Journal

of NUCLEAR SCIENCE and TECHNOLOGY, 15[7]

Nuclear

Characteristics Thermal

pp. 490~501

of D-D Fusion

Blanket

and

(July

1978).

Reactor

Blanket-Shield

Blanket,

(II)

Designs

Hideki NAKASHIMA, Masao OHTA and Yasuyuki NAKAO Department of Nuclear Engineering, Kyushu University* ReceivedJanuary 5, 1977 Revised June 15, 1977 The nuclear characteristics of the thermal blanket and blanket-shield designs are analyzed to provide a basis for optimizing the blanket design of D-D fusion reactors. The thermal blanket is devised to yield high energy deposition in a compact blanket through the use of neutron multiplier and energy converter with 1/v neutron absorption cross section. The blanket-shield design, on the other hand, aims at providing acceptably good shielding characteristics to protect the superconducting magnet by incorporating shielding substances within the blanket itself. The results of calculation reveal that the thermal blanket design provides only modest energy deposition in the blanket despite its use of beryllium, which is limited in availability. In contrast, the blanket-shield concept is found to offer attractive possibilities in terms of nuclear characteristics, and the results of this analysis point toward the blanketshield concept as the logical choice for D-D fusion reactor blankets. KEYWORDS: D-D fusion reactor, thermal blanket, blanket-shield, energy deposition, blankets, shielding characteristics

I.

INTRODUCTION

A fusion reactor operating on the deuterium (D-D) cycle can be considered to be of extremely great potential long range interest on account of the almost limitless supply of deuterium as fuel—available from the sea, and for the flexibility afforded to the design of the blanket, being free from requirements relevant to tritium breeding. The D-D reactor blanket, however, must satisfactorily fulfil at least two functions, which are (a) recovery of incident neutron energy and (b) shielding for the superconducting magnet. What is more, the plasma in a D-D reactor is governed by particularly severe conditions, which would tend to lower the power density in the fusion plasma(1). Hence, it is essential from the system economics point of view to aim at : (a) maximizing the power production in the reactor blanket, and (b) minimizing the volume occupied by the radiation shield for protecting

the superconducting magnet, with full advantage taken of the flexibility available for its design. A previous paper(2) covered neutronic and photonic calculations for several D-D fusion reactor blanket models, which threw some light on their nuclear characteristics. It proved that a total energy deposition of about 10 MeV per source neutron could be achieved as a typical case in the models considered, and it was established that to obtain a higher energy gain, it is essential to find a good combination between the material that constitutes the blanket and its geometrical configuration. This brings out the possible utility of the thermal blanket concept, which was first proposed by Darvas(3) for the D-T fusion reactor. The advantage offered by this scheme is that the two functions (a) and (b) cited above can both be satisfied by thermalizing the incident fast neutrons through neutron multiplication, the * Hakozaki

— 20 —

, Higashi-ku,

Fukuoka.

Vol. 15, No. 7 (July

1978)

491

blanket-shield concept is taken up in Chap. III and treated similarly. Chapter IV contains summarized conclusions.

resulting slow neutrons being at the same time more efficiently absorbed in materials presenting 1/u cross sections in order to convert the neutron energy into utilizable form, and more easily shielded to protect the superconducting magnet. The primary design criteria for assuring an effective radiation shield for the superconducting magnet(4)(5) are, among others, (a) to hold below —10-5 the ratio between the total heating rate in the superconducting magnet and that in the blanket, or equivalently, to hold below 1% the power consumed in refrigerating the superconducting magnet, in reference to the electric power output of the

II.

