use of the DANTSYS code9 topredict the overall TBR and Mn, combining 1-D estimateswith blanket coverage fraction. As the ARIES-CS designprogresses, a 3-D analysis is judged essential to confirmthe key nuclear parameters and to generate the exacttoroidal and poloidal neutron wall loading distribution10.

IV.A. Self-Cooled Flibe/FS/Be System

The molten salt system always needs a berylliummultiplier to meet the breeding requirement. The Flibecools the first wall (FW), turns around and flows throughthe beryllium and breeding zones. The system has acoolant outlet temperature of 700 oC and a thermalconversion efficiency (hth) of 45%.

IV.B. Self-Cooled LiPb/SiC System

This design utilizes the advanced SiC/SiC compositesas the main structural material. The LiPb flows throughthe SiC structure at a high speed, and then flows slowlywithin the breeding zone. The high operationaltemperature of the SiC makes it possible to achieve highLiPb outlet temperature (~1100 oC). The significance ofthe high operating temperature lies in the superior thermalconversion efficiency of the system (55-60%).

IV.C. Dual-Cooled LiPb/FS/He and Li/FS/He Systems

The dual coolant option holds the potential to operatethe FS-based design at higher temperatures to ensure highhth (~45%). Both LiPb/FS/He and Li/FS/He designs arevery similar. The helium cools the FS structure while theLiPb (or Li) flows slowly in the breeding zone. Serving asa liner for the structure, 0.5 cm thick SiC insert must beused to control the MHD effect and maintain the FStemperature below 600 oC. The impact of the SiC inserton breeding has not yet been assessed.

IV.D. He-Cooled Li4SiO4/FS/Be System

Generally, solid breeder (SB) designs fail to achievetritium self-sufficiency without neutron multiplier. Theproposed design11 features multiple Li4SiO4 and Be layerssandwiched between cooling channels to efficientlyremove the nuclear heating and operate within thetemperature windows for Be and SB. This design canhandle up to 4.5 MW/m2 peak neutron wall loading anddeliver electricity with hth approaching 45%.

V. BREEDING PERFORMANCE

As will be discussed shortly, the nominal radial buildvaries widely with the proposed blanket concepts. Wesized the blanket to essentially meet the breedingrequirement and protect the shield for the 40 full poweryear (FPY) plant life. Addressing the breeding issue, it isinformative to compare the breeding potential of allbreeders. Figure 3 shows that the beryllium multipliermakes a difference as the Flibe and SB systems offer thehighest breeding, resulting in thin blankets with excessbreeding margin. Note that beryllium may raise safety andeconomic concerns that need further investigation.

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0 10 20 30 40 50 60 70

Lo

cal T

BR

FW/Blanket Thickness (cm)

Flibe

LiPb/SiC

LiPb/FS

Li/FS

Flibe

LiPb/SiC

LiPb/FS

Li/FSSB

SB

Fig. 3. Breeding potential of proposed blankets.

To achieve an overall TBR of 1.1 (local TBR ~1.25),we recommend 30% enriched Li for the Flibe system,90% enriched Li for the LiPb system, natural Li for the Lisystem, and 20-90% enrichment for the multiple layers ofthe Li4SiO4 system. Excluding the back wall (thatsupports the blanket and serves as He manifolds), the FWand breeding zones of the Flibe/FS/Be, LiPb/SiC,LiPb/FS/He, Li/FS/He, and Li4SiO4/FS/Be/He systemsshould be 33, 50, 51, 67, and 45 cm thick, respectively.The arrows in Fig. 3 point to the recommended thicknessof the blanket that satisfies the breeding requirement(overall TBR = 1.1). The analysis assumes that a fewblanket modules can be installed behind the divertorplates (5 cm thick with 50/50 FS/He). The 1.1 overallTBR takes into account the 8% losses in the blanketcoverage for the shield-only zones that will be discussedshortly. This statement is true for the 3 FP configurationonly. A much larger fraction (~20% of FW area) could be

dedicated to the shield-only zones of the 2-FPconfiguration causing a breeding problem. The remainingarea occupied by the blanket may not supply enoughtritium. A uniform blanket everywhere appears likely tobe the most practical option for the 2 FP configuration.

VI. SHIELDING PERFORMANCE

The blanket along with the back wall provides animportant shielding function as it protects the shield forthe entire plant life (40 FPY). The life-limiting criteria forthe FS and SiC structural components have traditionallybeen the displacement of atoms (200 dpa) and burn-up(3%) of the Si and C atoms, respectively. An additionalshielding criterion relates to the reweldability of the VV.The blanket and shield must keep the neutron-inducedhelium at the VV below the reweldability limit (1 appm)at any time during plant operation. All three components(blanket, shield, and VV) help protect thesuperconducting magnets and externals. In our shieldinganalysis, we have considered a safety factor of three toaccount for the uncertainties in the computational toolsand design elements. The selection criteria for theshielding materials included several design parametersthat play an essential role in the acceptability of thematerials. These are the compatibility with the structureand other components, radiation stability, safetycharacteristics, and operating temperature windows.

