ORNL-CF-61-2-46.txt

ORNL S o
MASTER COPY

Distribution Limited to Recipients Indicated

EXTERNAL TRANSMITTAL AUTHORIZED
ORNL
Central Files Number
61-2-46 )

MOLTEN-SALT REACTOR EXPERIMENT

PRELIMINARY HAZARDS REPORT

NOTICE

This document contains information of a preliminary nature
and was prepared primarily for internal use at the Oak Ridge
National Laboratory. It is subject to revision or correction
and therefore does not represent a final report. The information
is not to be abstracted, reprinted or otherwise given public
dissemination without the approval of the ORNL patent branch,
Legal and Information Control Department.

OAK RIDGE NATIONAL LABORATORY
operated by
UNION CARBIDE CORPORATION
for the
U.S. ATOMIC ENERGY COMMISSION

-

Y

Contract No. W-T4O5-eng-26

MOLTEN-SALT REACTOR EXPERIMENT

PRELIMINARY HAZARDS REPORT

S. E} Beall

W. L. Breazeale B. W. Kinyon

February 28, 1961

OAK RIDGE NATIONAL LABORATORY
Oak Ridge, Tennessee
operated by
UNION CARBIDE CORPORATION

for the |
U.S. ATOMIC ENERGY CCMMISSION

External Transmittal
Authorized

ORNL-CF-£1-2-46
7. e e

. ’
1
=
)
[ g
-
. [

SUMMARY

‘The Molten Salt Reactor Experiment (MSRE) is a circulating-fuel,
low-pressure, high-temperature reactor. .The major objectives are the
demonstration of the safety, dependability, and serviceability-of”sueh
a reactdr and the obtalning of additionfil information about graphite

and fission-product gases in the environment of an operating molten-

salt reactor. .

The MSRE fuel is the tetrafluoride of enriched'U235 (UF@) dis-
solved in an LiF—Bng carrier, with ZrF4 and ThF, additlons The coolant
salt is an LiF-Bng mixture without additives. The core consists of

| unclad-graphite pieces held in position By'molybdenum bands. A nickel

molybdenum alloy, INQR;B, especially developed'as.a'container material
for molten fluorides, is used for all container and structural menmbers
in contact with either the fuel salt or the coolant salt. The heat

from the reactor is dissipated to the atmosphere through a radiator in

- the coolant-salt systemn.

The cover gas for ‘both the' fuel- and the eoolan%wsalt systems is
hellum The off-gas system is designed to hold up fission gases until
the activity level permits discharge to the atmosphere. |

The instrumentation and controls are designed to shut down the
reactor safely if excessive reactivity occurs. Periodié'Sampling per-
mits evaluation of fuel stability and corrosion rates. “

To demonstrate the serviceability of the system, provisions are
made for remote and. semidirect maintenance of the equipment in the
reactor cell and other regions of high residual activity. Direct
maintenance will be. performed in other areas, including the radiator
pit. -

The possible accidents consldered are reactivity excursions, fuel
separatien, loss of flow, controcl rod fallure, and several mechaniéa.]T
possibilities for contalnment failure. The maximum credible accident,
rupture of tfie fuel- and coolant-salt systems and subsequent spilling
into the cell of all of. the salt, would not burst the container. Any
escape of fission products from the:container should result in an ex-
posure of less than 25 rem to anyone in the reactor building or in the

surrounding area.

siii-
CONTENTS

Slmam ........0.‘...........9......‘...............‘.....O.......

iii

List Of Ta:bles L B B BB B B B BN B BN BN BE BN BN BE BN N BN BL BN BN L BN BN BN BN BN N B IR B B B BN BN BN B AL AR L BN BB B B AL L J vii

List Of Figuz‘es ......._..:’.............0..-.......0.90.............

l. Introduction .l.0.0l.v.l-'.;.....OQ.......‘.........0...."......

2,

Reactor comlex 0..0...0.....0..0..000..0......0.......‘...0..

2,1
2,2
2.3
2.4
2.5
2.6
2.7

2.8
2.9
2.10
2.11

General Description ...........Ol....OO....O....'...O...

F“lel and- Materia]-s ..0...0.‘...QQU..’.l‘...'."..0......:.000

Reactor vessel .....0..OO....CD..DO-0.0.0...0....0......

Reactor and Coolent-System PUmMPS cecoeecescccsccoscecss
Primary Heat Exchanger .,...1............L.............
Salt-to-Air Radiator‘;...,..f..........................
Drain and Storage TENKS '« v seeoesesasssossonseseososesss
2.7.1 Fuel Drain TANKS teeeeseescrcssessssessesbocsnes
2.7.2 Flush-Salt Tank ..ccosscescessscceccoscsccssscss
2.7.3 Coolant Drain Tank teeseanoncosesssasacaessnenes
2.T.4 Storage ToNK secesccososessvssessossscssssscsssss
Cover-Gas System ..o...........................;.......
Freeze VAlves .eccceccscsvcsecscscscsescscsccssssosssses

Smpler MeChanism .O...OO0.0;...0.0...OO....Q.....Q...l

'Nonnuclear Heating L3RI BN B BN BN BN NNCRE RN BN NN BN BN N BN NN BN BN BN BN BN NN B N NN BB BN BN

Instmentation and Contmls .....GO.O.AD...GG....O..........‘.

3.1

3.2

General ......\0'0.....000’9.0..0900.00’.0.0.0......._Q.O.

3.1.1 controlRequiremnts .'...DO..OG0.;.........‘...

(a) Xenon POidsonirlg LI B BB B B B BN B B BN BB BN AR BN B BN BN BN BN NN BB N

(.b.)f Power coefficient .0.000...;?....0.0..00..... '

(éj De]-ayed Neutmns ® 0 00 0800 99 S90S ST OCOEN PSSP PSS
(d-) Bumup [ B BN BN BN BN RN -BN BN BN BN BE BN BN NN BN BN N BN BN R RN BN R B BN BN NN BN BN BN BN BN BN N
3 .l.2 O-ther Contml Featums .0 .l. -.. 0060008 S0 PP Q0O RPOOSS

3.1.3 Nomnuclear Controls .cceccesessosccsssscssosssne

mstments .......9......0..0..00.......lv....QOOQCOOOO.

'3.2.1 Reactor-Power Measurement ....cccoeceeessccccsocscs

3.2.2 Fuel INVentory cececescevrecosscrscsccssccsocssssssce

-lve

viil
1l

3
h‘o
5.

T

+3.2.3 Nuclear-Instrumenté',,.........,.;..............
3.2.h Radiation Mbnitofing ...l.......,..,..,......,..
3.2.5 Pressure Measurements seo0000e00000s0000esssse e
I3.2.6 Flow MEasurements a,.;.,,.....a.................
3.2.7 Temperature Measurements .cscosceoevccsccsccssco
3.2.8 ‘Liquid-Level Measurements ..eeoceccecocosecccnsss

Reactor msics Data 000.0.00!06000.D0.0.0......0009‘.0;0.0.‘.{.‘

’Ihe Reactor conmlemnt DO00000D.OOO0.00.....0.0!000000;00.0..
5.1 Bui‘ldin.g 00.0.00000000.0o.000000..0.....000;00.0..0900000
5!02 Containment_WDOG000DG00.00‘00.0000..00000000'000OOOQ......O..

5.2,1 Fuel-System Contalner Desigh .ccceccoscseccccoces

522 O-bher Gells 000.0000.0.0.0....COQO.O.OODO.OOG..OGV

5 23 Penetrations .0....0....00.'.00..000000.0.0.00..0
503 Shieldin:g .00000O..0..00...00..00..000000000000000000000
501" Arrmlgement of Equipment'OOOOOO0.0000....GOO..0000.00...

5.5 Mailltenazlce BOOOOOOD....OOOOO0.00..O0.00...0000000{0000..

Construction, Startup, and Operation .eescocceeccoccccccssesse

601 Constmction .0.:..COOOOOGOOOOOOO0.00....OOOOOQObOOODQOQOO'_

602 Fl\lShwsalt Test o-‘oo_ooooooooo..oooooocoeoeo‘oqoooo‘o‘_ooooooo

6.3. Startup ooooooooonooooonoocoo.ooogoo,ooooo.o"o_ooo;t;’o'oo;oooo'

601I' APProaCh to Po;wer:-...0DOOOOODOOOOOGOOOO‘O‘;00-.0..,00000_..0

6.5 operations Personnell'00000000.0OOO..DO.0400000000000...00.

Hazafis Ana]-ysis .OD'OOO000.0..000.0....._.D..OO...‘OUOOGO00..00A..

T.1 Damage to the Primary Container ...... coesoesossacacene
T.1l.1 Reactivity ExcursionsS ..cscscccecscescecocssccsscs
(a) Star.tupAccid.en‘t».,....,“..,_.....,.o,“..”.....

| (b) Graphite Problems .cccesssoosccccsscccsaoees

T.1l.,2 Fuel Separation .,.,.....,........o.”..“._,..‘.....‘;,.

- T.1.3 Flow Stoppage
(a) Fuel-Circulating-Pump Failure .,-,....o..a;..

(b) C@olantPumpFailure cecosscoecsossesosssass

() Simultarne@ushmpFailures cosoo0sccacesceos

(d) FlO‘W‘ Stoppa»ge in F'Ulel I-O@P soecooecoecosesoo

e
T.2

7.3

7 L

T.1l.4 Control System Fallure cceoseseoscoceossocsccssse

701.5 Drain-Tank Hazards ooooooooooooooocooooo.oooooooo-

T.1.6 Other Possibilities for Primary-
Container_Dmge 00O OO GO0 0GOS SO OG0 S0 SO0 OSSOSO

(a) Freez.e-Vaive DEIAZE «ocococceccecccccecsssons
(b) Freeze FLETIZES 0 000 s's0aososcosossossossessss
(e) .Exceésive Wall Temperature .scococscocososos
(A) Excessive SLIESSES oeococsosscssscssoossssss
(e) Corrosion ..,
T.1.T Detection of Salt Spillage
Ru_ptgre of the Secondary Container cesoscccssccoetossoss

70203- MSSile Damge 0000000..000.0000.'0OOOOOOOOOOAGOO»O;.

702.2 -ExceSSive Pressure 6 0OD0D0O S0 OSSO OO0 SDODD SO G®CO0SOOSOOSPOOOS

(a) Sal‘b Spi-llage 00000;000'0000..0‘0‘60000000000.00

(-b) Oil"'Line Rupture 00000..00.300lOOOOOOOOOOOQ.O.

702.3 Acts Of Nature .QDOOOO...‘O0600.080.000000’0.000..
(a) Earthquake 00 0D0S$ 000G GO0 OSSO0 SOO000O0CO0O0S S SO0
(b) Flood ocoooooooooocoooooooo'oooo-ooooooooooooo

70201'I' Sabotage oooo.oo.!000.0‘0000000000.000000.00....0.

Consequencés of Radloactivity Release from
the Secondaw container 0O ® 0 000 O0®®0 & 000G O0C 00O &0 2DGS 0 @ &9

T.3.1 Rupture of the Secondary Container .cccccececocses
7.3.2 Maximm Credible Activity Release .oceocoococeses
T-3.3 Beryllium and Fluorine Hazards ceccecccosecsceses
{(a) Beryllium ........,...;;,.........,..,.;....
() Fluorine ...coovcecseoccccoscscocceosoocoosass

The Site ® 00 0O 00 PSOPO0OOCS OO0 O SO0 O0O00O0C0O0OSS®O0 SO0 S00O0O0S$080L0B00SD0

APBMH *CPROO0DODSSPOOGH GO0 OCHODBOOOTOSSO00O000O0CS0OOS®SDOSOROSSOSOOSO S SO

- A
B.

dhemistry and Corrosion cccccocococcoecoessscosccoscccssscao
Hot=Spot ANBLYSLE ooooosecoooooosossoasesoosooonsoossss
Specification £or Drain TaNK .eccooseccoccssoscoscoeccssosns
Component Development Program in Support of MSRE .cceeese

- Calculation to Support Maximum Credible Accident ..oecsoe

Graphite conmatibility With Salt 90 ¢ &8 0 0CHSOO0SQO000COSOSSOESS

WCES 0 O0 P OO OO TS OO O EPS®O0000O000CO00 S0 SO0 SO SO S 00000000000 O0DOO0ESES

-vi-

103
107
117

120
v Table
No.
1
2
3
b
- 5
“ 6
T
8
9
L
\i

/

4.
£
LISTiOF TABLES
o
&
i f Page
Composition and. Properties of Fuel and Coolant
forMSRE l.....o..l.ll..l.'.b...............ODOII‘.'. T
Composition and Properties of INOR=8 ..e.vveevecsencons 8
'Properties of MSRE- Core Gr&phite cesessecscscessserrees 9

Reactor-Vessel Design Data .........,.;...,;.;......... 13
Design Date for Fuel and Coolant PUIDS .u.ievssssessess 16
Design Data for Primary Heat Exdhangér'......a....;....-' 18

‘Design Data for Salt-to-Air RadiBtor ...ececeeeeciessss 21

Design Data for Fuel Drain Tank, Coolant Drain
Ta-n.k and. FluSh"Salt T&D.k oo--.olnoo-o.._‘_-oooooooonoo-o'o-o 25

Ré&C’bOI‘ PhyBiCS D&ta ....o.ogeoo.i-oo-on-'.-'o.oo.-ooooooo )'I-B

..vj_i.. ¢
Fig.

O 00 1 O I & w M M

)
O

A2

LIST OF FIGURES

MSRE Flow Diagram LA R AR R R R R R EEEEENENRNEBENEMENBEERNEESE RSN BN B

Artist's Impression of MSRE Arrangement ....csoceceevsoess

Reactor-Vessel ASsembly ..ceecescocosscossssssssssssscsss

Typical Graphite Stringer Arrangement ........c.eceeveees.

Circulation Pump and MOtOT ceeeeeecesocvesvcsccssossscnsss
Primary Heat EXChanNgZer ...ecsceesosceccssccassoscsscsasss
Salt-to-Air Radiator ....ccecceciecneccncnsncncencnnnnsnes
Fuel Drain and F111 Tanks .eecceesecsssscessocossoncascoas
Cooling Thimble for Fuel Drain and FAll Tanks ...........
Cover-Gas SUPPLY ceccosccssscctssacscsasssossssssssssnsoas
Fuel-System Off-Gas Disposal ..ecececececocesoscacncnanns
Coolant-System Off-Gas Disposal ..ecevecccrscccsccconanns
Radiant Heat Freeze Valve ..ccececoscsccrsssaccsscoscnsnas
Arrangement of Ion Chambers ccceeessccsccccscssssosccsssse
Plan of Building T503 «eeeeveeoesnnsesnsneosesoneessnnons
Shielding and Sealing Membrane for Top of Cell .ecevevess

North-South Sectional Elevation of Buillding 7503 eteeeeees

Arrangement of Equipment: Plan View ..ccceceecccsccaacss
Arrangement of Equipment: Elevation ....ecciv0e... coesas
Remote Manipulator and Shielded Control RoOm ...cececeass
Afterheat POVET GENETALION «.oeeseseeseossonseoscessonses

Map of Cities and Counties Surrounding _
Oak Ridge Area LI B NN B BN B RN BN B R R N R R R R B NN NN R R R R R RN R R R R R IR R )

Plot Plan: Molten-Salt Reactor EXperiment ......eeeeeo..
Estimated Vapor Pressure of Fuel Salt ................,{.
Phase Diagram of Coolant Salt ..ceececevisssrsccscssssennes

-viii-

«)
- ACKNOWLEDGMENT

The authors are indebted to many members of the Mblten—Sait
Reactor Program for their contributions to this report, and to the
Program Direétor; R. B. Briggs, for his advice and suggestions.

Initial studies of hazard evaluation were made by Messrs.
Remo Galvagni, Italy} John W. Holtzclaw, U.S5.A.; Osamn'Kawaguchi,
‘Japan; and Francisco Z. Pines, Spain, students of Oak Ridge‘School
of Reactor Technology. Their findings represent a significant

contribution to the material of the present report.

X -
L

ok

MOLTEN-SALT REACTOR EXPERIMENT PRELIMINARY
HAZARDS REPORT -

S. E. Beall

W. L. Breazeale- B. W. Kinyon

1. INTRODUCTION

One of the principal programs of the~0ak Ridge‘National Laboratory
is the development of liquid-fueled reactors. Since 1951 the Leboratory.
has constructed and operated two experimentallreactors fueled with uranium
in aqueous solution and one fueled with molten salt.

The first. experiment with each of these concepts demonstrated nuclear

‘feasibility only. IEngineering feasibility, dependability, and other

factors were to be determined in later experiments such as the current
Homogeneous Reactor Experiment No. 2 and the subject of this report, the
proposed Molten~Salt ReactOr Experiment (MSRE). : |

The development of molten-salt systems has beenrpursued continuously
since 1951, although the ma.jor effort was supported by the aircraft
reactor program. Application of the molten-salt reactor to stationary
power production has almays been considered desirable for three highly
important reasons: , _ ,

1. Molten-salt reactors have a great advantage because they have no

fuel elements and-consequently none of the problems associated with fuel

' elements. :Burnup is not limited by radiation damage or reactivity loss.

There are relatively simple methods for reprocessing fuel and blanket
salts, and their reconstitution involves only dissolution of UF4 or ThF4
in a carrier salt with no metallurgical, ceramic, or mechanical processing.
2. Molten-salt reactors can operate at very high temperatureito
produce steam at conditions comparable to those for the hest fossil-fuel
plants. The use of a fluid fuel, circulating at high‘rate, can be com-
bined with large temperature differences in the core and heat-transfer
systems to'produce very high power density. High power density and low

fuel inventory in the reactor and the processing plants combine to produce

' high specific power. In spite of the high temperature, the operating

pressure is <50 psig.
-2-

3. The nuclear and physical characteristics of the salt and the
use of unclad graphite as a moderator make possible”thé achievement of
very good neutron economy. Breeding ofi the t‘horium-U233 cycle with a
fuel yield of about 8 per year appears to be attainable.

In order to demonstrate that many of these desirable features can
presently be embodied in a practical reactor which can be 0perated safely
and reliably, and can be serviced without unusual difficulty, the Osk
Ridge National Laboratory has proposed recently this molten-salt reactor
‘experiment. An additional important objective of the experiment is to
- provide thé first large-écale test of unclad-graphite moderator, fuel
salt, and container materials in-lbng-term oPefation"at high temperature
and power. '

This is a preliminary report prefiared for review to obtaln approval
of the proposed site. \It is based on the present incomplete reactor
design and is primarilx‘concérnéd with the hazards of the eiperimfint as
it iéflpresently visualized. The hazards studies of the Aircraft Reactor
Experimentl and the proposed (but not built) Aircraft Reactor Test®
provide a good background for the prdblems presently foreseen and the
proposed solutions discussed in this study. Furthermore, experiéncq in
cperating three fluld-fuel reactors pyér a perio& of nine years pro-
vidés a good hasis for the design criteria and operating practices.
Although the general design of the reactor and its facilities has been
investigated for several months, detalls are still unsettled and
importan£ changes may be made before the designlis completed. |

lSu;perscript nurbers refer to similarly numbered items in the 1ist of
- references on page 120.
'y

1

-3-
2. REACTOR COMPLEX
2.1 General Description

The proposed Molten Salt Reactor Ekperxment (MSRE) is designed for -
a heat generation rate of lO Mw by use of principles which will apply to
the design of a much larger power reactor A flow diagram for the reactor
and coolant systems and an artist 's concept of the facility are presented
in Figs. 1 and 2. | | |

In the reactor primary system the molten-salt fuel is c1rculated
through a cylindrical reactor vessel which contains a graphite core
matrix._ Under design operating conditions,‘fuel enters the core at 1175°F
and leaves at 1225°F. Then it flows to a 1250-gpm sump~type pump mounted
directly above and concentric with the reactor vessel. The pump discharges
the fuel through the shell side of a cross-baffled shell-and-tube heat
exchanger and back to the reactor inlet. o _ N

A coolent salt is pumped through the tubes of the primary heat ex-
changer and then through tubes of an air-cooled radiator hy another sump ~

type pump. It flows at a rate of 830 gpm and cycles‘between 1Q259Exand

_llQO°F._ The coolant-salt pressure is kept higher than the fuel-salt_

pressure‘to prevent the escape of fuel in the event of a tube failure; |
Air is blown over the bare tubes of the radiatorlhy;two’axial blowers_ofi

164,000 cfm total:capacity. ~Electrical heaters on the piping and equipment

" of the fuel and coolant systems keep the salt'molten at all times.

A liquid-vapor interface is maintained in the reactor fuel systemlin
the sump of the pump. Fuel isvbypassed through the sump at a rate of
50 gpm, and the gaseous fission products in the bypass stream are trans-
ferred to a helium cover gas. There is a continuous flow of 7 liters/min
of helium through the sump to the off-gas system; the helium system is
used to pressurize the reactor to 20 psia.

In addition to the reactor and coolant systems, the plant is provided
with such auxiliaries as drain tanks for fuel and coolant salts, equipment
for sampling the fuel in the reactor, alhelium-cOver-gas system, facili-
ties for handling radiocactive wastes, and the usual nuclear and process .

control instrumentation and plant services.
PRIMARY SALT

HEAT EXCHANGER

1025 F:‘

UNCLASSIFIED
ORNL-LR-DWG 56870

COOLANT
PUMP

830gpm

LiF — 70%
BeF, — 23%
ZI'F4 - 5%
UF4 - 1‘70

REACTOR
VESSEL

|
i
|
:1
l ThE, — 1%
|
|
|
||

FREEZE
VALVE'

SPARE FiILL AND FLUSH
FILL AND DRAIN TANK TANK
DRAIN TANK (68 cu ft) (68 cu ft)

(68 cu ft)

REACTOR CELL

1400°F

LiF — 66%
BeFE— 34 %

AIR
167,000 ctfm
100°F

Fig. 1. Flow Diagram of Heat Transfer System.

SECONDARY SALT

— o | 300°F

COOLANT
DRAIN TANK
(40 cu ft)

-1-‘—
o)

i

o
w
u
7
7,
<
4
Iz}
=z
35

ORNL -LR-DWG 52876

‘Fig. 2. Artist's Impression of MSRE Arrangement. !

_
|
;
!
-

Under normal steady operating conditions the reactor is self-
controlling, as a result of the negative temperature coefficients of
reactivity of the fuel -and the graphite moderator. The temperature
coefficient of -3 x 1072 (Ak/k)/°F of the fuel provides for fast con-
trol. The total temperature coefficient is -9 x 10-> (2k/k)/°F, and
this coupled with'the'small amount of excess reactivity loafied‘into.the
reactor provides the margin of safety against nuclear excursions to
excessively high temperatures. . Nuclear control devices are'provided
primarily to hold the reactor suberitical below 1000°F during startup,
to compensate_for some fission product poisoning and burnup, and to
keep the critical temperature below 1300°F during abnormal periods of
operation. The reactor power is controlled by regulating the rate of

heat removal. The nuclear reaction can be stopped by the control

devices-and the system can be shut down completely by draining”the fuel.

Fuel addition in large amounts for the complete’loadings will take
place in the fuel drain tank. Subsequent additions to compensate for
burnup and fission-product poisoning will be made through a sampling
and enriching system communicating with the gas space in the pump bowl.

The system components are of all-welded construction. CompOnents
in the reactor fuel system are connected to the piping by specially de-
veloped freeze flanges which utilize frozen salt as a sealant for the
high-temperature fluid fuel. .Braaed'connections are planned for'the
radicactive auxiliary systems. The use of these joints makes possible
remote maintenance of the system following power operation. Except for
flanged connections to the primary heat exchanger, the coolant‘system
is of all-welded construction and can be maintained directly within a
few minutes after shutdown

No velves of the ordinary type are used in contact with the fuel
or coolant salts. Flow 1s prevented by freezing salt in designated' |
sections of pipe. The freeze valves can be thawed in a few minutes

and are the best choice for drain valves.
2.2 Fuel and Materials
Fuel for the MSRE is a solution of U235Fu ThF), and ZrF), in an

2

L1 F-BeF carrier salt. The composition and properties of the fuel are

a
, L -T- _ |

listed in Tableul..-Li7F_is a salt of. good fluid-flow and heat-transfer
properties and low neutron cross section. Low melting points are dbtained'.
in mixtures with BeF2 U235Fh is the primary fuel constituent ThFLL is

present as a fertile material. The fuel is representative of the. core

fuel for a two-region breeder or a one-region U235 burner reactor.