1. Comparative Study In what follows, a comparative study is presented of several materials offering promise for the D-D fusion thermal blanket, along with a detailed description of the procedure of calculation and nuclear data employed. (1) Models and Methods used in Calculation In Fig. 1 are shown the absorption cross sections that are relevant to the concept. For purposes of comparison, the 6Li(n, t) reaction cross section also is given in this figure. For a 1 eV neutron, the Na(n, g) reaction cross section is by four orders of magnitude smaller than the 10B(n, a) and 3He(n, p) reaction cross sections. The Q-value of Na(n, g) reaction is, on the other hand, the highest among these. It should be noted that the kinetic energy of the charged particles emitted by 10B(n, a) or 3He(n, p) reaction is assumed to be deposited locally, i. e., within a negligible distance from the site of reaction, while for the Na(n, g) reaction, the positions of energy deposition could not be determined without considering the transport of the emitted photons.

plant, and (b) to reduce below ~10-6 the ratio of high energy neutron fluxes (E>0.1 MeV) between the manget surface and that at the first wall surface, or equivalently, below~1O% the impairment by irradiation brought upon the critical current density of the magnet. These two criteria could be satisfied by the radiation shield design without requiring an additional magnet shield for the blanket if the blanket itself is loaded with shielding materials such as B4C, Pb or stainless steel. This arrangement could provide the key to a compact design of blanket combining the shielding, which we will call the "blanket-shield" concept. Such blanket-shield arrangements have already been considered in some D-T fusion reactor designs(6)~(9) for the inner part of the torus where only limited space is available to protect the magnet from radiation damage. And the same concept could be utilized to even better advantage in D-D reactors which do not require considerations for tritium breeding and hence the choice of blanket material is more flexible. The aim of the present study is to analyze the nuclear characteristics of both thermal blanket and the blanket-shield concepts for D-D reactors, and to establish basic principles for optimizing the blanket design of D-D fusion reactors. In Chap. II are presented examples of models based on the thermal blanket concept, together with comparative analyses of their nuclear characteristics, to bring out the salient features of this concept. The —

THERMAL BLANKET ANALYSIS

Fig.

1

Sodium 21

—

Absorption cross sections relevant to thermal blanket (BNL-325 Nucl. Data Tables, Sec. A, Vol. 11)

is the

material

to be

used

in fast

492

J. Nucl.

breeder reactors. Helium-3 has the disadvantage of extremely small natural abundance, but this material should become extractable from the exhaust of the D-D fusion plasma, and should possess unique advantages from the fuel cycle viewpoint as medium for circulation in a D-D fusion reactor blanket as neutron absorber at liquid-nitrogen or liquidoxygen temperature(10). In the present instance, this material is envisaged for use as

The material considered for the first wall is V for its low induced activity and its good thermal properties at high temperature. Figure 2 presents, in one-dimensional cylindrical geometry, the blanket configurations considered here. The materials to be used for Zones 4 and 5 are varied in four alternative designs, as specified in Table 1, the remaining zones being constituted of materials common to all four designs*. Design 1 uses He, Designs 2 and 3 10B, and Design 4 3Na as energy converter materials in Zone 5 (energy conversion). For Zone 4 (neutron multi-

gas coolant operating at high temperature. Boron-10 is widely used in fission reactors as material for control rods and shielding, for its large (n, a) cross section for low energy neutrons, absence of high energy capture g-ray emission, and freedom from radioactive decay products. The drawback of 10B is that it produces T from the reactions 10B(n, a)7Li followed by 7Li(n, in)a and 10B(n, t)2a, giving rise to contamination of the blanket com-

plier), Design 3 incorporates Pb, while Be is chosen for the remaining three designs. The transport calculations presented here have all been performed with the one-dimensional, discrete-ordinate transport theory code ANISN(12) using 63 energy groups (42-neutron and 21-g) in P3-S4, approximation. The code RADHEAT(13) developed in the Japan Atomic Energy Research Institute (JAERI) for nuclear heating calculation was used for obtaining the neutron and g-ray coupled cross section set (including KERMA factors), with truncation beyond the P5 Legendre coefficients from ENDF/B-III(4) and POPOP-4(15) libraries. The fusion source neutron is idealized as an isotropic source of 14.06 and 2.45 MeV neutrons, uniformly distributed throughout the

