ORNL-2378.txt

 

. 3 4456 03k1L85 7 | ORNL.2378

ENT HEPUHT C-84 — Reactors—Special Features
| of Aircraft Reactors

M-3679 (20th ed. Rev.)

    

  
   
   
  
 

 

Cy / 55 4

 

 

 

,
."J
-"‘; SALT REACTOR PROGRAM
RLY PROGRESS REPORT
# ' D ENDING SEP TEMBER 1, 1957
i,
A%

 

   

 

 

s
 

 

 

LEGAL NOTICE

 

This report was prepared as an account of Government sponsored wark. Neither the United States,

nor the Commission, nor ony person acting on behalf of the Commission:

A. Makes any worranty or representation, express or implied, with respect to the accuracy,
completeness, or usefulness of the information contained in this report, or that the use of
ony information, apparatus, method, or process disclosed in this report may not infringe
privately owned rights; or

B. Assumes any liabilities with respect to the use of, or for domages resulting from the use of
any information, apparatus, method, or process disclosed in this report.

As used in the above, '‘person acting on beholf of the Commission’’ includes any employee or

contractor cof the Commission 1o the extent that such employee or contractor prepares, hondles

or distributes, or provides access to, any information pursuant to his employment or controct
with the Commission.

 

 
 

s

ORNL-2378
C-84 — Reactors~Special Features
of Aircraft Reactors

M-3679 (20th ed. Rev.)

This document consists of 30 pages.

Copy /y-s—of 310 copies. Series A.

Contract No. W-7405-eng-26

MOLTEN SALT REACTOR PROGRAM
QUARTERLY PROGRESS REPORT

for Period Ending September 1, 1957

H. G. MacPherson, Program Director

DATE ISSUED

DEC 271957

 

DAK RIDGE NATIONAL LABORATORY
Operated by
UNION CARBIDE NUCLEAR COMPANY
Division of Union Carbide Corporation
Post Office Box X
Ock Ridge, Tennessee

 

MARTIN MARIETTA ENERGY SYSTEMS LIBRARIES

RN

3 4456 03L1185 7

 
-

PENG A O

10.
11,
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24,
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42,
43.
44.
45.

 

INTERNAL DISTRIBUTION

R. G. Atfel

L. G. Alexander

C. J. Barton

M. Bender

E. S. Bettis

D. S. Billington
F. F. Blankenship
E. P. Blizard

C. J. Borkowski
W. F. Boudreau

G. E. Boyd

M. A, Bredig

E. J. Breeding

R. B. Briggs

W. E. Browning

F. R. Bruce

A. D. Callihan

D. 0. Campbell

D. W. Cardwell

W. H. Carr

G. l. Cathers

C. E. Center (K-25)
R. .

R.
C.
J.
Ww.
R.
F
D.
J.
L.
D.
W.
L.
W.
J.
D.
A,
E.

. Clifford
Coobs

. Cotirell
Crouse

. Culler

. Cuneo

E
H.
B
S.
L
R
H. DeVan
M
A
K
B
K
Y

. Eister
« Emlet (K-25)
. Ergen
. Estabrook
E. Ferguson
P. Fraas
Franco-Ferreria
J. H. Frye, Jr.
W. T. Furgerson
R. J. Gray
J. L. Gregg
A. T. Gresky"

o
- Q&‘?_ P

 

ORNL-2378
C-84 ~ Reactors—Special Features
of Aircraft Reactors
M-3679 (20th ed. Rev.)

46. W.R. Grimes
47. A. G. Grindell
48. E. Guth

49. C. S. Harrill

50. M. R. Hill

51. E. E. Hoffman
52. H. W. Hoffman
53. A. Hollaender
54. A. S. Householder
55. J. T. Howe

56. W. H. Jordan
57. G. W. Keilholtz
58. C. P. Keim

59. F. L. Keller

60. M. T. Kelley
61. F. Kertesz

62. J. J. Keyes

63. B.W. Kinyon
64. J. A. Lape

65. R. B. Lindauer
66. R.S. Livingston
67. M. . Lundin

68. R. N. Lyon

69. W. B. McDonald
70. H. G. MacPherson

71. R. E. MacPherson
72. F. C. Maienschein
73. W.D. Manly

74. E. R. Mann

75. L. A. Mann

76. J. R. McNally

77. F. R. McQuilkin
78. R. V. Meghreblian
79. R. P, Milford

80. A. J. Miller

81. E. C. Miller

82. R. E. Moore

83. J. G. Morgan

84. K. Z. Morgan

85. E. J. Murphy

86. J. P. Murray (Y-12)
87. M. L. Nelson

88. G. J. Nessle
89- ' R- B.V‘Ql_iver
90, LG, ‘@erholser
91. P. Patriarca
92. S. K. Penny
93. A. M. Perry
94, D. Phillips
95. J. C. Pigg
96. P. M. Reyling
97. A. E. Richt
98. J. T. Roberts
99. M. T. Robinson
100. F. E. Ramie
101. W. Savage
102. Savelainen
103. Schultheiss
104. Scott
105. L. Scott
106. E. D. Shipley
107. A. Simon
108. O. Sisman
109. J. Sites
110. M. J. Skinner
111. A. H. Snell
112. C. D. Susano

H.
A.W.
R. D.
D.
J.

158-160.
161-162.
163.
164.
165.
166-168.
169.
170-171.
172.
173.
174-175.
176.
177.
178.
179-192.
193.
194-196.
197.
198.
199.
200.
201.
202-207.
208.
209.
210-211.
212,
213.

 

113. J. A. Swartout

114. E. H. Taylor

115. R. E. Thoma

116. D. B. Trauger

117. D. K. Trubey

118. F.C. YonderLage

119. G. M. Watson

120. A. M. Weinberg

121. M. E. Whatley

122. J. C. White

123. G. D, Whitman

124. E. P. Wigner (consultant)

125. G. C. Williams

126. J. C. Wilson

127. C. E. Winters

128. W. Zobel
129-130. ORNL — Y-12 Technical Library

Document Reference Section

131-153. Laboratory Records Department

154. Laboratory Records, ORNL R. C.
155-157. Central Research Library

EXTERNAL DISTRIBUTION

Air Farce Ballistic Missile Division

AFPR, Boeing, Seattle

AFPR, Boeing, Wichita

AFPR, Curtiss-Wright, Clifton

AFPR, Douglas, Long Beach

AFPR, Douglas, Santa Monica

AFPR, Lockheed, Burbank

AFPR, Lockheed, Marietta

AFPR, North American, Canoga Park

AFPR, North American, Downey

Air Force Special Weapons Center

Air Materiel Command

Air Research and Development Command (RDGN)
Air Research and Development Command (RDTAPS)
Air Research and Development Command (RDZPSP)
Air Technical Intelligence Center

ANP Project Office, Convair, Fort Worth
Albuquerque Operations Office

Argonne National Laboratory

Armed Forces Special Weapons Project, Sandia
Armed Forces Special Weapons Project, Washington
Assistant Secretary of the Air Force, R&D

Atomic Energy Commission, Washington

Atomics International

Battelle Memorial Institute

Bettis Plant (WAPD)

Bureau of Aeronautics

Bureau of Aeronautics General Representative

 
 

214. BAR, Aerojet-General, Azusa

215. BAR, Convair, San Diego

216. BAR, Glenn L. Martin, Baltimore

217. BAR, Grumman Aircraft, Bethpage

218. Bureau of Yards and Docks

219. Chicago Operations Office

220. Chicago Patent Group

221. Curtiss-Wright Corporation

222. Engineer Research and Development Laboratories
223-226. General Electric Company (ANPD)

227. General Nuclear Engineering Corporation

228. Hartford Area Office

229. ldaho Operations Office

230. Knolls Atomic Power Laboratory

231. Lockland Area Office

232. Los Alamos Scientific Laboratory

233. Marquardt Aircraft Company

234. Martin Company

235. National Advisory Committee for Aeronautics, Cleveland

236. National Advisory Committee for Aeronautics, Washington

237. Naval Air Development Center

238. Naval Air Material Center

239. Naval Air Turbine Test Station

240. Naval Research Laboratory

241. New York Operations Office

242. Nuclear Development Corporation of America

243. Nuclear Metals, Inc.

244, Office of Naval Research

245. Office of the Chief of Naval Operations (OP-361)

246. Patent Branch, Washington
247-250. Pratt & Whitney Aircraft Division

251. San Francisco Operations Office

252. Sandia Corporation

253. School of Aviation Medicine

254. Sylvania-Corning Nuclear Corporation

255. Technical Research Group

256. USAF Headquarters

257. USAF Project RAND

258. U. S. Naval Radiological Defense Laboratory
259-260. University of California Radiation Laboratory, Livermore
261-278. Wright Air Development Center (WCOS|-3)
279-303. Technical Information Service Extension, Oak Ridge
304-306. H. M. Roth, Division of Research and Development, AEC, ORO
307-309. S. R. Sapirie, AEC, ORO

310. F.C. Moesel, Division of Research and Development, AEC, Washington

*
 

o mem I -

CONTENTS

P RE F ACE et et e ettt e e et et e et e st et e ea st ere et eraea e sab et e aer et e neetareaere e e ereeaenteesaeteens vii
SUMMARY ettt ettt et et e ete et eer e et een et e eae et eet e e et e et e st e resae e et et e e st et et s eaee e s eereneensens 1
DYNAMIC CORROSION STUDIES ...ttt et s e e saeeeenesaensete st e st e et aese e enes e esste e teeen, 3
Thermal Convection Loop Tests ... ettt ettt et e et e e st e e e e eet e e ete e et e eeeneaeeseeees 3
PUMP Lo0p TeStS .ottt et st s et er et ene et esseransenee seetesteseemeereneesereaseane, 3
PROGRAM PLANNING ...ttt ettt et e e s e e s e et aeea et saseasesseaneasassssssessesessenseas 5
DESIGN OF IN-PILE NATURAL CONVECTION LOOPS e eeeeeee et ettt eee e e 7
GCAMMA HEATING OF CORE SHELL .ottt et et e e st ea e essreeataneeaseneens 9
HEat Generation .......cccoveciieioiiieieee et ee ettt e e easaasae e s eat s e bens e st et e teteeeess e em b saeseneeseaeeeeansees 9