Table 1. Composition and Properties of Fuel and Coolant for MSRE.'

Fuel Salt  Coolant Salt

Composition, mole %

LiF (99.97% 1i') _ | 70 66
BeF, | 3 3k
ZrF), . | 5
ThF), . | 1

U, - - ~L

. Physical properties , , l, ,

Temperature, °F | | 1200 - 1062

. Density, 1b/ft> 15L4.3 120.5
Viscosity, 1b/ft-hr 17.9 . 20.0
Specific heat, Btu/l1b-°F : 0.4 0.57
Thermal conductivity, Btu/hr-ft (° F/ft) 2.75 . . 3.5
Liquidus temperature, °F - 828 + 5 . 84T x5 .

Oxygen as 02 or in CO, H20, and other compouhds reacte\with the four-
component mixture to precipitate U02; however, if ZrFu.is present in an
amount such that Zr/U ~ 3/1, only Zr0, 1s precipitated by reaction with
oxygen-containing materials. During handling and while in the reactor,
the fuel must be blanketed by an inert cover gas such as heliqm, to pre-
vent contamination by gases and vapors containing oxXygen.

~ The coolant.selt'is an LiTF-Ber mixture of composition and properties
as shovn in Table 1. The same genefal considerations_that apply to hand-
ling of the fuel also apply to the coolant }

The principal materlal of construction for the resctor systems is
INOR-B, a n1ckel-molybdenum-chromium alloy developed,at‘ORNL Por use with
fluoride salts'at'high'temperature; ’The composition and properties of
-8~

INOR-8 are shown in Table 2. When the material 1s attacked, chromium is
leached from the elloy, resulting in the formation of subsurface voids.

Under most circumstences the rate of attack is governed by the rate of

diffusion of chromium in the alloy.

Measured rates of attack in typical

fuel and coolant salts have been less than 1 mpy at temperatures to at
least 1300°F. : |

I.

Table 2. Composition and Properties of INOR-8

Chemicel Composition”

II. Physlical Propertiés

Element - % | Element
Ni, nlino Ba.-l.o_( g 66 - 71) Mn, m&x-
MD, mmc. 1500 - ]_.8«;0 Si’ max.
CI‘ 6-0 - 8-0 Cu., III.B.X-
Fe, max. 5.0 B, max.
C " O-Oh‘ - 0008 W, max.
Ti + Al, max. 0.50 P, max.
S, max. 0.015 Co, mex.
Density, g/cc 3

1b/in7 .
Melting point, °F

Thermel conductivity, Btu/hréft2(°katju,

At 1112°F
" 1292°F

Modulus of elasticity, psi
At 1170°F
1290°F -

Specific heat (est.), Btu/1b=°F
Mean coeff. of thermal. expansion

%
0.80
0.50
0.35
0.010
0.50
0.010
0.20

8.79

0.317

2370 - 2430

12.20

13.01 E

26.2 x 102

24,8 x 10

0.095 at 212°F

°F in./in./°F_  AP(°F)  AL/L (in./in.)
70-1200 *¥-7.81 x 10 1130 8.82 x 10-5
III. Mechanical Properties
1/4 Min. Spec. 2/3 Min. Spec. 4/5 Rup. Str. Stress ' Max.

Tensile Yield for for Allow.
Temp. Strength Strength 10° hr 0.1 CRU ©Stress
(°F) (psi) (psi) (psi) (psi)  (psi)
1200 17,100 16,800 © 8,300 7,500 4,950
1250 16,100 16,600 6,200 5,400 3,600
1300 15,000 16,400 4,800 4,100 2,750
1350 13,800 - 16,300 2,050

3,600

3,100

c =0-

“Although the salt has moderating propertieés, use of a separate mod-

erator has the advantage of reducing the inventory of fuel in the reactor

and provides’ for ‘better neutron economy in a. breeder

)

Unclad graphite is

compatible with salt -at high temperature both in and out of radiatlon and

is the preferred moderator. _The‘properties‘of a graphite that satisfies

the requirements of the MSRE arerlisted in Table 3.

Table 3. Properties of MSRE Core Graphite '

Physical -properties

- . Bulk dens1ty, g/cc |
| Por051ty '

accessible, %
_ inaccessible, %
total, % :

Thermal conductiv1ty, at amblent temp,
unirradlated Btu/hr ft2(°F/ft)

parallel with grain
normal to grain

‘Temperature coeff1c1ent of linear
expansion, Ffl '

i
iV

” . parallel with grain
normal to grain

Matrix coefficlent of permesbility
to helium at 70°F, cm®/sec

Absorption of salt at 150 psig, vol %
Average specific heat at 1200°F, Btu/1b-°F

Mechanical strength properties

Tensile strength, psi
Flexureikstrength,'psi
Compressive strength, psi
Modulug of elasticity

paraliel with grain, psi
normal to grain, psi

1.87 - 1.89
n,'?
~ 8.9
~ 15.9
70
60

-5
1.3 x 10
1.6 x lO-6
1 x 10"lL - 1x107
0.50
Q.42
1500 - 2400
2000 - 3500
8600
1.9 x 102
1.5 x 10

- =10-
2.3 Reactor Vessel

The reactor consists of a cylindrical vessel approximately 5 ft in
digmeter and 7-1/2 ft high with an inner cylinder which forms the inner
wall of the shell-cooling annulus and serves to support the graphite
matrix with its positioning and suppbrting members and flow-regulating
orifices. Figure 3% is an assembly drawing of the reactor vessel and
core. Fluid enters the vessel at the top of the cylindrical section and
flows downward in a spiral path along the wall. With the design flow of
1250 gpm in the l-in. annulus, the Reynolds modulus is 22,000. At the
estimated heat generation rate of 0.2 w/cm5 in the wall, 23 kw of heat is
removed while maintaining the wall temperature at less than 5°F above the
bulk fluid temperature.

The fuel loses its rotational motion in the lower plenum and then
‘flows upward through the graphite core matrix, which cofistitfites‘about
77 .5% of the core volume. The moderator mafrix is constructed of
2- by 2- by 63-in.‘stringers of graphite which are loosely pinned to re-
straining beams at the bottom of the core. Molybdenum bands at the top
and center of the assembly restrain the bowing induced by the radial
neutron flux gradient.

Flow. passages in the matrix are 0.400- by 1.20-in. rectangular
channels machined in the faces of the stringers. A typical érrangement
of graphite stringers is shown in Fig. L. Fiow through the core is lam-
inaf, but because of the good thermal properties of the graphite and fuel,
the graphite temperature at the midpoint is only 4O°F above the fuel |
mixed-mean temperature at the center of the core. )

Provision is made for remote removal and replacement of five
stringers at the center of the core. They will be examined periodically
to determine whether the graphite deteriorates with increasing exposure
time. An INOR—Bnpiece is installed in the top dished head to displace
fuel and to provide a part of the shielding for equipment above the
reactor.

Design data for the reactor vessel are detailed in Table k.
-11-

UNCLASSIFIED

ORNL-LR- DWG 52034R

FUEL OUTLET

. GRAPHITE SAMPLE BLOCK

(o]
3
w
T
1]
Wi
o«
O
o
a
o
T

FUEL INLET

CORE YOKE

GRAPHITE-MODERATOR

STRINGER

REACTOR VESSEL

GRAPHITE-MODERATOR
STRINGER -

REACTOR GORE CAN

FUEL PASSAGE

CORE-POSITIONING
GRAPHITE BEAMS

VESSEL DRAIN LINE

CORE GRID SUPPORT

Fig. 3. Reactor - Vessel Assembly.
-]12-

UNCLASSIFIED
ORNL-LR-DWG 56874

PLAN VIEW

TYPICAL MODERATOR STRINGERS

SAMPLE PIECE

CROSS ~COMMUN! ~
CATING CHANNELS

NOTE: NOT TO SCALE

Fig. 4. Typical Graphite Stringer Arrangement.

-13-

Table 4. Reactor-Vessel Design Data -

Structural material
Core vessel

oD

ID
Wall thickness

Design pressure

Design temperature

Fuel inlet temperature .
Fuel outlet temperature
Inlet

Annulus ID

Annuwlus OD

Over-all height of core tank

-~ Head thickness
Inlet pipe
Outlet pipe
Core container
ID
oD

Wall thickness
Height

Graphite éore

Diameter

Core stringer

Number of fuel channels
Fuel-channel size:
Effective core length

Fractional fuel wvolume

INOR-8

. :59-1/8 1in.
58 in.
 9/16 in.

50 psi
1300°F

1175°F

1225°F

Constant-area distributor
56 in.

58 in.

94 in.

1 in.

6-in.-OD tubing, 0.205-in. wall

8-in. schedfhd pipe

68-1/2 in.

55-1/4 in. .

2 x 2x 63 in,
1064

1.2 x 0.4 in.

~ 65 in.

0.225

=1L
2.4 Reactor and Coolant-System Pumps

The fuel circulation pump is a sump-type centrifugal pump with a
vertical shaft and an overhung impellef. It has a 75-hp motor and is
capable of circulating 1450 gpm of salt against a head of ~ 50 ft.

Figure 5 is a drawing of the pump, and design data are presented in
Table 5.

| The pump assembly consists of motor and housing, bearing shaft and
impeller assembly, and sump tank. The sump tank is welded into the reac-
tor systeh Piping and serves as the expansion tank for the fuel and as a
place for fhe separation of gaseous fission products. The bearing housing
is flanged to the sump tank so that the rotating parts can be removed and
replaced. The motor is loosely coupled to the pump shaft, and the motor
housing is flanged to the upper end of the bearing housing to permit
separate removal of the motor.

The pump is equipped with ball bearings which are lubricated and
cooled with oil circulated by an external pumping system. The oil is con-
fined to the bearing housing by mechanical shaft seals. Helium is circu-
lated into a labyrinth between the lower bearing and the sump tank. Part
of the gas passes through the lower seal chamber to remove oil vapors
which lesk through the seal. The remainder flows downward along the shaft
to prevent radioactive gas from reaching the oil chamber. |

Massive metal sections are incorporated in the pump assembly as
shielding for the lubricant and the motor. The mobtor is enclosed and
sealed to prevent the escape of radioactive gas or fluids which might leak
through the pump aésembly under unusual conditions. Water cooling coils
are attached to the housing to remove heat generated by the motor. '

A similar pump is provided for the coolant system except:that fewer
provisions are required for protection agasinst radiation. The pump is
driven by a 125-hp motor and is designed to circulate 850 gpm of salt
against a head of 100 ft of fluid. The complete design data for the

coolant pump are included in Table 5.

|2
&

-15-

~ UNCLASSIFIED
ORNL~LR~-DWG 56876

MOTOR AND DRIVE
# COUPLING HOUSING

BUFFER GAS FOR GASKET\ e
LTI
FES 1 T L L
TR 1 18 LUBRICATING OIL INLET
R PO 1 : ' C
[ ", ; 1"
T l! ""‘lf. .] o
!i! 1113 | j PURGE GAS INLET
, : ' ol -
OIL LUBRICATED BEARINGS i ‘
- | T : GAS SEAL
ih ZHH
LUBRICATING OIL OUTLET SHIELDING
PURGE GAS OUTLET q | |
| f f GAS FILLED EXPANSION SPACE
NORMAL OPERATING LEVEL 1 | |

TO HEAT

. =\ EXCHANGER
DIMENSIONS: . ' _ -
36-in. DIA AT PUMP ! *gdoflm
BOWL x 88 in. HIGH “ 11l
. ) ,|'E "R
[ i

FROM REACTOR

FWg.5.FueICHcMafloh Pump.
-16-

Table 5. Design Data for Fuel and Coolant Pumps

-*Fuel Pump Coolant Pump '

Flow, gpm 1200 - 1450 850
Head, ft 50 * 5 at %&50 gpm 100
-Motor horsepowef, hp 75 ‘j o 125,
Pump speed, rpm 150 1750
Intake o 8-in. schéd-4o pipe 6-in.-OD tubing,
- ) o 0.205-in. wall
Discharge 6-in.-0D tubing, 5-in.-0D tubing,
0.205-in. wall 0.165-in. wall
Sump-tank volume,* £ 5.2 ' 4.8
Normal salt volume, £t° 2.61 . 2.4
Expansion volume, ** £1° 1.83 o | 1.6
Sump tank | S
Outside diameter, in. 36 o 36
Height, in. 15 b
Over-all assembly height, ft 7-3/4 7-1/3

Structural material INOR-8 _ ~ INOR-8

*Not including volute.
**Normal to maximum.

2.5 Primary Heat Exchanger

The primary heat exchanger_(Fig. 6) contains 165 tubes (1/2-in. OD,
0.045-1in. well, 13 ft long) and is designed to transfer 10 Mw from fuel
salt (in the shell) to coolant salt (in the tubes). The design is a con- -
ventional cross-baffled shell-and-tube configuration with emphasis placed
on ruggedness and reliasbility rather than high heat-transfer performance.
Design data are listed in Table 6. |

-Space limitations in the reactor cell require a short unit. The U-
tube configuration mekes this possible without greatly reducing the effi-
ciency of heat transfer as compared to a straight counter-flow unit and
eliminates a thermal expansion problem. The tubes are welded and back-
brazed to the tube sheet in order to greatly reduce the probability of
leakage between the fuel and coolant.
SSSSSSSSSSSS
RRRRRRRRRRRRRRRR

» Yo-in-0OD HEAT
EXCHANGER TUBE

THERMAL-BARRIER PLATE (25 % CUT)
o

COOLANT INLET
) 2,
X
5%)
NN )
83 N , :
Lo 164-in. OD x O0.2-in. WALL x B-ft LONG
" ‘|“ ., ) .“‘ . " " i
",'J'JL"‘;
gy 4
el
\
'fv i
AT X Al ' ,
* COOLANT-STREAM . ; .COOLANT OQUTLET
.| SEPARATING BAFFLE fl

FUEL OQUTLET

Fig. 6. Primary Heat Exchanger for MSRE.

CROSS BAFFLES

‘LT‘
=18~

Teble 6. Design Date for Primary Heat Exchanger

Structural material
Heat load
Shell-side fluid
"Tube-side fluid
Layout |

" Baffle pitch

Tube pitch

Active heat-transfer length of
shell ‘

Over-all length
Nozzles
Shell side
‘Tube side
Shell diameter
Shell thickness
Tube-sheet thickness
Number of tubes
Tube size
Tube length
Heat~transfer surface ares
Fuel holdup
Design temperature
Shell side
Tube side
- Design pressure
Shell side
Tube side

Terminal temperatures at
‘design point

Fuel salt
Coolant salt

Effective log mean temperature
difference

INOR-8

10 Mw

Fuel salt
Coolant salt

-25% cut, cross-baffled and U-tubes

12 in.
0.775 in., triasngular

6 £t
8 ft

6-in.-0D tubing, 0.205-in. wall
5-in.-0D tubing, 0.165-in. wall
16 in. ID

1/5 in.

1-1/2 in.

165

1/2-in. OD, 0.045-in. wall

13 ft
‘259 £t
~ 5.5 £t°

2

1300°F
1300°F

50 psi

75 psi

Inlet 1225°F; outlet 1175°F
Inlet 1025°F; outlet 1100°F

133°F.

..]_9_
2.6 Salt-to-Air Radistor

The thermal energy of the reactor is rejected to the atmosphere by
means of & salt-to-air radiator. The radiator contains 120 tubes .
(3/4-in. OD, 0.072-in. wall, 33 ft long) and is assembled as shown in
Fig. 7. Design data are listed in Table 7. o |

.Several features were inqbrporated in the design as protectioh.
agelnst freezing of coolant salt in the radiator:

1. Tubes are of large diemeter. |

2. The heat-removal rate per unit area is kept low by using tubes

without fins so that most of the temperéture drop is in the air
film.

3. The minimum salt temperature is kept 75°F above the freezing

point. |
., The headering system is designed to assure even flow distri-
bution between the tubes. o

5. 1In the event of flow stoppage, doors on the-radiator,housing
close within 30 sec, and heaters are turned on to prevent the
salt from freezing. , ‘

The layout of the tube matrix will allow movement of the tubes
with minimum restraint during thermal expansion. The tubes are pitched
to promote drainage.‘ ‘

The radistor is supported and retained in a structural steel frame
which 1s completely enclosed and insulated. Reflective. shields protect
structural members from excessive temperatures. The frame also provides
‘guides for the vertically sliding doors which are closed to thermally
~ isolate the radiator, should any situation develop which could cause salt
freézing in the radiator tubes.

Two doofs are employed, one each upstream and downstream; and they .
are raised and lowered at & speed of 7 ft/min during normal operation by
a gear-reduced motor driving an overheed line shaft. The doors are sus-
pended from roller chains which run over sprockets mounted on the line
shaft. The enclosure is capable of sustaining full blower pressure in
any position, and may be used to vernier air flow across the rgdiatOr'as

a control on the reactor load.
(s

UNCLASSIFIED
ORNL-LR-DWG 52037 R4

5-in. SECONDARY-SALT INLET

DIMENSIONS:
‘9-ft WIDE
6-ft HIGH
45-in. DEEP

\3/4-in.—OD X 0.072-in-WALL TUBING;
10 ROWS, 12 TUBES PER ROW

8-in. HEADER

5-in. SECONDARY -
SALT OUTLET

2%-in. TUBE
MANIFOLD

AIR FLOW

Fig. 7. Salt-to Air Radiator.

-0z-
~21-

Fmergency closure is effected by de-energizing a magnetic clutch
between the motor and the line shaft or, alternatively, by de-energlzing
magnets used to attach the roller chain to the door. Shock—absorbingm_
means are provided. Emergency closure 1s not contingent on door posi-
tion.

Table 7. Design Data for Salt-to-Air Radiator

Structural material |  INOR-8

Duty : 10 Mw

Temperéture differentials | _ | | _ |
‘Salt Inlet 1100°F; outlet 1025°F
Mr | . Inlet 100°F; outlet 300°F

Air. flow | - 164,000 cfm at 15 in. B,0

Salt flow | 830 gpm at ‘avg. temperature

Effective mean AT . 920°F

Over-all coefficient of heat transfer 53 Btu/ft?-hr-éF

Heat-transfer surface area | | 685 £2

Design temberature | 1300°F

Deéign bressure 15 pSi

Tube diemeter | ~ 0.750 in.

Well thickmess | © 0.072 1z.

Tube‘matfix | 12 tubes per row; 10 rows deep

Tube spacing o | 1-1/2 in., triasngular

Suvheaders | | 2-1/2 in., IPS

Main headers 8 in., IPS

Air-side AP , 11.6 in. H,0

Selt-side AP : - 6.5 psi

2.7 Drain and Storage Tanks .

Five tanks are provided for safe storage of salt mixturés when7
they are not in use in the reactor and coolant systems. They are: two
fuel drain tenks, a flush-salt tank, a fuel- and flush-salt storage
tank, and a coolant drain tank.
-22a

2.7.1 Fuel Drain Tanks

The fuel drein tanks have the important function of subcritical
storage of the fuel and must have means for femoving decay heat and for
meintaining selts molten when the internal heat generation rate is low.
Two tanks of the design shown in Fig. 8 are provided, each of 67 £t
capacity. Each tank can hold an entire fuel charge; so one is for
normal usé and the othér is a épare. The low moderating power of the
salt makes criticality impossible even with nearly double the planned
U235 loading.-

After long-term operation at 10 Mw, sudden draining of the fuel
requires that it be cooled at a rate of 100 kw to prevent excessive fuel
temperatures. Evaporative cooling was chosen over gas or other means on
the basié_of simplicity and independence of utilities. Heat is removed
by 4O bayonet cooling tubes (Fig. 9) inserted in thimbles in the tank. .
Water is fed through fihe center tube, and steam is generated in the
surrounding annulué. Heat is transferred from the thimble to the cooling-
tube through a 5as-filled annulus by radiation and conduction. Normally
the steam is condensed by ajwatér?cooled condenser, but it can be.ex-

_ Hausted to the stack in the event of failure of the coolant supply. A
300-gal feed-water reserve can provide cooling for 6 hr.

The drain tanks are insulated and are provided with electrical
heaters. They have dip—tube fill and drain lines ard gas connections for
maintaining a helium blanket for venting and for pressurizing to transfer
the salt. Design data for the drain tanks are presented in Table 8. -

2 07 ° 2 ° FluSh"sa:.Lt T&Dk

A salt of composition similar to the fuel salt but without thorium
or uranium is used to flush the reactor system to remove possible chemical
contaminants and to aid in preheating before the fuel is charged. During
shutdown it can be used to remove residual fuel after the primary system” |
has been drained but before the piping is opened for maintenance. The
 flush salt is stored in a tank (see Table 8) similar to the fuel drain
tank, but w1thout coollng tdbes.

P-4
@ iR
8 O &
o > 5
=2 - Y €T
g5 Z = 5 ® £
s O = 4 <+ T
m i} - o= | O
= | 0 oW T wo
M- _._N._ = .NI (] ™~ _H..
o z W = Ay
o =4 —
(@) oo X
w [t w a < - <1
a — o
| wQ o
= = -
=2 q
o wi
= =
3 )
L
T
[7)]
-~
c
o
T
L
he!
' c
o O
q f=
_ O
|
LCAC,~IC,;“Oa OO, o
=N
" o R s A 2oy /./l
- = .zflflfiwrww,%mp.wmww,#554444;.444444%%%4%%%%&%%444,414415441;55;.4/.1.41.4431.44%44%%/.4%##/.54%4..flml.%awfi..//., N /xf/////”. m
ol " F A N R A, At ///fi///////////////l\ /////// ”fi . ”
- QP. ,@//////w /fi o
4 D
9 < l,,\ .
-« LS ;
- & N Y R R RN R e nuw
N o —
Q _ : ,\ L
—Plv w \/ . J o G-r G
pM_ o J 2=
zc @ _y 1w o
oI Ogq no <
OO - w o a2 e’
©a o Qa - @
z8 2 == L 03
L& 0% o o, ga Wo
= e - w > 1 oo
Y p=d = > Z
=1 Jac w = m' ad
B wd a3 Y © o
mu._ 20 0 o Iy Jw
@ W [ > W O Rl qZ
o L= W > @
L ] )
ok~

UNCLASSIFIED
ORNL—-LR—-DWG 52038R¢

CQOLING-WATER INLET\\ /STEAM OUTLET

N
N
N
\
N
N
N
\
\
:
v

N

TANK HEAD

¥

y o

INTERMEDIATE
WELDING NIPPLE

OUTER CONTAINMENT TUBE

AIR SPACE

1
|
STEAM RETURN TUBE \)

¢ i|fllit! w— SALT CONTAINING TUBE

SPACER FINS
flSTEAM ANNULUS
| SPACER FINS
;. IN AIR ANNULUS

WATER FEED TUBE — J{\¥'F

AlIR SPACE

Fig. 9. Cooling Thimble for Primary-Salt Drain Tanks.
s

Table 8. Désign Data for Fuel Drein Tank, Coolent Drain Tank,
and Flush-Salt Tank

Fuel drain tank

Height
Diameter
Volume
Total
Fuel (normal)
Gas blanket (normel)
Wall thickness
Vessel
Dished head
Design temperature
Design pressure
Coolifig method
Cooling rate
Coolant thimbles
Number

Diameter
Coolant drain tank

Height
- Diameter
Volume
Total.
Coolant salt
Gas blanket
Wall thickness
Vessel
| Dished head.
Design temperature
Design pressure
Cooling method

INOR-8

81-1/2 in. (without coolant headers)
48 in. |

67.6 tt°
RN R %

~8 £t

1/2 in.
3/4 in.
1300°F
50 psi

Boiling water in double-walled thimbles

100 kw

40
2 in. OD

INOR-8

76 in.
36 in.