ponents. We here take up for calculation Na, He and 10B as possibly useful materials for 3 constituting energy converters. For the neutron multiplier, we take up, with a similar view, Be and Pb. Beryllium has the advantages of large (n, 2n) cross section, small parasitic capture cross section, and low threshold energy for (n, 2n) reaction, which implies small energy requirement for neutron multiplication. Its demerit is large He production, which gives rise to swelling. To avoid this, it must be used at a density lower than theoretical, thus impairing effective neutron multiplication. Davey et al.(11) have noted the potential application of liquid Pb for its advantages of excellent com-

plasma region (Zone 1) inside the blanket. There is assumed to be an equal number of source neutrons possessing these two energies. The values of atomic density have been taken from Ref. (16) except 3He, for which the same value was adopted as for 4He used in the JAERI conceptual design for a D-T fusion reactor i.e., N3He=0.0003x1024 at./cm3 corresponding to an atomic density of 4He at 500dc and 30 atm.

patibility with most materials, low neutron absorption cross section, high (n, 2n) cross section and its well established quality as gray shield. The disadvantages of this material include its requirement of high pumping power on account of its high density, and its adversely affecting the magnetic field. Lead can be considered a promising substitute for Be if reasons of limited availability restrict utilization of the latter material. As neutron reflector, we take up graphite, which is the standard material for fission reactors. —

Sci. Technol.,

(2) Results and Discussion Neutronic and photonic calculations were carried out with the methods described above, * Zone 7 (Shield) : 1 cmFe+9

cm (CH2) „ +17 cm 40%B4C(100%10B)]

[60%Fe+ +3 cmPb Zone 8 (Superconducting magnet) : 5%Nb +40%Cu+50%Fe+5%void Zone 9 (End Wall) : 1 cmFe 22

—

Vol. 15, No. 7 (July

1978)

Fig.

to obtain

a quantitative

deposition

expected

tions

of

what

follows

material. have

2

General

appraisal the

various

All

the

results

been

1

normalized Table and

Results

arrangement

of the energy

for

per unit source neutron. results for neutron. g-rav

Table

493

combinagiven to

in

values

1 shows the total nuclear

of neutronics

of thermal

blanket

design

heating rates for each zone of the models described above. These results give rise to the following observations : (1) Comparison of the neutronics for Designs 1 and 2 indicates that replacing He by 10B in the energy converter 3 Zone

calculations

23 —

on designs

shown

in Fig.

2

J.Nucl.

494

5 increases the total recoverable energy by about 28%, which reflects the larger number density and Q-value for 10B com-

(1) Description of Models Table 2 describes several designs taken up for neutronics analysis. These designs have been based on considerations given in Ref.

pared with those for 3He. (2) In the neutron multiplier Zone 4, replacement of Be by Pb—of lower (n, 2n) contribution—is seen to diminish by 31% the total recoverable energy, as indicated by a comparison between Designs 2 and 3.

(17). The similarity to the 6Li(n, t) cross section shown by those of the 10B(n, a) and Na(n, g) at low neutron energies (Fig. 1) should permit the salient energy production characteristics of the present thermal blanket to be demonstrated by analogy with the results re-

(3)

Replacement of 10B, in turn by Na in Zone 5 further enhances the total recoverable energy by about 22% (Designs 2 and 4). Design 4 has a small heating rate in Zone 5, but this is more than com-

ported in the above reference. The previously noted differences in mean free path between the reactions taking place in different materials has also been taken into consideration in these designs. As seen in Table 2, Designs1~5 contain 10B in the blanket , which is replaced by Na in Designs6~41. Design 1 is the same as Design 2 except for the division of B4C into two layers sandwiching the Be layer in the latter design. The same applies to the Na layer in Designs 6 and 7. The Be layer is reduced in thickness from 10 cm in Designs 2 and 7 to 5 cm in Designs 3 and 10. Design 4 differs from Design 1 by the introductions of an additional layer of B4C in the graphite zone. A homogeneous mixture of Be and B4C is employed for Design 5, the latter being replaced by Na in Design 11. Designs7~ 9 are intended for comparing the performance of Na,NaAlO2 and NaFeO2. The latter two compounds of Na have been pro-

pensated by greater heating values obtained in Zones 4 and 6, particularly the latter. This shift in heat generating zone is caused by neutrons and g-rays leaking out from the Na region. 2.