The blanket and shield have been sized to satisfy thedesign requirements and meet the radiation limits of TableI. There is only one blanket segment in all five designsexcept for the SiC system where two blanket segments areproposed, 25 cm thick each. The segmentation helpsreduce the replacement cost and minimize the wastestream. The outer segment along with the shield is alifetime component.

||

SOL

Vac

uum

Ves

sel

FS

Shie

ld

GapFW

Gap

+ T

h. I

nsul

ator

Win

ding

Pac

k

Pla

sma

Coi

l Cas

e &

Ins

ulat

or

Bla

nket

(LiP

b/F

S/H

e

Gap

Ext

erna

l Str

uctu

reThickness (cm) 5 4.8 47 9 1 32 2 28 > 2 2 31 18

Fig. 4. Schematic of LiPb/FS/He radial build.

149 cm

A representative radial build for the LiPb/FS/Heconcept is illustrated in Fig. 4. The reader is referred tothe following URL for the composition of all componentsand the radial build definition for other blanket systems:http://fti.neep.wisc.edu/aries-cs/builds/build.html. Beingthe closest component to the magnet, the composition ofthe VV (or LT shield) influences the radiation damage atthe magnet significantly. The double-walled VV wasfilled with shielding materials and optimized to achievethe necessary requirements for magnet protection. Severalfillers have been identified for evaluation: water, boratedwater, FS, and B-FS (FS with 3 wt% B). No structuralrole has been envisioned for the fillers. Water wasconsidered for its superior shielding characteristicsrelative to other coolants such as liquid breeders and Hegas. In fact, liquid breeders were excluded as the blanketand, to a lesser extent, the shield provide all the tritiumneeded for plasma operation. Tradeoff analyses of thefillers were conducted for the Flibe, LiPb, and Li4SiO4

concepts. The strong dependence of the damage at themagnet on the choice of the VV fillers is displayed in Fig.5 for the LiPb concept, showing the tradeoff betweenwater and B-FS filler. The optimal 65% water content isdriven by the more restrictive limit of the fast neutronfluence. Sandwiched between 3 cm thick face sheets, thecentral part of the VV consists of 5% FS ribs (dictated bythe structural requirements), 65% water, and 30% boratedFS filler, by volume. For the alternate design optionwhere an external VV surrounds the magnets, thecomposition of the LT shielding component couldresemble the internal VV described above, except for theLi system where water must be excluded for safetyreasons. Only B-FS filler has been considered for thehelium-cooled FS shield of the Li system. The use of analternate hydrogen-based filler (such as zirconiumhydride) could save tens of centimeters in the radial build,but may not be cost effective.

10-1

100

101

102

10 30 50 70 90 Flu

ence

(10

19 n

/cm

2 @ 4

0 F

PY

)

or

Hea

tin

g (

mW

/cm

3 )

Water Content in VV (%)

FluenceLimit

Fluence

Heating

HeatingLimit

Fig. 5. Sensitivity of peak radiation effects at magnet toVV composition, substituting B-FS for water.

0

50

100

150

200

Pla

sma

to M

id-c

oil

Dis

tan

ce (

cm)

Flibe/FS LiPb/SiC LiPb/FS SB/FS Li/FSLi/V

SPPS

SPPS

Fig. 6. Comparison of nominal distance betweenplasma boundary and mid-coil.

Figure 6 compares the nominal distance between theplasma boundary and the mid-coil (D ) for the fivesystems. An appreciable reduction of 40-65 cm in thedimension of the internals relative to the 1996 Li/V SPPSdesign4 has been observed for the Flibe and LiPb systems.The superior shielding performance of both breederscompared to Li and the ability to use water in the VV (orLT shield) along with the much thinner magnetcontributed to the smaller build. Table II lists the radialdimension of the components of all designs. The alloyingand impurity elements of all materials are available at:http://fti.neep.wisc.edu/aries-cs/builds/build.html. Thehelium coolant occupies 10-30 cm of the radial standoff.As noted, the Flibe system offers the thinnest radial build,followed by LiPb, SB, and Li.

TABLE II. Thickness of ARIES-CS Components

Thickness (cm)

FlibeFS/Be

LiPb/SiC

LiPb/FS/He

SB/FS/Be/He

Li/FS/He

FW/B-IBlanket-IIBack WallHT ShieldVV

Total**

33----4625

104

2525--3725

112

52--93228

121

45--203826

129

67--8

103--*

178* External VV.** Excluding assembly gaps.