39.5 £t°

~3l4 £t
~6 17

3/8 in. .
5/8 infb
1300°F
50 psi

None
26 -

Table 8. (Continued)

Flush-salt tank _ , INOR-8
Height 76-1/4 in.
Diameter _ 48 in.
Volume N |

Total ~ 67.1 £t°

Flush salt ~59 £t

Gas blanket ~8 Pt
Wall thickness

Vessel 1/2 in.

Dished head - 3/4 in.
Design temperature 1300°F
Design pressure 50 psi- '

Cooling method = None

2.7.3 Coolant Drain Tank

A tank (see Table 8) of 40-£t° capacity and similar to the drain
tanks but without cooling tubes is provided for the coolant salt.

2.7T.4 Storage Tank

Occasionally it will be necessary to remove the fuel charge from the
reactor‘for reprocessing and to add a fresh charge. A separate storage
tank of 67--ft5 capacity is provided for storing used fuel while it is
being removed in small batches to a reprocessing plant and for accumula-
ting small batches of new fuel until it can be charged to onme of the
drain tanks. ) |

The storage tank is like the fuel drain tanks except that it has no
cooling tubes and therefore cannot accept salt from the reactor until |
the aftefheat has decayed for about two weeks. The tank is equipped with
lines for transferring salt to the fuel drain and flush tanks and to
equipment provided for loading and unloaaing carriers.

2.8 Cover-Gas System

Because the fuel salt is sensitive to oxygen, it must be protected

by an oxygen- and moisture-free cover gas at all times. The principal

. -‘»27_

functions of the cover-gas system. are to supply an inert gas for blan-
keting. the salt and for the pressure transfer of salt between components,
to pnovide a means‘for disposing of radioactive gas, and to provide a
higher gas pressure in the coolant system than in the fuel system. A
simplified flow diagram of the system is presented in Fig. 10.

 The cover gas is helium supplied in cylinders at 2400 psig. It is
purified by passage through filters, dryers, and oxygen traps (possibly
titanium, zirconium, or uranium chips at high temperature). Purified gas
is then sent to two distribution systems. The total flow 1s about 10,500
liters/day (STP), and it is distributed at about 50 psig.

The largest flow of gas is directed to treated-helium storage tanks
and then to the primary distribution system, which supplies nurge for
the fuel circulation pump, the freeze-flange“buffer zones, &and the fuel
drain tanks, where it is in direct contact with the fuel salt. Gas which
-passes through the fuel pump is circulated through a series of pipes
where it is held and cooled for at least 2 hr to dissipate heat from the
decay of short-lived fission products. Then it passes through a .charcoal
bed where xenon and Krypton respectively are retained for at least 72 and
8'days, and through.a filter and blower to the off-gas stack. There it
is mixed with a flow of 20,000 cfm of air which provides dilution of
1: 15 000. The charcoal bed is a series of nipes packed with activated
carbon It and a spare bed are mounted vertically in a sealed, water-
filled secondary container; either or both beds may be used.

Fuel-salt transfers require more rapid venting of gas, but the heat
load is low. A third charcoal bed is provided for venting those gases
before they are sent to the stack. Fig. 11 is the off-gas disposa.l flow-
sheet. | '

Although not included in the initial installation, prov131on will
be made for recirculation of gases from the outlet of the carbon bed
through a purification system and into treated-gas storage tanks.

: The'cover—gas distribution for the coolant system (also shown on
Fig. 10) supplies a smell flow of helium to the coolant system, the
sampler-enricher system, and to the fuelwpump motor. That equipment
mustvbe'supplied with nonradiocactive gas and will not normally be contam-

inated by gaseous fission,products. Gas from the eoolant;system is
260 psig

FILTER

DRYER

FRESH HELIUM STORAGE TANKS

20-day SUPPLY

2400psig MAXIMUM PRESSURE

OXYGEN % <
REMOVAL .

FRESH HELIUM SURGE TANKS
4313, 260 psig OPERATING
PRESSURE.

Fig. 10. Cover-Gas Supply

IEEE

N/

RRIEERIRIARE

UNCLASSIFIED
ORNL-LR-DWG 56872
FUEL-PUMP SWEEP-GAS UPPER INLET
FUEL-PUMP SWEEP-GAS LOWER INLET
FUEL-SALT STORAGE (2)
FLUSH-SALT STORAGE
INTERMEDIATE TRANSFER TANK

SUPPLY HEADER FOR RADIOACTIVE
HELIUM (50 psig) ' '

SUPPLY HEADER FOR NON-RADIOACTIVE
HELIUM (50 psig)

COOLANT-PUMP SWEEP GAS
COOLANT-PUMP FLANGE BUFFER |
COOLANT-PUMP MAIN FLANGE GASKET
COOLANT-PUMP MOTOR BUFFER
COOLANT-PUMP MOTOR FLANGE GASKET
FUEL-PUMP FLANGE BUFFER
FUEL-PUMP MAIN FLANGE GASKET
FUEL-PUMP MOTOR BUFFER
FUEL-PUMP MOTOR FLANGE GASKET
FUEL-PUMP LIQUID-LEVEL BUBBLERS
FUEL-PUMP LUBE OIL SYSTEM
FREEZE-FLANGE BUFFER
SAMPLER-ENRICHER | .
COOLANT-SYSTEM SALT STORAGE
FUEL TRANSFER CARBOY

- 93-
cw

|
|
WA ™ l t
HELIUM |
SUPPLY 0 N VOLUME CHARCOAL BEDS (2)
—_— l o i o HOLDUP 4.2 liters/min — 72 DAYS Xe HOLDUP
L1 ) ' :
FUEL i iq MAX — °F
PUMP : 1Q psig MAX —-1200 ‘.
|
H—
Jd

-

REACTOR CELL

UNCLASSIFIED
ORNL-LR—-DWG 56877

ACTIVITY
MONITOR

5cfm MAX
PRESSURE

2-in. Hg VACUUM

EQUALIZING LINE —mm

7503 BUILDING

VENTILATION HDR
¢

< O psig—85°F

Y

27,400 cfm

)

FLUSH SALT STORAGE

TANK VENT }%D

FUEL STORAGE

TANK VENTS X

ACTIVITY ___? ' :
MONITOR ‘ 46'__
ACTIVITY
: MONITOR | — 1
CHARCOAL
BED 1 cfm -

INTERMITTENT FLOW
NEGLIGIBLE HEAT LOAD

~——t
) VENT -
HEADER\
“SNFUEL . SSFUEL
TRANSPORT INTERMEDIATE

CARBOY VENT

~—SAMPLER ENRICHER VENT

TRANSFER TANK VENT

| §‘<

Fig.11. Fuel-System Off-Gas Disposal.

FILTER STACK

- 68-
-30-
vented directly through filters to the off-gas stack as indicated on

Fig. 12, Monitors will stop the flow to the stack on indication of high
activity.

2.9 TFreeze Valves

The-molten salt in both the.fuel,and coolant circuits will be
sealed off from the respective drain tanks by meané of freeze valves
in the drain lines. These valves (Fig. 13) are simply short, flattened
sections of pipe which are cooled to freeze the salt in that section.
Calrod heaters surround each valve so that the salt can be thawed quickly,
when necessary,.to drain the system. The”salt can also be thawed slowly
without the heaters by stopping the cooling. The valves are mounted
with traps on both sides so that salt cannct be blown out of the line.
The £111 and discharge lines fof the fuel and the flush-salt tanks are
manifolded together and are connected to an outer stofage tank. There

is & freéze valve in each line.
2.10 Sampler Mechanism

Small quentities of fuel can be added or removed by means of the
sampler mechanism which is connected to the fuel-pump bowl. A special
container located outside the containmentlvessel above the pump elevation
" encloses the working parts. A cable assembly with a reel is used to
lower a small bucket into the salt pool in the pump bowl. Fuel samples
can be removed for analysis, and new fuel can be added in small (less
than 120 g) quantities to compensate for burnup.

Since this operation purposely breaches the secondary and primary
containers, it is extremely important that substitute protection be
provided. This is accomplished by building a special solder-seal dis-
connect and two isolated compartments into the sampler, with protection
against both being open at the same time. Furthermore, the sampler
enclosure 1s ventilated to the charcoal-filter system. With these
mechanical devices and with special attention being given to operating
procedures, it is belleved that the sampling operation can be handled
with complete safety.

HELIUM SUPPLY

ACTIVITY MONITOR

J

1000 liters/day —A\/ W\

{0 psig MAX

r--
I

UNCLASSIFIED

ORNL-LR-DWG 56873

111
I
COOLANT PUMP

—

XD

B

£

COOLANT SYSTEM
SALT STORAGE TANK

Fig.12 Coolant-System Off-Gas Disposol‘.

F|LTER FAN

STACK

-Tg-
AlR BLAST COOLING JETS

(
RADIANT HEATING COIL

Fig.13. Radiant Heat Freeze Valve.

UNCLASSIFIED
ORNL—-LR-DWG 56875
-33-
2.11 DNonnuclear Heating

External heating of salt-bearing components of the reactor system is
necessary: |
1. to prevent freezing of the salts,
5. +to raise the reactor temperature to & subcritical value
for'experimental convenience,
"3, to heat the salts for reactor startup.

Replaceable'eleCtric heaters were chosen as the safest, most reliable
means of supplying the large (~300 kw) high-temperature requirements.
Diesel-driven generators and two separate TVA substations make a complete
failure of the heating system extremely unlikely. In the 1000 to 1509°F
range nearly all the heat is transferred by radiation. This makes it
| unnecessary that the heating elements be in contact with the vessel walls
and results in a safer system from the standpoint of overheating and
. arcing damage. | ‘ |
Different kinds of heaters are applied to different parts of the
reactor. The core vessel, drain tanks, flush tank, and storage tank are
equipped fiith hairpin-shaped resistance heaters which fit into wells sur-
rounding the vessels. The piping and the heat exchanger are covered with
clamshell-type resistance heater assemblles, which are designed for easy
removal. Reflective insulation is incorporated in the design.

The coolant system and the radiator are heated by Calrod-type
radiant elements with ordinary insulation, because these areas can be

maintained directly.
-34-

3. INSTRUMENTATION AND CONTROLS
.'3.1 General

The MSRE is & safe reactor because:

1. It has a good negative temperature coefficient.

2, Only a small amount of reactivity is fequifed to compensate

for xenon‘poisofiing, the negétive power coefficient, and
" burnup of fuel between fuel additions. |
3. The stability of the fuel increaées with increasing temperature.
4. There are no known mesns by which the amount of uranium in the
core can be increased rapidlye- |

Normally the reactor will operate at 1175 to 1225°F. However, it can
be cooled to 900°F and heated to 1300°F with almost complete freedom.
The fuel salt beginszto freeze at 830°F, and freezing should be complete
at T90°F. The serious consequence of ffeezing is that salt expands on
melting so that remelting must be done with extreme care or small pipes
will rfipture and salt will_be spilled into the containment cells.

In desligning the reactor, the maximum stress allowed 1s two-thirds
G6f that which will produce a minimm creep rate of 1% in 105 hr at
1300°F. The same sustained stress will produce 1% creep strain in
2000 t6 4000 hr and rupture in 5000 to'i0,000 hr at_1500°F; At 1700°F
the time to 1% strain is 40 to 100 hr and the time to rupture is 200
to 600 hr. The equipment and piping will be designed so that the
stresses, neglecting relaxation, will not change much with dhanging
temperature in an isothermal system. Heating to 1500°F for a thousand
hours or to 1700°F for a few hours should have little effect on the
life of the equipment. | |

Large temperature gradients are the main cause of excessive streés
on heating and cooling the reactor. For this reason normal heating and
.cooling will be done so that température'differences in the system are
less than sbout 100°F. To ensure this, the normal rates of change of
temperature will be kept below 1°F per minute for total changes greater
than sbout 100°F. If a nucleaf excursion causes the reactor core tem-

perature to rise 300 to 500°F very rapidly and to remain at the higher

~35 e

temperature, the thermal stresses induced in the piping and equipment

as they heat somewhat less rapidly to about the same temperature will

-cauge some ylelding and distortion.. The fuel pump probably will have

to be replaced. Several such severe;thermai shocks would be required
to breach thevsystem and permit the fuel to leak.

3.1.1 ControereQuirements

The MSRE has temperature coefficients asfifoIlows:
1. Tuel COEPFiCient .....ovieeseseses =3 x 1070 (2k/k)/°F

2. Graphite moderator coefficient ... -6 x 10> (2x/k)/°F
The fuel coefficient prevails during very rapid trensients. The iso-
thermal temperature coefficient is the total or -9 x 10~ (2&k/k)/°F.

Excess reactivity over that required for the reactor to be criti-:
cal while cleen and noncirculating at the design operating temperature
(1200°F)'must be provided in order to maintain the temperature wvhile
the reactor is operating steadily at full power. Thiswreactivitylmust
be sufficient'to compensate‘for xenon poisoning; loss of delayed neutrons

in the circulating system, end some burnup of fuel.

(a) Xenon Poisoning. Xenon will be removed continuously from
the 4% of the fuel flow which circulates through the pump bowl. Some
xenon will c1rculate with the fuel, and this will permit apprec1able ,
amounts of gas to diffuse into the voids in the graphite. Also there
will be some permeation of fuel 1nto the graphite. The exact amount .
has not been determined for the MSRE graphite, but enough work has been
done with 31milar graphites (see Appendix F) to form a good basis for
assuming that only 0.5% of the graphite volume will be occupied by
fuel. ZXenon produced in this fuel will contribute to the poisoning.,
The steady-state xenon poisoning at 10 Mw is estimated to be 1. 3% in
Mk/k. The peak poison level after 10-Mw.operation is 4 to 5% There
is no need to compensate for the peak poisoning because the stripping

operatlon continues to remove xenon from the fuel when the reactor
pover is reduced.

(b) Power Coefficient. Heating in the graphite as the power is.

raised causes the reactor to have a power coefficient of reactivity
estimated to be =0.02% (Ak/k)/Mw or -0.2% total at 10 Mw.
-36-

(¢) Delayed Neutrons. Decay of precursors in the piping and heat

exéhangersiresults in loss of delayed neutrons from the reactor core.
The reactivity changes by -0.1T% in Ak/k when fuel circulation is started
and increases again when circulation 1s stopped.

(@) Burn p. The removal of fuel and increase in fission-product
poisoning by burnup changes the reactivity by about -0.002% in Ak/k
per Mw-day. Fuel will be added to the circulating gystem at intervals
of 10 days or less. The total change in reactivity between additions
will be -0.2% in Ak/k or less. |

The total excess reactivity over the émount required for steady
operation at 10 Mw and 1200°F is the sum of the sbove or 1.9% in Ak/k.
In the absénce of nuclear control devices, the critical temperature
would eventually rise from 1200 to about 1400°F in the event of pump
stoppage. Also, the system would have to be heated to 1400°F to keep
the reactor subcritical during the recharging of fuel and system start-
up operations following a shutdown for maintenance. This temperature
is sufficiently above the design temperature that-nuclear control
devices are necessary during startup and to eliminate large temperature
changes that would accompany power changes during normsl operation.

The control should have a worth of sbout 4.6% in Ak/k so that the
reactor can be held subcritical down to 900°F. This will make it
unnecessary to heat the system above "1000°F during normal loading
perations and will provide some margin for varying the temperature
while at full power. Fuel will not be kept in the reactor at lower
temperature or while the containment cell is open; so no additional
shutdown margin is necessany '

Although the reactor requires a nuclear control system, the excess
reactivity is so small that its insertion at any possible rate will not
cause the fuel to escape the piping. Therefore the reactor does not
require and is not pioVided with an infallible, fast-acting safety
system; ‘ : | |

In the present reactor design, spaces are provided in the graphite
assembly for four l-in.-diam control thimbles near the center of the

core. Calculations indicate that as much as twice the reQuired control
..37;_

can be incorporated in.these thimbles if desired. Botfi;liquid and solid
control devices are being studied, and the most satisfactory will be |
adopted for this_reactor. . The control devices are hot.required or desired
to operate rapldly. _However,-it,will be . important to\havé‘accurate )
knowledge of the position of the poison at all times to provide a con-
.tinu6us,indiéation of changes in-reactivitj, Design details and. results
of full-scale mockup tests of the comtrol system will be included in

the Final Hazards Reporfi. v |

3.1.2 Other Control Features

As discussed above, the primary function of the nuclear control
system is to maintain the critical temperature of the reactor below the
system design temperature of 1300°F. Poison will be inserted if the
circulating pumps stop (see Sec. 7.1.3), if the mean temperature rises
too fast, if the power should rise more then 50% above normsl, or if the
escape of rédioactivity is detected. Simulafor studies have indicated -
that the self:comtrol features of the MSRE will not hold the power
steady at low levels. Automatic operation of the control system will be
-provided to hold ‘the power constant at any desired level up to 10 Mw..

Additional protection against the reactor temperature exceéaing
1300°F for long periods is provided by the drain system. The freeze
valve in the reactor drain line can be thawed and the fuel can be.
drained in sbout 15 min. |

3.1.3 Nomnuclear Controls

Although not associéted with nuclear safety, it is important to
keep the fuel and coolant salts from'freezing in the reactor piping.
As has been described, electric heaters with an extra reliable power
supply are provided on all circulating lines and on all vessels con-
taining salt. The heaters are always energized, except those on the
coolant salt-to-air radiator. - _

A special protective circuit prevents the coolant salt from
freezing in the radiator. Three actions are involved: shutting off
the air blowers, applying full.power to the radiator heaters, and
closing the\radiator doors to give thermal isolation. If the salt
-38-

temperature at.the radiator exit drops from the design point of 1025°F
to 975°F, the blowers are turned off and full.power is applied to the
heaters. A further drop of 50°F (to 925°F) causes the radiator doors
to close. This is about 75°F above the temperature, 850°F, at which
solids initially appear. : .

The reactor and coolant drain systems pfovide additional protection
against 'salt freeziné in the reactor or coblant systems piping in that
the saits can be drained if the temperatures approach the freezing
points.
-39~

3.2 Instruments

3.2.1 Reactor-Power Measurement

‘The reactor power is determined from the flow rate of coolant salt
and its temperature difference across the fuel heat exchanger. The flow
18 measured with a Venturi meter and differential-préssUre cell in the
coolant circuit " This equipment is located outside the reactor and can
be easily maintained during reactor shutdowns. '

Reactor power will alsoc be indicated by the neutron level of the
reactor, after the neutron indicator is calibrated by heat balances.

3.2.2 Fuel Inventory

One of the most important measurements in a mobile fuel system is
the fuel inventory measurement. Two pleces of infdrmation afe required:
(1) the total quantity of salt in the system and (2) the concentration
of U235 in the salt. The quantity of salt is determined from the mass
of fuel in the circulating system as indicated by the liquid level in
the pump bowl and by the mean temperature of the loop. To this must be
added any material remaining in the drain tanks, which are we1ghed by
pneumatlc load cells.

The concentration of uranium in the salt must be determined by
chemical analysis. Samples can be removed at any time from the sampler-

enricher device described in Sec. 2.10.

- 3.2.3 DNuclear Instruments

The nucieaf‘instrumentation is comprised of two count-rate circults
with U235ucoated chambers as the detectors and four high-level circuits
equipped with ion chambers. All six chambers are inserted in access
‘thimbles which terminate insidetthe thermal shield around the reactor
vessel. The arrangement is shown in Fig. 14.

3.2.4 Radiation Monitoring

Rediation detectors are requiredv(l) to protect against the escape
of radioactive liquids through the many service lines which penetrate

the containment shell, (2) to measure the background levels of radiation
{— SAFETY CHAMBER —={ )\

\

OUTER STEEL SHELL

CONTAINMENT
VESSEL

{—-SAFETY CHAMBER

/{‘I—COMF’ENSATED CHAMBER 4

REACTOR SHIELDING

Sn—Be SOURCE TUBE

Fig. 14. MSRE Neutron Chamber Arrangement,

UNCLASSIFIED
ORNL-LR-DWG 56878

2—-FISSION CHAMBERS
{ ~ COMPENSATED CHAMBER
1—SAFETY CHAMBER
=41
through shielding adjacent to occupied areas, and (3) to monitor con- -
tinuous samples of the atmosphere within the operating areas.and the
stack discharge° . o | - | |
The detectors which perform functions (l) take automatic safety
'action when preset safe limits are exceeded Those in categories (2)

and (3) only alarm. All indicate remotely and record so that the .
operator can immediately investigate the source of indicated radiation.

3.2.5 Pressure Measurements

Pressure measurements at various points in the cover-gas and off-
gas systems are required during operation and during salt transfers.
Seal-welded pressure elements are used in the fuel and coolant systems.
Other pressures are measured with conventional pressure devices located
outside the container. They include the lubricating-oil systems, the
water cooling systems, and the pressure differential between the con-

tainment cell and the atmosphere.

3.2.6 Flow Measurements

Tn addition to the coolant salt, flow measurements are required on
the off-gas and lubricating-oil gas streams at the fuel pump, and on
the sump lubricating-oil and coolsmt-water streams. Seal-welded dif-
ferential-pressure cells are installed to measure the pressure drops

across these orifices and capillaries.

3.2.7 Temperature Measurements

Temperature'determinations throughout the reactor plant are made by
Chromel-Alumel thermocouples sheathed in Tnconel. The data will be used
for control, for operational information, or for initiating safety .
actions. Temperature points of interest in the reactor system include
all salt containers and piping, reactor-vessel and pump-bowl areas that
might have local heating, freeze flanges and freeze valves that must be
kept cool, and many locations in the off-gas system. Also, lubricating-
oil ard cooling-water temperatures and the ambient temperature:. of the
cell atmosphere must be monitored. - Over 500 thermocouples will be used
to collect the desired information. |

3.2.8 LiquidJLevel Measurements

The level of salt in the primary pump is measured with a differ-
ential transformer actuated by a float in the pump 5ow1. The level
of salt in the drain tanks is determined by weigh cells vhich are
designed into the tank support system. The levels of other liquids,
such ‘as the 1ubrica£ing oll and the water in varlous cooling systems,
are determined by conventional floats or static pressure-head measure -

ments because these devices are located outslide the container.

_u3_.
L. REACTOR PHYSICS DATA

A l1list of reactor physics data for the currently proposed MSRE
core design (in the clean, hot condition) is given in Table 9. The
critical mass, system inventory, temperature coefficient, neutron
fluxes, and power-density estimates were obtained by using the IBM-TOk
one-dimensional maltigroup diffusion-theory program GNU3 end the 34-
group cross-section lfl.'brt—:i.rylL prepared for use in the phermal-bfeeder-
reactor evaluation program.5 The poison-tube worth was calculated
withAtheitwb4group, fiwo-dimensional diffusioh-theory program PDQ,6
using'an effective extrapoiation distance at.fihefibeundaries of the con-
trol regionsiend two-group constants preduced by the GNU calculations

elsevhere.

Table 9. Reactor Data for Clean Hot Condition, All Tubes Empty

Fuel volume fraction in core 'l 1 0.225
' Core critical mass, kg U= o 15.6
Circulating-system inventory, kg U235 | My
Temperature coefficient of reactivity, (8k/k)/°F -9 x 107
Fuel - ' : - -3 x 10~
Graphite | -6 x 10~
Mean neutron lifetime, sec | 3 x 107"
Effective delayed—neutron fraction - 0.0048
Total poison-tube worth (four tubes), % Bk/k - ke
Ffaction'of power generated in core ‘ 0.962
Per cent thermal fissions - . 86.6
Total power, Mw : ‘ 10
Mean core power density, w/em® . , 3.86
Peak core power density, w/cm® | 10.5
Mean salt pofier density, w/em® o 17.2.
Peak salt power density, w/cm® 46.7
Peak thermal flux, neutrons/cm3-sec 8.1 x 1013
Mean thermal flux, neutrons/cm®-sec | 2.6 x 1013

(8c/k)/ (/M) | - 0.23

e

5. THE REACTOR COMPLEMENT;

5:1 Building

The Molten'SaltiReactor Experiment will be conducted in ORNL
Building 7505,‘which originally.was constructed for the-Aircraflt
Reactor Experiment (AREj and iater wa.s extefisively modified to'house
the Aircraft Reactor Test (ART). The additional revisiéns required
for the MSRE are described here. | |

A plan view of the building is shown’in.Fig- 15. The reactor
contaimment cell, the drain-tank pit, and the coolant;eqfiipment pit
are located in the south end of the structu:ej pits’for fuel storage
and contaminated-equipment handling are in the nofth end. This por-
tion (the high bay) of the building is isolated from the other por-
tions by a ventilation system and is designated a "contamination
zone'" because it is likely to become slightly contaminatéd during
reactor maintenance operations. On the west Wall.is‘éOnstructed a
shielded, remote-maintenance control room from which hoists and
manipulators may be controlled. Access-to’the high bay is at the -
northeast corner through.a change»room where workers can chahge to
"contamination" clothing before entering and can be surveyed for
contamlnatlon before leaving.

