ORNL-4233.txt

’1

, 3 4456 0356873 H

AL LABOR#

ROCMI IRDADV
al¥ial SHA M

'r..lj'v'A';l

-

Contract No. W-T40O5-eng-26

Reactor Division

ZERO-POWER PHYSICS EXPERIMENTS ON THE
MOLTEN-SALT REACTOR EXPERTMENT

B. E. Prince J. R. Engel
S. J. Ball P. N. Haubenreich
T. W. Kerlin

FEBRUARY 1968

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

ORNL-423%%

MARTIN MARIETTA ENERGY SYSTEMS I|J|BRARIES

3 4456 0356873 4

{
f

ii%

PREFACE

This report is a revised and expanded version of an internal memo-
randum titled "Preliminary Report on Results of MSRE Zero-Power Experi-

" issued in the summer of 1965, immediately following the com-

ments,
pletion of these experiments. It contains a complete description of all
the measurements made of the neutronic behavior of the reactor, before
the reactor was operated with substantial nuclear heat generation. Many
of the sections in the original internal memorandum have been carried
over essentially intact. In addition to these results, however, several
other experiments had been performed which could not be properly evalu-
ated by the time of the preliminary report writing, and which were of
necessity omitted or mentioned only briefly in that writing. Subse-
quently, the priority given to analysis of data obtained from power
operation of the reactor necessitated a delay in preparing a formal re-
port on this work. Rather than issue a separate report pertaining only
to these experiments, we feel that a coherent account of all the MSRE
zero-power nuclear experiments, in a single report, might be interesting
to a wider range of readers as well as provide useful retrospect in
later stages of development of the molten-salt reactor program.

Both the performance of the zero-power nuclear exXperiments and the
writing of this report were a group effort, with the former involving a
much larger group than the latter. General areas of responsibility, how-
ever, were as follows: The rod calibration experiments and the "static"
reactivity coefficient measurement program were planned and analyzed by
J. R. Engel and B. E. Prince. Frequency response measurements and ex-
periments involving the dynamic effects of temperature and pressure on
reactivity were devised by 5. J. Ball and T. W. Kerlin. Flux noise
measurements were made by D. N. Fry and D. P. Roux. P. N. Haubenreich
supervised the initial critical experiment and coordinated the remainder
of the activities. Many members of the reactor operations staff provided

indispensable aid in carrying out all of these experiments.

The Authors

CONTENTS

Preface L A L A A N N I I N A N I R N A A N R E R E N E N * & ¢ 8 2 2 & 9
Abstract ® & & 5 2 08 s 000 s ® 4 8 8 & 8 8 s 0 0SS RS L EE YT e s LI A B R R B I 8 o o o 2 s
l L] IntrOduC tion LA LA 2 B N R N N N A N A N I A I I A T N R R T

2. Initial Critical EXperiment ..e..oeveeeeeseecsesoonsessennss

5. ContrOl-ROd Calibration o 4 6 o8 28 a L R R R R R Y R N I A N A N T R ]

5.1 General DesSCripPtion suiveeeeeeeeceesaceescenosscacocenss
3.2 Theoretical Guidelines ....ceeveeneas Peecesssssasenanas
3.5 Differential-Worth Measurement: Fuel Stationary .....
3.4 Reactivity Equivalent of 225U Additions ..e..veee.o...
5.5 Rod-Shadowing EXperiments ..ue.eeeseseececesescenenses
3.6 Reactivity Effects of Fuel Circulation seeeeeeee.ooo..
5.7 ROA-Drop EXPeriments seu.e.ieeeecevetecoceocennoneness

3.8 Comparison of Measurements with Theoretical Analysis
OfcontrOl"'ROdworth ...li...-.l..I..ll........‘......

L. Temperature and Pressure Reactivity Effects «.eeeeeeseee...
k.1 Isothermal Temperature Coefficient of Reactivity .....
L.2 Fuel Temperature Coefficient of Reactivity ee.eeveoe...
L.,3 Effect of Pressure on ReacCtivVity teeieeececeeecaneenss
5. Dynamics TeStsS seeeecieerannans S eesssecsseeatcetaettaanrane

5¢1l Purpose OFf TeStS cvecrerrerareescatsecscannseeesacnnnens

5.2 Frequency-Response MeaSUrements +.veeeesocsosesvosesos

5.3 Pulse Tests wieeeereevrrrrscersecocscennoosasnsanssacsns

5.4 Pseudorandom Binary Sequence TeStS v.veeeo.. Ceeeeeaaa
5.5 Neutron Fluctuation Measurements ...eeee.. cseeracun .o
5.6 TranSient FlOW-Rate TeStS [ N N A R E N T

5.7 Conclusions from Dynamics TesStS ceeeeeereserssocceeese

6. General Conclusions (IR L B 2 B D B Y B BN I BN Y N RN BN RN B B A A I A A B A I B

References .l.....l.l.....l...'...."..I....Il...l.l..'.l..l.Il

iii

11
11
12
15
18
19
23
25

31
37
37
41
42
45
45
45
46
47
50
51
54
54
56

Fig. No.

10

11
12
13
1k
15
16
17

18

vii
LIST OF FIGURES

Title

Source and Instrumentation in Initial Critical
Experiment ...iiveveecan.. teeens teeneen Ceees e et e s

Relation of Rod Position and Levels in Reactor
Vessel * & & & & & » & & % 4 5 8 & 9 b e s s s ® & & 8 8 5 5 8 " A B 8 o 8 & & ¢ & & 9 @ > 9

Count Rate Ratios After First Four Additions of
2357, (Vessel full, rods at 51 in., source at
829 ft-9 in., chamber locations as in Fig. 1.) .....

Graphical Description of Control-Rod Calibration
Experiments ...... et e s erestsarssat et asean e ccieean

Differential Worth of Control Rod No. 1, Measured
with Fuel Stationary. (Normalized to initial

critical Z2°U 10ading.) veeeeeeenneenn. et ..
Integral Worth of Control Rod No. 1 t.evivivvencenna.
Effect of °°U Mass on Reactivity ........... ceeaean
Change in Critical Position of Rod 1 as Shim Rods

2 and 3 are Inserted into COre ceveeerrireeceens ceess
Lattice Arrangement of MSRE Control Rods and

Sample Holder coveiveecenreonnes ce e ceerenncans csae

Differential Worth of Control Rod 1, Measured with
Fuel Circulating. (Normalized to initial critical
2350 10ading.) cievenronniinnn. Cecerrcecsereneracenn

Results of Rod-Drop Experiments After %0 Capsule
Additions .eeveass s eecsetticenteaconnnresna ceenaas

Results of Rod-Drop Experiments After 65 Capsule
AddIitions veeevvverenensones cer e ceicerersiserecnees

Results of Rod-Drop Experiments After 87 Capsule
Ad—d—itions ® & & o & & & & 0 & § 8 8 T S G O S .S P E O S S F O P S T e s . * 8

sensitivity of Rod-Drop Experiment to Changes in
Magnitude of Reactivity Insertion .e..eecececerenecs

Geometric Models of MSRE Core Used for Nuclear
CalculationNsS eeeeseeenceccncannnoen st cescsneseane e

Effect of Slow Changes in Core Temperature on the
Reactivity eoveeeess Ceeeasereasrtsasr a0t ass s e .o

Photograph of a Three-Dimensional Plot of the
Reactivity Measurement Data ..... beeasesas e ceaenns .

Conditions During Rapid Pressure Release While
Circulating Helium Bubbles .ieiieeeiicectanrscecnnces

Page

13

17
18
19

20

21

25
29
30
30
31
33
38
40

bty

viii

Fig. No. Title Page

19 Reactivity-Pressure Frequency Response with 2%
to 3% Void Volume in Circulating Fuel. (Calcu-
lated from pressure release experiment using
Samulon's method with 0.2 min sampling in-

te—r‘val ) ......... 4 % 9 0 5 2 9 & B L ] o & 8 8 9 ” & s & & 9 a ® & 8 8 * 5 8 2 45
20 Frequency Response of (6n/n )/(Ek/k ) at Zero

Power; Fuel Stationary ..v.ieeereensereenneeeennnnas 48
o1 Frequency Response of (En/n )/ ( (ok/ky,) at Zero

Power; Fuel Circulating vuueeiveeeeeennonnneennnn, .. 49
oD Pump Speed and Flow Startup Transients ..eeeeeeee... 52
2% Pump Speed and Flow Coastdown Transients ...... . 53
24 Control-Rod Response to Fuel-Pump Startup and

COASTAOWIL 4 vvevenvnnsosnnsenenns C et teieecee e 54

ZERO-POWER PHYSICS EXPERIMENTS ON THE
MOLTEN-SALT REACTOR EXPERIMENT

B. E. Prince J. R. Engel
o, J. Ball P. N. Haubenreich
T. W. Kerlin

ABSTRACT

This report describes the techniques and results of a
program of experiments designed to measure the important
neutronic characteristics of the MSRE, under conditions of
negligible nuclear heat generation. The program includes
the initial critical °°U loading, the control-rod cali-
bration (period~differential worth and rod drop-integral
worth measurements), determinations of the reactivity loss
due to fuel circulation, the "static" reactivity coef-
ficients of excess #2°U concentration and isothermal core
temperature, the fuel salt temperature reactivity coef-
ficient, the pressure effects on reactivity, and a series
of system dynamics tests (frequency response, transient
flow, and neutron flux noise measurements). These measure-
ments, carried out in June 1965, form much of the experi-
mental baseline for current evaluation of the nuclear
operation at full power. The report includes discussions
of the comparisons of the measurement results with the
corresponding neutronic characteristics calculated from
theoretical models.

1. INTRODUCTION

A program of zero-power nuclear experiments, including the initial
critical experiment, was conducted on the Molten-Salt Reactor Experiment
(MSRE) in June 1965. The purpose of this program was to establish the
basic nuclear characteristics of the reactor system and provide a base-
line for evaluation of the system performance in nuclear operation. A
secondary purpose was to evaluate the calculational techniques and models

used in predicting the properties of the MSRE.

The initial critical experiment established the minimum critical
concentration of #?U in the fuel under the simplest possible condition;
that is, with core isothermal, fuel salt stationary, and control rods
withdrawn to their upper limits. The remainder of the tests were de-
signed to provide information about control-rod worths, various re-
activity coefficients, and dynamic behavior of the system, all under
zero-power conditions.

With the initial critical concentration established, more 235U was
added to the circulating loop in increments to permit the attainment of
criticality with the salt circulating and with various control-rod con-
figurations. Measurements were made of the differential worth of one
control rod as a function of position, both with the fuel salt stationary
and with it circulating. In addition, rod-drop experiments were per-
formed to provide an independent determination of the integral worth of
various control-rod configurations. Measurements of the critical
control-rod configurations as a function of uranium concentration, both
with and without fuel circulation, provided information about the 235U
concentration coefficient of reactivity, the effect of circulation on
reactivity, and control-rod shadowing effects. At several fixed =335U
concentrations, the reactor system temperature was varied to provide data
on the isothermal temperature coefficient of reactivity. Tests were
also made in which the system overpressure was varied to evaluate the
Pressure coefficient of reactivity. Several types of measurements were
made to provide information about the reactor dynamics. These included
the response of the system to single and pseudorandom sequences of re-
activity pulses, the response to flow and temperature transients, and
neutron flux noise data.

sufficient excess uranium was added during this program to permit
calibration of one control rod over its entire length of travel. This
was expected to provide enough excess reactivity to compensate for all
transient effects associated with full power operation.

Since the principal independent variable in these experiments was
the ?°U concentration in the fuel, the various tests were scheduled

around the uranium additions. Thus, many of the experimental tests were

interwoven chronologically‘to provide the required data, The results
presented in this report deal with individual topics without regard for
the actual chronology of the tests.

In describing the results of the experiments, some reference to the
reactor physics theoretical background is often an indispensable aid in
interpretation., For the purpose of this report, we have limited this
either to brief qualitative descriptions or to summaries of calculated
core characteristics. Sources of details of the MSRE physics analysis
are ref, 1 and various MSRE semiannual progress reports cited in the

following sections.
2. INITIAL CRITICAIL, EXPERTMENT

The purpose of thilis experiment was to provide a check on the calcu-
lations of critical concentration under the simplest conditions, that is,
with the core isothermal, the control rods fully withdrawn, and the fuel
stationary. It also served to establish the basepoint from which the
2357 additions necessary to reach the operating concentration could be
made with confidence.

The fuel salt composition specified for power operation is 65LiF-
29,2BeFo-52rF,-0.8UF, (expressed as molar percentages). The total
uranium content is considerably above the minimum required for criti-
cality if highly enriched uranium were used, and was chosen for reasons
of chemistry. With this total uranium content, theoretical calculations™
predicted that the reactor would be critical at 1200°F, rods cut, fuel
stationary with 0.256 mole % 2°5UF, {0.795 mole % total UF,).Z

Instead of using 52%—enriched uranium to make up the fuel salt, we
decided to start with depleted uranium in the salt and add the required
amount of 27U as highly enriched uranium (93% =°°U). This permitted
preliminary operation with uranium in the salt before the beginning of
nuclear operation and alsc facilitated the manufacture of most of the

uranium-bearing salt. The salt was prepared in three lots: the carrier

x
A review of the basis of these calculations is included in Sec. %.8.

sait, containing the beryllium, zirconium and most of the lithium
fluorides; 73LiF-27UF, eutectic containing 150 kg of depleted uranium;
and eutectic containing 90 kg of U in the highly enriched form.