T emPEratUre Pattern.......oovi ittt e sttt s e se s st esae e sreasnteeatssassamessteaneseesesneeeseenaesansenneas 9
TREIMAl SHr@SS oottt ettt e et e et e et e et e terara e e eateanasaentestaaneeanesanseeratesran 9
MERCURY AS A SECONDARY HEAT TRANSFER FLUID ..o eveeee et n s aaeeeen s 10
NUCLEAR CALCULATIONS ...ttt st et e e ettt sttt eeteteee saeaas e e asessanssesnnenssnnoresaenennssens 13
The UNIVAC Program Bcusol ...ttt et s es e et aan et e sb e e tsa e senas 13
Description 0f Program ..ottt ettt e v et eeve s e e eeneereeaa e st ererataentrenesaeteeeene 13

OcUsOl Cross SECHIOMS ..ooiviiiiriiieeiireni e e et re et et taest e e eeessseess st sae et eareeeeaesesaraneerreseeasasenaeaes 13
Parametric Studies of Reference Design Type of ReaCtors ........ocvovevioiiieeeieeeeeeeeeeee e, 13
Clean Reactor CalaUlations ..ot ettt eee s st et e ae et st et e e saeseesseeeases et rans 13
Variation of Nuclear Characteristics with Period of Operation ....occovvoiieeiieceiceeeeeee e, 16
Steady-State Reactors ..o et b s e bt b s et s et s aeen e e eeeeetene 18

FUEL CYCLE ECONOMICS .ottt ettt s e e et e st st e et e s e e estean e eatean s neeereeassseeseeeseoen, 19
Fission-Product POISONING ..ottt et v et sr et aseneen et sansens 19
NATURAL CONVECTION REACTOR ..ottt ettt ste et ses e et e e seeeesena s e enesaaeeesaeons 20
 

 
 

PREFACE

The efforts of the Molten Salt Power Reactor study group have been summarized, to
April 29, 1957, in a report entitled A Preliminary Study of Molten Salt Reactors (ORNL
CF-57-4-27). The report is a review of the state of the pertinent technology as understood
by the group, and of the study of the performance of a Reference Design Reactor (RDR),
a two-region, homogeneous, 600-Mw central station power reactor.

The work of the Molten Salt Reactor Program will henceforth be reported in quarterly
progress reports. They will report work which will be an extension of the work reported

in ORNL CF-57-4-27, and it will be assumed that the reader is familiar with that report.

 

vii
 

MOLTEN SALT REACTOR PROGRAM QUARTERLY PROGRESS REPORT

SUMMARY
H. G. MacPherson

The primary purpose of the present program is
to obtain the information necessary to initiate a
reactor construction project. Thus the two major
efforts are a materials-compability program and
a continuation of reactor studies. The former will
determine the practicality of building a reactor,
and the study program will determine the most
suitable reactor for construction.

MATERIALS COMPATIBILITY

The materials-compatibility program has three
major facets: the metallurgy and chemistry of the
materials themselves, out-of-pile experiments on
the nature and extent of corrosion, and in-pile
experiments to test the effects of radiation and
fission products.

Work on the metallurgy and chemistry of mate-
rials, as sponsored by this program, is just getting
under way in significant amounts, and thus no
extensive writeups of work in these fields are
included in this report. However, a better appre-
ciation of the significance of the mechanism of
corrosion has been obtained.

Two types of mass transfer can occur in alloy-
salt systems. In one, free metal can be deposited
in a dendritic form in the cold portions of the
system. If this occurs, the tubing will become
plugged and a reactor based on this system will
soon be inoperative. There are some salts which
deposit pure metal in this way when in combination
with Inconel under the influence of a high-tem-
perature gradient.

The other type of mass transfer is one in which
pure metal cannot be deposited in the cold leg.
This type occurs with NaF-ZrF,-UF, salt in
inconel. At equilibrium, the salt cannot dissolve
enough chromium from the hot Inconel to precipi-
tate pure chormium in the cold zone; however,
chromium may be deposited as a chromium-rich
surface alloy. The mass transfer can proceed only
as fast as chromium can diffuse from the metal
interior to the surface at the hot end, or from the
surface to the interior at the cold end.

Two very significant deductions can be made
for situations where the latter type of mass

transfer holds. Since the rate-limiting step is

 

metallic diffusion, it should be very temperature-
dependent and it should proceed at a low rate at
projected power reactor temperatures. Furthermore,
although the chromium-leached hot portions will
be greatly weakened, they will not be made thinner
and will not leak salt. Thus it should be possible
to design a reactor system which would not fail
because of corrosion,

The results of the first 1200°F, 1000 hr thermal
convection loop are described.
rosion

A very low cor-
rate was found, tending to confirm the
expectation of good stability at power reactor
The status of the pumped loop test,
which is under way, is also given.

In the section ‘*Program Planning,”’ the plans for

temperatures.

a more complete materials-compatibility program are
outlined. The bulk of the initial work will consist
of out-of-pile thermal convection and pumped loops.
A thermal convection in-pile loop is being de-
signed for installation in the LITR. As funds
become available, the present Solid State Division
in-pile loop program will be supported and, where
possible, augmented.

The alloy that is most favored for power reactor
use is still INOR-8 (17 wt % Mo~7 wt % Cr=5 wt %
Fe~balance Ni). Production-sized heats of it have
now been made by two companies, and its fabri-
cation characteristics are being determined. lts
present cost is about $4.50 per pound, and its
eventual cost is estimated to be about $2.50 per
pound. A proposal for a metallurgical program
suitable for power reactors is being prepared.

REACTOR STUDIES

The reactor studies are proceeding along two
principal lines: further clarification of problems
with respect to the Reference Design Reactor,
and a study of smaller-powered reactors suitable
for early construction.

One of the important questions in the RDR de-
sign was that of the effect of internal nuclear
heating on the walls of the reactor core. This
problem has been examined and it has been found
that the nuclear heat source in the walls amounts
to 17.5 w/cc. This will result in thermal stresses
that are believed to be tolerable, but the point
MOLTEN SALT REACTOR PROGRAM PROGRUE REBORT

would require further study during the detailed
design stages of such a reactor.

The RDR heat-removal system was designed
with compatible fluids at ad{acent positions every-
where except at the boiler and superheater. The
use of sodium to heat water and steam is still
regarded as a problem because of the violence
of the sodium-water reaction and because the
products of the reaction ore corrosive both in
the sodium and in the water. The results of a
study of the use of mercury to replace sodium
indicate that mercury vapor for heat transfer has
a great deal of merit. In addition, if the mercury
is used to drive an auxiliary turbine, significant
increases in plant efficiency will result. It is
known that mercury introduces its own problems,
and the balance between these and the problems
of sodium must be analyzed more fully.

The nuclear calculations on which the RDR was
based do not represent an exhaustive study or an
optimization of conditions. The two-region, homo-
geneous reactor systems are being examined in
an orderly fashion. This has involved the improve-
ment of the UNIVAC code, the refinement of cross-
section data, and the invention of a new code for
the Oracle. With the latter refinement, it will be
possible to follow the course of o reactor from
a clean initial condition to one requiring reproc-
essing, and to keep track of the breeding ratios
and the requirements for fuel addition.

The most recent calculations are for clean re-
actors with diameters varying from 6 to 10 ft,
thorium contents varying from 0 to 1 mole %, and
with LiF-BeF, base salf. The minimum critical
masses were for 6- and 8-ft reactors at 47 and
49 kg of U233, With any appreciable external in-
ventory, such as is required for power generation,
a core diameter which is larger than the 6 ft pro-
posed for the RDR results in a lower critical

inventory. A diameter of about 8 ft seems to be
a good choice. The results of the calculation are
still under study, and full conclusions cannot yet
be drawn.

The effectof fission-product cross sections higher
than the values assessed by Hurwitz and Greebler
has been estimated.! The total effect on reactor
economics would be such as to increase the fuel

cycle costs by 0.25 to 0.5 mill/kwhr.

In the consideration of lower-powered reactors,
great simplicity of construction seems more impor-
tant than achieving an exceptionally low fuel
cycle cost or a high thermal efficiency.
small

in a
power plant the capital charges tend to
greatly outweigh fuel costs, and maintenance can
be a big item it simplicity is not kept paramount.
In a search for simple molten salt reactors, the
natural convection reactor has seemed the best
solution to date. In this reactor the primary heat
exchanger is placed at some height above the
reactor, and the difference in density between
two columns of salt, one heated and the other
cooled, provides the driving force to circulate the
fuel.

So far, little attempt has been made to optimize
a design, yet the indications are very encouraging.
For powers up to at least 100 thermal Mw, the
additional external volume that a thermal convec-
tion reactor requires over a forced circulation
system does not seem excessive. The attraction
of an all-welded system with no moving ports in
the fuel circuit is very great. It is onticipated
that a reactor of this type will be the most suitable
for early construction.

 

'H. G. MocPherson et al., A Preliminary Study of
Molten Salt Power Reactors, ORNL CF-57-4.27 (April 29,
1957).