East of the high bay on the ground level are the reactor control
room and the office area. On the basement level are installed auxiliary
reactor instrumentation and building service equipment.

At the south end of the building is the fan room, where the two
large cooling fans are mounted to discharge over the coolant-salt
radiator and to the stack outside.

The building has two extensions on‘the west side, where an

emergency diesel power statlon and electrical switch gear are housed.

5.2 Containment

The cofitainment philosophy which has been applied to the MSRE -
requires that a mifiimum of two barriers be provided as protection

against the escape of radiOdctivity. The first barrier is the reactor
UNCLASSIFIED
ORNL—LR—DWG 56879

OPERATING - :
/CREW CHIEF W
INSTALLATION [{NSTALLATION / | OPERATING . CRAFT
ENGINEER ENGINEER CREW
: : HEALTH
o FOREMAN
_ ENGINEER
-, INSTRUMENTATION SHOP | e
/ : LOCKERS
' REACTOR CONTROL ROOM S
CHANGE b . _ - i
ROOM. | | ZONE 11 T 5
A, ’ ] A B | o REACTOR | | )
: ] I ’ —i [l cCONTAINMENT )
- \ CELL )
: N —
R0 |
- —— C D DRAIN | . —| | RADIATOR =
CONTAMINATED ! /N, | . TANK PIT LI PIT
3 ' 1 LA | I / I_‘ . .
MAINTENANCE ~-- ‘ (MAINTENANCE | VENT
SHOP . ri| CONTROL HOUSE
- |__room _ ,
{EL.B62 '
. Lot
N #5—49_<i=: ::
' SWITCH .
, . HOUSE BLOWER HOUSE.
A-CONTAMINATED PARTS STORAGE R
B - UNLOADING: PIT .
'C - DECONTAMINATION AREA
D= FUEL STORAGE PIT .

Fig.15. First Floor Plan, Building 7503 .

‘Sfi'
-b6-.

piping assembly which was described in Sec. 2.1. The second barrier is
the seal-welded contalnment vessel described below. These two barriers
can be made éssentially leak-tight to the approximately 50 million
curies circulating within, as has been demonstrated in HRE-2; the leak
rate of the piping must be less than 1 cc/day, but the containment
vessel may leak 1%/day, as will be discussed later.

. Actualiy & third line of defense has been provided in the MSRE.
It serves principally during maintenance when the first two barriers' i
~are opened purposely. This third barrier is the isolation of the high-
bay area by means of a sealant on the sides, ends, and top of fhe
existing building. The high-bay will be kept at a negative pressure by

ventilating fans which discharge through the stack-filter arrangement.

5.2.1 Fuel-System Container Design

The reactor container is in two connected parts, but is operated
as a single vessel. The first part, the reactor containment cell, waé
designed and instelled for the ART. It is a 1 1/hk- to 2-in.-thick
carbon steel cylinder, 24 ft in diameter and 33 Pt deep, including its
hemispherical bottom. The top is closed by a flat "sandwich" consisting
of a 1/8-in. carbon steel sheet held between & lower layer and an upper
layer of 3 1/2-ft-thick concrete blocks. The steel sheet is seal-
welded to the circumference of the tank, and the upper layer of blocks
is bolted down, as illustrated in Fig. 16.

As may be seen in Fig. 17, the reactor containment cell is'érected
in a 30-ft-diam steel-lined pit. The annulus between the two cylinders
is filled with water and solid shielding material. The penetrations
through which service and electrical lines enter the container are
bridged across the annulus. The reactor vessel, the main heat exchanger,
and the fuel circulating pump are .all located within the reactor cell.

The drain tanks are installed in a second cell with an open passage
to the reactor cell. The dump or drain line is located in this passage, °
‘connecting the reactor circulating systeh to the drain tanks. The '
drain-tank containment cell is of concrete construction, rectangular in
cross-section and lined with steel sheet. The top is closed and sealed

in a menner similar to'the.reactor-cell éloSure.
UNCLASSIFIED
'ORNL-LR-DWG 56880

P n

'UPPER SHIELDING BLOCKS

| . -~ ANCHOR BOLTS\
@G0 40D an 0, 4D o OB, on, J0__a%, T 4R Gh - on AN oo Jn  d0 0 gn an

+

STEEL WELDED 17

MEMBRANE — ]

30-ft-DIA
TANK ——>»

......

FLOOR
O GO SE

ANNULUS FILLED WITH
WATER AND AGGREGATE

/—\V“"\

-Fig. 16. Shielding ond Sealing Membrdne for Top of Cell.

T~~~ SHIELDING
RING

- 24-f1-DIA REACTOR
CONTAINMENT CELL.

'Lfl'
UNCLASSIFIED
ORNL-LR-DOWG 56884

I\ : ' . i
CONTROLLED VENTILATION ZONE Ofw
STACK
CHANGE
HOUSE
MAINTENANCE CONTAINMENT CELL
SHOP

STORAGE

REACTOR
(VESSEL

DECONTAMINATION AREA

FUEL STORAGE AND
TRANSFER AREA

DRAIN TANK
CONTAINMENT PIT

(WATER \_THERMAL SHIELD
ANNULUS

Fig. 17. North-South Sectional Elevation—-Bldg. 7503
-49-

Since the drain-tank cell is connected to the reactor cell, both
cells will be pressure- and leak—testédjat the same time. When~con;? 
struction is completed but before equipment is installed, a hydrostatic
pressure of 45 psig will be applied. Later, after the reactor is. ready
for operation, pneumatic pressure will be used to determine -the leak -
rate and to lodéte the leaks. This test will be repeated-each time.
the cells are opened and closed for maintenance to be certain that -
the 1%/24 hr allowable leak rate is not exceedegd. Furthermore,_during-
operation the.cells will bé kept at a negative pressure of'2upsigi(13f
psia);; and the leak rate will be indicated continuously by. the rate’

of pressure rise.

5.2.2 Other Cells

As mefitioned previously, pits for fuel storage, loading and un-
loading, used-equipment storage, and decontamination are also located
in the high—bay area. The storage tank is provided with a secondary
container in the‘form'of another tank which surrounds the storage tank
and is designed for complete containment. The amount of activity in
the other cellsAis low by comparison to that in the reactor ceil, and
it is considered sufficient to protect against the escape of activity
by providing a strong draft through openings and by the use of absolute
filtérs to remove particuléte activity before the air is diséharged to
the stack. Approximately 8000 cfm of air is-available for the venti-

lation of these cells.

5.2.3 Penetrations

‘Each of the many service lines which penetrate the walls of the
secohdary containers is equipped with a seal or a closing device to
prevent the escape of radicactive fluids. |

All electrical and thermocouple wires are encased in metal tubing
and insulated with a dense packing of magnesiufi oxide. Leakage tests.
on this type of cable have indicated leak-tightnesé at pressures as
high as 1500 psig. _

| Water, oil, and air lines are designed with_solenoid, pnéfimatic,
or spring-loaded valves which may be actuated to close in case of back-

flow or the detection of radiocactivity. Radiation monitors will be
-50-

loaded sufficiently near each line to detect and actuéte-the closures
before dangerous radiation levels are attained. Each of :the separate
fluid service systems.is completely closed. The leakage of radioactive
material into one of these systems will constitute a contained hazard
rather than a release of activity. _

The coolant-salt lines are not equipped with closure devices. As
previously mentioned, the coolant-salt pressure is kept greater than
the fuel-salt pressure, and in the event of a heat-exchanger tube
failfire, the coolant salt-will be pushed into the fuel. Shduld the
differential pressure disappear, the reactor will be drained. Although
the-coolant;salt cell is not leak;tight, containmént protection provided
by a fldw of air maintains theficell at a negative pressure greater thgn

0.1l in. Héo, snd the air is monitored for radioactivity.

5.3 Shielding

The shielding arrangement for the réactor equipmént is shown in

Fig. 1T. The reactor is shielded around the sides and on the top by
a 16-in. iron-water (50% Fe - 50% HpO by volume) laminated thermal
shield. An INOR-8 casting located in the outlet dome of the reactor
vessel, in order to reduce the dose to the fuel circulating pump, is
also effective in reducing the dose from the reactor above the reactor
cell. ‘ |

' The reactor-cell shielding consists of T ft of ordinary concrete
covering the cell and aggregate and water in the 3-ft annulus between
the reactor containment vessel and the cell wall. An additional 2 ft
of ordinary concreté; which constitutes part of the cell wall, encloses
the cell except for the section which is adjacent to the radiator room.
The reactor-cell top shielding and sealing mémbrane were previously.
" ‘described- in Sec. 5.2.1. The drain—tank cell is covered by .6 ft of
ordinary concrete sandwiching a sealing membrane in the same manner
as that described for the reactor container in Sec. 5.2.1 (see Fig. 16).
| The penthouse (the portion of the radiator ceil extending abhove
the main operating floér) has 2 ft of ordinary concrete shielding to

provide shielding from the activated coolant during operation. The -
5] =

radiation here results from F1%(n,q)N'® with a 7.4t-sec half life and
from F'2(n,7)F2° with an ll-sec Half life and permits entry into the
radiator room a few minutes following shutdown.

. The fuel storage cell is covered by a L-ft-thick ordinary concrete
plug. This cell, the hot storage cell, the decontanmination cell, and
the fuel transfer cell are separated by B-ft-thick'ordinary-copcrete
shadow-shield walls. -

The estimated dose rate in a small area directly over the reactor
through the 7 ft Qf ordinary-concrete shielding plugs is ~90 mr/hr,
reducing to lower values in other locations over the cell. Should the
actual dose rate.berthis_high, concrete -blocks will be stacked on the
shielding to reduce the dose. In no'contifiuously occupied area. will

the dose rate exceed 1 mr/hr.

5.4 Arrangement of Equilpment

The arrangement of the reactor components in these containment
areas 1s presented in Figs. 18 and 19. The layout of the fuel‘loop in
the 24-ft-diam containment cell achieves piping stresses well below the
ASME-code allowable values and provides satisfactory accessibility for
maintenance. The reactor is anchored, but the fuel circulating pump
above 1t is supported by spring hangeré. The primary heat exchanger
is free to move horizontally and vertically; the inlet is aligned with
the pump discharge nozzle, and the outlet with the reactor inlet volute.
The components are joined by 5-in.-ID tubing (0.165~in.-thick wall), |
with freeze flanges to permit removal. ' '

The fuel drain line connects the reactor to the drain tanks located
in the rectangular container adjacent to and south of the_reattbr pifi.
The two drain tanks and the flush tank are also arranged for ma@ntenance
from overhead., The elevation of the tanks provides a head of 5 to 15 ft
of salt to drain the reactor. Freeze valves are positioned between each
tank to control the routing of the salts. The line extending:thpqugh the
north wall connects these tanks with the storage tank. ‘
UNCLASSIFIED
" ORNL—LR—DWG 52040

' —=F : CLEAR
WATER SHIELDING o NUCLEAR .
L
M — Ll
L —
} | | r | } REACTOR AND PRIMARY PUMP
|
| I l : | i PIPE SLEEVES
| | } I = } -~ FILL AND DRAIN TANK NUMBER 3
|| B
i | ' =
l 15 | ! 1
HOT STORAGE PITS } : ! l:
L— CELL I '
1l N X N !
1 } DRAIN TANK NEUTRON SOURCE TUBE/ i ”
NUMBER 2 — : WA —i
| I :
| | [ /
I ' 4 k
I t [ : g SECONDARY BuMP STACK AND AIR DUCT
Y NUM
L —=——= | FLusH TaNK] 4 P
J_L CABLE AND PIPE Ti Y s ,
S - n n 1
1 \\ /’I 0
{ \°1';" i
/ iL.—AIR DUCT !
) i PLENUM—"h
by ,f :: :
[ " :I
/ I
/! (L Ii
1
I
,l
EXISTING
FANS
C:___r
i i |1
| s BELLMOUTH INLET
' ES .
' =
! Fre BLOWER HOUSE @ |
' (o] L?_'i ./ - |
| Ql~
:: a wy I
fil /= TRACK ! L
:i E_f i
i |
L] *,
[

..a'g-
7ft Oin.

CONCRETE SHIELDING

STEEL -BARRIER

\ _

4

FLUSH
TANK

WATER
SHIELD

\

MOTOR

PRIMARY PUMP
8-in. SUCTION
HEAT EXCHANGER

—
REACTOR—1L

1'%-in. DRAIN AND
FILL LINE
N\

" EXISTING STEEL-—\

TOP OF PENTHOUSE~

UNCLASSIFIED
ORNL-LR-DWG 504104

_MOTOR

SECONDARY
PUMP

BESREh

TANK NO. 2

LOW POINT
OF TANK

¢ TEST CELL

1 AIR PLENUM WALLS

.1 Y5-in. DRAIN LINE

PIT

RADIATOR

FILL AND DRAIN
TANK NO. 3

Fig.19. Arrangement of Equipment: Elevation.

..Eg..
-5

Loading and unloading equipment is located in the pit directly
east of the storage—tank cell. Provision is made for a shiéldedlcarrier
on top of the concrete blocks covering this cell; manipulators through
. the blocks accomplish the loading and unloading of salt..

Outside the contalnment areas in the south end of the bullding is
the coolant_equlpment. The salt-to-air radiator is mounted ;n the
coolant air duct. Two 5-in.-OD (0.165-in.-thick wall) pipes comnecting
the radistor and the-heat exchanger enter the reactor container through
bellows-equipped seals. The coolanf cell is nof completely sealed,
because the N*© activity of the coolant salt does not éonstitute a
hazard and_because-the coolant—salfi préssure is kept greater than the

fuel-salt pressure to prevent the inleakage of fuel.

5«5 Maintenance _

The MSRE 1s designed for replacement maintenance instead of repair
in situ. All parts of the fuel system are located so that they'can be
placed by remote techniques,-and this arrangement also permits semi-.
direct maintenance.

Remote maintenance is accomplished from & shielded control room
from which s General Mills manipulator can be directed to do work within
the cell. After the shielding blocks are removed, the manipulator with
lighting and television cameras is positioned on a track on top of the
cell and above the equipment to be removed. The arm of the manipulator
operates the*toois necessary to disconnect the equipment. The control-~
room operator then moves the manipulator out of the way and brings the
high-bay bridge hoist into'position to 1ift the component and move it
to the storage pit in the north end of the building. The hoist can then
be used to replace the lower shielding plugs. 4'

In the semidirect approach, & mobile shield is positioned over one
block in the lower layer of conerete shielding blocks. The motorized
shield is opened to permit remofial of the block which coveré the failed
component, and is then closed, By using long-handled tools through
openings in the shield, an operator standing on top of the shield dis-

connects the component. Viewing is through lead-glass windows and
55
periscopes, The operator then retires to the shielded control room,
opens the motorized.shield, and‘remotely controls the high-bay holst
to 1ift the component and move it to the storage ceil.-.~

Figure 20 illuefrates the use of the remote manipulato; and
shows the posifiion of the'shielded control room, whieh is efiuipped
with lead-glass windows as'weli as teievision‘receivefs.

Both these methods have been demonstrated afi ORNL.to be feasible
means of maintenance. Additional experience under conditions of high
radioacfiVity is necessary'before either method could be chosen for
larger reactors.

Since the removal of components necessarily requires breaching
the containment cell and opening pipes in the fuel system, protection
agalnst the escape of radioactive geses and particles must be given
pafticular attentien. Meintenance procedures require that the fuel
be drained, the pipes flushed with clean salt, and the system cooled
before repair work is begun. When the secondary container is opened
to permit work with the manipulator.or the'motbrized‘shield, the
secondary container ie venfiilated at a rete of 10 to 15,000 cfm. If
it should be néceSsary to remove all the’shieldifig blocks, a velocity
of at least 30 ft/fiin could be maintained downward through the openirg.
'However,'the normal fiaximum opening will be only 100 ft?,'and the
velocity of air will be 100 to 150 ft/min.

When the reector pipes are epened, nitrOgen_Wili be purged into
the pipes until temporary closures can be fastened. Closures are also
attached to the-flanges on tfle failed component to prevent the escape
of ectivitj dufing removai. Similar techniques have beefi-developed
on HRE—2 and haje'proved satisfaetory;

The coolant-system pit may be entered for direct maintenance as
soon as the system is cooled. The aciivity'in the salt decays by'e
factor of 10° in 2 min, andeho other sigfiificant centamination is

expected.
-56-

_ UNCLASSIFIED
ORNL-LR-DWG 56882

1
T T T 1 =
n .
N - )
o ‘
]
| LK ";
& REMOTE CONTROL
ROOM FOR HOISTS
AND MANIPULATORS \
m
: CRANE BAY %
1 T £ ‘
REMOTE

CONTROL ROOM
AND SERVICE AREA

MANIPULATOR

ST TV EY =
hl,: n=4= ! ilLJ?
m =M=

m =l

Il SEAL - WELDED '
MEMBRANE-/

m=n=q

CELL

THERMAL

=y

SHIELD

m=
[l

—

—

CONTROL TUNNEL/";

=M=

ANNULUS FILLED WITH
WATER-AGGREGATE
MIXTURE

EYi
==

(HSI=0

REACTOR
CONTAINMENT

1

MOTOR

FUEL PUMP

INSUL ATION

| =11 ==
ZHEN =

i =

Zuy

9

Y

=
=7
l’sm =m

W=y =

7

Fig. 20. East-West Sectional Elevation—Building 7503.

-57-

6. CONSTRUCTION, STARTUP, AND OPERATION

6.1 Construction

- Although no special construction‘practices will be employed in
assembling the MSRE, many special precautions will be taken to ensure
a high-quality, clean, and leak;tight assembly. A detailed specifica-.
tion, requiring quality control better than existing commercial codes,’
has been prepared'for each cdmponent of the éysfem (seézAppendix C).
Nondestructivé inspection techniques such as ulfirasofiic.testing,.dye;
penetrant inspection, x-ray examinatién, and helium leak testing are
employed at each fabricator's plant, under the Supervisidn of ORNL
 inspectors.. Assembly at the reactor éite will be similarly examined.
After completion the separate systems will be leak—teéted using rate-
of-pressure fise and isotopic-tracer techniques. The'leékage rate
from the entire fuel-containing system must be measured to be less
than 1 cc/day. _ _

After construction is completed, a period of several weeks will

be occupied by remote-maintenance practice.

6.2 Flush-Salt Test

Tt is planned to demonstrate the mechanical performance of the
system by a several-month period of'testing with a flush salt in the
ffiel system. Each piece of equipment will be ‘examined to determine
whefher it pérforms as designed; insofar as this can be determined
without nuclear heat generation. The flush salt will also serve to
scavenge oxygen and to remove other i@purities. Another important
benefit of the flush-salt test will be the development of the operéting

skills necessary for satisfactory control of the system variables.

6.3 Startup

After the flush-salt tests are concluded, the neutron source will
be installed and counting rates will be determined on each of the two

count-rate circuits before the salt is drained. Then the drain tanks
..58_

will be loaded with fuel salt containing sbout three-fourths of the UZ%S
“estimated for criticality at the minimum temperature (1000°F). The fuel
salt will be slowly pressurized into.‘the reactor, which has previously
been heated to l2OQ°F. Only one-half of the available poison will be
inserted during this initial loading and for later fuel additions. The
partial insertion will allow experimental flexibility such as withdrawal
to test for criticality after loading. If criticality is not attained
after any addition, the system temperature will be gradually lowered to
1000°F if necessary, and counting rates will again be determined. This
procedure will be followed after each fuel addition until the reactor
is critical. Further fuel additions will increase the critical tempera-
ture and provide information on the over-all temperature coefficient.
Quantities greater than 120 g of U235 w111 be added to the drain tanks
and transferred to the reactor; quantities of 120 g or less can be added
at the bowl of the fuel circulating pump through the sampler-enricher

mechanism.

6.4 Approach to Power

After a series of zero-power experiments to determine the nuclear
characteristics of the system, the pbwer will be raised in increments of
a few hundred kilowatts over a period of several weeks until full power
is reached. At each successively higher power level, information will
be collected and fuel éamples will be analyzed for studies of xenon and
fuel permeation of the graphite, fuel stability, power coefficient,
radiolytic-gas handling, and power stability.

6.5 Operations Personnel

The operation of the MSRE facility will be the responsibility of
the Reactor Division Operations Department. The previous assignments
of personnel in this group include the construction, startup, and
operation of four other experimental reactors: the Low-Intensity Test
Reactor, Homogeneous Reactor Experiments. 1 and 2, and thé Aircraft
Reactor Experiment.
-59-

The experiment will be conducted on a three-shift basis, employing
four operatlng shifts and a day staff for- reactor analys1s and planning.
Each of the four shifts will be headed by a senior-level superv1sor for -
the flrst few months. A Junlor engineer and three or four nontechnlcal
operators many with the prev1ously mentioned reactor experience, will
complete the shift organization. |

. The Reactor Analys1s Group will be composed of four to six engineers
with a broad experience in fluid-fuel reactors. Tts function will be
principally to plan, supervise, and analyze the experimental program.

- Over the years:this organizatioh has developed training,loperatihg,.
and maintenance pfactices which especially contribute to ekperimenfal-
reacpor safety. The same methods and policies will'be applied to the

Molten-Salt Reactor Experiment.
-60_

T. HAZARDS ANATYSIS

The hazards inherent in the MSRE are considered on the basis of _
possible damage, first to the primary container and second to the sec-.
ondary container. Finally, assuming thgt activity escapes‘from the
secondary container, the danger is considered to personnel at the site

and in the surrounding territory.

T.1 Damage to the Primary Container

The possibilities for rupture of the primary'container were
investigated in several categories: (1) reactivity excursions, (2)
melting of walls, (3) failure by excessive stresses, (4) corrosion.

T.1.1 Reactivity Excursions

Excess reactivity can result from the following.unusual circum-

stances.

(a) Startfip Accident. The MSRE is started normally by transferring
hot (>1000°F) fuel into the preheated (>1000°F) circulating system.
‘The normal fuel concentration is sufficient to make the reactor chain-
reacting at 1200°F when'at full power with the full éffect of the power

coefficient and xenon poisoning. When these effects are not counter-

acting excess reactivity--that 1s, at startup--the poison must be suf-
ficlent to hold the reactor subcritical. If; by some instance, the
poison were ndt inserted when fuel is pushed intc the core, the reactor
would begin to generate power unexpectedly with the core only partly
full,‘ In the worst situation the core would continue to £ill, the

- reactor would continue to generate heat, and the reactor temperature
wofild rise to a final temperature of 1L0O0O°F. Althpugh such a tempera-
ture rise is undesirable, it should not damage the reactor.

This accident will be'analyzed in m?re detail on the reactor simu-
lator before the final design of the conirol circuits.. It is planned
to protect against the startup accident in several wajs, For routine
startups, the control circuits will require that all the poison be
inserted before f£illing can be begun. A large number of thermocouples
_61 -

distributed on-the reactor vessel, throughout.the heat removal system,
and on the drain tank must indicate at least 1000°F. The rate at which
fuel can be transferred from the drain tanks to the reactor vessel will
be limited so that the loading time will be approximately one hour. The
reactor will be filled in several steps, with sufficient delay between
steps. for the neutron multiplication to be determined. The reactor can
easily be drained if eriticality is approached unexpectedly, because
the drain line is also the fill line.