Detailed Analysisfor, Energy Production

Increasing

Detailed neutronic analysis aimed at increasing the energyproductionin the blanket have been performed for theBe-10B and BeNa systems which appear to be the systems offering the most promise. To provide a reference, comparisons will be made with the neutronics of the D-T fusion thermal blanket. The mean free path la for the 4He production reaction in boron carbide (B4C)* is only0.014cm for 1 eV neutrons, whereas it is more than 390 cm for the lg of g-ray production in Na(0.15cm for l of T in 6Li(17)). Hence, if B4C is used, only a 2 cm thick B4C layer should suffice to absorb almost all lowenergy neutrons in reactions producing 4He, whereas with Na its g-emitting reactions would necessitate a layer exceeding 4 m in thickness to obtain the same results. Such a requirement would detract seriously from its value in terms of system economics as noted in Chap. I. The large Na inventory in the blanket should also be undesirable from considerations of safety. In the designs taken up in this study, the thickness adopted for the B4C zone is 2 cm, and a corresponding thickness of only 10 cm has been reserved for the Na zone. —

Sci. Technol.,

posed in Ref. (2) as promising substitutes for Na. Vanadium is employed as the first wall material in all designs 1~11. (2) Results and Discussion The neutronic and photonic calculations were based on the same methods as for the comparative study described inSec.II-1. Table 3 summarizes the neutronics results for the designs described in Table 2. The results give rise to the followingobservations: (1) Comparison of the neutronics results between Designs 2 and 3 and Designs 7 and 10 shows that energy production is * In

the present analysis for the B-10B system , B is replaced by B4C(100%10B), since utiliza- 10 tion of pure1 10B is impractical from themetallurgical viewpoint.

24

—

Vol. 15, No. 7 (July

1978)

Table

Table

3

495

2

Results

Description

of

neutronics

of typical

designs

calculations

sensitive to the amount of Be, whereas the V(n, 2n) reaction rate is almost insensitive to such differences in the blanket configuration. (2) Comparing Designs 1 and 2, it is seen that the introduction of the 0.5cm B4C layer behind the first wall causes the V(n, g) reaction rate to drop by about two orders of magnitude. This can be explained from the strong competition of the V(n, g) with the 10B(n, a) reaction in the B neutron spectrum ,-softened by ;the e(n, 2n) reaction. (The same trend is observedwhen a Li (enriched with 6Li) region is introduced behindthe first wall of a D-T fusion reactor(17).) — 25 —

taken

on designs

(3)

up

for

analysis

specified

in Table

2

Between Designs 1 and 4, there is only modest change brought by the addition of the B4C layer in the graphite reflector region. (4) Between Designs 1 and 5, it is evidenced that even the relatively large amount of Be provided in Design 5 contributes only modestly to energy deposition, when this material is homogenized with B4C. A similar observation applies to Designs 6 and 11. (5) Designs 7~9 compare the neutronics performance of Na, NaAlO2 and NaFeO2 as energy converter. Among these three, Design 7 provides the highest heating rate. The presence of O, Al and Fe in

496

J. Nucl.

the compounds results in parasitic neutron capture, which prevents high energy incident neutrons from effectively bringing about (n, 2n) reactions in the Be.

perature—possess a Na atomic density substantially greater than sodium itself, which should make it impossible to shorten the mean free path for the Na(n, g) reaction. In addition, it should be noted that the presence of other materials in these compounds may affect the nuclear performance of the blanket, as noted previously. A summary of the nuclear characteristic parameters obtained for the thermal blanket design is given Table 5, together with those for the standard blanket. The relevant designs are illustrated in Fig. 3. The standard blanket has been derived by replacing with Na the Li in the D-T fusion reactor blanket adopted in a benchmark calculation of the tritium breeding ratio(18). A thickness of 1.04 m has been taken for the blanket thickness(2). The Na blanket is frequently cited in literature(19)~(20) as a promising D-D fusion reactor blanket, which is why we call it the "standard blanket". Density factors of 0.5 for the Be and B4C regions, and 0.8 for the graphite region, are adopted to take account of swelling that may result from excessive He production in these materials.(17) The structural component is assumed to have 10% of the physical volume of the blanket. Without undertaking detailed cost optimization, a composition of 40 cm B4C (100%10B)+10 cm