2 4 T
PERIOD ENDING SEPTEMBER 1, 1957

DYNAMIC CORROSION STUDIES

J. H. DeVYan

THERMAL CONVECTION LOOP TESTS

Thermal convection loops are now being operated
in the temperature range of 1200°F in order to
evaluate several fluoride salt and structural metal
combinations for possible application in a molten
fluoride power reactor. The first such test, which
utilized an Inconel loop and a salt mixture of
NaF-ZrF ,-UF,, was operated successfully for
1000 hr and has been examined. The operating
conditions for the test (Loop No. 1161) were as
follows:

Loop design Thermal convection harp
constructed of ‘?'/B-in.

sched-10 Inconel pipe

Fuel 122, NoF-ZrF4-UF4
(57-42-1 mole %)

Fluoride salt

Maximum fluoride-Inconel 1250°F
interface temperature

Maximum mixed mean 1200°F
fluoride temperature

Fluoride temperature 100°F

gradient

Following operation, the test salt was aliowed
to freeze in place in the loop. Test samples were
then cut from several sections of the loop, as
shown in Fig. 1, and the fluorides contained in
these samples were melted out under a helium
blanket. Samples of the salt were chemically
analyzed, and sections of the loop were examined
metallographically.

Hot-leg and cold-leg samples were quite similar
in appearance, both showing less than 1 mil of
corrosive attack. Corrosion in these sections oc-
curred in the form of surface roughening and
shallow subsurface void formation, as can be
seen in Figs. 2 and 3. Cold-leg sections were
found to be entirely free of metal deposits both
in visual and metallographic examinations.

Chemical analyses of the fused salt tcken be-
fore and aofter the test are shown below. Note

J. R. DiStefano

that only a slight increase in chromium, pre-
sumably contained as CrF,, occurred in the salt
during the test.

U Ni Cr Fe
(%) (ppm) (ppm) (ppm)
Before-test salt analysis 2.78 60 35 330
After-test salt analysis
Hot leg 2.66 25 105 80
Cold leg 265 20 110 70

PUMP LOOP TESTS

An inconel pump loop of the design shown in
Fig. 4 was placed in operation in February 1957.
The loop, designated as CPR No. 1, provides for
the operation of two fluid circuits, the primary
circuit containing fuel 122, and the secondary
circuit containing sodium.

UNCLASSIFIED
ORNL-LR-DWG 23454

 

" | METALLOGRAPHIC
SAMPLES

HOT—LEG CHEM-

E ISTRY SAMPLE

SHELL HEATERS

iISTRY SAMPLE

E
%
SIX 6-in. CLAM- %
%

6 j COLD—LEG CHEM-

 

 

| METALLOGRAPHIC
[ sampLES

 

 

 

 

 

 

Figa ]o
vection Loop with Location of Metallographic Samples.
(Secret with caption)

Diagram of a Standard Inconel Thermal Con-
MOLTEN SALT REACTOR PROGRAM PROGRESS REPORT

UNCLASSIFIED
PHOTO T-13045] |

 

 

 

Fig. 2 Photomicrograph of Inconel Tube Wall at Point
of Maximum Loop Temperature. Metal particles appearing
above specimen surface are burrs produced during metalle-
graphic preparation. 250X. Reduced 32%. (Secret with

caption)

Heat is supplied to the fuel circuit by direct
resistance and is transferred to the sodium circuit
by means of a U-bend heat exchanger. Heat is
then taken from the sodium by an air blower, as
shown in the lower right portion of the diagram
(Fig. 4). A centrifugal pump circulates the salt,
while an electromagnetic pump is used in the
sodium circuit.

Operating conditions for the loop are as follows:

UNCL ASSIFIED
PHCTO T-13046

 

"y ‘A" W . T L™
£ "% W& WNICKEL PLATE! 2
- Ty .
] ' . . Pl ’ B

 

FE

Maximum salt temperature 1190°F
Maximum salt tube temperature 1260°F
Minimum salt temperature 1060~ 1070°F
Maximum sodium temperature 1120°F
Minimum sodium temperature 1060-1070°F
Reynolds number for salt 5000
(in heat exchanger)
Reynolds number for sodium 97,700

   

 

o
b
o

  

o]

INCHE

 

 

 

_: (008
2 \ 007
a , i
’ . 7| Loos
* ’ 903
- -t & 4
010
01
. o ol '
. ‘ e i
* 012
o - .Ol‘.'o;
e
Fig. 3. Appearance of ColdlLeg Specimen. Specimen

was nickel-ploted prior to polishing to preserve any metal

deposits which might be present. 250X. Reduced 32%.

(Secret with caption)

UNCLASSIFIED
ORNL-LR-DWG 23452

HEATER CONNECTION LUG

a
0

SALT

  

Na ANNULUS

L\ SALT ~No

It is intended that the loop will operate for a
period of 10,000 hr. No serious difficulties were
encountered in the first six months of operation,
although a slight increase in the pressure drop
in the fused-salt circuit occurred during the initial
1500 hr. This chonge was apparently caused by
salt freezing at some point in the circuit, since
running the loop isothermally at 1200°F appeared
to correct the conditions and restored the loop
to its original conditions,

 

 

 

 

 

 

 

’ HEAT EXCHANGER
SALT
$1260°F EM PUMP
E:[ 1120°F 1060°F I COOLER
1190 °F -‘ —— Cl
Na

 

 

 

SALT PUMP

£

 

et

1060°F

 

Fig. 4. Schematic Drawing of Inconel Bifluid Loop
CPR No. 1. (Secret with caption)
PERIOD ENDING SEPTEMBER 1, 1957

PROGRAM PLANNING
F. C. VonderLage

The beginning of FY-1958 marked the start of
detailed planning of an experimental program based
on the findings and recommendations contained in
o previous report! by the Molten Salt Reactor
Program group. The plan assumes a moderately
increasing rate of financial support to reach a
level of several million dollars per year at the

end of two or three years.

The scope of work being planned is limited
accordingly to the following areas:

1. testing materials and evaluating, ot tempera-
tures and in the environments applicable to
power reactors, the structural properties of
materials which are or will become available
during the period;

2. studying maintenance requirements of alternate
reactors and devising and testing remote main-
tenance techniques;

3. conceiving, studying, and evaluating alternate
reactor embodiments in the molten salt class.

Materials testing and structural properties eval-
uation, in part 1 above, refer, in the main, to
evaluation of the structural integrity of container
materials in circulating molten salt or hot liquid-
metal circuits. A wealth of data for higher tem-
perature, short-life reactor operation is available;
very few data are available at the lower operating
temperatures which apply to power reactors. It
is virtually certain, on theoretical grounds, that
container materials at these lower temperatures
will show, on the whole, improved performance,
especially as related to corrosion-initiated weak-
ening. It is an essential objective of the materials
testing program to demonstrate that such improved
performance is sufficiently great to promise an
attractively long-lived reactor.

Tests needed for container materials fall naturally
into the following three categories, each suited to
its purpose:

1. tests for basic compatibility between circulating
fluid and container materials, which are in-

tended to demonstrate that there will be no

 

"H. G. MacPherson et al., A Preliminary Study of
Molten Salt Power Reactors, ORNL CF-57-4-27 (April 29,
1957).

mass transfer of metals to form metallic de-
posits in cold zones {which eventually would
cause plugging) and that corrosion of material
from within the walls of the containing metal
will result at most in the production of dis-
connected voids, thus leaving the walls im-
pervious to the fluid even after long use;

2. tests for determining the structural properties
of materials under reactor design conditions
and environments, which are intended to demon-
strate that the properties of the
material are sufficiently attractive to warrant
its use for the construction and operation of
reactor components which would have tolerable
life expectancy, and to procure, for design
purposes, quantitative

structural

material s-performance
data in environments and temperatures to which
they will be subjected in a reactor;

3. tests to demonstrate the adequacy of techniques
for joining materials and fabricating compo-
nents.

Container materials to be tested for compati-
bility with salts include Inconel, both because
it is commercially available and because cor-
relation with existing higher temperature data
is desired, and will include INOR as soon as it
becomes available in tube form. For liquid metals,
Croloy and type 316 stainless steel are being
considered, perhaps in bimetallic loops. The
choice of metals for liquid metals tests has not
yet been made.

The choice of salts for compatibility tests has
been guided by nuclear, heat transfer, and safety

requirements. The salts are listed in Table 1.

Planned also are limited tests to determine
possible additional changes in the structural
properties of container materials when the molten
salt is in contact with graphite also.

For testing basic compatibility of fuel salts,
successful in-pile tests simulating actual reactor
Such
tests are very costly, and design of experimental
equipment simulating reactor conditions is severely
fimited, if not impossible, principally because of
space and safety requirements which must be met
Accordingly, the plan is to do most of

conditions would be the most reassuring.

in-pile.
MOLTEN SALT REACTOR PROGRAM PROGRESS REPORT

Table 1. Salts to Be Tested for Basic Compatibility

Composition in mole %

 

LiF NaF KF ZrF BeF, UF, ThF

 

4 4
Fuel Salts
57 42 1
55 41 4
53 46 1
53 46 0.5 0.5
53 46 1

Blanket Salts
58 35 7

58 35 7
Coolant Salts

46.5 1.5 420
35 27 38

 

the basic compatibility testing out-of-pile and
to make a few linking in-pile tests.

Work is presently proceeding on the design of
a thermal convection in-pile loop to be installed
in available LITR space. In-pile thermal con-
vection loops are expected to provide substantially
the same information as that anticipated from
in-pile pumped loops. The Program maintains a
continuing high interest in the Solid State Division's
in-pile loop pump-development program, and plans
to support a forced-convection in-pile loop test
as soon as component equipment showing promise
of successful in-pile operation is available.