- When the system has been filled to the operating level, there will
be a time delay before the pump cen be started so that the drain line
can be frozen and the temperature distribution in the reactor system
can be checked. Any lafge temperature differences between the reactor ;
and the heat‘removal system will be reduced by natural-convection circu-
lation during this delay time. ‘

A variation of the startup accident would involve filling the
- reactor and withdrawing poison to make the reactor critical at the
operating temperature of 1200°F before starting the pump. Assume that
the heat removal system is operating so that all the fuel in the heat.
exchanger is cooled to 850°F., If the circulation pump could be started
to cause the 850°F fuel slug to traverse the core at the average circu-
latiqn rate, the reactivity would increase at the average rate of 0.15%
2Mk/k per second for 7 sec. Actually thermal-convection circulation
begins when the fuel in the heat exchanger is cooled so that the
reactor gradually rises in power to satisfy the demand. According to
reactor simulator studies, thermal-conveetion circulation is sufficient
to extract 9.4 Mw with a temperature rise of 200°F across the reactor.
The studies do not indicate a likelihood of damage from the cold- -slug
accident. ' Nevertheless the fuel pump will be started et.reduced speed,
and the rate at which the speed can be increased will be regulated to
limit the rate at which reactivity can be added by introduction of
cold fluid. | -

(o) Graphite Problems. Four potentiasl reactivity problems are
associated fiith~the3presence of bare graphite: ' | o
1. compatibility with the fuel salt,
2. fuel penetration into graphite voids,
62-

3. 1irradiation-induced shrinkage,
4, xenon penetration into graphite voids.

If the salt and graphite were chemically incompatible at any tem-
perature above the liquidus of the salt, there would be concern sbout
reactions which might result in uranium'separation. Since no reaction
has been observed in several thousand hours of loop tests or in labora-
tory studies, there 1s considerable confidence that none occurs, and
that graphite and salt can be judged completely compatible.

An associated pfdblem is that of penetrétion of fuel into the T%
of accessible volds in the graphite. Many out-of-pile experiments have
been done on the wetting and permeation of graphite by molten salts.
The tests indicate that graphite is not wet by the MSRE fuel salt and
that permeation of MSRE grades of graphite does not exceed 0.5 vol %
at pressures as high as 150 psi. One in-pile test has been done and
it indicated that radiation effects do not change this behavior. Both
in-pile and out-of-pile testing will be continued to determine whether
any condifiions that can be produced in the reactor will cause wetting
and appreciable penetration of the graphite by fuel. It is believed
that the tests will continue to show no significant penetration of
salt into the pores of the graphite. However, in spite of this infor-
mation, operation of the reactor will be monitored fdr long-term effects
in the large mass of graphite at high radiation levels. Permeation
greater‘than 0.5 vol % should occur slowly, and the accompanying rise
in reactivity would be slow. This would be indicated by a gfadual rise
in critical temperé.ture° The increase in reactivity can‘5é easily
counteracted by the control system, by omitting fuel addi#ions to com-
pensate for burnup, or by adding lithium or thorium tolthe fuel as a
poison. The question of graphite compatibility is discussed nore
fully in Appendix F. ‘

‘The only potentially hazardous situation that cculd result from
several per cent of fuel soaking into the graphite porés exists during
maintenance. The decay of fission broducts in the graphite could raise
the central graphite temperature to about 2000°F in about 200 hr after
shutdown. If the core vessel were opened in air with the gréphite at

2000°F, some burning could occur, the quantity depending on the amount
63-

of oxygen available. Undoubtedly the evolution of CO and COp from burning
would carry fission products into the reactor cell. - It is unlikely that
more than 1% of the graphite would burn or more than 0.06% of the fission
- products would escape the reactor core before thesreaction could be stopped
by closing the vessel. The rapid ventilation of the secondary container
"~ would deliver the activity to the filters; only noble gases would be.ree
leased to the stack; so the biological hazard would not be significant.
Protection against this mishap. is conceived to be:.. (1) thorough cooling
of the graphite with flush salt prior to beginning maintenance work,
(2) developing practices for rapid closure of lines which might cause a
"chimey" effect through the hot core, (3) insistence on good ventilation
and monitoring so .that the maintenance work can be halted and closures
can be made to’ stop the escape of activity, and (4) a purge system of
nitrogen or helium which provides a blanket of gas in the contalner and
rednces the likelihood of the entrance of air.

iIrradiationwinduced shrinkage of the graphite was studied to determine
its effect on neactivity.‘ Because the shrinkage results from neutron bom-
bsrdment, the flux variations across the reactor produce dimensional
changes which vary with position. This results in stresses within the
separate graphite pieces, and bowing as well as axial and transverse
shrinkage. The resultant stresses may be considerably relieved by creep
and annealing, but even without these mechanisms, the.graphite should
not begin to form cracks in less than two years' exposure at 10 Mw. As
shrinkage gradually occurs, the space will £i11 with fuel salt. This_
process is so slow that the reactivity increase (which will result if
 the spaces fill with fuel salt) will be almost exactly counterbalanced
by the buildup of long-lived fission product poisons. If there is.a
change in reactivity, it is predlcted to be slightly negative. Because
of the very slow rate of change, compensation can be easiiy accomplished
by adjustment of the rate at which fuel is added to the system to
compensate for burnup. | | | '

The description of the core graphite assembly (Sec. 2.3) included
a mention of molybdenum bands which restrain the deformed graphite.
Originally the arrangement of fuel and- graphite was such that failure
of the molybdenum bands would cause e‘reactivity excunsion, but with

the present design, any reactivity shift would be negative and not
-6h-

significant (0.1% Ak/k). The molybdenum could be omitted from the present
design, but 1is being retained as an experiment because the metal might
be used in future reactors. '

Thus it appears that neither breakage of the bands nor slow shrink-
age of the graphite bars will be readily detectable while the reactor is
operating with the original charge'of fuel. However, & possibly hazardous
situation would arise if either change took place and the initial fuel
charge was replaced with new fuel. Without the fission product poison of
the old fuel, the Same concentration of uranium in the new fuel would re-
sult in a reactivity approximately 4% greater than the original. This
is but one reason why a reloading of the reactor, should it ever be neces-
sary, will be handled as carefully as the original critical experiment.
The precautions for that experiment (see Sec. 6.3) should protect against
acclidents in future loadings.

Xenon penetration into the graphite voids will occur by transfer from
the circulating fuel and/or from any fuel which has penetrated the graphite.
For 0.5% fuel penetration thé total xenon effect at 10 Mw is estimated to
be 1.3% Ak/k. At its peak after power reduction the poisoning will amount
to ~4% Ak/k. Xenon will be removed by the stripping system as long as
the fuel is circulated. No excess reactivity will be provided to over-

come xenon buildup on reduction of bower.

7.1.2 Fuel Separation
One of the few weaknesses of the fuel composition selected for this

experiment is its vulnerability to large amounts of oxygen as gas or in
compounds. The ZrF4 component of the fuel is for the purpose of reacting
fiith‘any oxygen. and thus'preventing the precipitation of U0z. Extensive
laboratory tests have shown that eo;long as the ratio of'ZrF4/UF4 is 3

or greater,_Zan iz always precipitated in preference to U0s. The actual
ratio of ZrF,/UFs4 in the fuel is 5, which gives a good margin of safety
-to maintain the'fuel within its known limitsvof stability. Approximately
2.5 £t3 of water or 7000 ft2 of air (ST?) would have to react with the
salt to precipitate the excess zirconiumd These amounts are considered
to be very large in view of the care being used to prevent the system
from becoming contaminated. The periodic sampling of the fuel should

- reveal the presence of centaminents long before this level is reached.

More informstion on the fuel chemistry is presented in Appendix A.
-657

In spite of the apparently good protection by ZrF4, and in spite of
-the extraordinary.efforts to keep the system free of oxygen and water
vapor, the consequences of U0z separation must be‘examined.- After pre-
cipitation, U0- would tend to remain suspended and circulate with the
fuel. If it were uniformly distributed, it would be indistinguishable,
nuclearwise, from the normal fuel. However, ‘after a while it would be-
gin to collect in low velocity areas or at the points of lowest tempera-
ture, and the most likely location 1s the reactor-vessel plenum under
the graphite. At this location, 2.5 kg of UQ- deposited per square foot
of surface would incresse the temperature only 100°F. | ,

The worst possibility 18 to assume the sudden transfer of the sepa-
rated material into the core. The resulting excursion would depend
among other factors, on the quantity of U 235 involved; so it is worth-
vhile to estimate the limits of detectable fuel loss.
| Assuming that a 25°F reduction in critical temperature could be
easily detected, the equivalent loss of U3 yould be 160 g from the
core or 625 g from the entire fuel system. If this meterial collected
in the reactor-vessel plenum it might all be returned to the core at a
rate which would be limited by the T-sec fluid transit time through
the core. The system temperature would rise temporarily by 100 tollSOPF
but no damage shaild result. The control system would normally eliminate
most of the temperature transient. .

Frequent analysis of the fuel and calculation of the U235 inventory
will be enother check on uranium separation, but the limits of detection
by this method are approximstely +3% of the fuel inventory, which amounts
to 1500 g. Since Zr0O- appears in the salt before U0z and 1s -easily -
recognizable,‘an inspection of the frequent salt samples will provide a

warning vhen oxygen enters the system.

7 1. 3 Flow Stoppage

Several nonroutine situations, ranging from probable to nearly in-
credible, that involve flow stoppage in the feed or coolant circuits
have been analyzed on an analog computer. ‘The necessary protective
actions and potential hazards are discussed below ' -

(a) Fuel-Circulation-Pump Failure. Failure of the fuel circulation
pump is highly credible because it could result from an electrical failure
in the circuit or pump motor, or from a mechanical defect, such as a

- ~66-

bearing failure. Instantaneous introduction of the delayed neutrons
normall& genérated'out of the core amounts to O,lfi%fiék/ko_ All this
amount is not effective because of the gradual stoppage of the pump and
the exponential decay of the precursors; the expected temperature rise
is less than 150°F. Failure of the pump causes automatic insertion¥of
poison. Convectilve circuiation is adequate to remove the afterheat, but
if the reactor'tem@erature continues to increase, the galt systems will
be drained. |

. (b) Coolant-Pump Failure. Failure of the coolant pump is also
highly credible and for the same reasons. Agaln the pcison is auto-
matically inserted, and dréinage of the salt in both systems 1s con-

venient protection.

(¢) Simultaneous Pump Failures. Simltaneous failures of pumps

in the fuel and the coolant loops is also credfble, because a power out-

age or a burnout of a main power bus would stop both motors. Automatic
action of thé poison is the same as.'before° A power outage automatically
starts the diesel-generators (within 15 sec), and operation could be
resufied. For other conditions which result in pump stoppages of long
duration, thermal-convection flow in the two systems is: sufficient to
handle the afterheat, and both loops can be drained.

(d) Flow Stoppage in Fuel toop. Flow in the fuel loop might be
stopped by plugging somewhere in the circuit; this 1is incfedible as an
instantaneous event and not very probasble even ag a gradual occurrence.

In the case of gradual plugging, as indicated by the temperature drop
across the heat exchanger, the reactor could be shut down and drained
rbutinely. In the unlikely event of instantemeous flow stoppage, the _
situation would be similar to that already mentioned for fuel-circulation-
pump failure. The increase in neutron flux would dutomatically cause

the polson to be inserted, shutting down the reactor. The reactor

would be drained, because no cooling would take place and the afterheat
would cause the reactor temperature to rise. '

In the situations considered above, there would be no” concern if
the fuel could be drained. Thus the most hazardous situation is a
plugged drain line combined with Interference with heat dissipation
67~

to the radiator. Any one of four possible‘circumStancesmcould~produce‘:
these conditions: fuel circuit plugged or partiaslly drained, the cool-
ant cifcuit»plugged or partially or completely drained. The power
productlon from afterheat within the core and the bulk mean temperature
of the fuel, graphite, and INOR-8 associated with the core vessel are
shown in Fig. 21 as a function of time after reactor shutdown. With
the reactor heaters off the temperature would rise to 1500°F in 16 hr,
to a maximum of 1800°F in 105 hr and then would slowly decline. The
reactor pfobably would be damaged and would have to be replaced if held
~ at 1800°F for several hundred hours. Methods of providing cooling to
maintain the temperature below 1500°F in an_emefgency are being con-

sidered for incorporation in the final design.

7.1.4 Control-System Failure

The“control system is not a safety system and is not required to
protect the reactor against calamity. Although relisbility will be an
impbrtant»criterion in designing the control system, the consequence of
complete failure was exemined. With the small amount of excess reac-
tivity present, sudden removal of all the control poison would result
in a final temperature near 1400°F. It is unlikely that the thermal
stresses produced 5y this sudden rise in temperature would cause
seriofis,damage to the assembly, or that salt would be spilled into the

reactor cell.

7.1.5 Drain-Tank Hazards

Two possibly hazardous problems are encountered when the drain tank
is filied with fuel salt from the operating reactor. These prdblefis are:
afterheat and potentially-critical fuel configurations. If the fuel is
drained a short time after shutdown, provisions for removal of after-
heat are necessary. Without heat removal, the temperature of the salt
would rise approximately 500°F in a 3-hr period starting 1/2 hr after
shutdowh, assuming the reactor had been ppérating af full powér for
1000 hr prior to shutdown. This condition is avoided by pfoviding a
100-kw heat removal system (see Sec. 2.7.1) to keep the bulk mean tem-
perature of the salt below 1400°F during storage in the tank T
POWER GENERATION (kw)

150

UNCLASSIFIED

ORNL-LR-DWG 56883

TIME AFTER SHUTDOWN (hr)

100 / < TEMPERATURE
4 POWER GENERATION —
20 30 40 50 60 70 80 90 100 10

1800

{700

1600

1500

1400

1300

Fig. 21. Afterheat Power Generation and Temperature Rise of Core Vessel vs. Shutdown Time.

BULK MEAN TEMPERATURE (°F)

_'89-
-69-

Water was selected as the coolant because of the relative simplicity
of the associated cooling system.~-The water and the salt are never in
contact with a common wall, as is evident from Fig. 9. Water is fed
through the center tube, and the steam forms in the annulus. The circuit
.is completely closed and requires no pumps. In an emergency, the steam
can be vented up the stack and fresh water added to the system. An
emergency reservoir is installed to provide cooling for 6 hr.

Several configurations of fuel inside the fuel drain: tank were
investigated for the possibility of some configuration having an effec-
tive multiplication constant greater than unity

The first configuration investigated was the flooding of the cell
with water, which is a possible measure in case all other cooling of the
drain tanks fails. The water would then act as a neutron reflector
around the drain tank. The k oo for this situation is O. 852. (The
Kope Without the water is 0.826. ) Flooding, therefore, does not present
8 criticality hazard. '

- The second problem considered was the precipitation of the uranium
by an oxidizing agent. If air should accidentally enter the drain tank
through the helium blanket system or an air leak in the drain-tank
| vessel and the salt should maintain contact with sufficient oxidizing
agent, the uranium, thorium, and zirconium would be precipitated. The
precipitate would form a semisolid mixture with the salt, and the LiF
and BeF» would exist as a liquid above the semisolid. The k. pp calcu-
lated for this configuration was only 0.185. Precipitation of the

uranium, therefore, does not create a criticality hazard.

7.1.6 Other Poeeibilities for Primary-Container Damage

There are several other accidents in which the integrity of the

primary containment might suffer.

(a) Freeze-Valve Damage.  The freezing and thawing of the freeze

valves could conceivably result in a rupture of the piping. This is
specifically minimized in the design by making the freeze-valve section
as short as possible so that there is a small danger of bursting as a
‘result of-efifirepment of liquid between the ends of a plug. Because
-T0=

the salt expands-as it thaws, special precautions are taken to apply
the heating so that expansion space is always avallable. A single

freeze valve has been frozen and thawed more than 100 times without
apparent damage.

(b) Freeze Flanges. The flanged joints in the fuel circulating
loop are also possibilities for failure. Freeze flanges were selected
as the simplest and most reliable Jjoint aveilaeble. The strength of the

bolting and the flange compression menbers are considerably in excess

of the strength necessary to maintain s tight joint. However, an

analysis of the original flange design8

suggested several improvements.
The recommendations of this study were incorporated in a new design
which will be thoroughly tested prior to use in the reactor. The old
flariges fiere cycled more than 100 times without any evidehce of failure,

and the new type is expected to be better.

(c) Excessive Wall Temperatures. Overheating of pipe and vessel

" walls might occur from the external electric heaters or from internal
ganma heating. The heater elements have a melting poini several hun-
dred degrees above that of the INOR-8, but it is not considered likely
that the INOR-8 could be melted by external heating as long as salt is
present inside the pipes. If salt were not present, the pipe wall
might melt before the heater element, but in this csse a small fraction
of the activity would be released. | |

During reactor operation various components are exposed to high .
gamma fluxes which result in gamma heating of the components. If this
heating should produce large temperature grasdients, excessive thermal
stresses may arise. The effects of gamma heating on the core vessel
and the grid structure vere investigated because these structures are
in regions of the highest gamma flux. |

The gamme heating of the core vessel results in a temperature
difference of only 1.3°F across the vessel walls and produces a calcu-
lated thermal stress of 300 psi, which is not serious.

The support grid structure experiences a.temperature difference
of 3.8°F across each grid. The resulting thermal stress is calculated
to be 850 psi. |
-1~

Gamms. and beta heating of the top of the pump bowl results in a

thermal. stress of 17,000 psi at a temperature of approximately 1000°F.
A static pressure stress of approximately 4000 psi exists at the same
point, resulting in a combined stress of approx1mately 21 000 psi.
The meximum allowsble static stress at this temperature is 16,000 psi}.
However, the ASME Unfired Pressure Vessel Code allows the combined ]
static and thermsl stress to be as much as 1.5 times the maximm allow-
sble static stress (24,000 psi at this temperature). Hence, excessive
thermal stresses do not exist at the junction in'qnestion. ' '

Thus a wall failnre as a result of beta-gamms heating does not
appear reasonsble., |

Another possibility for melting of a primazybcontainer wall is in
the fission—product adsorption beds in the off-gas system. There have |
been'1nstances9 where carbon beds became ignited in the presence of :
oxygen and consumed a portion of the charcoal in the beds. This acci-
dent is much less likely in the MSRE because of the special efforts to
exclude’oxygen-from.any part of the reacfor system. Furthermore, the
high temperatures necessary for ignition could not normally cceur )
because the beds are submerged in a pool of cooling water. In the.
unlikely event that ignition does occur, the resulting high bed tem-
peratures will alarm, and the inlet and exit valves will be closed.

The blanket of CO- resulting from the fire will extinguish the fire.
As final protection the beds are enclosed in a secondary container.

() Excessive Stresses. In a normal thermal cycle the témpera-

ture‘of‘the reactor varies between TO°F and 1300°F. With such & range
there are possibillties for excessive stresses as the piping expands o
and contracts. Particular attention has been given to providing a
flexible layout. Analysis of the extreme conditions indicate that

the maximum stress caused'by expansion or contraction is only 7050

psi. Instiumentation is provided to observe the normal rate of heatup
or cooldown, which will not exceed 100°F/hr. | |

(e) Corrosion. Another posaiblllty for failure of the primary

container is by corrosion. As reported in detail in Appendix A, the
corrosion rates experienced W1th the INOR-8 alloy have been very low

-

(less than 1 mil/yr) for periods as long as 15,000 hr. All available
evidence indicates that corrosion is not likely to be a cause of
piping failfireé.

Numerous in-pile capsules and two L0O-hr in-pile circulating corro-
sion tests have been examined for evidence that corrosion under irradi-
ation was different from that out of pile. Particular attention was
paid to the possible effects of free fluoride. No evidence was found
that indicates that high irradiation altered the normal corrosion

pattern.

7.1.7 Detection of Salt Spillage

The escape of activity from the primary container would be detected
by radiation monitors. If the spillage were into the secondary con-
tainer, the activity would be indicated by-monitors on a system which
cohtinfiously samples the cell atmosphere at several locations. Leakage
into a service line (e.g., the cooling water) would be detected by
monitors attached to the line just outside the cell.

In eithér_case, the action of the monitors would be to stop power
removal and insert poison. The salt would be drained unless the leak

were in the drain tanks.

(.2 Rupture of the Secondary Container

Assuming that radioactifie material has escaped the primary contain-
ment and has spilled into the secondary containment, the next concern is
preventing the escape of the activity from this second barrier. Two means
by which the secondary container walls might be ruptured are by missiles

and excessive internal pressure.

7.2.1 Missile Damage

It has not been possible to devise a situation in which damage by
missiles appeared likely. The maximum pressure expected in thé reactor
system is less than 100 psig, and the INOR-8 material is very ductile at
the normal operating temperature. No very large pressure excursion can

be envisioned without assuming that the large inlet and exit lines are

~73=
both frozen.. In any caée, the vessel and component sections gre rela-
tively thin.. Furthermore, the vessel is completely -surrounded by the
steel thermal shield, which is good protection against the possibility.

of missiles tearing through the wall of the container.

7.2.2 Excessive Pressure

(a) Salt‘Spillage.: The spillage of the salt at a high temperature

does'have pqséibilities‘for raising the ceil_pressure to high values.
A rapid spill into the cell of all the salt in both the fuel and coolant
systems would heat the Qell atmosphere sufficiently to produce a 2.fiApsig
final pressure. If the fuel were released as a fine spray,-the‘maximum
' pressure would be 16.L psig.

 The worst situation would be the simultaneous release of the salt
and the inleakage of the correct amount of water to allow the generation
of steam without subsequent cooling from additional inleakage of water.
Taking into account the large heat capacity of the secondary container-
and reascnable heat transfer coefficients, this worst accident wouid-pro—
duce a peak pressure of 39 psig, which is less than the U5-psig test
pressure of the container. ‘

The details of these calculations are presented in Appendix E.

(b) Oil-Line Rupture. The fuel-pump lubrication system contains a

maximum of 28 gal of oil, which, in the event of an oil-line rupture,
could come into contact with the hot pump bowl and reactor vessel. _With
an atmosphere containing thé normal 21% of oxygen in the cell,‘the oil.
~would burn, producing an excessive pressure 1n the cell, or would form
an explosive mixture which might later be 1gnited.

To ensure against containment damage by these possibilities, the .
oxygen content of the cell is kept below 5% by dilution with nitrogen.‘
Nitfogen'will be fed into the cell'continuously to maintain the oxygen

content below this level.

7.2.3 Acts of Nature

(a) Earthquake. Information on the frequency and severity of earth-

quakes in East Tennessee has been obtained both from-Lynch (letter from

J. Lynch to M. Mann, Nov. 3, 1948, quoted in A Report on the Safety

<Th=

'. Aspects of the'HomOgeneous Reactor Experiment, ORNL 731) of the Fordham

University Physics Department and'from Moneymaker (B. C. Moneymaker, a
private communication to W; B. Cottrell,_Oct. 27, 1952) of the Tennessee
Valley Authority. Bofh'souices indicated that such shocks as occaéionally
occur in the region are gquite common in the world and do not indicate undue
seismic activity. Consequently, earthquakes should be of little concern
in connection with the MSRE. _

The TVA,records show that the Appalachian Valley from Chattanooga
to Virginia has an average of only one or two earthquakes a year. Further-
more, the maximum intensity of any of these shocks is 5 on the Woods-Neuman
scale. This intensity is barely noticeable by ambulatory or stationary
individuals. For any one location, such as Oak Ridge, the expectancy of
an earthquake would be'oné in every few years.

The Fordham University records indicaté-even lower quake frequency;
however, the severity of the observed quakés‘is the same. Lynch further
concluded that "it is highly improbable that a major shock will be felt

in the area (Tennessee) for several thousand years to come.”

