ORNL-4069.txt

EARCH LIBRAR

DOCUMENT COLLECTION

cut
I

i ORNL-4069
3 445L 0487570 3 UC-80 — Reactor Technology

DEVELOPMENT OF A MODEL FOR COMPUTING

135y e MIGRATION IN THE MSRE

R. J. Kedl
A. Houtzeel

OAK RIDGE NATIONAL LABORATORY
CENTRAL RESEARCH LIBRARY
DOCUMENT COLLECTION

LIBRARY LOAN COPY
DO NOT TRANSFER TO ANOTHER PERSON
1f you wish someone else to see this

document, send in name with document
and the library will arrange a loan.

OAK RIDGE NATIONAL LABORATORY
operated by

UNION CARBIDE CORPORATION
for the

U.S. ATOMIC ENERGY COMMISSION

4"

Ll

Contract No. W-7405-eng-26

- : Reactor Division

i

DEVELOPMENT OF A MODEL FOR COMPUTING
135%e MIGRATION IN THE MSRE

R. J. Kedl A. Houtzeel

JUNE 1967

OAK RIDGE NATIONAL LABORATORY
Oak Ridge, Tennessee
operated by
UNION CARBIDE CORPORATION
for the
U.S5. ATOMIC ENERGY COMMISSION

ORNL-4069
43
iii

ADSETECT teveereessrossnusssssaesarsessssnesssassssssassssnssass
Description of the MSRE ...t ivicirivsnnesnsannsnnnns e eseseaann
Krypton-85 Experiment ...... [P UL
Description of the Problem ...... ettt eetaeeren e
Description of the Experiment .....ccceveeeneesosnncrcesnnnes
Procedure and Description of RUNS ceveeveaseesococoenes e
Analysis of the %°Kr Experiment ..... Ceareasee cenene EERE R
General Apprbaéh ........... e e,
Pump Bowl DynamicCsS cvveeeessssssosnsceosssosoasnanosassossensns
‘Xenon Stripper Efficiency ...ccciveceeean. e PRSP
Mass Transfer to Graphite ......cce..0.., R R P PR
Capacity Considerations .........oeuieeuieiieninnerenioeneanen

Xenon-135 Poisoning in the MSRE ......ccvvieeincencnnnannennannns

General DiSCUSSION v eeeeeesseerossssssssossorasesocsossoasassass ‘

Dissolved Xenon Source Terms and Considerations ...cecse..
Dissolved Xenon Sink Terms and Considerations ............
Other Assumptions and Considerations ..ececeeeeeeeieeeeses
Xenon-135 Genération Rate ....... eheeeaas e eeeesseae e
Xenon-135 Decay Rate in Salt ...veieereessencsnnnns [
Xenon-135 Burnup Rate in Salt ...cveeiieivinrenieeccnsanens
Xenon-135 Stripping Rate .. .iiii it eiin i,
Xenon-135 Migration to Graphite .....ovceeivereennvenenns .
PR | Xenon-135 Migration Rate to Circulating Bubbles ..... e

Xenon-135 Concentration Dissolved in Salt ....covveeniesss
. | Xenon-135_Poisoning Calculations ..ccecevcorvensencenseronsons

Estimated 13°Xe Poisoning in the MSRE Without
Circulating Bubbles ..uveereeeeencersessvacersronronnocnos

Estimated 1°°Xe Poisoning in the MSRE With
Circulating BubbleS teeeeeeeereeneanrencessssssassnnnonns

CONCLUSIONS setevasesvoaseossosoosssssasasesoossssonasssaasssansass
Acknowledgments ........................................ e

References .v.oveeciesssssaasrssassssses e e s s e acseas e s e resses s s an s s ena

Page

12
18
18
25
29
29
40
41
41
41
42
43
42,
44,
44,
44,
45
4
48
49

49

53
57

59

60
Appéndix A.
Appendix B.
Appehdix C.
Appendix D.

iv

MSRE ParametersS .eveeeeecscescearescasonsotcaacosses

Salt-to-Graphite Coupling ............ P |

Theoretical Mass Transfer Coefficlents «.ieeveen ;..

Nomenclature

65
66
71

73

o
A

»

>
DEVELOPMENT OF A MODEL FOR COMPUTING 135Xe
‘ MIGRATION IN THE MSRE

- R. J. Kedl A. Houtzeel

~ Abstract

The Molten Salt Reactor Experiment (MSRE) is a fluid-
‘fueled reactor with a potential as a thermal breeder. - Because
of the importance of neutron economy to the breeder concept, it
is necessary to know the dynamlcs of 135Xe in the circulating-
fuel system. 'There are several "sinks" where xenon may be de-
posited from the fuel, notably in the gas space of the pump
bowl and in the voids of the unclad graphite of the reactor
core. Since *25Xe in the core impairs the neutron economy, it
is important to understand the mass transfer mechanism involved
and the parameters that may be varied to control it.

‘This report deals prlmarlly with developing a model for
computing the migration of '2°Xe in the MSRE and with experi-
ments conducted to establish the model. A preoperational ex-
periment ‘was run in the MSRE with 85Ky tracer, and many of the
gas-transport constants were inferred from the results. Equiva-
lent transport constants for calculating the 135%e migraticn

-gave a poisoning of about 1. 4% without circulating bubbles and
. well below 1% with bubbles. Preliminary measurements made on
the critical reactor show xenon poisoning of 0.3 to O. 4% Since
physical measurements confirm that there are bubbles in the sys-
.. tem, the conclusion is drawn that the .computation model, the.
- krypton experiment, and reactor operation agree.

The goal of the Molten Salt Reactorwfrogram is to develop an effi-
01ent power- producing, thermal- breedlng reactor.‘ The Molten Salt Reactor
Experlment (MSRE) is one step toward that goal although it is not a
hreeder. Nuclear poisons, notably_1?5Xe, can detract 51gn1f1cantly from
the breedlng potentlal It was therefore considered appropriate to in-.
vestlgate in some detall the dynamlcs of" noble gases 1n thls pllot—plant—
scaled reactor and with this information to predlct quantltatlvely the
xenon p01son1ng.l

llhe 135Xe p01son1ng is a function of the steady-state 135Xe concen-
tratlon in the reactor core. It is computed by balan01ng the rates as-
soc1ated with the varlous source and sink terms 1nvolved. Since the MSRE

is fluid fueled xenon and iodine are generated directly in the salt and
the source term is essentially a constant. The sink terms, however, are
more complex. Xenon may be removed from the system via a stripping de-
vice, it can decay or be burned up in the salt, or it may be absorbed by;
the graphite and ultimately decay or burn up. Xenon may also be absorbed
by circulating helium bubbles, which complicate the model because of
their reiatively unknown dynamics. 7

This report is concerned principally with developing a model for
estimating the 135%e poisoning in the MSRE. waevef, the first part dis-
cusses an experiment, referred to as the krypton experiment, in which

some of the more elusive rate constants were evaluated.-

‘DESCRIPTION OF THE MSRE

The MSRE is a circulating-fluid-fueled graphite-moderated single-
region reactor. The fuel consists of uranium‘fluoride dissolved in a
mixture of lithium, beryllium, and zirconium fluorides. The normal oper-
" ating temperature is 1200°F, and the thermal power level is 7.5 Mw. The
reactor system consists of a primary loop containing the core and a sec-
ondary loop to remove the heat. Our cohcern is only with the primary
loop,>a schematic diagram. of which is shown.ianig. 1. Essentially it
consists of a pump, heat -exchanger, and reactor core. A detailed de-
scription of the MSRE is contained in Ref. 1, and pertinent design pa-
rameters are listed in Appendix A. | |

Figure 2 shows detalls 6f the fuel pump. If is rated for 1200 gpm
at a 48.5-ft head. fhe volute is cdmpletely enclosed in a vessel re-
ferred to as the pump bowl, which serves primarily as an expansion vol-
ume for the fuel salt. The overflow tank serves as an additional exXpan-
sion volume for the system and is fed by an overflow line that penétfates
up into the pump bowl. The normal operating helium pressure in the pump

bowl 1s 5 psig, which is also the pump suction pressure. There is a

continuous. flow of salt and helium through the pump bowl. The pfincipal'

salt flow is through the xenon stripper, which is a toroid containing

——

numerous small holes that spray salt through the helium atmosphere. The

salt flow is controlled with an orifice and has been calculated to be

o)

-P(‘l
3
v
ORNL-DWG 67-1955
T
TOTAL FLOW - { BUBBLER .PRESSURE Ref. FURGE GAS w24 'STD liter/min
) i in] - : ' A
09 STO liter/min BUBBLER (2) O o8 33 STD liter/min
. . ———
YXENON STRIPPER SAMPLER-ENRICHER PENETRATION -
SALT LEVEL
PUMP BOWL
VOLUTE SECONDARY
'SYSTEM
PIPING
OVERFLOW TANK| ‘

. CONTROL ROD

CONNECTIONS

2, .. .

SALT FLOW RATE,
N 00 gpm
«
= I //
’{/ ercese rance
UhFREEZE FLANGE
CORE
RN
FREEZE VALVE FOR
JFILLING AND DRAINING
Fig. 1. Schematic Diagram of MSRE Primary Loop.
about 50 gpm, but it has not been measured directly. The resulting high-
1 velocity jets impinging on the molten salt cause a large amount of splash-

\1ng "and turbulence, consequently, bubbles are transported into the loop.

N

‘It will be shown later that a very small quantity of” c1rculat1ng bubbles
has a very pronounced effect on xenon dynamlcs. Tn addition to the strip-
per.there’is salt flow of about 15 gpm from behind the impeller, through
a lahyrinth aleng the shaft, and into the pump bowl. The principal.helium
flow through the pump bowl is 2.4 std‘liters/min purge down the shaft to

‘ prevent ‘radioactive’ gases ‘and salt mist from reachlng the bearlng region

of the pump . There is an additional helium flow of 0.9 std llters/mln

. , ~
B
4 -
R
ORNL-LR-DWG-56043-BR!
. | ]
T
C — 3
—
SHAFT ' gi:; —3 .. SWATER
COUPLING . 0 ) A, o, COOLED
. A MOTOR
SHAFT SEAL e — D )
NS S e e vty § R R
LEAK = _ 1
DETECTOR— : : LUBE OIL BREATHER
LUBE OIL IN } } ‘ -
BALL BEARINGS: ; S e
F R
(Face to Face) BEARING HOUSING =~
BALL BEARINGS ! GAS PURGE IN
- $
(Back to Back ) SHAFT SEAL
LUBE OIL OUT _ SHIELD COOLANT PASSAGES
SEAL OIL LEAKAGE ' S | | (In Parallel With Lube 0Qil)
DRAIN SHIELD PLUG s
By : : I 4
'LEAK DETECTOR GAS PURGE - OUT
SAMPLER ENRICHER GAS FILLED EXPANSION )
(Out of Section) ] ‘ i SPACE
g | , XENON STRIPPER
HELIUM BUBBLER ‘ (Spray Ring)
= = d— SPRAY
OPERATING v | T M7 =1 [
LEVEL o - el _//// >
S E
HELIUM BUBBLER ' ' PUMP BOWL
CONFIGURATION \ I
To Overflow Tank )

Fig. 2. MSRE Fuel Pump.

from two bybblers and one pressure-referenée leg, which comprise the bub-
‘bler level indicator. Helium for each bubbler goes through a semitoroid "
located in the pump bowl, as shown in Fig. 2. Helium enters the semi-
toroid at the end and leaves in the middle;. therefore, half the semi-
toroid is stagnant gas. This stagnant kidney will be referréd_to,in the
analysis of the krypton experiment. ‘ | | |

Figure 3 is an isometric view of the reactor core. Fuel salt enters

the core vessel through a flow distribution volute and proceeds down an

%
3

annular region bounded.by the vessel wall and the moderator container.

The fuel then travels upward through the graphite moderator region and

F

out the top exit pipe.

The graphite is unclad and in intimate

igure 4 shows how the moderator bars fit to--

gether to form fuel channels.

ORNL-LR- DWG 61097RS

n
T
(]
T
o
o 3
n 7]
Fw w o )
-2 z z
=1 - =
oo - @ o
Zz - o
OCo < o
O J® -
o
W z o)
o - a
O . o
b O wn
wZ O - N
Lw _.n.w_
w Q
-4

REACTOR ACCESS PORT

1<
o
-
>
@
1
-
a
o
=
o
-
TR

— QUTLET STRAINER

MODERATOR

AR 2 J \) {0 \\;W\v,”“;.rlfl.fl.rl%pa@géfififigw

e . B .

L ACRA, o) | s .

B\ fl\'tv L. . YA e B i
ST e g | o fl@ it
wll)) BN s O | =
\,...!.‘nn.\ ! Gm 5 =)

" & i &\m *h A b -y

GRAPHITE SAMPLE ACCESS PORT

FUEL OQUTLET

CONTROL ROD THIMBLES

/
i
4
{

o &
fo
i
:
|

i
¢
o

;

VESSEL DRAIN LINE

CORE

CENTERING GRID
FUELINLETJ/fi

GRAPHITE-MODERATOR\

_ STRINGER !

REACTOR CORE CAN—"
REACTOR VESSEL—
ANTI-SWIRL VANES

SUPPORT GRID

Reactor Vessel and Access Nozzle.