Thirty-five cans of carrier salt and two cans of eutectic containing
the depleted uranium were blended as they were charged into a drain tank.
This mixture of salt was then circulated for 10 days at 1200°F while the
sampler-enricher was tested and 18 samples were analyzed to establish
the initial composition. The critical experiment then consisted of
adding enriched uranium in increments to bring the 235U concentration up
to the critical point,

Nuclear instrumentation for the experiment consisted of two fission
chambers, two BFs chambers, and an 2%41Am-242Cm-Be source, located as
shown in Fig. 1. The fuel salt itself also constituted a neutron source,
due to reaction of alpha particles from 2°%U with beryllium and fluorine.

The enriching salt was added in two ways: by transfer of molten
salt from a heated can into a drain tank, and by lowering capsules of
frozen salt into the pump bowl via the sampler-enricher. The latter
method was limited to 85 g 225U per capsule, only 0.0012 of the expected
critical loading. Therefore the bulk of the 23°U was added in four
additions to the drain tank. After each addition the core was filled
and count-rate data were obtained to monitor the increasing multiplication.

The amount of 2°°U expected to make the reactor critical was calcu-
lated to be 68.7 kg, using the volumetric concentration from the criti-
cality calculations and the volume of salt in the fuel loop and drain
tank.

Before the addition of enriched uranium, count rates had been deter-
mined with barren salt at several levels in the core. Then as the core
was filled after each “°°U addition, the ratio of count rates at each
level was used to monitor the multiplication. (Figure 2 shows elevations;
count rates were determined with salt at 0.4, 0.6, 0.8, and 1.0 of the
graphite matrix and with the vessel full.)

Count-rate ratios with the vessel full after each of the four ma jor
additions are shown in Fig. 3. Each addition, fill, and drain took

between one and two days, so only four major additions had been planned.

ORNL-DWG 65-7575

REACTOR INLET
LINE 102 <

SOURCE TUBE {-in.BF3 CHAMBER TUBE

~— REACTOR QUTLET
LINE 100

34°

REACTOR
VESSEL

2~in. BF3 CHAMBER

FISSION CHAMBER
NO. 2

THERMAL SHIELD

INSULATION FISSION CHAMBER
NO. ¢

NUCL. INST.

PENETRATION
PLAN
SOURCE TuBE T
 com—{ QUTLE
. 1 /

| —THERMAL SHIELD

~
\

|~ INSULATION

2-in. BF3
CHAMBER

TOP OF GRAPHITE
ELEV. 8331t 5% in.

ELEV. B3O ft 8in.
(CORE MIDPLANE)

REACTOR VESSEL

ELEV. 829 ft 9in.
NUCL. INST.

PENETRATION

ELEV. 8281t 3Y2in BOTTOM OF GRAPHITE FISSION CHAMBERS

ELEV 828ft QY% in.

vy

1-in. BF3 CHAMBER
BOTTOM ELEV.
828 f1 3in.

ELEVATION

Fig. 1. Source and Instrumentation in Initial Critical Experiment.

ELEVATION (ft)

ORNL-DWG 65-7573

836 836
I L
REACTOR OUTLET
835 ft —3%in. - - - -
835 (- 835
(o  B34ft—2.15in. |
834 — // 834
0.9 1o — 833f1=5%in. TOP _OF MOST GRAPHITE
833 f1—3 in- 51 — 1l UPPER ROD LIMIT
833 —— 48
os L 0.9 8 : 833
3 -
< — 42 — :J- 1
o 3 08 - =
Fo7l K u =
832 —— € = 36 | > 832
& £ 07 . - -
o6 = £ u w
6 — =z il -
@ B = 30f =
O ~ 08 O - =
O x = - T
831 —— w @ E 24 - 831
> 05 |- = 0 -
= < O o o
5 s 0.5 a - o
Ll W o 18 '
= - O o 3
c =
W o4 - T oa - =
830 ——§ < 12 |- = 830
a - O
b o [ Q
O 0.3 C
E 03} 5 6 -
— 1
I z - DRIVEN ROD
= u LOWER LIMIT
829 —— " | E 02 = 829
0.2 O
[ § -
15 —g L.
[T
0.4 SCRAMMED ROD
LOWER LIMIT
0.1 | - o
| 828ft—0Ygin. HORIZONTAL GRAPHITE
828 —— 0 828
oL 827 ft—7.15in. \
827 —~— | 827

Fig. 2.

Relation of Rod Position and Levels in Reactor Vessel.

ORNL-DWG 65-7574
06 T R T “" B

2-in. BF 3 CHAMBER
-in. BFy CHAMBER
FISSION CHAMBER NO.1

FISSION CHAMBER NO.2
05 - — -— - — — -

Q [ J [~

/i/

COUNT RATE RATIO (GCR, /CR)
o
w
\
T

\
~

40 44 48 52 56 60
MASS OF 239U IN TOTAL SALT CHARGE { kg)

|
/
e

Fig. 3. Count Rate Ratios After First Four Additions of ?3°U. (Ves-
sel full, rods at 51 in., source at 829 ft-9 in., chamber locations as
in Fig. 1.)

After the third addition, with 6L.54 kg 235U in the salt, the projected
critical loading was T70.0 + 0.5 kg £2°U. (The l-in. BFs chamber, located
in the thermal shield, whose count rates extrapolated to a higher value
was known to be strongly affected by neutrons coming directly from the
source.) The fourth addition was intended to bring the loading to about
1 kg below the critical point. After 4.38 kg of 235U was added, the
count rates showed the loading was within 0.8 kg of critical when the
rods were withdrawn and circulation was stopped. Preliminary estimates
of rod worth and circulation effect, based on changes in subcritical
multiplication, were approximately the expected values.

In the final stage, enriching capsules were added through the pump
bowl to bring the loading up 85 g at a time. After each addition, circu-

lation was stopped, the rods were withdrawn, and count rates were measured.

With the reactor within 0.2% ak/k of critical, slight variations in temper-
ature caused considerable changes in multiplication. (Variations in the
voltage of the area power supply change the heater inputs slightly, re-
quiring fine adjustments of the heater controls to keep the temperature
precisely at a specified value.) After seven capsules, it appeared that
after one more, the reactor could be made critical., The eighth was added,
circulation was stopped, and the rods were carefully withdrawn. At ap-
proximately 6:00 p.m., June 1, 1965, the reactor reached the critical
point, with two rods at full withdrawal and the other inserted 0.0% of
its worth. Criticality was verified by leveling the power at successively
higher levels with the same rod position. The £°°U loading was 69.6 kg.
During the approach to critical, a substantial internal source of
neutrons was observed. The MSRE fuel mixture has an inherent source of
neutrons produced by the interaction of alpha particles (primarily from
2347) with the beryllium and fluorine. Measurements were made with the
reactor only slightly subcritical to evaluate the intensity of this
source. Count-rate determinations with and without the external neutron
source in place, under otherwise identical conditions, showed that the
internal source supplied 0.0% to 0.0 as many neutrons to the core as the
external source. The external source at that time had an absolute in-
tensity of 1 X 10® neutrons/sec. However, because of the source location
in the thermal shield (Fig. 1), some distance from the reactor vessel,
only a small fraction of these neutrons are effective in reaching the
core. If this fraction were 10%,* the effective external source con-
tribution would be 1 X 10’ neutrons/sec, and thus the internal source
strength would be in the range of 3 X 10° to 5 X 10° neutrons/sec. The
calculated intensity of the internal source was within this range.l
Predicted and observed 2°°U requirements for criticality are com-
pared most logically on the basis of volumetric concentration. The
required volumetric concentration of #°°U is nearly invariant with re-

gard to the fuel salt density (unlike the required mass fraction, which

*The offset location of the Am-Cm-Be source makes it difficult to
calculate this fraction reliably. However, the value of 10% is com-
patible with the results of diffusion-theory calculations.?t

varies inversely with salt density) and depends not at all on system
volume or total inventory. The observed 22°U concentrations, however,
are obtained in the first instance on a weight basis, either from in-
ventory records or from chemical analyses. These weight fractions must
therefore be converted to volumetric concentrations by multiplying by
the fuel salt density.

The amounts of ©°°U and salt weighed into the system gave a 223U
mass fraction of 1.4l + 0.005 wt % at the time of the initial criticality.
The chemical analyses during the precritical operation and the zero-power
experiments gave uranium mass fractions which were 0.985 of the "book"
fractions. Applying this bias to the book fraction at criticality gave an
"analytical™ #3°U mass fraction of 1.39% wt %. On a statistical basis,
the uncertainty in the mass fractions obtained from chemical analyses is
about +0.007 wt %.

At the time of the zero-power experiments, we recognized that a
small amount of dilution of the fuel salt should occur, due to residues
of flush salt left in freeze valves and drain-tank heels when the fuel
salt was charged, (During the initial fill operations, ’'LiF-BeF, flush
salt was admitted to the fuel circulating system.) Experience with drain-
flush-fill cycles obtained from MSRE operation subsequent to the zero-
power experiments has indicated that the fuel salt would have been
diluted by 20 + 10 kg of flush salt, If we assume that this amount of
dilution occurred, the corrected value of the book mass fraction of Z3°U
would be 1.408 + 0,007 wt %.

The density of the fuel salt at 1200°F was determined after the
uranium was added to the fuel drain tank, using pre-calibrated drain
tank weigh cells and salt level probes within the tanks.” The average
of four measurements was 145.1 1b/ft®, with a maximum deviation of 1.1
lb/ft3. These weigh cell measurements were in close agreement with an
indirect determination of the density, inferred as follows. The density
of the fuel carrier salt (65LiF-30BeF.-5ZrF,) was measured as the salt
was charged to the fuel drain tank. This measured density, computed from
externally measured weights and the volume between the level probes within

the tanks was 140.6 1b/ft® at 1200°F. Addition of all the uranium added
10

during the zero-power experiments would be expected to increase the den-
sity to about 145.9 1b/ft>.

Concurrent with the zero-power experiments, laboratory glove-box
measurements of the fuel salt density were made.,* These experiments gave
an average density very slightly larger than the MSRE measurements, but
the statistical uncertainty was sufficiently large that little additional
information could be provided. TFor the calculations given below, we have
used 145 = 1 1b/ft> as our best estimate of the density of the fuel salt
at 1200°F, and with the uranium concentration at the time of initial
‘criticality.

In comparing the observed and calculated critical concentration of
235U, a small temperature correction should be applied to the salt den-
sity given above, since the core temperature at the time of criticality
was 1181°F instead of 1200°F. Based on a fractional change in density of
—1.2 X 107%/°F (see discussion in Sec. 4.1), the density at 1181°F would
have been 145.3 + 1.0 lb/ft3. Finally, corrections must also be applied
to the calculated critical concentration, both for the lower temperature
and the fact that one rod was at 46.6 in., compared to the reference
conditions of 1200°F and all three rods at maximum withdrawal, 51 in.
These two effects nearly compensated for one another. The calculated 235U
concentration for criticality at the reference conditions was 3%2.87
g/liter; corrected to the actual conditions, using measured values of
the temperature coefficient of reactivity and the control-rod worth
increment (see later sections), it is 32.77 g/liter. This "predicted"”
value is compared with "observed" £°°U concentrations in Table 1. Con-
centrations corresponding to both the book mass fraction, corrected for
the flush salt dilution, and the analytical mass fraction, described
above, are listed in Table 1. The predicted concentration was found to
be in remarkably close agreement with the observed concentration cor-
responding to the corrected boock value of the mass fraction, and to be
very slightly higher than the concentration calculated from the analyti-

cal mass fraction.

11

Table 1. Comparison of Critical 235U Concentrations

(1181°F, pump off; 0.08% &k/k rod poisoning)

235 -
235U Mass Fraction Fuel Density U Concen

= tration
Predicted 32T
Corrected book 1.408 + 0.007 145 + 1 32.8 + 0.3
Analytical 1.39% + 0,007 145 + 1 32.4 £ 0.3

3. CONTROL-ROD CALIBRATION

3.1 General Description

The addition of Z2°U beyond the minimum critical loading had a two-
fold objective: to end with enough excess reactivity to permit operation
at full power and in the process to make measurements which could be
analyzed to give control-rod worth and various reactivity coefficients.
The final amount of 2°°U was to be enough to be critical at 1200°F with
the fuel stationary and one rod fully inserted. The general method was
to add 85 g 275U at a time through the sampler-enricher, after the
addition determine the new critical rod position, and at longer inter-
vals do other experiments.