To further minimize the overall size of the machine,we developed a high-performance, compact shield for thecritical area surrounding Dmin. The main idea is to useshielding materials at a few local spots and deploy thenominal blanket elsewhere. For the 3 FP configuration,Dmin occurs twice per FP and the transition region betweenDmin and the nominal blanket covers ~8% of the FW area.Looking beyond conventional materials (such as steel,water, and borides), tungsten and its compounds possesssuperior shielding performance. Tungsten carbide, inparticular, offers the most compact radial build when usedin the blanket (replacing the breeder) and in the HT shield(replacing the B-FS). While more expensive than thebreeder and steel, the material cost difference is notprohibitive for such limited space.

Figure 7 displays the sensitivity of the neutronfluence at the magnet to replacing the breeder of theblanket and B-FS filler of the shield with WC for theLiPb/FS/He system. The analysis assumes that all shieldsare helium-cooled and the VV dimension (28 cm) and itsoptimal composition remain fixed. Of interest is the FScase where FS filler replaces the blanket breeder. The WCshield is superior and helps reduce the overall thicknessby 30 cm. Considering the positive impact on the overallmachine and economics, it pays to incorporate thecompact WC radial build at Dmin. The radial arrangementof the shield-only components is shown in Fig. 8.Examining the other four blanket concepts, it is found thatDmin varies within 10 cm (see Fig. 9) as the WC dominatesthe performance of the shield-only zones. This meansswitching from one blanket concept to the other wouldchange the major radius of the machine within 1.2 m. Achallenging task would be the integration of the localWC-shields with the surrounding blanket system. Ablanket with variable thicknesses has been envisioned forthe transition region as demonstrated in Fig. 10.

1018

1019

1020

1021

1022

1023

0 20 40 60 80 100

Pea

k F

ast

Neu

tro

n F

luen

ce

(n/c

m2 @

40

FP

Y)

Thickness (cm)

FluenceLimit

WCShield

FSShield

Blanketand Shield

Fig. 7. Sensitivity of magnet damage to shield type.

0

50

100

150

200

Min

imu

m P

lasm

a to

Mid

-co

il D

ista

nce

(cm

)

Flibe/FS LiPb/SiC LiPb/FS SB/FS Li/FSLi/V

SPPS

SPPS

Fig. 9. Comparison between Dmin of the five systems.

Fig. 8. Schematic of radial build of WC-shield for LiPb/FS/He concept.

Ext

erna

l Str

uctu

re

FW

|

Gap

|

Thickness(cm)

SOL

Gap

+ T

h. I

nsul

ator

Coi

l Cas

e &

Ins

ulat

or

Win

ding

Pac

k

Pla

sma

WC

Shi

eld

Vac

uum

Ves

sel

Bac

k W

all

5 4.8 47 11 2 28 2 2 31 18

118 cm

TABLE III. Summary of ARIES-CS Parameters

BreederStructureMultiplierHe Coolant

FlibeFSBe

LiPbSiC---

LiPbFS--

He

SBFSBeHe

LiFS--He

Dmin (m)Nominal D (m)Overall TBRMn

FW lifetime (FPY)hth

1.111.32

1.11.26.5

45%

1.141.4

1.11.16

~55%

1.181.49

1.11.15

5

45%

1.291.55

1.11.34.4

45%

1.162.04

1.11.13

7

45%

Comparing the nominal and minimum radial buildsof Figs. 4 and 8, approximately 20-90 cm saving can beachieved by replacing the blanket and FS-shield withWC-shield at the critical area near Dmin. This translatesinto 20% or more reduction in the major radius,depending on the blanket system.

Table III summarizes the key parameters needed forthe systems code to estimate the cost of electricity. Notethat the solid breeder system offers the highest Mn withthe largest Dmin while the Flibe system offers the thinnestDmin with a moderate Mn. An integrated economic analysiscan self-consistently assess the impact of Dmin, Mn, and hth

on the overall cost of the machine.

Fig. 10. Layout of shield-only zones and nominal blanket,shield, and VV inside the magnet.

VII. COMPARISON BETWEEN STELLARATORDESIGNS

The value of the compact feature of ARIES-CS isfully recognized when comparing the major radii of allstellarator power plants developed to date (refer toSection II). The six designs are shown in Fig. 11. Notethat over the past two decades, the major radius more thanhalved by the advanced physics and technology, droppingfrom 24 m in the early 1980s to 7-8 m for ARIES-CS.Thus, the compact stellarator designs are approaching theadvanced tokamak size. Incorporating the latestadvancements in physics and technology and means ofradial dimension control, ARIES-CS achieved thecompactness that other stellarator designs had not beenable to accomplish before.