(6)

Overall examination of the foregoing results indicate that very large values of energy deposition, such as 18 MeV expected in D-T fusion reactor blankets, should be very difficult to obtain with the designs considered here. This is due to the relatively low Q-value preventing the Be-10B system from high energy deposition despite the short mean free path for the 10B(n, a) reaction, while in the case of Be-Na system, energy deposition is again impaired in such a compact blanket by the exceedingly large mean free path for the Na(n, g) reaction, which offsets the advantage of the high Q-value available for this reaction. (In contrast, a D-T reactor embodying Be-6Li system(17) is favored by a relatively short mean free path for the 6Li(n , t) reaction combinded with relatively high Q-value and incident neutron energy, resulting in a value of recoverable energy that may reach 20 MeV.) (7) The atomic densities of Na in sodium compounds are given in Table 4. It is seen that none of the compounds except Na2O —which is thermally unstable at high tem Table

4

Physical

characteristics

of possibly

promising energy converter rials containing Na

Sci. Technol.,

mate-

Table

5

Characteristic thermal

— 26 —

and

nuclear standard

parameters blanket

designs

of

Vol. 15, No. 7 (July 1978)

497

(9) The internal spectral shifter and energy converter (ISSEC)concept(22)is one based on the idea of shifting or tailoring the neutorn spectrum incident on the first structural wall and can be regarded as a variaiton of the thermal blanket concept . In this context, the designs based on this ISSEC concept are also examined. The results indicate that the ISSEC concept is not attractive from the standpoint of energy multiplicaiton in D-D fusion thermal blankets, since the spectral shifter is a nonstructural element that has to be made of graphite or silicon carbide, or three-dimensionally woven carbon fiber(22),in every case involving a large amount of graphite to be contained in this region, which would impede the (n, 2n) reaction in Be.

Pb has been adopted for the shield of the standard blanket, while that for the shield of the thermal blanket has been made the same as described in Sec. II -1.

Fig. 3

(8)

Schematic presentation of (a) thermal blanket and (b) standard blanket (reference design)

.

The results show that the total energy deposition in the thermal blanket is 11.4 MeV, which is higher by about 0.8 MeV than that in the standard blanket. The neutron and g-ray energy leakages to the shield in the thermal blanket are 0.031 and 0.082 MeV, respectively, which are almost comparable with the corresponding values for the standard blanket. In addition, it is found that the fast neutron flux (E>0.1 MeV) evaluated at a distance of 70 cm from the first wall is smaller by a factor of about 50 in the thermal blanket than in the standard blanket. The foregoing observations indicate that a relatively high energy deposition can be obtained in a compact blanket without seriously affecting the shielding funciton when the thermal blanket concept is adopted for a D-D fusion reactor. The limited availability of Be is the most serious impediment to economic utilization of the advantages of high energy production and effective fast neutron attenuation in the thermal blanket D-D fusion reactors.

BLANKET-SHIELD III

ANALYSIS

In what follows we first perform neutronics calculations to systematically investigate and analyze the nuclear characteristics of the selected blankets. Then, for purposes of illustration, a nuclear design is presented, that meets the design criteria noted in Chap. I, and which is obtained by taking account of the results of the preceding systematic calculations. Finally, a comparison of neutronic characteristics is made between three blanket designs, (a) standard blanket design(2), (b) design based on the thermal blanket concept, and (c) based on the blanket-shield concept considered here, to make clear the salient features of the blanket-shield concept. 1. Methods of Calculation and Models Adopted The blanket configuration adopted in the present calculations is shown in Fig. 4, where no consideration is given to the thermal insulator installed between blanket and superconducting magnet. The method of calculation is the same as in the previous chapter. All the results are normalized to unit source neutron.