(b) Flood. Due to the topography of the MSRE site, flooding is highly
unlikely. The primary and secondary containment vessels are sealed to keep

watervfrom the fuel.
T7T.2.4 Sabotage

Severe damage to the reactor by sabotage would be difficult; arson
is probably the best possibility. The consequences of fire are minimized
by providing fireproof structures. Water for fire protection is supplied
to the building sprinklers by the main line from X—lO,'backed u@ by an
on-site sto:age tank. It is concluded that only a pefsoh with intimate
knowledge of the reactor would be capable of inflicting serious damage.
This possibility is minimized by adequate personnel policies and strin-

gent security regulations, particularly with respect to visitors.

=15~

7.3 Consequences of. Radioactivity Release from
the Secondary Container

7.3.1 Rupture of the Secondary Container

The incredible accident in which the container is ruptured and .a’
large fraction of the fission-product activity escapes is not studied-
again in this report. It was analyzed‘in great detail for the case of
the 10:Mw HRE-2 and the 60-Mw ART, both of which were assumed to be
- located in the same velley at ORNL. In each case it was concluded that
the consequences would be catastrophic. The same conclusion would apply

in the case of the MBRE.

7.3.2 Maximum Credible Activity Release

The leakage experience with other reactor containers is good evi-
dence that a loss of l% per day of the total volume is an attalnable -
leakage rate, and this mate is assumed for the MSRE at the 39-psig
pressure reported in Sec. 7.2.2(a).

Other assumptions are necessary to arrive at the quantity of rédiou
activity which might be released from the container imtolthe'building.
The isotbpes considered for the maximum permissible exposufe of 25‘rem»;
are the 35 selected by T. H. J. Burnett in the ART Hazards Repor‘to2 |
The building volume above the reactor cell is 13.6 x 10° liters: ‘Ten
per cent of the solid fission products and 10% of the iodine are assumed
to eséape the salt and to be dispersed in the seéondary_contginerc o
(Actually, on the basis of laboratory-scale experiments, there is good
evidence that only'l% of the iodine would be released.from‘ the-salt‘)-
Furthermore, the assumption is made that 90% of the iodine condemses-on
the. container walls, as indicated by the British Dido experiment,fiAEREfR—

3412, Removal of Radiocactive Iodine Vapor from Air). The container pres-

sure and the fission-product concentrations are considered to be at the.
maximum throughout the escape period. A final assumption is that a
person inside the building would be breathing at the abnormally high

rate of 30 liters/min.
-6~

The quantity of solid fission products air—borne‘in the cell is
calculated to be 6.8 X 10° curies. At the 1% per day leakage rate the
maximum permissible intake of the bone-seeker group (133.6 uc to give
25 rem) would be reached by a person in the building in 5.1 minutes,
if the bfiilding ventilation is not operating.j With the normal air
change every 12 minutes, a 12.7-minute occupancy would be tolerable.

On'the basis of the iodineé, sihce a 278—pc intake is allowed
(for 25 rem), a 20-min escape time is;allowed without exceeding the
maximum permissible intake. |

. The noble gases xenon and krypton, amounting to 0.87 x 10° and
2.88 x 10° curies, respeétively,‘afe assumed to be released 100%. The
leakage'into'the building would be 2;6 c/fiin which is only about one-
hélf as much as the mbre~importan£'bone—seekers. | |

| Thus it is concluded that the 7503 building personnel would have

sufficient time to escape without exceeding the 25-rem maximum per-

missible dose. -

7.3.2 Beryllium and Fluorine Hazards

(a) Beryllium. Assuming that the volume of fuel salt in the primary

loop is 59_ft3, the calculated weight of beryllium in the primary system
is 178 kg. If the beryllium is released by the same-mechanisfi as the
solid fission products, lO% of the total beryllium is released fiith the
escaping fission products; Therefore the beryllium source is 1.78 g/day
for the maximum credible accident, 0.89 kg/day fbr reactor-cell rupfiure,_
and 178 kg for feactor explosion. | _

The beryllium tolerance level is 2 pg/m® for steady éxposure and
25 pg/m® for short exposure (8 hr or less) (C. O. Smith, Reactor |
Materials Engineering for ORSORT Students, Vol. II, p 663, July 1958).

If the air inside the building (assumed volume is 1.36 x 10% m®)
is changed every 12 min, the beryllium'concentratiOn wbuld be only
1.13 pg/m® dquring the maximum credible accident. Hence, there is no
beryllium hazard in this case. _

In all cases the release of beryllium is accompanied by the release
of fission products, and the hazards assoclated With'the fissibfi products

are far more serious than those associated with the beryllium.
T7=

(b) Fiuorine. Fluorine gas is not released under ény circumstances
of salt spillage. In the case of salt-water contact, some HF is formed,
but the quantity is not sufficient to be hazardous to personnel or to

corrode the container significantly.

7.4 The Site

The MSRE site, the 7500 area of X-10, was previously approved for
the ARE and ART. Two operating reactors, HRE-2 and the Tower Shielding
FacilitY, are nearby. The High-Flux Isotope Reactor is to be built on.
a site 1 mile frofi the MSRE, and the EGCR is under construction 3 miles
away. The meteorology, climatology, geology, hydrology, and siesmology
of the site are reported in detail in ART and ARE hazards summary
reports’’? and are not changed. Population distribution and industrial
installations changed little in the intervening time. The potential
hazard in the event of a disaster is less serious for the MSRE than it
was for the ART: the power is less by a factor of 6; the coolant salt
of. the MSRE is nof explosive in contact with the water as was the ART
codlant, NeK; and the lifetime of the induced activity in the fluorine
is very short compared to that in sodium in the NaK. :

For these reasons, the data included in the previous hazards reports
are not reproduced here, but reference is made to them and to a report
on the meteorology of the Oak RidgelArea.lo

Maps of counties surrounding the Oak Ridge Area and of the area
surrounding Building 7503 are shown' in Figs. 22 and 23.
-78-

UNCLASSIFIED
DWG 17019 R

K E N T U C K Y

/ .?ELLICNJTHRAS
°NEWCOMB
. 1
o ONEIDA . OELK VALLEY o ' |
WESTBOURNE
\-lELENWOOD C A M
o Fs

. FINCASTLE
oHUNTSVILLE 36\ °
- SCOTT LAFOLLETTE o< ¢
. HIGHLAND PARK®
©ROBBINS : JACKSBORO
) ‘ *

A Y
o]
CARYVILLE
SUNBRIGHT, / i \
r( ‘ ' b | . LAKE CITY : LUTTRELL :
ML . © NORRIS > 4{

. 20 -/ STAINVILLE
KN O X \ wmascor T
]
FOUNTAIN CITY

] WINDROCK

\ BRICEVILLE
¥ GOBEY, FORK MT. SER s ON v
O WARTBURG g peTROS N A N CLINTON &
{+]

OLIVER e
°MAYL‘.‘ND M O R G A I(?JOALFIELD ) ; SR[NG o WHITTLE SP \
CUMBERLAND . W KNOXVILLE
CROSSVILLE \ /‘9 ////‘
e : \ . HARRIMAN /’4’7\""

CRAB ORCHARD S. HARRIMAN

»EMORY GAPoO

g AROCKWO0D
; A

NEUBERT
o

~

KingsTon HFIR4

<]
MARTEL ROCKFORD
LENOIR CIT “ \
i o QD °
/

! 1RING CITY
’ o

oS\v‘-'EETWATEFl

oMADISONVILLE

MONROE

-~ Q D
) o MOUNT VERNON Q) | \/\
© TELLICO PLAINS' < O ,
/ e
B4
Fig.22. O

MAP OF CITIES AND COUNTIES
SURROUNDING OAK RIDGE AREA

10 0 10 20 30 MILES
= = T S e o s e e |

o 0
§ E g, o S
o -
“‘__) w W " 1 :
: NI Oe
} 2
- [ ]
X g “ ’ D '9
: . M N-18,800
DIESEL ¥u‘sg
TAN K.
x
GENERATOL
0
H_\_J,fi,_fi_/ 1 N- 18,700
| JL1 1750
lSWITEH uoufii/flHE. /"‘
&
N-18,000
‘ CoMPLESSOn
HLOWENL WOLUSE-S
CHALCOAL PIT-
]
/ N-l&,soo
TSNEW
P -
T BLO
N-18,%00)
N-18,200
N-18,100

_PLoT PLAN

SCALE "= (00"

NOT CLASS|\FIEDR

PLOT PLAN
MOLTEN SALT REACTOR EXPERIMENT
BUILPING 7503 |

-80=

Appendix A
CHEMISTRY AND CORROSION

Tuel Composition and Stability

The MSRE fuel 1s a mixture of molten fluerides with the composition
LiF-BeFo-ZrF4-ThF4~UF, (70-23-5-1-1,mole %). The latest experiments |
' show that all three quadrivaelent cations behave similarly and that 15%
of these cations can be contained in a liquid solution of LiF and BeFo
at 40°C (824°F). Research is being continued in order to have a com-
plete phase diegram for the system. ‘ i

The purpose of ZrF4 in the fuel 1is to prevent the uranium from pre-
cipitating as U02 in the presence of small amounts of an oxidizing agent.
When solid BeOo is added to a molten mixture of LiF, BeF,, ZrF4, and UFy
with a ZrF./UF, ratio of 2, UOs is precipitated and no ZrO, is found.
When the ratio ZrF4/UFs is increased to 3, ZrO, is precipitated and no
UO» is found. This indicates that the reaction UFs + ZrOp 2 ZrFy + UOp
has an equilibrium constent between 2 and 3; that is,- ‘ -

(zrF, 1 (U0, ] ZeE,]
K = . - = . ~ 2.5
[UF), 1lzr0, | [uqJ ~

(the activities of Uo,, and Zr0,, both being solids, are taken as one)
- Therefore, to prevent precipltatlon.of vo,, a ratio ZrFL/UEh of not less
than 3 is necessary. The actual ratio in the fuel 1s 5, which gives a

good safety margin, and maintaips the fuel well within its known.limits
of stebility. o |
~ The vapor pressure of the fuel salt at operating temperatures is
about 10 -2 mm Hg, as shown in Fig. A.l.
The liquidus of the fuel salt is 440°C (828°F), At this tempera-
ture a solld phase appears having LiF, BeF,, and ZrF), in as-yet-unknown'.
proportions. At 431°C (800°F), an edditional solid, 2LiF:BeF,, forms.

The next known phase appears at 425°C (797°F). This is TLiF-6(U,Th)FL,

PRESSURE {(mm Hg)

1072

-81-

UNCLASSIFIED
ORNL-LR-DWG 56885

—-——————l———-—-—-—-—-—iatm.———-—-"—-— — T ——

l0g,q #= 32—

15,450
—Tk_ —-6.039 logy, 7;

P=PRESSURE {mm of Hg)

7= TEMPERATURE {°Kelvin)

1000 2000 3000 4000 5000 6000

TEMPERATURE (°K)
l A I | | | J { | i I
1000 2000 3000° 4000 5000 6000 -
TEMPERATURE (°C)
| L I l ! | ! } | | P
1000 3000 5000 7000 9000 11,0600

TEMPERATURE (°F)

Fig. A.A. Estimated Vapor Pressure of Fuel 3alt.
-82-

which is 15 mole % U, 85 mole % Th, Thus it may be noted that there is
no tendency for uranium to concentrate at the freeze flanges or freeze
valves. |

The compositions of the fuel andlcoolant salts are matched so that
it is impossible to freeze the ffiel salt by removing heat through the
coolant salt. The coolant-salt liquidus is 450°C (842°F). Below this
point about 96% of the salt forms 2LiF-BeF,, which would not circulate.
This is 18°F sbove the liquidus of the fuel. The phase diagram of the
coolant salt is shown in Fig. A.2.

Corrosion of INCR-8 by Fuel

Numerous corrosion tests have been comyleted‘with fuel mixtures
of the type to be utilized in the MSRE. Results of 37 INOR-8 thermal-
convection loops, 17 6f which operated in excess of one year,; show
complete compatibility between INOR-8 and the beryllium-based fluoride
systems,ll“l3 Experiments conducted in INOR-8 forced-convection loops
for one year or more similarly show low corrosion rates in fluoride
mixtures of this type.ll'l3 The operating conditions of these experi-

ments are shown below:

Forced Thermal
Variable Convectior Loops Coavection Loops
Fluid-metal ' o o
interface temp. 1300°F 1350°F
Fluid temperature ° o
gradient 20C°F 170°F
Flow rate ~ 2 gpm ~ T fom

Metallographic examinations of INOR-8 surfaces following salt
exposure in these loop experiments reveal no corrosion effects in time
periods up to 5000 hr. At times ionger than 5000 hr & thin (less than
1/2 mil) continuous surface layer develops at the salt-metal interfsce.

A quantitative measurement of the corrosion rate occurring in an
INOR-8 forced-convection system containing I..:I.}.i‘»Be__F‘E-uUFLF (62=37=1 mole %)
A

TEMPERATURE (°C)

UNCLASSIFIED

ORNL-LR-DWG 56886

900

800

700
600 Lif + LIQUID
\
N / BeFy+ LIQUID
400} 2LiF-Ber\\ e
+
~LLiQuiD \/
b
LiF +2LiF -BeF, o
. : W 2LiF -BeF, +BeF2 (HIGH QUARTZ TYPE)
300} "3
[3V]
L o :
2LiF -BeF2 & LiF -BeF, +BeF, (HIGH QUARTZ TYPE)
L ' & ] |
560 | LiF-BeF2 3 LiF -BeF, + BeF> (LOW QUARTZ TYPE) -
- LiF 10 40 50 60 70 80 90 BeF, -

BeF2 (mole %)

A.2. Phase Diagram of Coolant Salt.

..Eg_
-8)4-

was carried out by means of carefully fieighed and measured inserts
located at the point of maximm salt temperature (1300°F). %13 mme
inserts, removed after test intervals of 5000, 10,000 and 15,000 hr,
reveal relatively small weight losses, as shown below:

Timé (br) 2Wéight Loss .

' mg/cm mg/cn” /mo
5,000 1.8 0.26
10,000 2.1 0.15
15,000 1.7 0.08

The weight losses do not increase measurably after the first 5000 hr.
No changes in the wall thickness of the inserts are detected based on
before~ and after-test dimensions.

In some loops where there is evidence of contamination by water or
some other oxidizing agent, greater attack is found in the hot regions,
with roughening and pitting of the surface to depths of 1-1/2 mils or
more. In the cold-leg regions, magnetic metal crystals loosely adherent
to the cold-leg wall are found, composed pfedaminantly of nickel and
containing only-minor amounts of chrdmium and iron.

‘Several brazing materiasls have been developed for possible use in
the heat exchanger. Three nickel-base alloys, one gold~-base alloy, and
pure copper were tested in thermal-convection loops at l300°F without
showing any signs of attack after 10,000 hr of operation when they were
used to join INOR-8.

. Corrosion Reactions

The main corrosion mechanism is selective leaching of chromium,
not because of physical solubility of chromium metel in molten fluorides,
but by chemical reaction of this metal with oxidizing agents present in.
the melt or in the original metal surface.

Typical impurities produce corrosion by the following reactions:
@

+ s * 5

NlFé + Cr - CrFé + Ni

FeF,  + Or - CrF, + Fe

| 2FeF3 + 30r - 3CrF, + 2Fe
2CrF,  + COr - 3CrF,

Oxide films on the metal walls can react with the fuel constituents
(ZrFL or-UFh) to yield structural metal fluorides:

2Ni0 + ZrF4> - 2NiF +:  Zr0
2Fe203 + 3ZrFL N hFeFé + 37Zr0
2Cr 0 + ZrF, - u4CrF, + Zr0
These metal fluorides are then available for reaction with chromium as
shown above.
It is therefore necessary that both the melt and the structural

metal be of high purity. Even in this case, a possible cor;051on reac-
tion is 2UF4 + Cr & CrF. + 2UF..

2 3°
Sampling systems designed to provide periodic analyses of salts

“during corrosion tests are utilized in conjunction. with several forced-

" convection loop experiments.llF Samples taken over a 20,000-hr period

in loops operated under temperature conditions listed above show only

slight increases in the chromium concentration of the salt during test.

~In an experiment containing the mixture LiF-BeF JUFh-ThFh (62=36.5=0,5=1

2
mole %) the chromium concentration increased from an initial level of

400 ppm to 500 ppm during the first 1000 hr of operation and remained
at approximately the latter value during the remainder of the test.‘

Contamination of the Molten Fuel by Moisture or Air

In the case of moisture contamination, the possible reactlons are:

H 0 + OLiF - LiO0 + CHF

2 2
2HF + Cr ~»  CrF, + Hy
ZrF) + 210 - Zro,, +  LLiF
-86-

These reactions are complete and very rapild, causing both corrosion
of INOR-8 and precipitation of Zr0, in the fuel. |
Contamination by oxygen of the air has a worse effect, although the

reactions are not so fast as above:

0. + 2UF, -~ 2UOF, - - - - (strong oxidamt)

2
UOF, + Ni ~ NO 4+ UR,
2Ni0 + + ZrF, - Zr0, + Z2NiF | | NiF, + Cr - CrF2 + Ni+¢

2 2

The last two reactions are relatively slow but'they cause nickel transfer
from hot to cold regions. |
The Zroziis not dense enough to settle and stays in the fuel as.a
slurry. Although its particles are hard, erosion in the pump blades is

not evident.

A test program is in progress to evaluate the effects of contamina-
tion on the corrosion behavior of fused fluoridé mixtures, and to ascertain
the limits of the various contaminants which cen be tolerated without
seriously increasing the corrosiveaness of the fuel salt. The contaminants
under studj ificlude HF, metal oxides, and oil vapors. Results of this
program will be applied to specifications of the cover-gas purity as well
as to salt purity requirements. | |

Corrosion by the Coclanfi

- The coolant, being a mixture of LiF and BeF

09 does not present prob-

lems of precipitation of U02 or Zr02.
Although the coolant is very semsitive to moisture and air, the
oxides are 150 times more scluble in the coolant than in the fuel. Some

possible reactions are:

0 + 2Ni -  2Ni0o
. ' ' _' $ .
NiO + BeFé BeO ¥ + NiF,

NiFé + Cr « CrF2 + TNi
-87-

This causes selective leaching of chromium, precipitation of BeO,
and deposition of nickel. in the cold regions of -the loop, although the
reaction NiF, + Cr « Nij+ CrF, is not as féfiperafiure,sensitive as
the UFA oxidation reaction. ,

Since the probability of contamination in the coolant circuit is
higher than in the fuel eircuit, caution must be taken to avoid undue

corrosion.
~88-
Appendix B
HOT SPOT ANALYSIS

An ana;ysis was made to estimate the temperature that the MSRE
grephite and fuel may attain if one of the fuel passages becomes blocked
so that mo flow occurs. The estimate is based on a camiparison between
the temperatures resulting from a unidirectional heat-flow case and those
given by a relaxatiqn-solution techniqué of twomdifiensiénal heat flow
from one face to the other three faces of a square:rod.. The relexation
solution is found to be 36%, or about one-third, of that for the uni-
directional case. |

The following cases are superposed:

a. Temperature rise in an infinite slab of fuel with & uniform
volume heat .source.

b. Temperature drop for éonduction of this heat from the fuel
across a graphite slab. ‘

c. Temperature drop caused by heat génerated.within the graphite.

d. Film drop between the graphite wall and the bulk mean tempera-
ture of the fuel in the adjacént open channel due to the heat inflow and
its own volume heat source.

The expansion of the fuel'réduges the power generation rate slightly.
This adjustment is made at the end of the calculations. |

The following conditions and properties are used in the calcula-

tions:
Reactor power 10 Mwt
Fission-product decay heat outside core 4%
Fraction of core power in graphite ) 7%
Peak-to-average power ratio 2.8
Thermal conductivity of fuel (Kf) 2.75 Btu/hr-ft-°F
Thermal conductivity of graphite (Kc) 12 Btu/hr-ft-°F
Half fuel-channel width (xf) 0.20 in. or 0.0167 ft
Craphite block thickness (xc) 1.60 in. or 0.1333 ft
Bulk mean temperature of fuel in 1210°F

adjacent channel
Core size

,-89-

Fraction of core volumefiin fuel - 0.225
Symbols | ,

q” Heat generation rate, Btu/hr-ft5

q”  Heat flow rate, Btu/hr-ft

| nmi' Temperature difference,'°F

K Thermal conduct1v1ty, Btu/hr- ft-°F

X Thlckness or distance, ft
Subscripts

f Fuel N

bf  Fuel in blocked channel

ff Fuel in flowing ohannel

c Carbon -or éranhifie A

7 Wall of flowing fuel channel

W

4.5-ft diam, 5.5-ft height

The temperature drop across a slab with uniform heat generation and

cooling on one. side is:

The bulk mean temperature of the fuel is:

2

AT = £ AT.

m )

The tefiperature drop across & slab with Uniform heat flow is:

/” '
Ar = X

K

(1)

(2)

(3)

The film drop for laminar flow in flat chennels having uniform.heat

generation is:

/4

q

The heat generated within a slab and flowing across the face is:

X (17_ @+a” w/a” x) - 14)

35

(4)
=90~

-qg} = qg; = 4.3 x 106 Btu./hr-ft5 5
V7 : 6 2
Q) = 0.095 x 10~ Btu/hr-ft° .
For part (&):
AIII - q X?‘ ‘= 1"03 X 106 X (00016112 = 220°F
Féi péit (b):
‘. q” X : 6 S |
AT = 0.36 x "t o 0i36 x 4.3 x 10" x 0.0167.x 0.133 _ 287°F.
c S Kc . 12 4
For part (g): |
| X 0.095 x 10° x (0.133)2 o
ATC = 0.36 X ——2?{-:—- = 10.36 X 5 x 1o = | 25‘F_.-
For part (d):
qM,AX?f , 17(1 + q” ‘w/q”’ X ')wlh
AT . - fIf , : £ff"
£f Kf ’ [ 35 ‘] )

Id4

q, is taken to be one-third of the heat generated-ih one-half of
the blocked fuel channel plus four-thirds of that generated in the

graphite rod normally flowiiig through thé;fééé.
V7 7z
7’ qT Xf + _l_l- q(.‘, X,
W 3 3 Tk
6 6 o
. %3 x10 x0.0167 , 0.095 x 10" x 0.133

24,100 + 4,200 = 28,300 .
-0l -

6 x 0.0167 = 72,000 .

Q” XKoo = k.3 x 10

1 +q)/a¥ Xpe =1 + 28,300/72,000 = 1.39 .

AT '=.-_ ]'"'5 X 106 X 0.01672 <_‘_|_7 x 1.39 - 1k >

ff 2.75 35
= Lho x 0.274 = 121°F
The bulk mean temperature of the fuel above the wall is:
.2 .2 . 1R
Amm = 3 Ambf = 3 pd 220‘¢ 1&7 F

With the fuel density changing -1.25 X lO_h/qF, héat generation in
the blocked fuel channel is-‘about 93.5% of the base gsed in these calcula-

tions. Making this adjustment, the temperatures become as shown:

Temperature in the adjacent fuel channel, °F 1210

Temperature rise in fuel, °F - . _113 |
Graphite.wall temperature, °F ; | | 1323 B
Temperature rise in g;afihite, °F "f 4 : 292
Top graphite wall temfiérature, °F | - 1615
Temperature rise in fuel channel, °F ‘ 206
Peak fuel temperature, °F | | - agel

These temperatures are attainable only in the center of the core and

should not damagefeither'theffuel or the graphite.
_92 -
" Appendix C

Number JS-80-123
Date 1/24/61
Revised

Page 1l of 11

JOB SPECIFICATION
REACTOR DIVISION
OAK RIDGE NATIONAL IABORATORY
UNION CARBIDE NUCLEAR COMPANY
Division of Union Carbide Corporation

ORNL, Oak Ridge, Tennessee

Subject: Specification for Primary Drain and Fill Tanks, Primary Flush
Tank, Secondary Drain Tank and Fuel Storage Tank for Molten
Salt Reactor Experiment

am— — —
= —

—

1. SCOPE

This specification covers the requirements for materials, fabrication,
inspection and testing of the primary drain and fill tanks, primary
flush tank, secondary drain tank and fuel storage tank for the Molten
Salt Reactor Experiment. :

2. APPLICABLE SPECIFICATIONS) CODES, DRAWINGS, AND QTHER PUBLICATIONS
2.1 The latest revisions of the following dccuments shall form e
part of this specification to the exvent stated in subsequent
sections
ASME Boiler and Pressure Vessel Code, Sections VIIT and IX
ASME Code Case Interpretations 1270N-2 and 1273N;3

ORNL Specification MET-RM- 4 for INOR-8 Welding Fittings,
Shapes, etc ;

ORNL;Spec1fication MET-RM-B163 for INOR-8 Tubing

ORNL Specification MET-RM-B167 for INOR-8 Pipe

ORNL Specification MET -RM-B304 for INOR-8 Weld Filler Material
ORNL Specification MET-RM-B334k for INOR-8 Plate
ORNLlSpecifIcation MET-RM-2 for INOR-8 Forgings

ORNL Specification MET-NDT-E165 for Liquld Penetrant Inspection

ORNL Specification MET-WR-2 for INOR-8 Welding Requirements

- 93..