Following the initial critical experiment, another eight capsules
were required before the reactor could be made critical at 1200°F with
the fuel pump running. This was a consequence of the effective loss of
delayed neutrons due to precursor decay in the part of the circulating
system external to the core. Once this #2°U concentration had been
reached, the critical position of the control rod to be calibrated
(designated as the regulating rod) was measured after addition of each
capsule, with the fuel pump running. At intervals of four capsules,

period measurements to determine control-rod differential worth were also
12

made with the pump running, Then the pump was turned off, the critical
rod position with the fuel stationary was determined, and period measure-
ments were made with the fuel stationary. This went on until a total of
87 capsules had been added. Three times during the course of additions
of 2°°U (after 30, 65, and 87 capsules) rod-drop effects were observed.
The results of all these experiments are considered separately in the

following sections.

3.2 Theoretical Guidelines

oome useful theoretical guidelines in interpreting the control-rod
experiments described in the sequel can be obtained by reference to the
curves in Fig. 4., Bach curve is a qualitative graphical description of
the change in the static reactivity”® as a function of regulating-rod
position, with the other two rods withdrawn to their upper limits
(position 0). The various curves represent different total loadings of
2357, increasing in the direction shown by the arrow, The static re-

activity, p_, corresponding to each specific rod position and 235y

loading is defined by the equation:

P, = 7 > (1)

where p is the actual number of neutrons emitted per fission, and p is

the fictitious value for which the reactor with the specified rod position

*Experimental measurements of the reactivity effects associated with
substantial changes in core conditions, such as control-rod insertions,
fuel additions, and temperature variations, require some care in inter-
pretation. Thisg is particularly true here, where the results of using a
mixture of techniques, such as static measurements by compensating re-
activity effects, and dynamic measurements by period and rod-drop experi-
ments, are to be interpreted on a consistent basis. We have used the
static reactivity concept and scale as a basis for in integrated and
unified interpretation of the measured reactivities. This was done by
introducing normalization corrections, wherever necessary for consistency,
in the manner described in this section, and also by avoiding instances
where important differences between the static reactivity and the re-
activity inferred from experiment can occur. The problems of reactivity
measurement and interpretation have been quite thoroughly explored in the
reactor physics literature. The discussions given in refs. 5, 6 and 7
are particularly relevant to the present work.
13

ORNL-DWG 67-12322

FINAL 23%U LoaDiNG

INCREASING 23%y

INITIAL 235U LOADING

Fd — — — ——— — ——— — — —— —— — — —— — —

FRACTION OF INSERTION OF REGULATING ROD
(CIRCULATION STOPRED)

Fig. 4. Graphical Description of Control-Rod Calibration Experi-
ments.

and material composition, and with the fuel stationary, would be just

critical. An equivalent expression is:
p. = — , (2)

where ke is the effective multiplication constant of the reactor. Since

ke (or equivalently, pS) is the quantity normally calculated in reactor

14

physics analysis programs, it is convenient to attempt to interpret the
experimental measurements of reactivity on a basis consistent with the
theoretical analysis (Sec. 3.8).

One may observe from ¥Fig. 4 that, if the reactivity equivalent of
the 23°U addition is known, a direct means of calibration of the reac-
tivity worth of the regulating rod is provided, simply by relating the
critical position of the rod (solid points shown as examples in Fig. 4)
to the 225U loading. Alternatively, calibration of the rod by independent
experiments provides an empirical determination of the reactivity worth
of the additional #7°U, or the concentration coefficient of reactivity.
This latter approach was chosen, and the experiments specifically aimed
at determining rod worth were the stable-period measurements and the rod-
drop experiments. In Fig. 4 a typical measurement of the stable period
corresponds to a motion from the critical position upward and to the left
along the short segment marked (p). The measured change in reactivity
along the vertical axis, pp, is divided by the increment in rod motion,
and this sensitivity, or differential worth, is ascribed to the mean
position. A typical rod-drop experiment is indicated in Fig. L4 by the
segment marked (d), extending from the initial critical position into the
subcritical region. The purpose of this experiment is to measure the
negative reactivity inserted by the drop, marked Pq*

One other characteristic of some importance is indicated in Fig. L.
Because the reactivity worth of the rods is affected by the 225U concen-
tration in the core, one finds that Py, < ,psol’ or equivalently, that
the curves representing different fuel loadings are not exactly parallel,
Although the 275U loading was continually being increased during the
course of these experiments, it is useful for purposes of consistency to
interpret the combined reactivity measurements, over the whole range of
rod movement, on the basis of a single mass of “°°U., Theoretical calcu-
lations of the rod worths, summarized later, were used to determine the
effects of the 235U concentration on total worth, and these corrections
were used as an aid in normalizing the experimental reactivity measure-

ments to a single 27U loading.

15

3.3 Differential-Worth Measurement: Fuel Stationary

Pericd measurements were generally made in pairs. The rod being
calibrated was first adjusted to make the reactor critical at about
10 w. Then it was pulled a prescribed distance and held there until the
power had increased by about two decades. The rod was then inserted to
bring the power back to 10 w and the measurement was repeated at a some-
what shorter stable period. Two fission chambers driving log-count-rate
meters and a two-pen recorder were used to measure the period. The
stable period was determined by averaging the slopes of the two curves
(which usually agreed within about 2%). Periods observed were generally
in the range of 30 to 150 sec.

For the measurements with the pump off, the standard inhour re-
lation® was used to calculate the reactivity increment corresponding to w,

the observed stable inverse period, viz.,
i
o= WA+ ) —— (3)

The decay constants, Ai’ and the effective delay fractions, Bi’ used in
these calculations, are listed in the second and fourth columns of Table
2. These delay fractions contain approximate corrections for the in-
creased importance of delayed neutrons because of thelir emission at lower
energies relative to the prompt fission neutrons in the MSRE.® The
neutron generation time, A, was 2.6 X 10™% sec for the initial critical
loading, obtained from theoretical analysis. When applied to the analysis
of period-rod sensitivity measurements, Eq. (3) is quite insensitive to
neutron generation time.

Prior to pulling the rod for each period measurement, the attempt
was made to hold the power level at 10 w for at least % min, in an effort
to. help insure initial equilibrium of the delayed neutron precursors.
Generally, however, it was difficult to prevent a slight initial drift
in the power level (as observed on a linear recorder), and corrections
were therefore introduced for this initial period. The difference between

the reactivity during the stable transient and the initial reactivity,

16

Table 2. Delayed Neutron Fractions in the MSRE

10%* x Delay Fraction (n/n)

Decay Constant

Group - .
(sec 1) Effective
Actual (static fuel)
1 0.0124 2.11 2.23
2 0.0305 14.02 14.57
3 0.111% 12.54 13.07
Y 0.3013 25.28 26.28
5 1.140 740 7.66
6 3.010 2.70 2.80

both computed from Eq. (3), was divided by the rod movement to obtain
the rod sensitivity at the mean position.

The differential-worth measurements made with the fuel pump off are
plotted in Fig. 5. As discussed above, theoretical corrections have
been applied to these measurements to put them all on the basis of one
235U concentration, arbitrarily chosen as the initial critical concen-
tration at the beginning of the rod-calibration experiments. Theoretical
calculations described in Sec. 3.8 indicated that the static reactivity
worth of a single rod is reduced by nearly 9% of its total worth for the
total addition of 2°°U made during the course of these experiments. The
approximate correction factors which were applied to the rod sensitivity
measurements summarized in Fig. 5 increase linearly with 222U concen-
tration, up to 1.087 for the measurements made near the final concentra-
tion (corresponding to the points between 1 and 2 in. withdrawal).

Some imprecision (scatter) is evident in the data shown in Fig. 5,
because the differential worth is the ratio of the increment in re-
activity to the increment in rod withdrawal, each of which is subject to
experimental error. The root-mean-square deviation of the data points
of Fig. 5 from the curve is about 2.8 X 107%*% &k /k/in., or ~0.7% of the
mean differential worth; the maximum deviation of a single point was

8.7%. TFor the short term type of experiments described in this report,

17

ORNL-DWG 65-10292

0.07 | I

[ )
0.06 o ~3

7{'

0.05 o7

0.04 @ e,

0.03 . - . B W -

/ .
0.02 //4 N

0.01}

DIFFERENTIAL WORTH [(% 84/4)/in]

0 4 8 12 {6 20 24 28 3z 36 40 44 48 52
ROD POSITION {in. withdrawn)

Fig. 5. Differential Worth of Control Rod No. 1, Measured with Fuel
Stationary. (Normalized to initial critical “2°U loading.)

the precision of determining the rod position was about £0.01 in. Probably
the most important source of imprecision in the differential worth was in
the measurement of reactor period. As described above, only the con-
ventional reactor instrumentation was used in recording this data. De-
termination of the period in each measurement involved laying a straight-
edge along the pen line record of the log n chart and reading the time
interval graphically along the horizontal scale which corresponded to a
change of several decades in the neutron level. Since these charts, to-
gether with the pen speeds, are subject to variations, this was a prob-
able source of error in the rod sensitivity measurements.

Figure 6 shows a curve of the magnitude of rod reactivity vs position,
which is the integral of the differential worth curve in Fig. 5. An in-
tegral worth curve is also shown which is normalized to the final £2°U
concentration, This latter is simply the first curve reduced by a factor

of 1.087.
18

ORNL—DWG 65—-10293

(65.25 Kg

'8 —  ~- 1 1 1  CRITICAL LOADING ::::: - —

FINAL LOADING
235

REACTIVITY WORTH (% 84/k)
ro

1.71 Kg U IN LoorP) T
1.0 — -— —t -—
ogr— { —o{ 4 LA —t 4 b1 ] ] - —
06 — - — -+
.4 |— — —— —
0.2 |— - — -
0
0 4 8 12 16 20 24 28 32 3% 40 44 4B 52

ROD POSITION (in. withdrawn)

Fig. 6. Integral Worth of Control Rod Neo. 1.

3.4 Reactivity Equivalent of 225U Additions

Following the initial achievement of criticality with the fuel pump
rumning, the effect of each capsule addition on the critical position of
the control rod was measured. The critical position of the control rod
was measured with the pump off after every fourth capsule. Critical rod
positions at each 22U level were then converted to reactivity by using
Fig. 6 and linearly interpolating between the initial and final loadings
to correct for the °°U concentration effect on the total rod worth. The
results are shown in Fig. 7. In this figure, the ordinate is the total
excess static reactivity which would result from withdrawing the rod from
its critical position at that 22U loading to the upper limit of rod
travel, The separation between the two curves reflects the net reduction
in reactivity due to emission of delayed neutrons in the external piping

and heat exchanger (see later discussion).

19

ORNL-DWG 65—10291

20 - e : :
o

FUEL NOT CIRCULATING~Y

o FUEL CIRCULATING
10 s e e /i;#”flffl e ]
. |
08 : - P -

Sk/k ()

65 66 &7 68 69 70 71 72
MASS OF 233U IN FUEL LOOP (kg)

Fig. 7. Effect of 227U Mass on Reactivity.

The 235U concentration coefficient of reactivity is given by the
ratio of the change in reactivity to the fractional change in 2350 con-
centration, or circulating mass, as a result of a small addition. This
is the slope of the curves in Fig. 7 at any particular concentration,
multiplied by that concentration. The value of (&k/k)/(3m/m) obtained
from the experimental curves was 0.22%, which was very nearly independent
of 2357 mass over the range shown in Fig. 7. The theoretically calcu-
lated value of this quantity was 0.248 for the approach to the initial
critical mass, and 0.234 for the average during the excess uranium

additions.

3.5 Rod—Shadowing BExperiments

During the course of additions of enriching capsules, three separate
experiments were performed in which the change in the critical position
of the regulating rod (rod No. 1) was recorded as the shim rods (rods
Nos. 2 and 3) were inserted into the core, These experiments were per-
formed with the pump running. They included observations of both the
effect of inserting a single shim rod (rod No. 2) with rod No. 3 held
fixed in the fully withdrawn position, and the effect of inserting rods

20

2 and 3 as a bank, that is, with their tips at identical elevations.
Figure 8 shows the data obtained from these experiments. Each experiment
was terminated at the 45-deg line, where the tips of all three control
rods are at equal insertions.