WC-Shield

Blanket

Magnet

Pla

sma

Dmin

Shield/VV

0 5 10 15 20 25

m

8

6

4

2

Major Radius (m)

UWTOR-M24 m

ASRA-6C20 m

HSR-G18 m

SPPS14 m

FFHR-J10 m

ARIES-CS2 FP 3 PF

ARIES-STSpherical Torus

3.2 m ARIES-ATTokamak

5.2 m Stellarators

Fig. 11. Major and minor radii of stellarator power plant designs. Typical advanced tokamak and spherical torus arealso included for comparison.

VIII. GENERAL REMARKS AND CONCLUSIONS

Developing an advanced stellarator that meets thedemanding compactness goal required a collaborativeeffort between the physicists and engineers. Recently, theongoing ARIES-CS study demonstrated the potential forcompactness, with specific emphasis on plantmaintainability12. The engineering activities focused onbuilding a compact machine with a lifetime of 45-50years, a thermal efficiency of 45-60%, and an availabilityof 80-90%.

All five candidate blanket concepts proposed forARIES-CS meet the tritium breeding requirement if theblanket covers most of the FW area. A high performanceWC-shield has been developed to fit locally in the highlyconstrained areas where the magnet moves closer to theplasma. The distance from the plasma to the mid-coilvaries widely with blanket concepts, while the WC-shieldregion has much less variability. There will be a non-uniform thickness blanket surrounding the WC-shieldzones. The Flibe option provides the thinnest blanketwhile retaining sufficient breeding margin. Water is a costeffective, efficient shielding material and is highlyrecommended for the VV and LT components. Theengineering and safety limitations of all materials will bean important issue that needs further consideration duringthe continuing study. An integrated economic analysiswill self-consistently assess the impact of the blanket andshield parameters on the overall cost of the machine.

The design benefited substantially from the well-optimized compact radial build that employs WC-shield.This new approach helps minimize the major radius andoverall size of the machine. The economic benefit of thisapproach is yet to be determined. No attempt has beenmade during the first phase of the study to address theintegration issues. Future work present some challenges:integration of the shield-only zones with the rest of theblanket, assessing the need for a separate decay-heatremoval loop for the WC-shield, and handling of therelatively massive WC modules during maintenance.

The historic background provided perspective on theimpressive benefits of the new shielding approach as wellas the advancements in physics and technology. Having7-8 m major radius approaching that of advancedtokamaks, ARIES-CS is considerably smaller thanUWTOR-M2, the first stellarator power plant developedin the early 1980s. Achieving the compactness goal willcertainly improve the economic viability of stellarators.The positive trends in the ARIES-CS physics andengineering are positioning the compact stellarator for abright future.

ACKNOWLEDGMENTS

This work was performed under the auspices of theUS Department of Energy (contract # DE-FG02-98ER54462).

REFERENCES

[1] F. NAJMABADI, “Exploration of CompactStellarators as Power Plants: Initial Results fromARIES-CS Study,” These proceedings.

[2] B. BADGER et al., “UWTOR-M, A ConceptualModular Stellarator Power Reactor,” Universityof Wisconsin Fusion Engineering ProgramReport UWFDM-550 (1982).

[3] G. BOHME, L.A. EL-GUEBALY, G.A.EMMERT et al., “Studies of a ModularAdvanced Stellarator Reactor ASRA6C,” FusionPower Associates Report FPA-87-2 (1987).

[4] R. MILLER and the SPPS Team, “TheStellarator Power Plant Study,” University ofCalifornia San Diego Report UCSD-ENG-004(1996).

[5] C.D. BEIDLER, E. HARMEYER, F.HERRNEGGER et al., “The Helias ReactorHSR4/18,” Nuclear Fusion, 41, 1759 (2001).

[6] A. SAGARA, KUNUGI TOMOAKI, WUYICAN et al., “Studies on Flibe Blanket Designsin Helical Reactor FFHR,” Fusion Technology,39, No. 2, 753-757 (2001).

[7] L. BROMBERG, Massachusetts Institute ofTechnology, Private Communications (January2004).

[8] R. RAFFRAY, L. EL-GUEBALY, S.MALANG, and X. WANG, “Attractive DesignApproaches for a Compact Stellarator PowerPlant,” These proceedings.

[9] DANTSYS: A Diffusion Accelerated NeutralParticle Transport Code System, Los AlamosNational Laboratory Report, LA-12969-M(1995). The FENDL-2 175 neutron 42-gammagroup coupled cross-section library is availableat: http://www.iaea.org/.

[10] M. WANG, D. HENDERSON, T. TAUTGES,L. EL-GUEBALY, and X. WANG, “Three-Dimensional Modeling of Complex FusionDevices Using CAD-MCNP Interface,” Theseproceedings.

[11] R. RAFFRAY, S. MALANG, L. EL-GUEBALY, and X. WANG, “Ceramic BreederBlanket for ARIES-CS,” These proceedings.

[12] X. WANG, S. MALANG, and R. RAFFRAY,“Maintenance Approaches for ARIES-CS PowerCore,” These proceedings.