Fig. 3.
(3

contact with the fuel salt, and therefore ;35Xe may diffuse into its
porous structure; the graphite acts as a *?°Xe concentrator in the core.
The theoretical void percentage of MSRE graphitel(grade CGB) is 17.7%,
and slightly over half of it is accessible to. a gas such as xenon. ©
Other pertinent properties of this graphite are listed in-Appendix A,
and more detailed information is available in Refs. 3 and 4. The rate
at which 12%Xe diffuses to the graphite is a function of the salt-to-
graphite mass transfer coefficient, which is, in turn, a function of the

fuel-salt Reyndlds number.

v

The moderator region can be divided into three fluid dynamic regions "
of interest. First there is the bulk of the graphite (~95%), which is
characterized by salt velocities of about 0.7 ft/sec and a Reynolds num- ®

ber of about 1000. One would expect laminar flow; however, the entrance

ORNL-LR-DWG 56874 R

PLAN VIEW

TYPICAL MODERATOR STRINGERS

SAMPLE PIECE

Fig. 4. Typical Graphite Stringer Arrangement.

.‘v
'

to these channels is orificed and quite tortuous because of a layer of
graphite grid bars across the bottom of the moderator that are used to
space and support the core blocks. The effective mass £fahsfer coeffi-
cient is probably somewhere between the coefficients for laminar flow

and turbulent flow. The second fluid dymamic regién.is éompdsed of the
centermost channels in the core.(about 18). They do not have orificing
grid bars below them, so the fuel velocities are higher, about 1.8 ft/sec,
and give a Reynolds number of about 2500. Accordingly the mass transfer
coefficient is higher than for the bulk of the graphite and is for tur-
bulent flow. This region comprises about 1.5% of the graphite and is in
a zone of high nuclear importance. The third fluid dynamic'region is

the lowerhlayer of graphité'grid bars mentioned above, which do the ori-
ficing. These grid bars are subject to high salt velocities, a maximum
of about 5 ft/sec, and comprise a fluid dynamic entrance region. In ad-
difiion, the jéfs'forméd-by tfie grid bars impinge on the bottom of the
core blocks. The éntire région then is.- subject to much higher mass trans-
fer coefficients than the bulk graphite. This region is not.too well
defined but probably comprises about 3 5% of the total graphlte It is

1n a zone of very low nuclear 1mportance.

KRYPTON~85 EXPERIMENT

- Description of the Problem

_ Xenon-135 poisoning in the MSRE was considered prev1ously,5 7 but
these calculatlons were generally of an approximate design nature be-
cause of lack of 1nformatlon on the values of the rate constants 1n-
volved. In order to calculate the steady-state 135Xe poisoning in the
reactor, it was first necessary to compute. the }?5Xe_concentrat10n dis-
éolvéd in the salt. This was done by equating the various source and
sink rate terms involved and- solving for the .xenon concentration.. The
most significant_135Xe source term is that_whiqh comes from the decay

of .72°I; in addition a small amount is produced directly from fission.

The sink terms are discussed in some detail later, but we will initially
consider only the following terms and their associated rate constants:

Principal Rate

135 :
. Xe Sink Term Constants Involved

1. Dissolved '?°Xe that may be Stripping efficiency
transferred to the off-gas ' '
system via the xenon strip-

per
2. Burnup of dissolved 135%e as Burnup constant
it passes through core ‘
3. Decay of dissolved '3°Xe ' Decay constant
Migration of dissolved 135%e Mass transfer coefficient, dif-
to the graphite; ultimately fusion coefficient of xenon
this 13°Xe will either de- in graphite, decay constant,
cay or be burned up and burnup constant
5. Dissolved *3°Xe that may be = Mass transfer coefficient, de-
transferred into circula- cay constant, burnup constant,
ting helium bubbles, if and bubble stripping effi-
present; this 135%e will ul- ciency in the pump bowl

timately be burned up, de-
cay, or be stripped in the
pump bowl
Thé stripping efficiency of the pump bowl spray ring was méasured
at the University of Tennessee as part of a-masters degree thesis.®»?
This work was done with a CO,-water system maintained bubble free and
later confirmed with and O,-water system, also maintained bubble free.
A prototype of the xenon stripper was used in these tests. It was felt
desirable to check the results with a xenon and salt system, particularly
with circulating bubbles present. |
Xenon-135 burnup and decay rates are relatively well known. Migra-
tion of xenon to graphite is controlled by the mags transfer coefficient
and by the diffusion coeffiéient of xenon in graphite; The mass trans-
fer coefficient can be estimated from heat-mass transfer analogies (see
Appendix C), but the unknown mode of fluid flow (laminar or turbulent),
the unwettability of graphite by moltep salt, the question of mass trans-
fer to a porous surface rather than a continucus surface, and some natu-
ral resistance toward assuming a high degree of reliability for the heat-
mass transfer analogies made the estimated coefficients seem questionable.

The quantitative effect of circulating bubbles was almost completely

L))

>
"

unknown, except that their effect would be prominent because of the ex-

tremely low solubility of xenon -in salt. In. addition there may be other

-effects not considered. Generally the state of knowledge of the rate

constants was considered somewhat wénting. Fach of these rate constants
could be investigated individually in the laboratory, but this would be
too expensive and time consuming. Rather, after other approaches were
considered, it was decided to conduct a single summary experiment on the
reactor and extract as many of the rate constants as possible, or at
least set limits on them. The experiment was referred to as the krypton

experiment.

Description of the Experiment

Essentially the experiment was divided into two phases and took place
during the precritical period of MSRE operations. The first phase was
an addition phase and consisted of‘adding 85Kr to one of the pump-bowl
level-indicator bubbler lines at a steady rate for a period of time. Dur-
ing this phase-the pump bowl reached some equilibrium 85Ky concentration

almost immediately; then the salt dissolved krypton via the xenon-strip-

per spray ring; and the graphite absorbed krypton from the salt. The

second phase began by turning off the krypton flow but maintaining all
bubbler and purge helium flows. Then the reverse processes took place.
The pump bowl-purged clean of krypton;'fhe salt was stripped; and finally
the graphite was leached. During the entire experiment the off-gas line
was monitored continuously with a radiation counter. By analyzing the
krypton concentration decay rate in the off-gas during the stripping
phase of the experiment, we evaluated éome of the rate constants in-
volved. The experiment had the advantage of evaluating the actual reac-
tor under operating conditiohs rather than models under simulated con-
ditions. The experiment had.thg limitation that several parameters had
to be e#aluated from essentially a single set of data and were therefore
subject to a certain amount.of personal interpretation. Also, transient
experiments are inherently more difficult to analyze than steady-state
experiments. Krypton-85 was'chosen for the experiment primarily for ease
of continuous monitoring‘ét low concentrations in the off-gas line; also

its low cost and availability were considerations.
10

~ Figure 5 is a schematic diagram of the krypton experiment facilities.
Basfically, it consists of an addition station and a monitoring station.
The additiofi station controls the flow of an ®°Kr-He mixture into bubbler
line 593. The normal bubbler flow of fiure helium (0.37 std liters/min)
was maintained to transport the krypton-helium mixture into the pump bowl.
The reactor contains two bubblers. The second bubbler was used to per-
form its various reactor control functions. |

The krypton-helium container was made from 12-in. sched.-80 carbon
steel pipe and pipe caps and was about 5 ft long. It was hydrostatically
tes%ed at 520 psig. On one end was a U-tube and valve afrangement that
Was!used‘to transfer 85Kr from its shipping container to the experiment
container. The transfer was accomplished by first evacuating the experi- | | N
ment container and then opening the valve on the shipping container. This
resulted in about 95% transfer. The remaining krypton was transferred by

using the U-tube as a cold trap and freezing it with liquid nitrogen.

ORNL-DWG 67-1956

ROCKER DRIVE

|, —MSRE BUBBLER . ¥
LIMITING FLOW VALVE ' ' R:\%gll\l?TTéJ%N ¥

' e I

PRESSURE REGULATOR A o

E A|_|:‘| - Kr-He
\. /' CONTAINER FREEZE , -4 5 o
i -

X

e

MONITOR
SYSTEM

CONTROL VALVE

ADDITION SYSTEM

VALVING SET UP FOR ADDITION PHASE OF EXPERIMENT

|m. . f
I, MSRE OFF GAS

LINE 522

Fig. 5. Schematic Diagram of Experimental Equipment.

(2.

11

This two-step process resulted in the almost perfect transfer of the 120
curies of %%Kr purchased. The experiment container was then pressurized

to 180 psig with helium. After the first run it was further pressurized

‘with helium to 275 psig. Dilution was necessary in order to have enough

gas to measure and control adequately. The-original. 120 curies of 85Ky

amounted to only about two“liters, and this had to be added continuously

to:the.reactor for a period of several days. |
Based on expefience of the personnel in the Isotopes Division of

ORNL, krypton mixed with helium will tend to settle out over a period

of time. To counter this effect the krypton-helium contéiner was equipped

with a hermetically sealed agitator. It consisted of an 8-in. aluminum

‘ball inside the tank that rolled back and forth as the tank was rocked.

A large coil of 1/4-in. stainless steel tubing was located between the
krypton-helium container and flow control equipment to compensate for
the rocking motion. The limiting flow valve was set to limit the flow
from the container to about 20 std liters/hr'in case of a complete rup-
ture downstream. - The remainder of the flow control system consisted of
conventional filters‘(5 to 9 u), pressure gages, and low-capacity valves.
The flowmeter was a Hanover matrix type and was calibrated for various
outlet pressures.

As shown in Fig. 5, all the reactor off-gas from the pump bowl went
through the monitoring station. It could pass through either one of two
identical monitors or a bypass line. The monitors were labeled A and B.
Monitor B was used for all runs. Monitor A was intended as a spare but
was never needed. Théy were designed for a range of five decades of ac-

tivity. ZEach consisted of four amperex 90NB GM tubes, which were shielded

as follows:

. GM Tube

No. Shielding -

1 - None

100 mg of plastic per cm?

~100 mg/cm? plastlc window (7.62 X 2. 54 cm)
in 5.9 g/cm brass container

4 5.9 g/cm brass container
12

The four GM tubes were suspended in a 2-liter stainless steel labo-
ratory beaker. The monitors were calibrated with small samples of 8 5kr.
During the first run.of the experiment, it was found that the plastic
shielding on GM tube No. 2 absorbed 85kr and gave a false count rate;
also it affected other tubes in the array. . To correct for this, the plas-
tic waé removed and GM tube No. 2 became identical with tube No. 1. The
GM output was fed into a decade scaler and a count rate meter; The de-
cade scaler was used for recording data, and the count rate meter was |
used for éxperiment control assistance.

Much consideration was given to the safety aspect of handling 120
curies of 85Kr. The half-life of 8%Kr is 10.3 y and it gives off 0.695-
and 0.15-Mev beta particles and a 0.54-Mev gamma ray. The daughter prod- l

L)

uct is 85Rb, which is stable. The area in which the experiment was con-
ducted was equipped with radiation detectors énd air monitors. A con-
tinuous flow of air (17,000 to 20,000 cfm) was maintained through the re-
actor building and released to the atmosphere through a 100-ft stack.
Bricks were stacked around the krypton-helium tank, and the activity level
outside the bficks was negligible. Special beta-sensitive monitor badges
were worn by personnel operating the experiment. Detailed procedures for
transferring 8?Kr, pressurizing the container, and conduéting the addition
and stripping phases of the experiment were written and approved by ap-

propriate personnel.

Procedure and Description of Runs

The procedure used to start the addition phase was to adjust the

krypton-helium container regulator so that the pressure gage just up-

O

- stream of the main flow control valve was about 10 psig, that is, about

5 psi over the pump bowl pressure. The flow rate was then controlled , .
with the main flow control valve. During the addition phase the system

was checked every 1/2 to 1 hr, and the flow control was adjusted as nec-

essary to maintain a constant activity in the off-gas line. The krypton-

helium container was agitated for about 15 min every 2 to 4 hr. .For

various runs. the krypton-helium injection rate ranged from 2 to 6.3 std

liters/hr but was held constant for each run.
)

13

'Zero time in the procedure was defined as the time when the krypton-
helium flow was turned off. This was accomplished by closing the krypton-
off valve and then the regulafing valve. It‘took a minute or so before
the monitor started droppihg because all the lines had to be purged. At
the start of the stripping phase, a l-min count was taken every 1 1/2 min.
The times gradually increased until at the end of the long runs (2 and
3) a l/2—hr count was taken every hour. Note from Figs. 1 and 2 that
there are two essentially stagnant lines entering'the pump bowl, the sam-
pler-enricher line and the overflow line. These lines were purged free
of 87Kr before the stripping phase started and at various times during
the stripping operation. |

Six 8%Kr addition and stripping runs were made. Table 1 summarizes
the operational parameters in these runs. Figures 6 through 11 show the
results of these runs. The count rate in the off-gas.monitor is plotted
against time during the stripping phase and has been corrected for dead
time of the GM tubes. Nb correction was necessary for the decay of 8 5Ky
because its half-life is so long compared with the time scale of eéfih
run. As pointed out previously, the data from run 1 are erroneous be-
cause of 8°Kr absorbed.on the plastic shielding a GM tube 2. This plas-
tic was removed for subsequent runs. Nevertheless, as an added check,
the monitors were purged periodically with pure helium, and a background
count was measured. In all cases after run 1 the background count for
tubes 3 and 4 was less than 15 cpm.