— 27 —

492

J. Nucl.

2.

Results

and

Sci. Technol.,

Discussion

The ten tentative blanket designs examined here are summarized in Table 6, together with the results of neutronics calculations. The shielding materials in the blanket have been selected with due necessity of withstanding

consideration to the the high operating

temperatures prevailing in the blanket. Materials that should be too expensive for general use in power reactors (e.g. Ta, W) have Fig.

4

Schematic shield

presentation

design

Table

6

Compositions and

servations

of

been excluded. The results

blanket•

concept

results

of alternative of nuclear

blanket-shield

give designs

rise

to the

taken

following

ob-

up,

calculations

:

(1)

A mixture of light and heavy materials presents considerably better shielding characteristics than light or heavy material used alone. This is accountable to mutual complementation of the different mechanisms provided by these two kinds of material for attenuating high energy neutrons : The light material is effective for slowing down high energy neutrons below the threshold energy of inelastic scattering, while above it the heavy material becomes effective. Low inducedactivation blankets constituted of a light material such as Al or graphite alone will inevitably increase the thickness re-

quired for the radiation shield to protect the superconducting magnet. It is seen from the table that greater difficulties would be foreseen for satisfy— 28 —

ing the criterion pMG/pw ?? 10-6 than HTMG/HT ?? t10-5 (2) The total heating rate obtained in a blanket-shield is, on the whole, comparable with those of the standard(2) and thermal blankets, a typical value being 9~10 MeV per incident neutron, depending on the blanket material combination. This means that one of the envisaged functions of the blanket (recovery of incident neutron energy) can be satisfactorily fulfilled by this concept. (3) Of the total heating produced in the superconducting magnet, about 85~98% is due to g-rays. For the g-ray heating, the rays originating within the magnet contributes 98% in the case of the 70% Pb+30% B4C system, whereas with 100% 2O(B) the corresponding contribution H is only 14%. When a 10 cm thick Pb layer is

Vol. 15, No. 7 (July

499

1978)

substituted for the outermost part of the blanket in the 100%H2O(B)system, the gray heating in the superconducting magnet is reduced by a factor of about 5, with an accompanying reduction of HTMG/HT from 1.3x10-2 to 3.3x10-3 But there is a concomitant increase of pMG/pW from 2.4x10-3 to 5.7x10-3 It is common knowledge that Pb is effective for the absorption of g-rays, but is not adequate for attenuating high energy neutrons(23). (4) Among the ten alternative designs, the 70%Fe+30%H2O(B) system would appear to be the most promising in terms of the above-mentioned criteria, although the shielding characteristics of this system are far from satisfying the shielding criteria. We have considered this system to merit further detailed analysis. A 20 cmFe 50 cm (60% Fe 40%H2O(B)) system would have a homogenized composition approximating the 70%Fe+30% H2O(B) system, but HTMG/Hr would be reduced from 1.0x10-4 to 8.2x10-5 and G/r from 7.2x10-5 to 2.4x10-6 compM pared with the original 70%Fe+30% H2O(B) system. This indicates that the spatial arrangement of the materials is very important for improving the shielding characteristics.(24) If a 10 cm thick Pb layer is substituted for the outermost part of the blanket in the 70% Fe+30%H2O(B) system, HTMG/HT is increased to 2.3x10-4 and pMG/pW to 2.8x 10-4. As stated above, Pb does not adequately attenuate the high energy neutrons, and the resulting higher leakage of neutrons into the superconducting magnet results in the greater heating rate in this region. (5) The stratified blanket shown in Fig. 5 is an attempt to meet the above criteria through spatial arrangement of the shielding material which was found to be effective. A summary of the nuclear characteristics parameters of this design is given in Table 7. The total heating rate is 9.8 MeV, pMG/pW 2.4x10-6 and HTMG/HT 7.4x10-6. The arrangement of the shield— 29 —

ing

material

in

this

design

is

not

the

result of exhaustive optimization, but it suffices to prove that a 90 cm thick blan-

Fig.