The planning of out-of-pile basic compatibility
tests is virtually complete. Tests will proceed
in two steps: first, the use of thermal convection
loops (costing from $800 to $1000 per test) for
screening; second, the wuse of pumped loops
(costing from $11,000 to $12,000 per test plus
an additional $12,000 if additional test stands
are needed). Thermal convection tests will not
be made on liquid metals; ANP experience has
shown them to be too insensitive to operate
satisfactorily.

Initially, thermal convection loops will be used
for screening each of the salts listed above
against Inconel and INOR (the latter, when tubing

available). These tests will be run
between maximum and minimum tube wall tem-
peratures for 1000 hr at design temperatures,
1250°F maximum wall temperature and a tem-
perature difference of 170°F. The purpose of
these 1000-hr tests is to provide for the early
elimination of those salt-metal combinations which
demonstrate a hopeless basic incompatibility and,
in the case of Inconel, to provide correlation with
existing higher temperature ANP data.

The initial 1000-hr screening tests will be
backed up by thermal convection tests at 1250
and 1350°F and a temperature difference of 170°F.
These tests will be run for a year or more unless
leaks appear or abnormal temperature differences
indicate excessive corrosion or flow restriction.
The purpose of these tests is to increase con-
fidence in long-term compatibility at temperatures
bracketing expected design temperatures.

becomes

Because of their greater cost, pumped loops
will be more limited in number. They will be
operated for one year or longer unless there are
indications of basic incompatibility, or unless
loop failures occur during operation. Pumped
loops will be run at the design hot Reynolds
Number 10,000 and somewhat above expected
reactor design wall temperatures, 1300°F maximum
and a temperature difference of 200°F for salts;
1100°F maximum and a temperature difference of
450°F for liquid metals. In order to achieve a
tolerable compromise between the cost and elapsed
time for the completion of compatibility testing,
pumped loops on salts showing the most promise
in 1000-hr thermal convection tests will be started
before the longer thermal convection loop test
results are available. Because of budget limita-
liquid-metals tests
priority; possibly liquid-metal—
container-metal combinations will be started during
the calendar year 1957.

More basic research on mass-transfer corrosion

tions, have been assigned

lesser three

in molten salt systems has been vigorously pur-
sued during past years in the ANP program.
Presently, there are strong indications that this
work is about to bear fruit of a very practical
nature. There is rapidly mounting evidence of the
correctness of the hypothesis that mass-transfer
corrosion in circulating loops with temperature
gradients is diffusion-controlled. Pending work
on the detailed theoretical andlysis of the course
of the corrosion process under this hypothesis
and further evaluation of chemical affinities of
the chemical constituents of salts and metals,
the way is open for controlling the location of
corrosion to harmless regions by a combination
of mechanical design and corrosion-products con-
centration control. The latter can be achieved
either by the inclusion of suitable sacrificial
elements in the circulating-fluid system or by
chemical additives directly. As an alternative,

PERIOD ENDING SEPTEMBER 1, 1957

there is the possibility of desigining salt mixtures
whose equilibrium constants for reaction with
suitable metals will have @ small temperature
dependence. The Molten Salt Reactor Program has
a deep interest in this basic work now in progress,
and it seems likely that within the work period
which is presently being planned, anticorrosion
design tests aimed at exploiting this research
will be initiated.

 

DESIGN OF IN-PILE NATURAL CONYECTION LOOPS
F. E. Romie

For corrosion tests of container—metal-salt com-
binations in which the corrosion rate is governed
by diffusion processes within the metal, the
velocity of the salt flowing over the metal surface
has practically no detectable effect on the long-

term corrosion rate. Thus a natural convection

loop, despite its characteristically low flow
velocities, is a suitable test system for such
materials. The natural convection loop is particu-

tarly attractive for in-pile corrosion tests, espe-
cially in consideration of present experience with
small-scale pumped in-pile loops, because of its
apparent potential for long, trouble-free operation.

Design of a loop for use in the LITR requires
that the loop fit in a fuel element space of about
3-in.-square cross section. The legs of the loop
are thus close together, separated by not more
than 1.5 in. center line to center line. The height
of the loop is determined primarily by the vertical
flux distribution in the reactor. For any specified
average power generation in the salt, the maximum
salt temperature will be minimized if the ratio
of the peak to the average power generation is
small.  This condition can be realized by main-
taining the flux incident on the loop uniform,
which means for the LITR that the loop should
not exceed a height of about 1 ft.

The difference in mean salt temperature between
the two legs is maintained by unequal cooling.
The air, which is used as the coolant, flows
through an annular passage along the loop tube.
A schematic diagram of the convection loop is
shown in Fig. 5.

If uniform heat generation and laminar flow of
the salt are postulated, onalysis of this loop
gives the Reynolds modulus of the salt in terms
of four independent variables:

where
D = internal diameter of loop tube,
L = height of loop,
W = power generation per unit volume,
AT = difference between maximum and minimum

surface temperatures.

Simulation of temperature conditions in the RDR
requires a AT of about 150°F with o maximum
wall temperature of 1250°F.

Inspection of the above power relation indicates
that ottainment of appreciable Reynolds moduli
is primarily dependent on the diameter of the
loop tube and insensitive to the powergeneration,
W. The average power generation for the RDR,
50 w/cc, can be realized in the LITR with
1 mole % UF, in mixture 74 (Li-Be-F) and appears
suitable for o convection loop. The source
strength having been fixed at 50 w/cc, the di-
ameter of the loop is limited only by space
considerations, since the temperature difference
between the maximum salt temperature and the
tube wall is not excessive for any reasonable
diameter. A diameter of l/2 in. appears appropriate,
although fabrication of the bends at the top and
bottom of the loop may require a smaller diameter.
MOLTEN SALT REACTOR PROGRAM PROGRESS REPORT

Specification and performance of a possible
ORNL LR - WG 23453 convection loop consistent with the foregoing
considerations are given in Table 2.

Table 2. Specification and Performance of an In-Pile

Convection Loop

 

   
   

Mixture 74 LiF-Ber-UF4
(69.5-29.5-1 mole %)
ANNULAR AIR Tube diameter (internal) 0.50 in.
PASSAGE
Height of loop 12 in.
Coolant air rate 18 cfm (STP)
Reynolds modulus of salt Order of 100
Annular gap for air flow 0.04 in.
Source strength 5 w/cc
Difference between center 170°F

line and tube surface tem-

perature in no-flow condi-

 

tion

 

 

Difference between maximum  150°F
and minimum surface tem-

peratures

Maximum wall temperature 1250°F

 

=
s
=
- \
T R
I

It is anticipated that provision for off-gases
will be provided by a charcoal trap connected
to the top of the loop. During reactor shut-down
the salt will be kept in the molten state by electric
heating, which, with proper design, can be ar-
ranged to allow continuved circulation of the loop
salt,

 

 

Control of the temperature level of the sait is
afforded by regulation of the air flow rate. Partial
. control of the difference between maximum and
Al minimum surface temperatures can be realized
¥ by vertical positioning of the loop in the reactor

W core. A lesser degree of surface temperature
control can be obtained by adjustment of the

/.I//
"//// electrical input to the loop heaters.

MII 1 It is not possible, with present knowledge, to

 

 

 

 

 

 

 

 

 

 

 

_

 

predict the nature of the flow in the loop. The

tluid shear stress at the wall is consequently

not predictable, and the value of the Reynolds

Fig. 5. Schematic Drawing of In-Pile Thermal Con-  modulus listed in Table 2 must be regarded as
vection Loop. highly approximate.
PERIOD ENDING SEPTEMBER 1, 1957

GAMMA HEATING OF CORE SHELL

L. A. Mann

HEAT GENERATION

Since the Reference Design Reactor chosen for
study includes a container-metal wall surrounding
the reactor core and separating the fuel salt from
the blanket salt, the integrity of this shell against
failure from thermal stress is important. The
temperature level and pattern in the core shell
wall are determined by the fuel and salt tem-
peratures and the gamma energy absorbed by the
wall. The gamma absorption was calculated for
a slightly idealized model approximating expected
conditions in order that stress levels could be
calculated and compared with the strength of the
container material. The idealization consisted
only of assuming spherical geometry, whereas
the core shell is expected to only roughly approxi-
mate a sphere.

For the model, a core vessel 6 ft in diameter
was assumed, having a nickel-based wall Y in.
thick, heated by gammas from the core, blanket,
and the wall itself, The distribution of gamma
sources from fission, beta decay, radiative cap-
ture, and inelastic scattering was computed by
means of a multigroup neutron diffusion calcu-
lation.! The gamma heating in the vessel wall
was then estimated by means of the straight-
ahead-scattering, integral-spectrum method.? One
numerical and two semianalytical procedures were
used and compared with regard to convenience
and accuracy. A spot check in which Goldstein’s
buildup factors® were used was also made. It
was estimated that at a power level of 600 Mw
(187 w/cc, average), core gammas would generate
13.7 w/cc, blanket gammas would generate
0.032 w/cc, and gammas resulting from neutron
capture in the vessel wall would generate, on
the average, 3.78 w/cc; the total is 17.5 w/cc.
Details of the calculations are given elsewhere. 24

 

L. G. Alexander et al., Operating Instructions for the
UNIVAC Program Ocusol-A, A Modification of the Eye«
wash Program, ORNL CF-57-6-4 (June 5, 1957).

2L, G. Alexander and L. A. Mann, Gamma Heating in
the Core Vessel of the Molten Fluoride Reference De-
sign Reactor, ORNL CF Memorandum, in preparation.

3H. Goldstein and J. E. Wilkins, Jr., Calculations of
the Penetrations of Gamma Rays. Final Report, NYO-
3075 (June 30, 1954).