Objectives associated with each run were the following:

Run No. ' Objective

1 Check adequacy of equipment and procedures

First of two long-term runs : get a feeling
for the mass transfer coefficient from salt
to graphite

3 Second long-term run: obtain good values for
mass transfer coefficient to graphite

4 " Determine stripping efficiency and other short-
term effects with salt level in pump bowl at
61% scale

Same as 4, with pump bowl level at 70% scale
6 Same as 4, with pump bowl level at 55% scale
Téble 1. Summery Description of Runs of Krypton. Experiment

Starting Time

Pressure In Pressure In

Mean Count Ratea

Total Total Time

Salt Ievel in

Total He Flow

CKrypton—helium container pressurized to 275 psig with helium between runs 1 and 2.

L2

of Addition oy con- Kr-He Con- Kr-He In Off-Gas Line  Kr-He of Pump Bowl Through Pump
Run Phase of . . Injection - s ‘s - from Bubbler :
) tainer at tainer at During Addition Addition Stripping : Bowl (Purge Plus
No. Experiment Flow Rate Level
Start of Run End of Run . Phase (hr) Phase - Bubbler Flows)
; : (std liters/hr) . Indicator td 1iters/mi
Time Date (951g) (psig) (counts/min) (hr) (% scale) (std liters/min)
1P 1420 2/5/65 180 179 2.03 3570 6 14 71 3.3
2% 1130 2/6/65 275 240 3.57 4470 57.5 62 60 to 70 3.3
3 1613 2/11/e65 240 81 3.67 4429 279 149 &0 3.3
4 1545 3/1/65 81 75 6.30 7340 5.9 5.0 61 3.3
5 1020 3/2/65 75, 70 6.2 7081 5.5 7e2 70 3.3
6 0920 3/3/65 70 66 6.33 7149 5.3 12.3 55.5 3.3
®Count rate as measured by monitor B4, corrected for dead time, and averaged over the en-
tire addition phase.
'bl20 curies 8%Kr added to krypton-helium container, and container pressurized to 180 psig .
with helium before run 1.

71

(.)
COUNT RATE (counts/min)

n *
ORNL-DWG 67-1957
1)
\
MONITOR B2
]
10 \
1
\
5
\MONITOR B3
| N\
2 4 ——
\MONITOR B4
4 L ]
10 A
‘ -
\
.
||
5 -
O . 50 100 150 200 250 300 350 400 450 500 550

. 800 650 700
TIME AFTER ®°Kr FLOW TURNED OFF (min)

Fig. 6. Results of Krypton Experiment Run 1. Count rate in off-gas
monitor corrected for dead time.

T
ORNL-DWG 67-1958

16

105

(Uwysgunod) 31y INNOD

TIME AFTER ¥ FLOW TURNED OFF (hr)

T T T I,
uw O¢ H04 440 dWNd T3N3
. O
. (o)
N ©
. 10
. '
u P .
—HES N b
ok .
u3 .
0,.= " :
1 o) M
m = . o
] e ]
0w . ¥
= .
[ & .
2y :
& Z H
8o o < .
B [ . <+
Laf SIWIL € MNVL MOTIH3N0 omom:_n_ _
o <1 | I
@ Eg 3NIT H3dWYS a39Hnd
= ES N
P4 W| .
| ] 2 =20 . o
S mm : +
|1 .
- .
ad .
. 1)
MO14 §3788N8 1SNNQY B o
!
o o
__. "
ONIL31HQ MOTd ¥378d9nd§ °,
. ©
- oJ
hd N
[ ol
L]
H
L] O
N aJ
g .
»
STNIL € INIT MOTIHIAO Q3o8Nd |
. o~
-o @
uw ¢ 440 dWnd_,. 1"
(NOILYNIWVLNOD ON) oz:0moxu<m w_o_:zoz m_mxozu...
1 il 1 » [~
_ KR4
SIWIL 9 HNVL -MOT4H3IA0 ouom:\n_
Y} -._-i-o
—- _
o
[Te] o 40, 10 ol ..._mwV [Te)

Results of Krypton Experiment Run 2.

Fig. 7.

(x

A 1
COUNT RATE (counts/min)

ORNL-OWG 67-1959

w ALLLL COUNT RATES CORRECTED FOR DEAD TIME
5 |—a ALL COUNT RATES CONVERTED TO MONITCR B3 READINGS
* <ZI STAGNANT OVERFLOW LINE AND SAMPLER LLINE WERE PURGED
— - WITH CLEAN He BEFORE 85Kr FLOW WAS TURNED OFF
= Z "PUMP BOWL LEVEL 60%
e =4
. ¥ Q
* g E Z O
w7 =Z
. R 18} <
2rx=z—z Z -
e - [@) | w O
Qo |-_I-|L|.I = % =
(Y= "'_", @xr = o =
wE Yo o z w
(3] = [FETY O =
%g O E:JZ ] 5
10 Hig o—2a Y g &%
—{f e Or 5 =
|\ ___ &= o W P 9
S5< w - S T w
N S — T ) [GREY o W s
o= o7 WD
_* S o >
) aw = o
e * ) w B—'
e, oS
"M.o.a... 0a.®" ‘ %q
LX) ........,...... aw
..'l.c...ooo.o...........
i Theten, ooo.-o"..'.'uuolo.'.o
® 0040000000000 0, o
> - , s "% ..ool..........r
10%: - : :
0 ; 10 - 20 30 _ 40 50 60 70 80 90 100 150 120 130 140 150

TIME AFTER 85Kr FLOW TURNED GFF (hr) -

Fig. 8. Results of Krypton Experiment Run 3.

LT
COUNT RATE (counts/min)

18

ORNL-DWG 67-1960

5
v ALL COUNT RATES CORRECTED FOR DEAD TIME
* ALL COUNT RATES CONVERTED TO MONITOR B3 READINGS
. STAGNANT OVERFLOW LINE AND SAMPLER LINE WERE PURGED WITH
* : CLEAN He BEFORE &°Kr FLOW WAS TURNED OFF
. PUMP BOWL LEVEL 61%
2 Py
L]
.
L ]
109 .
L ]
hd w
. =2
0 3 -
5 5 . ff.j
: :
.. q
wy
* [a)
b =
» <
L] w
. =z
2 .: 5
., g
.'. ':LJ
- :
4 * >
10 o 2
(N [m]
~~. o
e, . %
® Seee . . o
.
5 H L .‘ .
L .
. [] - . . . ) . .
| ®
0 30 60 90 120 150 180 210 240 270 300

TIME AFTER 83Kr FLOW TURNED OFF (min)

Fig. 9. Results of Krypton Experiment Run 4.

ANAIYSIS OF THE &°Kr EXPERIMENT

General Approach

Analysis of the 8°Kr experiment is concerned with the stripping

phase of the runs. The principal stripping processes involved, in order

of occurrence, are purging the pump bowl, stripping the salt and asso-
ciated circulating bubbles (if present), and then leaching the graphite.

Other leaching processes of no fundamental interest but of importance

because they contribute to the measured flux decay curve are diffusion

out of the stagnant bubbler kidney described earlier and leaching 8 5y
that may have been trapped in gas pockets located in the primary loop.
Locations of potential gas pockets in the loop are in the freeze flanges,

graphite access port,.and the spaces formed by the assembly of the core

&y

0
COUNT RATE (counts/min)

o
o

&)

" ORNL-DWG 67-1961

TIME AFTER ®5Kr FLOW TURNED OFF (min)

Fig. 10. Results of Krypton Experiment Run 5.

[, ALL COUNT_ RATES CbRRECTED FOR DEAD TIME
ALL COUNT RATES CONVERTED TO MONITOR B3 READINGS
> STAGNANT OVERFLOW LINE AND SAMPLER LINE WERE PURGED WITH
* CLEAN He BEFORE 85Kr FLOW WAS TURNED OFF
—* PUMP BOWL LEVEL 70%
~
'o
. .
. Y
. \ .|
o. 5
., —J
. a
LY
- s
bt oy
b 2
'.O.-... < - (n_'g
.....lo.......'.. E
(XN ] ®e g0 ssgle + . *
* . e__e e
* | * * . . ’
) . .
0 30 60 S0 120 150 180 210 240 270 300 330 360 390

420

6T
ORNL-D‘WG 67-1962
. ALL COUNT RATES CORRECTED FOR DEAD TIME
. ALL COUNT RATES CONVERTED TO MONITOR B3 READINGS I
|’ STAGNANT OVERFLOW LINE AND SAMPLER LINE WERE PURGED WITH
" CLEAN He BEFORE 83Kr FLOW WAS TURNED OFF
. PUMP BOWL LEVEL 555%
L ]
R
[ ]
L ]
..
[ ]
—— [ ]
c -
E N
S~
w .
= .
3 .
3 L ]
w °
= .
q L ]
o (Y
E 2 %
=} ]
o °
Q ‘\
\
104 \b,
o
.~.‘~0
oo
se o ..
[ ]
5 v e T
~ 8 ® - !
[ ]
L) * . - . .
L ] .
, | ‘
o} 50 100 150 200 250 300 - 350 400 . 450 500 550
TIME AFTER 85Kr FLOW TURNED OFF (min}

600
Fig. 11. Results of Kry‘ptbn Experiment Run 6.

3

650

700

€]

Oc
‘\ .

»

)

21

blocks. - There are various bits of evidence that pockets actually exist,
although their location and capacity are not certain..

.The fuel salt circuit time around the loop is 25 sec, which is short
compared with-the time scale of the stripping process involved, so the
fuel loop can be considered as a well-stirred pot. At any Specific time
therefore,-the-kryptbn concentration is considered to be constant through-
out- the loop.‘ In the simplest case, it can be shown that each transient

stripping process, when unaffected by any other stripping process, will

.obey an exponential decay law; that is at time t,

—Qt

Kr flux, = Kr flux_ e :

t

In the actual case, however, each stripping process will affect every
other stripping process to a greater or iessér extent. Note that pump
bowl purging, salt stripbing, and graphite leaching are series processes;
that is, they occur in the sequence given; while leachings of the several
graphite regions are barallel processes; that is, they occur simulta-

neously after the krypton concentration in the salt starts to drop. Quali-

" tatively, the measured decay curve would be expected to be the sum of the

contributions of each leaching process, as shown in Fig. 12. Note that

ORNL-DWG 67-1963

CUMULATIVE Kr FLUX FROM REACTOR

_PURGING Kr IN PUMP BOWL
STRIPPING Kr IN SALT

LOG COUNT RATE (counts/min)

TIME

- Fig. 12. Qalitative Breakdown of Cumulative Flux Decay Curve into
Its Components. ' '
22

the count rate in the off-gas line is plotted on the ordinate and is a
unit of concentration; however, since the off-gas 1is fiowing at a con-
stant rate, it also represents the 85Ky flux leaving the reactor. Each
of the component curves should approach an exponential decay after the
initial transient.

Now the problem is to separéte,these individuql processes from the
measured flux turve, with the realization that there may be dther leach-
ing processes not accounted for. It should be pointed out again that
the breaking down of a single composite data curve into several individual
processes and the determination of rate constants for each is quite com-
Plex and necessarily subject to a certain amount of personal interpreta-
tion.-

Two approaches were used in analyzifig the data. The most success-
ful method consisted of an exponential peeling technique, as shown quali-
tatively in Fig. 13 for run 3. The assumption is made that the tail of
the decay curve is determined completely by leaching the slow (bulk) rate
constant graphite. The procedure was fhen to determine the equation for
the slowest exponential that fit asymptotically on the curve and subtract
it from the data. ‘The next exponential equation that fit asymptotically

ORNL-DWG 67-1964

CUMULATIVE Kr FLUX FROM REACTOR

STRIPPING Kr IN SALT
PURGING Kr FROM PUMP BOWL

—‘(-‘

Kr LEACHED FROM VARIOUS . —————————————
GRAPHITE REGIONS

LOG COUNT RATE (counts/min)

TIME

Fig. 13. Actual Method Used to Break Down Cumulative Flux Decay Curve
into Its Components. '
N

23

on the rémaining data was then determined and again subtracted. This

‘procedure was repeated to a logical conclusion; that is, until continued

subtracting from the data gave negligible values. This procedure for

run 3 resulted in five exponentials, which is somewhat significant be-
cause of the five major stripping operations (pump bowl, salt, including
bubbles, and three graphite regions). This general approach;is in error
in that it assumes that each strippifig process starts at zero time and
proceeds independently at all others. It will be geefi, however, that this
approach is adequate for both the vefy slowest rate constant processes

(leaching bulk graphite) and the very fastest rate constant processes

(purging pump bowl), but it is inadequate for intermediate procesées

(stripping salt and faster rate constant graphite). Figure 14 shows the

results of this exponential peeling process on run 3. The rate constants

"are a function only of the slopes of fhe exponentials involved, so abso-

lute calibrations of the monitor and detailed knofiledge of the &%Kr con-

centrations are not necessary. Numerical results of peeling run 3 and

‘their'intefpretation are given in Table 2. This approach to analyzing

the data was the principal method used. It is of necessity confined to .
the fastest and slowest rate processeé involved. But when applied to
these processes,‘the results have a high degree of reliability, as will
be seen.