Some useful information concerning the reactivity worths of various
shim and regulating-rod combinations can be obtained from the results
shown in Fig. 8. To the critical position of the regulating rod at the
start of each experiment with a given 225U loading, there corresponds an
equivalent excess reactivity, relative to the reference conditions. This
reactivity may be determined from the curves in Fig. 6. Then, since
each curve in Fig. 8 represents control rod positions at conditions of
criticality, 1t follows that the curve pairs corresponding to a given
2357 loading represent various shim-regulating-rod combinations which are
equal in excess reactivity worth. In addition, these curves can be used
to determine the reactivity worth corresponding to full insertion of the
banks of two and three control rods, if use 1s made of an approximate
device. If we assume that the shape of the banked worth curve is suf-

ficiently close to that for the single regulating rod, shown in Fig. 6,

ORNL-DWG 65-10655

8
| | [ I | ! I I i ! I
®* ROD 3 FIXED AT 51 in. ROD 1 COMPENSATING FOR INSERTION

12} - OF ROD 2
5 a RODS 2 AND 3 AT EQUAL POSITIONS.
5 © ROD 1 COMPENSATING FOR : o R
£ INSERTION OF BOTH SHIM ////
o 20 b—— RODS — — —
[}
£=
2 /
24— - —— - - —_t —]

L ]
[ ]
828 —— — af ——.le . .
o el -
% 32 / I. A A l.
o) .. A'A* .l
(@] A ) A [
w A ® 4} ®
= 36 [—t . Al | A LI —_
g » A . A L4
3 A ) A L
W ° A ° A >
O 40 —_— S - e — - —
= fs '3 A Y A L]
I®) A - A e A ..
- A [
('% 44 —1  aA-p f\fi.q Qc. . —
o Ne Aw A &
a he Ale o W
fa’ J 1}. Ale
48 - " - e e fe -
/ & 6794 kg  €9.94 kg se 7474 kg
@ 235 » 235 235

52 48 44 40 36 32 28 24 20 16 2 8 4 0]
CRITICAL POSITION OF ROD 1 (inches withdrawn)

Change in Critical Position of Rod 1 as Shim Rods 2 and 3 are Inserted
info Core.

Fig. 8. Change in Critical Position of Rod 1 as Shim Rods 2 and 3
Are Inserted inteo Core.

21

a simple ratio converts each of the three reactivity levels to the cor-
responding reactivity with the rod bank fully inserted in the core. {This
relative invariance of the shape of the worth vs position curves is sup-
ported by theoretical calculations.) The conversion of the reactivity
measurements in the manner described above is summarized in Table 3.

In addition to the experiments described above, at each of the three
2357 levels, an experimental check was made to determine if there was any
asymmetry in the control-rod worths, depending on the rod designated as
the regulating rocd. The configuraticon of control rods and graphite sample
holder is shown in Fig. 9. ©Since the three control rods are of identical

design, they could only differ in relative poisoning effect by virtue of

OHNL WG (4-8414

TYPICAL FUEL PASSAGE

NOTE: STRINGERS NOS. 7, 60 AND &1 (FIVE) ARE
REMOVABLE.

¢ e < 59 4
N ' 2= CONTROL ROD
GUIDE TUBE
\g 60 ?iij)g /
g \\ //é GUIDE BAR
! -

REACTOR
CENTERLINE

THREE GRAPHITE AND
& = INOR-8 REMOVABLE
N SAMPLE BASKETS

e

REACTOR CENTERLINE

Fig. 9. Lattice Arrangement of MSRE Control Rocds and Sample Holder,

Table 3.

Worth of Control-Rod Banks Measured in Rod Shadowing Experiments

Experiment Rod Group

Total 235U Excess Re-

Banked Critical

Worth at Banked
Critical Position

Total Worth at

, i o ,
o (s RHP GSHORE s vionaram) OB atmaL T Terien
1 1-2 67.94 0.780 %6 .k 0.1903 4,099
2 1-2 69,94 1.460 28.8 0.3673 3,975
3 1-2 71.71 2.095 23,3 0.5142 4,075
Average 4,050
1 1-2-% 67.94 0.780 39.0 0.1394 5.5%
2 1-2-3% 69.94 1.460 3%, % 0.2602 5.611
3 1-2-% 71.71 2,095 28 .4 0.3761 5.570
Average 5.592

"Normalized to initial critical loading in loop (65.25 kg 235U), zero point of reactivity
with all three rods at 51 in,

cc
23

their position with respect to the graphite sample holder. At a given
235U loading and core temperature, the critical positions of each of the
three rods were measured and compared, with the other two rods held in
the fully withdrawn position. The amount of asymmetry in rod worths ob-
served in these experiments was negligible (<D.OO5% Sk/k).

3.6 Reactivity Effects of Fuel Circulation

The reactivity effect of circulation, given by the vertical dif-
ference between the two curves in Fig. 7, is 0.212 * 0.004% &k/k. Farly
in the theoretical studies of MSRE physics, the effect of the change in
delayed-neutron precursor distributions due to circulation had been pre-
dicted to be —0.30% &k/k.®? It was expected that another —0.2% &k/k
might be present due to entrained helium bubbles in the circulating salt.
As will be discussed later, the evidence indicates there were no circu-
lating gas bubbles except for a brief period when the fuel level in the
pump bowl was lowered. Therefore, the measured gas effect attending
circulation was practically nil,

Following these experiments, the discrepancy between the predicted
and observed delayed-neutron losses was subjected to study. Prior calcu-
lations of this effect had accounted for the subsequent contribution to
fission of only those delayed neutrons emitted in the graphite-moderated
(channeled) region of the reactor core. By extending this analysis to
include the contribution of the delayed neutrons emitted while the fuel
is in the upper and lower plenums, Just outside the graphite region, a
reactivity difference of 0.222% 8k/k was calculated from theoretical
analysis. The importance of including the upper and lower heads (par-
ticularly the former) appears to be a consequence of the relatively large
fuel volume fractions, or fluid residence times in these regions, as well
as the net displacement of the equilibrium precursor distributions in the
direction of the upper plenum, due to the fuel circulation.

This same theoretical model was also extended to include the case
when the flux is increasing on a stable asymptotic period.”® This results
in an inhour-type formula for the circulating condition, which must be

used in place of Eq. (3) to analyze the period measurements made with the

24

fuel pump running. The equations basic to this analysis are:

6 6
L By 09%, 0B8] = ) A (6%, e, ]
P = WA + == % == ’ (4)
dCi
BiP¢ — (Ki + W) Ci = VZ E;— , in reactor, (5)
dCi
- (%i + W) C; = VZ'E;— » in external piping, (6)

together with boundary conditions requiring that the precursor concen-
tration in the salt, Ci’ be continuous around the circulating path of the
fuel. The symbols not already defined in Egs. (2) and (3) are as follows:

P$¢ = local rate of production of prompt neutrons plus delayed
precursors (v X fissions per unit volume of fuel salt,
per unit time),

C. = concentration of precursors of ith group delayed neutrons,
per unit volume of fuel salt,

¢* = fast group adjoint flux, or importance,
@, = local volume fraction of fuel salt,

z = position variable in the direction of fuel circulation,
VZ = local velocity in direction of circulation,

[g,h] = symbol for the scalar product of functions g and h,
representing their multiplication and integration over
the spatial and energy variables of the neutron
population in the reactor.

The results of applying this analysis to the observed period-rod
sensitivity measurements made during fuel circulation are shown in Fig,
10. In this figure, the solid curve has been reproduced from Fig. 5,
in order to provide a direct basis for comparison of rod sensitivities
megsured with the fuel stationary and with it cireulating. The solid
curve does indeed give a fair representation of the latter measurements.
However, the problem of maintaining adequate precision in the preriod

measurements, also present to some extent in the data shown in Fig. 5,

25

ORNL-DWG 65-10657
0.07

I
! :
006 | - : s

| |
0.0% —‘» /Oa i}—

o
0.04 %/ —
003 | - - ——+ ' - :
FUEL STATIONARY
o FUEL CIRCULATING
0.02 e P | =

I
i [ '
Q | J i L i L L
0 4 8 2 16 20 24 28 32 36
ROD POSITION (in. withdrawn)

DIFFERENTIAL WORTH [(% 8%/ )/in ]

Fig. 10. Differential Worth of Control Rod 1, Measured with Fuel
Circulating. (Normalized to initial critical “2°U loading.)

is magnified for the measurements made with the fuel circulating. OSug-
gestions concerning the improvement of the precision of these measure-
ments are included in ref. 9. However, the agreement between the calcu-
lated and observed value of the gross reactivity loss due to circulation,
together with the stable-period measurement results shown in Fig. 10,
provides considerable support for the model used to calculate the trans-

port of delayed-neutron precursors due to fuel circulation.

3.7 Rod-Drop Experiments

The rod-drop experiment consists of the intentional scram of a rod,
or rod group, from an initially critical configuration and the recording
of the decay of the neutron flux as a function of time following the
scram., One then determines the amount of negative reactivity required
to produce this flux decay trajectory. The trajectory is characterized
by a sharp drop in the flux immediately following the scram, which cor-
responds closely with the actual fall of the rod. The curve rapidly and
continuously evolves into one with a much slower rate of flux decrease,
governed by the decay of the initial distribution of delayed-neutron
precursors.

Due to the requirement imposed by this experiment of accurately

recording the flux while it 1s rapidly falling by about two decades, the

26

graphical record obtained from the trace of a log n recorder is of
limited usefulness. The combined requirements of fast response, good
counting statistics, and reproducibility can be served, however, by re-
cording the integrated count as a function of time following the drop.

In order that the integrated flux-time curve could be recorded with-
out requiring scalers with special, accurately timed cycles for count
accumulation and recording, three scalers, together with a photographic
technique of rapid recording, were employed in the following way. One
of the scalers was driven by one of the fission-chamber channels. The
other two were operated as 60 cycles/sec timing devices. One of these
timers was synchronized to start with the scram of the rod, while the
other timer and the count-accumilating scaler were synchronized and
started a few seconds before the scram was initiated. The three scalers
were stacked together in a vertical array, and a rapid-action camera was
used to photograph the records on the three scalers simultaneously approxi-
mately once every second, starting a few seconds before the scram and
ending about 30 sec after the scram. From these photographs, the count
rate at the initial critical condition, the time of the scram signal,
and the integral of the flux, or count rate, as a function of time
following the scram could be accurately determined. In these experiments,
the position of the fission chambers was adjusted to obtain an initial
count rate at criticality of ~30,000 counts/sec, which was low enough
to result in quite small counting-rate losses due to dead-time. Also,
the remote position of the MSRE instrument shaft, relative to the
reactor core (see Fig. 1), would be expected to minimize errors due to
changes in the spatial flux shape during the rod-drop experiment. The
experiments described here were performed with the fuel stationary.

The method used to analyze the experiments was that of comparison
with theoretical flux decay curves, These latter curves corresponded to

a negative reactivity insertion with magnitude determined from the in-

tegral worth vs position curves already obtained from the period measure-
ments and rod-shadowing experiments described in the preceding sections.
In this way, the results of the rod-drop experiments were used as a

check for consistency with the rod calibrations obtained from these

former measurements. The rate of reactivity insertion corresponded to
27

the rate of the fall of the rod, or rod group, from the initial critical
positions.
The theoretical time-integrated flux decay curves were calculated

using the MATEXP program'

O to integrate the standard space-independent
reactor kinetics equations. ©Since this program is designed to integrate
a general system of first-order differential equations, the theoretical

time-integrated flux following the scram could be obtained by solving

the system:
dy(t)
— n('t,) ) (7)
dt
an(t) 2p(t) — B % :
ac, (t) p,n(t)
= — AC.(t) + , 1=1, 2, eaay 6, (9)
dt
where

y(t) = time-integrated count rate following the scram,
n(t) = detector count rate following scram,

Ci(t) = delayed-neutron precursor concentrations, normalized to
detector count rate,

Lo(t) = reactivity addition vs time following scram,

Bi = production fraction for ith group of delayed neutrons,
Ai = radioactive decay constant for ith group precursors,
A = prompt neutron generation time,
6
B=) B
i=1

The initial conditions for performing the integrations of Eas. (7), (8),

and (9) are:

y(0) =0, (10)

n(0) = initial count rate at critical condition , (11)

28

p,;n(0)
ci(o) =—, (12)

The time variation of the reactivity is governed implicity by the

equations;
2o(t) = p(y, ) — ply(t) ], (13)
= <t <
y(t) = v, 0<t<st,
=y-—lat2 . St<t ++%
2 e m
= <
Y b+t St (1k)
where

y = rod position, inches withdrawn (0 < y < 51 in.); initial
position, yo; scram position, Ygr

po(y) = magnitude of reactivity worth corresponding to rod position
¥, normalized to zero reactivity at fully withdrawn position,

t = effective lag time between scram signal and start of rod
drop (~20 msec),

a = acceleration during fall (~15 ft/sec2)¥,
tm = time to fall to scram position.

As described previously, the value of p(y) used in this analysis
was determined by integration of the period differential-worth measure-
ments. We have also used an algebraic representation of these experi-
mental curves, obtained by a least-squares curve-fitting procedure.l?l

In applying the above analysis to the experiments in which rod
groups were dropped, the magnitude of the total negative reactivity in-
sertion was determined by combining the calibration curve for the single
rod with the results of rod-shadowing experiments, in the manner already
described in Sec. 3.5, Typical ganged rod-drop experiments consisted
of the simultaneous scram of the regulating rod from its initial critical
position and one or both of the shim rods from the fully withdrawn posi-

tion. The normalized shape of the reactivity insertion curve for the

*Based on drop-time measurements on the MSRE control rods made
during pre-nuclear testing.