L. G. Alexander

TEMPERATURE PATTERN

lf temperatures on the inside and outside wall
surfaces are assumed to be equal, the temperature
difference between the center and the surface
of the wall is calculated by

where AT is temperature difference, w is the
heat generation density {(power/volume), ¢ is the
half-thickness of the wall, and & is the thermal
conductivity of the wall. The value of AT was
calculated to be 16.3°F.

THERMAL STRESS

An approximation of the short-time thermal stress
(before relief by creep) is approximated by

aE
o= ———AT ,
1 = v

where o is the maximum thermal stress component
caused by AT, AT is the difference between bulk
wall temperature and temperature at the surface,
a is the volumetric thermal coefficient of ex-
pansion, E is the modulus of elasticity, and v is
Poisson’s ratio.
given elsewhere.4

The stress, 0, was calculated to be approxi-
mately 3170 psi. [Note that o will be higher if
the two surfaces are not at the same temperature.
If the outside wall surface is insulated (not
cooled), AT will be greater. For the same w, ¢,
and k, AT will be 65.3°F, and the thermal stress
will be 12,700 psi.] In order to provide a basis
for judgment of these stresses, it can be stated
that the yield strength of Inconel,® at 1200°F,
is approximately 22,000 psi and that the creep
strength (0.000007 per hour) is shown as 2000 psi.
No attempt was made to evaluate the lowering of
stress by creep.

Details of the calculation are

 

4. A. Mann ond L. G, Alexander, Core Shell Heating
in @ Two-Region Homogeneous Circulating Fuel, Circu-
lating Blanket Reactor, ORNL CF-57-7-61 (July 18,
1957},

R. M. Wilson, Jr., and W. F. Burchfield, *Nickel and
High-Nickel Alloys for Pressure Vessels,'' Welding Re-
search Council Bull. Ser. No. 24 {Jan. 1956).
MOLTEN SALT REACTOR PROGRAM PROGRESS REPORT

MERCURY AS A SECONDARY HEAT-TRANSFER FLUID
B. W. Kinyon

Mercury, of all the heat-transfer liquids proposed,
appears to offer the greatest compatibility with
both the molten fluorides and water (or steam).
In the event of a leak, either a small crack or
a ruptured tube, there would be no explosion or
exothermic chemical reaction and no violent cor-
rosion reaction. Shutdown and repair or replace-
ment of the faulty equipment would be routine,
with precautions against the toxicity of mercury
vapor as the most serious problem.

Mercury as a heat transfer medium may be used
in three ways:

1. as a static medium which forms a thin, stagnant
fayer
wall heat exchanger,

2. as a forced-flow medium which is pumped be-
tween two exchangers, one serving as a heat
source, the other as a heat sink,

3. as a boiling heat-transfer medium in which
mercury is vaporized in a heat source and
condensed in a heat sink.

One of the desired features of a central station

between concentric tubes of a double

nuclear power plant would be to have the steam
and water system entirely outside the radiation
field in order to allow direct inspection, adjust-
ment, and maintenance at all times. This cannot
be achieved with stagnant mercury in a double-wall
exchanger if a primary coolant which emits gamma
radiation is used. Furthermore, a double-wall
exchanger, which is essentially two units built
info one, costs more than two separate units
and is subject to the probability that both wunits
will have to be scrapped if one unit becomes
defective.

The second arrangement, in which the mercury
is pumped between two exchangers, meets the
requirement of isolating radiation from the water
system, but pumping requirements are excessive
and the mercury investment is prohibitive.

A boiling mercury heat-transfer system allows
the steam system to be completely accessible,
and can be arranged so that natural circulation
eliminates the need for any mercury pumps. As
1 Ib of mercury vaporized at 1000°F and condensed
at 600°F transfers as much heat as 9 |b of liquid
mercury transfers between the two temperatures
and as vapor occupies much of the volume, the
mercury inventory is much reduced.

10

A temperature-enthalpy diagram of the fuel,
primary coolant, and water-steam for a possible
set of temperature conditions is shown in Fig. 6.
Initial water conditions are assumed to be heated
feed water mixed with recirculated boiler water.
Mercury vapor would be used to transfer heat from
the primary coolant to the water or steam. Figure 7
illustrates the mercury-condensation conditions,
with three temperature-pressure conditions assumed
for the superheater in order to minimize thermal
stresses. Mercury-condensation and water-boiling
conditions match very well, as may be seen from
this diagram. The temperature of the mercury
vaporization does not have as good a match with
the temperature of the primary coolant, since the
heat flux would be high at the hot end and low
at the cold end of the exchanger. However, the
high density of mercury makes it possible to
arrange a J-shaped exchanger to give a more
uniform flux throughout.

As a temperature difference of 25°F between
salt and mercury is sufficient for boiling the
mercury, it would be desirable to have beiling
from 1075 to 1000°F while the salt is giving up

UNCLASSIFIED
ORNL-LR-DWG 23454

 

 

 

 

 

1200
FUEL
1100
PRIMARY COOLANT

1000 | — ]
&
ol

200 —
5
'_
<L
o
u \-
g 8200 —
: 5

&

700 —— 7 g

600 — WATER —

500

ENTHALPY

Fig. 6. Temperature-Enthalpy Diagram for Fuel, Pri-

mary Coolant, and Water.
heat from 1100 to 1025°F. As 1075°F corresponds
to 275 psia, or 95 psi more than 180 psia, which
corresponds to 1000°F, an 18-ft head of mercury
would be sufficient to increase the boiling point
by this amount. An arrangement of a heat ex-
changer is shown in Fig. 8, and a temperature-
enthalpy diagram is shown in Fig. 9. It is assumed
that the mercury will recirculate and mix with the
returning condensate.

could be used to
control the condensation temperature of mercury
at desired levels, or the mercury could be put
through a turbine, developing power and the same
time reducing the pressure. This, incidentally,
increases the thermal efficiency of the system
over that attainable with steam alone at the same
temperature, even with reheat. The comparison
with RDR, based on 600 Mw of heat from the core
plus 10 Mw of heat from the blanket, is shown
in Table 3.

Mercury turbines have been in operation since
1922, the largest producing 20 Mw of electricity.
For use in a nuclear power plant, as outlined,
larger size turbines, with bleed-off at the desired
intermediate pressures, would have to be de-
veloped. This probably would not present any
great difficulties.

The metallurgical aspect is more uncertain.
Mercury boilers in Newark, New Jersey, and

Pressure reducing valves

UNCLASSIFIED
ORNL~-LR-DWG 23455

 

 

 

 

 

 

 

1200

100 —— —_—

1000 —— —] 180
¥ ©
~ w
W 900 f— a
2 "
< >
o wy
ul "
& 800 L &
s
ul a.
= >

@
3

700 }—— 2
w
=

Hg
600 —
/ WATER
500
ENTHALPY
Fig. 7. Temperature-Enthalpy Diagram for Mercury-

Water Heat Exchange.

PERIOD ENDING SEPTEMBER 1, 1957

UNCLASSIFIED
ORNL-LR-DWG 23456

Hg VAPCOR
1000 °F

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

| 7 e e
1000 °F Hg CONDENSATE
Ry AT A X 600 °F
COOLANT SALT
1025 °F J
!
950 °F
______
] COOLANT SALT
1025°F
COOLANT SALT
100 °F
Fig. 8. Mercury Boiler.
UNCLASSIFIED
ORNL-LR-DWG 23457
1200 w
1100 L OLANT —
E PRIMARY €2 \
—_— "/
W
< Hg
= 1000 "
P
w
a
=
uJ
—
200 (__, —
800
ENTHALPY
Fig. 9. Temperature-Enthalpy Diagram for Mercury

Boiler with Split Flow of Primary Coolant,

1
MOLTEN SALT REACTOR PROGRAM PROGRESS REPORT

Table 3. A Comparison of Mercury Topping Cycle with Sodium Heat Transfer System

 

Mercury Topping Reference Design

 

 

 

Cycle Reactor

Mercury power {Mw) 106
Steam power (Mw) 198 240
Total (Mw) 304 240
Turbine efficiency (%) 49.8 42.3
Plant efficiency (%) 46.8 39.3
Change in costs from RDR (mill s/kwhr)

Turbine investment cost +0.35

Reactor complex cost —-0.45

Fuel cost -0.40
Net change -0.50

Pittsfield, Massachusetts, were troubled with  furnace contributed to the difficulty and that the

periodic plugging of certain boiler tubes. The
Hartford, Connecticut, plant appears to have had
much less trouble. One explanation offered for
the plugging is
mercury had deposited downstream from an orifice
in a boiler tube, causing further constriction and
eventua! plugging. The orifice was caused in
joining tubes to get the length desired, the joint
being either a fusion butt weld or an arc weld
with a backup ring. Weld flash or the backup ring
caused the flow restriction. The treatment of
mercury to ensure good wetting and uniform heat
transfer involves the addition of a few ppm of
magnesium, to remove oxide, and the addition
of titanium, to enhance wetting. Since particular
tubes were reported to be plugging, it is possible
that the high temperatures in the fossil fuel

that magnesium oxide in the

12

uniform heat of the molten fluoride would not
induce precipitation of the magnesium oxide.
Further study will be made.

Another metallurgical problem in the use of
mercury is that nickel alloys, which are necessary
for molten salts, are readily attacked by mercury,
This makes duplex tubing necessary for the salt-
to-mercury heat exchanger. Fabrication of duplex
tubing and its use in heat exchangers is common
practice, but each new combination requires devel-
mental work.