A second method of analysis of the data was undertaken primarily as
an attémpt to determine rate constants for the intermediate processés in-
volved, such as stripping efficiency and mass transfer to the faster rate
constant graphite regions. In this approach, unsteady-state differential
equations were set up around the pump powl gas phase, fuel salt, and
three graphite regions. The resulting five equations could be solved
simultaneously for the rate constants involved. 'Physically this was done
with a computer, and the rate constants were solved for by the method of
steepest ascent. The approach was not too successffil,_probably for the
following reasons: , '

1. There were actually more than five 8°Kr sources in the system,
so the mathematipal'model.Was bverly simplified. Primarily the,effects

of circulating bubbles were not adequately accounted for.
COUNT RATE (counts/min)

ORNL-DWG 67-1965

5
d
4
2 L
!
10 '\
5 \
¢ "hooooou.oo 2e®, .o
L] [ 1 g o000 ....
T Y rvrey -
“q e o._.t..l0.oooou—"rr'o.oo-looouooo-......m cscee
] e L 'fl.&uoo.l.o....
2 \ tyo =198 hr LI % o
3 \
10 \
1\
1\ N
1\ N
\ \s—t172 = 352 br ; \\
\ \ {'11/2 =155 hr 4
\,/\1,/2 =1039 hr \
5 \\
t/p = 019 hr \
10° .
0 10 20 30 40 50 60 TO 8O S0 100 1o 120 130 140 150
: TIME (bhr)

Fig. 14. Results of Exponential Peeling of Krypton Experimerit Run 3.

v

e
a0

25

Table 2. Numerical Results of Peeling Run 3
of Krypton Experiment

Count-Rate

Peeled-Curve Intercept at

Half-Iife Rate Constant

Process

(hr) Zero Time Determined
(counts/min)
198 . 3,574 \ Mass transfer to slow- Mass transfer coeffi-
est rate constant cient
(bulk) graphite
15.5 2,178 Mass transfer to next Mass transfer coeffi-
faster rate constant cient
graphite
4,52 2,114 {a)
1.039 4,945 (b) ‘
0.119 520,000 Pump bowl purging Purging efficiency

aProbably influenced mainly by mass transfer. to fast rate constant
graphite but may also be biased by other processes; generally has a low
degree of confidence. :

bProbably influenced mainly by stripping of the salt but is also
probably biased by other processes; has a low degree of confidence.

2+ The approach required accurate knowledge of the initial concen-

tration of the krypton in all regions involved (boundary values). This

could not be done for the graphite for reasons to be discussed later in

the section oh Capacity Considerations.

The results of the second method will not be presented here.

Pump Bowl Dynamics

Schematically the pump bowl can be represented as follows:

Pump Bowl - 85y
N Monitor
He purge Vgp = Volume of gas phase Off-gas
Q = ft3/hr at 1200°F Cg = Average °°Kr eoncen- ’ Avout | 1ine
- 8P and 5 psig tration in gas phase 70°F | -
fi : fi"- — Xe stripper spray ring
' S = Salt stripping efficiency
Salt phase p . .
5/ = Bubble stripping efficiency
Qgp = £t salt/nr
C§ = 85Ky concentration in '
salt Salt back to

_ ) loop
Void fraction

1l
26

[ 73

The dilution of 8°Kr in the gas phase of the pump bowl is given by*

K

dc Q_E Q.S Q. 8’

& .8 Kk, sp Kk, Sp Kk

dt v Vv Vv B
gp P ep

In the first term Cz is the mean ®°Kr concentration in the gas phase of
the pump bowl and the product Ec: is the concentration in the off-gas
line; therefore, E can be thought of as a mixing efficiency. The second
term represents the rate at which krypton is stripped from the salt,
where S is the stripping efficiency. The third term is the rate at which
krypton is stripped from the circulating bubbles, where | is the void
fraction and S’/ is the bubble stripping efficieficy.

If these three terms are evalfiated at the beginning of the stripping
phase, the second term is approximately 1/500 of the first term, so it
can be neglected. The third term is about 1/20 or less of the first term
for expected values of ¥ and S/. It must be neglected because of inade-
quate knowledge of ¢ and S’/. The error introduced, however, will not be
great. The above equation also neglects the 85Kr contribution by the
stagnant kidney in the bubbler line semitoroid, but estimates indicate
that this is also an adequate assumption.

Neglecting the second and third terms, the equation is

k
acC E
._gz_EEP_Ck
dt v g’
gp

and, solving for Cg at time t, gives

- E/V - )t
Kk (QgpB/ V)
g £0 ?

which, evaluated at concentratioh half-life conditions, is

*See Appendix D for nomenclature.

2'7

)

g 1/2 = e_(QgPE/VéP)tl/z _ ,0.693

or, solving for‘E,

0.693V
E

tl/EQgp

Figures 15 and 16 show the initial transients for runs 1 through 6.

Since the bubbler line stagnant kidney and the salt stripping have little

ORNL-DWG 67-1966

A RUN 1, HALF-LIFE OF CURVE = 6.8 min
® RUN 2, HALF-LIFE OF CURVE = 6.3 min
O RUN 3, HALF-LIFE OF CURVE =83 min — |

COUNT RATE (counts/min)

TIME (min)

Fig. 15. Expanded Plot of Data for the First Half Hour of Runs 1,
. 2, and 3. All count rate data taken with monitor B4 and corrected for
dead time.
28

ORNL-DWG 67-1967

104
A RUN 4, HALF-LIFE OF CURVE = 7.6 min
® RUN 5, HALF-LIFE OF CURVE = 6.3 min
0 RUN 6, HALF-LIFE OF CURVE = 8.4 min

-6 fl :
A
5
2

COUNT RATE (counts/min)

0 5 10 15 20 25 30 35 40
TIME (min)

Fig. 16. Expanded Plot of Data for the First Half Hour of Runs 4,
5, and 6. All count rate data taken with monitor B4 and corrected for
dead time.
29

effect on the initial transients, this section of the curve is determined
almost completely by pump bowl dynamics. A tangent line is shown on each
.curve and - its half-life is given. Note that the half-life measured in
the monitor at about 7Q°F is identical to the reactor half-lives where |
the operating temperature is 1200°F. Runs 1 through 6 gave the results
listed in. Table 3. The average pump bowl purging efficiency is 69%, and

it is not a strong function of pump bowl level.

Table 3. Pump Bowl Purging Efficiencies Obtained in
Runs 1 Through 6 of Krypton Experiment

Indicated Pump Volume of Gas Half-Tife Pump Bowl
Run Bowl Level at Phase in of Curve Purging
No. Start of Strippin Pump Bowl . Efficiency
(% scale) (£t3) (min) (%)
1 71 .75 6.8 6l
2 60 223 6.3 84
3 60 2.23 ' 8.3 64
4 61 2.19 7.6 68
5 70 . 1.79 6.3 67
6 55.5 C 2.43 8.4 68

Xenon Stripper Efficiency

As pointed out previously, values for the stripping efficiency could
not be extracted from the data with any degree of accuracy, even though
runs 4, 5, and 6 were performed with this goal in mind. Very rough cal--
culations do indicate that the stripping efficiencies are moré or less
consistent with those.measured at the University of Tennessee in a COjz-

- water system, but the calculations are so approximate and dependent on

hazy assumptions that they will not be presented here.

Mass Transfer to Graphite

-

- To review briefly the graphite regions, recall that three regions

were identified from fluid dynamic considerations. First there is the
30

bulk graphite region (~95%) that is characterized by salt velocities of
about 0.7 ft/sec and a Réynolds number of about 1000. The mass transfer.
coefficient will be between that for laminar and that for turbulent flofi.
Second, there is the graphite associated with the centermost fuel chan-
nels, which comprises about 1.5% of the graphite. The fuel velocity and
the mass transfer coefficient in this region will be higher than in the
bulk graphite region. Third, there is a region of structural graphite
across the battom of the core. It is difficult to determine the exact
boundary of this region, but it probably consists of approximately 3.5%
of the graphite and is in a zone of low nuclear importance. It is char-
acterized by orificing effects, impingement of salt, and fluid dynamic
entrance regions; therefore it will ha&e the highest mass transfer coef-
ficients. : ' ‘ _ g
‘The first question td be resolved concerns salt-to-graphite coupling
via the mass transfer coefficient. The krypton flux from the' graphite

can .be expressed as |

Kr flux from graphite = hIiAG(Clgi - Cl;) 5

where Cgi is conventionally defined as the krypton concentration in the
salt and at the interface, where the interface is continuous. In this
case the salt-gas interfgce is inside a pore that occupies only a small
fraction of the total graphite surface area. Therefore we neéd a rela-
tionship between this concentration at the pore interface and the more
conventional CEi' This is discussed in Appendix B, where it is shown
that the mean concentration of krypton in a continuous salt film across
the graphite surface is épproximately equal to the krypton concentration
in the salt at the salt-gas interface inside a graphite pore.

The rate at which €°Kr is ieached from the graphite is a function
of several parameters; for instance, the diffusivity of krypton in graph-
ite, the mass transfer coefficient, and the 85Kr concentration dissolved
in the bulk sait, which is in turn a function of the xenon stripper effi-
ciency. The general approach in the graphite analysis will be to first
show that this rate is very insensitive to expected values of the dif-

fusion coefficient of krypton in graphite. This being true, we can

e
31

determine a relationship between the mass. transfer coefficient and the
stripping efficiency, any combination of which will result in a flux curve
as meesured. Thefi by weighting this.relationship'with’the value of strip-
ping.efficiency measured at the University of Tennessee, theoretical
velues for the mass transfer coefficient (see Appendix C), and other con-
siderations, we will obtain a very narrow range of‘possible values for

the mass transfer coefficient. Generally, this procedure will be fol-
lowed for all the graphite regions considered. Most of the calculations
wili be confined to run 3, which was concerned primarily with measuring
graphite rate constants.

Consider the ®°Kr flux from graphite as a function only of its in—
ternal resistance (Dg) and its external resistance (pi). Also specify
that, for times equal to or greater than zero, the krypton concentration
dissolved in the fuel salt is zero. Cylindrical geometry is used; that
is, each core block is considered to be a cylinder, the surface area of
which is eqfial £b the fuel ehannel area aseociated with a single core
block. The volume'of a graphite cylinder ef this sort is very close to
the volume of the actual core blocks of the same length. The differen-

tial equation that describes this case is

G G G
+__'='—k .
dr? r Or Dy ot
with Boundary conditions

k .k -

CG = CGO at t = 0,
ack
—[=0atr =0,
dr
de h_HRT
G N ck at r=r
dr Dk G b
32

The solution to this equation is

' ' k
ok 'ch i -J;' Jl(Mn) —[(Mn)2DG/r€€:|t i r
G Go o1 Mo Jg(Mh) N Ji(Mn) 0 ’

where the eigenfunction is

and:

where 7.ismthe eigenvalue. Now, differentiating with respect to r and

evaluating at r = r , we have

K k k
g\ g, = J2(Mn) —[(Mn)zDG/rée:lt
2
ar ’ r, n=1 J5(Mn) + JZ(Mn)

Flux == = — |—
b € \dr Ty
we obtain
kk o 5 _[ sk 2]
. 2DCos Jl(Mn) (Mn) DG/rbe t
Flux, = —2 ) e :

r 2 2
b r e n=1 J§(Mn) + Ji(Mn)

From this équation it can be seen that the krypton flux from any
given graphite region is the sum of a series of exponentials, with the
slope of each exponential being determined by the exponent of e. The

problem now is one of relating these exponents to the slopes of the peeled
- 33

flux curves. Considering run 3 it is obvious that the slowest exponent
(t1/2 = 198 hr) is related to the bulk graphite, and it is expected that
the next exponential (ti1,2 = 15.5 hr) is related to the graphite region
located at the center line'of_the‘cofe. The next exponential (ti,z = 4.52
hr) has a fairly low order of confidence in equating it to any specific
graphite region and will not be considered. It can be shown that for
each of the two graphite regions considered, only the first térm of the
above series is significant. Thérefore the exponent of e can be related

to the measured half-life, as follows:

-

Now, by evaluating this equation in conjunction with the eigenfunction

equation, we can relate values .of hm and Dk This was. done with the fol-

lowing paramgter values, and the results agpear in Fig. 17: -
t1/, (bulk graphite) = 198 hr (run 3),
t1/2 (center-line graphite) = 15.5 hr (run 3),
¢ = 0.10, " -
r, = 0.0905 ft, )
H = 8.5 X 107% moles/cc-atm,
HRT = 6.43 X 107%.

The ordinate represents the range in which»Dg is expected to lie. Note
that for bulk graphite the value of hi is almost completely independent
of Dé; therefore the mass transfer coefficient is controlling the krypton
flux from this graphite region. For the center-line graphite region,

the dependence of flux on Dg
Dk

GI

becomes significant only at low values of

It is difficult to extract hi information from run 2 because the
time intervals involved were too short. The krypton addition and strip-
ping phases were about 60 hr each..'During this relatively short addition
time the bulk graphite reached only about 20% of its saturated value, in
contrast to run 3, where it reached about 70% of its saturated value.