29

rod group was approximated by that corresponding to a single rod, falling
from the fully withdrawn position. The results of the rod-drop experi-
ments made after additions of 30, 65, and 87 capsules of enriching salt
are shown in Figs. 11, 12, and 13 respectively. In all three sets of
experiments, the analysis of the single rod drops shows very good con-
sistency with the rod-worth calibratiofi obtained by integration of the
differential-worth measurements. A similar consistency was obtained
from the experiments involving the scram of all three rods. In the case
of the two-rod experiments, results obtained after 65 and 87 capsules
appear to be slightly anomalous with respect to the experiment with the
same rod group after 30 capsules. Multiplication of the magnitude of the
negative reactivity insertion by a factor of about 1.05 brings the calcu-
lated and experimental curves for these experiments into better agree-
ment. However, another source of this discrepancy could be in the ap-
proximation of the shape of reactivity insertion vs time for the tanden
fall of the rods by that corresponding to a single rod, This source

would be expected to have its greatest influence on the two-rod experiments.

ORNL-DWG 66-11445

\ | ]

c e & EXPERIMENTAL POINTS

CALCULATED BY INTEGRATION OF REACTOR

10— KINETICS EQUATION; REACTIVITY DETERMINED
FROM INDEPENDENT CALIBRATION MEASUREMENTS.

/
Q<
0,
ROD NO4 SCRAMMED

RODS NO.1 AND 2 SCRAMMED

.

ROBS 1,2,AND 3 SCRAMMED

s

INTEGRAL COUNT
o

10 15 20
TIME AFTER SCRAM (sec)

(@)
[&)]

Fig. 11. Results of Rod-Drop Experiments After 30 Capsule Additions.

ORNL—DWG 66-11447

(x10%)
14 |
i
4 ORNL— DWG 66 —t1446
(x107) T T |
o,8,4 EXPERIMENTAL POINTS / ROD NO.! SCRAMMED
12 F CALCULATED BY INTEGRATION OF 12
REACTOR KINETICS EQUATIONS;
REACTIVITY DETERMINED
FROM INDEPENDENT CAL-
IBRATION MEASUREMENTS,
10 i ; 10
| . ROD NO.1 SCRAMMED
' : i -
! =z
; =]
: o
; /
o 8 i i
) = ‘ '
2 = RODS 1 AND 2 SCRAMMED/
g ‘
t'_(a 6 RODS NO.1 AND 2 SCRAMMED —mme=# 6 ]
Z / /
4 A 4 ‘/10 //
: = 1 RODS 1,2, AND 3
RODS NO. 1.2, AND 3 e |
SCRAMMED /./ o,8,a EXPERIMENTAL POINTS
2 | 2 s L ——— CALCULATED BY INTEGRATION OF ~——rf
s ‘ REACTOR KINETICS EQUATIONS;
‘ REACTIVITY DETERMINED FROM
; INDEPENDENT CALIBRATION
| i MEASUREMENTS,
| | | j |
0 . i 1 0 ‘
0 5 10 15 20 0 5 10 15 20
TIME AFTER SCRAM (sec) TIME AFTER SCRAM ({sec)
Fig. 12. Results of Red-Drop Experiments Fig. 13. Results of Rod-Drop Experiments
After 65 Capsule Additions. After 87 Capsule Additions.

0t
31

The sensitivity of these experiments to variations in the magnitude
of the reactivity insertion i1s illustrated in Fig. 14. Here the analysis
of the experiment where rod No. 1 was scrammed after addition of 30
capsules 1s reproduced from Fig. 11; the theoretical curves are added
which correspond to an increase and a decrease of 5% of the total magni-
tude of negative reactivity insertion. For this particular experiment,
this corresponds to an increment of 0.007% 6k/k in the magnitude of re-
activity. With the largest apparent differences occurring in the experi-
ments involving the scram of two rods, the results of all these experi-
ments were within the 5% band of self-consistency with the rod cali-

bration results obtained from the differential-worth measurements.

ORNL-DWG 66-11448

! -
o EXPERIMENTAL POINTS g 2
=== CALCULATED BY INTEGRATION GOF g B
REACTOR KINETICS EQUATIONS, \
gl — - :
-
s
2 6 — :
jog .
(&) ./ |
= z
<{ 4
: v
/,
- ROD NO.1 SCRAMMED AFTER 30 CAPSULE
ADDITICNS.
: NEGATIVE REACTIVITY INSERTION, Ap, DETER-
MINED FROM INTEGRATION OF DIFFERENTIAL
WORTH MEASUREMENTS.
: NEGATIVE REACTIVITY INSERTION,1035 A p.
- NEGATIVE REACTIVITY INSERTION,095 Ap.
!
1
1

10 s 20
TIME AFTER SCRAM (sec)

FPig. 14. Sensitivity of Rod-Drop Experiment to Changes in Magnitude
of Reactivity Insertion.

3.8 Comparison of Measurements with Theoretical
Analysis of Control~Rod Worth
In the period before test operations were begun on the MSRE, de-
tailed theoretical studies were made in order to predict the neutronic
characterisitics of the reactor, as closely as practical. These studies
first focused on such characteristics as the expected critical 225U con-

centration, the control-rod reactivity worth, and the various reactivity

32

coefficients, In these studies, use was made of a synthesis of results,
ranging from "zero-dimensional multigroup calculations of the average
neutron spectrum in the central region of the reactor core to few-group
diffusion calculations with two-dimensional mcdels made to represent the
actual MSRE core geometry.®

The programs which were used to calculate the expected neutron
energy spectrum over most of the reactor core were the GAM-IT and the
THERMOS codes.1®s13 The former calculates the epithermal energy spectrum
of the neutron flux (the flux above energies where thermal motion of the
moderator nuclei exerts a strong influence on the neutron spectrum). The
latter calculates the spectrum in this thermal energy region, together
with its dependence on position within the individual cells comprising
the lattice of fuel salt and graphite. In each of these programs, the
actual energy spectrum is approximated by a finite number of energy groups
(99 groups in GAM-II spanning the interval 15 Mev to 0.41k ev, and 30
groups in THERMOS, spanning the interval 0 to 0.876 ev). The programs
intrinsically utilize library tapes of neutron cross sections for each
nuclide and energy group, which represent the best current evaluation of
experimental cross-section data.

In the GAM-II model, the calculated flux spectrum corresponds to
that in a single-region reactor, consisting of a homogeneous mixture of
graphite and fuel salt, with corrections introduced for the fuel-channel
geometric self-shielding of the absorption resonances in 222U. The GAM-II
and THERMOS models sacrifice consideration of the gross variation in the
core geometry and material composition (such as correspond to the control
rod thimbles or the upper and lower salt plenums) and their effect on
the neutron flux distribution in order to gain detail in describing the

energy spectrum in the most important region of the core. The gross

*Some form of synthesis of calculations from various theoretical
models, each of which is used to examine particular details of the core
neutronics, 1s by now fairly standard in reactor physics studies. We
recognize that the rapid development and automation of these technigues
usually quickly outdates the description of any particular method of
synthesis., Nevertheless, we feel that a brief summary of the basis of
the theoretical calculations should add perspective to this report, as
a record of our experience with the MSRE.

33

spatial variation of the neutron flux was studied with the aid of one-
and two-dimensional group-diffusion models. Slightly simplified versions
of these latter models are shown in Fig., 15. These represent radial (R),
radial-axial (R-7), and radial-angular (R-6) models of the actual core.
The process of synthesis was one of first determining the detailed neutron
spectrum in the graphite-salt lattice with the GAM-II and THERMOS programs.
Individual nuclide cross sections were then averaged over this spectrum
to obtain "broad group" cross sections, and these were applied in group
diffusion calculations for the models shown in Fig. 15.

The names for the group-diffusion codes used for these calcula-

tionsl%4215,16 gre given under each model shown in Fig. 15. The predicted

ORNL-DWG 65-10653

(g} ONE DIMENSIONAL MODEL -
MULTIGROUP DIFFUSION CALCULATIONS
IN RADIAL GEOMETRY (MODRIC)

(&) TWO DIMENSIONAL MODEL~
TWO GROUP DIFFUSION
CALCULATIONS IN R-Z
GEOMETRY (EQUIPOISE)

(¢) TWO DIMENSIONAL MODEL- FOUR
GROUP DIFFUSION CALCULATIONS IN
R-8 GEOMETRY (EXTERMINATOR)

Fig. 15. Geometric Models of MSRE Core Used for Nuclear Calcula-
tions.

34

critical concentration reported in ref. 2 was based on the one-dimensional
(radial) MODRIC calculations. Calculations ranging from 33 to 4 broad
groups showed very small differences in the predicted critical 22U con-
centration. However, calculations with the radial model sacrifice de-
tail concerning the variation in core composition in the axial dimension,
by assuming a uniform rate of axial neutron leakage (characterized by

an approximate transverse geometric buckling). In addition, this model
requires that the control-rod group be represented as an annular region.
The first of these approximations was checked by two-group, two-
dimensional calculations with the R-Z model shown in Fig. 15b, using

the 27°U concentration found to be critical in the MODRIC calculations
Later, when the EXTERMINATOR program became available, we checked the
second approximation by employing the R-9 model of Fig. 15c. In this
calculation, use was still made of the transverse leakage approximation,
but now we were permitted a much closer representation of the actual
geometry of the three control rods and the graphite sample holder (des-
ignated by S in Fig. 15¢). We found that, for the same 235U concen-
tration, these two corrections in the calculated multiplication constant
very nearly cancelled one another., (In the R-Z model, ke was 1.0106
with an annular poison region inserted to an elevation corresponding to
the position of maximum rod withdrawal. In the R-0 model, k_ was 0.9903. )
Therefore, we concluded that no corrections to the MODRIC predictions of
the critical concentration were necessary.

The theoretical calculations of the control-rod worth were based on
the R-6 model of Fig. 15c. When the gadolinium poison rods are inserted
in the core, diffusion theory cannot be applied routinely within the
indicated rod regions. Approximate blackness coefficients at the region
boundaries (fraction of neutrons incident on the rod thimbles which are
ultimately absorbed in the control region) were obtained from a simpli-
fied transport-theory calculation for the control region geometry. These
boundary conditions were then applied in the R-0 group-diffusion calcu-
lations. Table L4 summarizes the results of these calculations. The
change in the effective multiplication constant between the conditions

of rods completely withdrawn from the core and rods inserted all the way

Table L.

Comparison of Theoretical Calculations of Control
Rod Effects with Experimental Measurements

Control Rods

a . .
Uranium Concentration Inserted Ske 6ps 6ps (51 in.) 8pexp (51 in.)
(Rod Nos.) (% &/k) (% &/k) (% 8k/x)
Initial critical
concentration o 0.0218 2,278 2,111 2.26
1-2-% 0.0546 5,894 5.46% 5¢59
1.1 X initial critical
concentration 2 0.0211 2.094 1.041 2.08
a

through a cylindrical core with an effective nuclear height of 78.2 in.

&k = k (0) — k. (1); (0) = rods completely removed, (1) = rods completely inserted

Ge

36

through the core are tabulated for a single rod and for the bank of all
three rods. For the single rod, calculations were made both for a
uranium concentration very nearly equal to the initial critical concen-
tration (at the beginning of the excess uranium additions and rod-
calibration experiments) and for a 23°U concentration increased by 10%.
This latter value corresponds very closely to the uranium concentration
at the termination of the zero-power experiments. Table 4, column k4,
also 1lists the change in static reactivity, 605, due to the rod insertions,
as determined from Eq. (2). As was discussed in Sec. 3.2, this quantity
should be used as a consistent basis of comparison of calculations with
the experimental measurements of reactivity.* To make this comparison,
however, the calculated values of Sps given in colum 4 of Table 4 must
be reduced by a factor to account for the available rod travel (51 in.)
relative to the calculated effectiveness of the rods inserted all the
way through the core. For this purpose, approximate calculations of the
worth of the partially inserted, banked rods were made using the R-7
model shown in Fig. 15b. These results indicated that the available
travel of 51 in. covered 0.927 of the worth of rods inserted all the way
through the core. (The effective, or "nuclear" height of the core, ex-
tending into the upper and lower heads, was 78.2 in.) The values of the
static reactivity worth of the rods corrected by this factor, together
with the measured values of the rod worths, are given in the last two
columns of Table L.

Farly studies” had indicated that the worth of rod Nos. 1 and
3 might be about 4% higher than that for rod No. 2, due to the positions
of these rods with respect to the graphite sample holder (see Fig. 9).
In these calculations, however, the difference between the composition
of the sample holder and the main core region was neglected. The later
calculations made with the R-6 model included a composition of graphite,
fuel salt, and INOR-8 specimens very nearly equal to that of the actual
sample holder. In this case, no asymmetry in the worth of the three

*For small changes in reactivity about the critical condition, the
values of 8ke and dpg are very nearly equal. For changes with magnitude
of the order of the control-rod worth, this is no longer a good approxi-
mation, however.

37

rods was found to occur, primarily because the sample holder tends to
cause a depression in the thermal flux which is about equal to that

caused by the empty INOR-8 control element thimbles. As described earlier,
this conclusion is further supported by experiment, since negligible

asymmetry was observed in the calibration experiments.