As the compatibility of mercury with both water
and the fused salts is so much greater than it
is for sodium and as the economics appears to
be satisfactory, further study of the use of
mercury will be made. Both two-phase heat
transfer and the topping cycle will be considered.
PERIOD ENDING SEPTEMBER 1, 1957

NUCLEAR CALCULATIONS

THE UNIVAC PROGRAM OCUSOL

Description of Program
L. G. Alexander

The Eyewash program,' a 30-group, 9-region re-
actor code written by J. H. Alexander and N. D.
Given, was selected for the anclysis of the nuclear
characteristics of the reactors of interest to the
Molten Salt Reactor group.? The program has
been extended and modified as follows. The
lethargy group widths were revised in order to
better treat fused sclt reactors containing large
amounts of thorium, and a lethargy for the thermal
group corresponding to 1150°F was selected. A
short program for computing and editing, via the
typewriter, the absorptions
a maximum of seven different

supervisory control
of neutrons in
nuclides was written by R. Van Norton of the
New York University Institute of Mathemetical
Sciences and incorporated info the program, which
was renamed Ocusol. This feature facilitates the
computation of breeding ratios and the preparation
of detailed neutron balances. These changes,
together with other operating instructions, are
described elsewhere.?

Ocusol Cross Sections
J. T. Roberts

The cross sections used in Ocuscl calculations
for molten salt reactors have been recorded and
discussed in an ORNL CF memorandum.? Since
that memorandum was issued, beryllium, carbon,
and sodium have been added to the cross-section
tape {July 8, 1957). These cross sections, and
other additional ones (e.g., Pu), will be recorded
and discussed in a supplementary report to be
issued later.

 

1. H Alexander and N. D. Given, A Machine Multi-
rozlifr Calculation. The Eyewash Program for UNIVAC,
ORNL-1925 (Sept, 15, 1955).

2H. G. MacPherson et al., A Preliminary Study of
Molten Salt Power Reactors, ORNL CF.57-4-27 (April 29,
1957},

3L. G. Alexander et al., Operating Instructions for the
UNIVAC Program Ocusol-A, A Modification of the Eye-
wash Program, ORNL CF-57-6-4 (June 5, 1957).

4). T. Roberts and L. G. Alexander, Cross Sections
for Ocusol-A Program, ORNL CF-57-6-5 (June 11, 1957).

The main differences between the Ocusel cross
sections and those previously used in the similar
Eyewash and Murine calculations are as follows:

different sets of thorium
sections calculated for five different
thorium concentrations (0, 1, 4, 10, and 25
mole % in mixtures of 69 LiF=31 E’Jel:2 and
75 LiF=25 ThF4), with corrections for reso-
nance self-shielding and Doppler broadening
at 1150°F,

2. more conservative values of a for U233 and
U235
i

1. provision of five

Cross

3. a set of average fission-product cross sections,
less the lowest resonances of Xe '35 and Sm 149,
based on the estimates of Greebler and Hurwitz.?

These differences are conservative with respect
to most other comparable reactor caleulations, ®
but the fission-product absorption cross sections
may net be conservative enough. Recent experi-
mental measurements’ give values of o, at 25 kev
which are two to three times as high as those
estimated by Hurwitz and Greebler. The implica-
tions of this in the fuel-cycle economics of molten
salt reactors are discussed elsewhere in this

progress report.

PARAMETRIC STUDIES OF REFERENCE
DESIGN TYPE OF REACTORS

Clean Reactor Calculations
J. T. Roberts

A new series of reference design type of re-
actors? were calculated with core diameter and
thorium content of core salt as variables. The
resuits are summarized in Table 4. The calculo-
tions were made for spherical, homogeneous, three-
region reactors, at 1150°F, as follows:

 

5p, Greebler and H. Hurwitz, Jr., Evaluation of Nu-
clear Absorption Cross Sections in the Intermediate and
High Energy Range as Applied to FissioneProduct
Poisoning, KAPL-1440 (Nov. 9, 1955).

6$ee, for example, J. Chernick, Nuclear Sci. and Eng.
1, 135 {1956).

R, L. Macklin, N. H. Lazar, and W, S. L.yon, Phys.
Semiann. Prog. Rep. March 10, 1957, ORNL-2302, p 27.

13
MOLTEN SALT REACTOR PROGRAM PROGRESS REPORT

Table 4. Summary of UNIVAC-Ocusol Calculations for Two-Region, Spherical, Homegeneous
Molten Salt Reactors: 1150°F, **Clean’’

Core: 69% LiF-31% BeF2 plus UF4 and ThF4 as shown, diameter as shown
Shell: ) in. of INOR-8
Blanket: 75% LiF-25% ThF4, 2 ft thick

 

Code number

Core
Diameter (ft)
Mole % Th232
Mole % U235

Neutrons thermalized (%)
Fissions at thermal energies (%)

If v = 2.47, table shows volues of

Q=

Neutron balance

Core
y235

u238
TH232

Li+ Be
g
Shell
Blanket
Li+F
Th232
Leakage

Total

Critical inventory (kg) for external
volume (ft3) shown
0
113
226

339

Conversion ratio

Internal
U238

Th232

External
Th232

Total

 

 

53 54 55 64 g3 61 38 84 62

& 8 10 6 8 10 6 8 10

0 0 0 0.28 0.28 0.28 1 1 1

0.11 0.047 0.033 0.30 0.096 0.062 0.63 0.34 0.21
25 45 55 4.8 22 35 0.41 4.6 7.2

38 59 67 8.3 34 51 0.49 8.2 12
.93 2.01 2.03 1.79 .92 1.98 1.74 179 1.82
0.28 0.23 0.22 0.38 0.29 0.25 0.42 0.38 0.36
0.5185 0.4984 0.4919 0.5590 0.5218 0.5052 0.5744 0.5594 0.5509
0.0322 0.0072 0.0059 0.0224 0.0130 0.0088 0.0261 0.0224 0.0202
0.0000 0.0000 0.0000 0.0902 0.1318 0.1522 0.1428 0.2051 0.2432
0.0717 0.1587 0.2254 0.0235 0.0684 0,1168 0.0104 0.0228 0.031
0.0347 0.0470 0.0596 0.0285 0.0362 0.0422 0.0264 0.0280 0.0312
0.0719 0.0741 0.0607 0.0381 0.0443 0.0408 0.0240 0.0226 0.0188
0.0085 0.0072 0.0053 0.0058 0.0023 0.0041 0.0043 0.0034 0.0026
0.2752 0.2030 0.1490 0.2253 0.1747 0.1268 0.1853 0.1321 0.0988
0.0073 0.0044 0.0022 0.0072 0.0075 0.0037 0.0063 0.0042 0.0032
1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000
47 49 67 132 100 126 279 358 418

94 70 82 264 141 153 558 509 508
142 90 26 396 183 181 837 660 599
189 1 m 528 225 208 1115 811 689
0.023 0.014 0.0612 0.040 0.025 0.018 0.046 0.040 0.037
0.000 0.000 0.000 0.161 0,252 0.301 0.249 0.367 0.441
0.531 0.408 0.321 0.403 0,335 0.251 0.322 0.236 0.180
0.554 0.422 0.313 0.604 0.612 0.569 0.617 0.643 0.658

 

14
Core: 69 LiF-31] BeF2 plus 0.035 to 0.68
mole % UF, (93% U?%%) plus 0, 0.284,
and 1 mole % ThF, for diameters of 6,
8, and 10 ft

Shell: The equivalent of % in. of INOR-8 spread

over one spacial increment (the over-all
radius of the reactor is divided into 60
spacial increments)

Blanket: 75 LiF-25 ThF,, 2 ft thick.

The '*most thermal’' of the reactors was No. 55
(10-ft core diameter, 0% Th in core), with 55% of
the neutrons being thermalized and 67% of the
fissions occurring in the thermal group. The
‘"least thermal’’ of the reactors was No. 38
(-ft core diameter, 1% Th in core), with only
0.41% of the neutrons thermalized and 0.49%
of the fissions at thermal energies. These two
cases also represent the extremes in this study
for parasitic captures by U233, with a values
of 0.22 and 0.42, respectively. These values
correspond to 7 values of 2,03 and 1.74, respec-
tively, if the ‘‘world’’ value of 2.47 is used for

v (for U235) in the equation:
v
T IT e

The neutron balances given in Table 4 show that
the most important losses of neutrons are:

P arasitic captures in U235 10-17%
P arasitic captures in core salt 4-28%
Parasitic captures in shell 2-7%

Hardening the neutron spectrum by using smaller
core size or higher thorium content reduces the
core salt and shell losses but increases the
non-fission captures in U233,

The critical mass of U?33 in the reactors studied
varied from 47 to 418 kg. These are shown in
Table 4 as critical inventory for 0 ft° external
volume, Critical inventories are also shown for
113, 226, and 339 f1° of external volume. Roughly,
these external volumes correspond to 200, 400,
and 600 Mw forced-circulation reactors, or to
60, 130, and 190 Mw natural-circulation reactors.
The data for 0 £+ and 339 3 external volume are
plotted in Fig. 10. Although the ratio of internal
to external conversion varies considerably, as
shown in Table 4, the total conversion ratio varies
only between 0.55 and 0.65, except for the 8- and

PERIOD ENDING SEPTEMBER 1, 1957

10-ft reactors with no thorium in the core. Figure 11
shows the conversion ratio and (p — 1) for the
various reactors studied. Adding thorium to the
core increases the conversion ratio, but the in-
crease is much less than might be expected since
(n ~ 1) decreases significantly at the same time.

UNCLASSIFIED
ORNL—LR—DWG 23458

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2000
1000 \L
\_
Ty,
~
L s
4
500 ‘
- AN
_ \
2 \
o
o
« \
o N\
= \\‘
g \\\ 4
E 200
i
>
= \
3 \
3 \
’_—
: ™ ‘
*._____ 28
100
j'
. /
GF——’
® = 0% THORIUM .
T
A = 0.28% THORIUM
B = 1% THORIUM
— CRITICAL MASS (NO EXTERNAL VOLUME)
——— CRITICAL INVENTORY FOR 339—“3 EXTERNAL
INVENTORY
20
6 8 10

DIAMETER (ft)

Fig. 10. Critical Inventories for Clean Reactors.