For both runs the center-line graphite region (t1/2 = 15.5 hr in run 3)
3.

ORNL-DWG 67-1968

1073
5
2 NEXT GRAPHITE REGION
N AFTER BULK GRAPHITE
z (CENTER-LINE REGION OF
L, CORE)
£ 10 }
x (B
| =] \
|
i 5 \\
\
? 2
| —BULK GRAPHITE
1075
5 ps

0 01 02 03 04 05 06 07 o8

h:‘noss (ft/hr)

Fig. 17. Relationship Between D and hi That Fits Graphite Leaching
Curves Peeled from Run 3. . ‘
was almost completely saturated. After stripping for 60 hr in run 2, the
slope of fihe flux curve is not determined by a single graphite region
but is still under the influence of two graphite regions, and it cannot
be peeled by the same technique as run 3. Nevertheless, run 2 was looked
at, and without presenting any results, it will be stated that it was
consistent with run 3.

Now, since hm is not a strong‘function of Dg, we can determine the
relationship between hfi and S (stripping efficiency). This will be done
for the bulk graphite region after sufficient time so that other tran-

sients are negligible. First, we will make the following rate balance:

Kr flux from xenon stripper = Kr flux from graphite
+ dilution rate of Kr in salt,

where

Kr flux from xenon stripper = 8Q C

35

. kB k k
Kr flux from graphite = h_ AG(CSi CS),
| | ack
Dilution rate of Kr in salt = -V —.
S at
Theh, substituting,
k
acC
sq cf = nBa (ck - ~v =,
Sp S m G si S S 4t

In order to solve this equation, krypton concentrations must be converted
to krypton fluxes because this is the formlof the data. The measured

krypton flux is related to the stripping rate as follows:

kK
Flux = SQ,SPCS ’
and
d flux dCi
— = 3Q —_—
at SP qat
Rearranging gives
Ck _ Flux
. 2
S SQ,-Sp

and

dcg 1 d flux

and now, confining ourselves to one graphite region, and specifically the

bulk graphite from run 3 (t1,2 = 198 hr), it can be shown that

Flux - Flux, ,—0-693t/t172
36

and

d flux —0.693 flux

0 e—(0.693/t1/2)t
dt - ti1/2 ’

Substituting into the above equations gives

_ W% ~(0.693/t1/2)t

and

k
ac;  —0.693 flux, ~0.693/t1 /)

dt SQ,Sptl/2

|

Fufther, substituting these into the original rate balance we get

‘ fl
Flug e (0+693/t1/2)t _ kB, [k _ ™o ~0.693/t1/5)t
: 0 m G i sp

-

0.693V  flux
N s

0 e-.(0.693/t1/2)t .

S t
QSP 1/2
. k. -
Solving for CSi gives
flux h A 0.693V
ok of, ,m’e_ " s e—(0.693/t1/2)t
si kB - ' ?
S S t ,
hm AG | QSP ' QSP 1/2
and in its differential form,
k kB
dCy;  —0.693 flux, . by Ag 0693V ~0.693/t1 /)t
= 1{B ‘ 2
o S t
@ Tt Ae \ Plp SOt

which relates Czi and dczi/dt to the measured slope of the flux curve and
the various physical parameters involved. At this point we will set up'

a rate balance on the bulk graphite, as follows:

Kr flux from graphite = dilution rate of Kr in graphite,

37

where
, . _ kB k _ .k
Kr flux from graphite = h_ AG(Csi Cs) s
ack
Dilution rate of Kr in graphite = — G Ti—t— R
or .
k kB
dCq by Be /x K
— . — ¢c.—C ’
dt VCT s1 S

where Clé is the mean krypton concentration in the graphite. Now in the

previous analysis it was shown that for the bulk graphite, hx}:lB‘is inde-

pendent of Dlé

sult is that the krypton concentration profile across a graphite core

over. the range of interest. One consequence of this re-

block is essentially flat. We can therefore say that

k. k
Ce = Cgi 2

and also, from Henry's law,

Substituting this into the graphite ré.te,balance, we have

k kB
dc. h

si _ hm AGI-IRIII (Ck _ Ck)
dt eVG si s/ '

Substituting previously derived relations for Cls{i’ dC}sii/dt, and Clg into

the above equation, and solving for hiB , we have,

0.693¢ ( O.693VS )
l — E—————————
B AGHRTt Q 5

h _ 1/2 t1/2S S

m 1 0.693 € 1 ’
—— + ==
V. T8¢t (HZRT VG)

G Sp 1/2

38

We now evaluate hfi? as a function of S for the following values Qf other

parameters:
tl/2 = 198 hr (bulk graphite),
. £ = OnlO,

A = 1450 ft° (channel surface area in bulk graphite region),
V, = 64.8 ft? (volume of graphite in bulk region),

G
vV, =70.5 £t° (volume of salt in loop),
Q,Sp = ?O gpm, |
H=28.5 x 10-° moles/cc-atm, .
HRT = 6.43 x 1074, : | .

The results of this calculation are shown in Fig. 18, where the mass trans-
fer coefficient (hm) is plotted agalnst stripping efficiency (S). In . -
this figure the superscript B refers to the bulk graphite region. Also

shown on this plot is the theoretical expected range of hiBlbased on

ORNL-DWG 67-1969

04 T I . ¥ 1 1 1 kCL
SHADED AREA - EXPECTED REGION WHERE h*“t vs
S CURVE SHOULD LIE FOR CENTER-LINE GRAPHITE
4 —
/ 7/ KkCL
0.3 “ 44— THEORETICAL h/’s FOR CENTER-LINE GRAPHITE
;% REGION IN TURBULENT FLOW
é% '

{ |

o2 SHADED AREA - EXPECTED REGION WHERE htB vs. s_
CURVE SHOULD LIE FOR BULK GRAPHITE

hX , MASS TRANSFER COEFFICIENT (ft/hr)

4 : | THEORETICAL h® FOR TURBULENT] ' |
U7 FLOW FOR BULK GRAPHITE | - S
T I
OR 9/
7 % THEORETICAL h® FOR LAMINAR .
YN FLOW _FOR BULK GRAPHITE
~ -EXPECTED RANGE OF S
.0 I I |
O 10 20 30 40 50 6 70 80 90 100

S, STRIPPING EFFICIENCY (%)

-

~

Fig. 18. Relatlonshlp Between hm and S for Bulk and Center-Line
Graphite Regions.

39

laminar and turbulent flow; it displays the uncertainty in the diffusivity
of krypton in salt. In addition the expected range of stripping effi-
ciehcy is outlined. Note that the calculated curves fall within the ex-
pected range. _ |

In order to pick out more explicit values of th, we can weight this
curve with the following two cons1deratlons. First, it is expected that
some 01rculat1ng bubbles were present as will be shown in the next sec-
tion on Capacity Considerations. Furthermore, it is expected that cir-

- culating bubbles increase the effective stripping efficiency to the high
end-of the indicated range, and'likely even higher. Second, most of the
salt:iu the bulk graphite region is of a laminarrcharaoter, rather than
turbulent, so h;? ought to be on the lower end of the‘expected range;
Therefore, from Fig. 18 and with the above weighting considerations, we
might pick out a probable range of hfi?'to bulk graphite as

0.05 < K < 0.09 £t /hr .

The next graphite region (tl/é = 15.5 hr), interpreted as the center-
line graphite region, is more difficult to handle. The previous deriva-
tion assumes that all the krypton dissolved in salt comes from the graph-
ite region under consideration. Now, when working with the center-line
‘region, we must also consider the krypton dissolved in salt that origi-
nates in the buik graphite. "An equation whose derivation is similar to

the above and which partially compensates for tlis is

0.693V>" BB
G l + m G
CL. oL
L _ AChe] ) HRT QS
- J
m ‘ nEBaB 0.693ev° L AN
m G G 1/2 9]
1+ - T L 1+ L,
T
a,S | Sty R 6, £luxg

where the superscrlpts CL and B 1nd1cate center- llne and bulk graphite
reglons, respectlvely The equatlon was evaluated w1th the same parame—
ter values as before and with the follow1ng additional ones, and for a

given value of S the value of hfi? was taken from the preV1ous calculation
40

for the same void fraction:

tg?z = 15.5 hr (center-line graphite region),

ACL = 14.5 £t? (channel surface area in center-line region),
VgL = 0.682 ft> (volume of graphite in center-line region),
Fluxg = 3574 counts/min (intercept of peeled curve at t = O fer

bulk graphite),
FluxgL = 2178 counts/min (intercept of peeled curve at t = O for

bulk graphite).

Th% results of this caig;latiop appear on Fig. 18. Also shown is the |
theoretical range of hm based on turbulent flow and displaying the un-
certainty in diffusivity of krypton in salt. The equation for hkCL is
1nvalld at low values of S because kkB approaches 1nf1n1ty, hence the
llnes are terminated as they approach the equivalent line for th Nev-
er?heless, 1t seems that a reasonable range of hkCL for the MSRE would

be between 0.25 and 0.4 ft/hr.

Capacity Considerations

In addition to rate constant determinations from slopes of the peeled
exponentials, we should be able to infegrate under the curves and deter-
mine information on the capacity of the system. For instance, the inte-
gral of the bulk graphite curve (t1/2 = 198 hr) should yield.the approxi-
mate &Kr capacity of this graphite.. Then, from knowledge of the 8 5kr
addition concentration, we could compute the‘approximate graphite void
fraction available to krypton. When this calculation was performed with
the.additioh concentration taken as the mean pump bowl concentration, _
the graphite void fraction came out to be about 0.40. This is obviously
incorrect, even when' the rough nature of the calculation is considered.
The reason for this is that kryp%on was added through a bubbler,'and it
bubbled up through the salt at approximately one order of magnitude higher
in concentration than in the pump bowl proper. These highly concentrated
krypton bubbles were caught in the turbulent and recirculating zone formed
by the spray ring. Very likely some of these bubbles, or micro bubbles,
were carfied into the primary loofi and circulated with the salt. The re-

sult was that the concentration of dissolved krypton in the fuel salt was .

41

much higher than it would have been based.on the mean pum@ bowl concen-
tfation. Now, if we compute the graphite void fraction based on the dis-
solved krypton in the salt being in equilibrium with the bubbler addi-
tion stream, the result is about 0.04, and this is lower than would be
expected. The true void fraction is between the limits of what can be
calculated, and insufficient information is available to compute it more
accurately. The same‘is true for all other graphite regions and the fuel
salt, therefore little useful  information on capacity 1is available from
the data. It should be pointed out that the choice of a bubbler line

for 85K; addition as opposed to the pressure reference leg, which’would
not have given this deleterious effect, was dictated by other consider-

ations. -

XENON-135 POISONING IN THE MSRE

General Discussion

To calculate the'steady—staté'l35Xe poisoning in the MSRE, it is
first necessary to compute the steady-state 135%e concentration dissolved
in the salt. This is because the ?%Xe is generated exclusively in the
salt, at least it is assumed to be. The.xenon concentration in the salt
is computed by equating the éourCe and- sink rate terms involved. The
most significanf of these terms and considerations involving them are

discussed below.

Dissolved Xenon Source Terms and Considerations

1. Xenon Direct from Fission. The 135%e generation rate direct

from fission is about 0.3% per\fiissibn.

2. Xenon from Todine Decay. The *3°I generation rate is 6.1% per
135
) Te-

fission, either direct or from the decay of It in turn decays to
135¥e with & 6.68-hr half-life. The total ?°Xe generation rate is there-
fore 6.1 + 0.3 = 6.4% per fission, and, as will be seen from the next
consideration, is confined completely to the salt phase. Since the prin-

éipal 135%e source is from the decay of iodine and since the 1351 nalf-
42

life is long compared with the fuel-loop cycle time (25 sec), it will be
assumed that 135Xe is generated homogeneously throughout the fuel loop.

‘3. Iodine and Tellurium Behavior. Since '2°I and '35Te are pre-

cursors of 3%Xe, their chemistry is important. In this model it is as-
sumed that both elements femain in solution as ions, and therefore will
‘not be removed from solution by the xenon stripper or diffusion into the
grabhite. Concerning iodine, its thermodynamic properties indicate this

10 Recent evidence from the reactor also indicates this to

to be true.
be true. Some iodine has been found in thé off-gas system (but very lit-
tle, if any,_1351), but this is due to the volatilization of the precur-
sor tellurium. Since 135Te has such a short half-life (<0.5 min), very
little of it will have a chance to volatilize; therefore this effect is
neglected. It is also assumed for this model that iodine and tellurium
will not be absorbed on any internal reactor surfaces, such as the con-

tainment metal and graphite.

Dissolved Xenon Sink Terms and Considerations

1. Xenon De¢ay. Xenon-135 decays with a half-life of 9.15 hr, and

decay takes place throughout the entire fuel loop.

2. Xenon Burnup. Xenon-135 has a neutron absorption cross section
11

of 1.18 X 10° barns averaged over the MSRE neutron spectrum.

3. ZXenon Stripper Efficiency. As noted earlier, the efficiency of

the xenon stripper was measured at the University of Tennessee with a
COs-water systemg and confirmed later with an Oj-water system.12 Both
tests were for a bubble-free system. The measured rate constants were

then extrapolated to a xenon-salt system.?