Table

5

7

Version of blanket-shield design satisfying shielding criteria

Characteristic

nuclear

blanket-shield

design

parameters shown

of

in Fig. 5

ket with shielding materials incorporated within could suffice to meet the shielding criteria and to obtain a reasonable recoverable energy such as 10 MeV. (Note that a more than 180 cm thick blanket-shield is required in some D-T fusion reactor conceptual designs(25)(26).) As criterion values for comparing the characteristics of the three blanket designs—(a) standard, (b) thermal and (c) blanket-shield concept, we present in Table 7 the neutron and g-ray leakages evaluated at a distance of 70 cm from the first wall. (6) Comparing the blanket-shield with the thermal blanket design, it is seen that

500

J. Nucl.

the total energy deposition is higher by about 1.6 MeV (16%) in the thermal blanket than that in the blanket-shield design, but this advantage is offset by the significantly poorer shielding characteristics of the former, its energy leakage and fast neutron flux being about two orders of magnitude higher. To reduce these values by one order of magnitude would require additional shielding at least 15 cm thick. (7)

Between the blanket-shield and the standard blanket designs, the latter design provides about 0.7 MeV (7%) higher total energy deposition, but it should require a more than 150 cm thick blanketshield to provide the desired shielding characteristics.

(8)

It should be expected that as it is the case of the D-T reactor, the cost of a D-D fusion reactor is in large part represented by that of the superconducting magnet. Assuming a constant Bt (toroidal magnet field on the plasma axis), the change DCM brought upon the cost of the magnet per unit length in the toroidal direction by a change Drm of the magnet inner radius hecomes(27)

where gm=gb+ts, gb=gw+tb, where again gw, is the first wall radius, tb the blanket thickness, rb the outer radius of the blanket (inner radius of the shield) and ts the shield thickness. The factor z is considered to be 0.8~1.2. For rw=500 cm*, blanket-shield concept is found to reduce the cost of the superconducting

magnet

by

8~12%

as com-

pared with the thermal blanket, and it would be lower by 16~24% compared with the standard blanket design. These discussions indicate the decisive advantage of the blanket-shield concept from the

neutronics

point

of view.

N.

Sci. Technol.,

CONCLUSIONS

The nuclear characteristics of the thermal blanket and the blanket-shield designs have been analyzed to provide the basis for the optimal blanket design for a D-D fusion reactor. From the system economics point of view, it is essential (a) to maximize the energy production per fusion reaction in the blanket, and (b) to devise a compact radiation shield to protect the superconducting magnet, with advantage taken of the flexibility in blanket design afforded by the D-D fusion reactor. No other compact scheme is likely to be found that should provide energy deposition excelling the thermal blanket concept, but the amount of additional energy deposition actually obtainable with this concept is not as large as to compensate the cost of the very expensive Be of limited availability. The fusion-fission hybrid concept could offer a possible exception, but it should not be a longterm concept on account of the large amount of radioactivity that should accumulate in the form of fission product isotopes. These results lead us to consider the blanket-shield concept as the only promising blanket design for the D-D fussion reactor. About 90 cm of blanket with the shielding substance contained therein should suffice to meet the shielding criteria, while providing at the same time a reasonable recoverable energy such as 10 MeV. In addition, comparison with other blanket designs has brought out the extremely promising qualities of the blanket-shield concept from the neutronics point of view. It has already been mentioned that this concept will without doubt be adopted in D-T fusion reactors, particularly in the experimental reactors to be constructed in the near future. The experience acquired with these reactors would then be directly available for the D-D fusion reactor. No serious difficulty should be encountered in such adoptation to the D-D reactor, considering the * This value has been taken as a typical example from the conceptual design of the "Catalyzed D" fusion reactor in Ref. (1).