15
MOLTEN SALT REACTOR PROGRAM PROGRESS REPORT

The trends of the critical mass curves shown
in Fig. 10 do not appear to be mutually consistent.
In paorticular, it was expected that the minima
would shift to the left with increasing thorium

UNCLASSIFiED
ORNL—LR—DWG 23459

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12 I ‘
1 HORIUM
1.4 IN CORE SALT
(mole n/o)
|
2 4o Pl
g C.28
2
2 L
w)
W (
=
Q
o
Z(, 0.8 /‘ !
< 0.
'_
O
<
II:. 'JJ
g 0.7 !
5 )
o
2
< 06 \
= 0.28
1%
[43)
&
Y 05 > C.R, —
z ;
O
Q
-
<
O
= 0.4
[T ]
a8
O
wl
I
- 0 J
= 03 =
oD
=
>
<
=
‘I: 0.2
[
0.4
0

 

 

 

 

 

 

 

6 8 10
DIAMETER (ft}

Fig. 11. Actual Conversion Ratio (C.R.) and Maximum
Theoretical Conversion Ratio (7 — 1) as a Function of
Core Diameter and Mole % Thorium in Core Salit,

16

concentration. The curve for 0.28% thorium did
not conform to this expectation. The redasons for
this anomalous behavior are being investigated;
it is thought that the behavior may be related to
shifts in the spectral distributions of the flux in
these reactors.

VYariation of Nuclear Characteristics with Period
of Operation

L. G. Alexander

The characteristics listed in the previous section
are those of initial or ‘‘clean’ reactor states.
The accumulation of fission products and of
uranium isotopes will modify these characteristics
markedly. In those reactors containing no thorium
in the core, the accumulation of fission products
will tend not only to increase the critical mass
and inventory of U233 but will also, by decreasing
the leakage, cause a decrease in the breeding
ratio. As the concentrations of absorbers in the
core, including U233, increase, the neutron-flux
spectral distribution will shift toward higher
energies, While this tends to reduce the relative
number of absorptions in fuel carrier and other
parasitic materials, it tends to increase the rela-
tive number of radiative captures in U235, and
thus it impairs the breeding ratio. On the other
hand, addition to the core of U233 which is formed
in the blanket tends to compensate for the buildup
of parasitic materials. Also, with U233 in the
core, increasing hardness of the neutron spectrum
has much less effect on the breeding ratio because
the relative number of radiative captures at epi-
thermal energies are much less in U233 than in
U235.

In reactors having thorium also in the core,
the accumulation of U%33 tends to be more rapid,
and hence the U235 replacement rate is smaller
and the conversion ratio does not fall so rapidly.
With a sufficiently large initial inventory of U?35,
it is conceivable that the total reactivity might
tend to increase temporarily even without the
addition of any U235, and that for some limited
period it might be possible to add thorium and
perhaps obtain an in the conversion
ratio. |t might then be possible to stabilize the
system in this favorable state by proper selection
of processing methods and rates.

In order to investigate these various possibilities,
a zero-dimensional, 31-group, time-dependent pro-
gram (Sorghum), designed to use the output data

increase
from Ocusol (or Cornpone), is being prepared for
the Oracle. The essence of the theoretical treat-
ment follows.

The group diffusion equations used in Eyewash®
may be put in the form

() €263 + vz [ 30 67 ,u) du
0

i

(250 + £31) 60 -

 

Viig) Ui,

3%,
where,
u = lethargy,
7 = vector coordinate,
@#(,u) = neutron flux at lethargy u, space
point 7,
qbi(?) = flux at upper bound of ith lethargy
. group, space point 7,
') U* = flux in ith lethargy group

= 1:0 t,‘b(u,?) du,

v_ = fission yield of neutrons required to
make the system critical,
fraction of fission neutrons born in
ith lethargy group,
and where the other symbols have the customary
connotation,
We now set

Z' =

—, 1 . .
(2) ¢'F) = -2-[<75'(?) + &1 .

By introducing Eq. 2 into Eq. 1 and rearranging,
the following equations are obtained:

 

- (S0 + 2£58) 8'0) + 7 VIEIR) UF +

I
33
P{emi=1 4 23 )pin10) 4

(3) . 00
+ v 7 fo 2 (upfwldu =0,

¢ = 2;'51' = pi=]

8J. H. Alexander and N. D. Given, A Machine Multi-
roup Calculation. The Eyewash Program for UNIVAC,
RNL-1925, p 5, Eq. 11.2 (Sept. 15, 1955),

 

PERIOD ENDING SEPTEMBER 1, 1957

The flux level is, of course, arbitrary. A power
level is chosen such that there will be one neutron
born per second in the core; that is,

@ f [ duay =1,

where V. denotes the volume of the core. It is

also noted at this time that

] - : :
o [ —vEe v -1,

1

Vc 32!?‘

where L! is the leakage from the core in the ith
lethargy group. If Eq. 3 is now integrated over
the core and Eqs. 4 and 5 are applied, Eq. 6 is
obtained:

L 26E\
(6) ~ (2; + ¢:: - L 4

 

U!

 

U!

52:“' + 52: . _
+ qbi_"‘ +2Z' = 0,

2;;:- - 95:_‘.-] ’
[ #awiav,
VC

] B
A= 0w
il i

©-
n .,
i

f;ccfi"(‘r’)ui av .

The major assumption is now introduced that
L* is proportional to ¢%:

” L

where the C’ are readily obtained from the output
of the Ocusol calculation,

(8) Cct = .

 

Here L. denotes the leakage in the ith lethargy
group of the Ocusol calculation, % o is the ab-
sorption cross section, and A} is the absorptions
in the core in the ith lethargy group. All these
are available in the Ocusol output.

17
MOLTEN SALT REACTOR PROGRAM PROGRESS REPORT

Substituting Eq. 7 into Eq. 6 and rearranging
results in

i i
Z' + A'¢

(9) Zi + B?

a
ol = 21— gi-l,
where,
ey
Al -

2 £3]
B =

 

+Ci.
Ut

Equation 9 constitutes a set that may be solved
consectively for the flux using the boundary condi-
tion that qSZ, is zero for i = 0.

The quantity X! is arbitrary and depends strictly
on the composition of the core:

. =16

i I5d et

Za E NG
=1

where N/ denotes the atomic density of the jth
element, and the 57'* are the corresponding group
absorption cross sections. Thus, specifying the
core composition determines the spectral dis-
tribution of the flux. It remains to satisfy the
criticality condition, which is achieved when
Eq. 4 is satisfied by a value of v equal to the
v of the fissionable material in the core. If more
than one fissionable species is present, a weighted
mean value must be used. This is readily de-
termined once the spectrum of the flux is known.

An iterative procedure is used to solve Eq. 9.
This is combined with the differential equations
expressing the rate of change of atomic densities
of the various species with burnup in order to
formulate a description of the time-dependent be-
havior of the system. The initial composition,
taken from a basic Ocusol calculation, yields,
of course, a critical system. This composition
is operated upon and changed by the burnup
equations, and the system departs slightly from
criticality. Equation 9 is then used to bring the
system back to the critical condition, either by
adding U233 or by removing thorium. It is assumed

18

that A' and B are dependent mainly on the con-
centrations of lithium, beryllium, and fluorine,
which do not vary, and are insensitive to changes
in the concentrations of other materials. This
assumption appears to be reasonable because
A' and B? are mainly dependent on the scattering
power of the medium, expressed as £2, to which
lithium, beryllium, and fluorine make the major
contribution.

The burnup equations are then applied again
and the procedure is repeated. A log of the con-
centrations of the 12 variable species (Pa, Th,
U233, U235, U234, Np237, U238’ Pu, and the
fission products in the core, and Pa and U233
in the blanket) together with required U?3% qod-
ditions is accumulated and edited at the con-
clusion of the computation.

The differential equations expressing burnup
changes provide for the independent specification
of processing rates for the protactinium, uranium,
and thorium, and fission fragments in the core,
and for the U233 in the blanket.

The program has been written and a program
tape typed. The constants tape (cross sections,
etc.) is being prepared. The present edit is on
paper tape. After debugging, a curve-plotter edit
will be written for the compositions, and a mag-
netic tape edit for the flux spectrum and total
absorptions in the various species integrated
over the period of operation of the reactor will
be added.

When completed, Sorghum will be wused for
studying the non-steady-state behavior of the
reactors described in the previous section.

Steady-State Reactors
J. T. Roberts

A series of steady-state critical inventories and
neutron balances have been given® to illustrate
the effect of even-number uranium isotopes and
tission-product various processing
cycles. These steady-state reactors were not
calculated directly on the UNIVAC, but were
obtained by ‘‘perturbing’’ the reactors actually
calculated. Steady-state cases A through F, with
the modification that the U233 and fission products
were left out of the blanket, were subsequently

poisons for

 

M. 6. MacPherson et al., A Preliminary Study of
Molten Salt Power Reactors, ORNL CF-57-4-27, p 96,
Table XIV (April 29,1957).
computed on the UNIVAC (July 8 to 12, 1957) to
check the validity of the perturbation calculation.
While the results have not been fully analyzed,
the values of 1_ expected (assuming perturbation
calculation correct) and actually found are sum-
marized below:

v v

Case c ¢
Expected Found
A 2,52 2.42
B 2.58 2.46
C 2.57 2,50
D 2.56 2,48
E 2.60 2.49
F 2.62 2,62

PERIOD ENDING SEPTEMBER 1, 1957

A value of vc-found which is less than v_-expected
means that either the conversion ratio was under-
estimated or the critical inventory was overesti-
mated by the perturbation method, and vice versa.
The UNIVAC data have not yet been analyzed to
the point of deciding whether the reason for
v_-found being lower than v_-expected in Cases A
through E is due to a being lower than expected
or to 0 being higher than expected.