The stripping efficiency is
defined as the percent of dissolved gas transferred’from the salt to the
gas phase in passing through the xenon stripper spray system. In magni-
tude it turns out to be between 8 and 15%.

4. Xenon Adsorption. Xenon is not adsorbed on graphite signifi-

cantly at these temperatures, 3,14 and it is very wnlikely that it will
be adsorbed on metal surfaces. '

5. Xenon Migration to Graphite. The amount of xenon transferred

to the graphite is a function of the mass transfer coefficient, diffusion

coefficieht of xénon in graphite, and the burnup and decay rate on the
43

graphite. During manufacture, this graphité was impregnated éeveral times
to obtain a low permeability. Diffusion experim.ents2 with a - single sam-
ple of CGB graphite yielded a diffusion coefficient of xenon in helium

at 1200°F and 20 psig of about 2.4 X 1075 cm?/sec (9.2 x 10-5 £t2/hr).
This was measured in a single sample of graphite and may not be repre- |
sentative of fhe reactor core; however, it will be seen later that the
poison fraction in the MSRE is not é strong function of the diffusivity.
As previously pointed out, the core grgphite_may.be divided into three
fluid dynamic regions, bulk,.center line, and lower grid. The krypton
experiment did not yield any reliable information on the lower-grid re-
gion, and since it is in a region of very low nuclear importance, it will
not be considered. The bulk and center-line regions will, however,_be
considered. ' |

6. Xenon Migration to Circulating Bubbles. As will be seen, the

effect of circulating bubbles is very significant because Xenon is so
insoluble in salt. Although information on circulating bubbles is meager,

they will be considered..

Other Assumptions and Considerations

7

1. The '3°Xe concentration dissolved in the fuel salt was assumed
to be constant throughout the fuel loop. From the éomputed results it .
can be shown to change less than 1%.
2. Tt was assumed that the 12°Xe isotope behaves independently of
all other xenon isotopes present.
" 3. The 13%%e generated in the laminar sublayer of salt next to the”
~ graphite was considered as originating in the bulk salt.

4, The core was considered as being composed of 72 annular rings,

©/— Core
- | Core- element
Q¥ |

N

as shown below

.HeH

44,
Average values of parameters such as neutron flux, percent graphite, mass
Itransfer coefficient, etc., were used for each ring. :
5. The reactor system was assumed to the isothermal at 1200°F.
Consistent with all previous assumptions, the rate balance of 13°Xe

dissolved in the fuel salt at steady state is the following:

Generation rate = Decay rate in salt + burnup rate in salt
+ stripping rate + migration'rate to graphite

+ migration rate to circulating bubbles P

where the units of each térm‘are ;35Xe atoms per unit time. ZFach term

will now be considered Separately.

Xenon-135 Generation Rate

The total 13%Xe yield is 644% per fission. The'cérresponding 135xe

generation rate is 5.44 x 10'° 1?Xe atoms per hour at 7.5 Mw.

Xenon-135 Decay Rate in Salt

The decay of 135%e dissolved in salt is represented as follows:

0.693V C~
S S

Decay rate in salt =
) +X
1/2

Xenon-135 Burnup Rate in.Salt

The burnup rate of 135%e dissolved in salt is expressed incrementally
by dividing the core into the 72 elements described above and is as fol-

lows:
72
: X X
i t = .
Burnup rate in sal eZl ¢2ea feVeCS

Xenon-135 Stripping Rate

Recalling that the stripping efficiency (S) is defined as the per-
centage of 135%e transferred from salt to helium in passing through the
45

pump bowl stripper, the stripping rate is expressed as
. _. X
Strlpplng‘rate = Q,SPSCS .

Xenon-135 Migration to Graphite

Each graphite core block will be assumed to be éylindrical. This
seems to be a good compromise between the true case and ease of computa-
tioh. The surface-to-volume ratio for a cylinder is very close to the
channel surface area-to-volume ratio of the actual core blocks. Diffu-
sion of 1?°Xe inside cylindrical éore blocks at-steady state and with a

sink term is as follows:

Solving for the following boundary conditions

% = finite at r = 0,
x: X =
CG CGi at r r o

we obtain

where
2 X X
6 - _X (¢2O_ +, 7\ ))
'; - |
Ib-= zero-order modified Bessel function of the first kind.

.
Differentiating and evaluating at r;, we have

fdac : T (Br.)
4 =c.5_%__1__.
dr Io(firl)

The 135Xe flux in the graphite and at the surface (r = r;) is given by
or, substituting the previous equation, we obtain

.4
DG Il(Brl) .

B—.
1 € IO(Brl)

The 13%Xe flux can also be represented by

Combining the last two equations by eliminating C and then solving fof

X

N Gi

Fhmr;weoMmfll -
1 .

hXCX

m s

1 h;HRT I (Br)
1+ 9

X
BD, Il(Brl)

in units‘of 135%e atoms/hr-ftz. Because of the flux distributions, graph-
ite distributions, and the various fluid dynamic regions, the core will
be handled incrementally as the 72 regions described earlier. We can now

solve for the total '3°Xe flux into the graphite, as follows:

| 72 pfyvp o
Migration rate to graphite = Z} me ees
e=1 WS HRT I (B r )
1 4+ fe 0' e 1

X
5e DG Ii(Berl)

in wnits of *?°Xe atoms per hour, where

1l

volume of core element,

il

graphite volume fraction at v,

v
e

F
. €
Y

fuel channel surface area-to-graphite volume ratio.

47

It is assumed that Y is constant throughout the core moderator region

and has a value of 22.08 ft~l.

Xenon-135 Migrafiion Rate to Circulating Bubbles

The rate of *35Xe migration to the circulating bubbles is represented

by
Migration rate to bubbles = h A (c: - c¥,

in units of '35Xe atoms per hour. The salt film is by far the control-
ling resistance; therefore the 135%e concentration in the bubble is uni-
form and at equilibrium with the concentration in the salt at the inter-

face. Consequently the previous equation can be written as

[
f

Migration rate to bubbles = hpA_ @: - HRTC%) :

At steady state the migration rate to the bubbles equals the rate

that_135Xe is removed from the bubbles, therefore

Migration rate to bubbles = Decay ratep

+ burnup rate_ + bubble stripping rate ,

B
where

' X
O.693VS¢CB

Decay rate in bubbles = - )
‘ t1/2
Bubble stripping rate = Qspws’Cg s
72 _
Burnup rate in bubbles = Z} ¢2e0xfevew0§

e=1

and the burnup rate of '2°Xe in the bubbles is handled the same as in the
salt, that is, by dividing the core into 72 elements of volume and adding
up the burnup in each element. Substituting the individual rate terms in

the removal rate balance equation we get
48

| 0.693V yCy
Migration rate to bubbles = - S
P
I's 72
¢ X X X
+ /
: ;ZL Q2eg fevech i QSPWS CB ’

Now, this equation may be solved simultaneously with the previous equation

to eliminate CE. This results in

hpAgCo -
hBABHRT

0.693V y 72
-'———S—'+ZQO'XfV\JI+Q 1stl
X 2e e e Sp

tl/z e=1

Migration rate to bubbles =

1+

in units of '35Xe atoms per hour.

Xenon~135 Concentration Dissolved in Salt

The 135Xe concentration dissolved in salt may now be solved for by
substituting the individual rate equations into the original rate balance.
This will yield

0.693V ¢° 72
S 8

5.44 x 109 = + X X
t, Z; ¢2eg feVels
/2 e=1
72 X X
h™ YF V C
+ QS SC}S{ + E I}r{le e e s
P e=1 h HRT I (8 r.)
1+ me 0 e 1

X
fieDG Il(?erl)

hgAgCy

4 .
hBABHRT ‘ ?
1 +
0.693V_y 72 «
—_— / -+
tx Q'SPIL[S Z ¢2eU fevew
_ 1/2 e=1

where the units of each term are 135%e atoms per hour.
49

Xenon-135 Poisoning Calculations

The xenon poisoning as obtained in this report is defined as the
number of neutrons absorbed by 135%e over the number of neutrons (fast
and thermal) absorbed by 235U and weighted according to neutron impor-

tance;l5 it is expressed as a percentage. The weighting function is the
adjoinfi flux. When considering the core incrementally it is given by

72

* ox | *
Px GZ"]_ ¢2e¢2efevecs " Z © ¢2e 2€ e e G Z © ¢2e¢2 feV Wc

‘ u
Z) (o 1 1e 1e 8% b ) £UC

ggege e e S

where the first term in the numerator is the rate dissolved '?7Xe is
burned up, the second is the rate 135%e in the graphite is burned up,
and the third is the rate 135¥e in the bubbles is burned up; all terms
are weighted with the adjoint flux. Now, the term representing the rate
135%e in the graphite is burned up can be replaced by the 135%e flux

into the graphite times the fraction burned, that is

n* yv F ¢® ¢ o
O‘QS X me e e s e
2ee e’ = W HRT I (B r ) \¢ o +A°
me g €1 2c
BeDG Il(Berl) : |

With this substitution the poisoning of 135Xe in the MSRE can be computed. -
The reactivity coefficient is related to the poisoning by a constant,

which is a function only of the nuclear parameters} This has been evalu-
ated’® and is

(8k/x)* = —0.752 P* .

Estimated '3°Xe Poisoning in the MSRE Without Circulating Bubbles

With the equations given above, 135%e poisoning has been computed
for the MSRE for a variety of conditions subject to the assumptions dis-

~cussed earlier. The procedure was to first solve for the steady-state
50

1é’!s}(e concentration dissolved in the salt, and then from this to compute
the poisoning. A code was.set up to do these calculations on a computer.
Thé neutron fluxes used were those reported in Ref. 15 and corrécted with
more up to date information.'! Values of many parameters used are given
in Appendix A, and others were taken from standard reference manuals.
Nominal values of various important rate constants and other variables
were chosen, and the variation of polson fraction with these parameters
was computed. These nominal values may be interpreted as approximate
expected values. The first case to be discussed will be the bubble-free
situation. Then the case of circulating bubbles will be discussed.

In the bubble-free case the following nominal values éf various pa-

rameters were chosen:

Available void fraction in graphite, € 0.10
Diffusion coefficient of Xe in graphite, 1 x 1074
DE, ft3 of void per hr per ft of graphite

Mass transfer coefficient to bulk graph- 0.0600
ite, hXB, ft/hr

Mass transfer coefficient to center-line 0.380
graphite, hfigL, ft/hr

Stripping efficiency of spray ring, S, % 12

Reactor thermal power level, Mw 75

The pump bowl mixing efficiency was found to have a negligible effect and
was not considered. A

The results of the calculation are given in Figs. 19 and 20. Each
plot shows the poisoning as a function of the parameter indicated, with
all others being held constant at their nominal values. The circle in-
dicates the nominal value. From these plots the following observations
can be made. _ — ,

1. For the bubble-free case, the 1?5Xe poisoning in the MSRE should
be 1.3 to 1.5%.

2. Generally speaking the poisoning is a rather shallow function
of all variabies plotted. DNote particularly the insensitivity of poison-
ing to available graphite void (Fig. 20) and the diffusion coefficient
(Fig. 20). The reason is that hz'controls the '?°Xe flux to the graph-

ite. This could be an important economic consideration in future reactors
51

ORNL-DWG 67-1970°

3
2
© 2
=z
zZ
2
pdd ___.--—'_'__
@ / .
>
[Ty]
2
(a)
0 -
0] 0.02 0.04 Q.06 008 ) 0.10 0.12
MASS TRANSFER COEFFICIENT - BULK GR‘_APHITE (€t/hr)
3
2 )
© 2
<
z
2
o ]
o
(]
g
ald
(#)
0
0 (OR 0.2 . 03 04 05 06
MASS TRANSFER COEFFICIENT - CENTER-LINE 'GRAPHIT.E (ft/hr}
3
&
o 2 M
Z
Z
&
—_ o N
o L ]
a
- \\
a1 —~—]
o \_._
bl — |
{c)
0 )
-0 10 20 30. 40 50 100

STRIPPING EFFICIENCY (%)

Fig. 19. Predicted '?°Xe Poison Fraction in the MSRE Without Cir-
culating Bubbles at 7.5 Mw(t). See body of report for values of other
parameters, .
52

of this type. For .instance, if xenon poisoning is the only consideration,
the permeability specifications may be rej;axed somewhat.
Other numbers of interest are given below. For the nominal case,

the 13%Xe distribution to 'its sink terms is

ORNL-DWG 67-1971

2
S
:ED 0
=
31
2
a
QQ
X (o)
0
0
0 5 10 15 20
AVAILABLE GRAPHITE VOID FRACTION (%)
3
g, ,
<
=4
z /
@ o/
2 .
g /
% -
-
(#)
; 0 -
1 10 100
REACTOR THERMAL POWER LEVEL {(Mw)
3
2
o 2
=
pd
3
o o]
o
>
X
22
) {¢)
0
10° 2 5 104 2 5 103

DIFFUSION COEFFICIENT OF Xe IN GRAPHITE (ffa/hr)

Fig. 20. Predicted *°Xe Poison Fraction in the MSRE Without Cir-
culating Bubbles at 7.5 Mw(t). See body of report for values of other
parameters. ‘ : k
53

Decay in salt 3.4%
Burnup in salt 0.9
Stripped from salt 31.0
Migration to graphite 64 .7
100.0%

Of the 135Xe that migrates to the graphite, 52% is bufned up and 48% de-
cays, averaged ovef the moderator region. Again, for the nominal case,
96.4% of the total poisoning is due to *?%Xe in the graphite and only
3.6% is due to the "?°Xe dissolved in the salt.