4, TEMPERATURE AND PRESSURE REACTIVITY EFFECTS

L.,1 TIsothermal Temperature Coefficient of Reactivity

The effect of temperature on reactivity was measured in three sep-
arate experiments in which we adjusted the electric heaters to change
the system temperature slowly (about l5°F/hr) while we observed the
critical position of the regulating rod. This experiment gave the over-
all temperature coefficient, that is, the sum of fuel and graphite
coefficients. These experiments were performed with the fuel circulating.

Figure 16 shows the results of the three experiments involving slow
changes in temperature. The changes in critical position of the regu-
lating rod was converted to reactivity by use of the rod calibration
results described in the preceding sections. Here, the magnitude of the
reactivity change about the nominal 1200°F operating point is plotted as
a function of the fuel, or system temperature (temperature changes were
induced so slowly that the fuel and graphite could be safely assumed to
remain in thermal equilibrium). This temperature was the average of a
preselected set of thermocouples used throughout the nuclear experiments.
Since the slopes of the reactivity curves in Fig. 16 are the quantities
of interest, the positions of the curves have been translated parallel
to the vertical axis in order that the three curves might be plotted on
the same scale and still be distinguished. The Tirst experiment, with
67.9 kg of 235U in the loop gave a line whose slope ranges from —(6.6 to
8.%) x 107° °F 1, At 69.9 kg 275U, the experiment gave a straight line
with a slope of —7.24 X 107° °F~1. 1In the last experiment, at 71.7 kg
235, the slope of the curve above about 11LO°F is —7.3 X 107> °F~t.
38

ORNL-DWG 65-8032R

1.8
K4
1.6 — - &
e
67.86 kg 235U IN LOOP ..¢*
14— [ — - - — 1
,.’
Il/.
12 — - - o - —
l'.
£ 10 [— ‘ A e
- < e
< 1 -
z o8 ——- — — — _7! — ,/_. P—
o *
71.71kg 235U IN Locgz,- e
0.6 {— : P = - —
./. ,',.
o ~"69.85 kg 23%U IN LOOP
o4l | - I N s
-~ |
o :
-1 :
0.2 |— e | -
o i
|
o |
1000 1050 1100 150 1200 1250

FUEL TEMPERATURE (°F)

Fig. 16. Effect of Slow Changes in Core Temperature on the Reac-
tivity.

Calculations consistent with the models described in the preceding
section gave a value of —8.1 X 107> °F~! for the total isothermal temper-
ature coefficient of reactivity. The associated components for the fuel
salt and graphite were —4.1 X 107> °F1 and —4.0 X 107> °F~1, respec-
tively. Since the fuel coefficient is very nearly proportional to the
value of the coefficient of thermal expansion of the salt, it is necessary
to qualify the calculated ccefficients with this value. These calcu-
lations were based on an expansion coefficient of —1.18 x 107% °pF~1,
obtained from an early empirical correlation of temperature with density.
This value was later found to be in good agreement with the expansion
coefficient determined from observations of the change in salt level in
the MSRE fuel pump bowl with loop temperature (two separate measurements
gave —1.09 X 107* and —1.15 X 10™% °p~1),3

The experiment at 71.7 kg °°U shows a lower slope below about 1140°F.
We do not believe that the temperature coefficient is lower in this range
but that another phenomenon became significant during this part of the

experiment, That is the appearance of an increasing amount of helium

39

bubbles in the circulating salt as the temperature was lowered. The evi-
dence for this is discussed in the section on pressure effects. The
effect, so far as the temperature experiment is concerned, was that the
bubbles tended to reduce the amount of fuel salt in the core, compensating
to some extent for the increase in density of the salt itself as the tem-
perature was lowered. Thus the slope of the lower part of the curve can-
not be interpreted as a temperature coefficient of reactivity in the usual
sense,

The influence of the core temperature on reactivity, as reflected in
the change in critical position of the regulating rod, is also illustrated
in Fig. 17. This is a photograph of a three-dimensional model™ showing
the critical position of the regulating rod (vertical axis) as a function
of the 275U concentration (horizontal axis, nearly in the plane of the
page), and core temperature (depth axis, front-to-back of the model). The
measured critical position of the regulating rod with fuel circulating at
1200°F vs the 2°°U concentration is represented by the relatively "dense"
curve sloping downward to the right. FEach point represents the addition
of one capsule of enriching salt. The critical position of the regulating
rod with circulation stopped, measured after every fourth capsule, is the
curve with the "sparse" points lying vertically below and in the same
plane as the first curve., The separation between the two curves repre-
sents the increment of rod insertion required to compensate for the loss
in delayed neutrons due to circulation.

The three experiments in which the core temperature was varied are
shown in Fig. 17 as the segments sloping upward from the front-to-back
of the model and crossing the upper curve of critical position vs 22°U
concentration at 1200°F. The points at the extreme upper left of the
model represent the data taken at the time of the initial critical

experiment (when the temperature was ~1180°F and the rods were poisoning

0.08% &8k/k).

*This model was constructed by J. A. Watts, as a visual aid for
demonstrating to visitors the important parameters influencing the
neutronic behavior of the reactor. The position of the beads in the
model corresponds as closely as possible to the actual data points
obtained during the rod calibration experiments.
Fig. 17. Photograph of a
Measurement Data.

Three=Dimensional

Plot of the Reactivity

Photo 86401

O%
41

4,2 Fuel Temperature Coefficient of Reactivity

An attempt was made to separate the fuel (rapid) and graphite (slug-
gish) temperature coefficients by an experiment in which the coolant
system was used to increase the fuel salt temperature rather abruptly.
This was done by stopping the fuel pump, raising the temperature of the
circulating cocolant salt and the stagnant fuel in the heat exchanger,
then restarting the fuel pump to pass the hotter fuel salt through the
core. The reactor power was controlled at 10 w by the flux servo.® TFor
these experiments, use was made of a Bunker-Ramo 340 on-line digital
computer and data logger, installed as part of the MSRE instrumentation.**
The output of a thermocouple in the reactor vessel outlet, logged digit-
ally at l/h-sec intervals, showed a brief increase of 5 to 6°F as the
hot salt first passed. It then leveled at about 3.5°F for a few loop
transit times before decreasing gradually. The noise in the analog-to-
digital conversion (+1°F) limited the accuracy of the measurement, but
by taking an average of 50 points during the level period after mixing
and before the graphite temperature had time to change significantly, a
value was obtained for the increment in fuel temperature. The reactivity
change was obtained from the change in rod position using the rod cali-
bration results, corrected for the decrease due to circulation, and
ascribed to the fuel-temperature increase. The resulting coefficient was
—(4.9 + 2,%) X 1075 °F~1, Predicted values of the fuel-temperature coef-
ficient lie in this range.

This experiment was later repeated in a slightly different mamner,
with the intent of improving upon the first measurement. The reactor
inlet and outlet temperature thermocouple signals were preamplified and

filtered before digitizing, reducing the noise to about +0.1°F. This

*For rapid changes in reactivity, small perturbations in the flux
do occur, even when the regulating rod is flux servo controlled. However,
calculated corrections for these effects were found to be insignificant
for these experiments.

**At the time of the zero-power experiments, the logger-computer was
still being debugged and was frequently unavailable. Thus, only the con-
ventional recording instrumentation was employed for most of the rod
calibration experiments described in the preceding sections.
42

time the fuel was kept in circulation and the coolant loop was stagnant.
After heating the stagnant coolant salt about 20°F hotter than the fuel
salt, the coolant pump was started, introducing a hot slug of fuel into
the heat exchanger and subsequently into the core., 1In this test, then,
the change in reactivity was due entirely to the change in temperature,
and the rod-induced reactivity required to keep the reactor critical was
approximately equal and opposite to that due to the fuel and graphite
temperature change., If the immediate effects could be attributed to the
fuel temperature coefficient alone, a value of —(L4.7 + 0.7) x 1072 °pF~1
was obtained from this test. However, noise effects relative to the
temperature changes involved and also special problems due to temperature
measurement lag effects still prevented us from obtaining a good assess-

ment of the uncertainty in this measurement.

4.3 Effect of Pressure on Reactivity

The possibility exists for a pressure coefficient of reactivity in
the MSRE, because undissolved helium can be entrained in the circulating
fuel through the action of the fission product gas stripping device. A
small fraction of the fuel pump discharge stream (about 50 gpm out of
1250 gpm) is diverted into a spray ring in the gas space in the pump bowl.
The purpose of the spray, or stripper, is to provide contact so that 13°Xe
in the salt can escape into the gas space, which is continuously purged.
balt jetting from holes in the spray ring impinges on the surface of the
liquid pool in the pump bowl with sufficient velocity to carry under
considerable guantities of gas, and a small fraction of the submerged
bubbles can be swept through the ports at the pump suction into the main
circulating stream of fuel, A steady state is reached when the helium
concentration in the main stream has increased to the point where loss
of helium through the stripper flow equals the rate of injection.

At steady state, the volume fraction of gas in the circulating
stream varies around the loop proportionally to the inverse of the local
pressure, which changes with elevation, velocity, and head losses. If,
after steady state is established, the loop pressure is changed rapidly,

the mass ratio of undissolved gas to liquid would be expected to remain

43

practically constant and the volume fraction of gas in the loop would
decrease with increasing pressure. This would give rise to a positive
pressure coefficient of reactivity. On the other hand, for very slow
increases in pressure, the volume fraction of gas at the pump suction
would tend to remain constant, and the volume of gas in the core would
increase, because the ratio of absolute pressures between the core and
purp suction is reduced. These slow changes could give rise to a small
negative pressure coefficient of reactivity.

We performed three tests to explore the effect on reactivity of
changing the system overpressure. In each of the three tests the loop
overpressure was slowly increased from the normal 5 psig to 15 psig and
then guickly relieved, through a bypass valve, to a drain tank that had
previously been vented to atmospheric pressure.

The first two tests were carried out at normal system temperature
with the normal operating level of salt in the fuel-pump bowl., No
change in control-rod position was required to maintain criticality and
no significant change in pump-bowl level was observed during either of
the tests. These indicated that the pressure coefficient was negligibly
small and that essentially no helium bubbles were circulating with the
salt. Further evidence of the lack of circulating voids was obtained
from a gamma-ray densitometer on the reactor inlet line; this instrument
showed no change in mean salt density during the tests.™

The third test was performed at an abnormally low pump-bowl level
which was obtained by lowering the operating temperature to 1050°F.
Pigure 18 shows the pressure transient and the responses of the regu-
lating control rod, densitometer, and fuel-pump level during the rapid
pressure release. For these conditions, a change of one unit on the
densitometer was equivalent to a change of about 0.15 vol % in the
circulating void fraction. Independent evaluations of the void fraction
from these three parameters all gave values between 2% and 3% by volume,

The frequency response characteristics of the effects of pressure on

*The use of this instrument is described in more detail in ref. 18,
The reported sensitivity of the measurement is 0.076 vol %.

4,

ORNL-DWG 65-8074R
65

I — : _ B, | - —. — - — —
55 S S A — - i — ]

FUEL PUMP LEVEL (%)

OVERPRESSURE (psig)

DENSITOMETER

CONTROL ROD POSITION (in.)

2€

TIME (min)

Fig. 18. Conditions During Rapid Pressure Release While Circulating
Hellum Bubbles.

reactivity were calculated, using Samulon's method,l® from the pressure
and rod motion. These results are shown in Fig. 19 along with the pre-
dicted high-frequency response for a void fraction of 1.2%. Extrapo-
lation to the observed curve gives a void fraction of 2% to 2 1/2%. The
low- and high-frequency pressure coefficients were +0.0003 and +0.01lL

(% Sk/k)/psi, respectively, for this particular condition.

ORNL-DWG 65-8304A

07— HER - 1T — SRR = 1 =1 e
- l% Fi = “j; I __T mel 1 gia| 't“l -1
- o 1 O O N ‘ [ A O 0 I

| i i _—t Jr« ] L __T fl»,, - ‘ Ll

e ’ T Li_ fi" \‘I J;» i l 7\l V } T . ./.Ji_L‘L- V 7fi7i ] h
I A A % ety

O e e R = TR | e T e

TG S 5 511 Y o 17,/’ T HIGH-FREQUENCY VALUE LA L1
I e e ““@J_[j' | ’iL+ Cfififififlfifilfg>/ 1]
EL& S 1T “L“
o S H LE' *:J jEs /"/E e == = _fi{
Sl i S S 54 S UD s - .+ S JN B £ S s a0 A S St i ey v

- - % - - I ] URE R L
T fil/% SEALH I M a1 B 11 A H'*

L o co . s

1444 [ — s A .t L E— et

il - i |

o i T I
0.0001 0.004 Q.01 oA { 10

w. FREQUENCY (radians/min}

Fig. 19. Reactivity-Pressure Frequency Response with 2 to 3% Void
Volume in Circulating Fuel. (Calculated from pressure release experiment
using Samulon's method with 0.2-min sampling interval.)