 

FUEL-CYCLLE ECONOMICS
J. T. Roberts

The UNIVAC-Ocusol calculations made during
the last quarter did not indicate any significant
errors in the interrelationship between conversion
ratio, fissionable-material inventory, processing
rate, and fuel-cycle costs assumed in the pre-
liminary study previously reported.! The latest
calculations do provide the basis for extending
the fuel-cycle cost study over a wider range of
possible reactor diameters and thorium concentra-
tions in the core.

FISSION-PRODUCT POISONING

The fact that experimental evidence indicates
that the assumed values for the fission-product
absorption cross sections may be lower than the
experimental values by a factor of 2 or 3 has
significant bearing on the fuel-cycle economics.
Since previous calculations did not take any credit
for removal of rare gas fission-product poisons,
this discussion will assume that the effective
absorption cross sections should be just double
those previously used.

 

. A. Mann and L. G, Alexander, Core Shell Heating
in a Two-Region Homogeneous Circulating Fuel, Circus
lating Blanket Reactor, ORNL CF-57-7-61 (July 18,
1957).

Reactor ‘‘D,”” discussed in an earlier report,?
assumed a core processing cycle of 0,75 year,
an inventory of 649 kg of U233.U235 4nd qn
effective breeding ratio of 0.56. The *‘variable”
fuel-cycle costs (in dollars per year) for a 600 Mw
reactor were given as:

U233 and U235 rental $ 440,000
y23s makeup 2,030,000
Core salt makeup 810,000

$3,280,000 (= 2.0 mills/kwhr)

tf the fission products are just twice as bad as
previously assumed, the ‘‘same’ reactor will
require just twice the processing rate, that is,
an additional $810,000 per year (~ 0.5 mill/kwhr).
Actually, increasing the fission-product cross
section will tend to shift the economic ‘‘optimum"’
in the direction of higher critical mass instead
of higher processing rate, with preliminary in-
dications being that the increased cost due to
higher fission-product cross sections should be

about $400,000 rather than $800,000 per year.

 

4, 6. MocPherson et al., A Preliminary Study of
Molten Salt Power Reactors, ORNL CF-57-4-27, p 96,
Table XIV (April 29, 1957).

19
MOLTEN SALT REACTOR PROGRAM PROGRESS REPORT

NATURAL CONVECTION REACTOR
F. E. Romie

One of the problems of a circulating-liquid-fuel
reactor is the provision of reliable, long-lived,
fuel-circulating pumps. This problem can be ob-
viated by going to a reactor in which the fuel is
circulated through the primary exchanger and re-
actor by natural convection. The advantages of
omitting the circulating pump and its attendant
problems of maintenance and
purchased at the price of increased fuel-salt

replacement are

volume in the primary exchanger and in the con-
vection risers and return lines. The consequent
relatively low specific power attainable in the
convection system is probably the major deterrent
to its adoption.

for a reactor system in which the premium placed

There are, however, applications

on reliability and ease of maintenance could make
the convection system attractive. For this reason,
exploratory calculations were conducted to de-
termine the fuel-salt-holdup volume external to
the core of a natural-convection reactor.

For these calculations the thermal output se-
fected was 40 Mw.
consisted of a spherical core and a

The system defined for
analysis
fuel-salt-to-sodium heat exchanger located above
the core. A schematic diagram of the system is
shown in Fig. 12.
culations, the system variables listed in Table 5

For these exploratory cal-

were assigned fixed values.

With these system variables fixed, the additional
specification of the salt velocity in the exchanger
tubes and of the leaving temperature of the salt
determines the heat exchanger design: length
and number of tubes, pressure losses of salt and
sodium, and the salt volume in the exchanger

tubes.

For any given exchanger, the hydrostatic head
for thermal convection is given in terms of the
change in salt temperature and the effective
height of the exchanger above the reactor core.
The friction losses that, in steady flow, will
equal the hydrostatic head are the sum of the
exchanger pressure loss, the riser and downcomer
wall friction, and losses in bends, expansions,
Thus for any height of a given
is one diameter of the

and contractions.

exchanger, there riser

and downcomer pipes for which the hydrostatic

20

head will equal the sum of the flow losses. More-
over, an optimum exchanger height exists at which
the salt volume in the riser and downcomer is a
minimum for the given exchanger. The foregoing
can be summarized by stating that the exchanger
design and the corresponding minimum salt volume
in riser and downcomer can be expressed in terms

UNCLASSIFIED
ORNL-LR-DWG 23460

Na

!

L

 

 

 

 

 

 

 

 

 

 

—

Fig. 12.

Reactor System.

Schemoatic Drawing of Thermal Convection

Table 5. Specified System Variables

 

Thermal rating 40 Mw
Sodium inlet temperature 850°F
Sodium outlet temperature 1025°F
Salt temperature entering exchanger 1250°F
Exchanger tube, ID 0.50 in.
Exchanger tube wall thickness 0.05 in.
Tube spacing in A array 0.825 in.
Core diameter 6 ft
Fixed length of salt pipe, L 25 ft

/

 
of the two independent variables: salt velocity
in the tubes and leaving temperature of the salt.

Table 6 presents partial results of calculations
with Flinak as the fuel-bearing salt. Inspection
of the table reveals that the lowest calculated
total external-holdup volume, 59.3 f3 (1.5 #3/Mw),
was realized at the lowest salt velocity and lowest
salt-leaving temperature considered in the calcu-
lations. The minimum external-holdup volume for
the conditions specified in Table 5 is not re-
presented in Table 6, but the trend of results
indicates that the minimum external holdup occurs
at o salt velocity of less than 2 fps and at a
leaving salt temperature of less than 950°F.

Salt 14 (Flinok) was selected originally on the
basis of its large coefficient of thermal expansion,
low melting temperature (846°F), and low viscosity.
While the physical properties of Flinak madke it
a desirable salt from the heat transfer point of
view, its nuclear properties reduce its usefulness
as a fuel carrier. For this reason, calculations
were carried out with a Li-Be-U fluoride salt
{mixture 75). Since this salt melts at a tempera-
ture of about 20°F above that of Flinak, the sodium

PERIOD ENDING SEPTEMBER 1, 1957

inlet temperature to the exchanger was increased
from 850 to 870°F. On the basis of calculated
results for Flinak, the temperature of the salt
leaving the exchanger was specified as 950°F
and the Reynolds modulus for the salt leaving
the exchanger tubing was fixed at 2500, which
corresponds to the lowest salt velocity for which
heat transfer data were readily available. Under
these conditions the external-holdup volume of
the Li-Be-U fluoride salt was 70 3 (1.75 f3/Mw),
of which 10 ft* was holdup in the exchanger. The
corresponding optimum height of the exchanger
above the core top was 33 ft, and the optimum
pipe diameter was |1 in. Additional calculations
would be required to establish the salt velocity
and salt-leaving temperature corresponding to the
minimum-hoidup volume.

Use of sodium as the coolant in the primary
exchanger is open to criticism on the basis of
its lack of compatibility with the fuel salt.
ever,

How-
substitution of a compatible salt for the
sodium should not produce large changes from
the results presented (in particular, the ft°/Mw
ratios), since the salt is an excellent heat transfer

Table 6. Free Convection System Calculation Results

 

Salt velocity in 2 2

exchanger (fps)

Salt temperature leaving 950 1025
exchanger (°F)

Number of exchanger 792 1056
tubes

Length of exchanger 14.1 7.5
tube (ft)

Salt holdup in 15.3 10.8
exchanger (ff3)

Minimum salt holdup 44 58.2
in piping (ffs)

Total external-holdup 59.3 69.0
volume (ft3)

Optimum diameter of 10 12.3
piping (in.)

Optimum height of 28 23

exchanger above

core top (ft)

2 3 3 3
1100 950 1025 1100
2310 528 704 1054
2.7 15.1 8.3 4.4
5.7 10.8 8.0 6.3
91 66 83 139
96.7 76.8 N 145
16.4 10 11.8 15.3
18.4 48 43 42

 

21
 

medium and, especially, because the volume of
the fuel salt in the exchanger is a small part of
the total external holdup.

Supplementary calculations have demonstrated
that the salt volume external to the reactor is
very closely proportional to the thermal output
of the reactor. Thus, to a good approximation,
the cubic feet of external salt per megawatt is
independent of the thermal output.

Nuclear calculations (see Table 4) for a salt
reactor core of 8-ft diameter give a hot, clean
critical mass of 49 kg (0% Th). For a forced-
convection reactor {the RDR) the external-holdup
volume has been reported as 0.56 ft3/Mw. With
an external-holdup volume of 1.75 #3/Mw used
for a free-convection reactor, the specific powers
of the free and forced-convection burner reactors
are compared in Table 7 for a range of thermol
outputs,

Inspection of Table 7 shows that at low thermal
outputs the free-convection reactor system compares
very favorably with the forced-convection system.

In particular, the exploratory calculations indicate
that for an 8-ft-dia, 40-Mw reactor the free-con-
vection system requires only 15% more fuel in-
ventory than the forced-convection reactor. These
preliminary results indicate that the thermal-
convection reactor merits further analysis in the
thermal output range up to perhaps 100 Mw.

Table 7. Comparison of Specific Powers for Free

and Forced Convection Reactors

 

Specific Power

{(kw/kg)

 

Thermal Output
{Mw) Free

Convection

Forced

Convection

 

40 648 740
100 1230 1690
250 1940 3350
600 2500 5440