Estimated 12%Xe Poisoning in the MSRE with Circulating Bubbles

Xenon, and all noble_gases_for that matter, is extremely insolfible

in molten salt. From the Henry's law constant,

2

c§ = 2.08 X 10~ cz at 1200°F

where the units of concéntration'are xenon atoms per unit volume. A sim-
ple calculation would show that with a circulating void fraction of 0.01
and the xenon in the liqfiid and gas phasesvin equilibrium, about 98% of
the xenon present would be in the bubbles, or.if the void fraction is

- 0.001, 83% of the xenon would be in the bubbles. We would therefore ex-
pect that a small amount of circulating helifim bubbles would have a pro-
nounced effect on *?°Xe poisoning.

Circulating bubbles have been observed in the MSRE. The most sig?
nificant indications come from "sudden pressure release tests." These
experiments consist of slowly increasinglthe system pressure from 5 to
15 psig'and then suddenly ventiné“the pressure off. During the pressure
release phase, the salt level in the pump bowl rises and the éontrol rods
are withdrawn; both motions indicate that circulating bubbles are present.
Void fractions can be computed from these tests that range from C to 0.03 |
but are generally less than 0.0l. The reactor operational parameters
thafi controi the void fraction are not completely understood, and it ap-
pears fo be a quite complex phenomenon. For instance, the void fraction
may be a function of how long the reactor has been operating.

The bubble diameter is extremely difficult to estimate. The only

direct source of information on this point is from a water loop used for
54

MSRE pump testing. In this loop the pump bowl is simulated with Plexiglas
so that the water flow can be observed. The bubbles that migrated from
the pump bowl into the pump suction could be seen and were about the size
of a "pinpoint." For lack of any better measurement they were taken to
be in the order of 0.010 in. in diameter. As will be seén, this is not a
critical parameter in these calculations.

Information in the literature on mass transfer to circulating bub-
bles -is meager. Nevertheiess, from Refs. 16 through 19 and other sources,
the.mass transfer coefficient was estimated to be in the range 1 to 4
ft/hr and practically independent of diameter. Again it turns out that
this is not a crifical,paraMeter, even over this fourfold range. It is
also assumed that the existence of circulating bubbles will have no ef-
fect on the salt-to-graphite mass transfer coefficient. This is equiva-
lent to saying that the circulating bubbles do not come in contact with
the graphite in any significant quantities.

A parameter thaf is quite critical is the bubble stripping effi-
clency. This is defined as the.-percentage of 135%e enriched bubbles that
burst in passing through the spray ring and are replaced with pure he-
lium bubbles. At this time there is no good indication as to what this
falue is. It is probably a complex parameter like the circulating-void
fraction and depends on many reactor opefational variables. For lack of
any better information, a nominal value was taken as 10%, because this
is about the salt stripping efficienc&,.but it could just as easily be
in the order of 100%. 4

- Xenon-135 poisoning has been computed for the following range of

variables pertaining to circulating bubbles.

Nominal Range

Parameter ) Value Considered
Mean circhlafing void volume 0-1.0
Mean bubble diameter, in. 0.0ld _ 0.005-0.020
Mean bubble mass transfer coeffi- 2.0 0.5-4.0
cient, ft/hr , | A
Mean bubble stripping effi- 10 0—100

ciency,

55

/

Again the nominal value can be interpreted as the expected value but with
much less certainty than in the bubble-free case. All other parameters
not pertaining to the bubbles were held cOfistant at the nominal value
given for the bubble-free case. Figures 21 and 22 show the computed
135%e poisoning as a function of circulating void volume with other pa-

rameters ranging as indicated. Parameters not listed on these plots were

ORNL-DWG 67-1972

1.4
1.2
BUBBLE MASS TRANSFER
~ 10 COEFFICIENT
g (ft/hr) -
9. 05
> .
s 08 1.0
8 :
£ o6 '
@
u)x ) ‘ R
A \\Q\\
0.2
{a)
0] "
0 0.25 0.50 075 1.00 1.25
MEAN BUBBLE CIRCULATING VOID (%)
1.4
(- !
1.2.
;;5' 1.0
:_,'.; MEAN BUBBLE DIAMETER
=2 (in)
== 08 |
§ . 0.020
bl 0010
o
>
0
g \ »
T 04 F —
\‘ .
|
0.2
()
. 0
0 025 ' 050 075 1.00 1.25

. MEAN BUBBLE CIRCULATING VOID (%)

Fig. 21. Predicted 135¥e Poison Fraction in ‘th:e MSRE with Circu-
lating Bubbles at 7.5 Mw(t). See body of report for values of other
parameters. -
ORNL-CWG 67-1973
BUBBLE STRIPPING EFFICIENCY (%)

0

|

35%e POISONING (%)

\\
~
I

(o) —— 50
0 100
0 0.25 0.50 075 100 1.25
MEAN BUBBLE CIRCULATING VOID (%)
2
10 .
NN
Y
S
\-"-'--_
S —
~ | TOTAL —|
_ 2 o
3 — ) BUBBLES
Z 0t = : : GRAPHITE
3 7
12 7
e 5 <;
Q
>
5, N
0.01 [
e
5 e SALT —
(6)

O o1 02 03 04 05 06 07 08 09 10 i
MEAN BUBBLE CIRCULATING VOID (%)

~ Fig. 22. Predicted 12°Xe Poison Fraction in the MSRE with Circu-
lating Bubbles at 7.5 Mw(t). See body of report for values of other
parameters. »

held constant at their nominal values. From these figures, the following
observations can be made: |

1. Circulating bubbles have a very pronounced effect on 3°Xe poi-
soning, even at very low void percentages.

2. The '3%Xe poisoning is a rather weak function of the bubble mass
transfer coefficient and diameter over the expected range.

3. The poisoning is a strong function of the bubble stripping effi-

ciency.

[}

57

Figure 22 also shows the contribution of each system (salt, graph-
ite, and bubbles) to the total 135Xe.poisoning. All parameters.are fixed
at their nominal values. This figure illustrates how bubbles'work to
lower the poisoning. - As the circulating void is increased, more and more
of the dissolved xenon migrates to the bubbles, as noted by the fapidly
increasing contribution to poisoning by the bubbles. In contrast to the
bubble-free case in which the *2°Xe in the graphite is the greatest con-
tribution to poisoning, the 135%¢ concentration of the salt is rapidly-
reduced by the bubbles and is thus not available to the'grapfiite.

At the time this report was written, there was no accurate knowledge
of the extent of 2Xe poisoning. Preliminary values based on reactivity
balances indicate it is in the range 0.3 to 0.4%. This is considerably
below the value calculated for the bubble-free case, but it is well within
the expected range when circulating helium bubbles are considered. We
conclude therefore that this model probably does accurately portray the
physical reactor, and good agreement depends only on reliable values of
the various parameters involved. Work is currently under way to esti-
mate more accurately the circulating void fraction and to determine what
operational variables affect it. An attempt will also be made to esti~
mate the bubble stripping efficiency, although this may be quite an elu-
sive parametér to evaluate. Actually, the most recent information seemsk
to indicate that the circulating void fraction (V) is in the order of
0.1 to 0.3% and the bubble stripping efficiency (S/).is in the range of
50 to 100%. Equipment is currently being built to measure 135%e poison

fractions more accurately.

CONCLUSIONS

The analyses presented indicate the follOW1ng

1. A transient experlment such as the ®°Kr experiment can be useful
in determining rate constants and other 1nformat10n for a complex process
such as noble gas dynamics in the MSRE. There are serious limitations,

however, and a detailed study should be made beforehand.
58

2. The krypfon'experiment indicated that mass transfer coefficients
computed from heat-mass transfer analogies are quite good for tfie molten-
salt poroungraphite system. :

3., If the MSRE could be operated bubble free, the computed 135%e
poisoning would be 1.3 to 1.5% at 7.5 Mw. However, the reactor does not
operate bubble free, so the poisoning should be considerably less. This
results from the extreme insolubility of xenon in salt. When the model
is modified to include circulating bubbles, the compfited values can be
made to agree with preliminary measured values (0.3-0.4%) by adjusting
bubble parameters used over reasonably expected ranges. It would seem
therefore that the model does portray the physical ‘reactor. However,
to prove this conclusively, we must have accurateAknowledge of thé bub-
ble parameters, and with these calculate precisely the l?SXe poisoning.
We think.thé model is quite representative of this system and can easily
be extended to_éther fluid-fueled reaétors of this type.

4. This model should not be taken as final. For instance, it was
assumed that iodine does not volatilize. If it is later determined that
iodine does volatilizé, the model will have to be adjusted accordingly.

s, The'circulating helium bubble concept should be considered se-
riously as a 135%e removal mechanism in future molten-salt reactors. He-
lium bubbles could be injected into the flowing salt at the core outlet
and be removed with an in-line gas.separator some distance downstream.

6. Thé_insensitivity of 135%e poisoning to the graphite void frac-
‘tion and diffusion coefficient should be noted. This indicates that the
tight specifications of these Variables for the sole purpose of lowering
the 135%e poisoning might not be necessary. Considerable savings could
be realized in future_reactors. This phenomenon occurred in thé MSRE be-
cause the film coefficient is the controlling mechanism for transfer of
l?5Xe to the graphite. Each future reactor concept would have to be
studied in detail to assure that this was .still true before the above

statement was applicable.

59

ACKNOWLEDGMENTS

We are indebted to H. R. Brashear of the Instrumentation and Con-
trols Division for designing, building, and calibrating the radiation
detectors fised to monitor 85Kr, fo T. W. Kerlin of the Reactor Division
for his assistance in the finsteady-state parameter evaluation used in
the krypton experiment analysis, and to the members of the MSRE operatQ

ing staff for assistance in preforming the kryptonkexperiment.
16.

60

%

REFERENCES:

R. C. Robertson, MSRE Design and Operation Report, Part I, Descrip-
tion of Reactor System, USAEC Report ORNL-TM-728, Oak Ridge National
Laboratory, January 1965.

Oak Ridge National Laboratory, Reactor Chemistry Div. Ann. Progr.
Rept. Jan. 31, 1965, USAEC Report ORNL-3789.

Oak Ridge National Laboratory, Molten-Salt Reactor Program Semiann.
Progr. Rept. July 31, 1964, USAEC Report ORNL-3708.

F. F. Blankenship and A. Taboada, MSRE Design and Operation Report,
Part IV, Chemistry and Materials, USAEC Report ORNI-TM-7/31, Oak
Ridge National Laboratory (to be published).

I. Spiewak, Xenon Transport in MSRE Graphite, Oak Ridge National
Laboratory, unpublished internal report MSR-60-28, Nov. 2, 1960.

G. M. Watson and R. B. Evans, III, Xenon Diffusion in Graphite: Ef- -
fects of Xenon Absorption in Molten Salt Reactors Containing Graph-

ite, Oak Ridge National Laboratory, unpublished internal document,

Feh. 15, 1961. N

H. S. Weber, Xenon Migration to the MSRE Graphite, Oak Rldge National
Laboratory, unpublished internal document.

J. R. Waggoner and F. N. Peebles, Stripping Rates of Carbon Dioxide
from Water in Spray Type Liquid-Gas Contactors, University of Ten-
nessee Report EM 65-3-1, March 1965.

Ietter from F. N. Peebles to Dunlap Scott March 31, 1965, Subject:
Converting Stripping Rates from COj-Water System to Xe-Salt System.

F. F. Blankenship, B. J. Sturm, and R. F. Newton, Predictions Con-
cerning Volatilization of Free Iodine from the MSRE, Oak Ridge Na-
tional Laboratory, unpublished internal report MSR-60-4, Sept. 29,
1960.

B. E. Prince, Qak Ridge National Laboratory, personal commmnication
to authors.

F. N. Peebles, University of Tennessee, personal communlcatlon to
authors.

F. J. Salzano and A. M. Eshaya, Sorption of Xenon in High Density
Graphite at High Temperatures, Nucl. Sci Eng., 12: 1-3 (1962).

M. C. Cannon et al., Adsorption of Xenon and Argon on Graphite,
Nucl. Sci. Eng., 12: 49 (1962).

P. N. Haubenreich et al., MSRE Design and Operation Report, Part III,
Nuclear Analysis, USAEC Report ORNL-TM-730, Oak Ridge National Labo-
ratory, Feb. 3, 1964.

E. Ruckenstein, On Mass Transfer in the Continuous Phase from Spheri-
cal Bubbles or Drops, Chem. Eng. Sci., 19: 131-146 (1964).

17.

18.

19.

20.

21,

61

Li et al., Unsteady State Mass Transfer For Gas Bubbles — Liquid
Phase Controlling, Amer. Inst. Chem. Engrs. J., 11(4): 581-587 (July
1965).