— 30 —

501

Vol. 15, No. 7 (July 1978)

flexibility in the selection of blanket materials afforded by D-D reactors, dispensed from provisions for tritium breeding in the blanket. It may thus be concluded that the blanketshield concept should be the logical choice for D-D fusion reactors. This concept will open the possibility of combining in a radiation shield design based on well-tested technology the compactness essential for practical reasons and good performance in terms of system economics. ACKNOWLEDGMENTS The authors wish to express their sincere thanks to Dr. Y. Seki of the Division of Thermonuclear Fusion Research, Japan Atomic Energy Research Institute, for many fruitful discussions. (Text edited grammatically by Mr. M. Yoshida.) -REFERENCES

(2)

(4)

(5)

(6) (7) (8)

(9) (10)

BONANOS, P., CITROLO, J.C. : IEEE Trans. Nucl. Sci., 18, 736 (1971). NAKASHIMA, H., et al.: J. Nucl. Sci. Technol., 14[2], 75 (1977). DARVAS, J.(3) : A design with low lithium and tritium inventories, Presented at the IAEA Workshop on Fusion Reactor Design, IAEA, Culham, (1974). SEKI, Y.: Evaluation of shielding design of super conducting magnet (I), JAERI-M 6046, (in Japanese), (1975). MCCRACKEN, G.M., BLOW, S.: The shielding of super-conducting magnets in a fusion reactor, CLM-R 120, (1972). CONN, R.W., JASSBY, D.L. : Trans. Amer. Nucl. Soc., 21, 24 (1975). CONN, R.W., GOHAR, Y. : ibid., 22, 16 (1975). SEKI, Y., et al.: Preprint 1975 Fall Meeting At. Energy Soc. Japan, (in Japanese), A25, (1975). ABDOU, M.A.: Trans. Amer. Nucl. Soc., 22, 19 (1975). OWEN, D.L., IMPINK, A.J., Jr. : Proc. Technol-

ogy of Controlled Thermonuclear Fusion Experiments and the Engineering Aspects of Fusion Reactors (Texas, 1972), CONF-721111, (1972). (11) DAVEY, W.G., et al.: Proc. 1st Topical Meeting Technology of Controlled Nuclear Fusion, p.155 (1974). (12) ENGEL, W.W., Jr. : A user's manual for ANISN, K-1693, (1967). (13) MIYASAKA, S., et al.: Code system for the radiation heating calculation on nuclear reactor, RADHEAT, JAERI-M 5794, (in Japanese), (1974). (14) DRAKE, M.K., (ed.) : Data formats and procedures for the ENDF neutron cross section library, BNL-50279, (1970). (15) FORD, W.F., III.: The POPOP4 library of neutron induced secondary gamma ray yield and cross section data, CTC-42, (1970). (16) HUNEGRFORD, H.E. : "Engineering Compendium on Radiation Shielding", Vol. II, (1968), Springer-Verlag. (17) ABDOU, M.A., et al.: Nucl. Technol., 26, 400 (1975). (18) STEINER, D. : Analysis of a bench-mark calculation of tritium breeding in a fusion reactor blanket : The United States contribution, ORNLTM-4177, (1973). (19) CARRUTHERS, R., et al.: The economic generation of power from thermonuclear fusion, CLM-R85, (1967). (20) POST, R.F. : Mirror system : Fuel cycles ; Loss reduction and energy recovery, "Nuclear Fusion Reactor" (Proc. Conf. Culham, England, 1969), p.88 (1969). (21) MCNALLY, J.R., Jr. : Fusion chain reactor prospects and problems, ORNL-TM-4575, (1974). (22) CONN, R.W., et al.: Nucl. Technol., 26, 125 (1975). (23) GABRIEL, T.A., et al.: Shielding design calculations for ORMAK-F/BX, ORNL-TM-4619, (1974). (24) KRIESE, J.T., STEINER, D.: Magnet shield design for fusion reactors, ORNL-TM-4256, (1973). (25) SAKO, K., et al.: Proc. 5th Int. Conf. Plasma Physics and Controlled Nucl. Fusion Research, IAEA, Tokyo, 1974, CN-33/G 1-5, (1974). (26) CONN, R.W., et al.: ibid., CN-33/G 1-2. (27) ABDOU, M.A., MAYNARD, C.W. : Ref. (11), p. 685.

— 31 —