5, DYNAMICS TESTS

5.1 Purpose of Tests

We performed a variety of dynamic tests during the operation at
zero power. These tests were the start of an extensive program to
evaluate experimentally the inherent nuclear stability of the MSRE at
all power levels. The reactor system was analyzed on a theoretical
basis,®C and the tests were designed not only to characterize the present
system but also to evaluate the techniques and mathematical models used
in the theoretical analysis. Results from the analysis of the zero-
power tests are presented below. An extensive discussion of the sub-

sequent at-power tests is given in ref, 21.

5.2 Frequency-Response Measurements

A series of tests was run to determine the frequency response of
neutron level to reactivity perturbations. These experiments included
pulse tests, pseudorandom binary reactivity-perturbation tests, and

measurements of the inherent noise in the flux signal. Tests were run
46

with the fuel pump on and with it off. Noise measurements were also
made during a special run with a low pump-bowl level where there were
entrained bubbles in the core.

The frequency response is a convenient measure of the dynamic
characteristics of a reactor system. Classically the frequency response
1s obtained by disturbing the reactor with a sinusoidal reactivity per-
turbation, and observing the resulting sinusoidal neutron-level variations.
The magnitude ratio is defined as the ratio of the amplitude of the out-
put sinusoid to that of the input sinusoid. The phase angle is defined
as the phase difference between the output sinusoid and the input sinus-
oid. Other procedures, such as those described in this report, can be
used to yield the same results as the classical method but with less ex-
perimental effort.

The zero-power frequency response tests serve to check the theoret-
ical, zero-power frequency-response predictions, but they do not furnish
direct information on the stability of the power producing reactor. The
zero-power tests, however, do serve as an indirect, partial check on the
at-power predictions because the dynamic behavior at power is simply the
zero-povwer case with the addition of reactivity feedback from the system
resulting from power-induced temperature changes. Thus, the verification
of the zero-power kinetics predictions lends some support to the pre-

dictions regarding power operation.

5.3 Pulse Tests

In the pulse tests a control rod was withdrawn 1/2 in., held there
for 3.5 or 7 sec., then returned to its original position. The precise
rod positioning was done using a special analog computer timing circuit.
The rod was located such that 1/2-in. travel caused a change in re-
activity of about 0.03%. The rod position signal and flux signhal were
recorded digitally at 0.25 sec intervals using the BR-3L0 on-line digital
computer, The frequency-response characteristics are generally deter-
mined from pulse-response tests by obtaining, numerically, the Fourier
transforms of the input and output signals. One reguirement, however,

is that the Fourier integrals rmust be closed; that is, both input and

47

output signals must eventually return to their initial values. The zero-
power reactor, however, is an "integrating" system, since a temporary
reactivity perturbation can cause the flux to level out at a new value.
To circumvent this problem, the flux output signal was filtered with a
50-sec time-constant, high-pass digital filter., The output of this filter,
which eventually returns to its initial value, was then Fourier-transformed,
and the resulting transform corrected for the filter response.

"Practice" tests on a simulated zero-power reactor indicated that
reasonably accurate results could be expected from these tests only if
the fuel-temperature drift during a test could be held essentially to
zero, and if the control rod could be repositioned to its initial loca-
tion within an accuracy of better than 0.01 in.

The results of the pulse tests, shown in Figs. 20 and 21 for the
cases of the fuel stationary and circulating, are in adequate agreement
with the theoretical frequency response curves. This agreement provides
indirect evidence that the stringent accuracy requirements described above

were actually achieved in these tests.

5.4 Pseudorandom Binary Sequence Tests

The pseudorandom binary sequence (PRBS) is a specially selected
series of positive and negative rod movements which repeats itself after
a certain number of basic pulse or bit times. The principle advantage
of a PRBS in frequency response testing is that its frequency spectrum
typically consists of a large number of harmonics of approximately equal
size.® This means that the response may be analyzed at many frequencies
generated in a single test, and the signal-to-noise ratio at these known
frequencies is typically very good.

A PRBS is characterized by the number of bits in the sequence and
the bit duration., The bit time 1s defined as the mininum possible pulse
duration in the sequence. All pulses in a PRBS are minimm width or
integral rmultiples thereof. Numerous sequences may be generated, but
they are restricted to certain specific numbers of bits. For a sequence
of Z bits and a bit duration at At sec, the PRBS pericd is ZAt sec. The

lowest harmonic radian frequency W, and the spacing between harmonics Aw
48

ORNL-DWG 66-10708A

104 <
ZERO POWER
6 AN
)y 7 sec PULSE, TEST NO. 1 .
l N 7 sec PULSE, TEST NO. 2 A
> A 35 sec PULSE, TEST NO. 3 &
< 35 sec PULSE, TESTNO. 4 o
103 AN NOISE ANALYSIS v
\‘
6 h A NG
A s
R 2%
gn/n v ?
Sk/k, . ?}?
2 ST,
o
3 34 n 80 Yy ?-V.\i
102
i
6 !
R
2 A%
10'
0001 0.01 0.1 10 10 100
w, FREQUENCY (radians/sec)
ORNL-DWG 65-8909A
20
0
ik Ll =
2 l f'o g < AM”' Ny |
~ | T e
o - J ,,Jl,éf'L .
[+]
E o A ? opn a‘ - 4]
5 40 | et lpln) L
0 Pe !
pt - s A7 4«
a i , ZERO POWER
-60 ot A1 7 sec PULSE , TEST NO. | @ il
//’/ 7 sec PULSE , TEST NO. 2 4 | |||
, 1(/ 3.5 sec PULSE ,TEST NO. 3 &
L1
- 1 A= PULSE ,TEST NO. 4 O 11
80“ " :/r“ 3.5 sec PU ,TEST NO
I [
-100 l ‘ l t ‘ |
0.001 0.01 0.1 1.0 10

w, FREQUENCY ( radians /sec)

Fig. 20. Frequency Response of (&n/ng)/(8k/kg) at Zero Power; Fuel
stationary.

49

19 BIT PRBS—INDIRECT ANALYSIS @

1ot ORNL-DWG 66-10286
RN ZERO POWER
5 LAY 19 BIT PRBS DIRECT ANALYSIS @
19 BIT PRBS INDIRECT ANALYSIS &
\\\ 63 BIT PRBS DIRECT ANALYSIS &
2 - 63 BIT PRBS INDIRECT ANALYSIS ¥
5 q 7 sec. PULSE, TEST NO. | v
103 7 sec, PULSE, TEST NO. 2 ]
"Eui ~ NOISE ANALYSIS o]
5 A4+
SM“Q
Sk
102 ot
o
5 (9]
o\
2
10'
0.00t 0005 001 002 005 O1 02 05 10 20 50 100 200 500 1000
w, FREQUENCY (radions/sec)
ORNL—DWG &5-89074
20 i T T T T ! AR
ZERQ POWER f 11;! | i ;H}
10 + 19 BIT PRBS—DIRECT ANALYSIS A - r__“47T+YT! : ‘|yfi

° |63 BIT PRBS—DIRECT ANALYSIS ¥ | J | ‘l
—-10 163 BIT PRBS—INDIRECT ANALYSIS ¥ SR
—20 | 7 sec. PULSE, TEST NO. | o ‘\fifl

7 sec PULSE, TEST NO. 2 A N
-30 |— P T T T NG

t .
! ;
\ A
| - |

!

PHASE (deg}
& B
o O
B
fiA
R )

—-60 - dehe J7/<; . ?‘

a/ o | L o
70 — 1 oww — | wl
w0 | AT e
—g0 4=} —
{ i
-100

0.001 0.1 OR 1 10
w, FREQUENCY (rodians/sec)

l'ig. 2L. Frequency Response of (®&n/ng)/(dk/kg) at Zero Power; Fuel
Circulating.
are given by

) :Am:g—rf— (15)

o) yAA

Sequences of 19 and 63 bits were used in the zero-power tests with
bit times of 6.58 and 3.%5 sec, respectively. The 19-bit sequence was
generated manually, while the 63-bit test sequence, which was run after
the on-line computer was operational, was generated automatically by a
speclal shift-register algorithm. In both cases, analog computer timer
circuits were used to control the rod-drive motor insert and withdraw
"on" times.

The rod motion about the average position was adjusted to give a
reactivity perturbation of about +0.015%. As with the pulse tests, the
rod-position and neutron-level signals were recorded digitally at 0.25-
sec intervals using the MSRE computer. The frequency response was ob-
tained by two different methods. The direct method used a digital
filtering technique to obtain the power spectrum of the input and the
cross-power spectrum of the input and output. The frequency response at
some frequency is the ratio of the component of the cross-power spectrum
at that frequency to the component of the input power spectrum at that
frequency. The indirect method involved calculation of correlation
functions and subsequent numerical Fourier transformation. Both methods
gave essentially the same results. A description of the indirect method
is given in ref. 23, and the direct method is discussed in ref. 21.

The results of the PRBS tests for the circulating fuel case are
compared in Fig. 21 with the results from both the pulse-tests and the
predicted curves. The agreement is reasonably good, considering that
the stringent conditions of no temperature drift and extremely accurate
rod positioning, described in Sec. 5.3 for the pulse tests, had also to

be imposed on the PRBS tests.

5.5 Neutron Fluctuation Measurements

Flux-noise measurements were made by Roux and Fry=% at three dif-
ferent zero-power conditions: fuel-salt stationary, normal fuel-salt

circulation, and circulating fuel with entrained bubbles. The

51

fluctuations were recorded on an analog band-pass filter power-spectral-
density (PSD) analyzer. One run was also digitized and analyzed by auto-

correlation and digital filter techniques for comparison. All methods
gave essentially equivalent results. Results of the PSD analysis for
stationary and circulating fuel are shown in Figs., 20 and 21. These
points were obtained by taking the square root of the measured PSD after
subtracting the asymptotic high-frequency value of PSD due to noise in
the neutron detection process.=® This computation results in an ampli-
tude-ratio vs frequency curve if one assumes that the input reactivity
noise is white, that is, has a constant PSD in the frequency range of
interest. Since the magnitude of the input PSD is unknown, only the
relative magnitude of the noise results are available; thus the curves
were arbitrarily normalized to the theoretical results at ¢ radians/sec.
Based on the good correspondence of noise response and theoretical re-
sponse, the assumption of white noise input appears valid.

The results of the noise analysis of the case where bubbles are
entrained in the fuel salt showed a substantial increase in PSD in the
frequency range of 1 to 10 fadians/sec over the no-bubble case, indicating
an increase in the input reactivity fluctuation. Previous experiments
with the MSRE core hydraulic mockup indicated that random, hydraulically
induced pressure fluctuations in the core would probably cause a signif-
icant modulation of the core void fraction, thus causing reactivity
fluctuations., Hence, additional flux noise in this frequency range was
expected, although it was not possible to predict the "shape" or charac-

teristics of this spectrum.

5.6 Transient Flow-Rate Tests

Several transient flow-rate tests were run in order to: (1) obtain
startup and coastdown characteristics for fuel- and coolant-pump speeds
and for coolant-salt flow rate; (2) infer fuel-salt flow-rate coastdown
characteristics from the results of (1), since there is no flowmeter in the
fuel loop; and (3) determine the transient effects of fuel flow-rate

changes on reactivity.

52

Figures 22 and 23% show the fuel-pump speed, cooclant-pump speed, and
coolant-salt flow rate vs time for pump startup and coastdown. Data were
taken with the on-line computer and with a Sanborn oscillograph. The
output of a differential-pressure cell across the coolant-salt venturi
was recorded directly on the oscillograph, and the square root of that
signal was taken as flow rate. The lag in the response of the computer
flow signal is due to the response characteristics of the EMF-to-current
converter and the square-root converter between the differential-pressure
transmitter and the computer input.

It was hoped that the coolant-pump speed and coolant flow rate would
coast down in unison so that the fuel flow-rate coastdown could be in-
ferred directly from the fuel-pump speed coastdown curve. This was not
the case, however, and attempts to infer the fuel-flow coastdown transient
were abandoned.

Reactivity effects of fuel flow-rate transients were measured by
letting the flux servo controller hold the reactor critical during the
transients; the reactivity added by the rod is then equal (and opposite)
to the reactivity change due to the flow perturbations. Due to the
absence of voids in the fuel loop during normal operation, this transient

is due entirely to flow effects. Figure 24 shows the response of the

CRNL—DWG 65-8S11A

120

100 i——-—- I R - T - i—ew —0—0=-0
FUEL o-0-00-8-8-0=8~ e-8-¢
PUMP /COOLANT WP e
w SPEED PUMP ol
3 a0 e| ® SPEED _* G)’}
O Pl
< 7 / ¥ /
_ o -
E o [+] / /
&
> . COOLANT FLOW RATE
wl
o
wl
[«
o LOGGER S
® OSCILLOGRAPH

q 5 6 T 8 9 10
TIME (sec}

Fig. 22. Pump Speed and Flow Startup Transients.