P. Harriott, Mass Transfer to Particles, Part I — Suspended in Agi-
tated Tanks; Part II — Suspended in a Pipeline, Amer. Inst. Chem.
Engrs. J., 8(1): 93-102 (March 1962).

S. Sideman et al., Mass Transfer in Gas-Liquid Contacting Systems,
Ind. Eng. Chem., 58(7): 3247 (July 1966). .

R. B. Bird, W. E. Stewart, and E. N. Lightfoot, Transport Phenomena,
Wiley, New York, 1960.

C. V. Chester, Oak Ridge National Laboratory, personal communication
to authors.

.mr_ ,

i
M
%

etk e v

63

APPENDICES
R S—

N

>

65
Appendix A
MSRE PARAMETERS

General
Normal thermal power level, Mw
Nominal operating temperature, °F
Operating.pressure in pump bowl, psig
Fuel salt flow rate, gpm
Fuel salt volume, ft3
Graphite volume, ft3
Xenon stripper flow rate (estimated), gpm
Salt flow along shaft to pump bowl, gpm
Total bypass flow (sum of above), gpm
Fuel loop circuit time, sec
Graphite
Grade )
Bulk dénsity, g/cm3
Porosity, accessible to kerbsene; %
Porosity, theoretical, %
Porosity, available to xenon and kryptom,* %
Fuel salt absorption at 150 psig (confined to surface), %
Wettability
Graphite surface area in fuel channels, £t ;

Diffusivity of Kr at 1200°F in graphite (filled with He),
Kr atoms/hr.ft graphite (Kr atoms/ft’ gas)

Diffusivity of Xe at 1200°F .in graphite (filled with He),
Xe atoms/hr.ft graphite (Xe atoms/ft’ gas)

Equivalent diameter of fuel channels in bulk graephite, ft
Fuel Salt
Liquidus temperature, °F
Density at 1200°F, 1b/ft>
Viscosity at 1200°F, 1b/ftshr
Diffusivify of Kr at 1200°F (based on several estimated values), ftz/hr
Diffusivity of Xe at 1200°F (based on several estimated values), ft2/nr

Henry's law constant for Kr at 1200°F, moles of Kr per cc of salt
per atmosphere :

Henry's law constant for Xe at 1200°F, moles of Xe per cc of salt
per atmosphere :

* : '

From Ref. 2. : .
T .

Not wet by fuel salt at operating conditions.

CGB
1.82-1.87
4.0

17.7

~10

0.20

+

1520

~1.0 x 107%

~0.92 x 1074

¢.0519

840

130

18

4.3 x 107°=7.0 x 1077
3.9 x 107%=6.4 x 10~°
8 x 1079-9 x 10-°

2.75 x 107°
66

Appendix

B

SALT-TO-GRAPHITE COUPLING

A question exists concerning the process of mass transfer from a

fluid to a porous medium as opposed to a continuous or homogeneous me-

dium. For example, the krypton flux from graphite to salt is given by

k

' .k (k _
Flux~ = h A, (C., — C

v

where A, is the total channel surface area of the graphite.

G

hk is defined so that Ck
m sf

k)

S

The term

is not the krypton concentration in salt at

the salt-gas phase interface (located at a pore opening), but rather is

some continuous concentration across the entire surface of the graphite.

This is shown schematically below and would be similar for fhe case of

xefion flowing from salt to graphite.

Direction of Kr Flow

P — —

' ' Boundary Layer

A2

Flowing
Salt
(ck)

Salt-Gas Phase Interface (Cgi

Region of Conventional Mass Transfer

%,

)
"™ schematic Pore Contalnlng He and Kr

_ Where hfi Is Applicable

Equal Kr Concentration Profile (Ckf)

In this region Kr is transported from
the pore interface to the equal con- !
centration layer by a process of pure
diffusion

//j:;;;E;EEEj;// Schematlc G%gphlte
Matrix
//C/(////,/ /

7/

e

10{
— —— — ——— d— —— — — c—— A — ——

67

The mean pore entrance diameter for CGB graphite.isvless than O.1 H .
and probably closer to 0.02 i, which is extremely small compared with the
boundary layer thickness. It seems reasonable, therefore, that krypton
-is transported from the pore opening to the contihuous concentration layer
by a process of pure diffusion. Then from the continuous layer to the
bulk salt, krypton will be transported by‘conventional fluid dynamic mass
transfer. Note that since salt will not wet graphite, and because of
the small pore size, the salt-gas interface will certainly not peqetrate
inside a pore. Also inherent in this analysis is the idea that salt will
tofich the solid graphite matrix, even though it .will not.wet it, and
therefore the salt-gas phése interface will exist only at the pore open-
ing. In order to couple the salt to the graphite, it is necessary to

k Actually it will be shown

develop a relationship between Cgf and Csi'

that Ck. N,Ck .
si sf | o
In this development we consider a simplification of the previous

figure, as follows:

Boundary Layer

Voo
7,

k r.
. r i
s . f -y
Unit Transfer 1 (L Interface
Cell \ ~=——- Pore
-l T R ‘ Ei;// . 2 ; .
Csf \

Cor 7/—//

P

‘The pure diffusion region\@Ssociatéd with a single graphite pore is ap-
proximated in spherical geometry. The inner hemisphere at constant con-
centration Cii,and radius r. is the source for krypton that diffuses-

_through salt to the outer hemisphere at constant concentration sz and
68

radius rf.' Associated with each pore is a unit transfer cell of cross-
sectional area WT%, thrbugh which krypton is transferred by conventional

mass trans?er from the position of Cgf to thg bulk salt at condéntration
Cg. The term r. is taken as half the mean pore entrance diameter and ro

1s related to it with the graphite void fraction.
The general equation for diffusion in spherical coordinates at steady
state, and ‘when concentration is a function only of the radius, is

2
d Ck 2 de
S S

+=—— =0 .
dr? r dr

t

Solving with boundary values, as discussed above,

Yy
¢~ o, ==
S si T
kK k r.
| Csf Csi 1 —-;i
| i
Differentiating with respect to r,
ack  r, cf - cF
_ i st si
dr 2 Ir.
rT _.;l
f

k. dcls{,
Flux = —DYA
r S Tr dr
gives

S e
Flux = —D'a —= Sf 81

r r o2 r,

l——

r

69

Then, solving at r = ), we obtain

r. for a hemisphere (Ar = 2wr§
k Cif - Cii
Flux = —27r
T i’s Ty
_ l_I‘—
f

Without going into the considerations, we will state that a reasonable

relationship between r. and ra is

Substituting this into the equation for flux  gives
(E)¥/2
Fl1 —2wrka E (l( k‘)
e S 1/2 st si
c -
1"(3)

Now, at steady state, this diffusion flux at r_, must equal the convective

flux through the unit cell, where

f

.k : k K
. Fluxunit cell hmAunit cell chf Cs) ¢
or
o k(k _ k)
Flu'xu.nit cell thm(csf Cs )

Therefore, equating fluxr and fluxunit cell’ vC obtain

f

k k k . fe\/? ,
Csf —'Cs 2Ds (3)

k - k k 1/2 )
Csi Csf rfhm_l - (%)

Solving for the following parameter values,
70

DS = 4.26 x 107° £t2/hr,
h% = 0.06 ft/hr (approximate for bulk graphite region),
r; = 0.1 u (actually probably closer to 0.02 i),
€ = 0.10, '
r. = 0.446 1,
we obtain

sz "C:
* P
k. —¢

which says that,the'concentration difference between Cgi and Cgf is neg-
ligible compared with the difference between CK, and C¥. Another way of
putting it is that Cgi ] Cgf. The equation at the beginning of this ap-

pendix can now be written

&

71

Appendix C

THEORETICAL MASS TRANSFER COEFFICIENTS

Theoretical mass transfer coefficients between fuel salt and graph-
ite may be estimated by using standard heat transfer coefficient. relation-

20 These conversions

ships and the analogy between heat and mass transfer.
are brought about by substitution of equivalent groups from the following
table into the appropriate heat transfer coefficient relationship. They

apply for either laminar or turbulent flow.

Heat Transfer Mass Transfer
Quantity - Quantity
pd v pd Vv
Re = <q . Re = €4 .
35 35
, deq h.mdeq
Nuh = Nu =
k D
C u
Pr = B Sc = 1=
K eD

Applying these substitutions to the Dittus-Boelter equation for turbulent
flow, we have,

1. for heat transfer,

2. for mass transfer,

For laminar flow the following equations can be used:

1. for heat transfer,

k ( d_eq 1/3 k ‘pdgqvcp 1/3
hh ='l.86 d_‘ Re Pr T = 1.86 3 ‘ 5
eq eq

72

2. for mass transfer,

7 . D d 1/3 D d2 Vv 1/3‘
h = 1.86 T (Re Sc —Eg) = 1.86 ( eq)

m
eq eq

Calculated values of hm for krypton would then be as follows:

h,, Mass Transfer
Coefficient (ft/hr)

Center-Line
Bulk Graphite Graphite Region

Turbulent flow 0.115-0.155 0.250~0.338
Laminar flow 0.048-0.067

where the range in hIn reflects the range of D}; given in Appendix A. This

range in DS represents an expected range as determined from three sources:

1. as measured ind.irectly'21 from ané,logy'of the noble gas-salt system
to the heavy-metal ion-water system,
2. | estimated from the Stokes-Einstein equation,

3. estimated from the Wilke-Chang equation.

L
»

Term

€q

H O o

HEF'D_P" I ‘G—(D"ljm’-b‘m

@F O " > B =g

O
)
o)

73
Appendix D

NOMENCLATURE

Definition

Sfirface érea

Noble gas concentration. ,
Noble gas concentration in gas phase
Noble gas concentration in.gréphité‘xl |
Noble gas concefitration in salt

Heat capacity

Equivalent diameter of flow -channel
Diffusivity .

Pump bowl purging efficiency

Void fraction of graphite

Volume fraction of salt in element'Ve

Volume fraction of graphite in element Vg
Neutron flux
Henry's law constant for salt

Heat transfer coefficient

‘Mass transfer coefficient

Modified Bessel functions of the first
kind

Bessel functions of the first kind
Thermal conductivity

Lehgth of fuel channel

Radioactive decay constant
Poisoning

Volumetric flow rate

Volumetric flow rate of helium at 1200°F
and 5 psig through pump bowl

Volumetric flow rate of salt through
xenon stripper -

Radius

Radius at equivalent core block in cylin-
drical geometry

" Units

a1
atoms/ft3

atoms/ft3

.atoms/ft3

atoms/ft3
Btu/lb-°F
£t

ft2/hr

%

neutrons/cmz-sec
moles/cc-atm

Btu/hr.ft2.°F

ft/hr

Btu/hr.ft.°F
£t '
Term

ta o ™

S/

* 2 K® X

T4

Definition

Universal gas constant

Density of salt

Stripping efficiency, defined as the per-
centage of dissolved gas transferred from

liquid to gas vhase as salt is sprayed
through the xenon stripper

Bubble stripping efficiency, defined as
the percentage of 1*?°Xe containing bub-
bles that burst in passing through the
stripper and are replaced with pure
heljum bubbles.

Absorption cross section

Time

Half-1ife

Absolute temperature

Fluid velocity

Volume of salt in primary lodfi
Vblume of graphite in core
Volume at gas phase in pump bowl
Vélume of core element & .

Aibitrary constant

‘Viscosity

Ratio of fuel channel surface area to
graphite volume in core

Average void fraction of helium bubbles
circulating with salt in primary loop

Superscripts

135Xe

Adjoint flux
Mean
Bulk graphite region

Center-line graphite region

Units

cc.atm/°K-mole

1b/ft3
%

barns
75

Subscripts

Fast neutrons
Thermal neutrons
Boundary
Circulating bubbles

Element of core volume _

‘Film

Gas phase

Graphite

Heat

Interfacé

Mass

Initial conditions
Pump bowl

At radius f

Salt phase
L T

»

T L

(]

ORI NEC I T S

76.
7.

78.
79-80.
81-358.

ORNL-4069
‘UC-80 — Reactor Technology

Internal Distribution

C. F. Baes, Jr. 27. R. E. MacPherson

S. J. Ball 28. A. P. Malinauskas

W. P. Barthold 29. C. D. Martin, Jr.

S. E. Beall 30, H. A. Mclain N

'F. F. Blankenship 31. C. E. Miller, Jr.

E. G. Bohlmann 32. A. M. Perry

H. R. Brashear 33. M. W. Rosenthal

W. L. Carter 34, A. W. Savolainen

C. E. Center (K-25) 35=%40. Dunlap Scott

E. L. Compere , 41. M. J. Skinner

W. H. Cook 42. R. E. Thoma

F. L. Culler ‘ 43. A. M. Weinberg

J. R. Engel 44y M. E. Whatley /

R. J. Kedl 4547, Central Research Library

G. W. Keilholtz 48=49, Y-12 Document Reference Section
T. S. Kress 50—74. Laboratory Records Department
C. E. Larson (K-25) 75. Laboratory Records, RC

H. G. MacPherson | :

~ External Distribution

A. Giambusso, AEC, Washington

T. W. McIntosh, AEC, Washington

Division of Research and Development, AEC, ORO

Reactor Division, AEC, ORO

Given distribution as shown in TID-4500 under Reactor Technology

Category (25 copies — CFSTI)
c