Number JS-80-123

Date 1/24/61
Revised
Page 2 of 11

‘ORNL Specifications P.S.-23, P.S.=25, P. S.-26 for INOR-8
Welding Procedures (for information only)

ORNL Specifications QI'S-23, QTS-25, QTS-26 for Welder
Qualification Tests (for information only)

2.2, Drawings

The following Company's fabrication drawings form a part of thiS“
specification -

Assembly

D-FF-.-Ahoh'j'j Primary Drain and Fill Tank

DéFFgAhOh56 Primery Drain and Fill Tank - Steam Dome
- Assembly and Details

D-FF-A4OUST Primary Drain and Fill Tank - Assembly and

Details

D-FF-A4O458 Primary Drain and Fill Tank

Bayonet Exchanger ~
Assembly and Details ' - :

D-FF-ALOL59 Primary Drain and Fill Tank

Bayonet Exchanger
Bracing '

D-FF-AUOLEO  Primary Drain and Fill Tank - Cooling Water

Header Assembly and Details
D-FF-AkOLEL  Secondary Drain Tank - Assembly and Details

D-FF-A4OL62  Primary Flush Tank and Fuel Storage Tank -
; Assembly and Details

3. REQUIREMENTS

3.1 Design

The tanks to be furnished under this specification shall be
fabricated in accordance with the Company‘'s designs, and as shown
on the Company's fabrication drawings accompanying and forming
a part of this specification, except that the Seller may adopt
his own standards for weld-end preparation, provided they are
submitted to and approved by the Company in writing prior to
start of fabrication. The tanks will contain molten fluoride.
salts. ’ ' '

3.1.1

3.1.2,

Number JS-80-123

Date 1/24/61

Revilsed

Page 3 of 11

Primary Drain and Fill Tanks

Two (2) primary system drain and £ill tanks, similar in.
design, with the exception of certain nozzle locations,
will be required.

Each fill and drain tank consists of a lower tank, which
will contain the molten fluoride salt, and an ‘upper steam
dome. Heat will be removed from the molten salt in the
lower tank by introducing water into concentric thimbles

- which penetrate the top head of the lower tank. The steam

generated in the thimbles will be collected in the upper
steam dome through tubing and flexible connectors. Details
of the tank support designs bhave not been finalized. The
Seller shall be responsible for furnishing tank support
brackets and steam dome supporting steel after designs
are completed and drawings furnished to the Seller.

The following design criteria are included for the Seller's
information:

Tank désign pressure 50 psig

Tank design temperature '1300°F

Cooling system capability 100 kw at 1300°F
Cooling water temperature 100°F

Steam dome design pressure | 50 psig

Steam dome design temperature 300°F

Primary Flush Tank

One (1) primary system flush tank will be required. This
tank will contain the molten. salt necessary for flushing
the primary piping system after the radiocactive salts
have been drained from the.systen. '

Design conditions are as follows:
Design temperature 13009F

Design pressure . 50 psig

3.1.3

3.1.4

-95

Number Js-80-123
Date 1/24/61
Reviged '
Page 4 of 11

Secondary Drain Tank

‘One (1) secondary -drain tank will be required to contain

the molten salt drained from the secondary piping system.

"Design conditions are as follows:

Design temperature | : 13009F
Design pressure _ 50 psig

Fuel Storage Tank

one (1) fuel storage tank will be required. Design conditions

are as Tollows:
Design temperature 1300°F

Design pressure | 50 psig

3.2 Materials

" 3.2.1 All tanks, tubing, piping, etc., to be furnished under this

Temp. Op

specification shall be fabricated,from INOR-8, a nickel-
molybdenum-chromium material, unless otherwlse specified
in the Parts Lists which appear on the Company's fabrication
drawings. ‘ |

Design data for INOR-8 are given in the following table:

Modulus of Mean Coeff.

Allowable Elagticity - of Expansion Thermal

- Stress psi ' in./in. OFx10~6 Conductivity
 psi 10° psi TO%F 0 T _ Btu/ft2-hr-OF /£t
24,000 314 1T
24,000 30.7 | 8.1
22,800 30.1
21,700 29.5 6.45 9.0
20, 800 28.9 '
20,000 28,5 6.76 9.9

19,300 28.0
Number JS=-80-123
Date 1/25/61

Revised
Page 5 of 11
Modulus of Mean Coeff.
. Allowable Elasticity of Expansion Thermal
Temp. °F Stress - psi ' in./in. Opx10-6 Conductivity
T psi 100 psi 0% to T Btu/ft2-hr-SF/ft
800 - 18,700 37.7 - 7.09 | | 10.8
900 18,150 27.2
1000 16,000 | 26.8 o T.43 11.7
1050 13,250 g6,6
1100 9,600 26.4
1150 6,800 - 26.3
1200° 4,950 26.1 7.81 12.6
1250° 3,600 25.8
1300 2,750 25 .4
1350 . 2,0% ah.g | | |
100~ 1,600 2L, Y | - 8:16 | ‘13.5

Density, 0.3L7 lb/in3

Specific heat, 0.095 Btu/1lb °F

3

3.2.2 All material received by the Seller for use in fabrication
. .of the:equipment to be furnished under. this specification
shall be inspected by the Seller for damage during ship-
ment. All material found to be defective shall be-
rejected and reported to the Company. The methods of
inspection shall be approved in writing by the Company.

3.3 Fabrication

3 3 l All tanks shall be fabricated in accordance with the applicable
sections of the ASME Boller and Pressure Vessel Code, Sections
VIIT and IX, including Code Case Interpretations 12TON-2
and 1273N-3 for primary nuclear vessels, except that
materials of construction and allowable stresses shall

Number JS5-80-123
Date 1/24/61
Revised ‘
Page = 6 of 1L

be in accordance with this specification. In addition,
the supplementary requirements of this specifieation'Shall
be met. Where conflicts or inconsistencies occur, the
requirements of this specification shall govern. Code
stamping will not be required. .

3.3.2 .Forming of INOR-8 Materials

All procedures to be used by the Seller for forming of
INOR-8 materials shall be submitted in writing for the
approval of the Company prior to start of fabrication..

All materials after‘belng formed.by any method ‘such as
bending, drawing, or swaging shall be subaected to liquid
penetrant inspection of all surfaces in accordance with .
specification No.. MET-NDT-E165. Any type of crack, fissure,
fold, or other injurious defect shall be cause for re jection
of the part.

Removal or repair of injurious defects shall be permitted
only after written approval of the Company. -

- Formed heads shall be made in accordance with ORNL spe01fi-
cation No. MET-RM-6, .

3.3.3 Heat Treatment

The Seller shall furnish the Company with written heat
-treatment procedures which he proposes to adopt during the
fabrication of the tanks. These procedures shall be
~ approved In writing by the Company prior to start of tank
fabrication.

A record of each heat treatment shall be made and shall
form a: part of the fabrication and inspection report.

3.3.4 Welding

All welding of INOR-8 material shall conform to the Com--
pany's Welding Specification MET-WR-2, attached hereto.

Welding of INOR-8 material may be performed in accordance
with the Company's Procedure Specifications PS- 23, PS-25,
and PS-26. at the Seller's discretion, however, the Company
assumes no responsibility for the adequacy of these speci-
fications to meet the requirement of this specification.

All welding procedures used by the Seller in weldingiINOR-B
materials shall be submitted to and approved in writing by
the Company prior to start of faebrication. -The procedures
3.3.5

3.3.6

3.3.7

Number JS-80-123
Date 1/24/61
Revised .
Page T _of 11

and welders shall ‘be- qualified in accordance with ORNL
Specification MET-WR-E

Cleaning

Tmmediately following any operation that imposes any unclean
condition and before assenbly, all parts and subassenmblies
shall be cleaned free of all oxides, grease, oil, filings,
dust or any other foreign material. All internal and external
surfaces shgll have a bright finish. Precuations shall be
taken to insure that all parts and subassemblies are kept in

a clean conditiom throughout fabrlcat ot of tce subject tanks.

It is especilally important that oxide scale shall be removed
from all parts that camnot be reached for direct ‘inspection
and cleaning after assembly. Discoloration of su*glcal gauze
after wiping metal surfaces shall be used as a check for
cleanliness;7

Compounds containing sulphur, lead or mercury shall not be
permitted to come into contact with uurfacem of INOR~8 material.

Workmanship

The Seller:shall fabricate all tanks and appurtenances in a

menner consistent with the standards of high gquality workmanship.
Inferior quality of workmanskip, azs determined by visual inspection
by the Compeny's inspector, shall.bé couse for rejection of‘the work .

The quality of workmanship as approved in the Welder's qualifi-
cation tests shall be maintzined throughout performance of all

work on the subject tanks. Inferior workmanship as determined

by testing and inspection of the welds in accordance with this

specification shall be cause for rejection of the work and re-

gqualification of the welder.

Identification

The Seller shall affix an INOR-8 nameplate to the outside
shell of each tank. The following data shall be stamped or
engraved on each nameplate:

Fabricator's Name
Specification JS=80-123
Year Completed
Design Pressure
Design Temperature
Hydrostatic Test Pressure

A nameplate giving similar data shell also be affixed to
the steam dome of each primary drain and fill tank.
-99- o
Number JS-80-123
Date’  1/24/61
'Revised
‘Page 6 of 11

. INSPECTION AND TEST REQUIREMENTS

L.1

h.2

4.3

L.b

Tn addition to welds, each forging shall be- liquid penetrant
inspected in accordance with ORNL Spec1f1cation MET-NDT-E165

'Inspection of welds shall be in accordance with ORNL Welding _

Specification MET-WR 2.

The Seller shall arrange for the Company's representative to -

have access to such parts of all plants as are concerned with

the supply, manufacture, and assembly of parts for the subject
tanks when requested, including Seller's own plants and those

of his suppliers.. Where reference is made to the Purchaser

in the ORNL specifications for INOR-8, it shall be interpreted
to mean Company .

The Seller shall notify the Company at least three (3) working
days in advance of . the start of tests and inspections so the
Company representative may be present. The tests and inspections

| _referred to include:

L.5

h.6

(1) Welding procedur'e qualification. (3.3.4)
(2) Welder qualifications. (3.3.k) |
(3) Any repairs of defects. (4.5)
(4) Hydrostatic test. (L4.6) |

 -(5)‘ Leak tests.' (4.7, 4.8)
(6) Any other tests. (4.9)

(T) - Preparation for shipment. (5)

'"No waiver of inspection observation or any requirement of  this

spec1fication will be made unless confirmed in writing by the

Company .

Repalrs necessitated by defects in material or workmanship shall
not be made without full knowledge and approval of the Company.

The tanks shall be subjected to hydrostatic tests as follows:

The lower tank sections of the primary drain and f£ill tanks shall
be subjected'to a hydrostatic test of 655 psig.

The secondary drain tank, the primary system flush tank and the fuel
storage tank shall each be subjected to a hydrostatic test of 655 psig.
L7

L.8

h.9

k.10

=100-

Number JS-80-123
Date . 1/24/61
Revised -
Page 9 of 11

The primary drain and fill tank steam dome and interconnecting
piping to and including the tank thimbles shall be given a
hydrostatic test of 80 psig.

The hydrostatic test pressures shall be held for one hour, during
vbich time all surfaces and joints shall be visibly inspected for
leéaks. Repairs shall be made only after notifying end receiving
the approval of the Company.

After the hydrostatic test, a helium mass spectrometer leak test
shall be applied to the lower tank sections of the primary drain

and f£ill tenks, and to the primary flush tank, fuel storage tank

and secondary drain tank. The tanks shall be tested for leakage

to the inside by bagging each tank in a plastic bag filled with
helium and evacusting the tank. Each tank shall be tested separately.

The leak detector shall be demonstrated under test conditions

~to be sensitive to 1 x 10-8 §TD cc/sec of helium when using

8 standard leak of 1 x 10~9 STD cc/sec connected to the most remote
portion of the tank. Indicated 1§qkage‘into each tank under test
shall not be greater than 1 x 10-° STD ce/sez. The leak test
shall be run for a minimum of 30 minutes on each tank,or for
sufficient time to detect the standard lesk, whichever 1s greater.
If, in the cpinion of the Company, the background reading of the
leak detector has changed sufficiently during the above test to
create a doubt regfirding the absolute lesk-tigbtness of the tanks,
the test shall he repeated. :

The presence of the Company representative is required during all
helium leak testing.

After the hydrostatic test a halogen leak detection test shall be
applied to the upper tank sections, thimbles and inter-connecting
tubing to the primary drain and fill tanks. The volumes to be
tested shall be pressurized with a mixture of st least 25% freon in
air, and the exterior surfaces surveyed with a halogen leak detector,
using a technique demonstrated under test conditions, to be capable
of detecting a lesk of 1 x 102 STD ce/sec of the pressurizing
mixture. Leaks giving indications greater than 1 x 1072 STD cc/sec
shall be cause for rejectlon. '

The Seller shail make and report such other tests and inspections
as are necessary to satisfy himself of the integrity and good
condition of the tanks. - E

If at any stage of teéting or inspection, physilcal failfire,
deformation or mechanical damage occurs or is observed, it shall

be deemed as failure of the tank to meet these specifications.
(-

5.

6.

.11

<101~ o
Number JS-80-123
Date 1/24/61
Revised o
Page 10 of 11

A test fallure which requires repair and/or corrective measures
to be made will automatically necessitate repetition of the -
previous inspections and/or tests on the part requiring repair.

PREPARATION FOR DELIVERY

5.1

NOTES -

6.1

Preservation and Packaging

Prior to shipment, all tank openings shall be closed with
gas-tight closures and the tanks shall be evacuated and charged o
to 25 psig with welding quality helium, argon or dry nitrogen gas.
A valved pressure gauge shall be furnished and shall be securely
affixed to one of the closures in each tank in such a manner

as to permit reading of the internal gas pressure.

Each tank shall be securely mounted on a skid and suitably blocked

and strapped to prevent shifting and/or damage while in transit.

Engineering Information

6 1.1 Design Approval

Within five weeks after receipt of order, “the Seller shall
furnish the following Engineering Information for approval

by the Company

" No. of
Copies ‘ Descrlption
T Leak Test Procedures .
T o Welding Procedures
.7 ‘Heafi Treatment Proceoures
T Record of Welder s Qualification Tests
T INOR 8 Inventory System

After receipt of Engineering Information; the Company will -
require a minimum period of four (4) working days for its
review. TFabrication of work shall not be started and de-
livery of equipment and material shall not be authorized
until written approval of the Engineering Information has
been extended by. the Company. :
6.2

6.3

6.4

=102 - . ,
Number JS-80-123
Date '~ _1/24/61
Revised =
Page I of 11

The Seller shall furnish four (h) certified coples of the
Engineering Informetion within four (4) weeks after receipt
of approval.

Seller's Data

Certified éopies of Seller's Data shall be furnished in the |
quantities specified prior to shipment of equipment. .

No. of .
Copies ' Description
1 Radiographs
1 Heat Treating Charts
'h Leak Test Reports
A Liquid Penetrant Test Reports
L Weld Identification
b INOR-8 Inventory Report
Y Supplementary Shop Dr&wings
L - ASME Form U-1

Manufacturing, Inspection and Test Schedule

Within four (4) weeks after receipt of order, the Seller shall
submit to the Company the manufacturing, inspection and test
schedule for all material and equlpment furnished under this
specification.

TNOR-8 Material Inventory

Throughout fabrication of the tanks, the Seller shall maintain a
system of materlal identification and control sufficient to establish
the complete identity and history of all INOR-8 material, including
weld filler material. The system shall be approved in writing by
the - Company prior Yo shipment of material to the Seller.

4
7N

-103-

- Appendix D
COMPONENT DEVELOPMENT PROGRAM IN SUPPORT. OF THE MSRE

The reliasble performance of components;and-auxiliaries used in the
circulation ofimoltenwsalts has been established in;oyer QOO,QQO hr of..
accumulated loop operations, and forms. the basis,forAtne specification of
components for the MSRE. | ; e '

As an added insurance of reliability and safety, prototypes of
critical MSRE- components will be operated out—of-pile under conditions
resembling‘those of the reactor. Facilities to be used for this testing
include salt systems of various sizes and different degrees‘of'complex-

ity and model tests in which hydraulic and mechanical processes are

:analyzed

Core Flow

A1/5- scale plastic model of the reactor has been operated with :

.water as the fluid. Fluid velocity distribution in the entrance plenum

in the region next to the cylindrical wall and in the lower plenum has
been experimentally determined to be satisfactory_for purposes of B
cooling those regions..\On the basis of measurements made in'tfie 1/5-
scale model, the des1gns of the MSRE core and of a full-scale model
were establiShed The full scale model. Wlll be operated w1th 1250 gpm
of glycerol solution to reproduce the expected Reynolds nuflber of the
reactor salt. .The adequacy of flow distribution in every portion of
the core-vessel assembly will be proved or required modifications will

be devised and demonstrated

The possible interactions of fuel salt and full-size core graphite
stringers will be investigated in an B-in -diameter vessel, part of
the Engineering Test Loop (a facility for testing many MSRE prototype

components .and 0perating procedures)
=10k~
Fuel Circulation Pump

The conceptual design of the MSRE pump is similar to that of the
- pump developed for the Alrcraft Reactor Experiment. These pumps have
beén virtually trouble—free in out-of—pile use. rThe prototype will be
tested with vater and with molten salt. -

Hydraulic performance will be investigated in the water tests,
later the design of “the experiment will be modified as reqnired for the
stripping of dissolved gases in the pump bowl. rThe'investigation of
gas kinetics in.a pump bowl has been started with COo being used as a
tracer. B .

Hot tests with molten salt will provide-a{final test of the-pump
geometry derived from the water tests. Krypton-85 tracer will be used
to develop an-effective means of purging fission gases from the pump
and of excludirng them from the motor. The pump support arrangement
designed for the MSRE will be utilized in the bot test stand. The
MSRE pumps will be tested at operating temperstures in this equipment
prior to installation at the reactor. ‘ ‘

Freeze Flanges

Freeze flanges have been & part of moltennsalt engineering devel—
opment for several years. They have been found satisfactory in a large
circulating salt system, the Remote-Maintenance Demonstration Facility.
At was shown that the flanges could withstand repeated thermal cycling
and could be broken and reassembled. '

. Two flange pairs were cycled between room temperature and 1300 F
for more than lOQ_times without failure. Salt leakage from a freeze
flange has not been experienced to date, slthough leaksge of helium
through the secondary gas buffer seal is not uncommon. '

The design of the flanges has been improved to provide grester
strength toward axial loads, to reduce the thermal stress, and to
improve the tightness of the buffer seal. Flanges of the improved .
type will be tested under simulated reactor service conditions to
establish their reliability.

-105-
Heat Exchangers

Heat exchangers as complex as the MSRE exchanger were fabricated of .
Inconel for the ARE and the ART. The fabricability of INOR-8 inmto tubes,
and tube-tubesheet assemblies has been demonstrated.. The MSRE heat ex-
changer and radiator assemblies, ee now designed, do not require further

development.

 Freeze Valves

Plugs of frozen salt havegbeen used in many loops to isolate circu-
lating molten salt from the environment. Two prototypes of "valves" to
be used in the MSRE drain lines have been frozen and thawed 100 times
without damage or incident. . Ability of the valve to fall safe on loss
of heating or 1oss of cooling was demonstrated. |

| Testing of prototype Preeze valves will be continued as part of

the operation of other loop and component tests.

Sampler-Enricher

Part of the sampler-enricher mechanism has been mocked up, and
mechanical components and instruments are being added to the mockup as .
the deSign evolves. The basic mechanical parts thus far are functioning
reliably. | o ‘ |

A complete sampler-enricher mockup w1ll be fabricated when draw1ngs i
of the MSRE device are finished. After the mechanical debugging of -
this prototype,‘its ability to obtain representative samples without -
contaminating the system with oxygen, or the environment with activity,
will be demonstrated on the Engineering Test Loop

Heaters

Prototype MSRE heaters for pipes and vessels are being fabricated
for testing on mocknps and loops. .The heaters will be sflbjected to life
tests, and their temperature distribution and heat loss will be measured.
Possible_damage to pipe walls by overheating will be investigated also.
~106-

Poison Tube

‘A mockup of the MSRE poison tube will be fabricated and tested to
establish its reliability, response times, and control characteristics.

Gas-Handling System

Helium cleanup traps, based on the technology developed for.gas-
cooled reactors, will be tested under flow conditions. At the same time,
methods of chemical analysis to detect small quantlties of contaminating
oxygen will be developed and demonstrated

Maintenance

The practicality of the two general methods of maintenance designed
into the MSRE has been thoroughly demonstrated during the past 3 years.
Entirely remote maintenance, with the aid of stereo-television and a
General Mills manipulator; was demonstrated on & special salt system
(Remote-Maintenance Demonstration Facility) of about the size and degree |
of,complexity'of the MSRE. After this facility was operated with molten
salt, the pump, the dummy core, the heat exchanger, and heateérs were
removed and replaced. ,The:system was shown to be operable following
the replacements. : f '

Semidirect maintenance with long-handled tools operated through
,portable shielding has been used successfully on a number of operations
at Hbmogeneons Reactor Experiment No. 2, a reactor of the general size
and activity level of the MSRE. BSome of the operations performed with
these techniques were repair of the core vessel and replacement of
pumps, valves, ‘and a filter.

Although the feasibility of maintenance is regarded as having been
established, additional development and practice is required to work
out detailed procedures. Specific maintenance problems are being solved
with the use of- appropriate mockups in the Remote=Maintenance Demon- .
stration Facility. As an aid to the designer of maintenance procedures,
- a 1/12-scale model of the reactor system and maintenance area is being

built. This model will provide insurance that every item in the MSRE

can be repaired or replaced after operations have started.
_107-

Appendix E
'ACCIDENTS INVOLVING RELEASE OF MOLTEN SALT INSIDE THE REACTOR CELL

Several incidents involving the release of the hot molten fuel salt
and coolant salt into the reactor cell were studied. The first postu-
lated accident involved an instantaneous drop of the molten salt'into
the reactor cell. The second accident 1s a rapid dispersion into the
cell, such that the air within is quickly heated. The third accident is
similar to the first except that the molten salt falls into water in the
- bottom of the cell. o i S

Many simplifying assumptions are made, but all are conservative;
therefore the calculated. results of these accidents are more severe than
would be experienced in an actual accident. o

The calculations for the. three postulated accidents 1ndicate that
the third, invwhich water leaked into the reactor cell, is the most
severe. However, an automatic sump pump ‘removes all inleakage of water.
Furthermore, . any presence‘of7water is immediately alarmed to the operator,

who may then tske corrective action if the pump fails to work.

Air Only in Cell

Let it be assumed that- both fuel and coolant systems rupture and
all fuel and coolant fall into the bottom of the reactor cell For sim-
plicity in calculations it is assumed that the bottom ‘of - ‘the reactor cell
is flat. The follOWing parameters are used in the- analySis of such an
accident: . :
a. The total volume 6f the fuel and coolant mixture is 85 £,

b. The mixture covers'the bottom of the cell ih a layer 0.8 ft
thick. , .