PERCENT OF FULL SCALE

53

ORNL-DWG 65—-8942A

T T T T [ T

COOLANT SALT F%OW RATE
| |

— -

: | |
o LOGGER
‘ ® OSCILLOGRAPH ‘

“| - = | 1
I
L* | |
20— """ _COOLANT PUMP SPEED ~'
‘ P ! ! "
FUEL PUMP SPEED- e i
o | | |
0 2 4 6 8 10 12 14 16 18 20
TIME (sec}
Fig. 23. Pump Speed and Flow Coastdown Transients.
ORNL-DWG &7-12323
20
So g
19 S .
° . FUEL PUMP STARTUP
L © e * o'...o l..
’g folele] .0....0..‘. L . '0.0
(o]
S 18 —o *
£
< o _
. [ ]
=
g o
D o K
e
- * %
o Qn
= 16 Q0o L
] o
S): %%
0000000
. 0%0 © FUEL PUMP COASTDOWN
%65°%0% o
15 ——o0-
L 00000 o° 00} 60400% °
. 02%000%60004
Te
®
14
0 10 20 30 40 50 60 70
TIME (sec)

Fig., 24.

Control-Rod Response to Fuel-Pump

Startup and Coastdown.

54

control rod in attempting to maintain the reactor critical during fuel-
pump startup and coastdown. The positive reactivity effect of the re-
circulated precursors entering the core 13 sec after pump startup is

clearly shown.

5.7 Conclusions from Dynamics Tests

The two most significant conclusions obtained from the dynamic tests
were: (1) these tests gave results which show good agreement with theo-
retical predictions, giving increased confidence in the theoretical model
and in the predictions for stable power operation, and (2) the selected
testing procedures were, on the whole, quite satisfactory. Iater tests
performed over the full range of operating power levels have provided

additional evidence in support of these conclusions.2%t

6. GENERAL CONCLUSIONS

This completes our account of the techniques and results of the
program of zero-power physics measurements on the MSRE. Because the
predicted neutronic characteristics were untested by experiment prior
to the beginning of nuclear operation on June 1, 1965, these experiments
provided an interesting test of the ability of then-available calcu-
lational programs and neutron data for the nuclear design predictions of
an unusual new system. The principal interest in the results of these
measurements, however, stems from their applications in interpreting the
neutronic behavior during operation of the reactor at power. The es-
sential conclusions of this program can be summarized as follows:

1. The predicted minimum critical loading of 22U in the clean
fuel salt agreed with the observed 27U loading, within the limits of
uncertainty in the measured fuel-salt density at operating conditions.

2; In comparison to the calculated values of the remaining impor-
tant functional neutronic characteristics of the reactor, the measured
integral reactivity worths of the control rods were within 10%, the 2°°U

concentration coefficient of reactivity was within 5%, and the temperature

55

coefficients of reactivity were within 20% of the calculated quantities.
This provides confidence in pre-operational nuclear-safety studies which
were based on these calculations.

3. The loss of reactivity due to the transport of delayed-neutron
precursors by circulation at first apfieared to be smaller than would be
expected from calculations. ©Subsequent analysis revealed the importance
of including the top and bottom plenums of the reactor core in the theo-
retical calculations of this effect. Although these are regions of low
neutron importance, they are not "out" of the neutron flux, and the
increased fuel-salt volume fraction and relatively long residence times
in these regions both tend to enhance the reactivity contribution of
delayed neutrons emitted in transit through these regions.

k, At normal operating levels of salt in the fuel-pump bowl, and
with the "'clean" fuel-salt conditions throughout the zero-power experi-
ments, no evidence of circulating helium bubbles, entrained in the salt
by the action of the pump-bowl spray-ring apparatus, was detected. In
pressure transient tests at abnormally low pump-bowl levels, both the
control-rod reactivity response and direct densitometer measurements
indicated that 2% to 3% of voids by volume were in circulation with the
fuel salt.

5. Dynamics tests at zero power, which were the initial measure-
ments in a continuing program to assess the MSRE dynamic behavior at
various stages of operation, gave no evidence to support any expectation

of operational stability problems in the reactor system.

lOI

11.

12.

13.

56

REFERENCES

P. N. Haubenreich et al., MSRE Design and Operations Report: Part
ITT. DNuclear Analysis, USAEC Report ORNL-TM-730, Oak Ridge National
Laboratory, February 3%, 196k,

MSR Program Semiann. Progr. Rept. Jan. 31, 1964, USAEC Report ORNL-
%626, p. 54, Oak Ridge National Laboratory.

MSR Program Semiann. Progr. Rept. Aug. 31, 1965, USAEC Report ORNL-
3872, pp. 30—31l, Oak Ridge National Laboratory.

MSR Program Semiann. Progr. Rept. Aug. 31, 1965, USAEC Report ORNL-
3872, pp. 119121, Oak Ridge National Laboratory.

A. F. Henry, "The Application of Reactor Kinetics to the Analysis of
Experiments,™ Nucl. Sci. Fng., 3(1): 52—70 (January 1958).

T. Gozani, "The Concept of Reactivity and Its Application to Kinetics
Measurements," Nukleonik, 5(2): 55 (February 19%63).

T. Auerbach, T. Gozani, and P. Schnid, "Determination of Excess
Reactivity and Control-Rod Worth," Nucl. Sci. Eng., 21(2): 186-193
(February 1965).

P. N. Haubenreich, Predictions of Effective Yields of Delayed
Neutrons in the MSRE, USAEC Report ORNL-TM-380, Oak Ridge National
Laboratory, October 13, 1962,

B. E. Prince, Period Measurements on the MSRE During Fuel Circu-
lation: Theory and Experiment, USAEC Report ORNL-TM-1626, Oak Ridge
National Laboratory, October 19566.

5. J. Ball and R. K. Adams, MATEXP — A General Purpose Digital Com-
puter Program for Solving Ordinary Differential Equations by the
Matrix Exponential Method, USAEC Report ORNL-TM-193%%, August 1967.

MSR Program Semiann. Progr. Rept. Feb. 28, 1966, USAEC Report ORNL-
3936, pp. 32=87, Oak Ridge National Iaboratory.

G. D. Joanou and J. S. Dudek, GAM-II — A B Code for the Calculation
of Fast-Neutron Spectra and Associated Multigroup Constants, GA-#265,
General Atomic, September 1963,

H. C, Honeck, THERMOS — A Thermalization Transport Theory Code for
Reactor Lattice Calculations, USAEC Report BNL-5826, Brookhaven
National Laboratory, September 1961.

1k,

15.

16.

17‘

18.

19.

20.

2l.

22.

23

2k,

25

57

J. Replogle, MODRIC — A One-Dimensional Neutron Diffusion Code for
the IBM-7090, USAEC Report K-1520, Oak Ridge Gaseous Diffusion
Plant, September 6, 1962.

T. B. Fowler and M. L. Tobias, EQUIPOISE-3: A Two-Dimensional, Two-
Group, Neutron-Diffusion Code for the IBM-7090 Computer, USAEC
Report ORNL-%199, Oak Ridge National Laboratory, February 7, 1962,

T. B. Fowler, M. L. Tobias, and D. R. Vondy, EXTERMINATOR: A Multi-
group Code for Solving Neutron Diffusion Equations in One and Two
Dimensions, USAEC Report ORNL-TM-842, February 1965.

MSR Program Semiann. Progr. Rept. Aug. 31, 1961, USAEC Report ORNL-
3215, p. 83, Oak Ridge National Laboratory.

MSR Program Semiann. Progr. Rept. Aug. 31, 1965, USAEC Report ORNL-
3872, pp. 6265, Oak Ridge National Laboratory.

H. A. Samulon, "Spectrum Analysis of Transient Response Curves,"
Proc. I. R. E., 39: 175-186 (February 1951).

5. J. Ball and T. W. Kerlin, Stability Analysis of the Molten-Salt
Reactor Experiment, USAEC Report ORNL-TM-1070, Oak Ridge National
Laboratory, December 1965.

T. W. Kerlin and 5. J. Ball, Experimental Dynamic Analysis of the
Molten-Salt Reactor Experiment, USAEC Report ORNL-TM-1647, Oak Ridge
National Laboratory, October 13, 1966,

T. W. Kerlin, The Pseudorandom Binary Signal for Frequency Response
Testing, USAEC Report ORNL-TM-1662, Oak Ridge National Laboratory,
September 1966.

T. W. Kerlin and J. L. Lucius, CABS — A Fortran Computer Program

for Calculating Correlation Functions, Power Spectra, and the
Frequency Response from Experimental Data, USAEC Report ORNL-TM-1663,
September 1966,

D. N. Fry and D. P. Roux, Oak Ridge National Laboratory, personal
communication, October 1965. These measurements are also reported
in "Neutron Noise, Waves, and Pulse Propagation," ARC Symposium
Series, No. 9, pp. 463474 (May 1967).

C. W. Ricker, 5. H. Hanauer, and E. R. Mann, Measurement of Reactor
Fluctuation Spectra and Subcritical Reactivity, USAEC Report ORNL-
T™-1066, Oak Ridge National Laboratory, April 1965.
O 0= OVWJ1 W o

59

ORNL-4233
UC-80 — Reactor Technology

Internal Distribution

R. K. Adams 6. P, H. Harley

R. G. Affel 47, P. N. Haubenreich
R. F. Apple 48, P. G. Herndon

D. S. Asquith 49, T. L. Hudson

C. F. Baes 50. E. B, Johnson

S. J. Ball 51l. W. H. Jordan

H. F. Bauman 52. S. I. Kaplan

5. E. Beall 53. P. R. Kasten

M. Bender 54. R. J. Kedl

E. 5. Bettis 55. M, T. Kelley

F. F. Blankenship 5%6. T. W. Kerlin

R. Blumberg 57. S. 5. Kirslis

E. G. Bohlmann 58. A. I. Krakoviak
C. J. Borkowski 59. dJ. W. Krewson

G. E. Boyd 60. R. C. Kryter

R. B. Briggs 61. J. A. Lane

R. H. Bryan 62. C. E. Larson

J. M. Chandler 6%. R. B. Lindauer
R. H. Chapman 6k. M. I. Lundin

F. H. Clark 65. R. N. Lyon

C. E. Clifford 66. H. G. MacPhnerson
W. R. Cobb 67. R. E. MacPherson
W. H. Cook 68. C. D. Martin

L. T. Corbin 69. T. H. Mauney

W. B. Cottrell 70. H. E. McCoy

J. L. Crowley 71, H. C. McCurdy

F. L. Culler, Jr. 2. H. . McDuffie
D. G. Davis 73. C. K. McGlothlan
R. J. DeBakker Th. A, J. Miller

5. J. Ditto >. R. L. Moore

F. A. Doss 76. E. L. Nicholson
W. P. BEatherly 77. L. C. Oakes

J. R. Engel 78. A. M. Perry

E. P. Epler 79. H. B. Piper

D. E. Ferguson 80-84, B. E. Prince

A. P, Fraas 85. G. L. Ragan

D. N. Fry %. J. L. Redford

J. H. Frye, Jr. 87. M. Richardson

H. A. Friedman 88. R. C. Robertson
C. H. Gabbard 89. J. C. Robinson
R. B. Gallaher 90-114., M. W. Rosenthal
W. R. Grimes 115. A. W. Savolainen
A. G. Grindell 116. D. Scott, Jr.

E. D. Gupton 117. J. H. Shaffer

R. H. Guymon 118. E. G. Silver
60

119, M. J. Skinner 132. A, M, Weinberg
120, A. N. Smith 133, J. R. Weir
121, 0. L. Smith 134, K. W. West
122. P. G. Smith 135, M. E. Whatley
123, I. Spiewak 1%6. J. C. White
124. D. A. Sundberg 137. G. D. Whitman
125. R. C. Steffy, Jr. 138. Gale Young
126. H. H. Stone 1%39-141. Central Research Library
127. J. R. Tallackson 142-143, Y-12 Document Reference
128, R. E. Thoma Section
129. D. B. Trauger 144-209., Laboratory Records Department
130. C. 5. Walker 210. Laboratory Records Department,
131. B. H. Webster ORNL-RC
External Distribution
2l11. C. B. Deering, USAEC, ORO, Oak Ridge
212-213, K. E. Elliott, USAEC, ORO, Oak Ridge
214, H. M. Roth, USAEC, ORO, Oak Ridge
215, J. W. Lewellyn, USAEC Division of Reactor Development and

Technology, Washlngton, D. C.
216-22%3, T. W. McIntosh, USAEC, Division of Reactor Development and
Technology, Washington, D. C.
224, E. E. Purvis, USAEC, Division of Reactor Development and
Technology, Washington, D. C.
225. T. G. Schleiter, USAEC, Division of Reactor Development and
Technology, Washington, D. C.
226, M. Shaw, USAEC, Division of Reactor Development and Technology,
Washington, D. C.
227. Laboratory and University Division, USAEC, ORO, Oak Ridge
228-L48L. Given distribution as shown in TID-4500 under Reactor Technology
category (25 copies — CFST1)