¢. The temperature ofgthe mixture is;l500°F iflitially. ]
-d. The temperature of the air in the cell is*100°F initially.
e. The volume of the reactor cell is 11, 55O'ft3.

i The total heat transfer coefficient is- calculated to be
= 6.3 Btu/hr-°F-£t°.
=-108-

The produced decay heat, q(t),'is chosen to be 7% of total power at

1l sec after shutdown.

Q

o q@ (L sec) = (0.07)(20 Mw)(3.%15 x 10

Q'O

it

2.39 x 10° Btu/hr .

The decay heat per unit surface

6
q - 2P x 20 Btu/hr ,

107 ft

2.2k Btu/hr--ft2

o
]

The Way-Wigner equation for decay heat is:

. -002
a(t) = q t .

Writing (5) with t in hours and assuming t = O at 1 sec after

shutdown,

a(t) = q (3600 t + 1)'0'2 .

EEEPR - ~0.2
a(t) = g (3600 €)% .
The parficular equation for this system is
= q(t) = hte 2

where 6 = teyperature‘in °F

and ¢ QITcppfl- ,

(0.468)(133)(0.8) ,

- 49.8 Btu./°F-ft2 .

6

Btu/Mw-hr) ,

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)
-109-
'The Laplacian transformation of Eg. (8) is:

o) = —2 L) —2 . (9)
s + (ht/c) C s + (ht/C) AR |

Let. M =hfc = 0.126/nr

Returning to the time domain, and‘using the convélution integral, B

o) = 6™ + 2 [q(p) e"“"'] , | (10)
t -
- é)ie"')Lt + ;=./fi a(t) é-(t’T)k ar ,
£ J | o
S 4.35 x 10° f 0.2 e-_xt"a (11)
- i 49.8 i o )
| Yo
Equation (11) is not readily integrablej. therefore let‘i%rbe
approxifiatédiby'. | B |
6(t) = e v [ 6, + 87 .k Z ti-o-.e e At:} L (12)
L .=1

The surface temperature of the molten salt is-initially at 1300°F,
and the)shield water outside the reactor cell is initially at lOO°F. If
it is assumed that after,thé accidental falling of the fgel and coolant
these températures remain cdnstant, a meafilsteady-staté air fempefature
may be compfited as follows. The heat flow out of the'salt t0 the_air in
the steady-state condition equals the heat flow from the air through the
steel shell to the water. This tacitly assumes that thé‘salt is an in-
finite source whilé the water is an infinite sink. In equation form°

the heat flow is written as follows:
hsAS(tS - ta)

At =

Subscripts

where

t

p :
a

hAt +hAt
8 8- 8 W WW

-110:-

(1k)
thermal, conductivity, Btu/ft2-°F-hr ,
area, ft2 R |
temperature differential .
'thw(ta -t) . (15)

s, a, and v refer to salt, air, and water, respectively.

R

h A + h A
g s W W

i

h
s

ja g
]

Substitute

214°F .

(16)

2 Btu/ft°-°F-hr ,

0.86 ,

107 £t°

2

2370 £t°

1300°F ,

100°F .

in (16):

(17)

Initially the temperature of the air = 100°F, P = 14.2 pSia.}

The steady-state préssfire of the air within the reactor cell is:

560

(1b.2)(674) | |  (18)

17.1 psia |, - | - (19)

2.4 psig. | - (20)
=111~

Sudden and Minute Dispersion .of Fuel and Coolant Salt
Within the Reactor Cell

Let it be assumed thatpthe fuel and coolant circuits are ruptured
resulting in a very sudden and minute dispersion‘of thé molten salt into
the reactor cell.fl Let it also bevassumed that the molten salt and air
within the cell constitute an insulated system and that the air then
attains thermal equilibrium with the salt. '

In the steady state, the heat loss of the molten salt equals the
heat gained by the air. ' : |

(Tfuel B Tfinal) Ve cpf = (Tfinal B To) Ya cpa A (21)
where Tfuel temperature of fuel = 1300°F ,,.j
Tfinal = final steady state of fuel and glr s
T0 = initial temperature of air = 100°F ,
W - weight of molten salt = (85) £t (133) 1b/ft’
= 11,320 1b
cp' = specific heat of molten salt = 0.468 Btu/1b-°F
¢ : _ . o : |
w, = weight of air = (2.378 x 10'5)(52.17)(11,550) = 884.0 1b,
e, = specific heat of air = 0.24 Btu/lb .
a A
Substitute in (21):
(1300 - Tfiqa;Q(ll,ieo)(0-468) = (Tpypgy - 100)(88)(0.24) ,  (22)
6,890,000 - 5300 Ty oo 212 Tpy oy - 21,200

Teinal = ©,911,200/5512

1250°F . | | | (23)
<112 -

The idéalized gas relationship is: -

- >
R~
where - Pi = initial pressure in vessel = 1h.2 psia at 100°F ,
T, = 100°F = S559°R ., .
T2 = 1250°F = 1709°R ., ..
Substitute in (3):
5 _ (1k.2)(1709) -
T 2l
P 559 N P ( )
= L43.4 psia o * (25)

28.7 psig . ' - (26)

Assumé now that, after a steady-étate condition is reached within
the pressfire vessel, the drain-tank cell volume is suddenly added to the
reactor-cell volume. Initially, the reactor-cell steady-state pressure
is P, = 28.7 psig, and T = 1250°F, and the drain-tank cell is at P, =

2 3
14.2 psig and T3 ‘= 100°F.

-7/ ’ -/ = (M s . NN
(TE Tfinal) ?f Cpf M (TQ J Tfinal) Ya cpa (Tfinal _15)_wae cpa;’(ET)
(T2 Tfinal) (Wf.cp t Wy Cp ) (Tfinal_ T5) Ve Sp (28)

, e £ ! 2 “a
where .?Tgr = 1250°F ,
Cwe = 11,320 b,
c. = 0.L68 ,
Pp .
W = 884 1b ,
-113=

T, = 100°F ,
w = (2378 x 157)(32.17)(7010) = 5%
2 ‘ :
c = 0.24 Btu/lb
Pa

" Substitute in (8):

‘ (1250 - T% nal)[(11,320)(0 468) + (88&)(0 eui]= (Tfifiél - 100)(536)(0 2k) ,
| (29)
(1250 -1, 0 )(5502) = (Tfi g1 - 100)(128.6) ,
. _ 6,880,000 + 12,860 - 6,892,860 . _ . ;ep
Teinal =~ 565 _' S5 - R
T < TOVF
, A R -  (0)
Ty Lo : L '
. .2 ;Ei_'
559 ~ 1223 ’ |
'P3 = Blfl ésia ,
- 16.4 psig . | - (31)

Water in the Reactor Cell

In the third postulated accident, the ffiel system was considered
ruptured and molten salt released into water already in the bottom of
the containment shell, producing steam and increasing the pressure. The
worst situation was determined to be thé“%fifiture'of the fuel line between
the heat exchanger and the thermal shield, which would discharge about

- 3

3 ft© of fuel in 1l sec, under pump pressure, with another 2k ft3 draining

v
-11k-

out in 60 sec. Ruptures withln the thermal, shleld or in the drain line
would supply more salt but at a slower rate Likew1se, a rupture of a

3

tube in the heat exchanger would allow most ;of the 30 £t~ of coolant
salt to drain, but again at a slower rate, for the addltlonal salt.

The heat d1ss1pat10n rate of the contalnment-vessel walls when
condensing steam is 1.65 x; 106 Btu/mln, whrgh is the equivalent of a -
hot-salt leak rate of 21 ft3/m1n Thus,’after convective'circulation
is establlshed within the reactor cell, condensatlon beglns and the |
pressure within the contalnment cell w1ll begln to decrease For. the
calculations, it is assumed that there w1ll be no heat loss from the
vessel for the first 350 sec, -and that it w1ll increase 11nearly from
0 to 1.65 x 106 Btu/mln 1n the next 30 sec, . . | |

The pressure rise is. calculated on the,assumption of" the optlmum
amount of water being present The max1mum pressure results when thére
is Jjust sufficient water to be converted ; to saturated steam, surplus
water absorbs heat, and superheating the steam produces less increase
in pressure than the same; heat put into evaporatlng water Thus, if
' more water is present than” the optimum calculated below, ‘the initi&l"
pressure rise would be less than calculated and the contlnued dralning
of the salt would' evaporate more water, but the pressure would be® less
than the peask of the initial pressure surge. L
- The following is a table of symbols and definitions of the variables

and fixed parameters:

P, -Initial temperature of fuel = 1225°F.

1
T, Initial temperature of ‘water and air. lOO i
Te Final temperature oquuel, vater, and‘alr, op

p - Density of fuel = 154.5 1b/ft3

W, VWelght of fuel released 4170 1b

f - .
wa‘. Welght of air in the reactor contalnment shell = 834 1b
L e | :
wa‘ Weight of'air in drain-tank containment volume = 579 1b
‘o
1 ]

_lls -

v.OW. o+ W = total weight of air, 1b
a a a
1 2
L Welght of water, 1b
e Average specific heat of fuel over the temperature range
£ | | = 0.544 Btu/1b-°F
c, Average specific heat of water = 1.0 Btu/1b-°F | |
W : '
e, Average specific heat of air = 0.24 Btu/1b-°F
8 o
Mww - Number of moles of steam
Mwa Number of moles of air
hg' Enthalpy of steam at T, Btu/1b
Pa Partial pressure of air
PS | Partial'pressure of HéO

The energy balance after 1 min is:

(T, -32) w.c_ + (T -32)(w c +w c. )
L. £ pf} °© w pw 1 Py

= (Tf 52) (v, e, Wy © ) +w_h_ - 0.41 x 10 (32)
| £ "1 Pa
The sum of partial pressures is:
(Tf + 460) :
P = P T+ 1%0) + P | o (33)

also,
=116~

P - P (34)
Subspitutipg known values and simplifying,
2.39 x 100 - T, (2468) | o
WW‘ = - 68 ) I . | . (35)
g
P = :"0}025561(Tf + 460) + P, o ” | (36)
P = P [1+Ww ] . . o (37)

The solutiohs to Hs. (35), (36), and’ (37) show the maximum pressure
to be 39 psig, the temperature to be 260°F, and the optimum emount of
water to be 1590 1b. |

[ 3]
PrYg

W,

-11%-
~ Appendix F
GRAPHITE COMPATIBILITY WITH SALT
Assessment of the general question of.compatibility of graphite with
molten fluoride reactor fuels has required experimental study of several

possible problems. The present status of this'experimentai.program is

descrihed briefly in this appendix.

;Chemical Interactions

Intercalation compounds. of graphite with a variety of . -pure chlorides,

_bromides, and lodides are known. Graphite 1s severely damaged when such

compounds .are formed; a 2-hr treatment with FeCl, at 300°C, for example,

3
reduces graphite to powder by formation of an intercalate.- ,

- The maJor constituents, 1iF, BeF Zth, h’ h’ of the MSRE fuel
mixtures are known from a large number of experiments to form ne such com-
pounds with: graphite., This situation is not changed when such materials
as NiFg, FeFe, and CrF2,

rides are added in apprec1able concentrations. . T

alkaline earth fluorides, -and rare-earth fluo-

A few of the possible- fission-product fluorides (MoF5, for example)

might in the pure state, form intercalation compounds with.graphite.

,However, intercalation byicompounds which react readiiy when pure can be

.firevented by dilution with nonreactive salts. The possible'intercalate
formers among the fission-product fluorides will occur only.at concentra-
tions belowflQ.Ol mole %f“ifi appears rery unlikely"that,compound formation
can occur between the graphites and any constituent of the.molten fluoride
solution. R | »

" Some cesium and rubidium isotopes of a variety of half-lives will be
formed in ‘the graphite through decay ‘of xenon and krypton isotopes which
have diffused into the graphite. ' The moderator temperatures are such that
some stability of such compounds’ as CSCEh or CsCB'must be ‘expected. Such
compounds will tend to disrupt the graphite structure and would, if
present in sufficient concentratiohs, probably disintegrate the moderator
blocks. ' The' absolute amounts of these elements so-introduced into the

graphite are 'so small that their efféct can hardly be important.
-118-

Chemical reaction of the fuel salts with some of the contaminants
which desorb from the graphite must be expected. Mixtures containing
1iF, BeF ZrFu, ThFu, and UFh in concentrations typical of the MSRE

2
fuel redgt slowly if at all with CO,, CO, and 055 though reactions such
as
002 + Fe - FeO + CO
and

Zth + 2Fe0 Ha'”2FeF2 + ZrO2

would introduce some oxide contamination into the fuel salt. Reaction
of the salt mixture with HEO as by

. 2
is quite rapid. It will probably prove impossible to remove all chemi-

2

h ) .

ZrFu + 200 - LWHF + Zr0

'sorbed oxygen-bearing species from the moderator graphite before addi-
tion of the MSRE fuel. However, the MSRE fuel mixture can apparently
accommodate up to 1300 ppm of O in solution without precipitation of a
solid oxide; this is more by a factor of 3 than that available (assuming
complete desorption and reaction) from the graphite. Moreover, if the
ZrFu/UFu ratio in the MSRE mixture exceeds 2, the first fiaterial.to pre-
, which contains nmo UO,. Since the Zth/UFu
ratio ifi?the MSRE fuel will be at least 5;Ithere seems to be no fear of.
precipitation of U’O2 from the circulating fuel.

cipate is triclinic  -ZrO

Permeation of Graphite by Fuel Mixture

Graphite is apparently wefted by some fluorides in fhemolten _
state. Treatment of graphite with molten SnF2 at 300°C and at atmos-
pheric pressure, for example, results in virtually complete penetration
of the speéimen by the salt and in a continuous film of salt over the
graphite surface. | - :

The results of a considerable series ‘of experiments indicate, how-
ever, that graphife is not wetted by mol£en fluoride mixtures containing
LiF, BeFé;‘Zth, ThFu and UFh“ This behavior is not changed by addition
of minor constituents suchas the fluprides:of structural metals. This

'behaviéf is essentially unchanged on treatment of the molten salt mixture
+ll9;

with anhydrous HF,or'stronglreducing agents such as Zr°. Superficial
evidence of wetting is obtained if the graphite;salt system at‘700°q is
exposed to the air; in that case the screen which forms on the salt ap-
pears QO'promote wetting of the graphite surfaces. ©Since the graphite -
is not wetted by the salt mixture under nonoxidizing conditiohs, the
graphite tends to resist?pehetraticn by the fuel salt into its pore
structure. : | ) ' o o '

Moderator graphite is, however, of much less than theoretical
density. About 10% of the volume of a specimen of moderator graphite
consists of' a network of interconnecting capillary pessages of a variety
of sizes. “Even though the graphite is not wetted by ‘the salt, therefore,
the salt can be forced into the graphite pores by application of external
pressure. - The amount ofrfienetration to be expected is a function of the
external pressure, the interfacie1~tension of the liquid-solid interface,
and the poreisize spectrum of the'gréphite. The penetration observed for
a given.specimen should be independent of treatment time.

Permeétion.by fused fluorides has been studied with a large. number
of graphite’specimens which were degassed under conditions that could. be
matched in the reactor and were subsequently exposed to LiF-BeF, -ThFu—UFh
mixtures at” 1500 F and at 95 and 150 psig of argon pressure. Low-density
and qulte—permeable graphltes such as AGOT are permeated to the extent
of nearly l5°vol % at elther pressure. Of a total of 31 grades of
graphite tésted, four grades (B-1, S-4LB, GT-123-82, and CS-112-8) show
salt permeation of less than 0.5% of the bulk graphite volume under
150 psig, while four others (CT-150, CEY-1350, CT-158, and CEY-G) also
show less than 0.5% permeation under ‘95 psig. No effect of treatment

time was observed in these tests.

. Compatibility in. Long-Term Forced-Flow Test

In.an*efigineering téstof'the compatibility cf*graphitefiwith‘molten
salts, 31 spécimens of a”speéiai'graphite (National  Carbon Co. GT-123-82)
‘were exposed at 1300°F in & flowing stream of LiF- BeF »~UF), - (62 -37-1
mole %) for one year in a forced-c1rculat10n loop of. INOR 8. A flow .rate

of 1.1 gpm was maintained ‘over the spec1mens with an - effectlve pumping .
120~

head of 10 péi.‘ An additional pressure of 3 psig was maintained by
pressurizing the helium cover gas, so that the total pressure on the
salt-graphite system was 13 psig. The graphite specimens (of two sizes:
11 in. long by 1/2 or 3/8 in. in diameter) were degasséd for 24 hr at
1100°F by evacuation of the test loop and were flooded with argon be-
fore the loop was charged with salt.

After the year of operation the.specimens were recovered for
examination. The graphite was clearly not wetted by the fluoride melt;
the specimens drained clean except for a few tiny spherical particles
of salt which loosely adhered to the specimens. Dimensional changes.
for the rods averaged less than 0.5 mil on the diameter; this figure
1is close to the probable error of the measurements. No weight gains
were observed. Weight losses varied from negligible to 0.05% and
averaged 0.02%; these losses may represent desorption of residual gases |
from the samples or may, perhaps, be evidence of slight erosion.
Analysis of the graphite for uranium indicated an average of 15 ppm.

The graphite was in excellent condition; it is clear that exposure

under these conditions is not deleterious to the system.

In-Pile Testing

‘The results of an extensive program of out-of-pile testing are
generally quite reassuring.. The in-pile testing has disclosed no evi-
dence which contradicts the out-of-pile tests, but the in-pile tests
have been too few to be reassuring. | | E

Two graphite crucibles (each 1.5 in. long, 0.10 in. in inside dia-
meter, with 0.025 in. wall thickness) of a high-density graphite from
National Carbon Company (similar to GT—123-82) were irradiated ih the
MIR at 1250°F for 1610 and 1492 hr, respectively, while charged with
LiF-BeFEwUFu (62-57—1 mole %) mixture containing fully enriched U255.
The crucibles were plugged top and bottom with caps of the same graphite
material,afld were enclosed in containers of Inconel. Irradiations to
integrated dosages of 1520 and 1375_kw-hr/cm§ were given to the cap-
sules. Postirradiation sectioning and inspection of the specimens re-

vealed no evidence of damage to the graphite nor any evidence that the
X

-121-

graphite structure had been perméated by the salt. It may be concluded
that no gross damage to the MSRE graphite will occur, at least during
short-term irradiation.

Two éttempts at a considerably more sophisticated'experiment*gave-
only partial success. Graphite specimens, enclosed in flexible cap-
sules fabricated from INOR-8 bellows and filled with LiF-BeF,-ThF -UF)
mixture, were exposed for 1600 hr at 1300°F in the MIR. - The fléxible

capsules were immersed in a bath of molten sodium which served as the

heat removal sink and also transmitted a pressure of lOO psig to the

graphlte -salt system. Only one of elght capsules so exposed;suryiyedf

the numerous thermal cycles, with sudden freezing and thawing of the

'salt, imposed by the MIR operation. The ome surviving capsule con-

tained a specimen of S-UA graphite. After this exposure, in which the
power density in the fluoride fuel was about 200 w/cc, the external
appearance-of the graphite was relatively-good_and the physical dimen-
sions had changed relatively little. From the increase in weight it
appeared that 0.71 vol % of the specimen was permeated with salt; this
is to be compared with out-of-pile tests in which S-4A was permeated
to 1.0% at 150 psia. |

Subsequent sectioning of the specimen and a combination of auto-
radiography, micro core drilling and subsequent analysis by counting
techniques, -and metallographic examination showed relatively high con- "
centrations of fuel near the external surface of the graphite and
along an internal chord of the specimen. Except in these surface re-
gions and along this band of apparenfly‘high-porosity’graphite within
the specimen, the graphite interior is relétively clean; in‘:general,
the fission species which can be identified in the interior are alkali
metals presumably deposited from noble-gas precursors which permeated -
the moderator. | .

It would be unwise to conclude too much from this single. specimen-
(which is far from the best available graphite), and more studies are
clearly needed. It appears, however, that no marked differenceS’be-‘-

tween the in-pile and out-of-pile behavior have yet been encountered.
l.

2.

~122-

REFERENCES

W. B. Cottrell, ed., ORNL-140T (Nov. 24, 1952)(Secret).

W. B. Cottrell et al., The Aircraft Reactor Test Hazards Summary
Report, ORNL-1835 (Jan. 19, 1955).

C. L. Davis, J. M. Bookston, and B. E. Smith, GNU IT - A Multigroup
One-Dimensional Diffusion Program for the IBM-TOL, General Motors
Report GMR-101 (1957). |

Cc. W. Nestor, Jr., Multigroup Neutron Cross Sections ORNL CF-60-3- 35
and revision (March 1960). o

L. G. Alexander et al., Preliminary Report on Thermal Breeder
Reactor Evaluation, ORNL CF-60-T-1 (July 1960).

G. G. Bilodeau et al., PDQ - An IBM-T0l Code to Solve the Two-
Dimensional'Few-Group Neutron-Diffusion Equations, WAFD-TM~TO (1957).

L. F. Parsly, MSRE Drain Tank - Heat Removal Studies, ORNL CF-60-9-55
(Sept. 1960)

Sturm-Krouse, Inc., Analyses and Design Suggesticns for Freeze
Flange Assembllies for MSRE, Nov. 30, 1960.

S. E. Beall et al., HRP Quar. Prog. Rep. CGet. 31, 1958, ORNL-265Lk,
. T. ' fi

A Meteorological Survey of the Oak Ridge Area, OR0O-99.

Met. Ann. Prog. Rep. Oct. 10, 1958, ORNL-2632, p 89-92 (Classified).

Met. Ann. Prog. Rep. Sept. 1, 1959,’0RNL=2839, é 150-153.

. Met. Amn. Prog. Rep. July 1, 1960, ORNL-2983, p 197-199.
14,

MSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p 88.

O @O~ O\ Wi

10.

H. G. MacPherson

~123=

EXTERNAL TRANSMITTAL AUTHORIZED

ORNL CF-61-2-46

INTERNAL DISTRIBUTION

Manly

L. G. Alexander - W. D,

S. E. Beall 37. E. R. Mann

C. E. Bettis 38. 'W. B. 'McDonald

E. S; Bettis 39. C. K. McGlothlan

F. F. Bla.nkenship 40. R. L. Moore

A. L. Boch . J, C. Moyers

S. E. Bolt k2, C. W. Nestor '

W. L. Breazeale 3. T. E. Northup

R. B. Briggs L4y, L. F, Parsly

F. R. Bruce 45. H. R. Payne

R. A. Charpie 46. R. E. Ramsey

W. H. Cook 47. W. D. Reel

G. A, Cristy 48, M. Richardson

J. L. Crowley 49, R. T. Santoro

F. L., Culler. 50. H. W. Savage

E. P. Epler . 51. D. Scott

W. K. Ergen 52. M. J. Skinner.

W. H. Ford 53. A. N. Smith -

A, P, Fraas 54, P. G. Smith

C. H. Gabbard 55, I. Spiewak

W. R. Gall 56. J. A. Swartout

R. B. Gallaher 57. A. Taboadsa

W. R.- Grimes 58. J. R. Tallackson

. A. G. Grindell 59.. W. C. Ulrich

P. N. Haubenreich 60. A. M. Weinberg

J. W. Hill - 6l. J. H.  Westsik

E. C. Hise 62, L. V. Wilson

P. P, Holz 63. C. 'E'o";Win't;erS

L. N. Howell 64, C. H. Wodtke

W. H. Jordan 65-66. R. D. Library

P. R. Kasten 67-68. ORNL - Y-12 Technical Library,

. R, J. Kedl - \ Document Reference Section

B. W. Kinyon 69-94. Laboratory Records Department

J. A. Lane ' 95. Laboratory Records, ORNL R.C.
96-98.

Central Research Library

T

EXTERNAL DISTRIBUTTION

99-135. H. M. Roth, AEC, ORO